CLINICAL GUIDELINESON THEIDENTIFICATION,EVALUATION, ANDTREATMENT OFOVERWEIGHT ANDOBESITY IN ADULTSThe Evidence ReportN A T I O N A L I N S T I T U T E S O F H E A L T HN A T I O N A L H E A R T , L U N G , A N D B L O O D I N S T I T U T EO b e s i t y E d u c a t i o n I n i t i a t i v e
CLINICAL GUIDELINESON THEIDENTIFICATION,EVALUATION, ANDTREATMENT OFOVERWEIGHT ANDOBESITY IN ADULTSThe Evidence ReportNIH PUBLICATIONNO. 98-4083SEPTEMBER 1998NATIONAL INSTITUTESOF HEALTHNational Heart, Lung,and Blood Institute incooperation withThe National Institute ofDiabetes and Digestive andKidney Diseases
iiNHLBI Obesity Education InitiativeExpert Panel on the Identification,Evaluation, and Treatment ofOverweight and Obesity in AdultsF. Xavier Pi-Sunyer, M.D., M.P.H.Chair of the PanelChief, Endocrinology, Diabetes, and NutritionDirector, Obesity Research CenterSt. Lukes/Roosevelt Hospital CenterProfessor of MedicineColumbia University College of Physicians andSurgeonsNew York, NYDiane M. Becker, Sc.D., M.P.H.DirectorCenter for Health PromotionAssociate ProfessorDepartment of MedicineThe Johns Hopkins UniversityBaltimore, MDClaude Bouchard, Ph.D.Professor of Exercise PhysiologyPhysical Activity SciencesLaboratoryLaval UniversitySainte Foy, QuebecCANADARichard A. Carleton, M.D.Professor of MedicineBrown University School of MedicinePawtucket, RIGraham A. Colditz, M.D., Dr.P.H.Associate Professor of MedicineHarvard Medical SchoolChanning LaboratoryBoston, MAWilliam H. Dietz, M.D., Ph.D.DirectorDivision of Nutrition and Physical ActivityNational Center for Chronic Disease Preventionand Health PromotionCenters for Disease Control and PreventionAtlanta, GAJohn P. Foreyt, Ph.D.Professor of Medicine and DirectorNutrition Research ClinicBaylor College of MedicineHouston, TXRobert J. Garrison, Ph.D.Associate ProfessorDepartment of Preventive MedicineUniversity of Tennessee, MemphisMemphis, TNScott M. Grundy, M.D., Ph.D.DirectorCenter for Human NutritionUniversity of TexasSouthwestern Medical Center at DallasDallas, TXBarbara C. Hansen, Ph.D.Professor of PhysiologyDirector of Obesity and Diabetes Research CenterUniversity of Maryland School of MedicineBaltimore, MDMillicent Higgins, M.D.Department of EpidemiologySchool of Public HealthUniversity of MichiganAnn Arbor, MI
iiiJames O. Hill, Ph.D.Associate Director of ResearchCenter for Human NutritionUniversity of Colorado Health Sciences CenterDenver, COBarbara V. Howard, Ph.D.PresidentMedlantic Research InstituteWashington, DCRobert C. Klesges, Ph.D.ProfessorUniversity of Memphis Prevention CenterUniversity of MemphisMemphis, TNRobert J. Kuczmarski, Dr.P.H., R.D.Nutrition AnalystNational Center for Health StatisticsCenters for Disease Control and PreventionHyattsville, MDShiriki Kumanyika, Ph.D., R.D., M.P.H.Professor and HeadDepartment of Human Nutrition and DieteticsThe University of Illinois at ChicagoChicago, ILR. Dee Legako, M.D.Prime Care Canyon Park Family Physicians, Inc.Edmond, OKT. Elaine Prewitt, Dr.P.H., R.D.Assistant ProfessorDepartment of Preventive Medicine andEpidemiologyLoyola University Medical CenterMaywood, ILAlbert P. Rocchini, M.D.Chief of CardiologyUniversity of Michigan Medical CenterAnn Arbor, MIPhilip L. Smith, M.D.Professor of MedicineDivision of Pulmonary and Critical CareMedicineThe Johns Hopkins Asthma and Allergy CenterBaltimore, MDLinda G. Snetselaar, Ph.D., R.D.Associate ProfessorHead of Preventive Nutrition EducationDepartment of Preventive MedicineUniversity of IowaIowa City, IAJames R. Sowers, M.D.Professor of Medicine and PhysiologyDirectorDivision of Endocrinology, Metabolism, andHypertensionWayne State University School of MedicineUniversity Health CenterDetroit, MIMichael Weintraub, M.D.DirectorOffice of Drug Evaluation VFood and Drug AdministrationRockville, MDDavid F. Williamson, Ph.D., M.S.EpidemiologistDivision of Diabetes TranslationCenters for Disease Control and PreventionChamblee, GA
ivG. Terence Wilson, Ph.D.Oscar K. Buros Professor of PsychologyDirector, Rutgers Eating Disorders ClinicPiscataway, NJEx-Officio MembersClarice D. Brown, M.S.Project ManagerCODA Research, Inc.Silver Spring, MDKaren A. Donato, M.S., R.D.*Executive Director of the PanelCoordinatorNHLBI Obesity Education InitiativeNational Heart, Lung, and Blood InstituteNational Institutes of HealthBethesda, MDNancy Ernst, Ph.D., R.D.*Nutrition CoordinatorOffice of the DirectorDivision of Epidemiology and Clinical ApplicationsNational Heart, Lung, and Blood InstituteNational Institutes of HealthBethesda, MDD. Robin Hill, Ph.D.*Social Science AnalystBehavioral Medicine BranchDivision of Epidemiology and ClinicalApplicationsNational Heart, Lung, and Blood InstituteNational Institutes of HealthBethesda, MDMichael J. Horan, M.D., Sc.M.*DirectorDivision of Heart and Vascular DiseasesNational Heart, Lung, and Blood InstituteNational Institutes of HealthBethesda, MDVan S. Hubbard, M.D., Ph.D.Director, NIH Division of Nutrition ResearchCoordinationChief, Nutritional Sciences BranchNational Institute of Diabetes and Digestive andKidney DiseasesNational Institutes of HealthBethesda, MDJames P. Kiley, Ph.D.*DirectorAirway Biology and Disease ProgramDivision of Lung DiseasesNational Heart, Lung, and Blood InstituteNational Institutes of HealthBethesda, MDEva Obarzanek, Ph.D., R.D., M.P.H.*Research NutritionistPrevention Scientific Research GroupDivision of Epidemiology and ClinicalApplicationsNational Heart, Lung, and Blood InstituteNational Institutes of HealthBethesda, MDConsultantDavid Schriger, M.D., M.P.H., F.A.C.E.P.Associate ProfessorUCLA Emergency Medicine CenterUniversity of California at Los AngelesSchool of MedicineLos Angeles, CASan Antonio Cochrane CenterElaine Chiquette, Pharm.D.V.A. Cochrane Center at San AntonioAudie L. Murphy Memorial Veterans HospitalSan Antonio, TX* NHLBI Obesity Education Initiative Task Force Members
vCynthia Mulrow, M.D., M.Sc.V.A. Cochrane Center at San AntonioAudie L. Murphy Memorial Veterans HospitalSan Antonio, TXStaffAdrienne Blount, B.S.R.O.W. Sciences, Inc.Rockville, MDMaureen Harris, M.S., R.D.R.O.W. Sciences, Inc.Rockville, MDAnna Hodgson, M.A.R.O.W. Sciences, Inc.Rockville, MDPat Moriarty, M.Ed., R.D.R.O.W. Sciences, Inc.Rockville, MDThe panel acknowledges the assistance of Dr.Rashid Chotani, Johns Hopkins University; Dr.Robert Klesges, University of Memphis; Dr.Walter Pories, East Carolina University; Dr. IvanBaines, NHLBI; Dr. Christine Kelly, NHLBI;Glen Bennett, NHLBI; Dr. Fred Heydrick,BioReview; Debbie Lurie, Prospect Associates;Estelle Schwalb, Prospect Associates; LoriMcCray, R.O.W. Sciences, Inc.; andNiyati Pandya, R.O.W. Sciences, Inc.
viContentsForeword viiExecutive Summary xi1. Introduction 1A. Rationale for Guidelines Development 1B. Objectives of the Guidelines 2C. Guideline Development Methodology 2D. Statement of Assumptions 5E. Intended Users of These Guidelines 52. Overweight and Obesity: Background 6A. Health and Economic Costs 61. Prevalence and Time Trends 62. Demographic Variations inOverweight and Obesity Prevalence 93. Economic Costs of Overweightand Obesity 9B. Prevention of Overweight and Obesity 11C. Health Risks of Overweight and Obesity 121. Morbidity 122. Overweight/Obesity and Morbidityin Minority Populations 233. Obesity and Mortality 23D. Weight Loss and Mortality 25E. Environment 26F. Genetic Influence in the Developmentof Overweight and Obesity 273 Examination of Randomized ControlledTrial Evidence 29A. Why Treat Overweight and Obesity? 291. Blood Pressure 292. Serum/Plasma Lipids andLipoproteins 333. Impaired Glucose Toleranceand Diabetes 394. Decreases in Abdominal Fat withWeight Loss 41B. What Treatments are Effective? 421. Dietary Therapy 422. Physical Activity 443. Combined Therapy 474. Behavior Therapy 485. Pharmacotherapy 536. Surgery 547. Other Interventions for Overweightand Obesity Treatment 554. Treatment Guidelines 56A. Overview 56B. Assessment and Classification ofOverweight and Obesity 561. Assessment of Overweight andObesity 562. Classification of Overweight andObesity 58C. Assessment of Risk Status 62D. Evaluation and Treatment Strategy 65E. Exclusion from Weight Loss Therapy 70F. Patient Motivation 70G. Goals of Weight Loss andManagement 711. Weight Loss 712. Weight Maintenance atLower Weight 723. Prevention of Further Weight Gain 73H. Strategies for Weight Loss andWeight Maintenance 731. Dietary Therapy 732. Physical Activity 773. Behavior Therapy 814. Combined Therapy 835. Pharmacotherapy 836. Surgery 86I. Smoking Cessation in the Overweightor Obese Patient 91J. Role of Health Professionals inWeight Loss Therapy 925. Summary of Recommendations 956. Future Research 98A. Intervention Approaches 98B. Causes and Mechanisms ofOverweight and Obesity 99C. Abdominal Fat, Body Weight andDisease Risk 99D. Assessment Methods 997. Appendices 100Appendix I.A.1 Guidelines DevelopmentMethodology 100Appendix I.A.2 Literature Review 109Appendix II Description of Evidence 112Appendix III Special Populations 117Appendix IV Obesity and Sleep Apnea 137Appendix V Body Mass Index Chart—How to Measure Obesity 139Appendix VI Practical Dietary TherapyInformation 141Appendix VII Resource List 165Appendix VIII Glossary of Terms 168List of Abbreviations 179Reference List 181Publication List 226
viiFOREWORDIn 1995, the National Obesity EducationInitiative of the National Heart, Lung, andBlood Institute (NHLBI), in cooperation withthe National Institute of Diabetes and Digestiveand Kidney Diseases (NIDDK), convened thefirst Expert Panel on the Identification,Evaluation, and Treatment of Overweight andObesity in Adults to develop clinical practiceguidelines for primary care practitioners.The impetus for these guidelines was the recog-nition that the prevalence of overweight andobesity in the United States is increasing, andthat practitioners need to be alerted to theaccompanying health risks. According to the lat-est statistics from the third National Health andNutrition Examination Survey, 97 millionAmericans are overweight or obese. Excessweight is often accompanied by high blood pres-sure, high blood cholesterol, type 2 diabetes,coronary heart disease, and other health prob-lems. The total costs attributable to obesity-relat-ed disease approach $100 billion annually in theUnited States.The panel used the principles of evidence-basedmedicine, including an evidence model and evi-dence categories. It was the first time a panelthoroughly examined the scientific evidence forrisks associated with overweight and obesity, andtheir treatments, and developed clinical practicerecommendations based on their conclusions.The panel was headed by Dr. F. Xavier Pi-Sunyer, of St. Luke’s/Roosevelt Hospital Centerin New York City. He and the other 23 panelmembers methodically and critically examined avast amount of published scientific evidence.The panel also obtained scientific input fromapproximately 115 outside reviewers. The resultwas The Clinical Guidelines on the Identification,Evaluation, and Treatment of Overweight andObesity in Adults: The Evidence Report. Dr.David Schriger of the University of California atLos Angeles, who is a methodologist consultantin the area of evidence-based practice guidelines,and Dr. Elaine Chiquette of the San AntonioCochrane Center in Texas, who developed theevidence tables, served as key advisors to thepanel.Evidence examined by the panel included:research on the relationship of overweight andobesity to high blood pressure, high blood cho-lesterol, type 2 diabetes, stroke, congestive heartfailure, coronary heart disease, various cancers,osteoarthritis, and sleep apnea; risks associatedwith the distribution and amount of body fat;and various treatment strategies, including diet,physical activity, behavior therapy, pharma-cotherapy, and surgery. The resulting guidelinesdocument how different treatment strategiesaffect weight loss and how weight control affectsthe major risk factors for heart disease andstroke.The guidelines have been reviewed and endorsedby many professional organizations. In fact,because of the associated risks between highblood pressure and high blood cholesterol andoverweight and obesity, the document represents
viiithe first clinical practice guidelines to bereviewed and endorsed by members of the coor-dinating committees of both the NationalCholesterol Education Program and theNational High Blood Pressure EducationProgram which comprise approximately 52 pro-fessional societies, government agencies, andconsumer organizations. Two additional groupsendorsing the guidelines are the North AmericanAssociation for the Study of Obesity and theNIDDK National Task force on the Preventionand Treatment of Obesity.The report, the evidence model and its accom-panying evidence tables, and a body mass indexcalculator are available on the NHLBI website atthe following location: http://www.nhlbi.nih.gov./nhlbi/cardio/obes/prof/guidelns/ob_home.htmAn abbreviated version of the evidence report isbeing prepared and will be distributed to prima-ry care physicians, nurses, registered dietitiansand nutritionists, and other health care practi-tioners.It is our hope that these clinical guidelines willnot only help the health care practitioner under-stand the importance of weight management butalso provide them with the tools to assess andtreat their patients more effectively. Equallyimportant, we hope that the guidelines lead to agreater public understanding of obesity and agreater appreciation for the persistent efforts ofmillions of people to lose weight.Claude Lenfant, M.D.DirectorNational Heart, Lung,and Blood Institute
EVIDENCE REPORT ENDORSEMENTSixNATIONAL CHOLESTEROL EDUCATION PROGRAM (NCEP)Coordinating Committee Member OrganizationsTheodore G. Ganiats, M.D., American Academy of Family Physicians, Gary Graham, M.D., American Academy of InsuranceMedicine, Ronald Kleinman, M.D., American Academy of Pediatrics, Ann Smith, R.N., C.O.H.N.-S., American Association ofOccupational Health Nurses, Richard C. Pasternak, M.D., F.A.C.C., American College of Cardiology, Gerald T. Gau, M.D., AmericanCollege of Chest Physicians, Ruth Ann Jordan, M.D., American College of Occupational and Environmental Medicine, Thomas E.Nolan, M.D., American College of Obstetricians and Gynecologists, Alan J. Garber, M.D., Ph.D., American Diabetes Association, Inc.,Linda Van Horn, Ph.D., R.D., American Dietetic Association, Scott Grundy, M.D., Ph.D., American Heart Association, SandraCornett, R.N., Ph.D., American Hospital Association, Yank D. Coble, Jr., M.D., American Medical Association, Joan Watson, R.N.,Ph.D., F.A.A.N., American Nurses’ Association, Michael Clearfield, D.O., American Osteopathic Association, Joanne Mitten, M.H.E.,Association of State and Territorial Health Officials, Gerald J. Wilson, M.B.A., Citizens for Public Action on Blood Pressure andCholesterol, Inc., Linda Burnes-Bolton, Dr.P.H., R.N., National Black Nurses’ Association, Inc., Luther T. Clark, M.D., NationalMedical Association, Darlene Lansing, M.P.H., R.D., Society for Nutrition EducationAssociate Member Coordinating Committee OrganizationsStanley Wallach, M.D., American College of Nutrition, Donald O. Fedder, Dr.P.H., M.P.H., Society for Public Health EducationFederal Agency Liaison Coordinating Committee RepresentativesYvonne L. Bronner, Sc.D., R.D., L.D., NHLBI Ad Hoc Committee on Minority Populations, Francis D. Chesley, M.D., Agency forHealth Care Policy and Research, William H. Dietz, M.D., Ph.D., Centers for Disease Control and Prevention, Thomas M. Lasater,Ph.D., Coordinating Committee for the Community Demonstration Studies, Alanna Moshfegh, M.S., R.D., Department of Agriculture,Col. Michael Parkinson, M.D., M.P.H., Department of Defense, Pamela Steele, M.D., Department of Veterans Affairs, Celia Hayes,M.P.H., R.D., Health Resources and Services Administration, Clifford Johnson, M.P.H., National Center for Health Statistics, LindaMeyers, Ph.D., Office of Disease Prevention and Health PromotionNATIONAL HIGH BLOOD PRESSURE EDUCATION PROGRAM (NHBPEP)Coordinating Committee Member OrganizationsLee A. Green, M.D., M.P.H., American Academy of Family Physicians, Jack P. Whisnant, M.D., American Academy of Neurology,Barry N. Hyman, M.D., F.A.C.P., American Academy of Ophthalmology, Lisa Mustone-Alexander, M.P.H., P.A., American Academyof Physician Assistants, Henry Guevara, B.S.N., R.N., C.O.H.N.-S., American Association of Occupational Health Nurses, Edward D.Frohlich, M.D., American College of Cardiology, Sheldon G. Sheps, M.D., American College of Chest Physicians, Ron Stout, M.D.,American College of Occupational and Environmental Medicine, Jerome D. Cohen, M.D., American College of Physicians, CarlosVallbona, M.D., American College of Preventive Medicine, James R. Sowers, M.D., American Diabetes Association, Inc., Mary C.Winston, Ed.D., R.D., American Dietetic Association, Daniel W. Jones, M.D., American Heart Association, Roxane Spitzer, Ph.D.,F.A.A.N., American Hospital Association, Nancy Houston Miller, B.S.N., American Nurses’ Association, Linda Casser, O.D., AmericanOptometric Association, William A. Nickey, D.O., American Osteopathic Association, Raymond W. Roberts, Pharm.D., AmericanPharmaceutical Association, Pamela J. Colman, D.P.M., American Podiatric Medical Association, Nancy McKelvey, M.S.N., R.N.,American Red Cross, Barry L. Carter, Pharm.D., F.C.C.P., American Society of Health-System Pharmacists, Norman M. Kaplan, M.D.,American Society of Hypertension, Jackson T. Wright, M.D., Ph.D., Association of Black Cardiologists, Gerald J. Wilson, M.B.A.,Citizens for Public Action on Blood Pressure and Cholesterol, Inc., Joseph L. Izzo, Jr., M.D., Council on Geriatric Cardiology, James W.Reed, M.D., F.A.C.P., F.A.C.E., International Society on Hypertension in Blacks, Rita Strickland, Ed.D., R.N., National Black Nurses’Association, Inc., William Manger, M.D., Ph.D., National Hypertension Association, Inc., Murray Epstein, M.D., National KidneyFoundation, Inc., Otelio S. Randall, M.D., F.A.C.C., National Medical Association, Edwin Marshall, O.D., M.P.H., NationalOptometric Association, Harold W. “Pete” Todd, National Stroke Association, Kathryn M. Kolasa, Ph.D., R.D., L.D.N., Society forNutrition EducationFederal Agency Liaison Coordinating Committee RepresentativesKeith Ferdinand, M.D., F.A.C.C., NHLBI Ad Hoc Committee on Minority Populations, Francis D. Chesley, M.D., Agency for HealthCare Policy and Research, H. Mitchell Perry, Jr., M.D., Department of Veterans Affairs, Jay Merchant, M.H.A., Health Care FinancingAdministration, Vicki Burt, R.N., Sc.M., National Center for Health Statistics, Elizabeth H. Singer, M.S., National Institute ofDiabetes and Digestive and Kidney Diseases (NIDDK)NIDDK NATIONAL TASK FORCE ON THE PREVENTION AND TREATMENT OF OBESITYNORTH AMERICAN ASSOCIATION FOR THE STUDY OF OBESITY (NAASO)
xiIntroductionAn estimated 97 million adults in the UnitedStates are overweight or obese, a condition thatsubstantially raises their risk of morbidity fromhypertension, dyslipidemia, type 2 diabetes,coronary heart disease, stroke, gallbladder dis-ease, osteoarthritis, sleep apnea and respiratoryproblems, and endometrial, breast, prostate, andcolon cancers. Higher body weights are also asso-ciated with increases in all-cause mortality. Obeseindividuals may also suffer from social stigmati-zation and discrimination. As a major contributorto preventive death in the United States today,overweight and obesity pose a major publichealth challenge.Overweight is here defined as a body mass index(BMI) of 25 to 29.9 kg/m2and obesity as a BMIof ≥ 30 kg/m2. However, overweight and obesityare not mutually exclusive, since obese personsare also overweight. A BMI of 30 is about 30 lboverweight and equivalent to 221 lb in a 60"person and to 186 lb in one 56". The numberof overweight and obese men and women hasrisen since 1960; in the last decade the percent-age of people in these categories has increased to54.9 percent of adults age 20 years or older.Overweight and obesity are especially evident insome minority groups, as well as in those withlower incomes and less education.Obesity is a complex multifactorial chronic dis-ease that develops from an interaction of geno-type and the environment. Our understanding ofhow and why obesity develops is incomplete, butinvolves the integration of social, behavioral, cul-tural, physiological, metabolic and genetic factors.While there is agreement about the health risksof overweight and obesity, there is less agreementabout their management. Some have arguedagainst treating obesity because of the difficultyin maintaining long-term weight loss and ofpotentially negative consequences of the fre-quently seen pattern of weight cycling in obesesubjects. Others argue that the potential hazardsof treatment do not outweigh the known hazardsof being obese. The intent of these guidelines isto provide evidence for the effects of treatmenton overweight and obesity. The guidelines focuson the role of the primary care practitioner intreating overweight and obesity.Evidence-Based GuidelinesTo evaluate published information and to deter-mine the most appropriate treatment strategiesthat would constitute evidence-based clinicalguidelines on overweight and obesity for physi-cians and associated health professionals in clini-cal practice, health care policy makers, and clini-cal investigators, the National Heart, Lung, andBlood Institute’s Obesity Education Initiative incooperation with the National Institute ofDiabetes and Digestive and Kidney Diseases con-vened the Expert Panel on the Identification,Evaluation, and Treatment of Overweight andObesity in Adults in May 1995. The guidelinesare based on a systematic review of the publishedscientific literature found in MEDLINE fromJanuary 1980 to September 1997 of topics iden-tified by the panel as key to extrapolating thedata related to the obesity evidence model.EXECUTIVE SUMMARY
xiiEvidence from approximately 394 randomizedcontrolled trials (RCTs) was considered by thepanel.The panel is comprised of 24 members, 8 ex-officio members, and a methodologist consul-tant. Areas of expertise contributed to by panelmembers included primary care, epidemiology,clinical nutrition, exercise physiology, psycholo-gy, physiology, and pulmonary disease. Therewere five meetings of the full panel and twoadditional meetings of the executive committeecomprised of the panel chair and four panelmembers.The San Antonio Cochrane Center assisted thepanel in the literature abstraction and in orga-nizing the data into appropriate evidence tables.The center pretested and used a standardized25-page form or “Critical Review Status Sheet”for the literature abstraction. Ultimately, 236RCT articles were abstracted and the data werethen compiled into individual evidence tablesdeveloped for each RCT. The data from theseRCTs served as the basis for many of the recom-mendations contained in the guidelines.The panel determined the criteria for decidingon the appropriateness of an article. At a mini-mum, studies had to have a time frame fromstart to finish of at least 4 months. The onlyexceptions were a few 3-month studies related todietary therapy and pharmacotherapy. To con-sider the question of long-term maintenance,studies with outcome data provided at approxi-mately 1 year or longer were examined.Excluded were studies in which self-reportedweights by subjects were the only indicators usedto measure weight loss. No exclusions of studieswere made by study size. The panel weighed theevidence based on a thorough examination ofthe threshold or magnitude of the treatmenteffect. Each evidence statement (other thanthose with no available evidence) and each rec-ommendation is categorized by a level of evi-dence which ranges from A to D. Table ES-1summarizes the categories of evidence by theirsource and provides a definition for each category.s Who is at Risk? All overweight and obeseadults (age 18 years of age or older) with aBMI of ≥25 are considered at risk for devel-oping associated morbidities or diseases suchas hypertension, high blood cholesteral, type2 diabetes, coronary heart disease, and otherdiseases. Individuals with a BMI of 25 to29.9 are considered overweight, while individ-uals with a BMI ≥30 are considered obese.Treatment of overweight is recommendedonly when patients have two or more risk fac-tors or a high waist circumference. It shouldfocus on altering dietary and physical activitypatterns to prevent development of obesityand to produce moderate weight loss.Treatment of obesity should focus on produc-ing substantial weight loss over a prolongedperiod. The presence of comorbidities in over-weight and obese patients should be consid-ered when deciding on treatment options.s Why Treat Overweight and Obesity? Obesity isclearly associated with increased morbidityand mortality. There is strong evidence thatweight loss in overweight and obese individu-als reduces risk factors for diabetes and car-diovascular disease (CVD). Strong evidenceexists that weight loss reduces blood pressurein both overweight hypertensive and nonhy-pertensive individuals; reduces serum triglyc-erides and increases high-density lipoprotein(HDL)-cholesterol; and generally producessome reduction in total serum cholesterol andlow-density lipoprotein (LDL)-cholesterol.Weight loss reduces blood glucose levels inoverweight and obese persons without dia-betes; and weight loss also reduces blood glu-cose levels and HbA1c in some patients withtype 2 diabetes. Although there have been noprospective trials to show changes in mortalitywith weight loss in obese patients, reductionsin risk factors would suggest that develop-
xiiiEVIDENCE CATEGORIESEvidenceCategoryASources ofEvidenceRandomizedcontrolled trials(rich body ofdata)DefinitionEvidence is from endpoints of well-designed RCTs (or trialsthat depart only minimally from randomization) that providea consistent pattern of findings in the population for whichthe recommendation is made. Category A therefore requiressubstantial numbers of studies involving substantial numbersof participants.B Randomizedcontrolled trials(limited bodyof data)Evidence is from endpoints of intervention studies thatinclude only a limited number of RCTs, post hoc or subgroupanalysis of RCTs, or meta-analysis of RCTs. In general,Category B pertains when few randomized trials exist, they aresmall in size, and the trial results are somewhat inconsistent, orthe trials were undertaken in a population that differs from thetarget population of the recommendation.C NonrandomizedtrialsObservationalstudiesEvidence is from outcomes of uncontrolled or nonrandomizedtrials or from observational studies.D PanelConsensusJudgmentExpert judgment is based on the panel’s synthesis of evidencefrom experimental research described in the literature and/orderived from the consensus of panel members based on clinicalexperience or knowledge that does not meet the above-listedcriteria. This category is used only in cases where the provisionof some guidance was deemed valuable but an adequatelycompelling clinical literature addressing the subject of the rec-ommendation was deemed insufficient to justify placement inone of the other categories (A through C).Table ES-1:
xivment of type 2 diabetes and CVD would bereduced with weight loss.s What Treatments Are Effective? A variety ofeffective options exist for the management ofoverweight and obese patients, includingdietary therapy approaches such as low-caloriediets and lower-fat diets; altering physicalactivity patterns; behavior therapy techniques;pharmacotherapy*; surgery; and combinationsof these techniques.Clinical GuidelinesTreatment of the overweight or obese patient is atwo-step process: assessment and treatment man-agement. Assessment requires determination ofthe degree of overweight and overall risk status.Management includes both reducing excess bodyweight and instituting other measures to controlaccompanying risk factors.Assessment: When assessing a patient for riskstatus and as a candidate for weight loss therapy,consider the patient’s BMI, waist circumference,and overall risk status. Consideration also needsto be given to the patient’s motivation to loseweight.s Body Mass Index. The BMI, which describesrelative weight for height, is significantly cor-related with total body fat content. The BMIshould be used to assess overweight and obesi-ty and to monitor changes in body weight. Inaddition, measurements of body weight alonecan be used to determine efficacy of weightloss therapy. BMI is calculated as weight(kg)/height squared (m2). To estimate BMIusing pounds and inches, use: [weight(pounds)/height (inches)2] x 703. Weight clas-sifications by BMI, selected for use in thisreport, are shown in Table ES-2. A conversiontable of heights and weights resulting inselected BMI units is provided in Table ES-3.s Waist Circumference. The presence of excessfat in the abdomen out of proportion to totalbody fat is an independent predictor of riskfactors and morbidity. Waist circumference ispositively correlated with abdominal fat con-tent. It provides a clinically acceptable mea-surement for assessing a patients abdominalfat content before and during weight losstreatment. The sex-specific cutoffs noted onthe next page can be used to identifyTABLE ES-2:CLASSIFICATION OF OVERWEIGHT AND OBESITY BY BMIObesity Class BMI (kg/m2)Underweight < 18.5Normal 18.5 – 24.9Overweight 25.0 – 29.9Obesity I 30.0 – 34.9II 35.0 – 39.9Extreme Obesity III ≥ 40* As of September 1997, the Food and Drug Administration (FDA) requested the voluntary withdrawal from the market of dexfenfluramine and fenfluramine dueto a reported association between valvular heart disease and the use of dexfenfluramine or fenfluramine alone or combined with phentermine. The use of thesedrugs for weight reduction, therefore, is not recommended in this report. Sibutramine is approved by FDA for long-term use. It has limited but definite effects onweight loss and can facilitate weight loss maintenance (Note: FDA approval for orlistat is pending a resolution of labeling issues and results of Phase III trials.)
xvincreased relative risk for the development ofobesity-associated risk factors in most adultswith a BMI of 25 to 34.9 kg/m2:These waist circumference cutpoints lose theirincremental predictive power in patients witha BMI ≥ 35 kg/m2because these patients willexceed the cutpoints noted above. Table ES-4adds the disease risk of increased abdominalfat to the disease risk of BMI. These cate-gories denote relative risk, not absolute risk;that is, relative to risk at normal weight. Theyshould not be equated with absolute risk,which is determined by a summation of riskfactors. They relate to the need to instituteweight loss therapy and do not directly definethe required intensity of modification of riskfactors associated with obesity.s Risk Status. Assessment of a patient’s absoluterisk status requires examination for the pres-ence of:Disease conditions: established coronary heartdisease (CHD), other atherosclerotic diseases,type 2 diabetes, and sleep apnea; patients withthese conditions are classified as being at veryhigh risk for disease complications and mor-tality.Other obesity-associated diseases: gynecologi-cal abnormalities, osteoarthritis, gallstonesand their complications, and stress inconti-nence.Cardiovascular risk factors: cigarette smoking,hypertension (systolic blood pressure ≥ 140mm Hg or diastolic blood pressure ≥ 90 mmHg, or the patient is taking antihypertensiveagents), high-risk LDL-cholesterol (≥ 160mg/dL), low HDL-cholesterol (< 35 mg/dL),impaired fasting glucose (fasting plasma glu-cose of 110 to 125 mg/dL), family history ofpremature CHD (definite myocardial infarc-tion or sudden death at or before 55 years ofage in father or other male first-degree rela-tive, or at or before 65 years of age in motheror other female first-degree relative), and age(men ≥ 45 years and women ≥ 55 years orpostmenopausal). Patients can be classified asbeing at high absolute risk if they have threeof the aforementioned risk factors. Patients athigh absolute risk usually require clinicalmanagement of risk factors to reduce risk.Patients who are overweight or obese oftenhave other cardiovascular risk factors.Methods for estimating absolute risk status fordeveloping cardiovascular disease based onthese risk factors are described in detail in theNational Cholesterol Education Program’sSecond Report of the Expert Panel on theDetection, Evaluation, and Treatment of HighBlood Cholesterol in Adults (NCEP’s ATP II)and the Sixth Report of the Joint NationalCommittee on Prevention, Detection,Evaluation, and Treatment of High BloodPressure (JNC VI). The intensity of interven-tion for cholesterol disorders or hypertensionis adjusted according to the absolute risk sta-tus estimated from multiple risk correlates.These include both the risk factors listedabove and evidence of end-organ damage pre-sent in hypertensive patients. Approaches totherapy for cholesterol disorders and hyper-tension are described in ATP II and JNC VI,respectively. In overweight patients, control ofcardiovascular risk factors deserves equalemphasis as weight reduction therapy.Reduction of risk factors will reduce the riskfor cardiovascular disease whether or notefforts at weight loss are successful.Other risk factors: physical inactivity andhigh serum triglycerides (> 200 mg/dL).When these factors are present, patients canH I G H R I S KMen > 102 cm ( > 40 in)Women > 88 cm ( > 35 in)
xviTable ES-3:SELECTED BMI UNITS CATEGORIZED BY INCHES (CM) AND POUNDS (KG)Body weight in pounds (kg)Height in inches (cm) BMI 25 kg/m2BMI 27 kg/m2BMI 30 kg/m258 (147.32) 119 (53.98) 129 (58.51) 143 (64.86)59 (149.86) 124 (56.25) 133 (60.33) 148 (67.13)60 (152.40) 128 (58.06) 138 (62.60) 153 (69.40)61 (154.94) 132 (59.87) 143 (64.86) 158 (71.67)62 (157.48) 136 (61.69) 147 (66.68) 164 (74.39)63(160.02) 141 (63.96) 152 (68.95) 169 (76.66)64 (162.56) 145 (65.77) 157 (71.21) 174 (78.93)65 (165.10) 150 (68.04) 162 (73.48) 180 (81.65)66 (167.64) 155 (70.31) 167 (75.75) 186 (84.37)67 (170.18) 159 (72.12) 172 (78.02) 191 (86.64)68 (172.72) 164 (74.39) 177 (80.29) 197 (89.36)69 (175.26) 169 (76.66) 182 (82.56) 203 (92.08)70 (177.80) 174 (78.93) 188 (85.28) 207 (93.89)71 (180.34) 179 (81.19) 193 (87.54) 215 (97.52)72 (182.88) 184 (83.46) 199 (90.27) 221 (100.25)73 (185.42) 189 (85.73) 204 (92.53) 227 (102.97)74 (187.96) 194 (88.00) 210 (95.26) 233 (105.69)75 (190.50) 200 (90.72) 216 (97.98) 240 (108.86)76 (193.04) 205 (92.99) 221(100.25) 246 (111.58)Metric conversion formula = Non-metric conversion formula =weight (kg)/height (m)2[weight (pounds)/height (inches)2] x 703Example of BMI calculation: Example of BMI calculation:A person who weighs 78.93 kilograms and is A person who weighs 164 pounds and is177 centimeters tall has a BMI of 25: 68 inches (or 5 8") tall has a BMI of 25:weight (78.93 kg)/height (1.77 m)2= 25 [weight (164 pounds)/height (68 inches)2] x703 = 25
xviiCLASSIFICATION OF OVERWEIGHT AND OBESITY BY BMI, WAISTCIRCUMFERENCE AND ASSOCIATED DISEASE RISK*Disease Risk* Relative to Normal Weightand Waist CircumferenceObesity Men ≤ 102 cm (≤ 40 in) > 102 cm (> 40 in)BMI (kg/m2) Class Women ≤ 88 cm (≤ 35 in) > 88 cm (> 35 in)Underweight <18.5 — —Normal+18.5 –24.9 — —Overweight 25.0 – 29.9 Increased HighObesity 30.0 – 34.9 I High Very High35.0 – 39.9 II Very High Very HighExtreme Obesity ≥40 III Extremely High Extremely High* Disease risk for type 2 diabetes, hypertension, and CVD.+ Increased waist circumference can also be a marker for increased risk even in persons of normal weight.be considered to have incremental absoluterisk above that estimated from the precedingrisk factors. Quantitative risk contribution isnot available for these risk factors, but theirpresence heightens the need for weight reduc-tion in obese persons.s Patient Motivation. When assessing thepatient’s motivation to enter weight loss ther-apy, the following factors should be evaluated:reasons and motivation for weight reduction;previous history of successful and unsuccess-ful weight loss attempts; family, friends, andwork-site support; the patient’s understandingof the causes of obesity and how obesity con-tributes to several diseases; attitude towardphysical activity; capacity to engage in physi-cal activity; time availability for weight lossintervention; and financial considerations. Inaddition to considering these issues, thehealth care practitioner needs to heighten apatient’s motivation for weight loss and pre-pare the patient for treatment. This can bedone by enumerating the dangers accompany-ing persistent obesity and by describing thestrategy for clinically assisted weight reduc-tion. Reviewing the patients’ past attempts atweight loss and explaining how the new treat-ment plan will be different can encouragepatients and provide hope for successfulweight loss.Evaluation and Treatment: The general goals ofweight loss and management are: (1) at a mini-mum, to prevent further weight gain; (2) toreduce body weight; and (3) to maintain a lowerbody weight over the long term. The overallstrategy for the evaluation and treatment ofoverweight and obese patients is presented in theTreatment Algorithm on the next page. Thisalgorithm applies only to the assessment foroverweight and obesity and subsequent decisionsTable ES-4:
xixbased on that assessment. It does not includeany initial overall assessment for cardiovascularrisk factors or diseases that are indicated.s Goals of Weight Loss and Management.The initial goal of weight loss therapy is toreduce body weight by approximately 10 per-cent from baseline. If this goal is achieved,further weight loss can be attempted, if indi-cated through further evaluation.A reasonable time line for a 10 percent reduc-tion in body weight is 6 months of therapy.For overweight patients with BMIs in the typ-ical range of 27 to 35, a decrease of 300 to500 kcal/day will result in weight losses ofabout 1⁄2 to 1 lb/week and a 10 percent loss in6 months. For more severely obese patientswith BMIs > 35, deficits of up to 500 to1,000 kcal/day will lead to weight losses ofabout 1 to 2 lb/week and a 10 percent weightloss in 6 months. Weight loss at the rate of 1to 2 lb/week (calorie deficit of 500 to 1,000kcal/day) commonly occurs for up to 6months. After 6 months, the rate of weightloss usually declines and weight plateausbecause of a lesser energy expenditure at thelower weight.Experience reveals that lost weight usually willbe regained unless a weight maintenance pro-gram consisting of dietary therapy, physicalactivity, and behavior therapy is continuedindefinitely.After 6 months of weight loss treatment,efforts to maintain weight loss should be putin place. If more weight loss is needed, anoth-er attempt at weight reduction can be made.This will require further adjustment of thediet and physical activity prescriptions.For patients unable to achieve significantweight reduction, prevention of furtherweight gain is an important goal; suchpatients may also need to participate in aweight management program.s Strategies for Weight Loss and WeightMaintenance.Dietary Therapy: A diet that is individuallyplanned and takes into account the patient’soverweight status in order to help create adeficit of 500 to 1,000 kcal/day should be anintegral part of any weight loss program. Apatient may choose a diet of 1,000 to 1,200kcal/day for women and 1,200 to 1,500kcal/day for men. Depending on the patient’srisk status, the low-calorie diet (LCD) recom-mended should be consistent with theNCEP’s Step I or Step II Diet (see page 74 ofthe guidelines). Besides decreasing saturatedfat, total fats should be 30 percent or less oftotal calories. Reducing the percentage ofdietary fat alone will not produce weight lossunless total calories are also reduced.Isocaloric replacement of fat with carbohy-drates will reduce the percentage of caloriesfrom fat but will not cause weight loss.Reducing dietary fat, along with reducingdietary carbohydrates, usually will be neededto produce the caloric deficit needed for anacceptable weight loss. When fat intake isreduced, priority should be given to reducingsaturated fat to enhance lowering of LDL-cholesterol levels. Frequent contacts with thepractitioner during dietary therapy help topromote weight loss and weight maintenanceat a lower weight.Physical Activity: An increase in physicalactivity is an important component of weightloss therapy, although it will not lead to sub-stantially greater weight loss over 6 months.Most weight loss occurs because of decreasedcaloric intake. Sustained physical activity ismost helpful in the prevention of weightregain. In addition, it has a benefit in reduc-ing cardiovascular and diabetes risks beyondthat produced by weight reduction alone. Formost obese patients, exercise should be initiat-ed slowly, and the intensity should be
xxincreased gradually. The exercise can be doneall at one time or intermittently over the day.Initial activities may be walking or swimmingat a slow pace. The patient can start by walk-ing 30 minutes for 3 days a week and canbuild to 45 minutes of more intense walkingat least 5 days a week. With this regimen, anadditional expenditure of 100 to 200 caloriesper day can be achieved. All adults should seta long-term goal to accumulate at least 30minutes or more of moderate-intensity physi-cal activity on most, and preferably all, daysof the week. This regimen can be adapted toother forms of physical activity, but walking isparticularly attractive because of its safety andaccessibility. Patients should be encouraged toincrease “every day” activities such as takingthe stairs instead of the elevator. With time,depending on progress and functional capaci-ty, the patient may engage in more strenuousactivities. Competitive sports, such as tennisand volleyball, can provide an enjoyable formof exercise for many, but care must be takento avoid injury. Reducing sedentary time isanother strategy to increase activity by under-taking frequent, less strenuous activities.Behavior Therapy: Strategies, based on learn-ing principles such as reinforcement, that pro-vide tools for overcoming barriers to compli-ance with dietary therapy and/or increasedphysical activity are helpful in achievingweight loss and weight maintenance. Specificstrategies include self-monitoring of both eat-ing habits and physical activity, stress manage-ment, stimulus control, problem solving, con-tingency management, cognitive restructur-ing, and social support.Combined Therapy: A combined interventionof behavior therapy, an LCD, and increasedphysical activity provides the most successfultherapy for weight loss and weight mainte-nance. This type of intervention should bemaintained for at least 6 months before con-sidering pharmacotherapy.Pharmacotherapy: In carefully selectedpatients, appropriate drugs can augmentLCDs, physical activity, and behavior therapyin weight loss. Weight loss drugs that havebeen approved by the FDA for long-term usecan be useful adjuncts to dietary therapy andphysical activity for some patients with a BMIof ≥ 30 with no concomitant risk factors ordiseases, and for patients with a BMI of ≥ 27with concomitant risk factors or diseases. Therisk factors and diseases considered importantenough to warrant pharmacotherapy at aBMI of 27 to 29.9 are hypertension, dyslipi-demia, CHD, type 2 diabetes, and sleepapnea. Continual assessment by the physicianof drug therapy for efficacy and safety is nec-essary.At the present time, sibutramine is availablefor long-term use. (Note: FDA approval oforlistat is pending a resolution of labelingissues and results of Phase III trials.) Itenhances weight loss modestly and can helpfacilitate weight loss maintenance. Potentialside effects with drugs, nonetheless, must bekept in mind. With sibutramine, increases inblood pressure and heart rate may occur.Sibutramine should not be used in patientswith a history of hypertension, CHD, conges-tive heart failure, arrhythmias, or history ofstroke. With orlistat, fat soluble vitamins mayrequire replacement because of partial malab-sorption. All patients should be carefullymonitored for these side effects.Weight Loss Surgery: Weight loss surgery isone option for weight reduction in a limitednumber of patients with clinically severe obe-sity, i.e., BMIs ≥ 40 or ≥ 35 with comorbidconditions. Weight loss surgery should bereserved for patients in whom efforts at med-ical therapy have failed and who are sufferingfrom the complications of extreme obesity.
xxiGastrointestinal surgery (gastric restriction[vertical gastric banding] or gastric bypass[Roux-en Y]) is an intervention weight lossoption for motivated subjects with acceptableoperative risks. An integrated program mustbe in place to provide guidance on diet, phys-ical activity, and behavioral and social supportboth prior to and after the surgery.s Adapt Weight Loss Programs To Meet theNeeds of Diverse Patients. Standard treat-ment approaches for overweight and obesitymust be tailored to the needs of variouspatients or patient groups. Large individualvariation exists within any social or culturalgroup; furthermore, substantial overlapamong subcultures occurs within the largersociety. There is, therefore, no “cookbook” orstandardized set of rules to optimize weightreduction with a given type of patient.However, to be more culturally sensitive andto incorporate patient characteristics in obesi-ty treatment programs: consider and adaptthe setting and staffing for the program; con-sider how the obesity treatment program inte-grates into other aspects of patient health careand self care; and expect and allow for pro-gram modifications based on patient respons-es and preferences.The issues of weight reduction after age 65involve such questions as: does weight lossreduce risk factors in older adults; are thererisks associated with obesity treatment thatare unique to older adults; and does weightreduction prolong the lives of older adults?Although there is less certainty about theimportance of treating overweight at olderages than at younger ages, a clinical decisionto forgo obesity treatment in older adultsshould be guided by an evaluation of thepotential benefit of weight reduction and thereduction of risk for future cardiovascularevents.In the obese patient who smokes, smokingcessation is a major goal of risk factor man-agement. Many well-documented health ben-efits accompany smoking cessation, but amajor obstacle to cessation has been theattendant weight gain observed in about 80percent of quitters. This weight gain averages4.5 to 7 lb, but in 13 percent of women and10 percent of men, weight gain exceeds 28 lb.Weight gain that accompanies smoking cessa-tion has been quite resistant to most dietary,behavioral, or physical activity interventions.The weight gained with smoking cessation isless likely to produce negative health conse-quences than would continued smoking. Forthis reason, smoking cessation should bestrongly advocated regardless of baselineweight. Prevention of weight gain throughdiet and physical activity should be stressed.For practical reasons, it may be prudent toavoid initiating smoking cessation and weightloss therapy simultaneously. If weight gainensues after smoking cessation, it should bemanaged vigorously according to the guide-lines outlined in this report. Although short-term weight gain is a common side effect ofsmoking cessation, this gain does not rule outthe possibility of long-term weight control.SUMMARY OF EVIDENCE-BASED RECOMMENDATIONSADVANTAGES OF WEIGHT LOSSThe recommendation to treat overweight andobesity is based not only on evidence that relatesobesity to increased mortality but also on RCTevidence that weight loss reduces risk factors fordisease. Thus, weight loss may not only helpcontrol diseases worsened by obesity, it may alsohelp decrease the likelihood of developing thesediseases. The panel reviewed RCT evidence todetermine the effect of weight loss on bloodpressure and hypertension, serum/plasma lipidconcentrations, and fasting blood glucose andfasting insulin. Recommendations focusing onA
xxiithese conditions underscore the advantages ofweight loss.1. Blood PressureTo evaluate the effect of weight loss on bloodpressure and hypertension, 76 articles reportingRCTs were considered for inclusion in theseguidelines. Of the 45 accepted articles, 35 werelifestyle trials and 10 were pharmacotherapy tri-als. There is strong and consistent evidence fromthese lifestyle trials in both overweight hyperten-sive and nonhypertensive patients that weightloss produced by lifestyle modifications reducesblood pressure levels. Limited evidence existsthat decreases in abdominal fat will reduce bloodpressure in overweight nonhypertensive individ-uals, although not independent of weight loss,and there is considerable evidence that increasedaerobic activity to increase cardiorespiratory fit-ness reduces blood pressure (independent ofweight loss). There is also suggestive evidencefrom randomized trials that weight loss pro-duced by most weight loss medications, exceptfor sibutramine, in combination with adjuvantlifestyle modifications will be accompanied byreductions in blood pressure. Based on a reviewof the evidence from the 45 RCT blood pressurearticles, the panel makes the following recom-mendation:Weight loss is recommended to lower ele-vated blood pressure in overweight andobese persons with high blood pressure.Evidence Category A.2. Serum/Plasma LipidsSixty-five RCT articles were evaluated for theeffect of weight loss on serum/plasma concentra-tions of total cholesterol, LDL-cholesterol, verylow-density lipoprotein (VLDL)-cholesterol,triglycerides, and HDL-cholesterol. Studies wereconducted on individuals over a range of obesityand lipid levels. Of the 22 articles accepted forinclusion in these guidelines, 14 RCT articlesexamined lifestyle trials while the remaining 8articles reviewed pharmacotherapy trials. Thereis strong evidence from the 14 lifestyle trials thatweight loss produced by lifestyle modificationsin overweight individuals is accompanied byreductions in serum triglycerides and by increas-es in HDL-cholesterol. Weight loss generallyproduces some reductions in serum total choles-terol and LDL-cholesterol. Limited evidenceexists that a decrease in abdominal fat correlateswith improvements in lipids, although the effectmay not be independent of weight loss, andthere is strong evidence that increased aerobicactivity to increase cardiorespiratory fitnessfavorably affects blood lipids, particularly ifaccompanied by weight loss. There is suggestiveevidence from the eight randomized pharma-cotherapy trials that weight loss produced byweight loss medications and adjuvant lifestylemodifications, including caloric restriction andphysical activity, does not result in consistenteffects on blood lipids. The following recom-mendation is based on the review of the data inthese 22 RCT articles:Weight loss is recommended to lower ele-vated levels of total cholesterol, LDL-cho-lesterol, and triglycerides, and to raise lowlevels of HDL-cholesterol in overweightand obese persons with dyslipidemia.Evidence Category A.3. Blood GlucoseTo evaluate the effect of weight loss on fastingblood glucose and fasting insulin levels, 49 RCTarticles were reviewed for inclusion in theseguidelines. Of the 17 RCT articles accepted, 9RCT articles examined lifestyle therapy trialsand 8 RCT articles considered the effects ofpharmacotherapy on weight loss and subsequentchanges in blood glucose. There is strong evi-dence from the nine lifestyle therapy trials that
xxiiiweight loss produced by lifestyle modificationreduces blood glucose levels in overweight andobese persons without diabetes, and weight lossreduces blood glucose levels and HbAlc in somepatients with type 2 diabetes. There is suggestiveevidence that decreases in abdominal fat willimprove glucose tolerance in overweight individ-uals with impaired glucose tolerance, althoughnot independent of weight loss; and there is lim-ited evidence that increased cardiorespiratory fit-ness improves glucose tolerance in overweightindividuals with impaired glucose tolerance ordiabetes, although not independent of weightloss. In addition, there is suggestive evidencefrom randomized trials that weight loss inducedby weight loss medications does not appear toimprove blood glucose levels any better thanweight loss through lifestyle therapy in over-weight persons both with and without type 2diabetes. Based on a full review of the data inthese 17 RCT articles, the panel makes the fol-lowing recommendation:Weight loss is recommended to lower ele-vated blood glucose levels in overweightand obese persons with type 2 diabetes.Evidence Category A.MEASUREMENT OF DEGREE OFOVERWEIGHT AND OBESITYPatients should have their BMI and levels ofabdominal fat measured not only for the initialassessment of the degree of overweight and obe-sity, but also as a guide to the efficacy of weightloss treatment. Although there are no RCTs thatreview measurements of overweight and obesity,the panel determined that this aspect of patientcare warranted further consideration and thatthis guidance was deemed valuable. Therefore,the following four recommendations that areincluded in the Treatment Guidelines were basedon nonrandomized studies as well as clinicalexperience.1. BMI To Assess Overweight and ObesityThere are a number of accurate methods toassess body fat (e.g., total body water, total bodypotassium, bioelectrical impedance, and dual-energy X-ray absorptiometry), but no trial dataexist to indicate that one measure of fatness isbetter than any other for following overweightand obese patients during treatment. Since mea-suring body fat by these techniques is oftenexpensive and is not readily available, a morepractical approach for the clinical setting is themeasurement of BMI; epidemiological andobservational studies have shown that BMI pro-vides an acceptable approximation of total bodyfat for the majority of patients. Because there areno published studies that compare the effective-ness of different measures for evaluating changesin body fat during weight reduction, the panelbases its recommendation on expert judgmentfrom clinical experience:Practitioners should use the BMI to assessoverweight and obesity. Body weight alonecan be used to follow weight loss, and todetermine efficacy of therapy. EvidenceCategory C.2. BMI To Estimate Relative RiskIn epidemiological studies, BMI is the favoredmeasure of excess weight to estimate relative riskof disease. BMI correlates both with morbidityand mortality; the relative risk for CVD risk fac-tors and CVD incidence increases in a gradedfashion with increasing BMI in all populationgroups. Moreover, calculating BMI is simple,rapid, and inexpensive, and can be applied gen-erally to adults. The panel, therefore, makes thisrecommendation:B
xxivThe BMI should be used to classify over-weight and obesity and to estimate rela-tive risk of disease compared to normalweight. Evidence Category C.3. Assessing Abdominal FatFor the most effective technique for assessingabdominal fat content, the panel consideredmeasures of waist circumference, waist-to-hipratio (WHR), magnetic resonance imaging(MRI), and computed tomography. Evidencefrom epidemiological studies shows waist cir-cumference to be a better marker of abdominalfat content than WHR, and that it is the mostpractical anthropometric measurement forassessing a patient’s abdominal fat content beforeand during weight loss treatment. Computedtomography and MRI are both more accuratebut impractical for routine clinical use. Based onevidence that waist circumference is a bettermarker than WHR—and taking into accountthat the MRI and computed tomography tech-niques are expensive and not readily available forclinical practice—the panel makes the followingrecommendation:The waist circumference should be used toassess abdominal fat content. EvidenceCategory C.4. Sex-Specific MeasurementsEvidence from epidemiological studies indicatesthat a high waist circumference is associatedwith an increased risk for type 2 diabetes, dys-lipidemia, hypertension, and CVD. Therefore,the panel judged that sex-specific cutoffs forwaist circumference can be used to identifyincreased risk associated with abdominal fat inadults with a BMI in the range of 25 to 34.9.These cutpoints can be applied to all adult eth-nic or racial groups. On the other hand, if apatient is very short, or has a BMI above the 25to 34.9 range, waist cutpoints used for the gen-eral population may not be applicable. Based onthe evidence from nonrandomized studies, thepanel makes this recommendation:For adult patients with a BMI of 25 to34.9 kg/m2, sex-specific waist circumferencecutoffs should be used in conjunction withBMI to identify increased disease risks.Evidence Category C.GOALS FOR WEIGHT LOSSThe general goals of weight loss and manage-ment are to reduce body weight, to maintain alower body weight over the long term, and toprevent further weight gain. Evidence indicatesthat a moderate weight loss can be maintainedover time if some form of therapy continues. Itis better to maintain a moderate weight loss overa prolonged period than to regain from amarked weight loss.1. Initial Goal of Weight Loss from BaselineThere is strong and consistent evidence fromrandomized trials that overweight and obesepatients in well-designed programs can achieve aweight loss of as much as 10 percent of baselineweight. In the diet trials, an average of 8 percentof baseline weight was lost. Since this averageincludes persons who did not lose weight, anindividualized goal of 10 percent is reasonable.The panel, therefore, recommends that:The initial goal of weight loss therapyshould be to reduce body weight by approx-imately 10 percent from baseline. Withsuccess, further weight loss can be attempt-ed if indicated through further assessment.Evidence Category A.C
xxvD2. Amount of Weight LossRandomized trials suggest that weight loss at therate of 1 to 2 lb/week (calorie deficit of 500 to1,000 kcal/day) commonly occurs for up to 6months.Weight loss should be about 1 to 2 lb/weekfor a period of 6 months, with the subse-quent strategy based on the amount ofweight lost. Evidence Category B.HOW TO ACHIEVE WEIGHT LOSSThe panel reviewed relevant treatment strategiesdesigned for weight loss that can also be used tofoster long-term weight control and preventionof weight gain. The consequent recommenda-tions emphasize the potential effectiveness ofweight control using multiple interventions andstrategies, including dietary therapy, physicalactivity, behavior therapy, pharmacotherapy, andsurgery, as well as combinations of these strate-gies.1. Dietary TherapyThe panel reviewed 86 RCT articles to deter-mine the effectiveness of diets on weight loss(including LCDs, very low-calorie diets(VLCDs), vegetarian diets, American HeartAssociation dietary guidelines, the NCEP’s StepI diet with caloric restriction, and other low-fatregimens with varying combinations ofmacronutrients). Of the 86 articles reviewed, 48were accepted for inclusion in these guidelines.These RCTs indicate strong and consistent evi-dence that an average weight loss of 8 percent ofinitial body weight can be obtained over 3 to 12months with an LCD and that this weight losseffects a decrease in abdominal fat; and,although lower-fat diets without targeted caloricreduction help promote weight loss by produc-ing a reduced caloric intake, lower-fat diets withtargeted caloric reduction promote greaterweight loss than lower-fat diets alone. Further,VLCDs produce greater initial weight lossesthan LCDs (over the long term of >1 year,weight loss is not different than that of theLCDs). In addition, randomized trials suggestthat no improvement in cardiorespiratory fitnessas measured by VO2 max appears to occur inobese adults who lose weight on LCDs alonewithout physical activity. The following recom-mendations are based on the evidence extractedfrom the 48 accepted articles:LCDs are recommended for weight loss inoverweight and obese persons. EvidenceCategory A. Reducing fat as part of anLCD is a practical way to reduce calories.Evidence Category A.Reducing dietary fat alone without reduc-ing calories is not sufficient for weight loss.However, reducing dietary fat, along withreducing dietary carbohydrates, can facili-tate caloric reduction. Evidence Category A.A diet that is individually planned to helpcreate a deficit of 500 to 1,000 kcal/dayshould be an intregal part of any programaimed at achieving a weight loss of 1 to 2lb/week. Evidence Category A.2. Physical ActivityEffects of Physical Activity on Weight LossTwenty-three RCT articles were reviewed todetermine the effect of physical activity onweight loss, abdominal fat (measured by waistcircumference), and changes in cardiorespiratoryfitness (VO2 max). Thirteen of these articleswere accepted for inclusion in these guidelines.A review of these articles reveals strong evidencethat physical activity alone, i.e., aerobic exercise,in obese adults results in modest weight loss andthat physical activity in overweight and obese
xxviadults increases cardiorespiratory fitness, inde-pendent of weight loss. Randomized trials sug-gest that increased physical activity in over-weight and obese adults reduces abdominal fatonly modestly or not at all, and that regularphysical activity independently reduces the riskfor CVD. The panel’s recommendation on phys-ical activity is based on the evidence from these13 articles:Physical activity is recommended as part ofa comprehensive weight loss therapy andweight control program because it: (1)modestly contributes to weight loss in over-weight and obese adults (EvidenceCategory A), (2) may decrease abdominalfat (Evidence Category B), (3) increasescardiorespiratory fitness (Evidence CategoryA), and (4) may help with maintenance ofweight loss (Evidence Category C).Physical activity should be an integral partof weight loss therapy and weight mainte-nance. Initially, moderate levels of physicalactivity for 30 to 45 minutes, 3 to 5 daysa week, should be encouraged. All adultsshould set a long-term goal to accumulateat least 30 minutes or more of moderate-intensity physical activity on most, andpreferably all, days of the week. EvidenceCategory B.Effects of Physical Activity and Diet on WeightLoss (Combined Therapy)Twenty-three RCT articles were reviewed todetermine the effects on body weight of a com-bination of a reduced-calorie diet with increasedphysical activity. Fifteen of these articles wereaccepted for inclusion in the guidelines. Thesearticles contain strong evidence that the combi-nation of a reduced-calorie diet and increasedphysical activity produces greater weight lossthan diet alone or physical activity alone, andthat the combination of diet and physical activi-ty improves cardiorespiratory fitness as measuredby VO2 max in overweight and obese adultswhen compared to diet alone. The combinedeffect of a reduced calorie diet and increased-physical activity seemingly produced modestlygreater reductions in abdominal fat than eitherdiet alone or physical activity alone, although ithas not been shown to be independent of weightloss. The panel’s following recommendations arebased on the evidence from these articles:The combination of a reduced calorie dietand increased physical activity is recom-mended since it produces weight loss thatmay also result in decreases in abdominalfat and increases in cardiorespiratory fit-ness. Evidence Category A.3. Behavior TherapyThirty-six RCTs were reviewed to evaluatewhether behavior therapy provides additionalbenefit beyond other weight loss approaches, aswell as to compare various behavioral tech-niques. Of the 36 RCTs reviewed, 22 wereaccepted. These RCTs strongly indicate thatbehavioral strategies to reinforce changes in dietand physical activity in obese adults produceweight loss in the range of 10 percent over 4months to 1 year. In addition, no one behaviortherapy appeared superior to any other in itseffect on weight loss; multimodal strategiesappear to work best and those interventions withthe greatest intensity appear to be associatedwith the greatest weight loss. Long-term follow-up of patients undergoing behavior therapyshows a return to baseline weight for the greatmajority of subjects in the absence of continuedbehavioral intervention. Randomized trials sug-gest that behavior therapy, when used in combi-
xxviination with other weight loss approaches, pro-vides additional benefits in assisting patients tolose weight short-term, i.e., 1 year (no additionalbenefits are found at 3 to 5 years). The panelfound little evidence on the effect of behaviortherapy on cardiorespiratory fitness. Evidencefrom these articles provided the basis for the fol-lowing recommendation:Behavior therapy is a useful adjunct whenincorporated into treatment for weight lossand weight maintenance. EvidenceCategory B.There is also suggestive evidence that patientmotivation is a key component for success in aweight loss program. The panel, therefore,makes the following recommendation:Practitioners need to assess the patient’smotivation to enter weight loss therapy;assess the readiness of the patient to imple-ment the plan and then take appropriatesteps to motivate the patient for treatment.Evidence Category D.4. Summary of Lifestyle TherapyThere is strong evidence that combined inter-ventions of an LCD, increased physical activity,and behavior therapy provide the most success-ful therapy for weight loss and weight mainte-nance. The panel makes the following recom-mendation:Weight loss and weight maintenance therapyshould employ the combination of LCDs,increased physical activity, and behavior ther-apy. Evidence Category A.5. PharmacotherapyA review of 44 pharmacotherapy RCT articlesprovides strong evidence that pharmacologicaltherapy (which has generally been studied alongwith lifestyle modification, including diet andphysical activity) using dexfenfluramine, sibu-tramine, orlistat, or phentermine/fenfluramineresults in weight loss in obese adults when usedfor 6 months to 1 year. Strong evidence alsoindicates that appropriate weight loss drugs canaugment diet, physical activity, and behaviortherapy in weight loss. Adverse side effects fromthe use of weight loss drugs have been observedin patients. As a result of the observed associa-tion of valvular heart disease in patients takingfenfluramine and dexfenfluramine alone or incombination, these drugs have been withdrawnfrom the market. Weight loss drugs approved bythe FDA for long-term use may be useful as anadjunct to diet and physical activity for patientswith a BMI of ≥ 30 with no concomitant obesi-ty-related risk factors or diseases, as well as forpatients with a BMI of ≥ 27 with concomitantrisk factors or diseases; moreover, using weightloss drugs singly (not in combination) and start-ing with the lowest effective doses can decreasethe likelihood of adverse effects. Based on thisevidence, the panel makes the following recom-mendation:Weight loss drugs approved by the FDAmay be used as part of a comprehensiveweight loss program, including dietarytherapy and physical activity for patientswith a BMI of ≥ 30 with no concomitantobesity-related risk factors or diseases, andfor patients with a BMI of ≥ 27 withconcomitant obesity-related risk factors ordiseases. Weight loss drugs should never beused without concomitant lifestyle modifi-cations. Continual assessment of drug ther-apy for efficacy and safety is necessary. Ifthe drug is efficacious in helping thepatient to lose and/or maintain weight loss
xxviiiand there are no serious adverse effects, itcan be continued. If not, it should be dis-continued. Evidence Category B.6. Weight Loss SurgeryThe panel reviewed 14 RCTs that examined theeffect of surgical procedures on weight loss; 8were deemed appropriate. All of the studiesincluded individuals who had a BMI of 40kg/m2or above, or a BMI of 35 to 40 kg/m2with comorbidity. These trials provide strongevidence that surgical interventions in adultswith clinically severe obesity, i.e., BMIs ≥ 40 or≥ 35 with comorbid conditions, result in sub-stantial weight loss, and suggestive evidence thatlifelong medical surveillance after surgery is nec-essary. Therefore, the panel makes the followingrecommendation:Weight loss surgery is an option for careful-ly selected patients with clinically severeobesity (BMI ≥ 40 or ≥ 35 with comorbidconditions) when less invasive methods ofweight loss have failed and the patient is athigh risk for obesity-associated morbidityor mortality. Evidence Category B.GOALS FOR WEIGHT LOSS MAINTENANCEOnce the goals of weight loss have been success-fully achieved, maintenance of a lower bodyweight becomes the challenge. Whereas studieshave shown that weight loss is achievable, it isdifficult to maintain over a long period of time(3 to 5 years). In fact, the majority of personswho lose weight, once dismissed from clinicaltherapy, frequently regain it—so the challenge tothe patient and the practitioner is to maintainthe weight loss. Successful weight reduction thusdepends on continuing a maintenance programon a long-term basis. In the past, obtaining thegoal of weight loss has been considered the endof weight loss therapy. Observation, monitoring,and encouragement of patients who have suc-cessfully lost weight should be continued longterm. The panel’s recommendations on weightloss maintenance are derived from RCT evi-dence as well as nonrandomized and observa-tional studies.1. Weight Maintenance PhaseRCTs from the Behavior Therapy section abovesuggest that lost weight usually will be regainedunless a weight maintenance program consistingof dietary therapy, physical activity, and behaviortherapy is continued indefinitely. Drug therapyin addition may be helpful during the weightmaintenance phase. The panel also reviewedRCT evidence that considered the rate of weightloss and the role of weight maintenance. TheseRCTs suggest that after 6 months of weight losstreatment, efforts to maintain weight loss areimportant. Therefore, the panel recommends thefollowing:After successful weight loss, the likelihoodof weight loss maintenance is enhanced bya program consisting of dietary therapy,physical activity, and behavior therapywhich should be continued indefinitely.Drug therapy can also be used. However,drug safety and efficacy beyond 1 year oftotal treatment have not been established.Evidence Category B.A weight maintenance program should bea priority after the initial 6 months ofweight loss therapy. Evidence Category B.Strong evidence indicates that better weight lossresults are achieved with dietary therapy whenthe duration of the intervention is at least 6months. Suggestive evidence also indicates thatduring dietary therapy, frequent contactsE
xxixbetween professional counselors and patientspromote weight loss and maintenance.Therefore, the panel recommends the following:The literature suggests that weight loss andweight maintenance therapies that providea greater frequency of contacts between thepatient and the practitioner and are pro-vided over the long term should be utilizedwhenever possible. This can lead to moresuccessful weight loss and weight mainte-nance. Evidence Category C.SPECIAL TREATMENT GROUPSThe needs of special patient groups must beaddressed when considering treatment optionsfor overweight and obesity. The guidelines focuson three such groups including smokers, olderadults, and diverse patient populations.1. SmokersCigarette smoking is a major risk factor for car-diopulmonary disease. Because of its attendanthigh risk, smoking cessation is a major goal ofrisk-factor management. This aim is especiallyimportant in the overweight or obese patient,who usually carries excess risk from obesity-asso-ciated risk factors. Thus, smoking cessation inthese patients becomes a high priority for riskreduction. Smoking and obesity together appar-ently compound cardiovascular risk, but fear ofweight gain upon smoking cessation is an obsta-cle for many patients. Therefore, the panel rec-ommends that:All smokers, regardless of their weight sta-tus, should quit smoking. EvidenceCategory A. Prevention of weight gainshould be encouraged and if weight gaindoes occur, it should be treated throughdietary therapy, physical activity, andbehavior therapy, maintaining the primaryemphasis on the importance of abstinencefrom smoking. Evidence Category C.2. Older AdultsThe general nutritional safety of weight reduc-tion at older ages is of concern because restric-tions on overall food intake due to dieting couldresult in inadequate intake of protein or essentialvitamins or minerals. In addition, involuntaryweight loss indicative of occult disease might bemistaken for success in voluntary weight reduc-tion. These concerns can be alleviated by provid-ing proper nutritional counseling and regularbody weight monitoring in older persons forwhom weight reduction is prescribed. A reviewof several studies indicates that age alone shouldnot preclude treatment for obesity in adult menand women. In fact, there is evidence fromRCTs that weight reduction has similar effects inimproving cardiovascular disease risk factors inolder and younger adults. Therefore, in thepanel’s judgment:A clinical decision to forego obesity treat-ment in older adults should be guided byan evaluation of the potential benefits ofweight reduction for day-to-day function-ing and reduction of the risk of future car-diovascular events, as well as the patient’smotivation for weight reduction. Caremust be taken to ensure that any weightreduction program minimizes the likeli-hood of adverse effects on bone health orother aspects of nutritional status. EvidenceCategory D.F
xxx3. Diverse Patient PopulationsStandard obesity treatment approaches should betailored to the needs of various patients orpatient groups. It is, however, difficult to deter-mine from the literature how often this occurs,how specific programs and outcomes are influ-enced by tailoring, and whether it makes weightloss programs more effective. After reviewingtwo RCTs, four cross-sectional studies, and fourintervention studies, as well as additional pub-lished literature on treatment approaches withdiverse patient populations, the panel recom-mends the following:The possibility that a standard approach toweight loss will work differently in diversepatient populations must be consideredwhen setting expectations about treatmentoutcomes. Evidence Category B.
1IRATIONALE FOR GUIDELINES DEVELOPMENTAn estimated 97 million adults in the UnitedStates are overweight or obese, 1a condition thatsubstantially raises their risk of morbidity fromhypertension, 2-6type 2 diabetes, 7-10stroke, 11-13gallbladder disease, 14, 15osteoarthritis, 16-18sleepapnea and respiratory problems, 19-21andendometrial, breast, prostate, and colon cancers.22-24As a major contributor to preventivedeath in the United States today, 25overweightand obesity pose a major public health challenge.Not only is the prevalence of this serious medicalcondition soaring among adults (between 1960and 1994, overweight increased from 30.5 to 32percent among adults ages 20 to 74 and obesityincreased from 12.8 percent to 22.5 percent),but it is also affecting ever greater numbers ofAmerican youth and exacting a particularly harshtoll from low income women and minorities.The Third National Health and NutritionExamination Survey (NHANES III) estimatedthat 13.7 percent of children and 11.5 percent ofadolescents are overweight, while a number ofsmaller, ethnic-specific studies suggest that over-weight and obesity may afflict up to 30 to 40percent of children and youth from minoritypopulations. 26 , 27The prevalence of overweight and obesity inadults in the United States increased markedlyduring the last decade. According to NHANESIII data, 54.9 percent of U.S. adults aged 20years and older are either overweight or obese;32.6 percent are overweight, defined as having abody mass index (BMI)* of 25.0 to 29.9 kg/m2;and 22.3 percent are obese with a BMI of ≥ 30kg/m2. 1The panel acknowledges that overweightand obesity are not mutually exclusive; obesepersons are also overweight. Since overweightand obesity lead to increased morbidity and mor-tality, these figures demonstrate the enormity ofthe public health problem, as well as the clinicalproblem, of overweight and obesity in this country.In this report, overweight is defined as a BMI of25.0 to 29.9 kg/m2and obesity as a BMI of ≥ 30kg/m2. The rationale behind these definitions isbased on epidemiological data that show increas-es in mortality with BMIs above 25 kg/m2.28-32The increase in mortality, however, tends to bemodest until a BMI of 30 kg/m2is reached. 28, 31,32For persons with a BMI of ≥ 30 kg/m2, mor-tality rates from all causes, and especially fromcardiovascular disease, are generally increased by50 to 100 percent above that of persons withBMIs in the range of 20 to 25 kg/m2. 28, 31, 32Overweight and obesity result from a complexinteraction between genes and the environmentcharacterized by long-term energy imbalance dueto a sedentary lifestyle, excessive caloric con-sumption, or both. 33They develop in a sociocul-tural environment characterized by mechaniza-tion, sedentary lifestyle, and ready access toabundant food. Attempts to prevent overweightand obesity are difficult to both study andachieve. Indeed, few research efforts have investi-gated either individual or community-basedprevention strategies. 34INTRODUCTIONA* The BMI is calculated as follows: BMI = weight (kg)/ height squared (m2). Conversion: [weight (pounds)/height (inches) 2] x 703 (1 lb = 0.45 kg) (1.in= 2.54 cm = 0.0254 m). A BMI of 25 is equivalent to 184 lb in a 6’0” person and to 155 lb in one5’6”. A BMI of 30 is equivalent to 221 lb in a 6’0”person and to 186 lb in one 5’6”. (The conversion of BMI according to weight for height is provided in Appendix V.)
2Chapter 1: IntroductionA substantial body of research, however, doesexist on the health risks of overweight and obesi-ty, and on methods for treatment. This report,which bases its recommendations primarily onpublished evidence, emphasizes the importantrole of primary care practitioners in evaluating alloverweight and obese adults and promotingweight control through the use of multiple inter-ventions and strategies tailored to particularpatient needs. Although the recommendationsand guidelines included in this report focus onthe clinical assessment and treatment of over-weight and obese patients, a second importantgoal is to encourage primary care practitioners totake an active role in preventing inappropriateweight gain among all their patients.OBJECTIVES OF THE GUIDELINESs To identify, evaluate, and summarize pub-lished information about the assessment andtreatment of overweight and obesity;s To provide evidence-based guidelines forphysicians, other health care practitioners, andhealth care organizations for the evaluationand treatment of overweight and obesity inadults; ands To identify areas for future research.GUIDELINE DEVELOPMENT METHODOLOGYThe National Heart, Lung, and Blood Institute’s(NHLBI) Obesity Education Initiative, in coop-eration with the National Institute of Diabetesand Digestive and Kidney Diseases, convenedthe Expert Panel on the Identification,Evaluation, and Treatment of Overweight andObesity in Adults. These guidelines address thetreatment of overweight and obesity only inadults, but it is the judgment of the panel thatguidelines for treating obesity in children areequally important and should be drafted as soonas possible (see Appendix III for information onoverweight and obesity in children). The panel’scharge was to develop evidence-based clinicalguidelines for primary care practitioners; howev-er, the guidelines should also be useful for certainspecialists. The decision to develop “evidence-based” guidelines was based on the increasedattention being paid to clinical practice guide-lines from methodologists, professional associa-tions, third-party payers and policy makers, andthe NHLBI’s mission to analyze research resultswith the goal of providing information that mayenable health care practitioners to enhance theirability to detect, treat, and prevent disease. 35Inkeeping with this approach, the panels defined topics to be included in the guide-lines;s developed an evidence model depicting thestrategy of inquiry for each area of scientificinterest;s established criteria for searching and abstract-ing the literature;s constructed and reviewed evidence tables ofindividual studies and summary tables of stud-ies falling within a specific category of evi-dence; ands identified the level or strength of the evidencethat served as the basis for the recommenda-tions.A complete description of the methodology usedto develop the guidelines is included in thereport as Appendix I.A.1.The guidelines are based primarily on a systemat-ic review of the published scientific literature inEnglish found in MEDLINE from January 1980through September 1997. This was done in theinterest of time and economy. This informationwas supplemented by material provided by thepanel and an ancestral search of appropriate ref-erences in eligible articles. The literature wassearched and systematically reviewed bys establishing a priori eligibility criteria forinclusion of studies;BC
3and may require surgery. These guidelines arenot intended for pregnant women. Excludedfrom the analysis were adults with pharmacologi-cally induced obesity and those with specificgenetic syndromes associated with overweightand obesity.The selection of weight loss interventions to beconsidered was determined by the literaturereview. Namely, the panel considered any topicfor which articles meeting inclusion criteria werefound, including diet, physical activity, behaviortherapy, pharmacological therapy, surgery, andcombinations of these modalities. No clinicalinterventions were excluded at the outset.However, the panel did not consider other inter-ventions such as acupuncture or hypnosis, forwhich no randomized trial articles were available.Clinical interventions to prevent further weightgain in individuals already overweight were alsoconsidered relevant.The panel also evaluated population factors andclinical situations that might potentially influ-ence the physiological, medical, behavioral, orsociocultural context for obesity identificationand treatment. Evidence on special populationsand situations was provided from RCTs andnon-RCTs when available, but in many casessuch evidence was meager. Population factorsand clinical situations selected for special consid-eration for obesity classification and treatmentwere age, gender, race/ethnicity, socioeconomicstatus, pregnancy, eating disorders, sleep apnea,extreme obesity (BMI ≥ 40), concurrent treat-ment of other major conditions (such as heartdisease or diabetes), and treatment of obesity inconjunction with smoking cessation. When evi-dence on these special populations or clinicalconditions was insufficient to meet standards forinclusion in the main text of the guidelines, rele-vant issues are identified for the user, and insome cases are cross-referenced to an appendix(see Appendix III) or discussed in sidebar text asa commentary.s reviewing titles and abstracts to select promis-ing articles;s reviewing these full articles; ands compiling evidence tables summarizing thosearticles that met the inclusion criteria.As a priority, the panel identified randomizedcontrolled trials (RCTs) as the strongest level ofevidence for the evaluation of treatment efficacy.Only RCTs lasting 4 months or more were con-sidered. The only exceptions were a few 3-month trials in the diet and pharmacotherapysections. With the assistance of the San AntonioCochrane Center*, 394 publications of RCTswere reviewed for data abstraction. RCT evi-dence serves as the basis for the many recom-mendations contained in these guidelines relatedto treatment efficacy. Instances when the panelhad to make recommendations where RCTswere insufficient or absent are clearly indicatedin the text. These instances most often pertain toissues of obesity assessment, classification, andmeasurement where RCT evidence would not beappropriate to answer the question. These issuesare best addressed by epidemiological/observa-tional studies of large population groups. Inthose few cases where the literature could notcredibly support a recommendation, panel mem-bers relied on clinical experience and knowledge.The panel recognizes that by relying primarilyon only published literature, a publication bias(a positive result is more likely to be publishedthan is a negative result) may exist, so that treat-ment efficacy may be overstated. 36-39However,no other reliable sources of information wereavailable.The targeted population for the guidelines is alloverweight and obese adults (18 years of age andolder) with a BMI ≥ 25, with particular empha-sis on those with cardiovascular risk factors.While the guidelines are appropriate for patientswith a BMI ≥ 40, their care is often complicated* The San Antonio Cochrane Center is one of 12 centers around the world that comprise the Cochrane Collaboration. The Cochrane Collaboration is an interna-tional organization established in 1993 whose mission is to prepare, maintain, and disseminate systematic reviews and meta-analyses of health care interventions.
4Chapter 1: IntroductionThe panel recognized the possibility of an advo-cacy “bias” due to the large number of panelmembers drawn from organizations with anadvocacy role in the treatment of obesity. As aresult, it was agreed to obtain formal externalreviews of the document from 59 professionalsocieties, consumer groups and goverment agen-cies representing a wide spectrum of expertiseand concern about obesity.The format for those sections of the report basedon the RCT evidence begins with an evidencestatement followed by the rationale for thatstatement. At the end of a series of related evi-dence statements, a recommendation is given.Each evidence statement (other than those withno available evidence) and each recommenda-tion is categorized by a level of evidence (Athrough D) as described below. Statements forwhich there is no available evidence are so indi-cated.Category A: Evidence is from endpoints of well-designed RCTs (or trials that depart only mini-mally from randomization) that provide a con-sistent pattern of findings in the population forwhich the recommendation is made. Category Atherefore requires substantial numbers of studiesinvolving substantial numbers of participants.Category B: Evidence is from endpoints of inter-vention studies that include only a limited num-ber of RCTs, post-hoc or subgroup analysis ofRCTs, or meta-analysis of RCTs. In general,Category B pertains when few randomized trialsexist, they are small in size, and the trial resultsare somewhat inconsistent, or the trials wereundertaken in a population that differs from thetarget population of the recommendation.Category C: Evidence is from outcomes ofuncontrolled or nonrandomized trials or fromobservation studies.Category D: Expert judgment is based on thepanel’s synthesis of evidence from experimentalresearch described in the literature and/orderived from the consensus of panel membersbased on clinical experience or knowledge thatdoes not meet the above-listed criteria. This cat-egory is used only in cases where the provisionsof some guidance was deemed valuable but anadequately compelling clinical literature address-ing the subject of the recommendation wasdeemed insufficient to justify placement in oneof the other categories (A through C).In applying these guidelines, the reader shouldnote some caveats:s The emphasis of these guidelines was to iden-tify effective interventions, not to rank-orderthem in terms of their efficacy or effective-ness. The panel chose not to emphasize com-parisons among interventions because therewere few studies that compared long-termoutcomes. Also, since individual preferencesand circumstances often dictate choice oftherapy, the panel wished to present a menuof options rather than a ranked list of choices.s When no evidence was available on the effica-cy of a treatment, the panel usually renderedno opinion. An absence of studies should notbe confused with an absence of effect. Whileclinicians may prefer to use proven therapiesrather than untested ones, the lack of testingdoes not denote that the untested therapydoes not work.s The limitations of RCTs must be kept inmind. The RCT is the primary method fordemonstrating efficacy. Often, participantsenrolled in RCTs differ from the individualsin a primary care practice, and effectiveness inthe community may differ from efficacy asmeasured in an RCT.s The potential exists for misinterpretation ofclinical trial results. Analysis of endpoints notspecified at the outset, or post hoc or sub-group analysis, should be viewed as hypothe-sis-generating rather than hypothesis-testing.
5STATEMENT OF ASSUMPTIONSThe panel has made every attempt to base itsrecommendations on published evidence, withparticular attention to RCTs. Data from RCTsprovide the strongest evidence regarding theimpact of an intervention. The RCT literaturepredominantly describes short-term outcomes(< 1 year), although there are a small number ofpublished RCTs of intermediate and long-termtreatment and maintenance of weight loss. Thepanel chose to examine RCTs lasting 4 monthsor longer as their first priority. However, thereare some 3-month studies included in the dietand pharmacotherapy sections.Evidence of beneficial effects of weight reductionon risk factors and on diseases with which obesi-ty is associated, and evidence of the associationof obesity and mortality, are also available in thenon-RCT epidemiological literature. Therefore,although the treatment recommendations inthese guidelines are derived primarily from RCTevidence, they also come in part from the con-sidered judgment of the expert panel memberswho weighed the non-RCT epidemiological evi-dence.In setting forth its recommendations, the panelassumes that, for most individuals, the benefitsof weight loss on overall health outweigh theharmful effects, and that weight loss can bemaintained in many individuals with resultinglong-term health benefits. The recommendationsapply to all segments of the adult population,although apparent differences in applicabilityhave been considered and some guidance is pro-vided with respect to certain sectors of the popu-lation such as the elderly and in clinical situa-tions. This additional guidance is in the form ofnotations within the text and is sometimes sup-plemented by an appendix (Appendix III).INTENDED USERS OF THESE GUIDELINESThese guidelines were developed primarily foruse by physicians and associated health profes-sionals in clinical practice. They should also beuseful to managed care organizations or othergroups that define benefit plans for patients orhandle health care resources. Users of theseguidelines are encouraged to note text andappendix references to situations in whichweight reduction treatment may be contraindi-cated or may involve special treatment tech-niques or safety considerations (e.g., in olderadults or in certain sociocultural contexts).These guidelines also provide a state-of-the-artreview of the scientific basis of the relationbetween obesity and major disease endpointsand of the scientific rationale for the manage-ment of the overweight and obese patients. Thesystematic assessment of the literature containedin this document should be a valuable resourceto health care policy makers and clinical investi-gators.D E
62OVERWEIGHT AND OBESITY:BACKGROUNDHEALTH AND ECONOMIC COSTSIn 1973, and again in 1977, the John E. FogartyInternational Center at the National Institutes ofHealth (NIH), as part of its preventive medicineseries, sponsored two conferences that dealt withobesity as a public health problem; controversywas apparent regarding the cause-and-effect rela-tionship between obesity and ill health. 40, 41In1985, an NIH Consensus DevelopmentConference was held on the health implicationsof obesity. This conference provided importantnational recognition that obesity is a serioushealth condition that leads to increased morbidi-ty and mortality. The Consensus DevelopmentConference concluded that both prevention andtreatment of obesity were medical priorities inthe United States 42. In that conference, theterms ‘overweight’ and ‘obesity’ were defined aspart of a continuum of increasing health risk.In 1990, the Nation’s health goals for the year2000 were set forth with the release of HealthyPeople 2000 43, in which a national goal toreduce the prevalence of overweight was articu-lated. In 1993, the Deputy Assistant Secretaryfor Health (J. Michael McGinnis) and the for-mer Director of the Centers for Disease Controland Prevention (CDC) (William Foege) co-authored a journal article, “Actual Causes ofDeath in the U.S.” It concluded that a combina-tion of dietary factors and sedentary activity pat-terns accounts for at least 300,000 deaths eachyear, and, obesity was a key contributor. 25In1995, the Institute of Medicine issued a reportthat expressed concern about the growing preva-lence of overweight and obesity in this country,and suggested ways to evaluate various weightloss and weight maintenance programs availableto U.S. consumers. 441. Prevalence and Time TrendsNationally representative U.S. health examina-tion surveys, in which weight and height weremeasured in samples of the population, dateback to 1960. Beginning with the SecondNational Health and Nutrition ExaminationSurvey (NHANES II) (1976-1980), the defini-tion of overweight that has been used to com-pare these epidemiologic surveys has been a sta-tistical one that corresponded to the 85th per-centile of body mass index (BMI) for men andwomen aged 20 through 29 years in NHANESII with no particular relation to a specificincrease in disease risk. 45Adults in these surveyshave been categorized as overweight with a BMI≥ 27.8 kg/m2for men and ≥ 27.3 kg/m2forwomen. 45The rationale for using persons aged20 to 29 years as the reference population is sup-ported largely by the observation that theincreases in body weight after age 29 that com-monly occur with aging are attributable primari-ly to fat accumulation. 46,47However, the BMIlevels used for the definition of overweight andobesity are somewhat arbitrary, since the rela-tionship between body weight and disease risk iscontinuous with the exception of the extremelyunderweight: disease risk increases as weightincreases.Figure 1 depicts data from several NHANESsurveys using the panel’s definition of over-A
7weight as a BMI of 25 to 29.9 kg/m2and ofobesity as a BMI of ≥ 30 kg/m2. From 1960 to1994, the prevalence of overweight increasedslightly from 37.8 to 39.4 percent in men andfrom 23.6 to 24.7 percent in women (NationalCenter for Health Statistics/CDC) 48. In menand women together, overweight increased from30.5 to 32.0 percent. 48During the same timeperiod, however, the prevalence of obesityincreased from 10.4 to 19.9 percent in men andfrom 15.1 to 24.9 percent in women. In menand women together, obesity increased from12.8 to 22.5 percent. Most of the increaseoccurred in the past decade. In addition toadults, obesity in U.S. children increasedmarkedly as well 49(see Appendix III) and, ifunchecked, portends an even greater increase inadult obesity in the future.Table II-1 shows the combined prevalence ofoverweight and obesity, defined as a BMI of ≥25.0 kg/m2, among persons aged 20 to 80 plusyears, by age, race/ethnicity, and gender in theUnited States, 1960 to 1994. 48The increase inoverweight and obesity appears to have occurredamong U.S. adults across all ages, genders, andracial/ethnic groups. The most recent NHANESIII surveys, conducted from 1988-1994, reportedthat 59.4 percent of men and 50.7 percent ofwomen in the United States are overweight orobese. The prevalence is much higher in non-Hispanic Black women (66.0 percent), inMexican-American women (65.9 percent), andin Mexican-American men (63.9 percent).Using the definition of obesity as a BMI of ≥ 30kg/m2, Table II-2 shows that in the UnitedStates, 19.5 percent of men and 25.0 percent of50Source: CDC/NCHS, United States, 1960-94(ages 20-74 years)Figure 1. Age-Adjusted Prevalence of Overweight(BMI 25-29.9) and Obesity (BMI ≥30)40PercentBMI 25-29.9302010Men0Women37.841.139.1 39.4NHES I (1960-62)NHANES I (1971-74)NHANES II (1976-80)NHANES III (1988-94)23.6 23.6 24.3 24.710.411.8 12.219.915.116.1 16.324.9BMI ≥30Men Women
8Chapter 2: Overweight and Obesity: BackgroundTABLE II-1:NHES I NHANES I NHANES II HHANES NHANES IIIGender, race/ethnicity, age 20 1960-62 1971-74 1976-80 1982-84 1988-94years and older, age adjusted: (age 20-74) (age 20-74) (age 20-74) (age 20-74) (age ≥20)Both Sexes 43.3 46.1 46.0 54.9Men 48.2 52.9 51.4 59.4Women 38.7 39.7 40.8 50.7White men 48.8 53.7 52.3 61.0White women 36.1 37.6 38.4 49.2Black men 43.1 48.9 49.0 56.5Black women 57.0 57.6 61.0 65.8White, non-Hispanic men 52.0 60.6White, non-Hispanic women 37.6 47.4Black, non-Hispanic men 48.9 56.7Black, non-Hispanic women 60.6 66.0Mexican-American men 59.7 63.9Mexican-American women 60.1 65.9Age and gender-specific categories:Men20-29 39.9 38.6 37.0 43.130-39 49.6 58.1 52.6 58.140-49 53.6 63.6 60.3 65.550-59 54.1 58.4 60.8 73.060-69 52.9 55.6 57.4 70.370-79 36.0 52.7* 53.3* 63.180+ N/A** N/A** N/A** 50.6Women20-29 17.0 23.2 25.0 33.130-39 32.8 35.0 36.8 47.040-49 42.3 44.6 44.4 52.750-59 55.0 52.2 52.8 64.460-69 63.1 56.2 56.5 64.070-79 57.4 55.9* 58.2* 57.980+ N/A** N/A** N/A** 50.1COMBINED PREVALENCE OF OVERWEIGHT AND OBESITY (BMI≥25.0 KG/M2)AMONG ADULTS AGE 20 TO 80+ YEARS, BY GENDER, RACE/ETHNICITY, ANDAGE: UNITED STATES, 1960-1994.48* Prevalence for age 70 to 74 years ** Not available
9women are obese. 48The prevalence of obesity ismuch higher in minority women, being 36.7percent in non-Hispanic Black women and 33.3percent in Mexican-American women.2. Demographic Variations in Overweight andObesity PrevalenceAlthough NHANES III data show that theprevalence of overweight and obesity is muchhigher in African-American and Mexican-American women than in white women or inmen, these data provide ethnicity-specific esti-mates of overweight and obesity prevalence foronly three racial-ethnic groups: non-Hispanicwhites, non-Hispanic blacks, and Mexican-Americans. Examination survey data indicating ahigh overweight and obesity prevalence in otherethnic groups (e.g., for Puerto Ricans andCuban-Americans) are available from theHispanic HANES (HHANES) (1982-1984) 27and for American Indians 26and Pacific-IslanderAmericans, 50from smaller population-specificstudies (see Appendix III). The prevalence ofoverweight and obesity is generally higher formen and women in racial-ethnic minority popu-lations than in U.S. whites, with the exceptionof Asian-Americans, for whom overweight andobesity prevalence is lower than in the generalpopulation. 51In the 1982-1984 HHANES, theage-adjusted prevalence of a BMI of ≥ 27.3 inPuerto Rican women was 40 percent.27TheStrong Heart Study reported the average preva-lence of overweight using BMI ≥ 27.8 or ≥ 27.3for men and women, respectively, in threegroups of American Indians studied during1988-1989 as follows: in Arizona, 67 percent ofthe men and 80 percent of the women; inOklahoma, 67 percent of the men and 71 per-cent of the women; and in South Dakota andNorth Dakota, 54 percent of the men and 66percent of the women. 52Women in the United States with low incomesor low education are more likely to be obesethan those of higher socioeconomic status; theassociation of socioeconomic status with obesityis less consistent in men 53(Appendix III).Obesity is less common after the age of 70among both men and women, possibly due to aprogressive decrease in BMI with increasing agepast the fifth decade or to an excess in mortalityassociated with increasing BMI in the presenceof increasing age. 13. Economic Costs of Overweight and ObesityAlarm about the increasing prevalence of over-weight and obesity in the United States in recentyears 54, 55centers on the link between obesity andincreased health risks, 42, 56which translates intoincreased medical care and disability costs. 46,57The total cost attributable to obesity amountedto $99.2 billion in 1995. Approximately $51.6billion of these dollars were direct medical costsassociated with diseases attributable to obesity.The direct costs also associated with obesity rep-resent 5.7 percent of the national health expen-diture within the United States. 58The indirectcosts attributable to obesity are $47.6 billionand are comparable to the economic costs of cig-arette smoking. 58, 59Indirect costs represent thevalue of lost output caused by morbidity andmortality, and may have a greater impact thandirect costs at the personal and societal levels. 58Although a comprehensive cost analysis of obesi-ty is beyond the scope of this panel, a systematicreview of the literature identified studies estimat-ing the current economic burden of obesity inseveral Western countries. 57, 60-63Published esti-mates of the economic costs of obesity such asthose noted above use the prevalence-basedapproach, assuming that obesity is causally relat-ed to a range of chronic illnesses. Estimating theeconomic benefits of weight loss requires detailsof long-term weight maintenance and the timecourse of risk reduction following weight loss,and ultimately must also consider the costs oftreatment to reduce weight.
10Chapter 2: Overweight and Obesity: BackgroundP R E V A L E N C E O F O B E S I T Y ( B M I ≥ 3 0 . 0 K G / M 2) A M O N GA D U L T S A G E 2 0 T O 8 0 + Y E A R S , B Y G E N D E R ,R A C E / E T H N I C I T Y , A N D A G E : U N I T E D S T A T E S , 1 9 6 0 - 1 9 9 4 . 4 8NHES I NHANES I NHANES II HHANES NHANES IIIGender, race/ethnicity, age 20 1960-62 1971-74 1976-80 1982-84 1988-94years and older, age adjusted: (age 20-74) (age 20-74) (age 20-74) (age 20-74) (age ≥20)Both Sexes 12.8 14.1 14.4 22.3Men 10.4 11.8 12.2 19.5Women 15.1 16.1 16.3 25.0White men 10.1 11.4 12.0 20.0White women 13.7 14.7 14.9 23.5Black men 13.9 15.9 15.2 20.6Black women 25.0 28.6 30.2 36.5White, non-Hispanic men 12.0 19.9White, non-Hispanic women 14.8 22.7Black, non-Hispanic men 15.0 20.7Black, non-Hispanic women 30.0 36.7Mexican-American men 15.4 20.6Mexican-American women 25.4 33.3Age and gender-specific categories:Men20-29 9.0 8.0 8.1 12.530-39 10.4 13.3 12.1 17.240-49 11.9 14.2 16.4 23.150-59 13.4 15.3 14.3 28.960-69 7.7 10.3 13.5 24.870-79 8.6 11.1 13.6 20.080+ N/A N/A N/A 8.0Women20-29 6.1 8.2 9.0 14.630-39 12.1 15.1 16.8 25.840-49 17.1 17.6 18.1 26.950-59 20.4 22.0 22.6 35.660-69 27.2 24.0 22.0 29.870-79 21.9 21.0 19.4 25.080+ N/A N/A** N/A 15.1* *********** **TABLE II-2:* Prevalence for age 70 to 74 years ** Not available
11PREVENTION OF OVERWEIGHTAND OBESITYPrevention of overweight and obesity is asimportant as treatment. Prevention includes pri-mary prevention of overweight or obesity itself,secondary prevention or avoidance of weightregain following weight loss, and prevention offurther weight increases in obese individualsunable to lose weight. 44 , 64National and international observational datasuggest that environmental and behavioral fac-tors are likely to be important in the tendency ofindividuals within and between populations tobe obese during childhood or to gain weightprogressively with age during adulthood. 65These factors are also influenced by the geneticmakeup of individuals. There has been a paucityof intervention research to demonstrate howthese factors can be manipulated to prevent obe-sity. 64In two community studies, namely theMinnesota Heart Health Program and theStanford Five City Study, multifaceted weightloss and weight control programs within thecommunity were not associated with preventionof weight gain in longitudinally followed cohorts. 66In another community study, the PawtucketHeart Health Program, BMI levels did notchange in the intervention cities while theyincreased in the comparison cities. 67One obesi-ty prevention study of American Indian childrenwho are at high risk of becoming obese is underway. 68Otherwise, the only long-term reportsuggesting an effective approach to obesity pre-vention is from follow-up of obese children inan experimental study in which they had beentreated with or without a family-oriented treat-ment program. Long-term follow-up (10 years)of these children supported the importance offamily involvement in reducing the progressionof obesity. 69One population-based randomizedcontrolled pilot study of obesity prevention sug-gests that programs for weight gain preventionare feasible and effective in adults. 34Anotherstudy in China has shown that the prevention ofweight gain through diet, physical activity, andtheir combination can help prevent diabetes. 70It has been suggested that primary prevention ofobesity should include environmentally basedstrategies that address major societal contributorsto over-consumption of calories and inadequatephysical activity such as food marketing prac-tices, transportation patterns, and lack of oppor-tunities for physical activity during the workday. 71,72People at lower socioeconomic levels living inurban areas also lack access to physical activitysites. Such strategies will be essential for effectiveinitial and long-term prevention of obesity forlarge numbers of individuals and for the com-munity at large. Research is needed to clarify therole of societal policies, procedures, laws, andother factors that serve as disincentives to life-long caloric balance. The importance of obesityprevention needs to be brought to the attentionof health care payors and practitioners, employ-ers, educators, and public officials as an impor-tant priority to be addressed in policies, pro-grams, and direct services to individuals andfamilies. The development and implementationof appropriate policies and programs will requireoutcomes research that identifies effective weightgain prevention approaches. These programsmust be useful for multiple settings, includinghealth care facilities, schools, worksites, commu-nity and religious institutions, and be applicableto a broad population. In the end, efforts shouldbe made to make the general public more awareof the need to prevent overweight and obesity.Efforts to understand the genetic, developmen-tal, environmental, and behavioral underpin-nings of obesity and to mount successful preven-tion strategies are particularly critical for popula-tions in which overweight and obesity and relat-ed health problems such as diabetes are dispro-portionately prevalent; for example, women inlower socioeconomic groups and women andsometimes men in many racial/ethnic minorityB
12Chapter 2: Overweight and Obesity: Backgroundpopulations as described in Chapter 2.A.2 ofthis report. Public health approaches for pre-venting obesity, that is, approaches designed toreduce the difficulty for any given individual ofadopting healthful eating and activity patterns,will particularly benefit the socially disadvan-taged, who—compared to the more advan-taged— may have less access to preventive healthservices and fewer feasible options for makingchanges in their daily routines and lifestyles. 73-75Primary care practitioners are an important ele-ment in preventing and managing obesity in theUnited States. Prevention of overweight andobesity in primary care settings is compatiblewith efforts to prevent their health conse-quences, through control of dyslipidemia, highblood pressure, and type 2 diabetes. Thus, boththe quality and quantity of life may be enhancedthrough preventive strategies. As detailed else-where in this report, high blood pressure, highblood cholesterol, and type 2 diabetes should beaggressively treated in overweight patients andmay be treated prior to and in conjunction withweight loss.HEALTH RISKS OF OVERWEIGHTAND OBESITY1. MorbidityAbove a BMI of 20 kg/m2, morbidity for a num-ber of health conditions increases as BMI increas-es. Higher morbidity in association with over-weight and obesity has been observed for hyper-tension, 2-6, 76-80type 2 diabetes, 7, 8, 10, 81, 82, 84-89coro-nary heart disease (CHD), 11,42,86,88,90stroke, 11-13gallbladder disease, 14,15osteoarthritis, 16-18, 91-95sleep apnea and respiratory problems 21, 96-98andsome types of cancer (endometrial, breast,prostate, and colon). 107-115Obesity is also associ-ated with complications of pregnancy, menstrualirregularities, hirsutism, stress incontinence, andpsychological disorders (depression). 112, 116-128The nature of obesity-related health risks is simi-lar in all populations, although the specific levelof risk associated with a given level of over-weight or obesity may vary with race/ethnicity,and also with age, gender, and societal condi-tions. For example, the absolute risk of morbidi-ty in chronic conditions such as CHD is highestin the aged population, while the relative risk ofhaving CHD in obese versus nonobese individu-als is highest in the middle adult years. 129-131A high prevalence of diabetes mellitus in associa-tion with obesity is observed consistently acrossraces/ethnicities, while the relative prevalence ofhypertension and CHD in obese versusnonobese populations varies between groups.The health risks of overweight and obesity arebriefly described below:1.a.HypertensionData from NHANES III show that the age-adjusted prevalence of high blood pressureincreases progressively with higher levels of BMIin men and women (Figure 2). 2High bloodpressure is defined as mean systolic blood pres-sure ≥ 140 mm Hg, or mean diastolic bloodpressure ≥ 90 mm Hg, or currently taking anti-hypertensive medication. The prevalence of highblood pressure in adults with BMI ≥ 30 is 38.4percent for men and 32.2 percent for women,respectively, compared with 18.2 percent formen and 16.5 percent for women with BMI< 25, a relative risk of 2.1 and 1.9 for men andwomen, respectively. The direct and indepen-dent association between blood pressure andBMI or weight has been shown in numerouscross-sectional studies 3-5, including the largeinternational study of salt (INTERSALT) carriedout in more than 10,000 men and women. 6INTERSALT reported that a 10 kg (22 lb) high-er body weight is associated with 3.0 mm Hghigher systolic and 2.3 mm Hg higher diastolicblood pressure. 6These differences in blood pres-sure translate into an estimated 12 percentincreased risk for CHD and 24 percentC
13increased risk for stroke. 132Positive associationshave also been shown in prospective studies. 76-80Obesity and hypertension are co-morbid riskfactors for the development of cardiovasculardisease. The pathophysiology underlying thedevelopment of hypertension associated withobesity includes sodium retention and associatedincreases in vascular resistance, blood volume,and cardiac output. These cardiovascular abnor-malties associated with obesity are believed to berelated to a combination of increased sodiumretention, increased sympathetic nervous systemactivity, alterations of the renin-angiotensin sys-tem and insulin resistance. The precise mecha-nism whereby weight loss results in a decrease inblood pressure is unknown. However, it isknown that weight loss is associated with areduction in vascular resistance, total blood vol-ume and cardiac output, an improvement ininsulin resistance, a reduction in sympatheticnervous system activity, and suppression of theactivity of the renin angiotensin aldosteronesystem. 764-7691.b.Dyslipidemia, manifested by:High total cholesterolThe relationship of the age-adjusted prevalenceof high total cholesterol, defined as ≥ 240mg/dL (6.21 mmol/L), to BMI from NHANESIII is shown in Figure 3. 2At each BMI level, theprevalence of high blood cholesterol is greater inwomen than in men. In a smaller sample, higherbody weight is associated with higher levels oftotal serum cholesterol in both men 133andwomen 134at levels of BMI > 25. Several largelongitudinal studies also provide evidence thatoverweight, obesity and weight gain are associat-50Figure 2. NHANES III Age-Adjusted Prevalence of Hypertension*According to Body Mass Index40PercentBMI Levels302010Men0Women* Defined as mean systolic blood pressure ≥140 mm Hg, mean diastolicas ≥90 mm Hg, or currently taking antihypertensive medication.18.222.538.425.2BMI <25BMI 25-26BMI 27-29BMI ≥3016.521.924.032.2Source: Brown C. et al. Body Mass Index and the Prevalence of RiskFactors for Cardiovascular Disease (submitted for publication).
14Chapter 2: Overweight and Obesity: Backgrounded with increased cholesterol levels. 135-137Inwomen, the incidence of hypercholesterolemiaalso increases with increasing BMI. 138In addi-tion, the pattern of fat distribution appears toaffect cholesterol levels independently of totalweight. Total cholesterol levels are usually higherin persons with predominant abdominal obesity,defined as a waist-to-hip circumference ratio of≥ 0.8 for women and ≥ 1.0 for men. 139High triglyceridesThe strong association of triglyceride levels withBMI has been shown in both cross-sectional andlongitudinal studies, for both sexes and all agegroups. 133, 134, 140, 141In three adult age groups,namely 20 to 44 years, 45 to 59 years, and 60 to74 years, higher levels of BMI, ranging from 21or less to more than 30, have been associatedwith increasing triglyceride levels; the differencein triglycerides ranged from 61 to 65 mg/dL(0.68 to 0.74 mmol/L) in women 134and 62 to118 mg/dL (0.70 to 1.33 mmol/L) in men. 133Low high-density lipoprotein cholesterolThe age-adjusted prevalence of low high-densitylipoprotein (HDL)-cholesterol in relation toBMI levels, based on NHANES III data, isshown in Figure 4. 2HDL-cholesterol levels atall ages and weights are lower in men than inwomen. Although low HDL-cholesterol in thisstudy was defined as < 35 mg/dL (0.91 mmol/L)in men and < 45 mg/dL (1.16 mmol/L) inwomen 2, the panel accepts the definition of lowHDL-cholesterol as < 35 mg/dL for men andwomen used by the National CholesterolEducation Program’s Second Report of the ExpertPanel on the Detection, Evaluation, and Treatmentof High Blood Cholesterol in Adults (AdultTreatment Panel II Report). 142Cross-sectionalstudies have reported that HDL-cholesterol lev-els are lower in men and women with higherBMI. 143,144Longitudinal studies have found thatchanges in BMI are associated with changes inHDL-cholesterol. A BMI change of 1 unit isassociated with an HDL-cholesterol change of1.1 mg/dL for young adult men and an HDL-cholesterol change of 0.69 mg/dL for youngadult women. 145Normal to elevated low-density lipoproteincholesterolThe link between total serum cholesterol andCHD is largely due to low-density lipoprotein(LDL). A high-risk LDL-cholesterol is defined asa serum concentration of ≥ 160 mg/dL. Thislipoprotein is the predominant atherogeniclipoprotein and is therefore the primary target ofcholesterol-lowering therapy. Cross-sectionaldata suggest that LDL-cholesterol levels arehigher by 10 to 20 mg/dL in relation to a 10unit difference in BMI, from levels of 20 to 30kg/m2. 133,134According to extensive epidemiolog-ical data, a 10 mg/dL rise in LDL-cholesterolcorresponds to approximately a 10 percentincrease in CHD risk over a period of 5 to 10years. 146Small, dense low-density lipoprotein particlesFew large-scale epidemiological data are availableon small, dense LDL particles. 147-149Clinicalstudies have shown that small, dense LDL parti-cles are particularly atherogenic and tend to bepresent in greater proportion in hypertriglyceri-demic patients with insulin resistance syndromeassociated with abdominal obesity. 148-1521.c. Diabetes MellitusThe increased risk of diabetes as weight increaseshas been shown by prospective studies inNorway 7, the United States 8, Sweden 9, andIsrael. 10More recently, the Nurses’ HealthStudy, using data based on self-reported weights,found that the risk of developing type 2 diabetesincreases as BMI increases from a BMI as low as22. 81Since women in particular tend to under-report weight, the actual BMI values associatedwith these risks are likely to be higher than theNurses’ Health Study data would suggest. Anassociation between type 2 diabetes and increas-ing relative weight is also observed in popula-
15tions at high risk for obesity and diabetes, suchas in American Indians. 153,154In recent studies,the development of type 2 diabetes has beenfound to be associated with weight gain after age18 in both men 82and women. 81The relativerisk of diabetes increases by approximately 25percent for each additional unit of BMI over 22kg/m2. 83In addition, in a prospective study rep-resentative of the U.S. population, it was recent-ly estimated that 27 percent of new cases of dia-betes was attributable to weight gain in adult-hood of 5 kg (11 lb) or more.84Both cross-sec-tional 85-87and longitudinal studies 82,88,89showthat abdominal obesity is a major risk factor fortype 2 diabetes. 82,871.d.Coronary Heart DiseaseObservational studies have shown that over-weight, obesity, and excess abdominal fat aredirectly related to cardiovascular risk factors,including high levels of total cholesterol, LDL-cholesterol, triglycerides, blood pressure, fibrino-gen and insulin, 86and low levels of HDL-cho-lesterol. 42Plasminogen activator inhibitor-1causing impaired fibrinolytic activity is elevatedin persons with abdominal obesity. 763Over-weight, obesity, and abdominal fat are also asso-ciated with increased morbidity and mortalityfrom CHD. 11,42,155-161Recent studies have shown that the risks ofnonfatal myocardial infarction and CHD deathincrease with increasing levels of BMI. Risks arelowest in men and women with BMIs of 22 orless and increase with even modest elevations ofBMI. In the Nurses’ Health Study, which con-trolled for age, smoking, parental history ofCHD, menopausal status, and hormone use, rel-50Figure 3. NHANES III Age-Adjusted Prevalence of High Blood Cholesterol*According to Body Mass Index40PercentBMI Levels302010Men0Women*Defined as ≥240 mg/dLSource: Brown C. et al. Body Mass Index and the Prevalence of RiskFactors for Cardiovascular Disease (submitted for publication).14.717.520.220.4BMI <25BMI 25-26BMI 27-29BMI ≥3015.727.9 28.224.7
16Chapter 2: Overweight and Obesity: Backgroundative risks for CHD were twice as high at BMIsof 25 to 28.9, and more than three times as highat BMIs of 29 or greater, compared with BMIsof less than 21. 90Weight gains of 5 to 8 kg (11to 17.6 lb) increased CHD risk (nonfatalmyocardial infarction and CHD death) by 25percent, and weight gains of 20 kg (44 lb) ormore increased risk more than 2.5 times in com-parison with women whose weight was stablewithin a range of 5 kg (11 lb). 90In British men,CHD incidence increased at BMIs above 22 andan increase of 1 BMI unit was associated with a10 percent increase in the rate of coronaryevents. 162Similar relationships between increas-ing BMI and CHD risk have been shown inFinnish, Swedish, Japanese, and U.S. popula-tions. 90, 163, 164A relationship between obesity and CHD hasnot always been found. Two reasons mayaccount for this: the first is an inappropriatecontrolling for cholesterol, blood pressure, dia-betes, and other risk factors in statistical analysis;and the second is that there was not an adequatecontrol for the confounding effect of cigarettesmoking on weight. 88People who smoke oftenhave a lower body weight but more CHD.1.e.Congestive Heart FailureOverweight and obesity have been identified asimportant and independent risk factors for con-gestive heart failure (CHF) in a number of stud-ies, including the Framingham Heart Study. 11, 165-169CHF is a frequent complication of severe obe-sity and a major cause of death; duration of theobesity is a strong predictor of CHF. 170Since50Figure 4. NHANES III Age-Adjusted Prevalence of Low HDL-Cholesterol*According to Body Mass Index40PercentBMI Levels302010Men0Women*Defined as <35 mg/dL in men and <45 mg/dL in women.22.214.171.1243.1BMI <25BMI 25-26BMI 27-29BMI ≥3016.527.0 27.241.5Source: Brown C. et al. Body Mass Index and the Prevalence of RiskFactors for Cardiovascular Disease (submitted for publication).
17hypertension and type 2 diabetes are positivelyassociated with increasing weight, the coexis-tence of these conditions facilitates the develop-ment of CHF. 171Data from the Bogalusa HeartStudy demonstrate that excess weight may leadto acquisition of left ventricular mass beyondthat expected from normal growth. 171Obesitycan result in alterations in cardiac structure andfunction even in the absence of systemic hyper-tension or underlying heart disease. Ventriculardilatation and eccentric hypertrophy may resultfrom elevated total blood volume and high car-diac output. Diastolic dysfunction from eccen-tric hypertrophy and systolic dysfunction fromexcessive wall stress result in so-called “obesitycardiomyopathy”. 172, 173The sleep/apnea obesityhyperventilation syndrome occurs in 5 percentof severely obese individuals, and is potentiallylife-threatening. Extreme hypoxemia induced byobstructive sleep apnea syndrome may result inheart failure in the absence of cardiac dysfunc-tion. 1741.f. StrokeThe relationship of cerebrovascular disease toobesity and overweight has not been as wellstudied as the relationship to CHD. A reportfrom the Framingham Heart Study suggestedthat overweight might contribute to the risk ofstroke, independent of the known association ofhypertension and diabetes with stroke. 11Morerecently published reports 12, 13are based on larg-er samples and delineate the importance ofstroke subtypes in assessing these relationships.They also attempt to capture all stroke events,whether fatal or nonfatal. These studies suggestdistinct risk factors for ischemic stroke as com-pared to hemorrhagic stroke, and found over-weight to be associated with the former, but notthe latter. This may explain why studies that useonly fatal stroke outcomes (and thus overrepre-sent hemorrhagic strokes) show only weak rela-tionships between overweight and stroke. Theserecent prospective studies demonstrate that therisk of stroke shows a graded increase as BMIrises. For example, ischemic stroke risk is 75 per-cent higher in women with BMI > 27, and 137percent higher in women with a BMI > 32,compared with women having a BMI < 21. 121.g.GallstonesThe risk of gallstones increases with adultweight. Risk of either gallstones or cholecystec-tomy is as high as 20 per 1,000 women per yearwhen BMI is above 40, compared with 3 per1,000 among women with BMI < 24. 14According to NHANES III data, the prevalenceof gallstone disease among women increasedfrom 9.4 percent in the first quartile of BMI to25.5 percent in the fourth quartile of BMI.Among men, the prevalence of gallstone diseaseincreased from 4.6 percent in the first quartile ofBMI to 10.8 percent in the fourth quartile ofBMI. 151.h.OsteoarthritisIndividuals who are overweight or obese increasetheir risk for the development of osteoarthritis.16-18,91,92The association between increased weightand the risk for development of kneeosteoarthritis is stronger in women than in men. 92In a study of twin middle-aged women, it wasestimated that for every kilogram increase ofweight, the risk of developing osteoarthritisincreases by 9 to 13 percent. The twins withknee osteoarthritis were generally 3 to 5 kg (6.6to 11 lb) heavier than the co-twin with no dis-ease. 16An increase in weight is significantlyassociated with increased pain in weight-bearingjoints. 175There is no evidence that the develop-ment of osteoarthritis leads to the subsequentonset of obesity. 91A decrease in BMI of 2 unitsor more during a 10-year period decreased theodds for developing knee osteoarthritis by morethan 50 percent; weight gain was associated witha slight increase in risk. 93A randomized controlled trial of 6 months’duration examined the effect of weight loss on
18Chapter 2: Overweight and Obesity: Backgroundclinical improvement in patients withosteoarthritis. 176Patients taking phenterminehad an average weight loss of 12.6 percent after6 months while the control group had an aver-age weight loss of 9.2 percent. There wasimprovement in pain-free range of motion and adecrease in analgesic use in association withweight loss; patients with knee disease showed astronger association than those with hip disease.Similarly, improvement of joint pain wasobserved in individuals who had undergone gas-tric stapling, resulting in an average weight lossof 45 kg (99 lb). 94, 951.i. Sleep ApneaObesity, particularly upper body obesity, is a riskfactor for sleep apnea and has been shown to berelated to its severity. 19,20The major pathophysi-ologic consequences of severe sleep apneainclude arterial hypoxemia, recurrent arousalsfrom sleep, increased sympathetic tone, pul-monary and systemic hypertension, and cardiacarrhythmias. 21Most people with sleep apneahave a BMI > 30. 96,97Large neck girth in bothmen and women who snore is highly predictiveof sleep apnea. In general, men whose neck cir-cumference is 17 inches or greater and womenwhose neck circumference is 16 inches or greaterare at higher risk for sleep apnea. 98Additionalinformation on sleep apnea is included asAppendix IV.1.j. CancerColon CancerMany studies have found a positive relationbetween obesity and colon cancer in men but aweaker association in women. 8,22-24,99-106Morerecent data from the Nurses’ Health Study sug-gest that the relationship between obesity andcolon cancer in women may be similar to thatseen in men. Twice as many women with a BMIof > 29 kg/m2had distal colon cancer as womenwith a BMI < 21 kg/m2. 107In men, the relation-ship between obesity and total colon cancer wasweaker than that for distal colon cancer.Other data from the Nurses’ Health Study show asubstantially stronger relationship between waist-to-hip ratio and the prevalence of colon polypson sigmoidoscopy, than with BMI alone. 108Evenamong leaner women, a high waist-to-hip ratiois also associated with significantly increased riskof colon polyps. 107Breast CancerEpidemiologic studies consistently show thatobesity is directly related to mortality frombreast cancer, predominantly in postmenopausalwomen, 8but inversely related to the incidenceof premenopausal breast cancer. 109-112Ten ormore years after menopause, the premenopausal“benefit” of obesity has dissipated. 113Amongpostmenopausal women, peripheral fat is theprimary source of estrogens, the major modifi-able risk factor for postmenopausal breast cancer.This crossover in the relationship of obesity withbreast cancer, pre- and postmenopausally, com-plicates prevention messages for this commonfemale cancer. Recent data from the Nurses’Health Study, however, show that adult weightgain is positively related to risk of post-menopausal breast cancer. This relation is seenmost clearly among women who do not usepostmenopausal hormones. A gain of more than20 lb from age 18 to midlife doubles a woman’srisk of breast cancer. Even modest weight gainsare positively related to risk of postmenopausalcancer. 114Endometrial CancerObesity increases the risk of endometrial cancer.The risk is three times higher among obesewomen (BMI ≥ 30 kg/m2) compared to normal-weight women. 115However, the absolute risk ofthis condition is low when compared to breastcancer, heart disease, and diabetes. Adult weightgain is also related to increased risk. 115
19Gallbladder CancerObesity is related to the risk of gallbladder can-cer, particularly among women. 177Using aweight index of 100 as the average weight with acorresponding mortality ratio of 1.0 for thecohort, mortality ratios were 1.16 at a weightindex of 120 to 129, 1.22 at 130 to 139, and1.53 at ≥140.1.k.Obesity and Women’s Reproductive HealthMenstrual Function and FertilityObesity in premenopausal women is associatedwith menstrual irregularity and amenorrhea. 112,116As part of the Nurses’ Health Study, a case con-trol study suggested that the greater the BMI atage 18 years, even at levels lower than those con-sidered obese, the greater the risk of subsequentovulatory infertility. 117The most prominentcondition associated with abdominal obesity ispolycystic ovarian syndrome, 118a combinationof infertility, menstrual disturbances, hirsutism,abdominal hyperandrogenism, and anovulation.This syndrome is strongly associated with hyper-insulinemia and insulin resistance. 119PregnancyPregnancy can result in excessive weight gainand retention. The 1988 National Maternal andInfant Survey observed that 41.6 percent ofwomen reported retaining ≥ 9 lb of their gainedweight during pregnancy, with 33.8 percentreporting ≥14 lb of retained weight gain. 120Theretained weight gain associated with pregnancywas corroborated by the study of CoronaryArtery Risk Development in Young Adults(CARDIA). As a result of their first pregnancy,both black and white young women had a sus-tained weight gain of 2 to 3 kg (4.4 to 6.6 lb) ofbody weight. 121Another study on a nationalcohort of women followed for 10 years reportedthat weight gain associated with childbearingranged from 1.7 kg (3.7 lb) for those having onelive birth during the study to 2.2 kg (4.9 lb) forthose having three. 178In addition, higherprepregnancy weights have been shown toincrease the risk of late fetal deaths. 179Obesity during pregnancy is associated withincreased morbidity for both the mother and thechild. A tenfold increase in the prevalence ofhypertension and a 10 percent incidence of ges-tational diabetes have been reported in obesepregnant women. 122Obesity also is associatedwith difficulties in managing labor and delivery,leading to a higher rate of induction and prima-ry Caesarean section. Risks associated with anes-thesia are higher in obese women, as there isgreater tendency toward hypoxemia and greatertechnical difficulty in administering local or gen-eral anesthesia. 123Finally, obesity during preg-nancy is associated with an increased risk of con-genital malformations, particularly of neuraltube defects. 123A certain amount of weight gain during preg-nancy is desirable. The fetus itself, expandedblood volume, uterine enlargement, breast tissuegrowth, and other products of conception gener-ate an estimated 13 to 17 lb of extra weight.Weight gain beyond this, however, is predomi-nantly maternal adipose tissue. It is this fat tissuethat, in large measure, accounts for the postpar-tum retention of weight gained during pregnan-cy. In turn, this retention reflects a postpartumenergy balance that does not lead to catabolismof the gained adipose tissue. In part, this mayreflect reduced energy expenditure throughdecreased physical activity, even while caring foryoung children, but it may also reflect retentionof the pattern of increased caloric intakeacquired during pregnancy. 180One difficulty in developing recommendationsof optimal weight gain during pregnancy relatesto the health of the infants. A balance must beachieved between high-birth-weight infants whomay pose problems during delivery and whomay face a higher rate of Caesarean sections andlow-birth-weight infants who face a higher
20Chapter 2: Overweight and Obesity: Backgroundinfant mortality rate. 181However, data from thePregnancy Nutrition Surveillance System fromthe CDC showed that very overweight womenwould benefit from a reduced weight gain dur-ing pregnancy to help reduce the risk for high-birth-weight infants. 181The 1990 Institute of Medicine report maderecommendations concerning maternal weightgain. 182It recommended that each woman haveher BMI measured and recorded at the time ofentry into prenatal care. For women with a BMIof less than 20, the target weight gain should be0.5 kg (1.1 lb) of weight gain per week duringthe second and third trimester. For a womanwhose BMI is greater than 26, the weight gaintarget is 0.3 kg (0.7 lb) per week during the lasttwo trimesters.Women who are overweight or obese at theonset of pregnancy are advised to gain less totalweight during the pregnancy (see box above). 1821.l. Psychosocial Aspects of Overweight andObesityA number of reviews have been published on thepsychosocial aspects of obesity. 124-128,183The spe-cific topics that will be reviewed here includesocial stigmatization, psychopathology, bingeeating, and body image perceptions.Social stigmatizationIn American and other Westernized societiesthere are powerful messages that people, espe-cially women, should be thin, and that to be fatis a sign of poor self-control. 125, 126, 128, 184, 185Negative attitudes about the obese have beenreported in children and adults, 186-191in healthcare professionals, 192-194and in the overweightthemselves. 195, 196People’s negative attitudes toward the obeseoften translate into discrimination in employ-ment opportunities, 197-199college acceptance, 200less financial aid from their parents in paying forcollege, 195, 201job earnings, 202rental availabilities,203and opportunities for marriage. 204Much of the research on the social stigma ofobesity has suffered from methodological limita-tions. For example, a number of the early studiesrelied on line drawings rather than more lifelikerepresentations of obese people and on checkliststhat forced one to make YES or NO choices.More importantly, there has been a lack ofresearch that has looked at the impact of obesityin the context of other variables, such as physicalattractiveness, the situational context, and thedegree of obesity. 184,185In addition, social stigmatoward the obese has primarily been assessedamong white individuals. There is some evi-dence that members of other racial and ethnicgroups are less harsh in their evaluation of obesepersons. One study assessed 213 Puerto Ricanimmigrants to the United States, and found awide range of acceptable weights among them. 205Crandall found that Mexican students were sig-W E I G H T G A I N D U R I N G P R E G N A N C YBMI Kilograms Pounds<19.8 12.5 to 18 28 to 4019.8 to 26 11.5 to 16 25 to 35>26 to 29 7 to 11.5 15 to 25>29 ≤6 ≤13
21nificantly less concerned about their own weightand were more accepting of other obese peoplethan were U.S. students. 206In addition, thedegree of acceptance of obesity among people oflower education and income has not been wellstudied. Thus, these data are very incompletewith respect to racial and ethnic groups otherthan whites.Psychopathology and ObesityResearch relating obesity to psychological disor-ders and emotional distress is based on commu-nity studies and clinical studies of patients seek-ing treatment. In general, community-basedstudies in the United States have not found sig-nificant differences in psychological statusbetween the obese and nonobese. 126,183,207,208However, several recent European studies in gen-eral populations do suggest a relationshipbetween obesity and emotional problems. 209-211Thus, it may be premature to state that there isno association between obesity and psy-chopathology or emotional distress in the gener-al population. More focused, hypothesis-driven,and long-term studies are needed.127, 212Overweight people seeking weight loss treatmentmay, in clinic settings, show emotional distur-bances. 213-215In a review of dieting and depression,there was a high incidence of emotional illnesssymptoms in outpatients treated for obesity. 213However, several factors influenced these emo-tional responses, including childhood onset ver-sus adult onset of obesity (those with childhoodonset obesity appear more vulnerable). Anotherstudy that compared different eating disordergroups found that obese patients seeking treat-ment showed considerable psychopathology,most prominently mild to severe depression. 214Sixty-two percent of the obese group seekingtreatment showed clinically significant elevationson the depression subscale of the MinnesotaMultiphasic Personality Inventory, and 37 per-cent of this same group showed a score of 20 orhigher (indicating clinical depression) on theBeck Depression Inventory. Focusing on depres-sion was considered an important component ofthe weight loss program. Another study com-pared obese people who had not sought treat-ment to an obese group that had sought treat-ment in a professional, hospital-based program,and to normal weight controls. 215Again, obeseindividuals seeking treatment reported more psy-chopathology and binge eating compared to theother groups. Both obese groups reported moresymptoms of distress than did normal weightcontrols. The authors suggest that the obesepopulation is not a homogenous group, andthus, may not respond in the same way to stan-dardized treatment programs. In particular,obese individuals seeking treatment in clinic set-tings are more likely than obese individuals notseeking treatment and normal controls to reportmore psychopathology and binge eating.Binge Eating DisorderBinge eating disorder (BED) is characterized byeating larger amounts of food than most peoplewould eat in a discrete time period (e.g., 2hours) with a sense of lack of control duringthese episodes.762It is estimated to occur in 20 to50 percent of individuals who seek specializedobesity treatment. 216 -218In community-basedsamples, the prevalence is estimated to beapproximately 2 percent. 219Comparisons havebeen made between BED and bulimia nervosa(BN), an eating disorder characterized by recur-rent and persistent binge eating, accompanied bythe regular use of behaviors such as vomiting,fasting, or using laxatives. Studies comparingnormal weight individuals who have BN withobese BED individuals have found that obesebinge eaters are less likely to demonstrate dietaryrestraint and show few if any adverse reactionsto moderate or severe dieting. Most obese bingeeaters do not engage in inappropriate compen-satory behaviors such as purging. 220Comparedwith BN, the demographic distribution of BED
22Chapter 2: Overweight and Obesity: Backgroundis broader with respect to age, gender, and race218, 219, 221-225while data suggest that BED is ascommon in African-American women as inwhite women. 226The difference between BEDand BN is dramatic regarding gender. Very fewmen have BN, 227whereas the distribution isclose to equal in BED. 225, 228, 229Compared to obese nonbingers, obese individu-als with BED tend to be heavier, 230reportgreater psychological distress, and are more likelyto have experienced a psychiatric illness (espe-cially affective disorders). 225,231-236They alsoreport an earlier onset of obesity and a greaterpercentage of their lifetime on a diet. 237, 238Somestudies have shown histories of greater weightfluctuation or weight cycling in obese bingeeaters compared with nonbingers, 219, 237, 238butothers have not. 239These individuals are alsomore likely than nonbinging obese people todrop out of behavioral weight loss programs, 233and to regain weight more quickly. 220,233,240Critics of behavioral treatment of obesity haveargued that caloric restriction may cause or con-tribute to the episodes of binge eating and BN. 241Three studies have tested this hypothesis. 218, 242,243Neither moderate nor severe caloric restrictionexacerbated binge eating. All three studiesshowed that weight control treatment featuringcaloric restriction significantly reduced the fre-quency of binge eating in these patients.Body ImageBody image is defined as the perception of one’sown body size and appearance and the emotion-al response to this perception. 183,244Inaccurateperception of body size or proportion and nega-tive emotional reactions to size perceptions con-tribute to poor body image. Obese individuals,especially women, tend to overestimate theirbody size. 245-249People at greater risk for a poor body image arebinge eaters, women, those who were obese dur-ing adolescence or with early onset of obesity,and those with emotional disturbances. 127, 235, 244,250-253It is no surprise, then, that in some groupsof obese persons, these individuals are more dis-satisfied and preoccupied with their physicalappearance, and avoid more social situations dueto their appearance. 254, 255Body image dissatisfac-tion and the desire to improve physical appear-ance often drives individuals to seek weight loss.However, obese persons seeking weight reduc-tion must come to terms with real limits in theirbiological and behavioral capacities to loseweight. Otherwise, weight loss attempts mayonly intensify the sense of failure and strugglethat is already present among many obese indi-viduals. For this reason, psychosocial interven-tions which incorporate strategies to improvebody image may be helpful for those who wantto lose weight and are very concerned abouttheir physical appearance. A review of bodyimage interventions in obese persons can befound in Rosen (1996). 256Body image perceptions of individuals in variousethnic and racial groups may be different, onaverage, from those of the mainstream culture.There may be a similar range of attitudes but ona different scale; for example, it may take amuch greater degree of overweight to elicit nega-tive reactions. 257Differences in body image andweight-related concerns between black and whitegirls and women have been observed. 257In gen-eral, black girls and women report: less socialpressure to be slim, 258-260fewer incidences ofweight-related discrimination, 261less weight andbody dissatisfaction, and greater acceptance ofoverweight than their white counterparts. 259,262-266College-age black women report less concernand fear about fatness, less drive to be thin, andless concern about dieting than do college-agewhite women. 267In addition, black women mayascribe some positive qualities to being large,such as having stamina, strength, and solidity,and are less likely to link body size to health
23than white women. Black elementary school andhigh school girls were more likely to be trying togain weight 268, 269and less likely to be trying tolose weight as compared to white girls. 269Because of the above, it is possible that weightcontrol initiatives may elicit different reactionsfrom black and white women. Less is knownabout the relationship between obesity and bodyimage disturbance in other racial and ethnicgroups. 2702. Overweight/Obesity and Morbidity inMinority PopulationsThe data on overweight and obesity in minoritypopulations include men and women across awide age range and geographic area. Relevantstudies increasingly consist of well-designed,population-based surveys and longitudinal stud-ies. These studies have standardized, objectivemeasurements of overweight and obesity andrisk factors or disease outcomes. There is now awealth of evidence to demonstrate that over-weight and obesity incidence (both generalizedand abdominal) predisposes to chronic diseasesin racial/ethnic minority populations as it doesin whites, though the absolute risk may differ. 51-52, 271-284Indications for treatment of overweight and obe-sity in minority populations are, therefore, thesame as those for non-Hispanic whites. Apparentdifferences in the strength of association betweenobesity and disease in various populations arenot necessarily relevant to individuals in clinicalsettings, and obesity should be treated in any sit-uation in which excess weight is associated withan observable or probable risk of morbidity. Inaddition, from a public health perspective, theneed for obesity prevention and treatment isparticularly pressing in racial/ethnic minoritypopulations because of the high proportion ofoverweight and obese persons in many suchpopulations.3. Obesity and MortalityAs stated in the introduction to the guidelines,in the majority of epidemiologic studies, mortal-ity begins to increase with BMIs above 25kg/m2. 28-32The increase in mortality generallytends to be modest until a BMI of 30 kg/m2isreached. 28, 29, 31, 32For persons with a BMI of 30kg/m2or above, mortality rates from all causes,and especially from cardiovascular disease, aregenerally increased by 50 to 100 percent abovethat of persons with BMIs in the range of 20 to25 kg/m2.28,31,32Three aspects of the associationbetween obesity and mortality remain unre-solved:3.a.Association of Body Mass Index WithMortalityMany of the observational epidemiologic studiesof BMI and mortality have reported a ‘U-’ or‘J-shaped’ relationship between BMI and mortal-ity. 28Mortality rates are elevated in persons withlow BMI (usually below 20) as well as in personswith high BMI. 28, 31, 32In some studies, adjust-ment for factors that potentially confound therelationship between BMI and mortality, such assmoking status and pre-existing illness, tends toreduce the upturn in mortality rate at low BMI,31but in a meta-analysis the higher mortality at lowBMIs was not eliminated after adjustment forconfounding factors. 32It is unclear whether theelevated mortality observed at low BMI is due toan artifact of incomplete control for confoundingfactors, 285inadequate body fat and/or inadequatebody protein stores that result from unintention-al weight loss, 286or individual genetic factors.Currently, there is no evidence that intentionalweight gain in persons with low BMIs will leadto a reduction in mortality.3.b.Association of Body Mass Index WithMortality in Older AdultsMany of the observational epidemiologic studiessuggest that the relationship between BMI andmortality weakens with increasing age, especiallyamong persons aged 75 and above. 287- 290Several
24Chapter 2: Overweight and Obesity: Backgroundfactors have been proposed to explain this obser-vation. Older adults are more likely thanyounger adults to have diseases that bothincrease mortality and cause weight loss leadingto lower body weight. 291-293In addition, as peo-ple age, they tend to have larger waist circumfer-ences that increase their risk of mortality even atlower BMIs.294Also, weight in middle age ispositively related to risk of mortality in old age.292The impact of smoking on body weight andmortality is likely to be much stronger in olderadults because of the cumulative health effects ofsmoking. 295BMI, which is an indirect estimate of adiposity,may underestimate adiposity in older adultswhose BMI is similar to younger adults. 296It isalso possible that persons most sensitive to theadverse health effects of obesity are more likelyto have died before reaching older ages, resultingin older cohorts that are more “resistant” to thehealth effects of obesity. Recently, a 20-yearprospective study of a nationally representativesample of U.S. adults aged 55 to 74 years sug-gested that lowest mortality occurs in the BMIrange of 25 to 30. 297 298After adjusting for smok-ing status and pre-existing illness, lowest mortal-ity occurred at a BMI of 24.5 in white men,26.5 in white women, 27.0 in black men, and29.8 in black women (see commentary on olderadults on pages 90-91).3.c. Association of Body Mass Index WithMortality in Ethnic MinoritiesThe levels of BMI associated with increasedmortality are based on epidemiological studies ofprimarily white populations. The interest in con-firming the association between BMI and mor-tality in other racial/ethnic groups stems partlyfrom observations that lower-than-average totalmortality has been observed among some popu-lations with a high BMI level, 299and partly fromobservations that within certain populationsthere appears to be no effect of obesity at all orat the BMI levels that are associated with highermortality in whites.African-Americans:Three small cohort studies of narrowly definedpopulations of African-Americans failed to showthe expected association of BMI and mortalitybased on data from white populations. 300-303Although the shape of the association of BMIand mortality in two large, representative U.S.data sets (the National Health and NutritionExamination Follow-up Study and the NationalHealth Interview Survey) is similar for black andwhite males and females, 304the BMI-relatedincrease in risk begins at a 1 to 3 kg/m2higherBMI level for blacks than for whites. For exam-ple, in the National Health and NutritionExamination Follow-up Survey, the estimatedBMI associated with minimum mortality was27.1 for black men and 26.8 for black women,compared with 24.8 and 24.3, respectively, forwhite men and women. On the basis of thesedata, the use of the cutpoint of BMI ≥ 30 kg/m2for defining obesity is clearly applicable toAfrican-Americans as well as to whites.Other Ethnic Minority Populations:Limited data relating obesity to mortality inAmerican Indians were identified, but no datawere found relating obesity to mortality inHispanic-Americans, Asian-Americans, or PacificIslanders. 305The lowest mortality rate amongPima men is observed at a BMI range of 35 to40 kg/m2for men, and no relationship betweenBMI and mortality is observed among Pimawomen. 306, 307Based on mortality data alone, itwould be hard to justify using the same standardfor defining obesity in populations, such asAmerican Indians, among whom the mean BMIis much higher than in the general U.S. popula-tion. However, diabetes-related morbidityamong obese American Indians is extremelyhigh, 154and the overall age-specific mortalityamong American Indians is generally higher
25than in the U.S. general population. 306,307Thus,obesity in American Indians is associated with acompromised overall survival of the population.Although the data on mortality are still fragmen-tary for many minority populations, there are nostudies that would support the exclusion of anyracial/ethnic group from the current definitionsof obesity. Secular trends in many populations inthe United States and throughout the worldhave demonstrated that longstanding overweightand obesity eventually leads to the emergence ofchronic diseases. Therefore, prevalent overweightand obesity cannot be ignored even where theassociated health problems have not reached thelevel that would be expected on the basis of datafor white populations.WEIGHT LOSS AND MORTALITYA number of studies of “generic weight loss”(cause of weight loss unknown), “weight cycling”(cycles of weight loss followed by weight regain),and mortality have been published. 308- 311Inmost, 308,309,311but not all, 310of these studies,generic weight loss and weight cycling are associ-ated with increases in mortality. None of thesestudies, however, differentiated between inten-tional and unintentional weight loss. 311Withthe exception of the studies below, very little iscurrently known about factors related to inten-tional and unintentional weight loss in the gen-eral population or about the relationshipbetween weight loss intention and mortality. 312Two studies of factors related to weight lossintention have been carried out in the generalpopulation: French and colleagues assessed cor-relates of intentional and unintentional weightloss of > 20 lb in the Iowa Women’s HealthStudy, a cohort study of approximately 29,000women with a mean age of about 65 years; 313-315and Meltzer and Everhart analyzed data on 1-year self-reported weight change from approxi-mately 9,000 participants in the nationally rep-resentative U.S. National Health InterviewSurvey, aged 45 years and above. 316The resultsof these two studies suggest the following:s In cross-sectional studies of weight loss recall,heavier persons are more likely to reportintentional weight loss than unintentionalweight loss, while the reverse is true for leanerpersons;s Intentional and unintentional weight lossesoccur with similar frequency in the U.S. pop-ulation and contribute similarly to long-termweight fluctuation;s The frequency of intentional weight loss islower at older ages, while the frequency ofunintentional weight loss is higher at olderages; ands Unintentional weight loss occurs more oftenin persons who report that their health statusis poor, who use medications for chronichealth conditions, and who smoke. 312To date, only three studies have examined therelationship between intentional weight loss andmortality. Singh and colleagues 317publishedresults from a 1-year randomized controlled trialof a “cardioprotective diet” in East Indianpatients hospitalized with recent myocardialinfarction (mean age 50 years, mean BMI about24 kg/m2). Although this study was notdesigned to specifically test the efficacy of inten-tional weight loss on lower mortality, theauthors found that those who lost at least 0.5 kg(1.1 lb) had a 50 percent lower incidence of car-diac events and a 54 percent lower risk of overallmortality compared with counterparts who lost< 0.5 kg (1.1 lb). 317Williamson and colleagues 318published a 12-year prospective observational study of weightloss and mortality that directly assessed weightloss intention. They analyzed data from 43,457overweight (BMI > 27), never-smoking, whitewomen ages 40 to 64 years. Mortality ratioswere compared for women who intentionallyD
26Chapter 2: Overweight and Obesity: Backgroundlost weight with those for women who had nochange in weight. In women with obesity-relatedcomorbidities, intentional weight loss of anyamount was associated with a statistically signifi-cant 20 percent reduction in all-cause mortality,primarily due to a significant 40 to 50 percentreduction in mortality from obesity-related can-cers; diabetes-related mortality was also signifi-cantly reduced by 30 to 40 percent in those whointentionally lost weight. In women with nocomorbidities, intentional weight loss was gener-ally unrelated to mortality; however, after subdi-viding intentional weight loss by time interval, itwas found that a loss of at least 20 lb thatoccurred within the previous year was associatedwith small to modest increases in mortality. Theauthors concluded that the association betweenintentional weight loss and longevity in middle-age overweight women depends on health status.In addition, preliminary evidence suggests thatintentional weight loss in middle-age overweightmen may be associated with a similar reductionof diabetes-related mortality as was observed inthe overweight women. 319The ongoing Swedish Obesity Study is a con-trolled trial of surgically induced weight loss andsubsequent morbidity and mortality over a 10-year follow-up period (1,006 participants aged37 to 57 years; initial BMI of 34 in men and 38in women 320, 321). Although the study is not ran-domized (participants self-select for surgery), thecontrols (who receive a behavioral weight lossprogram) are computer-matched to surgical par-ticipants on a large number of potential con-founders including weight. 320In a preliminaryabstract,322the study reported the 2-year inci-dence rates shown in Table II-3.These results for disease and risk factor inci-dences suggest that 10-year mortality will ulti-mately be lower in the surgical interventiongroup. Definitive mortality results have not beenreported to date.ENVIRONMENTThe environment is a major determinant ofoverweight and obesity. Environmental influ-ences on overweight and obesity are primarilyrelated to food intake and physical activitybehaviors 71. In countries like the United States,there is an overall abundance of palatable, calo-rie-dense food. In addition, aggressive andsophisticated food marketing in the mass media,supermarkets, and restaurants, and the large por-tions of food served outside the home, promotehigh calorie consumption. Many of our socio-cultural traditions promote overeating and thepreferential consumption of high calorie foods.For many people, even when caloric intake isT W O - Y E A R I N C I D E N C E R A T E SControl SurgeryDiabetes 16 percent 0.5 percentHypertriglyceridemia 23 percent 6.0 percentLow High-Density 16 percent 5.0 percentLipoprotein LevelsETABLE II-3:
27not above the recommended level, the numberof calories expended in physical activity is insuf-ficient to offset consumption. Mechanizationlimits the necessity of physical activity requiredto function in society. Many people areentrenched in sedentary daily routines consistingof sitting at work, sitting in traffic, and sitting infront of a television or a computer monitor formost of their waking hours.In this obesity-promoting environment, individ-ual attitudes and behaviors are critical in weightmanagement. Many individuals may needextended treatment in clinical or communitysettings to enable them to cope with the com-plexities of long-term weight management, espe-cially if there is a history of unsuccessfulattempts at self-treatment. 44When the typicaldaily routine is so strongly biased towards pro-moting and perpetuating overweight and obesi-ty, very high levels of knowledge, motivation,personal behavioral management skill, andlifestyle flexibility are required for an overweightor obesity-prone individual to avoid becomingoverweight, or progressing to moderate or severeobesity.Although there are undoubtedly some inter- andintrapopulation variations in the genetic predis-position to become overweight or obese, severallines of evidence suggest that genetic factorsalone cannot explain the demographic and eth-nic variations in overweight and obesity preva-lence. For example, there is a difference in obesi-ty prevalence among low- and high-incomewhite women in industrialized societies. 323,324Other studies of populations, including migra-tion studies, have shown an increase in averagebody weight in those who move from a tradi-tional to a Westernized environment. 325-328Culturally determined attitudes about food,physical activity, and factors that vary withincome, education, and occupation may increasethe level of difficulty in weight management.Body image concerns and other motivations foravoiding obesity or controlling weight withingiven limits also vary with ethnic background,age, socioeconomic status, and gender. Thus, thecompetence of practitioners in working withdiverse sociocultural perspectives can be a criticalfactor in the success of obesity treatment. 257Fora discussion of cultural issues in obesity treat-ment and related references, see Appendix III.GENETIC INFLUENCE IN THE DEVELOPMENTOF OVERWEIGHT AND OBESITYObesity is a complex multifactorial chronic dis-ease developing from interactive influences ofnumerous factors—social, behavioral, physiolog-ical, metabolic, cellular, and molecular. Geneticinfluences are difficult to elucidate and identifi-cation of the genes is not easily achieved infamilial or pedigree studies. Furthermore, what-ever the influence the genotype has on the etiol-ogy of obesity, it is generally attenuated or exac-erbated by nongenetic factors.A large number of twin, adoption, and familystudies have explored the level of heritability ofobesity; that is, the fraction of the populationvariation in a trait (e.g., BMI) that can beexplained by genetic transmission. Recent stud-ies of individuals with a wide range of BMIs,together with information obtained on their par-ents, siblings, and spouses, suggest that about 25to 40 percent of the individual differences inbody mass or body fat may depend on geneticfactors. 329-331However, studies with identicaltwins reared apart suggest that the genetic con-tribution to BMI may be higher, i.e., about 70percent. 332There are several other studies ofmonozygotic twins reared apart that yieldedremarkably consistent results. 333Some of thereasons behind the different results obtainedfrom twin versus family studies have beenreported. 334-336The relative risk of obesity forfirst-degree relatives of overweight, moderatelyobese, or severely obese persons in comparisonto the population prevalence of the conditionF
28Chapter 2: Overweight and Obesity: Backgroundreaches about 2 for overweight, 3 to 4 for mod-erate obesity, and 5 and more for more severeobesity. 337,338Support for a role of specific genes in humanobesity or body fat content has been obtainedfrom studies of Mendelian disorders with obesityas one of the clinical features, single-gene rodentmodels, quantitative trait loci from crossbreed-ing experiments, association studies, and linkagestudies. From the research currently available,several genes seem to have the capacity to causeobesity or to increase the likelihood of becomingobese. 339The rodent obesity gene for leptin, anatural appetite-suppressant hormone, has beencloned 340as has been its receptor. 341In addition,other single gene mutants have been cloned. 341, 342However, their relationship to human diseasehas not been established, except for one studydescribing two subjects with a leptin mutation. 342This suggests that for most cases of human obe-sity, susceptibility genotypes may result fromvariations of several genes.Severely or morbidly obese persons are, on theaverage, about 10 to 12 BMI units heavier thantheir parents and siblings. Several studies havereported that a single major gene for high bodymass was transmitted from the parents to theirchildren. The trend implies that a major reces-sive gene, accounting for about 20 to 25 percentof the variance, is influenced by age and has afrequency of about 0.2 to 0.3. 343However, nogene(s) has (have) yet been identified. Evidencefrom several studies has shown that some per-sons are more susceptible to either weight gainor weight loss than others. 344,345It is importantfor the practitioner to recognize that the phe-nomenon of weight gain cannot always beattributed to lack of adherence to prescribedtreatment regimens.
293WHY TREAT OVERWEIGHTAND OBESITY?The recommendation to treat overweight andobesity is based not only on the previously pre-sented evidence that shows overweight is associ-ated with increased morbidity and mortality, butalso on randomized controlled trial (RCT) evi-dence that weight loss reduces risk factors fordisease. Thus, weight loss may help control dis-eases worsened by overweight and obesity andmay also decrease the likelihood of developingthese diseases. The evidence meeting the panel’sinclusion criteria is presented in the followingsection in the form of evidence statements with acorresponding rationale for each statement. Thedetails of the studies reviewed but not discussedhere are provided in Appendix II.Many of the RCTs examined by the panelincluded the use of pharmacotherapy for weightloss. When the panel began its deliberations in1995, two weight loss drugs, fenfluramine anddexfenfluramine, were being used widely forlong-term weight loss, i.e., 14 million prescrip-tions were given out over 11⁄2 years. However, asof September 1997, the Food and DrugAdministration (FDA) requested the voluntarywithdrawal of these drugs from the market due toa reported association between valvular heart dis-ease and the use of dexfenfluramine and fenflu-ramine alone or combined with phentermine. 761In November 1997, the FDA approved the drugsibutramine hydrochloride monohydrate for themanagement of obesity, including weight lossand maintenance of weight loss when used inconjunction with a reduced-calorie diet. Thus, atthe present time only one weight loss drug isavailable for long-term weight loss. (Note: FDAapproval for orlistat is pending a resolution oflabeling issues and results of Phase III trials.)However, due to rapidly evolving informationregarding the use of pharmacotherapy for weightloss, the panel decided to present (below) theircritique of those pharmacotherapy trials meetingtheir criteria for consideration.1. Blood PressureTo evaluate the effect of weight loss on bloodpressure and hypertension, 76 articles reportingthe results of RCTs were potentially eligible forinclusion. Sixty articles included lifestyle trialsthat studied diet and/or physical activity, and 16articles were of pharmacotherapy trials. Dietaryinterventions included low-calorie diets and dietsthat promoted macronutrient compositionchanges, such as amount and type of dietary fat.Physical activity, when included, was used tohelp promote increased energy expenditure.These trials did not always control for otherdietary factors that lower blood pressure, such asdietary salt, and the degree to which those takingthe blood pressure measurements were blinded tothe patient’s change in weight.Of the 35 lifestyle RCT articles deemed accept-able, 16 included hypertensive patients, 346-361and19 were conducted in individuals with normal orhigh-normal blood pressure. 362-380EXAMINATION OF RANDOMIZEDCONTROLLED TRIAL EVIDENCEA
30Chapter 3: Examination of Randomized Controlled Trial Evidence1.a.(1). Lifestyle Trials in Hypertensive PatientsEvidence Statement: Weight loss pro-duced by lifestyle modifications reducesblood pressure in overweight hypertensivepatients. Evidence Category A.Rationale: A 1987 meta-analysis 381covering fiveof the acceptable studies 351, 352, 356-358in hyperten-sive patients concluded that weight loss accom-plished by dietary interventions significantlylowered blood pressure. In hypertensive patients,10 kg (22 lb) of weight loss was associated withan average reduction of 7 mm Hg systolic and 3mm Hg diastolic blood pressure compared withcontrols. 381Since publication of this meta-analysis in 1987,almost all relevant studies have reported thatweight loss reduces blood pressure or the needfor medication in hypertensive patients: 346- 348, 350,352-361s The Trial of Antihypertensive Interventionsand Management (TAIM), conducted inhypertensive individuals not taking medica-tion for 6 months, reported that, comparedwith controls, a mean net weight reduction of4.7 kg (10.4 lb) reduced systolic and diastolicblood pressure by 2.8 mm Hg and 2.5 mmHg, respectively; 355the effects on blood pres-sure were equivalent to drug therapy amongthose participants who lost 4.5 kg (9.9 lb) ormore. 360s One study in older (age 60 to 80 years)hypertensive adults whose blood pressuremedication was withdrawn (Trial ofNonpharmacologic Interventions in theElderly, TONE) showed that after 2 years,mean net weight loss of nearly 4 kg (9 lb)resulted in more participants free of trial end-points (occurrence of high blood pressure,resumption of blood pressure medication, oroccurrence of a cardiovascular-clinical compli-cation [39.2 percent versus 26.2 percent]). 382Furthermore, blood pressure control was simi-lar for men and women and for African-Americans and whites. 353s In other studies of primarily middle-agedhypertensive adults, compared with controls,weight loss significantly reduced a return tohypertension medication at 1 year 354and at 4to 5 years. 348, 359s The Multiple Risk Factor Intervention Trial,which recruited 12,866 high-risk men, 30percent of whom had hypertension, deliveredan integrated intervention addressing severallifestyle behaviors and included weight loss asan important component. Significant reduc-tions in systolic and diastolic blood pressurewere found over 6 years compared with theusual care group and were directly related toweight loss: 1 kg (2.2 lb) of weight loss wasassociated with a reduction of 0.4 mm Hgsystolic and 0.3 mm Hg diastolic blood pres-sure in men not taking antihypertensive med-ications. The effect was slightly lower for mentaking antihypertensive medications.s Only one RCT conducted in hypertensivepatients reported no significant change inblood pressure despite weight loss of 3.3 kg(7.3 lb). 351s Another study conducted in hypertensiveindividuals suggested that weight loss reducedblood pressure only when sodium intake wasalso reduced. 349This finding was not consis-tent with a prior study of hypertensivepatients 358and has not been corroborated insubsequent larger trials conducted in nonhy-pertensive individuals. 378, 379
311.a.(2). Lifestyle Trials in NonhypertensiveIndividualsEvidence Statement: Weight loss pro-duced by lifestyle modifications reducesblood pressure in overweight nonhyperten-sive individuals. Evidence Category A.Rationale: A semiquantitative review 383coveringfour of the acceptable studies 363, 366, 372, 374in non-hypertensive individuals concluded that weightloss through dietary interventions significantlylowered blood pressure. In nonhypertensives,excluding one outlier trial that showed very largereductions in blood pressure, 3721 kg (2.2 lb) ofweight loss was associated with a reduction of0.45 mm Hg in both systolic and diastolic bloodpressure. Since this review, published in 1991,almost all relevant studies have reported thatweight loss reduces blood pressure in nonhyper-tensive individuals. 364, 365, 367-371, 373, 375-380, 384The Trials of Hypertension Prevention Phase 1(TOHP I) and Phase 2 (TOHP II) are amongthe larger, well-designed randomized trials, hav-ing follow-up rates exceeding 90 percent. Theyfollowed the participants for 18 months 378and 3to 4 years. 379The sample size of the weight lossintervention arms ranged from approximately300 (TOHP I) to 600 (TOHP II). The popula-tions were diverse, consisting of 30 to 35 percentwomen and 15 to 18 percent African-Americans, and were recruited from 10 centersin TOHP I and 9 centers in TOHP II. Resultsfrom both TOHP I and TOHP II demonstratedthat weight loss reduced blood pressure and theincidence of hypertension. Compared with con-trols, in both trials 10 kg (22 lb) of weight losswas associated with a reduction of 7 mm Hg sys-tolic and 5 to 6 mm Hg diastolic blood pressureat 18 months. At 36 months, systolic and dias-tolic blood pressure was reduced by 6 and 4 mmHg, respectively, for every 10 kg of weight loss.379In addition, weight loss reduced the incidence ofhypertension at 18 months by 20 to 50 percent378, 379and at 3 years by 19 percent. 379Secondaryanalyses from TOHP I 376and analyses underway from TOHP II 385demonstrated that thegreater the weight loss the greater the bloodpressure reduction, and as long as weight losswas maintained, blood pressure remainedreduced. One study that focused on olderAfrican-American diabetics showed that weightloss of 2.4 kg (5.3 lb) at 6 months resulted in areduction of 3.9 and 4.0 mm Hg systolic anddiastolic blood pressure. 362Only one study innonhypertensive individuals 368showed inconsis-tent results, where weight reduction decreasedsystolic (significant) and diastolic (nonsignifi-cant) blood pressure at 12 months in womenbut not in men. Another study 377showed short-term (3 months) blood pressure reduction butno longer-term (9 months) blood pressurereduction despite maintenance of weight loss of9 kg (19.8 lb). 3841.b.Pharmacotherapy TrialsEvidence Statement: Weight loss producedby most weight loss medications (except forsibutramine) and adjuvant lifestyle modi-fications is accompanied by reductions inblood pressure. Evidence Category B.To determine the effects of pharmacologicaltherapy on weight loss and subsequent changesin blood pressure levels, 10 RCT articles wereexamined. 386-395Weight loss studies using phar-macotherapy were conducted predominantly inwhite women and assumed that dietary changeswere uniform in the active drug and placebogroups. All of these studies had a placebo-con-trol group for which dietary recommendationsfor weight loss were provided; weight loss alsooccurred in the placebo group. No studies exam-
32Chapter 3: Examination of Randomized Controlled Trial Evidenceined whether weight loss using pharmacotherapyresults in blood pressure reductions similar toweight loss produced by diet therapy, nor werethe results compared with a pure control groupwith no weight loss. In addition, none of thestudies that examined the effect of pharma-cotherapy on blood pressure controlled forweight loss. In general, the use of phentermine,fenfluramine, and dexfenfluramine resulted insimilar or better weight loss than that seen in thecontrol or placebo group, and reductions in sys-tolic and diastolic blood pressure were concomi-tantly similar to or better than those observed inthe diet plus placebo group. 386, 390-393, 395However,one study on dexfenfluramine 394showedincreased blood pressure compared to controls,even though weight loss was greater with dexfen-fluramine than with placebo. Another studyshowed blood pressure lowering consistent withweight reduction produced by dexfenfluraminein women with upper-body obesity, but not inwomen with lower-body obesity. 387In a com-bined summary of data on orlistat presented atthe FDA Endocrinologic and Metabolic DrugsAdvisory Committee Meeting (held in Bethesda,Maryland, in May 1997), only a small decreasein systolic (approximately 2 mm Hg) and dias-tolic (approximately 1 mm Hg) blood pressurewas observed compared to the placebo. 388In astudy of sibutramine, a net weight loss at 3months of approximately 2 kg (4.4 lb) did notresult in lower blood pressure compared withplacebo controls.389Eleven double-blind place-bo-controlled trials lasting from 12 to 52 weeksshowed that sibutramine is associated with meanincreases in systolic blood pressure of 1 to 3 mmHg, in diastolic pressure of 1 to 2 mm Hg, andmean increases in pulse rate of 3 to 5 beats perminute relative to placebo. In hypertensive obesepatients, the mean change in blood pressure wasthe same in the sibutramine group and theplacebo group; both dropped. Systolic pressuredecreased 5.4 mm Hg in the sibutramine groupand 5.8 mm Hg in the placebo group, while thediastolic drop was 5.9 mm Hg and 3.7 mm Hg,respectively. At the FDA recommended doses,45 percent of patients had an increase, 35 per-cent had a decrease, and 20 percent had nochange in blood pressure. 3891.c. Abdominal FatEvidence Statement: Limited evidence sug-gests that decreases in abdominal fat reduceblood pressure in overweight nonhyperten-sive individuals, although this has not beenshown to be independent of weight loss.Evidence Category C.Rationale: Five RCTs 365,369,373, 375,377testing theeffects of weight loss on blood pressure also mea-sured abdominal fat by waist circumference orby dual-energy X-ray absorptiometry. In all fivestudies, waist circumference was reduced alongwith weight, and blood pressure was decreased.These studies were not designed to test whetherreduction in abdominal fat reduces blood pres-sure independent of weight loss.1.d.FitnessEvidence Statement: Increased aerobicactivity to increase cardiorespiratory fitnessreduces blood pressure independent ofweight loss. Evidence Category A.Rationale: Seven RCTs testing the effects ofweight loss on blood pressure in overweight andobese individuals also had measures of cardiores-piratory fitness, as measured by maximal oxygenuptake 346, 363, 369, 375, 377, 380; or submaximal heart ratetests. 365Weight loss was accompanied byincreased fitness, primarily if the interventionincluded increased physical activity or physicalactivity combined with diet. 346, 363, 375, 377, 380These
33Rationale: Fourteen RCT articles that studiedthe effects of diet and/or physical activity onlipids were reviewed. 365, 368, 370, 373, 380, 384, 399-406Table III-1 summarizes RCTs in which dietalone was used to produce weight loss. Entriesunder percentage change were calculated as fol-lows: the percentage change (value at the end ofstudy minus the value at baseline, divided by thevalue at baseline) was determined for the inter-vention group and the control group; the per-centage change for the control group was thensubtracted from the percentage change for theintervention group, giving the numerical valuepresented. When a run-in period was used, the“baseline” subtracted was the value at the end ofthe run-in, which was when the interventionbegan.Reductions in body weight varied from 5 to 13percent compared to the control group. The fol-lowing changes in plasma lipids and lipoproteinswere observed: total cholesterol, 0 to -18 per-cent; triglycerides, -2 to -44 percent; LDL-cho-lesterol, -3 to -22 percent; and HDL-cholesterol,-7 to +27 percent. Most of the trials were of suf-ficient duration to ensure that these changeswere not the result of acute caloric deficit.Table III-2 summarizes RCTs in which diet plusphysical activity or physical activity alone wasused to produce weight loss. Entries under per-centage change were calculated as for Table III-1.Similar changes were observed as for diet alone,although there was perhaps a more consistenttrend for all the lipid parameters with the com-bination of diet and physical activity and withphysical activity alone.studies were not designed to test whetherincreased cardiorespiratory fitness reduces bloodpressure independent of weight loss in over-weight and obese adults. By contrast, in 3 meta-analyses of 68 controlled studies of physicalactivity conducted in normotensive and hyper-tensive individuals, obesity was not an outcomemeasure nor was it included in the eligibility cri-teria. These meta-analyses showed that aerobicexercise to increase cardiorespiratory fitness sig-nificantly reduces blood pressure independent ofor in the absence of weight loss. 396-398RECOMMENDATION: Weight loss is recom-mended to lower elevated blood pressure inoverweight and obese persons with highblood pressure. Evidence Category A.2. Serum/Plasma Lipids and LipoproteinsSixty-five RCT articles were evaluated for theeffect of weight loss on serum/plasma concentra-tions of total cholesterol, low-density lipoprotein(LDL)-cholesterol, very low-density lipoproteincholesterol, triglycerides, and high-densitylipoprotein (HDL)-cholesterol. Studies wereconducted on individuals over a range of over-weight and lipid levels. Weight loss in 52 ofthese trials was induced by various lifestyle mod-ifications, including diet modification (reducingcalories, or saturated fat and cholesterol, orboth) and increased physical activity.2.a.Lifestyle TrialsEvidence Statement: Weight loss producedby lifestyle modifications reduces serumtriglycerides and increases HDL-cholesterol,and generally produces some reductions inserum total cholesterol, and LDL-choles-terol. Evidence Category A.
372.b.Pharmacotherapy TrialsEvidence Statement: Weight loss pro-duced by weight loss medications and adju-vant lifestyle modifications produces noconsistent change in blood lipids. EvidenceCategory B.Rationale: The effects of pharmacological thera-py on weight loss and subsequent changes intotal serum cholesterol levels were evaluated byexamining eight RCTs 386, 390-393, 395, 407,408(Table III-3).The four trials of dexfenfluramine showed noconsistent effects on total cholesterol; three trialsshowed decreases in triglycerides ranging from14 to 40 percent and two trials showed increasesin HDL-cholesterol of 8 percent in women to27 percent in men. One trial of phentermineplus fenfluramine showed a 24 percent decreasein triglycerides and a slight change in HDL-cho-lesterol. 395The trial of orlistat showed a decreasein total cholesterol and an increase in triglyc-erides, while the trial of fluoxetine showed theopposite. 407, 408Other RCTs using orlistat haveshown a modest improvement in lipids withweight loss. 388In a combined summary of data presented at theFDA Endocrinologic and Metabolic DrugsAdvisory Committee Meeting (held in Bethesda,Maryland, in September 1996), use of sibu-tramine, which resulted in a greater net weightloss of 2.4 kg (5.3 lb) at 4 months, also resultedin modestly lower levels of total cholesterol,LDL-cholesterol, and triglycerides, along withslightly higher HDL-cholesterol levels, comparedto placebos. 3892.c. Abdominal FatEvidence Statement: Limited evidencesuggests that decreases in abdominal fatcorrelate with improvements in the lipidprofile of overweight individuals, althoughthese improvements have not been shown tobe independent of weight loss. EvidenceCategory C.Rationale: Four RCTs testing the effects ofweight loss on blood lipids also included mea-sures of abdominal fat, as measured by waist cir-cumference. 365,373,375, 399In each study, waist cir-cumference was reduced along with weight, andblood lipids were improved. These studies werenot designed to test whether reductions inabdominal fat reduce blood lipids independentof weight loss. However, evidence exists fromepidemiologic observational studies that abdom-inal fat is related to an adverse lipid profile,including higher levels of total cholesterol, LDL-cholesterol, and triglycerides, and lower levels ofHDL-cholesterol. 409,4102.d. FitnessEvidence Statement: Increased aerobicactivity to increase cardiorespiratory fitnessfavorably affects blood lipids, particularlyif accompanied by weight loss. EvidenceCategory A.Rationale: Nine RCTs testing the effects ofweight loss on lipoproteins measured cardiores-piratory fitness, as measured by maximal oxygenuptake. 369, 375, 380, 384, 399, 401, 404-406Weight loss wasaccompanied by increased fitness if the interven-tion included increased physical activity or phys-ical activity combined with diet. 375, 380, 401, 404-406These studies were not designed to test whetherincreased cardiorespiratory fitness reduces bloodlipids independent of weight loss in overweightand obese adults. However, considerable evi-dence exists that aerobic exercise decreasestriglycerides and total cholesterol and mayincrease HDL-cholesterol in men and women,particularly if accompanied by weight loss. 411
38Chapter 3: Examination of Randomized Controlled Trial EvidenceW E I G H T R E D U C T I O N B Y D R U G T H E R A P YTC= Total Cholesterol TG= Triglycerides BMI= Body Mass IndexLDL-C= Low-Density Lipoprotein CholesterolHDL-C= High-Density Lipoprotein CholesterolN, Sex, and Percentage ChangeStudy Duration Weight Loss TC TG LDL-C HDL-CBremer386Dexfenfluramine -6 -14 -40 -9 +1512 (8M/4F)experimental14 (8M/6F)control12 weeksMathus- Dexfenfluramine -4 +7 -14 — —Vliegen39117 experimental18 control12 monthsO’Connor392Dexfenfluramine -6 +5 -26 — +27M27 (6M/21F) +8Fexperimental24 (14M/10F)control6 monthsPfohl393Dexfenfluramine -2 -3 +26 — —19 experimental15 control12 monthsWeintraub395Fen/Phen -13 -6 -24 -2 -153 Fen/Phen51 placebo34 weeksDrent407Orlistat -2 -2 +8 — —20 (3M/17F)Orlistat19 (3M/16F)placebo12 weeksO’Kane408Fluoxetine -6 +9 -25 — —7 Fluoxetine9 placeboTABLE III-3:
39RECOMMENDATION: Weight loss is recom-mended to lower elevated levels of totalcholesterol, LDL-cholesterol and triglyc-erides and to raise low levels of HDL-cho-lesterol in overweight and obese personswith dyslipidemia. Evidence Category A.3. Impaired Glucose Tolerance and DiabetesForty-nine articles of RCTs were reviewed toevaluate the effect of weight loss on fastingblood glucose and fasting insulin. Studies wereconducted in individuals with normal blood glu-cose levels (fasting plasma glucose < 115 mg/dL[< 6.4 mmol/L]), in individuals with impairedglucose tolerance (fasting plasma glucose of< 140 mg/dL [7.8 mmol/L] or 2 hours post-prandial plasma glucose of ≥ 140 to < 200mg/dL [7.8-11.1 mmol/L]), or in individualswith diabetes (fasting plasma glucose of > 140mg/dL or 2 hours postprandial plasma bloodglucose ≥ 200 mg/dL).*The methods of weight loss included diet, physi-cal activity, or both; behavior therapy; and phar-macotherapy. The dietary interventions includedlow calorie and very low calorie, and those thatpromoted changes in diet composition, such asamount and type of dietary fat. Physical activityvaried from controlled individualized exerciseprograms to informal group sessions coupledwith behavior therapy and dietary changes. Thedegree of blinding of outcome measures was notalways well described, nor was there a systematicconsideration of concurrent effects of changes inmedication or adherence to medication amongdiabetic patients enrolled in these weight reduc-tion studies.Of the 17 RCT articles considered acceptable,10 included normoglycemic individuals, 367, 373, 386,387, 390-393, 403, 4122 included patients with impairedglucose tolerance, 70, 369and 5 includedpatients with type 2 diabetes. 362, 404, 408, 413, 4143.a.Lifestyle TrialsEvidence Statement: Weight loss producedby lifestyle modifications, reduces bloodglucose levels in overweight and obese per-sons without type 2 diabetes, and reducesblood glucose levels and HbA1c in somepatients with type 2 diabetes. EvidenceCategory A.Rationale: Nine RCTs examined lifestyle therapy.70, 362, 367, 369, 373, 403, 404, 412, 413Five RCTs among normo-glycemic overweight individuals examined theeffects of behavior therapy or lifestyle change onweight loss. Changes in fasting glucose or fastinginsulin are used as secondary outcome measures.Improvements in the levels of fasting insulinwith weight loss ranged from 18 to 30 percent.While weight loss was maintained, improve-ments were observed for fasting glucose, fastinginsulin, and glucose and insulin levels duringoral glucose tolerance testing compared tocontrols. 367, 373, 403, 412Three RCTs of lifestyle modifications wereconducted among overweight diabetic patients.362,404,413One RCT of diet and physical activity inAfrican-Americans aged 55 to 79 years with type2 diabetes showed a 2.4 kg (5.3 lb) weight lossat 6 months compared to the usual-care controlgroup. The intervention group also decreasedHbA1c by 2.4 percentage units at 6 months. 362In another study, weight loss of 5 kg (11 lb)through diet resulted in a reduction of HbA1c by2 percentage units at 6 months compared withthe controls. 413At 1 year, the difference inweight loss between the two groups was reduced,and diabetic control was similar in each group.In the third study, weight loss, through physicalactivity in 25 type 2 diabetes patients randomly* The reviewed articles used the “old” definitions of the American Diabetes Association (ADA) for impaired glucose tolerance and diabetes. As ofNovember 1997, the new ADA definitions define “impaired fasting glucose” as those individuals having a fasting plasma glucose of 110 to 125 mg/dL,and “diabetes” as those individuals having a fasting plasma glucose of ≥ 126 mg/dL or 2 hours postprandial plasma glucose of ≥ 200 mg/dL.
40Chapter 3: Examination of Randomized Controlled Trial Evidenceassigned to physical activity (13 patients) or con-trol (12 patients) groups, was associated withimproved HbA1c. 404One RCT of 530 individuals in China withimpaired glucose tolerance compared the effectsof diet, physical activity, and diet plus physicalactivity on the incidence of diabetes. The datawere analyzed for the group as a whole and forthe lean and overweight groups separately. Theoverweight individuals in the diet and diet plusphysical activity groups lowered their BMIs,while the lean individuals did not. After 6 years,the cumulative incidence rates of diabetes weresignificantly lower in all three interventiongroups than they were in the controls (43.8 per-cent in the diet group, 41.1 percent in the phys-ical activity group, 46 percent in the diet plusphysical activity group, and 67.7 percent in thecontrols). This was true also if only the over-weight persons were included in the analysis. 703.b. Pharmacotherapy TrialsEvidence Statement: Weight loss producedby weight loss medications has not beenshown to be any better than weight lossthrough lifestyle modification for improvingblood glucose levels in overweight or obesepersons both with and without type 2 dia-betes. Evidence Category B.Rationale: Eight RCT articles considered theeffects of pharmacotherapy on weight loss andsubsequent changes in blood glucose. 386, 387, 390-393408, 414Six trials in nondiabetic individuals reportedweight loss with anorexiant drugs. 386, 387 390-393One study of 75 subjects with refractory obesityshowed that those taking dexfenfluramine had asignificantly better improvement in their bloodglucose levels as compared to placebo controls. 390Another trial showed that after 6 months ondexfenfluramine, fasting serum insulin levels fellfrom 105 at baseline to 89 at 6 months, butglucose levels did not decrease. 392This studywas difficult to interpret. Four other RCTs ofdexfenfluramine failed to show any significantdifferences between the active drug and placebogroups for glucose or insulin, despite significantdifferences in weight reduction. 386, 387, 391, 393Among diabetic patients, there were two trials ofanorexiant drugs. One trial of dexfenfluraminecompared with standard clinic practice showedsignificantly greater weight loss at 3 months butnot at 6 or 12 months. 414Weight loss at 3months was 3.4 kg (7.5 lb) in the dexfenflu-ramine group and 1.59 kg (3.5 lb) in the clinicgroup. At 6 months, the weight loss was 3.13 kg(6.9 lb) in the dexfenfluramine group and 1.70kg (3.7 lb) in the clinic group. By 1 year, bothgroups had regained some weight. Despite morethan a 1 kg (2.2 lb) difference of weight betweenthe dexfenfluramine group and the controlgroup, there were no significant differencesbetween groups in changes from baseline forHbA1c at any of the follow-up times (3, 6, or 12months). In the trial of fluoxetine, those patientson 60 mg of fluoxetine daily compared with theplacebo group had significant weight loss at 6months, a finding that was associated withimprovements in glucose and HbA1c. At 12months, these differences were no longer signifi-cant. 408The RCTs of orlistat showed a drop inboth fasting glucose and HBAlc in diabeticpatients on a combination therapy of lifestylechange and orlistat for weight loss. 3883.c. Abdominal FatEvidence Statement: Decreases in abdom-inal fat improve glucose tolerance in over-weight individuals with impaired glucosetolerance, although this has not been shownto be independent of weight loss. EvidenceCategory C.
41Rationale: Two RCTs testing the effects ofweight loss on changes in glucose tolerance andinsulin also had measures of abdominal fat, asmeasured by waist circumference. 369,373In eachstudy, waist circumference was reduced alongwith weight loss and glucose tolerance improved.Neither study was designed to test whetherreductions in waist circumference improved glu-cose tolerance or diabetic control independentof weight loss.Evidence exists from observational studies thatabdominal fat is related to impaired glucose tol-erance or type 2 diabetes independent of BMI.For instance, a few studies have found a signifi-cant relationship between abdominal fat loss andimprovement in glucose tolerance and insulinaction. 415-4183.d. FitnessEvidence Statement: Increased cardiores-piratory fitness improves glucose tolerancein overweight individuals, but no evidenceshows this relationship to be independentof weight loss. Evidence Category C.Rationale: Two RCTs that examined the effectsof weight loss on glucose tolerance in older men369and in persons with diabetes 404also had mea-sures of VO2 max (oxygen consumption) as ameasure of fitness. In both studies, physicalactivity resulted in a net weight loss of approxi-mately 2.3 to 2.5 kg (2 to 3 percent) andincreased VO2 max by 9 to 18 percent. In onestudy 419most measures of glucose toleranceimproved, whereas in the other study 369only 2-hour insulin improved. Neither study wasdesigned to test whether increased fitnessimproves glucose tolerance independent ofweight loss.RECOMMENDATION: Weight loss is recom-mended to lower elevated blood glucose lev-els in overweight and obese persons withtype 2 diabetes. Evidence Category A.4. Decreases in Abdominal Fat With WeightLossEvidence Statement: Weight loss is asso-ciated with decreases in abdominal fat, asmeasured by waist circumference. EvidenceCategory A.Rationale: Six acceptable RCTs on weight lossthat measured waist circumferences all show thatweight loss is associated with reductions in waistcircumference. 365, 369, 373, 375, 384, 399Waist circumfer-ence is commonly used as a surrogate measurefor abdominal visceral fat. 420, 421Although noRCTs examined changes in visceral fat per se,one observational study 422showed significantdecreases in visceral fat with a mean weight lossof 12.9 kg (28.4 lb).In addition, many observational and non-RCTinterventional studies have reported that weightloss caused by a variety of treatments reducesabdominal visceral fat levels. 417, 423-428Fat locatedin the abdominal region is associated withgreater health risks than that in peripheralregions, e.g., the gluteal-femoral area. 155-159, 429, 430Many studies used the waist-to-hip ratio as themeasure of abdominal fat; however, the panelconsidered waist circumference to be a bettermarker of abdominal fat content than the waist-to-hip ratio. For more information on abdomi-nal fat, see Section 4.B.
42Chapter 3: Examination of Randomized Controlled Trial EvidenceWHAT TREATMENTS ARE EFFECTIVE?A variety of modalities may be used to treat over-weight and obesity. To assess their effectivenesson weight loss, loss of abdominal fat, and car-diorespiratory fitness, evidence from RCTs wasexamined. These trials considered dietary thera-py, physical activity, combined therapy, behaviortherapy, pharmacotherapy, and surgery. Thedetails of the studies reviewed and subsequentlynot included are provided in Appendix II.1. Dietary TherapyEighty-six RCT articles evaluated the effective-ness of many different diets on weight loss: low-calorie diets (LCDs), very low-calorie diets(VLCDs), vegetarian, American HeartAssociation guidelines and the NationalCholesterol Education Program’s Step I diet withcaloric restriction, and other low-fat regimenswith varying combinations of macronutrients.The trials that included participants who werenot overweight were rejected. Some trials weredone in patient populations with diabetes,hypertension, or hyperlipidemia. Many of thetrials do not mention whether concomitantphysical activity programs were promoted, andmany do not give information about the behav-ioral approaches that were used to deliver thediet or physical activity intervention.Approximately 60 percent of the trials are equalto or greater than 6 months in duration.1.a Low-Calorie DietsEvidence Statement: LCDs can reducetotal body weight by an average of 8 per-cent over 3 to 12 months. EvidenceCategory A.Rationale: Thirty-four RCT articles examinedthe impact on weight loss of an LCD consistingof approximately 1,000 to 1,200 kilocalories/day.Many of these LCDs also promoted low fatintake as a practical way to reduce calories.Twenty-five RCT articles covered interventionslasting 6 or more months. 346-348, 350, 354, 355, 359, 363-366,368, 369, 373, 375, 378-380, 399, 400, 406, 413, 431-433Another nine RCTs lasted 12 to 21 weeks. 349, 351,356, 370, 371, 377, 384, 434, 435All of the studies, regardless of the length of theintervention, showed that LCDs result in weightloss. From the 25 RCT articles with a durationof ≥ 6 months, compared to controls, LCDsbrought about a mean weight loss of approxi-mately 8 percent of body weight over a period of6 months and up to 1 year. The nine studieswith an intervention lasting from 3 months upto 6 months also averaged approximately 8 per-cent weight loss compared to controls. Fourstudies that included a long-term weight lossand maintenance intervention lasting 3 to 4.5years reported an average weight loss of 4 per-cent over the long term. 348, 359, 366, 379Evidence Statement: LCDs resulting inweight loss effect a decrease in abdominalfat. Evidence Category A.Four RCTs testing the effects of LCDs alone onweight loss also had measures of abdominal fat,as measured by waist circumference. 365, 369, 375, 399The studies showed consistently that waist cir-cumference decreases along with weight loss pro-duced by LCDs. After 6 to 12 months, LCDsresulted in a mean weight loss of as little as 0.3BMI unit to as much as 11 kg (24.3 lb) bodyweight, and a concomitant reduction in waistcircumference of 1.5 to 9.5 cm compared tocontrols.B
43Evidence Statement: No improvement incardiorespiratory fitness as measured byVO2 max appears to occur in overweight orobese adults who lose weight on LCDswithout increasing physical activity.Evidence Category B.Rationale: Ten RCTs examined whether weightloss through diet alone without increased physi-cal activity had an effect on cardiorespiratory fit-ness as measured by VO2 max when comparedto controls with no weight loss. Four RCTsreported no improvement in VO2 max, 363,375,399,435while six RCTs reported an improvement. 346, 369,380, 384, 406, 432In three of the six studies, improve-ment in those on the LCD was shown becausethe control group had a large decrease in VO2max. 346,406,432In all six studies that showedimprovement, VO2 max was expressed asml/kg/min rather than liters/min. Because theratio of ml O2/kg/min increases with weight lossalone and not necessarily from an improvementin oxygen uptake, results from studies reportingVO2 max in terms of ml O2/kg/min are equivo-cal. The three studies that reported VO2 max asliters/min consistently showed no improvementfrom diet alone. 375,399,4351.b. Very Low-Calorie DietsEvidence Statement: VLCDs producegreater initial weight loss than LCDs.However, the long-term (> 1 year) weightloss is not different from that of the LCD.Evidence Category A.Four RCTs compared VLCDs to LCDs. 436-439The VLCDs provided about 400 to 500kcal/day, whereas the LCDs provided approxi-mately 1,000 to 1,500 kcal/day. During theactive phase, VLCDs were given exclusively for12 to 16 weeks and then were followed byLCDs for a total duration ranging from 24weeks to 5 years. The participants in these trialswere primarily women who were extremelyobese.VLCDs either alone or, usually, combined withbehavioral therapy, promoted weight loss ofapproximately 13 to 23 kg (28.7 to 50.7 lb)during the active phase of the VLCD interven-tion, compared to 9 to 13 kg (19.8 to 28.7 lb)with LCDs. In three of the four studies, VLCDsresulted in 4 to 12 kg (8.8 to 26.5 lb) greaterweight loss than the LCDs at the end of theactive phase. 436-438Over the medium term of 6 to12 months, weight loss on the VLCDs rangedfrom 1.1 kg (2.4 lb) to as much as 10.4 kg (22.9lb) greater than weight loss on LCDs. One studydid not show a particular advantage of VLCDover LCD either during the active phase or at 24weeks. 439After 1 year, there was no long-termadvantage of VLCDs over LCDs. 436, 4371.c. Lower-Fat DietsEvidence Statement: Although lower-fatdiets without targeted caloric reductionhelp promote weight loss by producing areduced caloric intake, lower-fat diets cou-pled with total caloric reduction producegreater weight loss than lower-fat dietsalone. Evidence Category A. Lower-fatdiets produce weight loss primarily bydecreasing caloric intake. EvidenceCategory B.Rationale: Nine RCTs testing the effects of dieton weight loss promoted diets that varied in fatand caloric content. 317, 364, 369, 434, 440-444Lower-fatdiets varied from 20 to 30 percent of caloriesfrom fat, and calories ranged from 1,200 to2,300 calories.
44Chapter 3: Examination of Randomized Controlled Trial EvidenceThree RCTs, all 6 months or greater in dura-tion, promoted lower-fat diets with ad libitumcaloric intake. Two of the three RCTs reportedthat lower-fat diets with ad libitum energyintake resulted in a reduction in caloric intake of85 kcal and 300 kcal, 317, 444and all three RCTsproduced greater weight loss by 1 to 3.9 kg (2.2to 8.6 lb), compared to a higher-fat diet. 317, 441, 444The third RCT had a large disparity in baselinecaloric intake between the two diet groups, mak-ing the reported caloric intakes at the end of thestudy difficult to compare. 441Three RCTs compared lower-fat diets with tar-geted caloric reduction to lower-fat diets alone,and all three found that weight loss is greater inthe lower-fat diet with caloric reduction thanwith the lower-fat diet alone. 369, 434, 442When similar caloric levels occur between lower-fat and higher-fat diets, similar amounts ofweight loss were reported in two studies, 364, 443whereas one study showed 1.8 kg (4 lb) greaterweight loss on the lower-fat compared with thehigher-fat diet, despite similar reported calorielevels. 440Taken together, these studies show that lower-fatdiets ranging from 20 to 30 percent of caloriescan contribute to lower caloric intake even whencaloric reduction is not the focus of the inter-vention, 317, 444but when LCDs are targeted withlower-fat diets, better weight loss is achieved. 369,434, 442However, there is little evidence that lower-fat diets (per se) cause weight loss independentof caloric reduction.RECOMMENDATION: LCDs are recom-mended for weight loss in overweight andobese persons. Evidence Category A.Reducing fat as part of an LCD is a prac-tical way to reduce calories. EvidenceCategory A.2. Physical ActivityTwenty-four RCT articles were reviewed for theeffect of physical activity on weight loss, abdom-inal fat (measured by waist circumference), andchanges in cardiorespiratory fitness (VO2 max).Thirteen articles were deemed acceptable. 346, 363,365, 369, 375, 401, 404, 406, 432, 434, 445-447Only one of theseRCTs compared different intensities and formatwith a control group, although the goal was toincrease physical activity and not specifically toproduce weight loss. 401Results from this trialwere subsequently reported after 2 years, butthese no longer included the control group. 447One additional study did not have a no-treat-ment control group, but compared three activetreatment groups with each other: diet only,exercise only, and combination exercise plus diet. 448Most RCTs described the type of physical activi-ty as cardiovascular endurance activities in theform of aerobic exercise such as aerobic dancing,brisk walking, jogging, running, riding a station-ary bicycle, swimming, and skiing, preceded andfollowed by a short session of warmup and cooldown exercises. Some physical activity programsalso included unspecified dynamic calisthenicexercises. 363, 369, 406, 446The intensity of the physical activity was adapt-ed to each individual and varied from 60 to 85percent of the individual’s estimated maximumheart rate, or was adjusted to correspond toapproximately 70 percent of maximum aerobiccapacity (VO2 max). The measure of physical fit-ness included VO2 max. The frequency of physi-cal activity varied from three to seven sessions aweek and the length of the physical activity ses-sion ranged from 30 to 60 minutes. Some physi-cal activity programs were supervised, and somewere home-based. Adherence to the prescribedphysical activity program was recorded andreported in some studies and not mentioned inothers. Most studies did not estimate the caloricexpenditure from the physical activity or reportcalorie intake. The duration of the intervention
45varied from 16 weeks to 1 year; six articlesreported on trials that lasted at least 1 year. 346, 363,375, 401, 406, 432Evidence Statement: Physical activity,i.e., aerobic exercise, in overweight andobese adults results in modest weight lossindependent of the effect of caloric reduc-tion through diet. Evidence Category A.Rationale: Twelve RCT articles examined theeffects of physical activity, consisting primarilyof aerobic exercise, on weight loss compared tocontrols. 346, 363, 365, 369, 375, 401, 404, 406, 432, 434, 445, 446Ten ofthe 12 RCT articles reported a mean weight lossof 2.4 kg (5.3 lb) (or 2.4 percent of weight) 363,369, 375, 406, 419, 432, 434or a mean reduction in BMI of0.7 kg/m2(2.7 percent reduction) 346, 365, 401in theexercise group compared to the control group.In three of these ten studies, the weight loss was< 2 percent of body weight (< 2 kg) (4.4 lb). 369,375, 434In contrast, two RCTs showed no benefiton weight from exercise, reporting weight gainin the exercise group compared to the controlgroup. 445, 446In one of these studies, the controlgroup received only diet advice but neverthelesslost 9 kg (19.8 lb), whereas the exercise grouplost only 7 kg (15.4 lb). 445In the second study,there was a total of only 10 participants, all hav-ing noninsulin-dependent diabetes mellitus, andthe control group lost 3 kg (6.6 lb) whereas theexercise group lost only 2 kg (4.4 lb). 446A meta-analysis of 28 publications of the effect onweight loss of exercise compared to diet or con-trol groups showed that aerobic exercise aloneproduces a modest weight loss of 3 kg (6.6 lb) inmen and 1.4 kg (3.1 lb) in women compared tocontrols. 449Ten articles reported on RCTs that had a diet-only group in addition to an exercise-only group.346, 363, 365, 369, 375, 406, 432, 434, 445, 448In every case exceptone, 365the exercise-only group did not experi-ence as much weight loss as the diet-only group.The diet-only group produced approximately 3percent, or 3 kg (6.6 lb), greater weight loss thanthe exercise-only group.No single study examined the length of theintervention in relation to the weight loss out-come. Only one study compared the effect onmaximum oxygen uptake of different intensitiesand formats of physical activity over a 1-year fol-low-up 401and 2-year follow-up period. 447Betteradherence over 1 year was found if the exercisewas performed at home rather than in a groupsetting, regardless of the intensity level.Subsequently, the different exercise groups werecompared with each other over the longer term(2 years), and better long-term adherence wasfound in the higher intensity home-based exer-cise group compared to the lower intensityhome-based or higher intensity group-basedexercise groups. 447The question of whether physical activityenhances long-term maintenance of weight losshas not been formally examined in RCTs.Examination of long-term weight loss mainte-nance produced by physical activity interven-tions compared with diet-only interventions can-not easily be compared between RCTs becauseof numerous differences between studies withrespect to design, sample size, intervention con-tent and delivery, and characteristics of the studypopulation samples. However, a number ofanalyses of observational and post hoc analysesof intervention studies have examined whetherphysical activity has a beneficial effect onweight. Cross-sectional studies have generallyshown that physical activity is inversely related tobody weight 450-454and rate of weight gain withage. 455Longitudinal studies with 2 to 10 years offollow-up results have observed that physical activ-ity is related to less weight gain over time 456-459,less weight gain after smoking cessation inwomen, 460and weight loss over 2 years. 461In
46Chapter 3: Examination of Randomized Controlled Trial Evidenceaddition, post hoc analyses of several weight lossintervention studies reported that physical activi-ty was a predictor of successful weight loss. 454, 462, 463The results of these RCTs showed that physicalactivity produces only modest weight loss andobservational analyses from other studies suggestthat physical activity may play a role in long-term weight control and/or maintenance ofweight loss.Evidence Statement: Physical activity inoverweight and obese adults modestly reducesabdominal fat. Evidence Category B.Rationale: Only three RCTs testing the effect ofphysical activity on weight loss also had mea-sures of abdominal fat as assessed by waist cir-cumference. 365, 369, 375One study demonstratedthat physical activity reduced waist circumfer-ence compared with the control group, 365andanother study showed a small effect on waist cir-cumference (0.9 cm) in men but not women. 375One study in men showed a small increase inwaist circumference. 369Weight loss was modestin all of these studies. These studies were notdesigned to test the effects of physical activity onabdominal fat independent of weight loss.However, several large studies in Europe, 464Canada, 453and the United States 465-468reportedthat physical activity has a favorable effect onbody fat distribution. These studies showed aninverse association between energy expenditurethrough physical activity and several indicatorsof abdominal obesity, such as waist circumfer-ence and waist-to-hip and waist-to-thigh cir-cumference ratios.Evidence Statement: Physical activity inoverweight and obese adults increases car-diorespiratory fitness independent of weightloss. Evidence Category A.Rationale: Eleven RCT articles testing the effectof physical activity alone on weight loss in menand women also included measures of cardiores-piratory fitness, as measured by maximal oxygenuptake (VO2 max). 346, 363, 369, 375, 401, 404, 406, 432, 434, 445, 446All 11 showed that physical activity increasedmaximum oxygen uptake in men and women inthe exercise groups by an average of 14 percent(ml/kg body weight) to 18 percent (L/min).Even in studies with modest weight loss (< 2percent), physical activity increased VO2 max byan average of 12 percent (L/min) to 16 percent(ml/kg). 369, 375, 434One study that compared different formats andintensities of physical activity on VO2 maxreported that improvement in VO2 max wasrelated to adherence to the physical activityregime. In that study, the lower-intensity pro-gram was equally effective on VO2 max as ahigher-intensity program, largely as a result ofdifferent levels of adherence. 401The results of the RCTs strongly demonstratethat physical activity increases cardiorespiratoryfitness in overweight and obese individuals.RECOMMENDATION: Physical activity isrecommended as part of a comprehensiveweight loss therapy and weight mainte-nance program because it: (1) modestlycontributes to weight loss in overweightand obese adults (Evidence Category A),(2) may decrease abdominal fat (EvidenceCategory B), (3) increases cardiorespiratoryfitness (Evidence Category A), and (4)may help with maintenance of weight loss(Evidence Category C).
473. Combined Therapy (Diet and PhysicalActivity)Twenty-three RCT articles investigated theeffects on body weight of a combination of areduced calorie diet with increased physicalactivity. The control groups used diet alone orphysical activity alone.Of the 15 studies deemed acceptable, each ofthem compared the combined intervention withdiet alone, 346, 365, 375, 377, 380, 384, 434, 435, 445, 448, 469-473and 6of them also compared the combined interven-tion to physical activity alone. 346, 365, 375, 434, 448, 474The studies varied in terms of the length of theactive intervention period as well as the lengthof the follow-up. Many of the programs lastedfor more than 6 months, and five studies hadintervention or follow-up data for at least 1 year.346, 375, 380, 445, 473Two studies included patients withtype 2 diabetes. 470, 473Participants in the diet andphysical activity group most often participatedin supervised group-based physical activity con-sisting of 30- to 60-minute sessions 3 times aweek, and they were encouraged to maintain amaximum heart rate between 60 and 80 percent.There were three diet components: generaldietary advice, caloric reduction of 500 to 1,000kcal per day, or a diet containing 1,200 kcal perday.Evidence Statement: The combination ofa reduced calorie diet and increased physi-cal activity produces greater weight lossthan diet alone or physical activity alone.Evidence Category A.Rationale: Of the 15 RCT articles, 12 observedthat the combined diet and physical activitygroup had a mean greater weight loss of 1.9 kg(4.2 lb) and a mean greater BMI reduction of0.4 (range of 0.3 to 0.5) than the diet-alonegroup. 346, 365, 375, 377, 380, 384, 434, 435, 448, 469, 470, 473One study compared different forms of physicalactivity in combination with diet and foundthat, compared to diet alone, the greatest weightloss occurred when the combination interven-tion included both aerobic and resistance train-ing: 0.9 kg (2 lb) greater weight loss than aero-bic exercise plus diet. 435Another study reportedthat participants receiving resistance training incombination with diet had 2.2 and 5.0 kg (4.9and 11 lb) greater weight loss than the diet-alonegroup after 24 and 48 weeks, respectively. 469Five of the six studies that compared combinedintervention with physical activity aloneobserved that the combined intervention grouphad a 5.3 kg (11.7 lb) (range of 3.6 to 6.2 kg)(7.9 to 13.7 lb) greater weight loss and 0.9change in BMI unit than the physical activity-alone group. 346, 365, 375, 434, 448This greater weightloss was significant in three studies. 365, 375, 448Three RCTs compared the longer-term versus theshorter-term effects of the combination of physi-cal activity and diet versus diet alone. 384, 469, 473Allthree studies found that the combination result-ed in approximately 1.5 to 3 kg (3.3 to 6.6 lb)greater weight loss than diet alone over thelonger term of 9 months to 2 years.Evidence Statement: The combination of areduced calorie diet and increased physicalactivity produces greater reductions inabdominal fat than either diet alone orphysical activity alone, although it has notbeen shown to be independent of weight loss.Evidence Category B.Rationale: Three RCTs 365, 375, 471that examined thecombined effect of diet and physical activity onweight loss also had measures of abdominal fat,as measured by waist circumference. Two of thethree RCTs showed that the combination inter-vention resulted in greater reduction in waist cir-
48Chapter 3: Examination of Randomized Controlled Trial Evidencecumference than diet alone in men 365, 375and(only one) in women. 375In the other RCT, nodifferences in reduced waist circumference wereobserved. 471Of the three RCTs comparing com-bination therapy to physical activity alone, onestudy reported a greater waist reduction of 2.7to 3.8 cm in women and men, respectively, 375and by less than 1 cm in another study. 365Theeffects of combination therapy on abdominal fatin these studies are not separable from the effectson weight loss.Evidence Statement: A combination of areduced calorie diet and increased physicalactivity improves cardiorespiratory fitness asmeasured by VO2 max when compared todiet alone. Evidence Category A.Rationale: Nine RCTs examined the effect of acombination of diet and physical activity on fit-ness compared to diet alone. 346, 375, 377, 380, 434, 435, 445,448, 475In most studies, the physical activity was inthe range of 60 to 80 percent VO2 max. All nineRCTs showed that a combination of diet andphysical activity improved cardiorespiratory fit-ness, as measured by VO2 max, more than dietalone, whether reported relative to body mass inml O2/kg/min or in absolute values by mlO2/min. 346, 375, 377, 380, 434, 435, 445, 448, 475Studies oflonger duration tended to show the greatestimprovement. This improvement was significantin six of the studies. 346, 375, 377, 380, 434, 448RECOMMENDATION: The combination ofa reduced calorie diet and increased physi-cal activity is recommended, since it pro-duces weight loss, decreases abdominal fat,and increases cardiorespiratory fitness.Evidence Category A.4. Behavior TherapyBehavior therapy constitutes those strategies,based on learning principles such as reinforce-ment, that provide tools for overcoming barriersto compliance with dietary therapy and/orincreased physical activity. Two questions areimportant in assessing the effect of behaviortherapy in the treatment of overweight and obe-sity. First, does behavior therapy provide addi-tional benefits above and beyond the otherweight loss approaches—diet, physical activity,or drug therapy? Second, are there behavioralstrategies that are more effective than others?4.a. Additional Benefits Beyond Other WeightLoss ApproachesThirty-six RCT articles were examined to evalu-ate whether behavior therapy provides additionalbenefits beyond other weight loss approaches.The acceptable RCTs compared behavior thera-py plus another weight loss strategy to thatweight loss strategy without behavior therapy.Evidence Statement: Behavior therapy,when used in combination with otherweight loss approaches, provides additionalbenefits in assisting patients to lose weightshort term (1 year). Evidence Category B.No additional benefits are found at 3 to 5years in the absence of continued interven-tion. Evidence Category B.Rationale: Four RCTs addressed the question ofwhether behavior therapy, in order to reinforce ahealthy diet, provides additional benefits aboveand beyond other weight loss approaches. Threestudies compared behavior therapy plus dietaryintervention to a dietary intervention alone. 436,476, 477No studies were found that examined theadded benefit of behavior therapy when com-bined with either physical activity or diet andphysical activity. One study was found that com-pared behavior therapy to drug therapy. 478
49One study compared 36 obese women in one ofthree groups of 12 participating in a 16-weekintervention: (1) individual counseling; (2)group counseling; and (3) group counseling plusbehavior therapy. All groups lost weight at theend of treatment, and there were no significantdifferences among the groups. By the 1-year fol-low-up, participants in groups 1 and 3 wereequally successful in maintaining weight loss,while most participants in group 2 regainedmost of their weight. 476Other studies examined the effect of VLCDalone and in combination with behavior therapy.436, 477Subjects were randomly assigned to either:(1) VLCD alone; (2) behavior therapy plus LCD(1,000 to 1,200 kcal), denoted as the behaviortherapy-alone group; or (3) VLCD plus behaviortherapy (combined treatment). They wereassessed at the end of treatment and at 1- and 5-year follow-ups. At the end of treatment, and atthe 1-year follow-up, weight loss was significant-ly greater for the combination group as com-pared to the other two groups. About one-thirdof those receiving behavior therapy (either incombination with an LCD or VLCD) were ableto maintain their full end-of-treatment weightloss at 1 year, compared to only 5 percent ofthose in the diet-alone group. By the 5-year fol-low-up, the mean weight loss was not signifi-cantly different across groups.Thus, although the weight loss results of behav-ioral treatment in combination with either anLCD or VLCD were quite good at the end oftreatment and at the 1-year follow-up, they werenot maintained by 5 years. These results maywell have been due to the infrequent practice ofthe behavioral strategies by the subjects. Waddenand colleagues recommended that patients par-ticipate in a 6- to 12-month weight maintenanceprogram after weight loss (which subjects didnot do in the study described above), and that ifthey experience a gain of ≥5 kg (11 lb) whichthey cannot reverse on their own, that theyshould reenter a weight loss therapy program.One study assessed the short- and long-termweight loss of behavior therapy, fenfluramine, ora combination of the two in 120 obese women. 478Patients in the drug group and those in thecombined drug plus behavior therapy group lostsignificantly more weight than did the groupthat received behavior therapy only and the con-trol group; the drug-only group and combina-tion group did not differ from each other at theend of the 6-month treatment. However, at the1-year follow-up, significantly less weight wasregained by the patients in the behavior-therapy-only group compared to the combined treat-ment group and the pharmacotherapy group,resulting in significantly better overall net weightloss. Results of this study suggest that the vari-ous interventions had different effects at the endof treatment and at follow-up.To summarize, three of the four studies demon-strated that behavior therapy, when used in con-junction with other weight loss approaches, wasmore effective in reducing weight or delayingweight regain, either at the end of treatment orat the 1-year follow-up or both. 476-478Behaviortherapy was not better in its effect on weight lossat the 5-year follow-up.No RCTs examined the additional effect ofbehavior therapy combined with diet and physi-cal activity on cardiorespiratory fitness orabdominal fat.4.b. Comparison of Behavior Therapy StrategiesAnother way to examine the efficacy of behaviortherapy in the treatment of overweight and obe-sity is to evaluate studies that compare variousbehavioral techniques with one another. Of 31RCTs that compared one or more behavioralinterventions, 19 were deemed acceptable.
50Chapter 3: Examination of Randomized Controlled Trial EvidenceEvidence Statement: No one behaviortherapy appeared superior to any other inits effect on weight loss; rather, multimodalstrategies appeared to work best and thoseinterventions with the greatest intensityappeared to be associated with the greatestweight loss. Evidence Category A.Of the 19 RCTs, 17 produced weight loss. 367, 476,479-493Behavioral strategies to reinforce changes indiet and physical activity in obese adults pro-duce weight loss in the range of 10 percent ofbaseline over 4 months to 1 year. The greatestamount of weight loss was usually observed inless than 12 months of the intervention; howev-er, weight loss was observed for up to 21⁄2 years insome studies. Multimodal strategies worked best,and the intensity (number of contacts and dura-tion) and degree (albeit difficult to discern insome of these studies) of a particular reinforce-ment were most related to weight loss. Althoughinitial therapy proved successful, it also appearedthat when behavioral therapy ceased, weight wasregained on average, but not all subjects relapsedback to baseline levels. Ten of the 19 studieswere based on at least 1-year results. 367, 476, 479, 480,484, 486-488, 492, 493Fourteen trials compared different behaviortherapy strategies. Two trials compared cognitivebehavior therapy techniques, specifically cogni-tive rehearsal, social pressure, and cue avoidance(see Glossary) with other behavior therapies. 479, 480With regard to weight loss, cognitive rehearsalwas no more effective than the group support or“social pressure” condition and was less effectivethan cue avoidance. 480Another study indicatedno clear benefits for cognitive rehearsal aboveand beyond other behavioral strategies. 479One483of two studies that assessed the effect of mon-etary incentives reported greater weight loss. 367, 483The effect of extending treatment in variousways was evaluated in several studies. 486, 487, 489, 494One study evaluated 20 weeks versus 40 weeksof standard behavior therapy and found greaterweight loss with the extended therapy that wasmaintained at 72 weeks. 486Another study evalu-ated behavior therapy and the effect of a post-treatment therapist on weight loss. 487The studyevaluated the effect of extended treatment usingtherapist contact in general, therapist contactemphasizing social influence, and extendedtreatment emphasizing physical activity. Theemphasis of physical activity plus social influ-ence produced greater weight loss than not hav-ing contact emphasizing these components.Extended contact with a therapist producedabout 10 kg (22 lb) greater weight loss than noextended contact, even over the long term. At 18months, all strategies that combined behaviortherapy with a post-treatment maintenance pro-gram yielded a significantly greater long-termweight loss than behavior therapy alone.Another study examined the effect of extendingcognitive behavior therapy for 3 months in addi-tion to the initial 3 months of therapy for obesepatients with binge eating disorder, and foundthat extended cognitive behavior therapyimproved binge eating behavior, although notweight. 494A pilot study of women with type 2diabetes compared a 16-week group-behavioralweight loss program to the same group, but withthree individualized motivational interviewingsessions added. Both groups demonstrated sig-nificant weight loss with the motivational grouplosing 5.5 kg (12.1 lb) on average and the stan-dard behavioral group losing 4.5 kg (9.9 lb);there were no differences between the two.However, those in the motivational interviewinggroup demonstrated better attendance at meet-ings (attending on average 13.3 meetings versus8.9 meetings), had significantly better glucosecontrol, and recorded blood glucose and fooddiaries more frequently 489than did members ofthe standard behavioral group.
51A study that involved self-instruction plus incen-tives versus group instruction plus incentivesdemonstrated no effect of the group treatmenton weight loss. 481Weight loss was positivelyassociated with attendance; however, mandatoryattendance did not seem to increase programeffectiveness and seemed to discourage atten-dance among men. The program was equallyeffective when implemented through either self-help materials or by professionals. Another studycompared stimulus-narrowing during the rein-troduction of food (achieved by providingprepackaged food) for those on VLCDs to pro-viding regular food during the reintroduction. 495They also compared reintroduction of foodsbased on progress in losing weight or maintain-ing weight loss versus a time-dependent basis.Neither strategy was helpful in maximizingweight loss or weight maintenance.Two trials evaluated the impact of spousalinvolvement versus no spousal involvement inbehavior therapy. 488, 492One study found no over-all differences in weight loss between those treat-ed with spouses and those treated without spous-es; however, women lost more weight when treat-ed with a spouse, and men did better when treat-ed alone (P < .01 at end of treatment; P < .06 at1 year). 492Another study found a positive effectof spousal involvement on weight loss soon aftertreatment was completed; however, this effecthad largely dissipated by the 3-year follow-up. 488In summary, a variety of behavioral therapystrategies promote weight loss, with no one par-ticular behavioral strategy having the greatestefficacy. Multimodal strategies with greaterintensity appear to be most effective.Evidence Statement: Long-term follow-upof patients undergoing behavior therapyshows a return to baseline weight in the greatmajority of subjects in the absence of contin-ued behavioral intervention. EvidenceCategory B.Rationale: Studies have shown that while weightloss is achieved, it is very difficult to maintainover a long period of time (1 to 5 years) in theabsence of continued intervention. 436, 477, 486, 491This emphasizes the great importance of contin-uing a maintenance program on a long-termbasis.Evidence Statement: Little evidence existson the effect of behavior therapy in combi-nation with diet and physical activity oncardiorespiratory fitness.Rationale: One RCT examined the impact ofdifferent levels of behavior therapy in combina-tion with diet and aerobic exercise on cardiores-piratory fitness. 484Subjects were randomlyassigned to one of three groups which varied intheir degree of training in, and detailed applica-tion of, behavioral change principles (describedas the basic, extended, and maximal behavioraltherapy groups). All groups incorporated fre-quent intensive therapist contact, self-monitor-ing of behavior, regular objective assessment,and feedback of change in status. The threegroups differed in the degree of emphasis givento behavioral self-management training, amountof detail provided on individual risk factors andoverall coronary heart disease risk, and establish-ment of realistic short-term goals for coronaryrisk factor change. In terms of cardiorespiratoryfitness, the group receiving the extended behav-ioral therapy and the group receiving the maxi-mal behavior therapy showed improvements inVO2 max of 2.90 and 3.40 mL/kg/min, respec-tively, when compared with controls who
52Chapter 3: Examination of Randomized Controlled Trial Evidencereceived the basic behavior therapy. However, itis unlikely that behavior therapy caused theseimprovements independent of weight loss.No RCT evidence exists on the effect of behav-ior therapy combined with diet and physicalactivity on changes in abdominal fat as mea-sured by waist circumference.RECOMMENDATION: Behavior therapy isa useful adjunct when incorporated intotreatment for weight loss and weight main-tenance. Evidence Category B.Commentary: Do the same treatmentapproaches for overweight and obesityfit diverse patient populations?Whether or not the same weight reductionprogram will accomplish equivalent weightchange in ethnically diverse populationgroups can be questioned on theoreticalgrounds. Weight reduction programs addressmotivations and behaviors strongly influ-enced by sociocultural factors that vary withethnicity. 257, 496, 497Sociocultural factors mayalso influence the relative efficacy of behav-ioral programs in young, middle, and olderadulthood; men or women; or any otherpatient groups where the motivations andlifestyle factors targeted may vary from thosecommonly assumed to be in operation. Innateor acquired physiological factors that affectenergy metabolism may also influence therelative efficacy of behavioral weight reduc-tion programs. For example, some cross-sec-tional comparisons have suggested that therelative weight loss achieved for a givendegree of calorie restriction will be lower inAfrican-Americans than whites because of alower resting metabolic rate in African-Americans. 498-501Another study suggests meta-bolic influences on weight loss in premeno-pausal versus postmenopausal women. 502The diversity of populations in obesity treat-ment studies is limited, and there is not acoherent body of studies evaluating systematicdifferences in the outcomes of overweight andobesity treatment by age, socioeconomic sta-tus, or other participant characteristics thatmight be expected to influence outcome.However, the possibility of differences in theeffectiveness of behavioral weight reductionapproaches between African-Americans andwhites has been suggested by reports fromseveral intervention studies. 360, 503-505In someof these trials, weight reduction was the onlyintervention; in other instances it was com-bined with sodium reduction or otherdietary or lifestyle changes. Taken together,these studies suggest that, on average, withinthe same program, weight losses of African-Americans are less than those of their whitecounterparts, particularly among women. Aformal analysis of this issue in two RCTs—the Hypertension Prevention Trial and Phase1 of the Trials of Hypertension Prevention—suggested that the weight gain in controlparticipants was also higher among African-Americans than whites, but, in women, thenet weight loss (i.e., active interventionminus control) was still less in African-Americans. There is ample evidence to sup-port explanations of this differential weightloss on the basis of cultural or behavioral fac-tors related to program adherence. 257However, data to evaluate the alternative oradditional contribution of metabolic factorsto this differential are lacking. To date, nopublished studies permit separation of meta-bolic and behavioral influences on theobserved ethnic differences in weight loss.(Also see commentary on pages 89 to 91under treatment guidelines.)
53RECOMMENDATION: The possibility thata standard approach to weight loss willwork differently in diverse patient popula-tions must be considered when settingexpectations about treatment outcomes.Evidence Category B.5. PharmacotherapySince 1995, the use of the prescription drugsfenfluramine or dexfenfluramine for weight losshad increased greatly to 14 million prescriptionsin 11⁄2 years. The increased interest in drug treat-ment of obesity derives from the poor long-termresults often obtained with behavior therapy,including diet and physical activity, as noted ear-lier in this report. The rationale for the additionof drugs to these regimens is that a more suc-cessful weight loss and maintenance may ensue.However, as of September 1997, the FDArequested the voluntary withdrawal of fenflu-ramine and dexfenfluramine from the market,due to a reported association between valvularheart disease with the drugs dexfenfluramineand fenfluramine, alone or combined with phen-termine. 761In November 1997, the FDA pro-vided clearance for marketing the drug sibu-tramine hydrochloride monohydrate for themanagement of obesity, including weight lossand maintenance of weight loss when used inconjunction with a reduced-calorie diet. Due tothe rapidly evolving information regarding theuse of pharmacotherapy for weight loss, thepanel decided to present their critique of thosepharmacotherapy trials meeting their criteria forconsideration.Forty RCT articles evaluated the effect of phar-macotherapy on weight loss. In most studies,advice or behavioral therapy promoting reducedenergy intake and increased physical activity wasincluded in all treatment arms, including theplacebo group.Drugs that have been evaluated alone or in com-bination for at least 1 year by RCTs are phenter-mine, d,l-fenfluramine, dexfenfluramine, sibu-tramine, and orlistat. Two reports addressingdexfenfluramine 390, 506represent a subset of alarger multicenter trial, the InternationalDexfenfluramine Study (INDEX trial). 507Ofthe nine trials with dexfenfluramine, four hadinterventions of approximately 1 year, three ofapproximately 6 months, and two of 3 months.Most of the trials included a majority of whitewomen, and some included persons with dia-betes. Most of the trials have weight regain data.One RCT 508evaluated the effect of dexfenflu-ramine versus placebo on patients who had beenon an 8-week VLCD for an additional 24 weeks.One RCT 509evaluated the effect of phentermineversus placebo on patients with osteoarthritis.One RCT 510evaluated the effect of variousdoses of sibutramine on weight loss over a 24-week period. Two RCTs of 12 and 8 weeks induration studied phenylpropanolamine versusplacebo. 511, 512One used phenylpropanolaminewith diet, 512and the other with diet, exercise,and behavior therapy. 511Phenylpropanolamineis available as an over-the-counter drug. Theaverage weight loss after 8 weeks of treatmentwas 2.59 kg (5.7 lb) for the drug versus 1.07 kg(2.4 lb) for the placebo, a difference of 1.5 kg(3.3 lb). 511Short-term, but not long-term (over1 year), effectiveness has been documented.Evidence Statement: Pharmacotherapy,which has generally been studied alongwith lifestyle modification including dietand physical activity, using dexfenflu-ramine, sibutramine, orlistat, or phenter-mine/fenfluramine, results in weight loss inobese adults when used for 6 months to 1year. Evidence Category B.
54Chapter 3: Examination of Randomized Controlled Trial EvidenceRationale: Data reporting the results of 52 weeksof drug treatment are available only for dexfen-fluramine, orlistat, phentermine, d,l-fenflu-ramine, and sibutramine. The effect on weightloss from these drugs after 1 year is modest: 2.6kg (5.7 lb) difference on average betweendexfenfluramine and placebo (9.8 kg [21.6 lb]versus 7.2 kg [15.9 lb], baseline weights of98.0 kg [216 lb] and 96.6 kg [213 lb]). 507Forsibutramine, a difference of 2.8 kg (6.2 lb) wasseen at 1 year (4.4 kg [9.7 lb] versus 1.6 kg [3.5lb]). 513In another study of sibutramine, the dif-ference was a net loss of 5.5 kg (12.1 lb) at 24weeks with a 20 mg dose. 510In all of the weightloss studies that were up to 1 year in duration,the majority of the weight loss occurred in thefirst 6 months with a plateauing or actualincrease in weight in the following 6 months. 388,389, 507, 514For orlistat, a difference of 2.2 kg (4.9lb) (4.3 kg [9.5 lb] versus 2.1 kg [4.6 lb]) wasfound in a 16-week study. 407After a 210-weekstudy on phentermine and d,l-fenfluramine, par-ticipants averaged a weight loss of 1.4 kg (3.1 lb)from baseline weight at week 210.395, 514There are no known pharmacologic agentswhose specific effect is to reduce abdominal fator to improve cardiorespiratory fitness. Anyimprovements in abdominal fat or fitness wouldbe secondary to weight loss or other interven-tions.Sibutramine is approved for long-term use.(Note: FDA approval of orlistat is pending a res-olution of labeling issues and results of Phase IIItrials.) Such drugs can then be used as anadjunct to modifications in behavior, includingdiet and physical activity. Drugs should be dis-continued if significant weight loss is notachieved, i.e., a loss of < 4 pounds in the first 4weeks, or if serious adverse effects occur.RECOMMENDATION: Weight loss drugsmay only be used as part of a comprehen-sive weight loss program including diet andphysical activity for patients with a BMIof ≥30 with no concomitant obesity-relat-ed risk factors or diseases, or for patientswith a BMI of ≥27 with concomitantobesity-related risk factors or diseases.Evidence Category B.6. SurgeryFourteen RCTs compared the weight-reducingeffect of different surgical interventions. Theydid not compare surgery versus no intervention.Because of the nature of the intervention, thereis a reluctance to do such a randomized trial.Evidence Statement: Surgical interven-tions in adults with a BMI ≥ 40 or aBMI ≥ 35 with comorbid conditionsresult in substantial weight loss. EvidenceCategory B.Rationale: Of the 14 RCTs that examined theeffect of surgical procedures on weight loss, 8were deemed appropriate. All the studies includ-ed individuals who had a BMI of 40 kg/m2orabove, or a BMI of 35 to 40 kg/m2with comor-bidity; most of the study participants werewomen. Weight loss due to surgical interventionsuch as the gastric bypass ranged from 50 kg(110 lb) to as much as 100 kg (220 lb) over aperiod of 6 months to 1 year. Gastroplasty withdiet had a favorable net outcome on weight lossafter 2 years compared to diet alone. 515Vertical-banded gastroplasty was more effective than hor-izontal-banded gastroplasty. 516Gastric resectionwith a modest biliopancreatic diversion withoutintestinal exclusion resulted in significantlygreater weight loss than conventional Roux-en Y
55gastric bypass; this long-limb modification ofRoux-en Y gastric bypass was shown to be safeand effective in patients who were 200 lb ormore overweight and did not cause additionalmetabolic sequelae or diarrhea. 517The Swedish Obesity Study (SOS), 518a non-RCT, found that gastric bypass produced greaterweight loss than gastroplasty, 42.3 kg (93.3 lb)versus 29.9 kg (67 lb), at 1 year. The gastricbypass was deemed superior to gastroplasty orgastric partitioning in other RCTs as well. 518-522The SOS also evaluated the correlation betweenstoma size and weight loss in patients assigned togastric bypass or gastroplasty, and found no sig-nificant correlation in the gastric bypass group.Comorbidity factors associated with weight lossshowed improvement after surgery. One studyshowed that medical illnesses either improved(47 percent) or resolved (43 percent) in all butfour patients (9 percent), and these four hadunsatisfactory weight loss. 517The Adelaide Studyshowed that 60 percent of the patients who ini-tially had obesity-related comorbidity were freeof medication for these comorbidities 3 yearsafter surgery. 519Extremely obese persons often do not benefitfrom the more conservative treatments forweight loss and weight maintenance. Obesityseverely impairs quality of life, and these indi-viduals are at higher risk for premature death. 523The National Institutes of Health ConsensusDevelopment Conference consensus statement,“Gastrointestinal Surgery for Severe Obesity, 523”concluded that the benefits outweigh the risksand that this more aggressive approach is reason-able in individuals who strongly desire substan-tial weight loss and have life-threatening comor-bid conditions.The effects of surgery on abdominal fat or car-diorespiratory fitness independent of weight lossare unknown. One small study demonstratedthat surgery (adjustable silicone gastric banding)reduced abdominal visceral fat. No analyses wereperformed to examine whether visceral fatreduction was independent of weight loss. 524There are no data on the effects of surgery onchanges in cardiorespiratory fitness.RECOMMENDATION: Surgical interven-tion is an option for carefully selectedpatients with clinically severe obesity (aBMI ≥40 or ≥35 with comorbid condi-tions) when less invasive methods of weightloss have failed and the patient is at highrisk for obesity-associated morbidity andmortality. Evidence Category B.7. Other Interventions for Overweight andObesity TreatmentOther types of interventions considered for over-weight and obesity treatment included acupunc-ture, herbal remedies, supplements, and hypno-sis. However, these treatments did not fulfill thea priori inclusion criteria, and for that reasonwere not included in this review.
564TREATMENT GUIDELINESOVERVIEWThe presence of overweight and obesity in apatient is of medical concern because it increasesthe risk for several diseases, particularly cardio-vascular diseases (CVDs) and diabetes mellitus(see Chapter 2.C.) and it increases all-causemortality. Treatment of the overweight andobese patient is a two-step process: assessmentand management. Assessment requires determi-nation of the degree of obesity and absolute riskstatus. Management includes both weight con-trol or reducing excess body weight and main-taining that weight loss as well as institutingother measures to control associated risk factors.The aim of this guideline is to provide usefuladvice on how to achieve weight reduction andmaintenance of a lower body weight. Obesity isa chronic disease, and both the patient and thepractitioner need to understand that successfultreatment requires a life-long effort.ASSESSMENT AND CLASSIFICATION OFOVERWEIGHT AND OBESITY1. Assessment of Overweight and Obesitys Determination of total body fat. Overweight isdefined as a body mass index (BMI) of 25 to29.9 kg/m2. Obesity is defined as an excess oftotal body fat that is documented by a BMIof ≥30 kg/m2. Several methods are availablefor determining or calculating total body fat:total body water, total body potassium, bio-electrical impedance, and dual-energy X-rayabsorptiometry.Evidence Statement: Measures of body fatgive reasonably equivalent values for fol-lowing overweight or obese patients duringtreatment. Evidence Category D.Rationale: Even though accurate methods toassess body fat exist, measuring body fat contentby these techniques is often expensive and is notreadily available clinically. Although bioelectricalimpedance devices are becoming more readilyavailable, they lose accuracy in severely obesepersons and are of limited usefulness for trackingchanges in total body fat in persons losingweight. Thus, bioelectrical impedance offers nosignificant advantage over BMI in the clinicalmanagement of patients. No trial data exist toindicate that one measure of fatness is betterthan any other for following overweight andobese patients during treatment. No studies havebeen published to compare the effectiveness ofdifferent measures for evaluating changes inbody fat during weight reduction. BMI providesa more accurate measure of total body fat thanrelying on weight alone. It has an advantage overpercent above ideal weight (e.g., based on theMetropolitan Life Insurance Tables). Ideal bodyweight tables were developed primarily fromwhite, higher socioeconomic status populationsand have not been documented to accuratelyreflect body fat content in the public at large. Inaddition, separate tables are required for menand women. The weight tables also are based onAB
57mortality outcomes and do not necessarily pre-dict morbidity. BMI is recommended as a prac-tical approach for the clinical setting. BMI pro-vides an acceptable approximation for assess-ment of total body fat for the majority ofpatients. 525-527However, simply measuring bodyweight is a practical approach to follow weightchanges.RECOMMENDATION: Practitioners shoulduse the BMI to assess overweight and obesi-ty. Body weight alone can be used to followweight loss and to determine efficacy oftherapy. Evidence Category C.Rationale: The panel concentrated on toolsavailable in the office, i.e., weight, height andthe BMI. BMI is a practical indicator of theseverity of obesity, and it can be calculated fromexisting tables. BMI is a direct calculation basedon height and weight, regardless of gender. Thelimitations of BMI as a measure of total bodyfat, nonetheless, must be recognized. For exam-ple, BMI overestimates body fat in persons whoare very muscular and can underestimate bodyfat in persons who have lost muscle mass (e.g.,the elderly).The BMI is calculated as follows:BMI = weight (kg)/height squared (m2)To estimate BMI from pounds and inches use:[weight (pounds)/ height (inches)2] x 703 528(1 lb = 0.4536 kg)(1 in = 2.54 cm = 0.0254 m)A simple BMI chart is provided in Appendix V.A patient should be weighed with shoes off andclad only in a light robe or undergarments.s Determination of degree of abdominal obesity.The presence of excess fat in the abdomen,out of proportion to total body fat, is anindependent predictor of risk factors andmorbidity. Relatively accurate measures oftotal abdominal fat can be made by magneticresonance imaging 529or computed tomogra-phy. 500, 530These methods, however, are expen-sive and not readily available for clinical prac-tice. Research with these techniques, nonethe-less, has shown that the waist circumferencecorrelates with the amount of fat in theabdomen, and thus is an indicator of theseverity of abdominal obesity. In this docu-ment the term “waist” circumference is usedinstead of “abdominal” circumference becauseit more accurately describes the anatomicalsite of measurement.Evidence Statement: Waist circumferenceis the most practical anthropometric mea-surement for assessing a patient’s abdomi-nal fat content before and during weightloss treatment. Computed tomography andmagnetic resonance imaging are both moreaccurate but are impractical for routineclinical use. Evidence Category C.Rationale: Fat located in the abdominal region isassociated with greater health risks than that inperipheral regions, e.g., the gluteal-femoral area.155-159, 429, 430Abdominal fat is described as havingthree compartments: visceral, retroperitoneal,and subcutaneous. 529, 531, 532Several studies sug-gest that the visceral fat component of abdomi-nal fat is the most strongly correlated with riskfactors; 160, 161, 533, 534other studies, however, indi-cate that the subcutaneous component is themost highly correlated with insulin resistance. 529,535, 536Therefore, the relative contributions of dif-ferent components of abdominal fat to overallrisk remain to be determined with certainty.Nonetheless, the presence of increased totalabdominal fat appears to be an independent risk
58Chapter 4: Treatment Guidelinespredictor when the BMI is not markedlyincreased. 537Therefore, waist or abdominal cir-cumference, as well as BMI, should be measurednot only for the initial assessment of obesity, butalso as a guide to the efficacy of weight losstreatment.The waist-to-hip ratio (WHR) also has beenused in a number of epidemiologic studies toshow increased risk for diabetes, coronary arterydisease, and hypertension. 500However, waist cir-cumference has been found to be a better mark-er of abdominal fat content than is WHR. 85Whether WHR imparts any independent infor-mation about disease risk beyond waist circum-ference is uncertain, but between the two, thewaist circumference appears to carry greaterprognostic significance. Therefore, in clinicalpractice, abdominal fat content should beassessed and followed by measuring waist cir-cumference.RECOMMENDATION: The waist circum-ference should be used to assess abdominalfat content. Evidence Category C.2. Classification of Overweight and Obesitys According to BMI. The primary classificationof obesity is based on the measurement ofBMI. This classification is designed to relateBMI to risk of disease. It should be notedthat the relation between BMI and diseaserisk varies among individuals and among dif-ferent populations. Therefore, the classifica-tion must be viewed as a broad generalization.Individuals who are very muscular may have aBMI placing them in an overweight categorywhen they are not overly fat. Also, very shortpersons (under 5 feet) may have high BMIsthat may not reflect overweight or fatness. Inaddition, susceptibility to risk factors at agiven weight varies among individuals. Someindividuals may have multiple risk factorsInstructions for Measuring WaistCircumference, According to NHANES IIIProtocolTo define the level at which waist circum-ference is measured, a bony landmark isfirst located and marked. The subjectstands and the examiner, positioned at theright of the subject, palpates the upper hipbone to locate the right iliac crest. Justabove the uppermost lateral border of theright iliac crest, a horizontal mark is drawn,then crossed with a vertical mark on themidaxillary line. The measuring tape isplaced in a horizontal plane around theabdomen at the level of this marked pointon the right side of the trunk. The plane ofthe tape is parallel to the floor and the tapeis snug, but does not compress the skin.The measurement is made at a normalminimal respiration (see Figure 5).REF: U.S. Department of Health andHuman Services, PHS. NHANES IIIAnthropometric Procedures Video. U.S.Government Printing Office Stock Number017-022-01335-5. Washington, D.C.: U.S.GPO, Public Health Service; 1996. 538
59with mild obesity, whereas others may havefewer risk factors with more severe obesity. Itshould also be noted that the risk levelsshown for each increment in risk are relativerisks; that is, relative to risk at normal weight.They should not be equated with absoluterisk which is determined by a summation ofrisk factors. No randomized controlled trialstudies exist that support a specific system forclassification that establishes the degree of dis-ease risk for patients during weight loss orweight maintenance. The classification isbased on observational and prospective epi-demiological studies.Figure 5. Measuring tape position for waist (abdominal) circumference
60Chapter 4: Treatment GuidelinesRECOMMENDATION: The BMI * shouldbe used to classify overweight and obesityand to estimate relative risk for diseasecompared to normal weight. EvidenceCategory C.Rationale: Overweight and obesity are classifiedby BMI (see Table IV-1).Calculating BMI is simple, rapid, and inexpen-sive. The classification can be applied generallyto adults. See Appendix III for comments aboutobesity and its classification in children. Thebasis for this BMI classification scheme stemsfrom observational and epidemiologic studieswhich relate BMI to risk of morbidity and mor-tality. 11, 81-83, 131, 155, 278, 539-541For example, relative riskfor CVD increases in a graded fashion withincreasing BMI in all population groups, althoughabsolute risk in obese versus nonobese personsdepends on the summed contribution of all riskfactors.Evidence Statement: The same BMI cut-points (see Table IV-1) can be used to clas-sify the level of overweight and obesity foradult men and adult nonpregnant women,and generally for all racial/ethnic groups.Evidence Category C.Rationale: Clinical judgment must be used ininterpreting BMI in situations in which it maynot be an accurate indicator of total body fat.Examples are the presence of edema, high mus-cularity, muscle wasting, or for very short peo-ple. The relationship between BMI and body fatcontent varies somewhat with age, sex, and pos-sibly ethnicity because of differences in factorssuch as composition of lean tissue, sittingheight, and hydration state. 526, 542For example,older persons often have lost muscle mass andhave more fat for a given BMI than youngerpersons, women may have more fat for a givenBMI than men, and persons with clinical edemaSource (adapted from): Preventing and Managing the Global Epidemic of Obesity. Report of the World HealthOrganization Consultation of Obesity. WHO, Geneva, June 1997.65C L A S S I F I C A T I O N O F O V E R W E I G H T A N D O B E S I T Y B Y B M I *Obesity Class BMI (kg/m2)Underweight < 18.5Normal 18.5 – 24.9Overweight 25.0 – 29.9Obesity I 30.0 – 34.9II 35.0 – 39.9Extreme Obesity III ≥ 40* Pregnant women who, on the basis of their prepregnant weight, would be classified as obese may encounter certain obstetrical risks.However, the inappropriateness of weight reduction during pregnancy is well recognized. 44Hence, this guideline specifically excludespregnant women.TABLE IV-1:
61may have less fat for a given BMI than personswithout edema. Nevertheless, these differencesgenerally do not markedly influence the validityof BMI cutpoints either for classifying individu-als into broad categories of overweight and obe-sity or for monitoring weight status of individu-als in clinical settings. 526s According to waist circumference. Althoughwaist circumference and BMI are interrelated,waist circumference provides an independentprediction of risk over and above that ofBMI. Waist circumference measurement isparticularly useful in patients who are catego-rized as normal or overweight on the BMIscale. At BMIs ≥ 35, waist circumference haslittle added predictive power of disease riskbeyond that of BMI. It is therefore not neces-sary to measure waist circumference in indi-viduals with BMIs ≥ 35.Evidence Statement: Sex-specific cutoffsfor waist circumference can be used toidentify increased risk associated withabdominal fat in adults with a BMI in therange of 25 to 34.9 kg/m2. An increase inwaist circumference may also be associatedwith increased risk in persons of normalweight. Evidence Category C.Waist circumference cutpoints can generallybe applied to all adult ethnic or racialgroups. On the other hand, if a patient isvery short (under 5 feet) or has a BMIabove the 25 to 34.9 kg/m2range, waistcutpoints used for the general populationmay not be applicable. EvidenceCategory D.Rationale: A high waist circumference is associ-ated with an increased risk for type 2 diabetes,dyslipidemia, hypertension, and CVD inpatients with a BMI in a range between 25 and34.9 kg/m.2 82Monitoring changes in waist cir-cumference over time may be helpful, in addi-tion to measuring BMI, since it can provide anestimate of increased abdominal fat even in theabsence of a change in BMI. Furthermore, inobese patients with metabolic complications,changes in waist circumference are useful predic-tors of changes in CVD risk factors. 537Thewaist circumference at which there is anincreased relative risk is defined as follows:To evaluate health risks, it is not necessary tomeasure waist circumference in patients with aBMI ≥ 35 kg/m2; the measurement usually willbe greater than the cutpoints given above, andwaist measurements lose their predictive powerat very high BMIs.There are ethnic and age-related differences inbody fat distribution that modify the predictivevalidity of waist circumference as a surrogate forabdominal fat. 526These variations may partlyexplain differences between ethnic or age groupsin the power of waist circumference or WHR topredict disease risks. 429, 543In some populations, waist circumference is abetter indicator of relative disease risk than isBMI; examples include Asian-Americans or per-sons of Asian descent living elsewhere. 51, 273, 544Waist circumference also assumes greater valuefor estimating risk for obesity-related disease atolder ages. Table IV-2 incorporates both BMIand waist circumference in the classification ofoverweight and obesity, and provides an indica-tion of disease risk.H I G H R I S KMen > 102 cm ( > 40 in)Women > 88 cm ( > 35 in)
62Chapter 4: Treatment GuidelinesRECOMMENDATION: For adult patientswith a BMI of 25 to 34.9 kg/m2, sex-specif-ic waist circumference cutoffs should be usedin conjunction with BMI to identifyincreased disease risk. Evidence Category C.ASSESSMENT OF RISK STATUSThe patient’s risk status should be assessed bydetermining the degree of overweight or obesitybased on BMI, the presence of abdominal obesi-ty based on waist circumference, and the pres-ence of concomitant CVD risk factors or comor-bidities. Some obesity-associated diseases andrisk factors place patients in a very high risk cat-egory for subsequent mortality. These diseaseswill require aggressive modification of risk fac-tors in addition to their own clinical manage-ment. Other obesity-associated diseases are lesslethal, but still require appropriate clinical thera-py. Obesity also has an aggravating influence onseveral cardiovascular risk factors. Identifyingthese risk factors is required as a guide to theintensity of clinical intervention.1. Determination of relative risk status basedon overweight and obesity parameters.Table IV-2 defines relative risk categories accord-ing to BMI and waist circumference. It is impor-tant to note that these categories denote relativerisk, not absolute risk. They relate to the need toinstitute weight loss therapy, and do not directlydefine the required intensity of risk factor modi-fication. The latter is determined by estimationSource (adapted from): Preventing and Managing the Global Epidemic of Obesity. Report of the World HealthOrganization Consultation of Obesity. WHO, Geneva, June 1997.65CL A S S I F I C AT I O N O F OV E R W E I G H T A N D OB E S I T Y B Y BMI, WA I S TCI R C U M F E R E N C E A N D AS S O C I AT E D DI S E A S E RI S K *Disease Risk* Relative to Normal Weightand Waist CircumferenceObesity Men ≤102 cm (≤40 in) >102 cm (>40 in)BMI (kg/m2) Class Women ≤88 cm (≤35 in) >88 cm (>35 in)Underweight <18.5 — —Normal+18.5 – 24.9 — —Overweight 25.0 – 29.9 Increased HighObesity 30.0 – 34.9 I High Very High35.0 – 39.9 II Very High Very HighExtreme Obesity ≥40 III Extremely High Extremely High* Disease risk for type 2 diabetes, hypertension, and CVD.+ Increased waist circumference can also be a marker for increased risk even in persons of normal weight.CTABLE IV-2:
63of absolute risk based on the presence of associ-ated disease or risk factors.2. Identification of patients at very highabsolute risk.The following disease conditions or target organdamage in hypertensive patients denote the pres-ence of very high absolute risk that triggers theneed for intense risk factor modification as wellas disease management. For example, the pres-ence of very high absolute risk indicates the needfor aggressive cholesterol-lowering therapy. 1422.a. Established coronary heart disease (CHD)2.a.1. History of myocardial infarction2.a.2. History of angina pectoris (stable orunstable)2.a.3. History of coronary artery surgery2.a.4. History of coronary artery procedures(angioplasty)2.b. Presence of other atherosclerotic diseases2.b.1. Peripheral arterial disease2.b.2. Abdominal aortic aneurysm2.b.3. Symptomatic carotid artery disease2.c. Type 2 diabetes2.d. Sleep apnea (For more information, seeAppendix IV.)3. Identification of other obesity-associateddiseases.Obese patients are at increased risk for severalconditions that require detection and appropri-ate management, but that generally do not leadto widespread or life-threatening consequences.These include:3.a. Gynecological abnormalities3.b. Osteoarthritis3.c. Gallstones and their complications3.d. Stress incontinenceManagement options of risk factors for pre-venting CVD, diabetes mellitus, and otherchronic diseases are described in detail inother reports. For details on the managementof serum cholesterol and other lipoprotein dis-orders, refer to the National CholesterolEducation Program’s Second Report of theExpert Panel on the Detection, Evaluation,and Treatment of High Blood Cholesterolin Adults (Adult Treatment Panel II/ATP II)(1993). 142For the treatment of hypertension,the National High Blood Pressure EducationProgram recently issued the Sixth Report ofthe Joint National Committee onPrevention, Detection, Evaluation, andTreatment of High Blood Pressure (JNC VI)(1997). 545For the most recent recommenda-tions about type 2 diabetes from the AmericanDiabetes Association, see the ADA ClinicalPractice Recommendations (1998). 546Finally,smoking cessation is strongly recommended inobese smokers, with particular attention tomethods that diminish the weight gain associ-ated with cessation (see pages 91-92). Forstrategies and recommendations for supportingand delivering effective smoking cessationintervention, see the Agency for Health CarePolicy and Research’s Clinical PracticeGuidelines on Smoking Cessation. 547
64Chapter 4: Treatment GuidelinesAlthough obese patients are at increased risk forgallstones, this risk increases even more duringperiods of rapid weight reduction.4. Identification of cardiovascular risk factorsthat impart a high absolute risk.Patients can be classified as being at highabsolute risk for obesity-related disorders if theyhave three or more of the multiple risk factorslisted below. The presence of high absolute riskincreases the intensity of cholesterol-loweringtherapy 142and blood pressure management. 5454.a. Cigarette smoking4.b. Hypertension—A patient is classified ashaving hypertension if systolic blood pressure is≥ 140 mm Hg or diastolic blood pressure is≥ 90 mm Hg, or if the patient is taking antihy-pertensive agents.4.c. High-risk low-density lipoprotein choles-terol—A high-risk LDL-cholesterol is defined asa serum concentration of ≥160 mg/dL. A bor-derline high-risk LDL-cholesterol (130 to 159mg/dL) together with two or more other riskfactors also confers high risk.4.d. Low high-density lipoprotein cholesterol—A low HDL-cholesterol is defined as a serumconcentration of < 35 mg/dL.4.e. Impaired fasting glucose (IFG)—The pres-ence of clinical type 2 diabetes (fasting plasmaglucose of ≥ 126 mg/dL or 2 hours postprandialplasma glucose of ≥ 200 mg/dL) is a major riskfactor for CVD, and its presence alone places apatient in the category of very high absolute risk(see above). IFG (fasting plasma glucose 110 to125 mg/dL) is considered by many authoritiesto be an independent risk factor for cardiovascu-lar (macrovascular) disease, justifying its inclu-sion among risk factors contributing to highabsolute risk. Although including IFG as a sepa-rate risk factor for CVD departs from the ATPII and JNC VI reports, its inclusion in this listmay be appropriate. IFG is well established as arisk factor for type 2 diabetes.4.f. Family history of premature CHD—A posi-tive family history of premature CHD is definedas definite myocardial infarction or suddendeath at or before 55 years of age in the fatheror other male first-degree relative, or at or before65 years of age in the mother or other femalefirst-degree relative.4.g. Age4.g.1. Male ≥ 45 years4.g.2. Female ≥ 55 years (or postmenopausal)Methods for estimating absolute risk status fordeveloping CVD based on these risk factors aredescribed in detail in the ATP II and JNC VIreports. The intensity of intervention for highblood cholesterol or hypertension is adjusteddepending on the absolute risk estimated bythese factors. Approaches to therapy for choles-terol disorders and hypertension are described inthe ATP II and JNC VI, respectively.5. Other risk factors deserve special considera-tion for their relation to obesity. When these fac-tors are present, patients can be considered tohave incremental absolute risk above that esti-mated from the preceding risk factors. Quanti-tative risk contributions are not available forthese risk factors, but their presence heightensthe need for weight reduction in obese persons.5.a. Physical inactivity—A lack of physical activ-ity imparts an increased risk for both CVD andtype 2 diabetes. Physical inactivity enhances theseverity of other risk factors, but it also has beenshown to be an “independent” risk factor for all-cause mortality or CVD mortality. 411, 548Although physical inactivity is not listed as a riskfactor that modifies the intensity of therapyrequired for elevated cholesterol or blood pres-sure, increased physical activity is indicated formanagement of these conditions (see the ATP IIand JNC VI). Increased physical activity is espe-
65cially needed in obese patients because it pro-motes weight reduction and favorably modifiesobesity-associated risk factors. Conversely, thepresence of physical inactivity in an obese personwarrants intensified efforts to remove excessbody weight, because physical inactivity andobesity both heighten disease risks.5.b. High triglycerides—Obesity is commonlyaccompanied by elevated serum triglycerides.The relationship between high triglycerides andCHD is complex. Triglyceride-rich lipoproteinsmay be directly atherogenic. In addition, elevat-ed serum triglycerides are the most commonmanifestation of the atherogenic lipoproteinphenotype (high triglycerides, small LDL parti-cles, and low HDL-cholesterol levels). 142, 549Moreover, in the presence of obesity, high serumtriglycerides are commonly associated with aclustering of metabolic risk factors known as themetabolic syndrome (atherogenic lipoproteinphenotype, hypertension, insulin resistance andglucose intolerance, and prothrombotic states).Thus, in obese patients, elevated serum triglyc-erides are a marker for increased cardiovascularrisk. According to current guidelines (ATP IIand JNC VI), the presence of high triglyceridesdoes not modify the intensity of cholesterol orblood pressure-lowering therapy. Their presencein obese patients, however, calls for an intensi-fied effort to achieve weight reduction andincrease physical activity. Both will reduce thevarious risk factors characteristic of the metabol-ic syndrome, and thus should reduce overall car-diovascular risk as well as decrease the risk fortype 2 diabetes.According to the ATP II guide-lines, 142triglyceride levels are clas-sified as follows:Patients with very high triglycerides are atincreased risk for acute pancreatitis and mustundergo immediate triglyceride-loweringtherapy.EVALUATION AND TREATMENT STRATEGYWhen physicians encounter patients in the clini-cal setting, the opportunity exists for identifyingoverweight and obesity and accompanying riskfactors and for initiating treatment for both theweight and the risk factors, as well as chronicdiseases such as CVD and type 2 diabetes.When assessing a patient for treatment of over-weight and obesity, consider the patient’s weight,waist circumference, and the presence of risk fac-tors. The strategy for the evaluation and treat-ment of overweight patients is presented inFigure 6 (Treatment Algorithm). This algorithmapplies only to the assessment for overweightand obesity and subsequent decisions based onthat assessment. It does not reflect any initialoverall testing for other conditions and diseasesthat the physician may wish to do. Approachesto therapy for cholesterol disorders and hyper-tension are described in ATP II and JNC VI,respectively. In overweight patients, control ofcardiovascular risk factors deserves equal empha-sis as weight loss therapy. Reductions of risk fac-tors will reduce the risk for cardiovascular dis-ease whether or not efforts at weight loss are suc-cessful. The panel recognizes that the assessmentfor overweight and obesity may take place aspart of an overall health assessment; however, forclarity, the algorithm focuses only on the evalua-tion and treatment of overweight and obesity.Category Serum Triglyceride LevelsNormal triglycerides Less than 200 mg/dLBorderline-high triglycerides 200 to 400 mg/dLHigh triglycerides 400 to 1,000 mg/dLVery high triglycerides Greater than 1,000 mg/dLD
67Each step (designated by a box) in this process isreviewed in this section and expanded upon insubsequent sections.Box 1: “Patient Encounter”A patient encounter is defined as any interactionbetween a health care practitioner (generally aphysician, nurse practitioner or physician’s assis-tant) that provides the opportunity to assess apatient’s weight status and provide advice, coun-seling, or treatment.Box 2: “History of Overweight or RecordedBMI ≥ 25”The practitioner must seek to determinewhether the patient has ever been overweight.While a technical definition is provided, a sim-ple question such as “Have you ever been over-weight?” will accomplish the same goal.Questions directed towards weight history,dietary habits, physical activities, and medica-tions may provide useful information about theorigins of obesity in particular patients.Box 3: “BMI Measured in Past 2 Years”For those who have not been overweight, a 2-year interval is appropriate for the reassessmentof BMI. While this time span is not evidence-based, it is believed to be a reasonable compro-mise between the need to identify weight gain atan early stage and the need to limit the time,effort, and cost of repeated measurements.Box 4: “Measure Weight, Height, WaistCircumference; Calculate BMI”Weight must be measured so that the BMI canbe calculated. Most charts are based on weightsobtained with the patient wearing undergar-ments and no shoes. BMI can be manually cal-culated (kg/[height in meters]2), but is more eas-ily obtained from a nomogram (see AppendixV). Waist circumference is important becauseevidence suggests that abdominal fat is a particu-larly strong determinant of cardiovascular risk inthose with a BMI of 25 to 34.9 kg/m2. Increasedwaist circumference can also be a marker ofincreased risk even in persons of normal weight.The technique for measuring waist circumfer-ence is described on page 58. A nutrition assess-ment will also help to assess the diet and physi-cal activity habits of overweight patients.Box 5: “BMI ≥ 25, OR Waist Circumference >88 cm(F) or > 102 cm(M)”These cutpoints divide overweight from normalweight and are consistent with other nationaland international guidelines. The relationbetween weight and mortality is J-shaped, andevidence suggests that the right side of the “J”begins to rise at a BMI of 25. Waist circumfer-ence is incorporated as an “or” factor becausesome patients with BMI lower than 25 will havedisproportionate abdominal fat, and this increas-es their cardiovascular risk despite their low BMI(see page 61). These abdominal circumferencevalues are not necessary for patients with a BMI≥ 35 kg/m2.Box 6: “Assess Risk Factors”Risk assessment for CVD and diabetes in a per-son with evident obesity will include specialconsiderations for the history, physical examina-tion, and laboratory examination. Of greatesturgency is the need to detect existing CVD orend-organ damage. Since the major risk of obe-sity is indirect (obesity elicits or aggravateshypertension, dyslipidemias, and diabetes, whichcause cardiovascular complications), the manage-ment of obesity should be implemented in thecontext of these other risk factors. While there isno direct evidence demonstrating that addressingrisk factors increases weight loss, treating the riskfactors through weight loss is a recommendedstrategy. The risk factors that should be consid-ered are provided on pages 63-65. With regardto age, please see commentary on pages 90-91.
68Chapter 4: Treatment GuidelinesBox 7: “BMI ≥ 30, OR ([BMI 25 to 29.9 OR WaistCircumference > 88 cm(F) or > 102 cm(M)] AND≥ 2 risk factors)”The panel recommends that all patients meetingthese criteria attempt to lose weight. However, itis important to ask the patient whether or notthey want to lose weight. Those with BMIsbetween 25 and 29.9 kg/m2who have one or norisk factors should work on maintaining theircurrent weight rather than embark on a weightreduction program. The panel recognizes thatthe decision to lose weight must be made in thecontext of other risk factors (e.g., quitting smok-ing is more important than losing weight) andpatient preferences.Box 8: “Clinician and Patient Devise Goals”The decision to lose weight must be madejointly between the clinician and patient.Patient involvement and investment is crucialto success. The patient may choose not to loseweight but rather to prevent further weightgain as a goal. The panel recommends as aninitial goal the loss of 10 percent of baselineweight, to be lost at a rate of 1 to 2 lb/week,establishing an energy deficit of 500 to 1,000kcal/day. (See pages 71-72.) For individualswho are overweight, a deficit or 300 to 500kcal/day may be more appropriate, providinga weight loss of about 1/2 lb/week. Also, thereis evidence that an average of 8 percent ofweight can be lost in a 6-month period. Sincethe observed average 8 percent weight lossincludes people who do not lose weight, anindividual goal of 10 percent is reasonable.After 6 months, most patients will equilibrate(caloric intake balancing energy expenditure)and will require adjustment of energy balanceif they are to lose more weight. (See pages72-73.)The three major components of weight losstherapy are dietary therapy, increased physicalactivity, and behavior therapy. (See pages 73-82.) Lifestyle therapy should be tried for atleast six months before considering pharma-cotherapy. In addition, pharmacotherapyshould be considered as an adjunct to lifestyletherapy in patients with a BMI ≥ 30 with noconcomitant obesity-related risk factors or dis-eases, or for patients with a BMI ≥ 27 withconcomitant obesity-related risk factors or dis-eases. The risk factors or diseases consideredimportant enough to warrant pharmacothera-py at a BMI of 27 to 29.9 are hypertension,dyslipidemia, CHD, type 2 diabetes, andsleep apnea. However, sibutramine, the onlyFDA approved drug for long-term use, shouldnot be used in patients with a history ofhypertension, CHD, congestive heart failure,arrhythmias, or history of stroke. Certainpatients may be candidates for weight losssurgery. Each component of weight loss thera-py can be introduced briefly. The selection ofweight loss methods should be made in thecontext of patient preferences, analysis of pastfailed attempts, and consideration of the avail-able resources.Box 9: “Progress Being Made/Goal Achieved”During the acute weight loss period and at 6-month and 1-year follow-up visits, the patientsshould be weighed, BMI calculated, andprogress assessed. If at any time it appears thatthe program is failing, a reassessment shouldtake place to determine the reasons (see Box 10).If pharmacotherapy is being used, appropriatemonitoring for side effects is recommended (Seepage 86). If a patient can achieve the recom-mended 10 percent reduction in body weight in6 months to 1 year, this change in weight can beconsidered good progress. The patient can thenenter the phase of weight maintenance and long-term monitoring. It is important for the practi-tioner to recognize that some persons are moreapt to lose or gain weight on a given regimenand that this phenomenon cannot always beattributed to degree of compliance. However, ifsignificant obesity remains and absolute risk
69from obesity-associated risk factors remains high,at some point an effort should be made to rein-stitute weight loss therapy to achieve furtherweight reduction. Once a limit of weight losshas been obtained, the practitioner is responsiblefor long-term monitoring of risk factors and forencouraging the patient to maintain a reducedweight level.Box 10: “Assess Reasons for Failure to LoseWeight”If a patient fails to achieve the recommended 10percent reduction in body weight in 6 monthsor 1 year, a reevaluation is required. A criticalquestion is whether the level of motivation ishigh enough to continue clinical therapy. Ifmotivation is high, revise the goals and strategies(see Box 8). If motivation is not high, clinicaltherapy should be discontinued, but the patientshould be encouraged to embark on efforts tolose weight or to at least avoid further weightgain. Even if weight loss therapy is stopped, riskfactor management must be continued.Failure to achieve weight loss should prompt thepractitioner to investigate energy intake (dietaryrecall including alcohol intake, daily intakelogs), energy expenditure (physical activitydiary), attendance at behavior therapy groupmeetings, recent negative life events, family andsocietal pressures, or evidence of detrimentalpsychiatric problems (depression, binge eatingdisorder). If attempts to lose weight have failed,and the BMI is ≥ 40, surgical therapy should beconsidered.Box 11: “Maintenance Counseling”Evidence suggests that over 80 percent of per-sons who lose weight will gradually regain it (seepage 72). Patients who continue on weightmaintenance programs have a greater chance ofkeeping weight off. Maintenance consists of con-tinued contact with the health care practitionerfor continued education, support, and medicalmonitoring (see page 72 and 75).Box 12: “Does the Patient Want to LoseWeight?”All patients who are overweight (BMI 25 to29.9), or do not have a high waist circumfer-ence, and have few (0 to 1) cardiovascular riskfactors and do not want to lose weight, shouldbe counseled regarding the need to keep theirweight at or below its present level. Patients whowish to lose weight should be guided per Boxes8 and 9. The justification for offering these over-weight patients the option of maintaining(rather than losing) weight is that their healthrisk, while higher than that of persons with aBMI < 25, is only moderately increased (page62).Box 13: “Advise to Maintain Weight/AddressOther Risk Factors”Those who have a history of overweight and arenow at appropriate weight, and those who areoverweight and not obese but wish to focus onmaintenance of their current weight, should beprovided with counseling and advice so thattheir weight does not increase. An increase inweight increases their health risk and should beprevented. The physician should actively pro-mote prevention strategies including enhancedattention by the patient to diet, physical activity,and behavior therapy. For addressing other riskfactors, see Box 6, because even if weight losscannot be addressed, other risk factors should becovered.Box 14: “History of BMI ≥ 25”This box differentiates those who are not over-weight now and never have been from thosewith a history of overweight; see Box 2.Box 15: “Brief Reinforcement”Those who are not overweight and never havebeen should be advised of the importance ofstaying in this category.Box 16: “Periodic Weight, BMI, and WaistCircumference Check”Patients should receive periodic monitoring oftheir weight, BMI, and waist circumference.
70Chapter 4: Treatment GuidelinesPatients who are not overweight or have no his-tory of overweight should be screened for weightgain every 2 years. This time span is a reasonablecompromise between the need to identify weightgain at an early stage and the need to limit thetime, effort, and the cost of repeated measure-ments.EXCLUSION FROM WEIGHT LOSS THERAPYPatients in whom weight loss therapy is notappropriate include most pregnant or lactatingwomen, those with serious psychiatric illness,and patients who have a variety of serious ill-nesses in whom caloric restriction might exacer-bate the illness.PATIENT MOTIVATIONAssessment of patient motivation is a prerequi-site for weight loss therapy. Weight reduction inthe clinical setting represents a major investmentof time and effort on the part of the health careteam and expense to the patient. For these rea-sons, motivation for weight loss should be rela-tively high before initiating clinical therapy. Atthe same time, it is the duty of the primary carepractitioner to heighten a patient’s motivationfor weight loss when such is perceived to be ofsignificant benefit for risk reduction. This can bedone by enumerating the dangers associatedwith persistent obesity and by describing thestrategy for clinically assisted weight reduction.For patients who are not motivated (or able) toenter clinical weight loss therapy, appropriatemanagement of risk factors (e.g., high serumcholesterol, hypertension, smoking, and type 2diabetes) is still necessary. Sustained weightreduction may facilitate control of cardiovascularrisk factors and delay the onset of type 2 dia-betes.Evidence Statement: Patient motivationis a key component for success in a weightloss program. Evidence Category D.Rationale: Assessment of the patient’s motiva-tion for treatment by the practitioner is impor-tant in the decision to initiate a weight loss pro-gram. The following factors must be evaluated:s Reasons and motivation for weight loss—What is the extent of the patient’s seriousnessand readiness to undergo a sustained periodof weight loss at this time? What is thepatient’s current attitude about making a life-long commitment to behavior change?s Previous history of successful and unsuccess-ful weight loss attempts—What factors wereresponsible for previous failures and successesat weight loss or maintenance of normal bodyweight?s Family, friends, and work-site support—Whatis the social framework in which the patientwill attempt to lose weight, and who are thepossible helpers and antagonists to such anattempt?s The patient’s understanding of overweightand obesity and how it contributes to obesity-associated diseases—Does the patient have anappreciation of the dangers of obesity, and arethese dangers of significant concern to thepatient?s Attitude toward physical activity—Is thepatient motivated to enter a program ofincreased physical activity to assist in weightreduction?s Time availability—Is the patient willing tocommit the time required to interact withhealth professionals in long-term weight losstherapy?s Barriers—What are the obstacles that willinterfere with the patient’s ability to imple-ment the suggestions for change?s Financial considerations—Is the patient will-ing to pay for obesity therapy? This mayinclude having to pay for travel to the med-EF
71ical facility, time lost from work, and payingfor professional counseling that is not coveredby insurance.One of the most important aspects of the initialevaluation is to prepare patients for treatment.Reviewing patients’ past attempts at weight lossand explaining how the new treatment plan willbe different can encourage patients and providehope for successful weight loss. It is helpful todiscuss the proposed course of treatment anddescribe necessary behaviors, such as keepingdiaries of food intake and physical activity.Finally, given the social stigmatization that obesepatients often feel, even from health care profes-sionals, the initial evaluation is an opportunityto show the patient respect, concern, andhope.550A patient who has shared feelings aboutbeing overweight and previous attempts to loseweight with a sympathetic listener may be morewilling to consider new ideas and information. Apartnership in which the patient feels supportedand understood can help to sustain the necessarymotivation for the difficult task of weight lossand maintenance. 551,552RECOMMENDATION: Practitioners need toassess the patient’s motivation to enterweight loss therapy; assess the readiness of thepatient to implement the plan; and thentake appropriate steps to motivate thepatient for treatment. Evidence Category D.GOALS OF WEIGHT LOSS ANDMANAGEMENTThe general goals of weight loss and manage-ment are: (1) to reduce body weight; and (2) tomaintain a lower body weight over the longterm; or (3) at a minimum, to prevent furtherweight gain. Specific targets for each of thesegoals can be considered.1. Weight Loss1.a. Target levels for weight loss—The initialtarget goal of weight loss therapy for overweightpatients is to decrease body weight by about 10percent. If this target can be achieved, considera-tion can be given to the next step of furtherweight loss.Evidence Statement: Overweight andobese patients in well-designed programscan achieve a weight loss of as much as 10percent of baseline weight, a weight lossthat can be maintained for a sustainedperiod of time (1 year or longer). EvidenceCategory A.Rationale: The rationale for this initial goal isthat even moderate weight loss, i.e., 10 percentof initial body weight, can significantly decreasethe severity of obesity-associated risk factors. Itcan also set the stage for further weight loss, ifindicated. Available evidence indicates that anaverage weight loss of 8 percent can be achievedin 6 months; however, since the observed aver-age 8 percent includes people who do not loseweight, an individual goal of 10 percent is rea-sonable. This degree of weight loss can be achievedand is realistic, and moderate weight loss can bemaintained over time. It is better to maintain amoderate weight loss over a prolonged periodthan to regain from a marked weight loss. Thelatter is counterproductive in terms of time, costs,and self-esteem. Patients generally will wish tolose more weight than 10 percent, and will needto be counseled and persuaded of the appropri-ateness of this initial goal. 553, 554Further weightloss can be considered after this initial goal isachieved and maintained for 6 months.RECOMMENDATION: The initial goal ofweight loss therapy should be to reducebody weight by approximately 10 percentG
72Chapter 4: Treatment Guidelinesfrom baseline. With success, further weightloss can be attempted, if indicated throughfurther assessment. Evidence Category A.1.b. Rate of weight loss—A reasonable time linefor weight loss is to achieve a 10 percent reduc-tion in body weight over 6 months of therapy.For overweight patients with BMIs in the typicalrange of 27 to 35, a decrease of 300 to 500kcal/day will result in weight losses of about 1⁄2 to1 lb/week and a 10 percent weight loss in 6months. For more severely obese patients withBMIs ≥ 35, deficits of up to 500 to 1,000kcal/day will lead to weight losses of about 1 to2 lb/week and a 10 percent weight loss in 6months.Evidence Statement: Weight loss at therate of 1 to 2 lb/week (calorie deficit of500 to 1,000 kcal/day) commonly occursfor up to 6 months, at which point weightloss begins to plateau unless a more restric-tive regimen is implemented. EvidenceCategory B.Rationale: To achieve significant weight loss, anenergy deficit must be created and maintained.Weight can be lost at a rate of 1 to 2 lb/weekwith a calorie deficit of 500 to 1,000 kcal/day.After 6 months, this caloric deficit theoreticallyshould result in a loss of 26 to 52 lb. However,the average amount of weight lost actuallyobserved over this time period usually is in therange of 20 to 25 lb. A greater rate of weightloss does not yield a better result at the end of 1year. 437It is difficult for most patients to contin-ue to lose weight after a period of 6 months dueto changes in resting metabolic rates and diffi-culty in adhering to treatment strategies,although some can do so. To continue to loseweight, diet and physical activity goals need tobe revised to create an energy deficit at the lowerweight, since energy requirements decrease asweight is decreased. To achieve additional weightloss, the patient must further decrease caloriesand/or increase physical activity. Many studiesshow that rapid weight reduction almost alwaysis followed by regaining of weight. Moreover,with rapid weight reduction, there is anincreased risk for gallstones and, possibly, elec-trolyte abnormalities.RECOMMENDATION: Weight loss shouldbe about 1 to 2 lb/week for a period of 6months with the subsequent strategy basedon the amount of weight lost. EvidenceCategory B.2. Weight Maintenance at Lower WeightOnce the goals of weight loss have been success-fully achieved, maintenance of a lower bodyweight becomes a major challenge. In the past,obtaining the goal of weight loss has been con-sidered the end of weight loss therapy.Unfortunately, once patients are dismissed fromclinical therapy, they frequently regain the lostweight. This report recommends that observa-tion, monitoring, and encouragement of patientswho have successfully lost weight be continuedon a long-term basis.Evidence Statement: After 6 months ofweight loss treatment, efforts to maintainweight loss through diet, physical activity,and behavior therapy are important.Evidence Category B.Rationale: After 6 months of weight loss, therate of weight loss usually declines and plateaus.395, 507, 555The primary care practitioner andpatient should recognize that at this point, weight
73maintenance, the second phase of the weightloss effort, should take priority. Successfulweight maintenance is defined as a weight regainof < 3 kg (6.6 lb) in 2 years and a sustainedreduction in waist circumference of at least 4cm. If a patient wishes to lose more weight aftera period of weight maintenance, the procedurefor weight loss outlined above can be repeated.RECOMMENDATION: A weight mainte-nance program should be a priority afterthe initial 6 months of weight loss therapy.Evidence Category B.Evidence Statement: Lost weight usuallywill be regained unless a weight mainte-nance program consisting of dietary thera-py, physical activity, and behavior therapyis continued indefinitely. Drug therapy canalso be used; however, drug safety and effi-cacy beyond 1 year of total treatment havenot been established. Evidence Category B.Rationale: After a patient has achieved the goalsof weight loss, the combined modalities of thera-py (dietary therapy, physical activity, and behav-ior therapy) must be continued indefinitely; oth-erwise, excess weight likely will be regained.Numerous strategies are available for motivatingthe patient; all of these require that the practi-tioner continue to communicate frequently withthe patient. Long-term monitoring and encour-agement can be accomplished in several ways: byregular clinic visits, at group meetings, or viatelephone or E-mail. The longer the weightmaintenance phase can be sustained, the betterthe prospects for long-term success in weightreduction. Drug therapy may also be helpfulduring the weight maintenance phase.RECOMMENDATION: After successfulweight loss, the likelihood of weight lossmaintenance is enhanced by a programconsisting of dietary therapy, physical activ-ity, and behavior therapy, which should becontinued indefinitely. Drug therapy canalso be used. However, drug safety andefficacy beyond 1 year of total treatmenthave not been established. EvidenceCategory B.3. Prevention of Further Weight GainSome patients may not be able to achieve signifi-cant weight reduction. In such patients, animportant goal is to prevent further weight gainthat would exacerbate disease risk. Thus, preven-tion of further weight gain may justify enteringa patient into weight loss therapy. Prevention offurther weight gain can be considered a partialtherapeutic success for many patients. Moreover,if further weight gain can be prevented, thisachievement may be an important first steptoward beginning the weight loss process.Primary care practitioners ought to recognize theimportance of this goal for those patients whoare not able to immediately lose weight. Theneed to prevent weight gain may warrant main-taining patients in a weight management pro-gram for an extended period.STRATEGIES FOR WEIGHT LOSS ANDWEIGHT MAINTENANCE1. Dietary TherapyIn the majority of overweight and obesepatients, adjustment of the diet to reduce caloricintake will be required. Dietary therapy consists,in large part, of instructing patients on how tomodify their diets to achieve a decrease in caloricintake. A key element of the current recommen-dation is the use of a moderate reduction incaloric intake to achieve a slow but progressiveweight loss. Ideally, caloric intake should beH
74Chapter 4: Treatment Guidelinesreduced only to the level required to maintainweight at the desired level. If this level of caloricintake is achieved, excess weight will graduallydisappear. In practice, somewhat greater caloricdeficits are used in the period of active weightloss, but diets with very low calories are to beavoided. Finally the composition of the dietshould be modified to minimize other cardiovas-cular risk factors. 142The centerpiece of dietary therapy for weightloss in overweight patients is a low-calorie diet(LCD) (800 to 1,500 kcal/day). This diet is tobe distinguished from a very low-calorie diet(VLCD) (250 to 800 kcal/day), which has beenunsuccessful in achieving weight loss over thelong term. The LCD recommended in thisreport also contains nutrient compositions thatwill decrease other risk factors, notably, highserum cholesterol and hypertension.Evidence Statement: LCDs can reducetotal body weight by an average of 8 per-cent and help reduce abdominal fat con-tent over a period of approximately 6months. Evidence Category A.Rationale: A decrease in calorie intake is themost important dietary component of weightloss and maintenance. LCDs have been shownto reduce total body weight by an average of 8percent over a period of 6 months, accompa-nied by significant reductions in waist circum-ference. Since this represents an average thatincludes individuals who did not lose weight,an individual average goal of 10 percent is fea-sible. When weight loss occurs, the loss con-sists of about 75 percent fat and 25 percentlean tissue. 556, 557A deficit of 500 to 1,000kcal/day will produce a weight loss of 70 to140 grams/day, or 490 to 980 grams/week (1to 2 lb/week). A deficit of 300 to 500kcal/day will produce a weight loss of 40 to70 grams/day, or 280 to 490 grams/week (1/2to 1 lb/week) A patient may choose a diet of1,000 to 1,200 kcal for women and 1,200 to1,500 kcal for men.VLCDs (less than 800 kcal/day) are not recom-mended for weight loss therapy because thedeficits are too great, and nutritional inadequa-cies will occur unless VLCDs are supplementedwith vitamins and minerals. 558Moreover, clini-cal trials show that LCDs are just as effective asVLCDs in producing weight loss after 1 year. 437Although more weight is initially lost withVLCDs, more is usually regained. Further, rapidweight reduction does not allow for gradualacquisition of changes in eating behavior.Successful behavior therapy is the key to long-term maintenance of weight at a reduced level.Finally, patients using VLCDs are at increasedrisk for developing gallstones.Successful weight reduction by LCDs is morelikely to occur when consideration is given to apatient’s food preferences in tailoring a particulardiet. Care should be taken to be sure that all ofthe recommended dietary allowances are met;this may require use of a dietary supplement.Dietary education is a necessary ingredient inachieving adjustment to an LCD. Educationalefforts should pay particular attention to the fol-lowing topics:s energy value of different foods;s food composition—fats, carbohydrates(including dietary fiber), and proteins;s reading nutrition labels to determine caloriccontent and food composition;s new habits of purchasing—preference to low-calorie foods;s food preparation—avoiding adding high-calo-rie ingredients during cooking (e.g., fats andoils);
75s avoiding overconsumption of high-caloriefoods (both high-fat and high-carbohydratefoods);s maintain adequate water intakes reducing portion sizes; ands limiting alcohol consumption.The Step I Diet in ATP II provides an appropri-ate nutrient composition for an LCD diet. Thecomposition of the diet is presented in Table IV-3. Additional practical dietary advice for patientsis included in Appendix VI.RECOMMENDATION: A diet that is indi-vidually planned to help create a deficit of500 to 1,000 kcal/day should be an inte-gral part of any program aimed at achiev-ing a weight loss of 1 to 2 lb/week.Evidence Category A.RECOMMENDATION: Reducing dietaryfat alone without reducing calories is notsufficient for weight loss. However, reduc-ing dietary fat, along with reducingdietary carbohydrates, can facilitate caloricreduction. Evidence Category A.Evidence Statement: Optimally, dietarytherapy should last at least 6 months.Evidence Category A.Rationale: Many studies suggest that the rate ofweight loss diminishes after about 6 months.Shorter periods of dietary therapy usually resultin lesser weight reductions. Therapeutic effortsshould be directed toward behavior therapy aswell as maintaining LCDs. 486, 570, 571Evidence Statement: During dietary ther-apy, frequent contacts between professionalcounselors and patients promote weight lossand maintenance. Evidence Category C.Rationale: Frequent clinical encounters duringthe initial 6 months of weight reduction appearto facilitate reaching the goals of therapy.During the period of active weight loss, regularvisits of at least once per month and preferablymore often with a health professional for thepurposes of reinforcement, encouragement, andmonitoring will facilitate weight reduction.Weekly group meetings can be conducted at alow cost, and can contribute to favorable behav-ior changes. However, no clinical trials havebeen specifically designed to test the relative effi-cacy of different frequencies of encounters withphysicians, dietitians, or others in the weightloss team. 487Evidence Statement: The amount of timespent with the patient favorably affectsweight change in overweight or obeseadults given dietary therapy. EvidenceCategory D.Rationale: Training of a health professional intechniques of weight reduction, especially inbehavior therapy and dietary principles, isexpected to facilitate weight reduction. Further,adequate time must be made available to thepatient to convey the information necessary, toreinforce behavioral and dietary messages, and tomonitor the patient’s response. Despite thesejudgments, none of the studies reviewed weredesigned to specifically address the type or quali-fications of the health professional who imple-mented the various weight loss approaches.Many of the studies differed in the types of
76Chapter 4: Treatment GuidelinesNutrient Recommended IntakeCalories1Approximately 500 to 1,000 kcal/day reduction from usual intakeTotal Fat230 percent or less of total caloriesSaturated Fatty Acids38 to 10 percent of total caloriesMonounsaturated Fatty Acids Up to 15 percent of total caloriesPolyunsaturated Fatty Acids Up to 10 percent of total caloriesCholesterol3< 300 mg/dayProtein4Approximately 15 percent of total caloriesCarbohydrate555 percent or more of total caloriesSodium Chloride No more than 100 mmol per day (approximately 2.4 g of sodiumor approximately 6 g of sodium chloride)Calcium61,000 to 1,500 mgFiber520 to 30 g1. A reduction in calories of 500 to1,000 kcal/day willhelp achieve a weight loss of 1 to 2 lbs/week. Alcoholprovides unneeded calories and displaces more nutri-tious foods. Alcohol consumption not only increases thenumber of calories in a diet but has been associatedwith obesity in epidemiological studies 559-562as well as inexperimental studies. 563-566The impact of alcohol calo-ries on a person’s overall caloric intake needs to beassessed and appropriately controlled.2. Fat-modified foods may provide a helpful strategy forlowering total fat intake but will only be effective ifthey are also low in calories and if there is no compen-sation of calories from other foods.3. Patients with high blood cholesterol levels may needto use the Step II diet to achieve further reductions inLDL-cholesterol levels; in the Step II diet, saturated fatsare reduced to less than 7 percent of total calories, andcholesterol levels to less than 200 mg/day. All of theother nutrients are the same as in Step I.4. Protein should be derived from plant sources andlean sources of animal protein.5. Complex carbohydrates from different vegetables,fruits, and whole grains are good sources of vitamins,minerals, and fiber. A diet rich in soluble fiber, includ-ing oat bran, legumes, barley, and most fruits and veg-etables may be effective in reducing blood cholesterollevels. A diet high in all types of fiber may also aid inweight management by promoting satiety at lower levelsof calorie and fat intake. Some authorities recommend20 to 30 grams of fiber daily, with an upper limit of 35grams. 546, 567, 5686. During weight loss, attention should be given tomaintaining an adequate intake of vitamins and miner-als. Maintenance of the recommended calcium intakesof 1,000 to 1,500 mg/day is especially important forwomen who may be at risk of osteoporosis. 569LO W-CA L O R I E ST E P I DI E TTABLE IV-3:
77dietary intervention provided. Most programsinvolved dietitians and nutritionists as primarytherapists and used group therapy rather thanindividual sessions. The length of time spentduring each session and the nature of the practi-tioner/patient interaction tended not to be pro-vided.RECOMMENDATION: The literature sug-gests that weight loss and weight mainte-nance therapies that provide a greater fre-quency of contacts between the patient andthe practitioner and are provided over thelong term should be put in place. This canlead to more successful weight loss andweight maintenance. Evidence Category C.2. Physical ActivityAn increase in physical activity is an importantcomponent of weight loss therapy since it leadsto increased expenditure of energy. Increasedphysical activity may also inhibit food intake inoverweight patients. Physical activity can also behelpful in maintaining a desirable weight. Inaddition, sustained physical activity has the ben-efit of reducing overall CHD risk beyond thatproduced by weight reduction alone.Several experts believe that a progressive decreasein the amount of energy expended for work,transportation, and personal chores is a majorcause of obesity in the United States. Theseauthorities note that total caloric intake has notincreased over the last few decades; instead, thecaloric imbalance leading to overweight and obe-sity is the result of a substantial decrease in phys-ical activity and, consequently, a decrease indaily energy expenditure. However, this hypoth-esis is difficult to prove because appropriate dataare lacking. Successful restoration of normalweight in many overweight and obese personsrequires a higher level of energy expenditure.Increased regular physical activity is the way toachieve this goal of augmenting daily energyexpenditure.Evidence Statement: Physical activitycontributes to weight loss, both alone andwhen it is combined with dietary therapy.Evidence Category A.Rationale: Increased physical activity alone cancreate a caloric deficit and can contribute toweight loss. However, efforts to achieve weightloss through physical activity alone generallyproduce an average of a 2 to 3 percent decreasein body weight or BMI. Even so, increased phys-ical activity is a useful adjunct to LCDs in pro-moting weight reduction.Evidence Statement: Physical activity inoverweight and obese adults increases car-diorespiratory fitness independent of weightloss. Evidence Category A.Rationale: There is strong evidence thatincreased physical activity increases cardiorespi-ratory fitness, with or without weight loss. 346, 363,369, 375, 401, 404, 406, 432, 434, 445, 447, 448Improved cardiovas-cular fitness also improves the quality of life inoverweight patients by improving mood, self-esteem, and physical function in daily activities. 572Evidence Statement: Physical activityindependently reduces CVD risk factors(Evidence Category A), and reduces riskfor cardiovascular disease. EvidenceCategory C.
78Chapter 4: Treatment GuidelinesRationale: There is considerable evidence thatphysical inactivity is an independent risk factorfor CVD and diabetes. 572Furthermore, themore active an individual is, the lower the risk.By the same token, increased physical activityappears to independently reduce risk for CVDmorbidity and mortality, and diabetes. 411, 548, 573, 574Physical activity reduces elevated levels of CVDrisk factors, including blood pressure and triglyc-erides, increases HDL-cholesterol, and improvesglucose tolerance with or without weight loss. 572Evidence Statement: Physical activitycontributes to a decrease in body fat,including a modest effect on abdominalfat. Evidence Category B.Rationale: Physical activity appears to have afavorable effect on distribution of body fat. 572Several large cross-sectional studies in Europe, 464Canada, 575and the United States 466-468showedan inverse association between energy expendi-ture through physical activity and several indica-tors of body fat distribution, such as WHR andwaist-to-thigh circumference ratio. Fewer dataare available on the effects of physical activity onwaist circumference, although the ratio changesnoted above suggest a decrease in abdominalobesity. Only a few randomized controlled trialsthat have tested the effect of physical activity onweight loss measured waist circumference. Insome (but not all) studies, physical activity wasfound to produce only modest weight loss anddecreased waist circumference. 365, 369, 375However,it is not known whether the effects of physicalactivity on abdominal fat are independent ofweight loss.2.a. Strategies to increase physical activityMany people live sedentary lives, have littletraining or skills in physical activity, and are dif-ficult to motivate toward increasing their activity.For these reasons, starting a physical activity reg-imen may require supervision for some people.The need to avoid injury during physical activityis high. Extremely obese persons may need tostart with simple exercises that can gradually beintensified. The practitioner must decidewhether exercise testing for cardiopulmonary dis-ease is needed before embarking on a new physi-cal activity regimen. This decision should bebased on a patient’s age, symptoms, and con-comitant risk factors.For most obese patients, physical activity shouldbe initiated slowly, and the intensity should beincreased gradually. Initial activities may bewalking or swimming at a slow pace. With time,depending on progress, the amount of weightlost, and functional capacity, the patient mayengage in more strenuous activities. Some ofthese include fitness walking, cycling, rowing,cross-country skiing, aerobic dancing, and ropejumping. Jogging provides a high-intensity aero-bic exercise, but can lead to orthopedic injury. Ifjogging is desired, the patient’s ability to do thismust first be assessed. The availability of a safeenvironment for the jogger is also a necessity.Competitive sports, such as tennis and volleyball,can provide an enjoyable form of physical activi-ty for many, but again, care must be taken toavoid injury, especially in older people. As theexamples listed in Table IV-4 show, a moderateamount of physical activity can be achieved in avariety of ways. People can select activities thatthey enjoy and that fit into their daily lives.Because amounts of activity are functions ofduration, intensity, and frequency, the sameamounts of activity can be obtained in longersessions of moderately intense activities (such asbrisk walking) as in shorter sessions of morestrenuous activities (such as running).A regimen of daily walking is an attractive formof physical activity for many people, particularlythose who are overweight or obese. The patient
79can start by walking 10 minutes, 3 days a week,and can build to 30 to 45 minutes of moreintense walking at least 5 days a week andpreferably most, if not all, days. 577, 578With thisregimen, an additional 100 to 200 calories perday of physical activity can be expended. Caloricexpenditure will vary depending on the individ-ual’s body weight and intensity of the activity(see Table IV-5).This regimen can be adapted to other forms ofphysical activity, but walking is particularlyattractive because of its safety and accessibility.Washing and waxing a car for 45-60 minutes Less Vigorous,Washing windows or floors for 45-60 minutes More Time **Playing volleyball for 45 minutesPlaying touch football for 30-45 minutesGardening for 30-45 minutesWheeling self in wheelchair for 30-40 minutesWalking 13⁄4 miles in 35 minutes (20 min/mile)Basketball (shooting baskets) for 30 minutesBicycling 5 miles in 30 minutesDancing fast (social) for 30 minutesPushing a stroller 11⁄2 miles in 30 minutesRaking leaves for 30 minutesWalking 2 miles in 30 minutes (15 min/mile)Water aerobics for 30 minutesSwimming laps for 20 minutesWheelchair basketball for 20 minutesBasketball (playing a game) for 15-20 minutesBicycling 4 miles in 15 minutesJumping rope for 15 minutesRunning 11⁄2 miles in 15 minutes (10 min/mile) More Vigorous,Shoveling snow for 15 minutes Less TimeStairwalking for 15 minutesEX A M P L E S O F MO D E R AT E AM O U N T S O F AC T I V I T Y ** A moderate amount of physical activity is roughly equivalent to physical activity that uses approximately 150 calories ofenergy per day, or 1,000 calories per week.** Some activities can be performed at various intensities; the suggested durations correspond to expected intensity of effort.TABLE IV-4:
80Chapter 4: Treatment GuidelinesWith time, a larger weekly volume of physicalactivity can be performed that would normallycause a greater weight loss if it were not com-pensated by a higher caloric intake.Reducing sedentary time is another approach toincreasing activity. Patients should be encour-aged to build physical activities into each day.Examples include leaving public transportationone stop before the usual one, parking furtherthan usual from work or shopping, and walkingup stairs instead of taking elevators or escalators.New forms of physical activity should be sug-gested, e.g., gardening, walking a dog daily, ornew athletic activities. Engaging in physicalactivity can be facilitated by identifying a safearea to perform the activity, e.g., communityparks, gyms, pools, and health clubs. However,when these sites are not available, an area of thehome can be identified and perhaps outfittedwith equipment such as a stationary bicycle or atreadmill.Health professionals should encourage patientsto plan and schedule physical activity 1 week inadvance, budget the time necessary to do it, anddocument their physical activity by keeping adiary and recording the duration and intensityof exercise.RECOMMENDATION: Physical activityshould be an integral part of weight losstherapy and weight maintenance. EvidenceCategory A. Initially, moderate levels ofphysical activity for 30 to 45 minutes, 3 to5 days per week should be encouraged. Alladults should set a long-term goal to accu-Intensity Activity Approximatedurationin minutesModerate Volleyball, noncompetitive 43Moderate Walking, moderate pace (3mph, 20 min/mile) 37Moderate Walking, brisk pace (4mph, 15 min/mile) 32Moderate Table tennis 32Moderate Raking leaves 32Moderate Social dancing 29Moderate Lawn mowing (powered push mower) 29Hard Jogging (5 mph, 12 min/mile) 18Hard Field hockey 16Very Hard Running (6 mph, 10 min/mile) 13DU R AT I O N O F VA R I O U S AC T I V I T I E S T O EX P E N D 150 KI L O C A L O R I E S F O RA N AV E R A G E 70 K G (154 L B ) AD U LTSource: Surgeon General’s Report on Physical Activity and HealthTABLE IV-5:
81mulate at least 30 minutes or more ofmoderate-intensity physical activity onmost, and preferably all, days of the week.Evidence Category B.3. Behavior TherapyBehavioral strategies to reinforce changes in dietand physical activity can produce a weight lossin obese adults in the range of 10 percent ofbaseline weight over 4 months to 1 year. Unlessa patient acquires a new set of eating and physi-cal activity habits, long-term weight reduction isunlikely to succeed. The acquisition of newhabits is particularly important for long-termweight maintenance at a lower weight. Mostpatients return to baseline weights in the absenceof continued intervention. Thus, the physicianor staff members must become familiar withtechniques for modifying life habits of over-weight or obese patients.The goal of behavior therapy is to alter the eat-ing and activity habits of an obese patient.Techniques for behavior therapy have beendeveloped to assist patients in modifying theirlife habits.Evidence Statement: Behavior therapy,in combination with an energy deficit,provides additional benefits in assistingpatients to lose weight short-term (1 year).Its effectiveness for long-term weight main-tenance has not been shown in the absenceof continued behavioral intervention.Evidence Category B.Rationale: The primary assumptions of behaviortherapy are that:s by changing eating and physical activityhabits, it is possible to change body weight;s patterns of eating and physical activity arelearned behaviors and can be modified; ands to change these patterns over the long term,the environment must be changed.Behavior therapies provide methods for over-coming barriers to compliance with dietary ther-apy and/or increased physical activity, and arethus important components of weight loss thera-py. Most weight loss programs incorporatingbehavioral strategies do so as a package thatincludes education about nutrition and physicalactivity. However, this standard “package” ofmanagement should not ignore the need forindividualizing behavioral strategies. 5793.a. Behavior Therapy Strategies Used inWeight Loss and Weight MaintenanceProgramsStudies reviewed for this report examined arange of modalities of behavioral therapy. Nosingle method or combination of behavioralmethods proved to be clearly superior. Thus,various strategies can be used by the practitionerto modify patient behavior. The aim is to changeeating and physical activities behaviors over thelong term. Such change can be achieved eitheron an individual basis or in group settings.Group therapy has the advantage of lower cost.Specific behavioral strategies include the follow-ing:s Self-monitoring of both eating habits andphysical activity—Objectifying one’s ownbehavior through observation and recording isa key step in behavior therapy. Patients shouldbe taught to record the amount and types offood they eat, the caloric values, and nutrientcomposition. Keeping a record of the fre-quency, intensity, and type of physical activitylikewise will add insight to personal behavior.Extending records to time, place, and feelingsrelated to eating and physical activity willhelp to bring previously unrecognized behav-ior to light. 580
82Chapter 4: Treatment Guideliness Stress management—Stress can trigger dys-functional eating patterns, and stress manage-ment can defuse situations leading to overeat-ing. Coping strategies, meditation, and relax-ation techniques all have been successfullyemployed to reduce stress.s Stimulus control—Identifying stimuli thatmay encourage incidental eating enables indi-viduals to limit their exposure to high-risk sit-uations. Examples of stimulus control strate-gies include learning to shop carefully forhealthy foods, keeping high-calorie foods outof the house, limiting the times and places ofeating, and consciously avoiding situations inwhich overeating occurs. 580s Problem solving—This term refers to the self-correction of problem areas related to eatingand physical activity. Approaches to problemsolving include identifying weight-relatedproblems, generating or brainstorming possi-ble solutions and choosing one, planning andimplementing the healthier alternative, andevaluating the outcome of possible changes inbehavior. 580Patients should be encouraged toreevaluate setbacks in behavior and to ask“What did I learn from this attempt?” ratherthan punishing themselves.s Contingency management—Behavior can bechanged by use of rewards for specific actions,such as increasing time spent walking orreducing consumption of specific foods. 44Verbal as well as tangible rewards can be use-ful, particularly for adults. Rewards can comefrom either the professional team or from thepatients themselves. For example, self-rewardscan be monetary or social and should beencouraged.s Cognitive restructuring—Unrealistic goalsand inaccurate beliefs about weight loss andbody image need to be modified to helpchange self-defeating thoughts and feelingsthat undermine weight loss efforts. Rationalresponses designed to replace negativethoughts are encouraged. 580For example, thethought, “I blew my diet this morning by eat-ing that doughnut; I may as well eat what Ilike for the rest of the day,” could be replacedby a more adaptive thought, such as, “Well, Iate the doughnut this morning, but I can stilleat in a healthy manner at lunch and dinner.”s Social support—A strong system of socialsupport can facilitate weight reduction.Family members, friends, or colleagues canassist an individual in maintaining motivationand providing positive reinforcement. Somepatients may benefit by entering a weightreduction support group. Overweight patientsshould be asked about (possibly) overweightchildren and family weight control strategies.Parents and children should work together toengage in and maintain healthy dietary andphysical activity habits.3.b. Treatment of Obese Individuals with BingeEating DisorderIf a patient suffers from binge eating disorder(BED), consideration can be given to referringthe patient to a health professional who special-izes in BED treatment. Behavioral approaches toBED associated with obesity have been derivedfrom cognitive behavior therapy (CBT) used totreat bulimia nervosa. 227Among the techniquesare self-monitoring of eating patterns, encourag-ing regular patterns of eating (three meals a dayplus planned snacks), cognitive restructuring,and relapse prevention strategies. 581RECOMMENDATION: Behavior therapystrategies to promote diet and physicalactivity should be used routinely, as theyare helpful in achieving weight loss andweight maintenance. Evidence Category B.
834. Combined TherapyTo achieve the greatest likelihood of successfrom weight loss therapy, the combination ofdietary therapy with an LCD, increased physicalactivity, and behavior therapy will be required.Inclusion of behavior therapy and increasedphysical activity in a weight loss regimen willprovide the best opportunity for weight loss, andhopefully for long-term weight control. In orderto achieve weight loss, such a regimen should bemaintained for at least 6 months before consid-ering pharmacotherapy.Evidence Statement: Combined inter-vention of an LCD, increased physicalactivity, and behavior therapy provides themost successful therapy for weight loss andweight maintenance. Evidence Category A.Rationale: Clinical trials have demonstrated thatcombining behavior therapy, LCDs, andincreased physical activity provides better out-comes for long-term weight reduction than pro-grams that use only one or two of these modali-ties. A lower-fat diet markedly improves thepotential of physical activity to achieve a nega-tive energy balance. 317, 369, 434, 444, 582In addition,lower-fat diets that are also low in saturated fatsreduce serum cholesterol levels, which wouldreduce CVD risk. It is difficult to achieve a neg-ative energy balance and weight loss with physi-cal activity of moderate duration and intensityin individuals who consume a high-fat diet andalcohol. 583, 584RECOMMENDATION: Weight loss andweight maintenance therapy should employthe combination of LCDs, increased physi-cal activity, and behavior therapy. EvidenceCategory A.5. PharmacotherapyDrug therapy has undergone radical changes inthe last 2 years. With the publication of the tri-als with phentermine and fenfluramine byWeintraub in 1992 (210 weeks), drug therapybegan to change from short-term to long-termuse. Both dexfenfluramine and fenfluraminealone, as well as the combination of phenter-mine/fenfluramine, were used long term.However, concerns about recently reportedunacceptable side effects, such as valvular lesionsof the heart causing significant insufficiency ofthe valves, 658have led to the withdrawal of thedrugs dexfenfluramine and fenfluramine fromthe market in September 1997. 761No drugsremained that were approved by the Food andDrug Administration (FDA) for use longer than3 months. In November 1997, the FDAapproved a new drug, sibutramine, for use inobesity and is in the process of evaluating orlis-tat for long-term use.Evidence Statement: Appropriate weightloss drugs can augment diet, physical activ-ity, and behavior therapy in weight loss.Evidence Category B.Rationale: The purpose of weight loss andweight maintenance is to reduce health risks. Ifweight is regained, health risks increase oncemore. The majority of persons who lose weightregain it, 585so the challenge to the patient andthe practitioner is to maintain the weight loss.Because of the tendency to regain weight afterweight loss, the use of long-term medication toaid in the treatment of obesity may be indicatedin carefully selected patients.The drugs used to promote weight loss havebeen anorexiants or appetite suppressants. Threeclasses of anorexiant drugs have been developed,all of which affect neurotransmitters in the
84Chapter 4: Treatment Guidelinesbrain: those that affect catecholamines, thosethat affect serotonin, and those that affect both.They work by increasing the secretion ofdopamine, norepinephrine, or serotonin into thesynaptic neural cleft, or by inhibiting the reup-take of these neurotransmitters back into theneuron, or by both mechanisms. The new agentsibutramine has norepinephrine and serotonineffects. Another new agent, orlistat, has a differ-ent mechanism of action, the blockage of fatabsorption. Very few trials longer than 6 monthshave been done with any drug. The ones testedfor at least 1 year that received FDA approval forlong-term use are shown in Table IV-6.These drugs are effective but modest in theirability to produce weight loss. Net weight lossattributable to drugs generally has been reportedto be in the range of 2 to 10 kg (4.4 to 22 lb),although some patients lose significantly moreweight. It is not possible to predict how muchweight an individual may lose. Most of theweight loss usually occurs in the first 6 monthsof therapy.Adverse effects include primary pulmonaryhypertension with fenfluramine and dexfenflu-ramine, 395, 586, 587valvular heart disease withdexfenfluramine and fenfluramine, 658andincreases in blood pressure and pulse with sibu-tramine. 510With orlistat, there is a possibledecrease in the absorption of fat-soluble vita-mins; overcoming this may require vitamin sup-plementation. People with a history of highblood pressure, CHD, congestive heart failure,arrhythmias, or history of stroke should not takesibutramine, and all patients taking the medica-tion should have their blood pressure monitoredon a regular basis. Depression has been describedwith the serotonergic drugs but it is generallynot clinically significant. Neurotoxic effects withneuronal atrophy have been described with highdoses of dexfenfluramine in rats and primates,but not in humans. 588The risks for usingDrug Action Adverse Effectsdexfenfluramine* serotonin reuptake inhibitor valvular heart diseasefenfluramine* serotonin releaser primary pulmonary hypertensionneurotoxicitysibutramine norepinephrine, dopamine, increase in heart rate and blood pressureand serotonin reuptake inhibitororlistat± inhibits pancreatic lipase, decrease in absorption of fat-solubledecreases fat absorption vitaminssoft stools and anal leakagepossible link to breast cancerWE I G H T LO S S DR U G S ++ Ephedrine and caffeine, and, fluoxetine have also been tested for weight loss, but are not approved for use in the treatment ofobesity. Mazindol, phentermine, benzphetamine, and phendimetrazine are approved for only short-term use for the treat-ment of obesity.* FDA approval withdrawn± FDA approval pendingTABLE IV-6:
85appetite suppressant drugs during pregnancy isunknown.Evidence Statement: Weight loss drugsapproved by the FDA for long-term usemay be useful as an adjunct to diet andphysical activity for patients with a BMIof ≥ 30 with no concomitant obesity-relat-ed risk factors or diseases, and for patientswith a BMI of ≥ 27 with concomitantobesity-related risk factors or diseases.Evidence Category B.Rationale: If after at least 6 months on a weightloss regimen of an LCD, increased physicalactivity, and behavior therapy, the patient hasnot lost the recommended 1 lb/week, carefulconsideration may be given to pharmacotherapy.There are few long-term studies evaluating thesafety or effectiveness of many currentlyapproved weight loss medications. At present,sibutramine is available for long-term use.(Note: FDA approval for orlistat is pending aresolution of labeling issues and results of PhaseIII trials.) Their risk/benefit ratio is such thatthey can be recommended for use with thedegree of obesity outlined above. 589Weight lossmedications should be used only by patientswho are at increased medical risk because oftheir weight and should not be used for “cos-metic” weight loss. The risk factors and diseasesconsidered serious enough to warrant pharma-cotherapy at BMI of 27 to 29.9 are hyperten-sion, dyslipidemia, CHD, type 2 diabetes, andsleep apnea.Not every patient responds to drug therapy.Tests with weight loss drugs have shown thatinitial responders tend to continue to respond,while initial nonresponders are less likely torespond even with an increase in dosage. 395, 507Ifa patient does not lose 2 kg (4.4 lb) in the first 4weeks after initiating therapy, the likelihood oflong-term response is very low. 507This findingmay be used as a guide to treatment, either con-tinuing medication in the responders or stop-ping it in the nonresponders. If weight is lost inthe initial 6 months of therapy or is maintainedafter the initial weight loss phase, this should beconsidered a success, and the drug may be con-tinued. It is important to remember that themajor role of medications should be to helppatients stay on a diet and physical activity planwhile losing weight. Medication cannot beexpected to continue to be effective in weightloss or weight maintenance once it has beenstopped. 590, 591The use of the drug may be con-tinued as long as it is effective and the adverseeffects are manageable and not serious. There areno indications for specifying how long a weightloss drug should be continued. Therefore, aninitial trial period of several weeks with a givendrug or combination of drugs may help deter-mine their efficacy in a given patient. If a patientdoes not respond to a drug with reasonableweight loss, the physician should reassess thepatient to determine adherence to the medica-tion regimen and adjunctive therapies, or con-sider the need for dosage adjustment. If thepatient continues to be unresponsive to themedication, or serious adverse effects occur, thephysician should consider its discontinuation. 592Evidence Statement: Adverse side effectsfrom the use of weight loss drugs have beenobserved in patients. Evidence Category A.Rationale: The potential for side effects fromthe use of weight loss drugs is of great concern.Adverse effects with sibutramine includeincreased blood pressure and tachycardia. 590With orlistat there are oily and loose stools andsome fat-soluble vitamin malabsorption. 389Thus, a multivitamin supplement is recom-
86Chapter 4: Treatment Guidelinesmended. Side effects are generally mild and mayimprove with continued use, although theirpersistence may result in discontinuation ofthe drug.There is a great interest in weight loss drugsamong consumers. Because of the possibility ofserious adverse effects, it is incumbent on thepractitioner to use drug therapy with caution.Practitioners should be sure that patients are nottaking drugs that have been withdrawn from themarket for safety reasons, and herbal medica-tions are not recommended as part of a weightloss program. These preparations have unpre-dictable amounts of active ingredients andunpredictable and potentially harmful effects.In those patients with a lower level of obesity risk,nonpharmacological therapy is the treatment ofchoice. It is important for the physician to moni-tor both the effectiveness and the side effects ofthe drug.Evidence Statement: Using weight loss drugssingly (not in combination) and starting withthe lowest effective doses can decrease the likeli-hood of adverse effects. Evidence Category C.Rationale: Given the fact that adverse eventsmay increase in association with combinationdrug therapy, it seems wise that, until furthersafety data are available, using weight loss drugssingly would be more prudent. Some patientswill respond to lower doses, so that full dosage isnot always necessary. The short-term use ofdrugs (< 3 months) has not generally beenfound to be effective.Drugs should be used only as part of a compre-hensive program that includes behavior therapy,diet, and physical activity. Appropriate monitor-ing for side effects must be continued whiledrugs are part of the regimen. Patients will needto return for followup in 2 to 4 weeks, thenmonthly for 3 months, and then every 3 monthsfor the first year after starting the medication.After the first year, the doctor will advise thepatient on appropriate return visits. The purpos-es of these visits are to monitor weight, bloodpressure, and pulse, discuss side effects, conductlaboratory tests, and answer questions.Since obesity is a chronic disorder, the short-term use of drugs is not helpful. The health pro-fessional should include drugs only in the con-text of a long-term treatment strategy. 593Therisk/benefit ratio cannot be predicted at thistime, since not enough long-term data (> 1 year)are available on any of the available drugs.RECOMMENDATION: Weight loss drugsapproved by the FDA may only be used aspart of a comprehensive weight loss pro-gram, including dietary therapy and physi-cal activity, for patients with a BMIof ≥ 30 with no concomitant obesity-relat-ed risk factors or diseases, and for patientswith a BMI of ≥ 27 with concomitantobesity-related risk factors or diseases.Weight loss drugs should never be usedwithout concomitant lifestyle modifica-tions. Continual assessment of drug therapyfor efficacy and safety is necessary. If thedrug is efficacious in helping the patientlose and/or maintain weight loss and thereare no serious adverse effects, it can be con-tinued. If not, it should be discontinued.Evidence Category B.6. Surgery for Weight LossSurgery is one option for weight reduction forsome patients with severe and resistant obesity.The aim of surgery is to modify the gastroin-testinal tract to reduce net food intake. Mostauthorities agree that weight loss surgery shouldbe reserved for patients with severe obesity, inwhom efforts at other therapy have failed, and
87who are suffering from the complications ofobesity.Considerable progress has been made in devel-oping safer and more effective surgical proce-dures for promoting weight loss. Surgical inter-ventions commonly used include gastroplasty,gastric partitioning, and gastric bypass. Theseprocedures are designed primarily to reduce foodconsumption. They have replaced previous pro-cedures that were designed to promote malab-sorption of nutrients. The latter procedures werefraught with side effects that made their useimpractical or dangerous.Evidence Statement: Gastrointestinalsurgery (gastric restriction [vertical gastricbanding] or gastric bypass [Roux-en Y])can result in substantial weight loss, andtherefore is an available weight loss optionfor well-informed and motivated patientswith a BMI ≥ 40 or ≥ 35, who havecomorbid conditions and acceptable opera-tive risks. Evidence Category B.Rationale: According to the National Institutesof Health’s Consensus Development Conferenceon Gastrointestinal Surgery for Severe Obesity, 523the risk for morbidity and mortality accompany-ing obesity increases with the degree of over-weight. Thus, treatment of clinically severe obe-sity involves an effort to create a caloric deficitsufficient to result in weight loss and reductionof weight-associated risk factors or comorbidi-ties. Surgical approaches can result in substantialweight loss, i.e., from 50 kg (110 lb) to as muchas 100 kg (220 lb) over a period of 6 months to1 year. A major limitation of nonsurgicalapproaches is the failure to maintain reducedbody weight in many individuals.Surgical procedures in current use (gastricrestriction [vertical gastric banding] and gastricbypass [Roux-en Y]) can induce substantialweight loss, and serve to reduce weight-associat-ed risk factors and comorbidities. Compared toother interventions available, surgery has pro-duced the longest period of sustained weightloss. Assessing both perioperative risk and long-term complications is important and requiresassessing the risk/benefit ratio in each case.Patients whose BMI equals or exceeds 40 kg/m2are potential candidates for surgery if theystrongly desire substantial weight loss, becauseobesity severely impairs the quality of their lives.Less severe obese patients (BMIs between 35 and39.9 kg/m2) also may be considered for surgery.This group primarily includes those patientswith high-risk comorbid conditions (cardiovas-cular, sleep apnea, uncontrolled type 2 diabetes)or weight-induced physical problems interferingwith performance of daily life activities.A recent retrospective study of severely over-weight patients with noninsulin-dependent dia-betes, who were referred for consideration of agastric bypass procedure, allowed for a compari-son of those who opted for the surgical proce-dure versus those who did not undergo the pro-cedure because of personal preference or refusalof insurance payment. Patients undergoing thesurgical procedure had a decrease in mortalityrate for each year of follow up. 594This latterobservation provides initial documentation ofthe significant impact that reduction in weightmay have on mortality.Most of the surgery studies primarily includedwomen of childbearing age. Caution should beexercised in selecting candidates for surgery totreat obesity, as pregnancy demands increasenutritional needs and a normal need for weightgain. Women with reproductive potential shouldbe advised to avoid pregnancy until their weighthas stabilized postoperatively and potentialmicronutrient deficiencies have been identifiedand treated.
88Chapter 4: Treatment GuidelinesOf special note is that many of the studiesreported to date have not had population sam-ples representative of the general severely over-weight population with respect to race, ethnicity,cultural or socioeconomic background, or gen-der.Evidence Statement: Patients opting forsurgical intervention should be followed by amultidisciplinary team (medical, behavioral,and nutritional). Evidence Category D.Rationale: As for all other interventions for obe-sity, an integrated program should be in placethat will provide guidance concerning the neces-sary dietary regimen, appropriate physical activi-ty, and behavioral and social support both priorto and after the surgical procedure.Evidence Statement: Lifelong medicalsurveillance after surgical therapy is anecessity. Evidence Category C.Rationale: Since surgical procedures result insome loss of absorptive function, the long-termconsequences of potential nutrient deficienciesmust be recognized and adequate monitoringmust be performed, particularly with regard tovitamin B12, folate, and iron. Some patients maydevelop other gastrointestinal symptoms such as“dumping syndrome” or gallstones. Occasionally,patients may have postoperative mood changesor their presurgical depression symptoms maynot be improved by the achieved weight loss.Thus, surveillance should include monitoring ofindices of inadequate nutrition and modificationof any preoperative disorders. Table IV-7 illus-trates some of the complications that can occurfollowing gastric bypass surgery.Vitamin B12 deficiency 239 39.9 percentReadmit for various reasons 229 38.2 percentIncisional hernia 143 23.9 percentDepression 142 23.7 percentStaple line failure 90 15.0 percentGastritis 79 13.2 percentCholecystitis 68 11.4 percentAnastomotic problems 59 9.8 percentDehydration, malnutrition 35 5.8 percentDilated pouch 19 3.2 percentGA S T R I C BY PA S S SU R G E RY CO M P L I C AT I O N S : 14-YE A R FO L L O W UPData derived from source (Pories WJ595) and modified based on personal communication.TABLE IV-7:
89RECOMMENDATION: Weight loss surgeryis an option for carefully selected patientswith clinically severe obesity (BMI ≥ 40or ≥ 35 with comorbid conditions) whenless invasive methods of weight loss havefailed and the patient is at high risk forobesity-associated morbidity or mortality.Evidence Category B.COMMENTARY: Adapting Weight Loss ProgramsTo Meet the Needs of Diverse PatientPopulations.Standard obesity treatment approaches should betailored to the needs of patients. It is, however, dif-ficult to determine from the literature how oftenthis occurs and whether it makes weight loss pro-grams more effective. Very few published reports ofsuch adapted programs can be identified, particu-larly when a distinction is made from reports thatinclude or focus on special populations but do notreport any particular steps taken to modify the inter-vention for these populations. In addition, it isimpossible to compare directly the amount ofweight lost using specially adapted programs withthat achieved when more standard approaches areused. Studies reporting these programs are sometimespilot studies or descriptive reports that do not meetthe standards of evidence set forth elsewhere in theseguidelines. Where randomized controlled trials orquasi-experiments are available, they usually donot include an internal comparison with a programinvolving no adaptations. Appendix III illustratesthe types of adaptations that have been reported.Large individual variation exists within any socialor cultural group; furthermore, there is substantialoverlap among subcultures within the larger society.There is, therefore, no “cookbook” or standardizedset of rules to optimize weight reduction with agiven type of patient. A theoretical and qualitativeanalysis of cultural appropriateness in obesity treat-ment programs has been conducted, and it providessome guidance for incorporating patient character-istics and perspectives when designing and deliver-ing weight loss programs. 257Some examples follow:s Adapt the setting and staffing to the patientpopulation. The setting should:s be physically accessible to the patient;s have features likely to be familiar to thepatient;s be free of negative psychosocial connotations;s be devoid of aspects that create a largesocial distance among patients or betweenpatient and practitioner; ands promote active patient participation andhigh patient self-esteem and self-efficacy.The staff should be culturally self-awareand culturally competent in working withpersons of diverse cultural backgroundsand income or educational levels. Forexample, cultural adaptation of programshas been approached with the assumptionthat community centers may be preferableto hospitals or medical offices as venues forconducting lifestyle weight reduction pro-grams. Or, some programs have includedpeer educators as a possible way of helpingto overcome background and social classdifferences by providing a bridge of knowl-edge, experience, and perspective betweenpatients and practitioners (see Appendix III).s List assumptions about the type of patientfor whom the program will be best suitedand evaluate the extent to which theseassumptions are appropriate for prospectivepatients. Where appropriate, modify theprogram to avoid the need for certainassumptions. Consider patients’—s preexisting knowledge base;s day-to-day routine;s discretionary time;
90Chapter 4: Treatment Guideliness financial resources and living situation;ands cultural preferences for types of food andactivity.For example, redesign printed materialsto be suitable for patients with low lit-eracy skills or poor vision. Offer dietaryand physical activity recommendationsthat will be feasible for low-incomepatients living in inner-city areas withlimited access to supermarkets or withhigh crime rates.s Consider how the obesity treatment programfits in other aspects of the health care and self-care of the patient(s), and integrate otheraspects where appropriate. For example, forthose patients with diabetes, information aboutweight reduction should be aligned with otherdiabetes management advice.s Expect and allow for program modificationsbased on patient feedback and preferences.Program appropriateness can be increasedwhen patients can express their needs and pref-erences, and the program is then adapted tothose needs and preferences. This is especiallyapplicable when practitioners have limitedcommon experience with patients.In recent years, a fat acceptance, nondieting advo-cacy group has developed. This has emerged fromconcerns about weight cycling and its possibleadverse effects on morbidity and mortality.However, recent evidence suggests that intentionalweight loss is not associated with increased morbid-ity and mortality. For this reason, the guidelineshave been made explicit on the importance ofintervention for weight loss and maintenance inthe appropriate patient groups.COMMENTARY: Weight Reduction After Age65—What Are the Issues?s Are the indications for treating obesity in olderadults the same as those for younger adults?s Does weight loss reduce risk factors in olderadults?s Does weight reduction prolong the lives of olderadults?s Are there risks associated with obesity treatmentthat are unique to older adults?The higher prevalence of cardiovascular risk factorsin overweight versus nonoverweight persons isclearly observed at older ages. 596, 597In addition,obesity is a major predictor of functional limita-tions and mobility impairments in older adults. 596,598-600Both observational data 596and applicablerandomized controlled trials cited in these guide-lines suggest that weight loss reduces risk factorsand improves functional status in older persons inthe same manner as in younger adults.The question of whether weight reduction leads toincreased survival among older adults has beenraised because of observations that weight reductionafter age 60 or 65 years may be unable to reversethe deleterious effects of longstanding obesity.288, 596,601, 602Also, weight loss at older ages has been associ-ated with increased mortality. 318, 603-607Some of theassociation of weight loss with higher mortalitymay be due to the increased frequency at older agesof involuntary weight loss due to identified oroccult illness. In any case, the association of weightloss with higher mortality applies most clearly toindividuals who enter old age with a BMI in thelower part of the range. It is not clear that itapplies to overweight older persons with CVD riskfactors. This issue has been difficult to clarify inobservational data, but there are no randomizedtrials in which the effects of obesity treatment onmortality can be directly assessed, at any age.The wisdom or importance of treating obesity atolder ages has also been questioned because of epi-demiological observations suggesting a decreasedsignificance of obesity-related relative risks at olderages (see chapter 2.3.b.). However, these relative-risk data are not a useful reference point for con-sidering whether obese older persons will benefit.
91Relative risks are influenced by the characteristicsof the comparison group. In this case, the compari-son group is lower-weight older persons who havealso had high morbidity and mortality.Concerns about potential adverse effects of obesitytreatment in older adults have been raised withrespect to bone health and to dietary adequacy.Weight reduction may accelerate aging-related boneloss and thereby increase the risk of osteoporoticfractures in high-risk groups such as older whitewomen. 608-610This concern is more relevant toweight loss in thin persons than in obese persons.Some evidence suggests that including resistancetraining and moderate weight-bearing exercise as apart of a weight reduction program may helpmaintain bone integrity. 611, 612The general nutritional safety of weight reductionat older ages is of interest because restrictions onoverall food intake due to dieting could result ininadequate intakes of protein or essential vitaminsand minerals. In addition, involuntary weight lossindicative of occult disease might be mistaken forsuccess in voluntary weight reduction. These con-cerns can be alleviated by providing proper nutri-tional counseling and regular body weight moni-toring for older persons for whom weight reductionis prescribed.Evidence Statement: Age alone shouldnot preclude treatment for obesity in adultmen and women. Evidence Category D.Rationale: There is little evidence at present toindicate that obesity treatment should be with-held from adult men and women on the basis ofage alone up to 80 years of age.RECOMMENDATION: A clinical decisionto forego obesity treatment in an olderadult should be guided by an evaluation ofthe potential benefits of weight reductionfor day-to-day functioning and reductionof the risk of future cardiovascular events,as well as the patient’s motivation forweight reduction. Care must be taken toensure that any weight reduction programminimizes the likelihood of adverse effectson bone health or other aspects of nutri-tional status. Evidence Category D.SMOKING CESSATION IN THE OVERWEIGHTOR OBESE PATIENTCigarette smoking is a major risk factor for car-diopulmonary disease. Because of its attendanthigh risk, smoking cessation is a major goal ofrisk-factor management. This aim is especiallyimportant in the overweight or obese patient,who usually carries excess risk from obesity-asso-ciated risk factors. Thus, smoking cessation insuch patients becomes a high priority for riskreduction.Evidence Statement: Smoking and obesitytogether increase cardiovascular risk, butfear of weight gain upon smoking cessationis an obstacle for many patients whosmoke. Evidence Category C.Rationale: Both smoking and obesity are accom-panied by increased risks for cardiovascular dis-ease. Many well-documented health benefits areassociated with smoking cessation, but a majorobstacle to successful smoking cessation has beenthe attendant weight gain observed in about 80percent of quitters. This weight gain averages 4.5to 7 lb, but in 13 percent of women and 10 per-cent of men, weight gains in excess of 28 lb havebeen noted among quitters. 613-615Weight gain isan important barrier to smoking cessation, par-ticularly in women. 616, 617I
92Chapter 4: Treatment GuidelinesWeight gain that accompanies smoking cessationso far has been relatively resistant to mostdietary, behavioral, or physical activity interven-tions. 618, 619Postcessation weight gain has beenassociated with a reduction in energy expendi-ture of up to 100 kcal/day, accounting forapproximately one-third of the weight gain aftersmoking cessation. 620, 621The reduction in ener-gy expenditure appears to be the result of adecrease in the resting metabolic rate. 614In gen-eral, no differences in the level of physical activi-ty have been observed after smoking cessation.About two-thirds of the weight gain after smok-ing cessation appears to result from increasedcaloric intake. 621-623However, dietary counselingprograms combined with smoking cessation pro-grams have not been very successful. There areseveral products that reduce postcessation weightgain during drug administration, including nico-tine replacement therapy, 624phenylpropanol-amine, 625and bupropion. 626No matter what thedrug, however, it appears that these drugs merelydelay rather than prevent postcessation weightgain. That is, while providing weight gain sup-pression during drug administration, subjects onthese drugs experience a rebound of weight gainonce they go off the products and long-termweight gain is equal to that in those not receiv-ing these drugs. 624, 626The weight gained with smoking cessation is lesslikely to produce negative health consequencesthan would continued smoking. For this reason,smoking cessation should be strongly reinforcedin persons regardless of their baseline weight.Smoking is not an acceptable weight controltherapy, although it seems to be used for thispurpose by a great many people. Overweightpatients, as well as all others, should be coun-seled to quit smoking. For practical reasons, itmay be prudent to avoid initiating smoking ces-sation and weight loss therapy simutaneously.Prevention of weight gain through diet andphysical activity should be stressed. If weightgain ensues after smoking cessation, it should bemanaged vigorously according to the guidelinesoutlined in this report. Although short-termweight gain is a common side effect of smokingcessation, this gain does not rule out the possi-bility of long-term weight control. There are noclinical trials to test whether ex-smokers are lesslikely to successfully achieve long-term weightreduction than those who never smoked.RECOMMENDATION: All smokers, regard-less of their weight status, should quitsmoking. Evidence Category A. Preventionof weight gain should be encouraged and ifweight gain does occur, it should be treatedthrough dietary therapy, physical activity,and behavior therapy, maintaining theprimary emphasis on the abstinence fromsmoking. Evidence Category C.ROLE OF HEALTH PROFESSIONALS INWEIGHT LOSS THERAPYEvidence Statement: A number of healthprofessionals can play an important role ina weight loss and management program.Evidence Category B.Rationale: Various randomized controlled trialswere reviewed that highlighted the role of thenutritionists, 413, 476exercise physiologists andphysical education instructors, 401, 447nurses, 413and psychologists 476, 487in addition to the physi-cian in providing assessment, treatment, and fol-low up during weight loss. As a result, it is sug-gested that the health professionals avail them-selves of the various disciplines that offer exper-tise in dietary counseling, physical activity, andbehavior change. The relationship between thephysician and these disciplines can be a direct,J
93formal one or a more indirect referral. It isimportant to emphasize that a positive attitudeof support and encouragement from all profes-sionals is crucial to continuing success. 627RECOMMENDATION: A weight loss andmaintenance program can be conducted bya practitioner without specialization inweight loss so long as that person has therequisite interest and knowledge. However,various health professionals with differentexpertise are available and helpful to apractitioner who would like assistance.Evidence Category B.
955ADVANTAGES OF WEIGHT LOSSs Weight loss is recommended to lower elevatedblood pressure in overweight and obese per-sons with high blood pressure. EvidenceCategory A.s Weight loss is recommended to lower elevatedlevels of total cholesterol, low-density lipopro-tein cholesterol, and triglycerides and to raiselow levels of high-density lipoprotein choles-terol in overweight and obese persons withdyslipidemia. Evidence Category A.s Weight loss is recommended to lower elevatedblood glucose levels in overweight and obesepersons with type 2 diabetes. EvidenceCategory A.MEASUREMENT OF DEGREE OFOVERWEIGHT AND OBESITYs Practitioners should use the body mass index(BMI) to assess overweight and obesity. Bodyweight alone can be used to follow weightloss, and to determine efficacy of therapy.Evidence Category C.s The BMI should be used to classify over-weight and obesity and to estimate relativerisk for disease compared to normal weight.Evidence Category C.s The waist circumference should be used toassess abdominal fat content. EvidenceCategory C.s For adult patients with a BMI of 25 to 34.9kg/m2, sex-specific waist circumference cutoffsshould be used in conjunction with BMI toidentify increased disease risks. EvidenceCategory C.GOALS FOR WEIGHT LOSSs The initial goal of weight loss therapy shouldbe to reduce body weight by approximately10 percent from baseline. With success, fur-ther weight loss can be attempted, if indicat-ed, through further assessment. EvidenceCategory A.s Weight loss should be about 1 to 2 lb/weekfor a period of 6 months, with the subsequentstrategy based on the amount of weight lost.Evidence Category B.HOW TO ACHIEVE WEIGHT LOSS1. Dietary Therapys Low-calorie diets are recommended forweight loss in overweight and obese persons.Evidence Category A. Reducing fat as part ofa low-calorie diet is a practical way to reducecalories. Evidence Category A.s Reducing dietary fat alone without reducingcalories is not sufficient for weight loss.However, reducing dietary fat, along withreducing dietary carbohydrates, can facilitatecaloric reduction. Evidence Category A.s A diet that is individually planned to helpcreate a deficit of 500 to 1,000 kcal/dayshould be an integral part of any programaimed at achieving a weight loss of 1 to 2lb/week. Evidence Category A.2. Physical Activitys Physical activity is recommended as part of acomprehensive weight loss therapy and weightmaintenance program because it: (1) modestlycontributes to weight loss in overweight andSUMMARY OF RECOMMENDATIONSABCD
96Chapter 5: Summary of Recommendationsobese adults (Evidence Category A), (2) maydecrease abdominal fat (Evidence CategoryB), (3) increases cardiorespiratory fitness(Evidence Category A), and (4) may helpwith maintenance of weight loss (EvidenceCategory C).s Physical activity should be an integral part ofweight loss therapy and weight maintenance.Evidence Category A. Initially, moderate lev-els of physical activity for 30 to 45 minutes, 3to 5 days per week should be encouraged. Alladults should set a long-term goal to accumu-late at least 30 minutes or more of moderate-intensity physical activity on most, andpreferably all, days of the week. EvidenceCategory B.s The combination of a reduced calorie dietand increased physical activity is recommend-ed since it produces weight loss, decreasesabdominal fat, and increases cardiorespiratoryfitness. Evidence Category A.3. Behavior Therapys Behavior therapy is a useful adjunct whenincorporated into treatment for weight lossand weight maintenance. Evidence CategoryB.s Practitioners need to assess the patient’s moti-vation to enter weight loss therapy; assess thereadiness of the patient to implement the planand then take appropriate steps to motivatethe patient for treatment. Evidence CategoryD.s Behavior therapy strategies to promote dietand physical activity should be used routinely,as they are helpful in achieving weight lossand weight maintenance. Evidence CategoryB.4. Combined Therapys Weight loss and weight maintenance therapyshould employ the combination of low-caloriediets, increased physical activity, and behaviortherapy. Evidence Category A.5. Pharmacotherapys Weight loss drugs approved by the FDA maybe used as part of a comprehensive weightloss program including diet and physicalactivity for patients with a BMI of ≥ 30 withno concomitant obesity- related risk factors ordiseases, and for patients with a BMI of ≥ 27with concomitant obesity-related risk factorsor diseases. Drugs should never be used with-out concomitant lifestyle modification.Continual assessment of drug therapy for effi-cacy and safety is necessary. If the drug is effi-cacious in helping the patient lose and/ormaintain weight loss and there are no seriousadverse effects, it can be continued. If not, itshould be discontinued. Evidence Category B.6. Weight Loss Surgerys Weight loss surgery is an option in carefullyselected patients with clinically severe obesity(BMI ≥ 40 or ≥ 35 with comorbid condi-tions) when less invasive methods of weightloss have failed and the patient is at high riskfor obesity-associated morbidity or mortality.Evidence Category B.GOALS FOR WEIGHT LOSS MAINTENANCEs After successful weight loss, the likelihood ofweight loss maintenance is enhanced by aprogram consisting of dietary therapy, physi-cal activity, and behavior therapy, whichshould be continued indefinitely. Drug thera-py can also be used. However, drug safety andefficacy beyond 1 year of total treatment havenot been established. Evidence Category B.s A weight maintenance program should be apriority after the initial 6 months of weightloss therapy. Evidence Category B.E
97HOW TO MAINTAIN WEIGHT LOSSs The literature suggests that weight loss andweight maintenance therapies that provide agreater frequency of contacts between thepatient and the practitioner and are providedover the long term should be put in place.This can lead to more successful weight lossand weight maintenance. Evidence CategoryC.s A weight loss and maintenance program canbe conducted by a practitioner without spe-cialization in weight loss so long as that per-son has the requisite interest and knowledge.However, various health professionals withdifferent expertise are available and helpful toa practitioner who would like assistance.Evidence Category B.SPECIAL TREATMENT GROUPSs All smokers, regardless of their weight status,should quit smoking. Evidence Category A.Prevention of weight gain should be encour-aged and if weight gain does occur, it shouldbe treated through diet, physical activity, andbehavior therapy, maintaining the primaryemphasis on the abstinence from smoking.Evidence Category C.s A clinical decision to forego obesity treatmentin older adults should be guided by an evalua-tion of the potential benefits of weight reduc-tion for day-to-day functioning and reductionof the risk of future cardiovascular events, aswell as the patient’s motivation for weightreduction. Care must be taken to ensure thatany weight reduction program minimizes thelikelihood of adverse effects on bone health orother aspects of nutritional status. EvidenceCategory D.s The possibility that a standard approach toweight loss will work differently in diversepatient populations must be considered whensetting expectations about treatment out-comes. Evidence Category B.FG
98Obesity is a heterogeneous chronic disorder thathas many causes, although the fundamental basisis an imbalance between energy intake and ener-gy expenditure. Future research needs to exam-ine the most effective ways to treat and preventobesity, the causes of obesity and their mecha-nisms, the influence of fat distribution on healthrisk, and the development of better methods forassessing energy intake and energy expenditure.INTERVENTION APPROACHESConsiderable research is needed on interventionapproaches to treat and prevent obesity.Increased research on behavioral theory specifi-cally addressing obesity treatment and preven-tion for all individuals, including children andadolescents, needs to be conducted. Interventionmethods to prevent weight gain with smokingcessation are of particularly high priority inhelping achieve smoking cessation.More research is needed on behavioral interven-tion methods conducted in various settings, par-ticularly the primary care setting. Effective pro-grams to treat or prevent obesity in culturally,ethnically, and socioeconomically diverse popu-lations need to be developed and tested. Simplescreening tools should be tested for their predic-tive value in achieving lifestyle modificationsthat lead to weight loss or weight control prac-tices. Research is needed on identifying appro-priate and successful intervention content; forexample, magnitude of weight loss goals (smallerchanges versus larger changes), and goals for therate of weight loss (1 lb versus 2 lb per week;initial weight loss goal of 5 percent of bodyweight and, subsequently, an additional 5 per-cent versus a single initial goal of 10 percent atthe outset). Of particular importance is researchon the optimal amount of physical activity topromote weight loss, the maintenance of weightloss, and the prevention of obesity. Also impor-tant are strategies which preserve muscle andbone in the face of weight loss. More research isneeded on identifying the characteristics of indi-viduals who have successfully maintained theirweight loss over the long term.Research on surgical interventions for weightloss should include evaluating surgical risk,including not only complications, morbidity,and mortality, but also long-term postoperativesurveillance to monitor vitamin and mineralnutritional adequacy. Evaluation of the healthbenefits of weight loss from surgery shouldinclude changes in fat distribution; cardiorespi-ratory fitness; obesity-related comorbidities,including blood pressure, blood lipids, and glu-cose tolerance; and degree of success in long-term weight loss maintenance. Finally, researchis needed on techniques for integrating behav-ioral methods to promote long-term mainte-nance of weight loss after surgical treatment.Likewise, research on pharmacologic interven-tions for weight loss should include evaluatingchanges in fat distribution, cardiorespiratory fit-ness, obesity-related comorbidities, and thedegree of success of long-term weight loss main-tenance. Better methods for integrating behav-ioral methods, along with pharmacologic treat-ment, should also be investigated.FUTURE RESEARCHA6
99Finally, research is needed on environmental andpopulation-based intervention methods, includ-ing community- and school-based interventions,to augment public health approaches towardpromoting weight maintenance and preventingobesity in the general population.CAUSES AND MECHANISMS OFOVERWEIGHT AND OBESITYThe regulation of energy balance needs to beexplored, including the neuroendocrine factorsthat control energy intake, energy expenditure,and the differentiation of adipose tissue resultingfrom excess calories. The genes that are impor-tant in human obesity need to be identified.These include those that alter eating and physi-cal activity behaviors, those that affect thermoge-nesis, and those associated with the comorbidi-ties of obesity. The roles of environmental andbehavioral influences on metabolic factorsimportant in obesity, as well as gene-environ-ment interactions, need to be studied. Predictivefactors should be examined to identify who ismost at risk of developing obesity, and whetherthere are critical periods of life when these fac-tors are most operative. In addition, the influ-ence of the intrauterine environment on thedevelopment of obesity needs to be investigated,particularly to determine whether early depriva-tion leads to a later propensity for overweightand associated comorbidities, such as insulinresistance, or if high maternal weight gain andhigh birth weight are related to the risk of obesi-ty and its comorbidities.ABDOMINAL FAT, BODY WEIGHT ANDDISEASE RISKThe influence of abdominal fat independent oftotal body fat on health risk needs to be furtherdefined. More information is needed on the rela-tionship between differential body fat compart-ments and increased risk, the distribution ofbody fat compartments among various racialgroup populations, and the relationship betweenabdominal fat and disease risk in racial groups.Weight loss studies should include measure-ments of abdominal fat, as well as cardiorespira-tory fitness, to better assess health improvement.Intentional weight loss treatments need to beexamined in terms of their acute and chroniceffect on the development and progression ofdiabetes, heart disease, and overall mortality.Large prospective studies are needed to examinethe relationship of body mass index and body fatdistribution to overall mortality.ASSESSMENT METHODSMuch of the current research is hampered by thelack of good methods to accurately, objectively,and economically assess energy intake andexpenditure, including physical activity, bodycomposition and fat distribution, and behavioraland psychological variables. More research istherefore needed to focus on measures to assessintake of fat and other dietary components, lev-els of physical activity, energy metabolism, andbody fat and visceral obesity. In addition, bettermethods for assessment of psychological, behav-ioral, and psychosocial variables that may berelated to behavioral risk factors for obesity(such as poor diet and inactive lifestyle) areneeded, and particularly so for special popula-tion segments based on race, ethnicity, andsocioeconomic status. Methods for assessing cul-ture, social integration, and psychological stressshould also be developed.BCD
1007APPENDIX I.A.1. GUIDELINESDEVELOPMENT METHODOLOGYThe panel has attempted to provide primary carepractitioners with recommendations regardingeffective strategies to evaluate and treat over-weight and obesity in adults. The panel antici-pates that the recommendations also will be fol-lowed by the large proprietary weight loss indus-try in the United States. The guideline is basedon a systematic review of the scientific literaturepublished in MEDLINE from January 1980through September 1997. This was done in theinterest of time and economy. This informationwas supplemented by material provided by thepanel and a search of appropriate references ineligible articles. The panel identified randomizedcontrolled trials as the strongest level of evidencefor the evaluation of treatment efficacy. Thus,evidence from randomized controlled trials(RCTs) serves as the basis for many of the rec-ommendations contained in this guideline.However, in some instances, the panel had tomake recommendations in the absence of RCTs.Each evidence statement (other than those withno available evidence) and each recommenda-tion is categorized by level of evidence (Athrough D) as described on page 108.Statements for which there is no available evi-dence are so indicated.1.a. Panel SelectionThe National Heart, Lung, and Blood Institute(NHLBI) Obesity Education Initiative TaskForce recommended panel members represent-ing the various expertise required to deal withthe issues of cardiovascular disease and bodyweight. The executive director of the AmericanAcademy of Family Physicians (AAFP) also pro-vided recommendations for the primary carepractitioner who would represent AAFP. Thepanel members include primary care practition-ers, clinicians from academic medical centers,nutritionists, exercise physiologists, pulmonolo-gists, cardiologists, psychologists, behaviorists,epidemiologists, and experts on cost issues.Some of the panel members were also membersof the National Institute of Diabetes andDigestive and Kidney Diseases (NIDDK)National Task Force on the Prevention andTreatment of Obesity. The funding source forthe guidelines was primarily the NHLBI, withthe cooperation of the NIDDK of the NationalInstitutes of Health.1.b. Topics Included in the GuidelinesThe panel decided to focus on all adult (18 yearsof age and older) overweight and obese patientswith a body mass index (BMI) ≥ 25, and partic-ularly those with cardiovascular risk factors.Excluded from the analysis were patients withknown genetic or hormonal syndromes andpharmacologically induced obesity.In addition, since pediatric obesity contributesto about one-third of adult obesity, the panel feltthat some attention needed to be focused on theissue of overweight in children and adolescents(see Appendix III). Treatment issues surroundingoverweight children and adolescents are quitedifferent from the treatment of adults. The panelrecommended that this issue warrants its ownproject as soon as possible.APPENDICES1
101The therapeutic interventions requiring exami-nation included diet, physical activity, behaviortherapy, pharmacological therapy, surgery, andcombinations of these. No clinical interventionswere excluded at the outset. Because of theimportance and complexity of the primary pre-vention of obesity, the panel decided not to dealwith these concerns in detail and that the issuebe addressed in a separate document. However,clinical interventions to prevent further weightgain in patients already overweight, or thosepatients not currently overweight at high risk forbecoming overweight, were considered relevanttherapeutic interventions. The guideline alsocontains some information on the cost of obesity.In terms of reliable diagnostic measures, thepanel decided to concentrate on useful toolsreadily available in a physician’s office, i.e.,weight, height and BMI. The panel did notsearch for evidence to prove that measures ofbody composition or metabolic rate are usefultools; however, they did include studies usingthese diagnostic measures.1.c. Development of the Evidence ModelThe panel determined the evidence model usedto develop the guidelines for the evaluation andtreatment of obesity (Figure 7, and Figures 7aand 7b). The model contains two broad cate-gories, the association of weight to cardiovascu-lar and noncardiovascular mortality and morbid-ity in the population and the clinical issues relat-ed to the treatment of the overweight individual.The population portion of the model addressessuch questions as, “Is there an associationbetween weight and cardiovascular risk factors?”,and “Is there an association between weight andcardiovascular and noncardiovascular morbidityand mortality?” The patient portion of themodel focuses on the clinical setting andaddresses such questions as, “Does a reductionin weight result in a decrease in cardiovascularmorbidity and mortality?”The model illustrates the areas requiring a thor-ough examination of the relevant scientific liter-ature through data abstraction and evaluation(the hollow shaded arrows), the areas where theliterature has been firmly established (the hollowarrows), and the areas requiring a considerationof the preexisting evidence for minority popula-tions (the crosshatched arrows). The arrows inthe model represent the questions the panel con-sidered when looking at the evidence (see theEvidence Model questions on pages 110 to 111).1.d. Search and Review of the LiteratureThe literature was searched and systematicallyreviewed bys establishing a priori eligibility criteria forinclusion of studies;s reviewing titles and abstracts to select promis-ing articles;s reviewing these full articles; ands compiling evidence tables summarizing thosearticles that met the inclusion criteria.The panel decided on the parameters for the lit-erature search, particularly for the search dealingwith the model linking treatment to weight loss.The parameters of the search included year ofpublication, country, language, study design,length of the study, outcome measures (which,what, and when), and patient characteristics (ageand weight).The panel decided to limit its search to English-language studies, but included foreign studiesthat provide an English abstract and are ran-domized controlled clinical trials. The searchincluded human studies only published inMEDLINE from January 1980 to September1997. No editorials, letters, or case reports wereaccepted. The panel was aware that two bookspublished in the early 1980s coded the bulk ofthe obesity literature written to that point.
105Panel staff and librarians at the National Libraryof Medicine’s National Information Center forHealth Services Research determined the med-ical subject heading terms used in the literaturereview. The number of citations found by topicarea is provided in Appendix I.A.2.Figure 8 illustrates the various steps used to col-lect the evidence. Of the 43,627 titles acquiredfrom the MEDLINE search, 18,217 were dupli-cates, leaving 25,410 titles to be considered.Using the ProCite reference database, the 25,410titles were screened by two reviewers and ulti-mately marked for their appropriateness.Abstracts for the appropriate 8,040 remainingtitles were then screened for relevance in tworounds. Two individuals independently reviewedeach abstract using the inclusion and exclusioncriteria developed by the panel. A third reviewerwould sporadically check the quality of thescreening. This review led to 2,440 possibly rele-vant abstracts. Due to the greater quality of theevidence, the decision was made that random-ized controlled trials pertinent to the treatmentportion of the model be the top priority for dataabstraction. Ultimately, 394 articles of random-ized controlled trials were reviewed for dataabstraction.1.e. Literature Abstraction and Preparation ofEvidence TablesIn order to abstract the correct data related tothe treatment portion of the model in a consis-tent manner, a 25-page form called the “CriticalReview Status Sheet (CRSS)” was developed.The CRSS was reviewed and approved by thepanel. It took into account the inclusion andexclusion criteria described below and allowedfor (1) the collection of data on the studys pur-pose and design; (2) the intervention parame-ters, including the format, setting, practitioners,population demographics, and detailed descrip-tions of the intervention per se, be it dietary,physical activity, behavioral, pharmacotherapy,or other types of interventions (surgery,acupuncture, hypnosis, etc.); and (3) outcomemeasures, including the method of determiningthe outcome, and adverse events. The CRSS waspilot-tested by staff as well as by staff of the SanAntonio Cochrane Center.The San Antonio Cochrane Center is one of 12centers around the world that comprise theCochrane Collaboration. The CochraneCollaboration is an international organizationestablished in 1993 whose mission is to prepare,maintain, and disseminate systematic reviewsand meta-analysis of health care interventions.The San Antonio Center provides advice onmultiple aspects of systematic review, includingsearching and selecting materials, abstractingmaterials, organizing materials into evidencetables, and performing meta-analysis.Inclusion and exclusion criteria—The appropri-ateness of an article was determined by applyingcriteria determined by the panel. These criteriaincluded the time frame for the study, i.e., theminimum amount of time that must pass beforethe outcome measure is made; how body weightwas reported; and the type and size of the study.In defining the time frame, the panel needed toclarify if the important outcome was weight lossor sustained weight loss. Due to the importanceof both outcomes, the panel decided to includestudies that considered the effects during theacute phase of weight loss, and those that exam-ined effects during the maintenance phase. Bothtypes of studies were deemed important, andtwo cutpoints were considered appropriate forfollow-up. At a minimum, studies had to have atime frame from start to finish of at least 4months. However, in order to consider the ques-tion of long-term maintenance, studies with out-come data provided at approximately 1 year orlonger were examined. The panel decided toexclude studies that used only self-reportedweights for their measurement. No exclusion ofstudies was made by study size.
106Chapter 7: AppendicesFigure 8Obesity Guideline Development ProcessEvidence Collection SchemaFinal EvidenceModel defines ~ 39relationshipsNLM MEDLINE literature searchassesses each relationship inEvidence ModelNLM MEDLINE literature searchmerged into Procite referencedatabase (43,627 records)Print titles of Procite database(25,410 records)(Two rounds of screening)Procite reference databasemarked for duplicates(18,217 dupes)19,596 records not relevant2,440 abstracts markedas possibly relevant forarticle retrievalCritical review status sheetdatabase and 394 articles(primarily RCTs) sent to the SanAntonio Cochrane Center5,600 abstracts not relevant8,040 abstracts screenedfor relevance(Two rounds)Expert Panel memberliterature search performedand merged into Procitedatabase (2,226 abstracts)
107Preparation of evidence tables—Of the 394 arti-cles of randomized controlled trials consideredfor data abstraction by the San AntonioCochrane Center, 158 were rejected for a varietyof reasons; i.e., randomization was not adequateor subjects were not all overweight or obese.Ultimately, 236 articles were abstracted by twoseparate reviewers who independently read andabstracted each article according to the CRSS.The reviewers met, compared coding, and re-solved discrepancies. The data were then com-piled into individual evidence tables developedfor each RCT. In addition, summary tables weredeveloped to compile the evidence necessary toaddress the 23 questions relevant to the treat-ment portion of the model. All of the evidencetables will be available for online retrieval (seeinformation regarding availability of the fullobesity evidence report and other resources relat-ed to weight control on pages 165-167, and226-228.1.f. The Formulation of the Evidence into theGuidelinesIn order to consider the evidence for each of the23 questions, the panel met in groups of 6 to 10members. During the first round of consideringthe evidence, 12 small groups of panel memberswere randomly assigned to consider the evidencefor 2 or 3 questions. They developed evidencestatements and determined the strength of theevidence using the criteria noted on page 108.Their recommendations were presented to thefull panel, which then made additional recom-mendations regarding their conclusions.After considering additional studies for ques-tions where adequate RCTs were not available,another iteration of evidence statements and rec-ommendations was developed. The full panelmet again to consider this iteration, which wasconsidered by small groups assigned to focus oneither a specific treatment modality or outcomemeasure. The small groups brought their recom-mendations back to the full panel for their finalconsideration. The panel weighed the evidencebased on a thorough examination of the thresh-old or magnitude of the treatment effect.Criteria:Each evidence statement (other than those withno available evidence) and each recommenda-tion is categorized by a level of certainty (Athrough D) as described on the next page. Theconsensus process used for drawing conclusionsand writing the recommendations was a groupprocess that took into account all opinions.Conclusions reflect the widest possible agree-ment of the panel.1.g. Consideration of Special Populations andSituationsThe panel also evaluated population factors andclinical situations that might potentially influ-ence the physiological, medical, behavioral, orsociocultural context for obesity identificationand treatment. Evidence on special populationsand situations was captured from non-RCT evi-dence when available, but in many cases suchevidence was meager. Population factors selectedfor special consideration with respect to obesityclassification and treatment were age, gender,race/ethnicity, socioeconomic status, pregnancy,eating disorders, sleep apnea, extreme obesity,concurrent treatment of other major conditions(such as heart disease or diabetes), and treatmentof obesity in conjunction with smoking cessa-tion.These considerations were addressed in all stagesof guideline development. Persons with relevantspecial expertise were invited to serve on thepanel. The search for evidence was formulated toascertain studies that include special popula-tions. Studies evaluated were coded to permitanalysis of the extent to which the special popu-lations or situations of interest had been includ-ed and whether findings for these populationsand situations deviated from those for themajority populations in these studies. In thenumerous cases where evidence on these special
108Chapter 7: AppendicesE V I D E N C E C A T E G O R I E SEvidenceCategoryASources ofEvidenceRandomizedcontrolled trials(rich body ofdata)DefinitionEvidence is from endpoints of well-designed RCTs (or trialsthat depart only minimally from randomization) that providea consistent pattern of findings in the population for whichthe recommendation is made. Category A therefore requiressubstantial numbers of studies involving substantial numbersof participants.B Randomizedcontrolled trials(limited bodyof data)Evidence is from endpoints of intervention studies thatinclude only a limited number of RCTs, post hoc or subgroupanalysis of RCTs, or meta-analysis of RCTs. In general,Category B pertains when few randomized trials exist, they aresmall in size, and the trial results are somewhat inconsistent, orthe trials were undertaken in a population that differs from thetarget population of the recommendation.C NonrandomizedtrialsObservationalstudiesEvidence is from outcomes of uncontrolled or nonrandomizedtrials or from observational studies.D PanelConsensusJudgmentExpert judgment is based on the panel’s synthesis of evidencefrom experimental research described in the literature and/orderived from the consensus of panel members based on clinicalexperience or knowledge that does not meet the above-listedcriteria. This category is used only in cases where the provisionof some guidance was deemed valuable but an adequatelycompelling clinical literature addressing the subject of the rec-ommendation was deemed insufficient to justify placement inone of the other categories (A through C).
109issues was insufficient to meet standards forinclusion in the main text of the guideline,potentially important issues were identified forthe user and cross-referenced to an appendix ordiscussed in sidebar text.1.h. External Review of the GuidelinesThe external review of the guidelines included115 reviewers from 59 government agencies,professional societies, and consumer groups rep-resented on the Coordinating Committees of theNational Cholesterol Education Program andthe National High Blood Pressure EducationProgram, as well as the NIDDK’s National TaskForce on the Prevention and Treatment ofObesity, and selected members of the NorthAmerican Association for the Study of Obesity.Reviewers were asked to evaluate the guidelinebased on five criteria: validity, clarity, flexibility,completeness, and clinical applicability.Reviewers were also encouraged to provide addi-tional comments. Comments of the peer review-ers were evaluated by the panel and panel staff,and were incorporated into the guideline whereappropriate.1.i. Caveats to Recommendation UseIn applying these guideline recommendations,the reader should note some caveats:s The emphasis of these guidelines was to iden-tify effective interventions, not to rank-orderinterventions in terms of relative efficacy oreffectiveness. The panel chose not to empha-size comparisons among interventions,because there were few studies that comparedlong-term outcomes, and, since patient pref-erence often dictates choice of therapy, wewished to present a menu of options ratherthan a ranked list of choices.s When no evidence was available on the effica-cy of various treatments, the panel usuallyrendered no opinion. An absence of studiesshould not be confused with an absence ofeffect. While clinicians may wish to useproven therapies in preference to untestedtherapies, the lack of testing does not provethat the untested therapy does not work.s The limitations of RCTs must be kept inmind. The RCT is the primary method fordemonstrating efficacy. Often, patientsenrolled in RCTs differ from the patients in aprimary care practice, and effectiveness in thecommunity may differ from efficacy as mea-sured in an RCT. The potential exists for mis-interpretation of clinical trial results. Analysisof endpoints not specified at the outset, posthoc, or subgroup analyses should be viewed ashypothesis-generating rather than hypothesis-testing.APPENDIX I.A.2. LITERATURE REVIEWThe results of the literature search are providedbelow in terms of the population (top) and clini-cal (bottom) portions of the evidence model. Atotal of 43,627 records were obtained from thesearch; the following notes the number ofrecords per topic.Population model search citations by topic area(22,017 total):cardiovascular disease 3,675 citationshypertension 93 citationsdiabetes 65 citationsdyslipidemia 73 citationscardiovascular morbidityand mortality 225 citationsnoncardiovascular morbidityand mortality 4,272 citationsClinical model search citations by topic area(11,268 total):cardiovascular disease 1,866 citationshypertension 616 citationsdiabetes 380 citationsdyslipidemia 917 citations
110Chapter 7: AppendicesArrowNumber Question1 Not considered2 What is the evidence that relates weight to abdo