Diabetes Mellitus Complicating Pregnancy MARK B. LANDON, PATRICK M. CATALANO, AND STEVEN G. GABBE CHAPTER 37Pathophysiology 977 Congenital Malformations 986 Detection and Signiﬁcance of Normal Glucose Tolerance 977 Fetal Macrosomia 987 Gestational Diabetes Mellitus 992 Glucose Metabolism 977 Hypoglycemia 988 Treatment of the Patient with Type 1Diabetes Mellitus 979 Respiratory Distress Syndrome 988 or Type 2 Diabetes Mellitus 994 Type 1 Diabetes Mellitus 980 Calcium and Magnesium Ketoacidosis 996 Type 2 Diabetes/Gestational Metabolism 989 Antepartum Fetal Evaluation 997 Diabetes 980 Hyperbilirubinema and Timing and Mode of Delivery 999 Amino Acid Metabolism 983 Polycythemia 989 Glucoregulation During Labor and Lipid Metabolism 984 Maternal Classiﬁcation and Risk Delivery 1000 Maternal Weight Gain and Energy Assessment 989 Management of the Patient with Expenditure 984 Nephropathy 990 Gestational Diabetes 1000Perinatal Morbidity and Mortality 985 Retinopathy 991 Counseling the Diabetic Patient 1002 Fetal Death 985 Coronary Artery Disease 992 Contraception 1003 KEY ABBREVIATIONS The introduction of insulin therapy 85 years ago remains an important landmark in the care of pregnancy American College of Obstetricians and ACOG for the diabetic woman. Before insulin became available, Gynecologists pregnancy was not advised because it was likely to be Biophysical proﬁle BPP accompanied by fetal mortality and a substantial risk Continuous subcutaneous insulin CSII for maternal death. Over the past 35 years, management infusion (pump therapy) techniques have been developed which can prevent many Depomedroxyprogesterone acetate DMPA complications of diabetic pregnancy. These advances, Diabetic ketoacidosis DKA based on understanding of pathophysiology, now result Disposition index DI in perinatal mortality rates in optimally managed cases Gestational diabetes mellitus GDM that approach that of the normal population. This dra- Glucose tolerance test GTT matic improvement in perinatal outcome can be largely Glucose transporter GLUT attributed to clinical efforts to establish improved mater- Hemoglobin A1c HbA1c nal glycemic control both before conception and during High-density lipoprotein HDL gestation (Fig. 37-1). Excluding major congenital malfor- Hyaline membrane disease HMD mations, which continue to plague pregnancies in women Infant of the diabetic mother IDM with type 1 and type 2 diabetes mellitus, perinatal loss for Insulin-dependent diabetes mellitus IDDM the diabetic woman has fortunately become an uncom- Insulin-like growth factor IGF mon event. Low-density lipoprotein LDL Although the beneﬁt of careful regulation of maternal Maternal serum alpha-fetoprotein MSAFP glucose levels is well accepted, failure to establish optimal Maturity onset diabetes of youth MODY glycemic control as well as other factors continue to Nonstress test NST result in signiﬁcant perinatal morbidity. For this reason, Oral contraceptive OC both clinical and basic laboratory research efforts con- Phosphatidylglycerol PG tinue to focus on the etiology of congenital malforma- Respiratory distress syndrome RDS tions and fetal growth disorders. Clinical experience has Total urinary protein excretion TPE also resulted in a more realistic appreciation of the impact Tumor necrosis factor-α TNF-α that vascular complications can have on pregnancy and Urinary albumin excretion UAE the manner in which pregnancy may impact these disease Very-low-density lipoprotein VLDL processes. With modern management techniques and an organized team approach, successful pregnancies have976
Chapter 37 Diabetes Mellitus Complicating Pregnancy 977 45 40 35 Perinatal mortality rate % 30 25Figure 37-1. Perinatal mortalityrate in pregnancy complicated 20by insulin-dependent diabetesmellitus. 15 10 5 0 1925 1935 1945 1955 1965 1975 1985 Present Time periodbecome the norm even for women with the most com- postprandial glucose levels and increased insulin responseplicated diabetes. in late gestation. However, early gestation can be viewed Gestational diabetes mellitus (GDM), the most common as an anabolic state because of the increases in maternaltype of diabetes found in pregnancy, represents a continu- fat stores and decrease in free fatty acid concentration.ing challenge for both clinicians and investigators. After Weiss et al.1 have described signiﬁcant decreases in mater-40 years since the concept of GDM was introduced, the nal insulin requirements in early gestation. The mecha-clinical signiﬁcance of this disorder, particularly in its nism for this decrease in insulin requirements have beenmildest variety, sparks great debate. Controversy also ascribed to various factors including increased insulinremains concerning screening techniques, diagnostic cri- sensitivity, decreased substrate availability secondary toteria, thresholds for insulin initiation, and whether oral factors such as nausea, the fetus acting as a glucose sink,hypoglycemic agents are suitable treatment. or enhanced maternal insulin secretion. Longitudinal Before considering these clinical issues, it is important studies in women with normal glucose tolerance haveto understand the metabolic effects of pregnancy in rela- shown signiﬁcant alterations in all aspects of glucosetion to the pathophysiology of diabetes mellitus. metabolism as early as the end of the ﬁrst trimester.2 There are progressive increases in insulin secretion in response to an intravenous glucose challenge with advanc-PATHOPHYSIOLOGY ing gestation (Fig. 37-2A and B). The increases in insulin concentration are more pronounced in lean as comparedNormal Glucose Tolerance to obese women, most probably as a response to the greater decreases in insulin sensitivity in lean women as There are signiﬁcant alterations in maternal metabolism will be described later. Data regarding insulin clearance induring pregnancy, which provide for adequate maternal pregnancy are limited. In separate studies Bellman,3 Lindnutritional stores in early gestation in order to meet the et al.,4 and Burt and Davidson5 reported no differenceincreased maternal and fetal demands of late gestation in insulin disappearance rate when insulin was infusedand lactation. Although we are apt to think of diabetes intravenously in late gestation in comparison with non-mellitus as a disorder exclusively of maternal glucose gravid subjects. In contrast, Goodner and Freinkel,6 usingmetabolism, in fact, diabetes mellitus affects all aspects a radiolabeled insulin described a 25-percent increaseof nutrient metabolism. In this section, we will consider in insulin turnover in a pregnant as compared with amaternal glucose metabolism as it relates to pancreatic nonpregnant rat model. Catalano et al.7 using the eug-β-cell production of insulin and insulin clearance, endog- lycemic-clamp model reported a 20-percent increase inenous (i.e., primarily hepatic) glucose production and insulin clearance in lean women and 30-percent increasesuppression with insulin and peripheral glucose insulin in insulin clearance in obese women by late pregnancysensitivity. We also address maternal protein and lipid (Fig. 37-3). Although the placenta is rich in insulinase,insulin metabolism. Finally, the impact of these alterna- the exact mechanism for the increased insulin clearancetions on maternal metabolism are examined as they relate in pregnancy remains speculative.to maternal energy expenditure and fetal growth. Although there is a progressive decrease in fasting glucose with advancing gestation, the decrease is most probably a result of the increase in plasma volume in early gestationGlucose Metabolism and increase in fetoplacental glucose use in late gestation. Kalhan and Cowett,8,9 using various stable isotope meth- Normal pregnancy has been characterized as a dia- odologies in cross-sectional study designs, were the ﬁrstbetogenic state because of the progressive increase in to describe increased fasting hepatic glucose production
978 Section VI Pregnancy and Coexisting Disease 900 Lean control 800 Lean control Obese control Obese control 800 700 Insulin clearance (ml/m2/min) 700 600 600 Insulin ( U/ml) 500 500 400 400 300 300 200 200 100 100 0 0 Pregravid Early Late Pregravid Early Late A pregnancy pregnancy pregnancy pregnancy 8000 Lean control Figure 37-3. Longitudinal increases in metabolic clearance rate Obese control of insulin (ml/m2/min) in lean and obese women with normal glucose tolerance; pregravid, and early and late pregnancy. 6000 200 Lean control Obese control Insulin ( U/ml) 4000 175 Glucose (mg/min) 2000 150 0 125 Pregravid Early Late B pregnancy pregnancyFigure 37-2. Longitudinal increase in insulin response to anintravenous glucose challenge in lean and obese women 100with normal glucose tolerance, pregravid, and early and late Pregravid Early Latepregnancy. A, First phase: Area under the curve from 0 to 5 pregnancy pregnancyminutes. B, Second phase: Area under the curve from 5 to Figure 37-4. Longitudinal increase in basal endogenous (pri-60 minutes. marily hepatic) glucose production (mg/min) in lean and obese women with normal glucose tolerance; pregravid, early and late pregnancy.in late pregnancy. Additionally, Catalano et al.,10 usinga stable isotope of glucose in a prospective longitudinalstudy design reported a 30-percent increase in maternal Estimates of peripheral insulin sensitivity in pregnancyfasting hepatic glucose production with advancing gesta- have included the measurement of insulin response to ation (Fig. 37-4), which remained signiﬁcant even when ﬁxed oral or intravenous glucose challenge or the ratio ofadjusted for maternal weight gain. Tissue sensitivity to insulin to glucose under a variety of experimental condi-insulin involves both liver and peripheral tissues, pri- tions. In recent years, newer methodologies such as themarily skeletal muscle. The increase in fasting maternal minimal model12 and the euglycemic-hyperinsulinemic13hepatic glucose production occurred despite a signiﬁcant clamp have improved our ability to quantify peripheralincrease in fasting insulin concentration, thereby indicat- insulin sensitivity. In lean women in early gestation,ing a decrease in maternal hepatic glucose sensitivity in Catalano et al.14 reported a 40-percent decrease inwomen with normal glucose tolerance. Additionally, in maternal peripheral insulin sensitivity using the eugly-obese women, there was a decreased ability of infused cemic-hyperinsulinemic clamp. However, when adjustedinsulin to suppress hepatic glucose production in late for changes in insulin concentrations during the clampgestation as compared with pregravid and early pregnancy and residual hepatic glucose production (i.e., the insulinmeasurements, thereby indicating a further decrease in sensitivity index), insulin sensitivity decreased only 10hepatic insulin sensitivity11 in obese women. percent (Fig. 37-5). In contrast there was a 15-percent
Chapter 37 Diabetes Mellitus Complicating Pregnancy 979 0.20 Lean control resistance such as leptin, tumor necrosis factor-α (TNF- Obese control α), and resistin. Among these factors, TNF-α and leptin are known to be produced in the placenta and, therefore, could play a central role in the development of insulin 0.15 resistance. A recent study by Kirwan et al.23 reported Insulin sensitivity index that TNF-α was inversely correlated with the changes in insulin sensitivity before conception through late ges- tation. In combination with other placental hormones, 0.10 multivariate stepwise regression analysis revealed that TNF-α was the strongest independent predictor of insulin sensitivity in pregnancy, accounting for approximately half of the variance in the decrease in insulin sensitivity 0.05 during gestation. Placenta glucose transport is a nonenergy requiring process and takes place through facilitated diffusion. Glucose transport is dependent on a family of glucose 0.00 transporters referred to as GLUT glucose transporter Pregravid Early Late pregnancy pregnancy family. The principal glucose transporter in the pla- centa is GLUT 1, which is located in the syncytiotro-Figure 37-5. Longitudinal changes in the insulin sensitivity phoblast.24 GLUT 1 is located on both the microvillusindex (glucose infusion rate adjusted for residual endogenousglucose production and insulin concentrations achieved during and basal membranes. Basal membrane GLUT 1 maythe glucose clamp) in lean and obese women with normal be the rate-limiting step in placental glucose transport.glucose tolerance, pregravid, and early and late gestation. There is a two- to threefold increase in the expression of syncytiotrophoblast glucose transporters with advancing gestation.25 Although GLUT 3 and GLUT 4 expressionincrease in the insulin sensitivity index in obese women in have been identiﬁed in placental endothelial cells andearly pregnancy as compared with pregravid estimates.15 intervillous nontrophoblastic cells, respectively, the roleHence, the decrease in insulin requirements in early ges- they may play in placental glucose transport remainstation observed in some women requiring insulin may speculative.26,27be a consequence of an increase in insulin sensitivity,particularly in women with decreased insulin sensitivityprior to conception. DIABETES MELLITUS As compared with the metabolic alterations in earlypregnancy, there is a uniformity of opinion regarding the Diabetes mellitus is a chronic metabolic disorder char-decrease in peripheral insulin sensitivity in late gestation. acterized by either absolute or relative insulin deﬁciency,Spellacy and Goetz16 were among the ﬁrst investigators resulting in increased glucose concentrations. Althoughto report an increase in insulin response to a glucose glucose intolerance is the common outcome of diabeteschallenge in late gestation. Additionally, Burt17 demon- mellitus, the pathophysiology remains heterogeneous.strated that pregnant women experienced less hypoglyce- The two major classiﬁcations of diabetes mellitus aremia in response to exogenous insulin in comparison with type 1, formerly referred to as insulin-dependent dia-nonpregnant subjects. Later research by Fisher et al.18 betes or juvenile onset diabetes, and type 2, formerlyusing a high-dose glucose infusion test, Buchanan et al.19 referred to as non–insulin-dependent or adult-onset dia-using the Bergman minimal model, and Ryan et al.20 and betes. During pregnancy, classiﬁcation of women withCatalano et al.2 using the euglycemic-hyperinsulinemic diabetes has often relied on the White classiﬁcation,28clamp have demonstrated a decrease in insulin sensitiv- ﬁrst proposed in the 1940s. This classiﬁcation is based onity ranging from 33 percent to 78 percent. It should be factors such as the age of onset of diabetes and duration,noted, however, that all these quantitative estimates of as well as end organ involvement, primarily retinal andinsulin sensitivity are very likely overestimates due to renal (Table 37-1).non–insulin-mediated glucose disposal by the fetus and All forms of diabetes can occur during pregnancy. Inplacenta. Hay et al.21 reported that in the pregnant ewe addition to type 1 and type 2 diabetes, there are geneticmodel, approximately one third of maternal glucose uti- causes of diabetes, the most common of which is maturitylization was accounted for by uterine, placental, and fetal onset diabetes of the young (MODY). MODY is charac-tissue. Additionally, Marconi et al.22 reported that based terized by β-cell dysfunction and is an autosomal domi-on human fetal blood sampling, fetal glucose concentra- nant mode of inheritance, usually becoming manifest intion was a function of fetal size and gestational age in young adulthood. Mutations in the glucokinase gene areaddition to maternal glucose concentration. a frequent cause of MODY. Various mutations have been Historically, the decrease in insulin sensitivity during described, and each mutation is associated with varyingpregnancy has been ascribed to an increased production degrees of disease severity. The most common of theseof various placental and maternal hormones, such as mutations (MODY2) occurs in the European popula-human placental lactogen, progesterone, estrogen, cor- tion and involves the glucokinase gene. Because the agetisol, and prolactin. However, more recent evidence has of onset of diabetes in women with MODY coincidesfocused on the role of several new mediators of insulin with the reproductive years, it may be difﬁcult to distin-
980 Section VI Pregnancy and Coexisting Disease Table 37-1. Modiﬁed White Classiﬁcation of Pregnant tory epinephrine and glucagon response to hypoglycemia. Diabetic Women The deﬁciency in this counterregulatory response may be in part due to an independent effect of pregnancy. DIABETES The alterations in glucose metabolism in women with ONSET DURATION VASCULAR INSULIN type 1 diabetes are not well characterized. Because of CLASS AGE (Y) (Y) DISEASE NEED maternal insulinopenia, insulin response during gestation can only be estimated relative to pregravid requirements. Gestational diabetes Estimates of the change in insulin requirements are com- A1 Any Any 0 0 plicated by the degree of preconceptual glucose control A2 Any Any 0 + and potential presence of insulin antibodies. Weiss and Pregestational diabetes Hofman1 reported on the change in insulin requirements in women with type 1 diabetes and strict glucose control B >20 <10 0 + either before conception or before 10 weeks’ gestation. C 10–19 or 10–19 0 + There was a 12-percent decrease in insulin requirements D <10 or >20 + + F Any Any + + from 10 to 17 weeks’ gestation and a 50-percent increase R Any Any + + in insulin requirement from 17 weeks’ until delivery as T Any Any + + compared with pregravid requirements. After 36 weeks’ H Any Any + + gestation, there was a decrease in insulin requirements. A 5-percent decrease in insulin requirements after 36 weeks’Modiﬁed from White P: Pregnancy complicating diabetes. Am J Med gestation was also noted by McManus and Ryan.33 The7:609, 1949. decrease in insulin requirements was associated with a longer duration of diabetes mellitus but not with adverse perinatal outcome. The fall in insulin requirements inguish between the two. The glucokinase gene acts as a early pregnancy in women with type 1 diabetes may besensor in the β-cell, which leads to a secretory defect in a reﬂection of increased pregravid insulin sensitivity asinsulin response. Ellard et al.29 reported that 2.5 percent was described previously.of women with GDM in the United Kingdom have the Schmitz et al.34 have evaluated the longitudinal changesglucokinase mutation, whereas Stoffel30 in a small popu- in insulin sensitivity in women with type 1 diabetes inlation in the United States reported that 5 percent of early and late pregnancy as well as postpartum in com-patients had a glucokinase mutation. The implication parison with nonpregnant women with type 1 diabetes.is that if the mother has the mutation, the fetus is at an In the pregnant women with type 1 diabetes, there wasincreased risk for macrosomia. The implications for the a 50-percent decrease in insulin sensitivity only in latefetus, if the mutation is inherited from the father, are gestation. There was no signiﬁcant difference in insulina signiﬁcant decrease in growth secondary to relative sensitivity in pregnant women with type 1 diabetes ininsulinopenia. early pregnancy or within 1 week of delivery as com- pared with the nonpregnant women with type 1 diabetes. Therefore, based on the available data women with typeType 1 Diabetes Mellitus 1 diabetes appear to have a similar decrease in insulin sensitivity when compared with women with normal Type 1 diabetes mellitus is usually characterized by glucose tolerance.an abrupt onset at a young age and absolute insulinope- Relative to the issue of placental transporters (GLUTnia with life-long requirements for insulin replacement, 1), there is a report by Jansson and Powell35 describingalthough depending on the population, the onset of type 1 an increase in both basal GLUT 1 expression and glucosediabetes may occur in individuals in their third or fourth transport activity from placental tissue in women withdecades of life. Patients with diabetes mellitus may have White class D pregnancies.a genetic predisposition for antibodies directed againsttheir pancreatic islet cells. The degree of concordancefor the development of type 1 diabetes in monozygotic Type 2 Diabetes/Gestational Diabetestwins is 33 percent, suggesting that the events subsequentto the development of autoantibodies and appearance The pathophysiology of type 2 diabetes involves abnor-of glucose intolerance are also related to environmental malities of both insulin sensitive tissue (i.e., both a decreasefactors. Because of the complete dependence on exog- in skeletal muscle and hepatic sensitivity to insulin) andenous insulin, pregnant women with type 1 diabetes are β-cell response as manifested by an inadequate insulinat increased risk for the development of diabetic keto- response for a given degree of glycemia. Initially in theacidosis (DKA). Additionally, because intensive insulin course of development of type 2 diabetes, the insulintherapy is used in women with type 1 diabetes to decrease response to a glucose challenge may be increased relativethe risk for spontaneous abortion and congenital anoma- to that of individuals with normal glucose tolerance butlies in early gestation, these women are at increased risk is inadequate to maintain normoglycemia. Whether orfor hypoglycemic reactions. Studies by Diamond et al.31 not decreased insulin sensitivity precedes β-cell dysfunc-and Rosenn et al.32 have shown that women with type 1 tion in the development of type 2 diabetes continues todiabetes are at increased risk for hypoglycemic reactions be debated. Arguments and experimental data supportduring pregnancy because of diminished counterregula- both hypotheses. As noted by Sims and Calles-Escadon,36
Chapter 37 Diabetes Mellitus Complicating Pregnancy 981heterogeneity of metabolic abnormalities exists in any 900 Lean controlclassiﬁcation of diabetes mellitus. Obese control Despite the limitations of any classiﬁcation system, 800 Lean GDMcertain generalizations can be made regarding women Obese GDM 700with type 2 or GDM. These individuals are typicallyolder and more often heavier compared with individu- 600 Insulin ( U/ml)als with type 1 diabetes or normal glucose tolerance.The onset of the disorder is usually insidious, with few 500patients complaining of classical triad of polydipsia,polyphagia, and polyuria. Individuals with type 2 diabetes 400are often initially recommended to lose weight, increase 300their activity (i.e., exercise), and follow a diet that is lowin fats and high in complex carbohydrates. Oral agents 200are often used to either increase insulin response or, withnewer drugs, enhance insulin sensitivity. Individuals with 100type 2 diabetes may eventually require insulin therapy inorder to maintain euglycemia but are at signiﬁcantly less 0 A Pregravid Early pregnancy Late pregnancyrisk for DKA. Data from monozygotic twin studies havereported a lifetime risk of both twins developing type 2 8000 Lean controldiabetes that ranges between 58 percent and almost 100 Obese controlpercent, suggesting that the disorder has a strong genetic 7000 Lean GDMcomponent. Obese GDM Women with type 2 pregestational diabetes are usually 6000classiﬁed as class B diabetes according to the White clas-siﬁcation system. Women developing GDM (i.e., glucose Insulin ( U/ml) 5000intolerance ﬁrst recognized during pregnancy) sharemany of the metabolic characteristics of women with 4000type 2 diabetes. Although earlier studies reported a 10-to 35-percent incidence of islet cell antibodies in women 3000with GDM as measured by immunoﬂuorescence tech-niques,37,38 more recent data using speciﬁc monoclonal 2000antibodies have described a much lower incidence, on theorder of 1 to 2 percent,39 suggesting a low risk of type 1 1000diabetes in women with GDM. Furthermore, postpartumstudies of women with GDM have demonstrated defects 0in insulin secretory response40 and decreased insulin sen- B Pregravid Early pregnancy Late pregnancysitivity,41 indicating that typical type 2 abnormalities in Figure 37-6. A and B, Longitudinal increase in insulin responseglucose metabolism are present in women with GDM. Of to an intravenous glucose challenge in lean and obese womeninterest, the alterations in insulin secretory response and with normal glucose tolerance and gestational diabetes; pre-insulin resistance in women with a previous history of gravid, early and late pregnancy. A, First phase: Area under the curve from 0 to 5 min. B, Second phase: Area under the curveGDM as compared with a weight-matched control group from 5 to 60 min.may differ depending on whether or not the women withprevious GDM are lean or obese.42 Thus, in womenwith GDM, the hormonal events of pregnancy may repre- differences in insulin response may be related to thesent an unmasking of a genetic susceptibility to type 2 ethnicity of the various study groups. Although therediabetes. is an increase in the metabolic clearance rate of insulin There are signiﬁcant alterations in glucose metabolism with advancing gestation, there is no evidence that therein women who develop GDM relative to the changes in is a signiﬁcant difference between women with normalglucose metabolism in women with normal glucose toler- glucose tolerance and GDM.15ance. Decreased insulin response to a glucose challenge There is a signiﬁcant decrease in fasting glucose concen-has been demonstrated by Yen et al.,43 Fisher et al.,44 tration with advancing gestation in women developingand Buchanan et al.19 in women with GDM in late gesta- GDM. In late pregnancy, glucose and hepatic glucosetion. In prospective longitudinal studies of both lean and production increase in women with GDM in comparisonobese women with GDM, Catalano et al.14 also showed a with a control group.45 Whereas there was no signiﬁ-progressive decrease in ﬁrst-phase insulin response in late cant difference in either fasting glucose concentration orgestation in lean women developing GDM as compared hepatic glucose production in the longitudinal studies ofwith a weight-matched control group (Fig. 37-6A). In Catalano et al.,14,15 these differences may again be popu-contrast, in obese women developing GDM, there was lation speciﬁc or related to the degree of fasting hypo-no difference in ﬁrst-phase insulin response but rather glycemia. However, to date all reports indicate that ina signiﬁcant increase in second-phase insulin response late gestation, women with GDM have increased fastingto an intravenous glucose challenge as compared with insulin concentrations (Fig. 37-7) and less suppressiona weight-matched control group (see Fig. 37-6B). These of hepatic glucose production during insulin infusion,
982 Section VI Pregnancy and Coexisting Disease 45 Lean control 0.20 Lean control Obese control Obese control 40 Lean GDM Lean GDM 35 Obese GDM Obese GDM 0.15 Insulin sensitivity index 30 Insulin ( U/ml) 25 0.10 20 15 0.05 10 5 0 0 Pregravid Early pregnancy Late pregnancy Pregravid Early pregnancy Late pregnancyFigure 37-7. Longitudinal increase in basal or fasting insulin Figure 37-8. Longitudinal changes in the insulin sensitivity(µ/ml) in lean and obese women with normal glucose toler- index (glucose infusion rate adjusted for residual endogenousance and gestational diabetes; pregravid, and early and later glucose production and insulin concentrations achieved duringpregnancy. the glucose clamp) in lean and obese women with normal glucose tolerance and gestational diabetes; pregravid, early and late pregnancy.thereby indicating decreased hepatic glucose insulin sen-sitivity in women with GDM as compared with a weight-matched control group.14,15,45 In the studies of Xiang 1000et al.,45 there was signiﬁcant correlation between fastingfree fatty acid concentrations and hepatic glucose produc- Insulin secretion rate (ISR) 800tion, suggesting that increased free fatty acid concentra-tions may contribute to hepatic insulin resistance. Women with GDM have decreased insulin sensitivity 600 Normalin comparison with weight-matched control groups.Ryan et al.20 was the ﬁrst to report a 40-percent decrease GDM 400in insulin sensitivity in women with GDM in comparisonwith a pregnant control group in late pregnancy usinga hyperinsulinemic-euglycemic clamp. Xiang et al.,45 200found that women with GDM who had normal glucosetolerance within 6 months of delivery had signiﬁcantly 0decreased insulin sensitivity as estimated by the glucose 0.0 0.1 0.2 0.3 0.4clearance rate during a hyperinsulinemic-euglycemic Insulin sensitivity index (ISI)clamp, as compared with a matched control group.Catalano et al.,14,15 using similar techniques, described 3rd trimesterthe longitudinal changes in insulin sensitivity in both Postpartumlean and obese women developing GDM in comparison Figure 37-9. Insulin sensitivity index.with a matched control group. Women developing GDMhad decreased insulin sensitivity as compared with thematched control group (Fig. 37-8). The differences in The interactions of β-cell response and insulin sensitivityinsulin sensitivity were greatest before and during early are hallmarks of the metabolic adaptations of pregnancy.gestation, and by late gestation, the differences in insulin As described by Bergman,47 there is a ﬁxed relationshipsensitivity between the groups were less pronounced but between insulin response and insulin resistance in non-still signiﬁcant. Of interest, there was an increase in insulin pregnant individuals following a hyperbolic curve, i.e.,sensitivity from the time prior to conception through the disposition index. Buchanan48 described a similarearly pregnancy (12 to 14 weeks), particularly in those relationship between insulin response and insulin actionwomen with greatest decreases in insulin sensitivity prior during pregnancy. Indeed, when the disposition indexto conception. The changes in insulin sensitivity from the has been compared between women with normal glucosetime before conception through early pregnancy were tolerance and GDM both during and after pregnancy, thesigniﬁcantly correlated with changes in maternal weight failure of the β-cell to compensate for insulin resistancegain and energy expenditure.46 The relationship between in GDM has been similar to the hyperbolic changes inthese alterations in maternal glucose insulin sensitivity the control group (Fig. 37-9). This relationship betweenand weight gain and energy expenditure may help explain insulin sensitivity and insulin resistance, however, maythe decrease in maternal weight gain and insulin require- not hold in early pregnancy when there is both an increasements in women with diabetes in early gestation.1 in insulin sensitivity and insulin response.
Chapter 37 Diabetes Mellitus Complicating Pregnancy 983 IR Ras IRS-1/2 Glucose transport Raf P85 PI-3-K P110 GLUT4Figure 37-10. Schematic model of insulin signaling GLUT4cascade in skeletal muscle. GLUT, glucose trans- MAPKporter; IR, insulin receptor; IRS, insulin receptor GLUT4substrate. PKB Mitogenesis gene transcription GSK3 Glycogen synthesis Studies in human skeletal muscle and adipose tissue acid concentrations in early pregnancy prior to thehave demonstrated that postreceptor defects in the accretion of signiﬁcant maternal or fetal tissue.52 Theseinsulin signaling cascade are related to decreased insulin anticipatory changes in fasting amino acid metabolismsensitivity in pregnancy. Garvey et al.49 were the ﬁrst to occur after a shorter period of fasting in comparisondemonstrate that there were no signiﬁcant differences in with nonpregnant women, and may be another examplethe glucose transporter (GLUT 4) responsible for insulin of the accelerated starvation of pregnancy as describedaction and skeletal muscle in pregnant as compared with by Freinkel.53 Furthermore, amino acid concentrationsnonpregnant women. Based on the studies of Friedman such as serine correlate signiﬁcantly with fetal growthet al.50 in both pregnant women with normal glucose in both early and late gestation.54 Maternal amino acidtolerance and GDM as well as weight-matched nonpreg- concentrations were signiﬁcantly decreased in mothersnant control subjects, there appeared to be defects in the of small-for-gestational-age neonates in comparisoninsulin-signaling cascade relating to pregnancy as well as with maternal concentration in appropriately grownwhat may be additional abnormalities in women with neonates.55GDM. All pregnant women appeared to have a decrease Based on a review of various studies, Duggleby andin insulin receptor substrate-1 (IRS-1) expression. The Jackson56 have estimated that during the ﬁrst trimesterdown-regulation of the IRS-1 protein closely parallels the of a pregnancy, protein synthesis is similar to that ofdecreased ability of insulin to induce additional steps in nonpregnant. However, there is a 15-percent increasethe insulin signaling cascade, resulting in movement of in protein synthesis during the second trimester andthe GLUT 4 to the cell surface membrane and to facili- a further increase in the third trimester by about 25tate glucose transport into the cell. The downregulation percent. Additionally, there are marked interindividualof IRS-1 protein closely parallels the ability of insulin differences at each time point. These differences have ato stimulate 2-deoxyglucose uptake in vitro. In addi- strong relationship with fetal growth, that is, motherstion to the above-mentioned mechanisms, women with who had increased protein turnover in midpregnancy hadGDM demonstrate a distinct decrease in the ability of the babies who had increased lean body mass after adjust-insulin receptor β (that component of the insulin receptor ment for signiﬁcant covariables.57not on the cell surface) to undergo tyrosine phosphoryla- Amino acids can be used either for protein accrual ortion. The additional defect in the insulin signaling cascade oxidized as an energy source. Urea synthesis has beenresults in a 25-percent lower glucose transport activity estimated in a number of studies using stable isotopes.(Fig. 37-10). In general, there is a modest shift in oxidation in early pregnancy, with an accrual of amino acids for protein synthesis in late gestation.56 Furthermore, Kalhan et al.58Amino Acid Metabolism reported that there are signiﬁcant pregnancy-related adap- tations in maternal protein metabolism early in gestation Although glucose is the primary source of energy before any signiﬁcant increase in fetal protein accretion.for the fetus and placenta, there are no appreciable Preliminary studies by Catalano et al.59 have reportedamounts of glucose stored as glycogen in the fetus or that there is decreased insulin sensitivity as manifestedplacenta. However, accretion of protein is essential for by a decreased suppression of leucine turnover duringgrowth of fetoplacental tissue. There is increased nitro- insulin infusion in late gestation in all pregnant women.gen retention in pregnancy in both maternal and fetal There is evidence for an increase in basal leucine turn-compartments. There is an increase of approximately over in women with GDM as compared with a matched0.9 kg of maternal fat-free mass by 27 weeks.51 There control group. Whether these decreases in amino acidis a signiﬁcant decrease in most fasting maternal amino insulin sensitivity are related to decreased whole body/
984 Section VI Pregnancy and Coexisting Diseaseliver protein synthesis or increased breakdown are not infants of obese women had an increased birth weightknown at this time. and skinfold thickness, and higher free fatty acid levels Recently, Cetin et al.60 reported that placental amino when compared with infants of lean women.acid exchange is altered in pregnancies complicated Lipid metabolism in women with diabetes mellitus isby GDM. Ornithine concentrations were signiﬁcantly inﬂuenced by whether the woman has type 1 or type 2increased in women with GDM as compared with con- diabetes. This also applies when these women becometrols, and in the cord blood of infants of women with pregnant. In women with type 2 diabetes and GDM,GDM, there were signiﬁcant increases in multiple amino Knopp et al.64 reported an increase in triglyceride and aacids including phenylalanine and leucine but decreases decrease in HDL concentration. However, Montelongoin glutamate. The investigators speculate that in infants et al.65 reported little change in free fatty acid concentra-of women with GDM, the altered in utero fetal milieu tions through all three trimesters after a 12-hour fast.impacts fetal growth through multiple mechanisms, Koukkou et al.66 noted an increase in total triglycerideaffecting various nutrient compartments. but a lower LDL cholesterol in women with GDM. In Amino acids are actively transported across the placenta women with type 1 diabetes, there was no change in totalfrom mother to fetus through energy-requiring amino acid triglyceride but a lower cholesterol concentration, sec-transporters. These transporters are highly stereospeciﬁc, ondary to a decrease in HDL. This is of interest becausebut they have low substrate speciﬁcity. Additionally, they HDL acts as plasma antioxidant and thus may be relatedmay vary with location between the microvillus and basal to the increase in congenital malformations in womenmembranes.61 Decreased amino acid concentrations have with type 1 diabetes. Oxidative stress has been implicatedbeen reported in growth restricted neonates in compari- as a potential factor in the incidence of anomalies inson with appropriately grown neonates. Decreased amino women with type 1 diabetes.acid transporter activity has been implicated as a pos- Hyperinsulinemic-euglycemic clamp studies in preg-sible mechanism. However, the potential role, if any, of nant women with normal glucose tolerance and GDMplacental amino acid transporters in the development of revealed a decreased ability of insulin to suppress plasmafetal macrosomia in women with diabetes is currently free fatty acids with advancing gestation. Insulin’s abilityunknown.62 to suppress plasma free fatty acid was lower in women with GDM as compared to women with normal glucose tolerance.67Lipid Metabolism Taken together, these studies demonstrate that there is decreased nutrient insulin sensitivity in all women with Although there is ample literature regarding the changes advancing gestation. These decreases in insulin sensitiv-in glucose metabolism during gestation, the data regard- ity are further exacerbated by the presence of decreaseding the alterations in lipid metabolism are meager by pregravid maternal insulin sensitivity, which becomescomparison. Darmady and Postle measured serum choles- manifest in later pregnancy as GDM, resulting in greaterterol and triglyceride before, during, and after pregnancy nutrient availability and higher ambient insulin concen-in 34 normal women.63 There was a decrease in both trations for the developing fetoplacental unit, which maycholesterol and triglyceride at approximately 7 weeks’ eventually result in fetal overgrowth.gestation. Both of the levels increased progressively untilterm. There was then a decrease in serum triglyceridepostpartum. The decrease was more rapid in women who Maternal Weight Gain andbreast-fed compared with those women who bottle fed Energy Expendituretheir infants.63 Additionally, Knopp et al.64 have reportedthat there is a two- to fourfold increase in total trigly- Estimates of the energy cost of pregnancy range from aceride concentration and a 25- to 50-percent increase in cost of 80,000 kcal to a net saving of up to 10,000 kcal.66total cholesterol concentration during gestation. There is As a result, the recommendations for nutritional intake ina 50-percent increase in low-density lipoprotein (LDL) pregnancy differ and depend upon the population beingcholesterol and a 30-percent increase in high-density evaluated. Furthermore, based on more recent data, rec-lipoprotein (HDL) cholesterol by midgestation, which ommendations for individuals within a population maydecreases slightly in the third trimester. Maternal trigly- be more varied than previously believed, making generalceride and very-low-density lipoprotein (VLDL) levels guidelines for nutritional intake difﬁcult.68,69in late gestation are positively correlated with maternal The theoretical energy cost of pregnancy was originallyestriol and insulin concentrations. estimated by Hytten51 using a factorial method. The addi- Free fatty acids have been associated with fetal over- tional cost of pregnancy consisted of (1) the additionalgrowth, particularly of fetal adipose tissue. There is a maternal and fetoplacental tissue accrued during preg-signiﬁcant difference in the arteriovenous free fatty acid nancy and (2) the additional “running cost” of preg-concentration at birth much as there is with arteriovenous nancy (e.g., the work of increased cardiac output). Inglucose concentration. Knopp et al.64 reported that neo- Hytten’s model, the greatest increases in maternal energynatal birth weight was positively correlated with triglyc- expenditure occur between 10 and 30 weeks’ gestation,eride and free fatty acid concentration in late pregnancy. primarily because of maternal accretion of adipose tissue.Similar conclusions were reached by Ogburn et al.,61 who However, the mean increases in maternal adipose tissueshowed that higher insulin concentrations decrease free vary considerably among various ethnic groups. Forsumfatty acid concentrations, inhibit lipolysis and result in et al.70 reported a mean increase of more than 5 kg ofincreased fat deposition. Last, Kleigman62 reported that adipose tissue in Swedish women, whereas Lawrence
Chapter 37 Diabetes Mellitus Complicating Pregnancy 985et al.71 found no increase in adipose tissue stores in show that there is a relationship between the changes inwomen from the Gambia. maternal insulin sensitivity and accretion of adipose tissue Basal metabolic rate accounts for 60 to 70 percent in early gestation.78 The ability of women with decreasedof total energy expenditure in individuals not engaged pregravid glucose insulin sensitivity (obese women andin competitive physical activity and correlates well with women with GDM) to conserve energy, not signiﬁcantlytotal energy expenditure. As with the changes in mater- increase body fat, and make sufﬁcient nutrients avail-nal accretion of adipose tissue, there are wide variations able to produce a healthy fetus, supports the hypothesisin the change in maternal basal metabolic rate during that decreased maternal insulin sensitivity may have agestation, not only in different populations but again reproductive metabolic advantage in women when foodwithin relatively homogeneous groups. The cumulative availability is marginal. In contrast, decreased maternalenergy changes in basal metabolic rate range from a insulin sensitivity before conception in areas where foodhigh of 52,000 kcal in Swedish women72 to a net savings is plentiful and a sedentary lifestyle is more commonof 10,700 kcal in women from the Gambia71 without may manifest itself as GDM and increase the long-termnutritional supplementation. The mean increase in basal risk for both diabetes and obesity in the woman and hermetabolic rate in Western women relative to a nonpreg- offspring.79nant, nonlactating control group averages approximately20 percent.71 However, the coefﬁcient of variation ofbasal metabolic rate in these populations during gestationranges from 93 percent in women in the United Kingdom72 PERINATAL MORBIDITYto more than 200 percent in Swedish women.70 When AND MORTALITYassessing energy intake in relation to energy expenditure,however, estimated energy intake remains lower than Fetal Deaththe estimates of total energy expenditure. These discrep-ancies have usually been examined by factors such as In the past, sudden and unexplained stillbirth occurred(1) increased metabolic efﬁciency during gestation,73 (2) in 10 to 30 percent of pregnancies complicated by typedecreased maternal activity,74 and (3) unreliable assess- 1 diabetes mellitus insulin-dependent diabetes mellitus,ment of food intake.75 (IDDM).80,81 Although relatively uncommon today, such Data in nonpregnant subjects may help explain some losses still plague the pregnancies of patients who doof the wide variations in metabolic parameters during not receive optimal care. Stillbirths have been observedhuman gestation, even with homogeneous populations. most often after the 36th week of pregnancy in patientsSwinburn et al.76 reported that in the Pima Indian popula- with vascular disease, poor glycemic control, hydramnios,tion, subjects with decreased insulin sensitivity gained less fetal macrosomia, or preeclampsia. Women with vascularweight as compared with more insulin-sensitive subjects complications may develop fetal growth restriction and(3.1 versus 7.6 kg) over a period of 4 years. Furthermore, intrauterine demise as early as the second trimester. In thethe percentage weight change per year was highly cor- past, prevention of intrauterine death led to a strategy ofrelated with glucose disposal as estimated from clamp scheduled preterm deliveries for type 1 diabetic women.studies. Catalano et al.77 conducted a prospective longitu- This empiric approach reduced the number of stillbirths,dinal study in early pregnancy of the changes in maternal but errors in estimation of fetal size and gestational ageaccretion of body fat and basal metabolic rate in lean and as well as the functional immaturity characteristic ofobese women with normal GDM. Women with GDM the infant of the diabetic mother (IDM) contributed tohad decreased insulin sensitivity for glucose in early many neonatal deaths from hyaline membrane diseasegestation as compared with the control group and had (HMD).signiﬁcantly smaller increases in body fat than women The precise cause of the excessive stillbirth rate inwith normal glucose tolerance. In these lean women, there pregnancies complicated by diabetes remains unknown.was a signiﬁcant inverse correlation between the changes Because extramedullary hematopoiesis is frequentlyin fat accretion and insulin sensitivity (i.e., women with observed in stillborn IDMs, chronic intrauterine hypoxiadecreased pregravid insulin sensitivity had less accretion has been cited as a likely cause of these intrauterine fetalof body fat as compared with women with increased deaths. Studies of fetal umbilical cord blood samples inpregravid insulin sensitivity). These results are consistent pregnant women with type 1 diabetes have demonstratedwith a previous report showing that total weight gain in “relative fetal erythremia and lactic acidemia.”82 Mater-women with GDM was 2.5 kg less as compared with a nal diabetes may also produce alterations in red bloodweight-matched control group.77 cell oxygen release and placental blood ﬂow.83 In the basal state, lean women increase the use of car- Reduced uterine blood ﬂow is thought to contribute tobohydrate as a metabolic fuel, whereas in obese women, the increased incidence of intrauterine growth restrictionthere is an increased use of lipids for oxidative needs. observed in pregnancies complicated by diabetic vascu-However, with the decrease in insulin sensitivity in late lopathy. Investigations using radioactive tracers have alsogestation, all women lean or obese with normal glucose suggested a relationship between poor maternal meta-tolerance or GDM have an increase in fat oxidation and bolic control and reduced uteroplacental blood ﬂow.84a decrease in nonoxidative glucose metabolism (storage). Ketoacidosis and preeclampsia, two factors known toOf interest, these increases in lipid oxidation in pregnancy be associated with an increased incidence of intrauter-are positively correlated with the increases in maternal ine deaths, may further decrease uterine blood ﬂow. Inleptin concentrations, possibly accounting for a role of DKA, hypovolemia and hypotension caused by dehy-leptin in human pregnancy. The results of these studies dration may reduce ﬂow through the intervillous space,
986 Section VI Pregnancy and Coexisting Diseasewhereas in preeclampsia, narrowing and vasospasm of week of gestation.90 Central nervous system malforma-spiral arterioles may result. tions, particularly anencephaly, open spina biﬁda, and, Alterations in fetal carbohydrate metabolism also may possibly, holoprosencephaly, are increased 10-fold.90,91contribute to intrauterine asphyxia.85,86 There is consider- Cardiac anomalies, especially ventricular septal defectsable evidence linking hyperinsulinemia and fetal hypoxia. and complex lesions such as transposition of the greatHyperinsulinemia induced in fetal lambs by an infusion vessels, are increased ﬁvefold. The congenital defectof exogenous insulin produces an increase in oxygen con- thought to be most characteristic of diabetic embryopa-sumption and a decrease in arterial oxygen content.85,86 thy is sacral agenesis or caudal dysplasia, an anomalyPersistent maternal-fetal hyperglycemia occurs indepen- found 200 to 400 times more often in offspring of dia-dent of maternal uterine blood ﬂow, which may not be betic women (Fig. 37-11). However, this defect is notincreased enough to allow for enhanced oxygen delivery pathognomonic for diabetes, since it occurs in nondia-in the face of increased metabolic demands. Thus, hyper- betic pregnancies.insulinemia in the fetus of the diabetic mother appears to Impaired glycemic control and associated derangementsincrease fetal metabolic rate and oxygen requirement in in maternal metabolism appear to contribute to abnormalthe face of several factors such as hyperglycemia, ketoaci- embryogenesis. The notion of excess glucose as the singledosis, preeclampsia, and maternal vasculopathy, which teratogenic agent in diabetic pregnancy has thus beencan reduce placental blood ﬂow and fetal oxygenation. replaced with the view of a multifactorial etiology90 (see the box “Proposed Factors Associated with Teratogenesis in Pregnancy Complicated by Diabetes Mellitus”).Congenital Malformations Maternal hyperglycemia has been proposed by most investigators as the primary teratogenic factor, but hyper- With the reduction in intrauterine deaths and a marked ketonemia, hypoglycemia, somatomedin inhibitor excess,decrease in neonatal mortality related to HMD and trau- and excess free oxygen radicals have also been sug-matic delivery, congenital malformations have emerged gested.90 The proﬁle of a woman most likely to produceas the most important cause of perinatal loss in pregnan- an anomalous infant would include a patient with poorcies complicated by type 1 and type 2 diabetes mellitus. periconceptional control, long-standing diabetes, andIn the past, these anomalies were responsible for only10 percent of all perinatal deaths. At present, however,malformations account for 30 to 50 percent of perinatalmortality.81 Neonatal deaths now exceed stillbirths inpregnancies complicated by pregestational diabetes mel-litus, and fatal congenital malformations account for thischanging pattern. Most studies have documented a two- to sixfold increasein major malformations in infants of type 1 and type 2diabetic mothers. At The Ohio State University Diabetesin Pregnancy Program, we observed 29 congenital anom-alies in 289 (10 percent) diabetic woman enrolled over a10-year period.87 In a prospective analysis, Simpson et al.observed an 8.5 percent incidence of major anomalies inthe diabetic population, whereas the malformation ratein a small group of concurrently gathered control subjectswas 2.4 percent.88 Similar ﬁgures were obtained in theDiabetes in Early Pregnancy Study in the United States.89The incidence of major anomalies was 2.1 percent in389 control patients and 9.0 percent in 279 diabeticwomen. In general, the incidence of major malformationsin worldwide studies of offspring of diabetic mothers hasranged from 5 to 10 percent (Table 37-2). The insult that causes malformations in IDM impactson most organ systems and must act before the seventh Figure 37-11. Infant of a diabetic mother. Table 37-2. Frequency of Congenital Malformations in Infants of Diabetic Mothers Proposed Factors Associated with Teratogenesis in Pregnancy Complicated By Diabetes Mellitus Mills90 25/279 9.0 Greene262 35/451 7.7 • Hyperglycemia Steel and Duncan264 12/239 7.8 • Ketone body excess Fuhrmann et al259 22/292 7.5 • Somatomedin inhibition Simpson et al99 9/106 8.5 • Arachidonic acid deﬁciency Albert et al89 29/289 10.0 • Free oxygen radical excess
Chapter 37 Diabetes Mellitus Complicating Pregnancy 987vascular disease.91 Genetic susceptibility to the terato-genic inﬂuence of diabetes may be a factor. Koppe andSmoremberg-Schoorl as well as Simpson and colleagueshave suggested that certain maternal HLA types may bemore often associated with anomalies.92,93 Several mechanisms have been proposed by whichthe above-mentioned teratogenic factors produce mal-formations. Freinkel et al.94 suggested that anomaliesmight arise from inhibition of glycolysis, the key energy-producing process during embryogenesis. He found thatD-mannose added to the culture medium of rat embryosinhibited glycolysis and produced growth restriction andderangement of neural tube closure.94 Freinkel et al.94stressed the sensitivity of normal embryogenesis to altera-tions in these key energy-producing pathways, a processhe labeled “fuel-mediated” teratogenesis. Goldman and Figure 37-12. Two extremes of growth abnormalities.Baker95 suggested that the mechanism responsible for theincreased incidence of neural tube defects in embryoscultured in a hyperglycemic medium may involve a linemia, resulting in excessive fetal growth. Increasedfunctional deﬁciency of arachidonic acid, because supple- fetal β-cell mass may be identiﬁed as early as the secondmentation with arachidonic acid or myoinositol will trimester.100 Evidence supporting the Pedersen hypothesisreduce the frequency of neural tube defects in this exper- has come from the studies of amniotic ﬂuid and cordimental model.95 Pinter and Reece,96 and Pinter et al.97 blood insulin and C-peptide concentrations. Both arehave conﬁrmed these studies and demonstrated that increased in the amniotic ﬂuid of insulin-treated womenhyperglycemia-induced alterations in neural tube closure with diabetes at term101 and correlate with neonatal fatinclude disordered cells, decreased mitoses, and changes mass.102 Lipids and amino acids, which are elevated inindicating premature maturation. These authors have pregnancies complicated by GDM, may also play a rolefurther demonstrated that hyperglycemia during organ- in excessive fetal growth by stimulating the release ofogenesis has a primary deleterious effect on yolk sac insulin and other growth factors from the fetal pancreaticfunction with resultant embryopathy. β cells and placenta. Infants of mothers with GDM have Altered oxidative metabolism from maternal diabetes an increase in fat mass compared with fat-free mass.103may cause increased production of free oxygen radicals Additionally, the growth is disproportionate, with chest/in the developing embryo, which are likely teratogenic. head and shoulder/head ratios larger than those of infantsSupplementation of oxygen radical–scavenging enzymes, of women with normal glucose tolerance. This factor maysuch as superoxide dismutase to the culture medium of contribute to the higher rate of shoulder dystocia andrat embryos protects against growth delay and excess birth trauma observed in these infants.104malformations.98 It has been suggested that excess free The results of several clinical series have validated theoxygen radicals may have a direct effect on embry- Pedersen hypothesis inasmuch as tight maternal glyce-onic prostaglandin biosynthesis. Free oxygen radical mic control has been associated with a decline in theexcess may enhance lipid peroxidation, and in turn, incidence of macrosomia. In a series of 260 insulin-generated hydroperoxides might stimulate thrombox- dependent women achieving fasting plasma glucose con-ane biosynthesis and inhibit prostacyclin production, an centrations between 109 and 140 mg/dl, Gabbe et al.105imbalance that could have profound effects on embryonic observed 58 (22 percent) macrosomic infants. Kitzmillerdevelopment.90 and Cloherty106 reported that 11 percent of 134 women achieving fasting glucose concentrations between 105 and 121 mg/dl were delivered of an infant with a birthFetal Macrosomia weight in excess of 4,000 g. A more dramatic reduc- tion in the rate of macrosomia has been reported when Macrosomia has been variously deﬁned as birth weight more physiologic control has been achieved. Roversigreater than 4,000 to 4,500 g as well as large for ges- and Gargiulo107 instituted a program of “maximally tol-tational age, in which birth weight is above the 90th erated” insulin administration and observed macrosomiapercentile for population and sex-speciﬁc growth curves. in only 6 percent of cases. Jovanovic and coworkers108Fetal macrosomia complicates as many as 50 percent eliminated macrosomia in 52 women who achieved meanof pregnancies in women with GDM and 40 percent of glucose level of 80 to 87 mg/dl throughout gestation.pregnancies complicated by type 1 and type 2 diabe- Landon and colleagues,109 using daily capillary glucosetes, including some women treated with intensive glyce- values obtained during the second and third trimestermic control (Fig. 37-12). Delivery of an infant weighing in women requiring insulin, reported a rate of 9 percentgreater than 4,500 g occurs 10 times more often in women macrosomia when mean values were below 110 mg/dlwith diabetes as compared with a population of women compared with 34 percent when less optimal control waswith normal glucose tolerance.99 achieved. Jovanovic et al.110 have suggested that 1-hour According to the Pedersen hypothesis, maternal hyper- postprandial glucose measurements correlate best withglycemia results in fetal hyperglycemia and hyperinsu- the frequency of macrosomia. After controlling for other
988 Section VI Pregnancy and Coexisting Diseasefactors, these authors noted that the strongest prediction Table 37-3. Neonatal Body Compositionfor birth weight was third-trimester nonfasting glucosemeasurements. GDM (n = 195) NGT (n = 220) P value In a series of metabolic studies, Catalano et al.111 esti- Weight (g) 3,398 ± 550 3,337 ± 549 0.26mated body composition in 186 neonates using anthro- FFM (g) 2,962 ± 405 2,975 ± 408 .74pometry. Fat-free mass, which comprised 86 percent of Fat mass (g) 436 ± 206 362 ± 198 .0002mean birth weight, accounted for 83 percent of the vari- Body fat 12.4 ± 4.6 10.4 ± 4.6 .0001ance in birth weight, and fat mass, which comprised only FFM, fat-free mass; GDM, gestational diabetes mellitus; NGT,14 percent of birth weight, accounted for 46 percent of normal glucose tolerance.the variance in birth weight. There was also signiﬁcantlygreater fat-free mass in male as compared with femaleinfants. Using independent variables such as maternal these children at age 1 to 9 years and in adolescents agedheight, pregravid weight, weight gain during pregnancy, 14 to 16 years. Silverman and colleagues115 have reportedparity, paternal height and weight, neonatal sex and ges- that there is a strong correlation between amniotic ﬂuidtational age, the authors accounted for 29 percent of the insulin levels and increased body mass index (wt/ht2)variance in birth weight, 30 percent of the variance in in 14- to 17-year-old children, indicating an associationfat-free mass and 17 percent of the variance in fat mass.112 between islet cell activation in utero and developmentIncluding estimates of maternal insulin sensitivity in 16 of childhood obesity. This obesity present in childhoodadditional subjects, they were able to explain 48 percent then predisposes to obesity in the adult. Pettitt and col-of the variance in birth weight, 53 percent in fat-free mass leagues116 have shown that infants born to Pima Indianand 46 percent in fat mass.113 Studies by Caruso et al.114 women with impaired glucose tolerance were more obesehave corroborated these ﬁndings, reporting that women as children than infants of women with normal glucosewith unexplained fetal growth restriction had greater tolerance, even when they developed diabetes later in life.insulin sensitivity as compared with a control group of These data suggest that there are both in utero maternalwomen whose infants were appropriate weight for gesta- metabolic factors as well as genetic factors in the latertional age. The potential mechanisms for this relate to the development of type 2 diabetes and obesity.possibility that maternal circulating nutrients for glucose,free fatty acids, and amino acids available for placentaltransport to the fetus are decreased because of the relative Hypoglycemiaincrease in maternal insulin sensitivity. A positive cor-relation between birth weight and weight gain has been Neonatal hypoglycemia, a blood glucose less than 35observed in women with normal glucose tolerance. The to 40 mg/dl during the ﬁrst 12 hours of life, results fromcorrelation was strongest in women who were lean before a rapid drop in plasma glucose concentrations followingconception and became progressively weaker as pregravid clamping of the umbilical cord. Hypoglycemia is par-weight for height increased.77 In women with GDM, there ticularly common in macrosomic newborns, in whichwere no signiﬁcant correlations between maternal weight rates exceed 50 percent. With near-physiologic controlgain and birth weight, irrespective of pregravid weight for of maternal glucose levels during pregnancy, overall ratesheight. These studies emphasize the role of the maternal of 5 to 15 percent have been reported.108,109 The degree ofmetabolic environment and fetal growth. hypoglycemia may be inﬂuenced by at least two factors: Normalization of birth weight in infants of women (1) maternal glucose control during the latter half ofwith GDM, however, may in itself not achieve optimal pregnancy, and (2) control of maternal glycemia controlgrowth. In a study of approximately 400 infants of women during labor and delivery.117 Prior poor maternal glucosewith normal glucose tolerance and GDM, Catalano control can result in fetal β-cell hyperplasia, leading toet al.,79 showed that the infants of women with GDM had exaggerated insulin release following delivery. IDMsincreased fat mass but not lean body mass or weight as exhibiting hypoglycemia have elevated cord C-peptidecompared with a control group even after adjustment for and free insulin levels at birth and an exaggerated pan-potential confounding variables (Table 37-3). Similarly, creatic response to glucose loading.118when only infants who were appropriate-for-gestationalage (i.e., between the 10th and 80th percentiles) wereexamined, the infants of the women with GDM had sig- Respiratory Distress Syndromeniﬁcantly greater fat mass and percent body fat but hadless lean mass as compared with the control group but no The precise mechanism by which maternal diabetesdifference in birth weight. Of note, in the infants of the effect pulmonary development remains unknown. Exper-women with GDM, the strongest correlates with fat mass imental animal studies have focused primarily on thewere fasting glucose and gestational age. This accounted effects of hyperglycemia and hyperinsulinemia on pulmo-for 17 percent of the variance in infant fat mass. nary surfactant biosynthesis. An extensive review of the In addition to the perinatal association with fetal mac- literature conﬁrms that both of these factors are involvedrosomia in the infants of women with abnormal glucose in delayed pulmonary maturation in the IDM.119tolerance, there are signiﬁcant long-term risks. The In vitro studies have documented that insulin canincrease in birth weight of these infants tends to nor- interfere with substrate availability for surfactant bio-malize by 1 year of age before increasing again in early synthesis.120,121 Smith121 has postulated that insulinchildhood.115 There is an increase in the risk of obesity in interferes with the normal timing of glucocorticoid-induced
Chapter 37 Diabetes Mellitus Complicating Pregnancy 989pulmonary maturation in the fetus. Cortisol apparently proposed to explain these clinical ﬁndings, the pathogen-acts on pulmonary ﬁbroblasts to induce synthesis of esis of hyperbilirubinemia remains uncertain. In the past,ﬁbroblast-pneumocyte factor, which then acts on type the jaundice observed in the IDM often was attributed toII cells to stimulate phospholipid synthesis.122 Carlson prematurity. Studies that have analyzed morbidity care-and coworkers123 demonstrated that insulin blocks fully, according to gestational age, however, have rejectedcortisol action at the level of the ﬁbroblast by reducing this concept.131the production of ﬁbroblast-pneumocyte factor. Although severe hyperbilirubinemia may be observed Clinical studies investigating the effect of maternal dia- independent of polycythemia, a common pathway forbetes on fetal lung maturation have produced conﬂicting these complications most likely involves increased reddata. The role of amniocentesis in determining fetal lung blood cell production, which is stimulated by increasedmaturity is discussed with timing and mode of delivery. erythropoietin in the IDM. Presumably, the majorWith the introduction of protocols that have empha- stimulus for red cell production is a state of relativesized glucose control and antepartum surveillance until hypoxia in utero, as described previously. Althoughlung maturity has been established, respiratory distress cord erythropoietin levels generally are normal in IDMssyndrome (RDS) has become a less common occurrence whose mothers demonstrate good glycemic controlin the IDM. Several studies agree that in well-controlled during gestation, Shannon and colleagues found thatdiabetic women delivered at term, the risk of RDS is no hemoglobin A1c (HbA1c) values in late pregnancy werehigher than that observed in the general population.124,125 signiﬁcantly elevated in mothers of hyperbilirubinemicKjos et al.125 studied the outcome of 526 diabetic gesta- infants.130,132tions delivered within 5 days of amniotic ﬂuid fetal lungmaturation testing and reported HMD in ﬁve neonates(0.95 percent), all of whom were delivered before 34 MATERNAL CLASSIFICATION ANDweeks’ gestation. Mimouni et al.126 compared outcomes RISK ASSESSMENTof 127 IDMs with matched controls and have concludedthat diabetes in pregnancy as currently managed is not a Priscilla White133 ﬁrst noted that the patient’s age atdirect risk factor for the development of RDS. Yet, cesar- onset of diabetes, the duration of the disease, and theean delivery not preceded by labor and prematurity, both presence of vasculopathy signiﬁcantly inﬂuenced perina-of which are increased in diabetic pregnancies, clearly tal outcome. Her pioneering work led to a classiﬁca-increase the likelihood of neonatal respiratory disease. tion system that has been widely applied to pregnantWith cesarean delivery, most of these cases represent women with diabetes.133 A modiﬁcation of this schemeretained lung ﬂuid or transient tachypnea of the newborn, is presented in Table 37-1. Counseling a patient andwhich usually resolves within the ﬁrst days of life. formulating a plan of management requires assessment of both maternal and fetal risk. The White classiﬁcation facilitates this evaluation.Calcium and Magnesium Metabolism Class A1 diabetes mellitus includes those patients who have demonstrated carbohydrate intolerance during a Neonatal hypocalcemia, with serum levels below 7 mg 100-g 3-hour oral glucose tolerance test (GTT); however,per dl, occurs at an increased rate in the IDM when their fasting and 2-hour postprandial glucose levels areone controls for predisposing factors such as prematu- maintained within physiologic range by dietary regulationrity and birth asphyxia.127 With modern management, alone. Class A2 includes gestational diabetic women whothe frequency of neonatal hypocalcemia is less than 5 require insulin or oral hypoglycemic therapy in responsepercent in the infants of diabetic women.127 Hypocal- to repetitive elevations of fasting or postpartum glucosecemia in the IDM has been associated with a failure levels following dietary intervention.to increase parathyroid hormone synthesis following The Second and Third International Workshop Confer-birth.128 Decreased serum magnesium levels have also ences on Gestational Diabetes sponsored by the Americanbeen documented in pregnant diabetic women as well as Diabetes Association in cooperation with the Americantheir infants. Mimouni et al.128 described reduced amni- College of Obstetricians and Gynecologists (ACOG)otic ﬂuid magnesium concentrations in women with type recommended that the term gestational diabetes rather1 diabetes mellitus. These ﬁndings may be explained by a than Class A diabetes be used to describe women withdrop in fetal urinary magnesium excretion, which would carbohydrate intolerance of variable severity with onsetaccompany a relative magnesium deﬁcient state. Mag- or recognition during the present pregnancy.134,135 Thenesium deﬁciency paradoxically then may inhibit fetal deﬁnition applies whether insulin or only diet modi-parathyroid hormone secretion. ﬁcation is used for treatment and whether or not the condition persists after pregnancy. It does not exclude the possibility that unrecognized glucose intoleranceHyperbilirubinemia and Polycythemia may have antedates or begun with pregnancy. The term gestational diabetes fails to specify whether the patient Hyperbilirubinemia is frequently observed in the IDM. requires dietary adjustment alone or treatment with dietNeonatal jaundice has been reported in as many as 25 to and insulin. This distinction is important because those53 percent of pregnancies complicated by pregestational patients who are normoglycemic while fasting appeardiabetes mellitus and 38 percent of pregnancies with to have a signiﬁcantly lower perinatal mortality rate.136GDM.127,129,130 Although several mechanisms have been Women with GDM who require insulin are at greater risk
990 Section VI Pregnancy and Coexisting Diseasefor a poor perinatal outcome than those whose diabetes antihypertensive therapy consisting of either angiotensinis controlled by diet alone. converting enzyme inhibitors or angiotensin II receptor Patients requiring insulin are designated by the letters B, blockers are indicated prior to pregnancy in womenC, D, R, F, and T. Class B patients are those whose onset with microalbuminuria or overt nephropathy.140–143 Theseof disease occurs after age 20. They have had diabetes for agents should, however, be discontinued once pregnancyless than 10 years and have no vascular complications. is established because they are associated with fetal proxi-Included in this subgroup of patients are those who have mal tubular dysgenesis and oligohydramnios. It is impor-been previously treated with oral hypoglycemic agents. tant to note that women who exhibit microalbuminuria Class C diabetes includes patients who have the onset in early pregnancy have a 35- to 60-percent risk forof their disease between the ages of 10 and 19 or have superimposed preeclampsia.144,145had the disease for 10 to 19 years. Vascular disease is Women with diabetic nephropathy have a signiﬁcantlynot present. reduced life expectancy. Disease progression is character- Class D represents women whose disease is of 20 years ized by hypertension, declining glomerular ﬁltration rateduration or more, or whose onset occurred before age and eventual end-stage renal disease requiring dialysis10, or who have benign retinopathy. The latter includes or transplantation. In women with overt nephropathy,microaneurysms, exudates, and venous dilation. end-stage renal disease occurs in 50 percent by 10 years and in greater than 75 percent of cases by 20 years. Class F describes the 5 to 10 percent of pregnant patientsNephropathy with underlying renal disease. This includes those with reduced creatinine clearance and/or proteinuria of at least Renal disease develops in 25 to 30 percent of women 400 mg in 24 hours measured during the ﬁrst twentywith insulin-dependent diabetes mellitus with a peak weeks of gestation. Two factors present before 20 weeks’incidence after approximately 16 years of diabetes.137 gestation appear to be predictive of perinatal outcome inOvert diabetic nephropathy is diagnosed in women with these women (e.g., preterm delivery, low birthweight, orboth type 1 or type 2 diabetes mellitus when persistent preeclampsia). These areproteinuria exists in the absence of infection or other 1. Proteinuria greater than 3.0 g/24 h.urinary tract disease. The criteria for diagnosis in the 2. Serum creatinine greater than 1.5 mg/dl.nonpregnancy state including a total urinary protein In our series of 45 class F women, 12 women had suchexcretion (TPE) of greater than 500 mg/24 h or greater risk factors.146 Preeclampsia developed in 92 percent withthan 300 mg/24 h of urinary albumin excretion (UAE). a mean gestational age at delivery of 34 weeks com- Before the development of overt diabetic nephropathy, pared with an incidence of preeclampsia of 36 percentsome individuals develop incipient diabetic nephropathy in 33 women without these risk factors who reached andeﬁned by repetitive increases in UAE known as micro- average gestational age of 36 weeks. Remarkably, perina-albuminuria. The diagnosis is established from a 24-hour tal survival was 100 percent in this series, and no deliv-urine collection exhibiting UAE of 20 to 199 µgm/min eries occurred prior to 30 weeks’ gestation. Comparableor 30 to 299 mg/24 h. Without speciﬁc interventions, series detailing perinatal outcomes in Class F patients areapproximately 80 percent of individuals with type 1 dia- presented in Table 37-4.betes who develop sustained microalbuminuria will have The management of the diabetic women with nephro-their UAE increase at 10 to 20 percent/year to the stage pathy requires great expertise. Limitation of dietaryof overt nephropathy. In the nonpregnant individual, protein, which may reduce protein excretion in nonpreg-improvement of glycemic and blood pressure control have nant patients, has not been adequately studied duringbeen demonstrated to reduce the risk or slow the pro- pregnancy. Although the method is controversial, somegression of diabetic nephropathy.138,139 Renoprotective or nephrologists recommend a modiﬁed reduction in protein Table 37-4. Comparative Studies of Outcomes in Class F Diabetes Mellitus KITZMILLER147 GRENFEL276 REECE148 GORDON146 ROSENN150 (n) 26 20 31 45 61 Chronic HTN 31% 27% 22% 26% 47% Initial Creat >1.9 mg/dl 38% 10% 22% 11% — Initial proteinuria >3.0 gm/24 hr 8.3% — 22% 13% — Preeclampsia 15% 55% 35% 53% 51% Cesarean delivery — 72% 70% 80% 82% Perinatal survival (%) 88.9 100 93.5 100 94% Major anomalies 3 (11.1%) 1 (4.3%) 3 (9.7%) 2 (4%) 4 (6%) IUGR (%) 20.8 NA 19.4 11.0 11% Delivery <34 weeks (%) 30.8 27 22.5 15.5 25% 34–36 weeks (%) 40.7 23 32.3 35.5 28% >36 weeks (%) 28.5 50 45.2 49 47%Creat, creatinine; HTN, hypertension; IUGR, intrauterine growth restriction.
Chapter 37 Diabetes Mellitus Complicating Pregnancy 991intake for pregnant women with nephropathy. Control of ment and newer immunosuppressive regimens. Earlierhypertension in pregnant women with diabetic nephropa- reports of pregnancies after renal transplantation docu-thy is crucial to prevent further deterioration of kidney mented a signiﬁcant risk for perinatal loss (22 percent)function and to optimize pregnancy outcome. Although and preeclampsia (67 percent).153debatable, some cautiously use diuretics when patients Presently, many transplant centers strive to performare extremely nephrotic because this group may be prone combined kidney-pancreas transplants in diabetic indi-to volume-dependent forms of hypertension. Again, ACE viduals with end stage renal disease. McCrory and col-inhibitors and receptor blockers, which reduce intraglo- leagues154 reported 23 pregnancies following combinedmerular pressure and improve proteinuria in nonpreg- transplantation in which 25 percent of women developednant diabetic patients, should be discontinued during preeclampsia and 70 percent of pregnancies resulted inpregnancy. preterm birth. A report from the International Pancreas Several studies have failed to demonstrate a perma- Transplant Registry of 19 pregnancies after combinednent worsening of diabetic renal disease as a result of transplantation documented a 100-percent live-birth rate.pregnancy.146–148 Furthermore, it has been suggested that One woman each developed pancreas graft rejection andpregnancy itself does not increase the risk of develop- kidney graft rejection following pregnancy.155ing nephropathy, although development of proteinuriaand poor glycemic control are markers for subsequentrenal disease. Kitzmiller and colleagues147 reviewed 35 Retinopathypregnancies complicated by diabetic nephropathy. Pro-teinuria increased in 69 percent and hypertension devel- Class R diabetes designates patients with proliferativeoped in 73 percent. After delivery, proteinuria declined retinopathy, representing neovascularization or growthin 65 percent of cases. In only two patients did protein of new retinal capillaries. These vessels may cause vitre-excretion increase after gestation. In Gordon’s series, 26 ous hemorrhage with scarring and retinal detachment,women (58 percent) had more than a 1-g increase in resulting in vision loss. As with nephropathy, prevalenceproteinuria, and by the third trimester, 25 (56 percent) of retinal disease is highly related to the duration of dia-excreted more than 3.0 g/24 h.146 In the vast majority of betes. At 20 years, nearly 80 percent of diabetic individu-cases, protein excretion returned to baseline levels fol- als have some element of diabetic retinopathy. Excellentlowing gestation. glycemic control prevents retinopathy and may slow its Changes in creatinine clearance during pregnancy are progression. Parity is not associated with a risk for sub-variable in class F patients. Kitzmiller,149 in reviewing 44 sequent retinopathy.156 However, pregnancy does conveypatients from the literature, noted that about one third a greater than a twofold independent risk for progressionof women had an expected rise in creatinine clearance of existing retinopathy.157 Progression of diabetic reti-during gestation, compared with one third who had a nopathy during pregnancy is associated with hypertensivedecline of more than 15 percent by the third trimester. disease.158 Retinopathy may worsen signiﬁcantly duringIn Gordon’s series, 12 of 16 women in this category pregnancy in spite of the major advances that have beendeveloped preeclampsia. Of interest, most patients with made in diagnosis and treatment of existing retinopathy.a severe reduction in creatinine clearance (<50 ml/min) in Ideally, women planning a pregnancy should have a com-the ﬁrst trimester do not demonstrate a further reduction prehensive eye examination and treatment before concep-in clearance during pregnancy.146 However, a decline in tion. For those discovered to have proliferative changesrenal function can be anticipated in 20 to 30 percent of during pregnancy, laser photocoagulation therapy withcases. Several authors have suggested that any deteriora- careful follow-up has helped maintain many pregnanciestion of renal function after pregnancy is probably consis- to a gestational age at which neonatal survival is likely.tent with the natural course of diabetic nephropathy and In a large series of 172 patients, including 40 casesis not related to pregnancy per se.150 This ﬁnding has been with background retinopathy and 11 with proliferativeconﬁrmed by Rossing and colleagues151 who conducted changes, only one patient developed new onset prolif-an observational study of women who developed diabetic erative retinopathy during pregnancy.159 A review of thenephropathy between 1984 and 1989 who were followed literature by Kitzmiller160 conﬁrms the observation thatto 2000. Over the 10 years postpartum, the mean decline progression to proliferative retinopathy during pregnancyin creatinine clearance was 3.2 + 3.4 ml/min/year in 17 rarely occurs in women with background retinopathy orparous women compared with 3.2 + 5.1 ml/min/year in those without any eye ground changes. Of the 561 women42 women who never became pregnant. in these two categories, only 17 (3.0 percent) developed With improved survival of diabetic patients after renal neovascularization during gestation.160 In contrast, 23transplantation, a growing number of kidney recipients of 26 (88.5 percent) with untreated proliferative diseasehave now achieved pregnancy (class T). Armenti of the experienced worsening retinopathy during pregnancy.United States National Transplant Registry has described Pregnancy may increase the prevalence of some back-28 pregnancies in diabetic renal transplant recipients.152 ground retinal changes.161 Characteristic streak-blob hem-The majority of patients had underlying hypertension, orrhages and soft exudates have been noted, and suchalthough preeclampsia was diagnosed in only 17 percent retinopathy may progress despite strict metabolic control.of cases. Allograft rejection occurred in one case. Overall, At least two studies have related worsening retinal diseasedespite an increase in deliveries before 37 weeks, perina- to plasma glucose at the ﬁrst prenatal visit as well astal survival was 100 percent. These excellent results have the magnitude of improvement in glycemia during earlyresulted from improvements in both perinatal manage- pregnancy.162,163 In a subset of 140 women without pro-
992 Section VI Pregnancy and Coexisting Diseaseliferative retinopathy at baseline followed in the Diabe- prior to 1980.157 A high index of suspicion for isch-tes in Early Pregnancy Study, progression of retinopathy emic heart disease should be maintained in women withwas seen in 10.3, 21.1, 18.8, and 54.8 percent of patients long-standing diabetes because anginal symptoms may bewith no retinopathy, microaneurysms only, mild non- minimal and infarction may thus present as congestiveproliferative retinopathy, and moderate-to-severe nonpro- heart failure.166 Although there are in excess of one dozenliferative retinopathy at baseline, respectively. Elevated reports of successful pregnancies following myocardialglycosylated hemoglobin at baseline and the magnitude infarction in diabetic women, cardiac status should beof improvement of glucose control through week 14 were carefully assessed early in gestation or preferably beforeassociated with a higher risk of progression of retinopa- pregnancy. If electrocardiographic abnormalities arethy.163 Women with an initial glycohemoglobin greater encountered, echocardiography may be employed tothan 6 SD above the control mean were nearly three times assess ventricular function or modiﬁed stress testing mayas likely to experience worsening retinopathy compared be performed. The decision to undertake a pregnancy in awith those within 2 SD of the mean. Whether improving woman with type 1 or type 2 diabetes mellitus and coro-control or simply suboptimal control itself contributes nary artery disease needs to be made only after seriousto a deterioration of background retinopathy remains consideration. The potential for morbidity and mortalityuncertain. Hypertension may also be a signiﬁcant risk must be thoroughly reviewed with the patient and herfactor for the progression of retinopathy during preg- family. The management of myocardial infarction duringnancy.158 Rosenn and colleagues158 reported that wors- pregnancy is discussed in Chapter 34.ening of retinopathy occurred in 55 percent of womenwith a hypertensive disorder of pregnancy compared to25 percent of women without chronic hypertension or DETECTION AND SIGNIFICANCE OFpreeclampsia. Hypertension was associated with the pro- GESTATIONAL DIABETES MELLITUSgression of retinopathy after controlling for changes inglycemic status. It has been estimated that 2 to 3 percent of pregnan- For women with proliferative changes, laser photo- cies are complicated by diabetes mellitus and that ninetycoagulation is indicated and most will respond to this percent of the cases represent women with GDM.167therapy. However, those women who demonstrate severe An increased prevalence of GDM is found in womenﬂorid disc neovascularization that is unresponsive to laser of ethnic groups which have high frequencies of typetherapy during early pregnancy may be at great risk for 2 diabetes. These include women of Hispanic, African,deterioration of their vision. Termination of pregnancy Native American, Asian, and/or Paciﬁc Island ancestry.168should be considered in this group of patients. Women with GDM represent a group with signiﬁcant risk In addition to background and proliferative eye disease, for developing glucose intolerance later in life. Whereas,vaso-occlusive lesions associated with the development of O’Sullivan projected that 50 percent of GDM wouldmacular edema have been described during pregnancy.164 become diabetic in follow-up study of 22 to 28 years,Cystic macular edema is most often found in patients Kjos and colleagues have reported that 60 percent ofwith proteinuric nephropathy and hypertensive disease Latina women will develop type 2 diabetes, and this levelleading to retinal edema. Macular capillary permeability of risk may actually be manifest by 5 years after the GDMis a feature of this process. The degree of macular edema index pregnancy.169–171 The likelihood for subsequentis directly related to the fall in plasma oncotic pressure diabetes apparently increases when GDM is diagnosed inpresent in these women. In Sinclair’s series, seven women early pregnancy, and is accompanied by impaired β-cellwith minimal or no retinopathy before becoming preg- function and obesity.nant developed severe macular edema associated with As noted earlier, GDM is a state restricted to pregnantpre-proliferative or proliferative retinopathy during the women whose impaired glucose tolerance is discoveredcourse of their pregnancies. Although proliferation was during pregnancy. Because, in most cases, patients withcontrolled with photocoagulation, the macular edema GDM have normal fasting glucose levels, some challengeworsened until delivery in all cases and was often aggra- of glucose tolerance must be undertaken. Traditionally,vated by photocoagulation.164 Although both macular obstetricians relied upon historical and clinical risk factorsedema and retinopathy regressed after delivery in some to select those patients most likely to develop GDM. Thispatients, in others, these pathologic processes persisted, group included patients with a family history of diabe-resulting in signiﬁcant visual loss. tes, or those whose past pregnancies were marked by an unexplained stillbirth, or the delivery of a malformed or macrosomic infant. Obesity, hypertension, glycosuria,Coronary Artery Disease and maternal age older than 25 were other indications for screening. Interestingly, more than half of all patients Class H diabetes refers to the presence of diabetes of who exhibit an abnormal GTT lack the risk factors men-any duration associated with ischemic myocardial disease. tioned earlier. Coustan and colleagues172 have reportedThere is concern that the small number of women who that in a series of 6,214 women, using historic risk factorshave coronary artery disease are at an increased risk for and an arbitrary age cutoff of 30 years for screeningmortality during gestation. This is particularly true for would miss 35 percent of cases of GDM.women who suffer an infarction during pregnancy.165 In the summary and recommendations of the SecondThe maternal mortality rate exceeded 50 percent for and Third International Workshop-Conference oncases of infarction during pregnancy in cases reported GDM, screening was recommended for all pregnant
Chapter 37 Diabetes Mellitus Complicating Pregnancy 993 Table 37-5. Screening Strategy for Detecting GDM for the diagnosis of GDM failed to evaluate an associa- tion between mild carbohydrate tolerance and perinatal Risk assessment for GDM should be ascertained at the ﬁrst outcome has led many to question the overall signiﬁcance prenatal visit. of this diagnosis. Because of the lack of high-quality Low risk evidence concerning the beneﬁt of treatment of milder Blood glucose testing is not routinely required if all of cases of GDM, it is difﬁcult to determine the extent to the following characteristics are present: which screening impacts maternal and neonatal health Member of an ethnic group with a low prevalence outcomes.177 It has been suggested that the criteria for of GDM the diagnosis of GDM are conceptually ﬂawed in that No known diabetes in ﬁrst-degree relatives Age <25 years they represent a dichotomous deﬁnition of normal and Weight normal before pregnancy abnormal gestational glucose tolerance, when the risk No history of abnormal glucose metabolism of adverse maternal-fetal outcomes and later diabetes No history of poor obstetric outcome should be logically graded upward with higher values on the oral GTT and with the degree of fasting hyper- Average risk Perform blood glucose screening at 24–28 weeks using glycemia.178 Two studies have in fact addressed the rela- one of the following: tionship between mild degrees of carbohydrate tolerance Two-step procedure: 50 g GCT followed by a and rates of neonatal macrosomia. In a study of 3,637 diagnostic OGTT in those meeting the threshold women without GDM, Sermer and colleagues179 demon- value in the GCT strated a graded increase in adverse outcomes (includ- One-step procedure: diagnostic OGTT performed ing large infants) with increasing maternal carbohydrate on all subjects intolerance. Similarly, Sacks identiﬁed fasting and 2-hour High risk glucose values as independent risk factors for macroso- Perform blood glucose testing as soon as feasible, mia in a multivariate analysis of more than 3,500 preg- using the procedures described above. nant women. However, because no clinically meaningful If GDM is not diagnosed, blood glucose testing glucose threshold could be identiﬁed, Sacks180 concluded should be repeated at 24–28 weeks or at any time that the criteria for GDM will likely be established by a patient has symptoms or signs suggestive of consensus. hyperglycemia Both a recent large-scale retrospective cohort study andGCT, glucose challenge test; GDM, gestational diabetes mellitus; a randomized clinical trial have both suggested a beneﬁtOGTT, oral glucose tolerance test. to treatment of GDM in reducing the risk of perinatal Adapted from Fourth International Workshop Conference on complications. Langer and colleagues181 compared peri-GDM, Diabetes Care, Volume 21, Supplement 2, August 1998. natal outcomes in 555 GDM women diagnosed after 37 weeks with 1,110 subjects treated for GDM as well as 1,110 subjects matched for demographic features andwomen. Following the Fourth International Workshop- gestational age at delivery. A single feature of a compositeConference in 1997, universal screening was recom- adverse outcome (stillbirth, macrosomia, neonatal hypo-mended for women in ethnic groups with relatively high glycemia, erythrocytosis, and hyperbilirubinemia) wasrates of carbohydrate intolerance during pregnancy and present in 50 percent of untreated subjects, 18 percent ofof diabetes later in life.173 It was recognized that certain treated subjects, and 11 percent of nondiabetic subjects.features place women at low risk for GDM (Table 37-5), These signiﬁcant differences were present when untreatedand it may not be cost-effective to screen this subgroup GDM were stratiﬁed according to disease severity (fastingof women. Those at low risk include women who are not plasma glucose level on the diagnostic oral GTT).members of ethnic groups at increased risk for develop- Crowther and colleagues have conducted the only largeing adult-onset diabetes, who have no previous history of scale multicenter randomized trial to determine whetherabnormal glucose tolerance or poor obstetric outcomes treatment of GDM improves perinatal outcome.182 Overusually associated with GDM and who have all of the a 10-year period, 1,000 women were randomly assignedfollowing characteristics: age younger than 25 years, between 24 to 34 weeks’ gestation to receive dietarynormal body weight, and no family history of diabetes. advice, blood glucose monitoring, and insulin therapyDanilenko-Dixon and colleagues have reported that such as needed versus routine care. The authors found thelow risk women represent only 10 percent of their popu- composite rate of serious complications was lower inlation and this identifying such women may add com- the intervention group than in the routine care group (1plexity to screening in a busy clinic or ofﬁce practice.174 versus 4 percent). However, none of the individual com-The ACOG recognizes that while universal screening is ponents of the composite reached statistical signiﬁcance.the most sensitive approach, selective screening may be Among the components of the composite, only the rate ofemployed with the criteria cited earlier.175 shoulder dystocia (16 versus 7 cases) was signiﬁcantly dif- Despite the widespread acceptance of screening for and ferent among comparison groups. Thus, the differencestreating GDM in the United States, some expert panels in serious perinatal complications reported in this studyhave questioned the beneﬁt of GDM screening programs should be interpreted with caution.altogether.176 The criteria for the diagnosis of GDM origi- An international blinded observational study is cur-nally designated a population at increased risk for the rently in progress to assess the association betweendevelopment of type 2 diabetes in later life. The fact that maternal hyperglycemia and perinatal outcomes as wellO’Sullivan’s original work establishing the criteria used as a Maternal-Fetal Medicine Networks Unit random-
994 Section VI Pregnancy and Coexisting Diseaseized trial addressing the efﬁcacy of treatment of mild on GDM. Several studies have conﬁrmed that patientsGDM.183,184 Together, these studies should help guide diagnosed using the less stringent Carpenter criteria expe-thresholds for intervention in women with carbohydrate rience as much perinatal morbidity (macrosomia andintolerance during pregnancy.185 cesarean delivery) as subjects diagnosed by the National The 50-g glucose challenge may be performed in the Diabetes Data Group Criteria.191–193 Using either criteria,fasting or fed state. Sensitivity is improved if the test the patient must have at least two abnormal glucoseis performed in the fasting state.186,187 A plasma value determinations to be diagnosed with GDM.between 130–140 mg/dl is commonly used as a thresh-old for performing a 3-hour oral GTT. Coustan et al.186have demonstrated that 10 percent of GDM women have Treatment of the Patient with Type 1 orscreening test values between 130 and 139 mg/dl. This Type 2 Diabetes Mellitusstudy indicated that the sensitivity of screening wouldbe increased from 90 percent to nearly 100 percent if Because fetal glucose levels reﬂect those of the mother,universal screening were employed using a threshold of it is not surprising that clinical efforts aimed at optimiz-130 mg/dl. The prevalence of positive screening tests ing maternal control are considered the key componentrequiring further diagnostic testing increases from 14 responsible for the decline in perinatal death in preg-percent (140 mg/dl) to 23 percent (130 mg/dl), which is nancies complicated by diabetes mellitus over the lastaccompanied by an approximately 12-percent increase in few decades. Self–blood glucose monitoring combinedthe overall cost to diagnose each case of GDM. with intensive insulin therapy has resulted in improved Whereas most women can be screened for GDM at glycemia for many pregnant diabetic women (Table 37-approximately 24 to 28 weeks’ gestation, it is advis- 7). Women with pregestational diabetes should monitorable to screen earlier in pregnancy those with strong risk their glucose control 5 to 7 times daily using glucose-factors such as morbid obesity, a strong family history, oxidase impregnated reagent strips and a glucose reﬂec-previous GDM, prior macrosomic stillbirth, or an infant tance meter.194weighing more than 4,500 g.188 If initial screening is To achieve the best glycemic control possible for eachnegative, repeat testing is performed at 24 to 28 weeks. patient, during pregnancy conventional insulin therapyUsing the plasma cutoff of 135 to 140 mg/dl, one can is abandoned in favor of intensive therapy. An attemptexpect approximately 15 to 20 percent of patients with is made to stimulate physiologic insulin requirements byan abnormal screening value to have an abnormal 3-hour providing basal and prandial insulin needs through threeoral GTT. Patients whose 1-hour screening value exceeds or four daily injections or continuous insulin infusion190 mg/dl (10.5 mmol/L) will exhibit an abnormal oral (pump therapy) (CSII). Insulin regimens have classicallyGTT in 90 percent of cases.189 In women with a screen- included multiple injections of insulin usually prior toing value between 190 and 215 mg/dl, it is preferable to breakfast, the evening meal, and often bedtime compli-check a fasting blood glucose level before administering mented by self–blood glucose monitoring and adjustmenta 100-g carbohydrate load.190 If the fasting glucose is of insulin dose according to glucose proﬁles. Patients aregreater than 95 mg/dl, the patient is treated for GDM. instructed on dietary composition, insulin action, recog- The criteria for establishing the diagnosis of gesta- nition and treatment of hypoglycemia, adjusting insulintional diabetes are listed in Table 37-6. The U.S. National dosage for exercise and sick days, as well as monitoringDiabetes Data Group criteria represent a theoretic for hyperglycemia and potential ketosis. These princi-conversion of O’Sullivan’s thresholds in whole blood. ples form the foundation for intensive insulin therapy inCarpenter and Coustan189 prefer to use another modiﬁca- which an attempt is made to simulate physiologic insulintion of these data, which is supported by a comparison requirements. Insulin administration is provided for bothof the old Somogyi-Nelson method and current plasma basal needs and meals, and rapid adjustments are madeglucose oxidase assays. These criteria have been modi- in response to glucose measurements. The treatmentﬁed by the Fourth International Workshop Conference regimen generally involves three to four daily injections or the use of CSII devices. With either approach, frequent self–blood glucose monitoring is fundamental to achieve Table 37-6. Detection of Gestational Diabetes—Upper the therapeutic objective of physiologic glucose control. Limits of Normal Glucose determinations are made in the fasting state SCREENING TEST PLASMA (MG/DL) and before lunch, dinner, and bedtime. Postprandial and 50-G, 1-HOUR 130–140 nocturnal values are also helpful. Patients are instructed Oral GTT* NDDG192 Carpenter189 on an insulin dose for each meal and at bedtime, if nec- Fasting 105 95 1-hour 190 180 2-hour 165 155 Table 37-7. Target Plasma Glucose Levels in Pregnancy 3-hour 145 140 TIME MG/DLGTT, glucose tolerance test; NDDG, National Diabetes Data Group. *Diagnosis of gestational diabetes is made when any two values are Before breakfast 60–90met or exceeded. Before lunch, supper, bedtime snack 60–105 National Institutes of Health Diabetes Data Group: Classiﬁcation Two hours after meals ≤120and diagnosis of diabetes mellitus and other categories of glucose 2 a.m. to 6 a.m. >60intolerance. Diabetes 28:1039, 1979.
Chapter 37 Diabetes Mellitus Complicating Pregnancy 995essary. Mealtime insulin needs are determined by the in comparison with human regular insulin in 16 womencomposition of the meal, the premeal glucose measure- with GDM. Glycemic excursions were signiﬁcantly lowerment, and the level of activity anticipated following the with insulin aspart.meal. Basal or intermediate acting insulin requirements The long-acting insulin analogues glargine and detemirare determined by periodic 2-a.m. to 4-a.m. glucose mea- have been designed to more accurately mimic basal insulinsurements, as well as late afternoon values, which reﬂect secretion, yet neither has been adequately evaluated formorning intermediate-acting insulin action. During preg- safety or efﬁcacy during pregnancy. Insulin glargine has anancy, diabetic women should develop the self-man- ﬂat proﬁle when compared with NPH so that when admin-agement skills that are essential to an intensive insulin istered with short-acting insulin, unpredictable spikes intherapy regimen. insulin levels with resulting hypoglycemia appear to be In patients in whom diabetes is not well controlled, a less common.194,199 A concern with insulin glargine is itsbrief period of hospitalization is often necessary for the high afﬁnity for insulin-like growth factor (IGF) recep-initiation of therapy. Individual adjustments to the regi- tors. Theoretically, this might increase the progression ofmens implemented can then be made. It is gratifying for retinopathy in certain women. Further studies are neededmany patients to feel that they can take charge of their to establish both the safety and efﬁcacy of insulin glargineown diabetic control. Women who have previously fol- and detemir during pregnancy.lowed a prescribed dosage regimen for years gain conﬁ- Insulin is generally administered in two to three injec-dence in making adjustments in their insulin dosage after tions. We prefer a three-injection regimen, although mosta short period of time. Patients are encouraged to contact patients present taking a combination of intermediate-their physician at any time if questions should arise con- acting and short-acting insulin before dinner and break-cerning the management of their diabetes. During early fast. As a general rule, the amount of intermediate-actingpregnancy, patients are instructed to report their glucose insulin will exceed the short-acting component by a two-values by telephone, fax, or email at least weekly. to-one ratio. Patients usually receive two thirds of their Insulin therapy must be individualized, with dosage total dose with breakfast and the remaining third in thedeterminations tailored to diet and exercise. Semisynthetic evening as a combined dose with dinner or split intohuman insulin preparations and newer insulin analogues components with short-acting or rapid-acting insulin at(Table 37-8) are used during pregnancy. Insulin lispro dinner and intermediate-acting insulin at bedtime in anand insulin aspart are rapid-acting insulin preparations effort to minimize periods of nocturnal hypoglycemia.that have considerable advantages over regular insulin. These episodes frequently occur when the mother is in aInsulin lispro features reversal of proline and lysine at relative fasting state, whereas placental and fetal glucosepositions B28 and B29, and remains in monomeric form consumption continue. Finally, some women may requireand is thus rapidly absorbed. Its duration of action is a small dose of short-acting or rapid-acting insulin beforeshorter than that of regular insulin so that unexpected lunch, thus constituting a four-injection regimen.hypoglycemia hours after injection is avoided. Insulin Open-loop CSII pump therapy is preferred by manylispro appears to be safe for use during pregnancy, and IDDM women during pregnancy. The pump is a batteryit is a category B drug. An early report raised some ques- powered unit, which may be worn during most dailytion regarding a possible association with progression of activities like a beeper. These systems provide continu-retinopathy. However, recent experience suggests this is ous short-acting insulin therapy through a subcutane-not the case.195–197 ous infusion. The basal infusion rate and bolus doses to Insulin aspart has been studied in a limited number of cover meals are determined by frequent self-monitoringwomen with GDM during the third trimester.198 Pettit of blood glucose. The basal infusion rate is generally closeet al.198 assessed the short-term efﬁcacy of insulin aspart to 1 unit per hour. Table 37-8. Type of Human Insulin and Insulin Analogues SOURCE ONSET (H) PEAK (H) DURATION (H) Short acting Humulin R (Lilly) Human 0.5 2–4 5–7 Velosulin-H (Novo Nordisk) Human 0.5 1–3 8 Novolin R (Novo Nordisk) Human 0.5 2.5–5 6–8 Lispro Analogue 0.25 0.5–1.5 4–5 Aspart Analogue 0.25 1–3 3–5 Intermediate acting Humulin Lente (Lilly) Human 1–3 6–12 18–24 Humulin NPH (Lilly) Human 1–2 6–12 18–24 Novolin L (Novo Nordisk) Human 2.5 7–15 22 Novolin N (Novo Nordisk) Human 1.5 4–20 24 Long acting Humulin Ultralente (Lilly) Human 4–6 8–20 >36 Glargine Analogue 1 — 24 Determir Analogue 1–2 — 24
996 Section VI Pregnancy and Coexisting Disease Pregnant patients may require hospitalization before rated fatty acids, and the remainder derived from mono-initiation of pump therapy. Women must be educated saturated sources.203 Caloric intake is established basedregarding the strategy of continuous infusion and have on prepregnancy weight and weight gain during gesta-their glucose stabilized over several days. This requires tion. Weight reduction is not advised. Patients shouldthat multiple blood glucose determinations be made for consume approximately 35 Kcal/kg ideal body weight.the prevention of periods of hyperglycemia and hypo- Obese women may be managed with an intake as lowglycemia. Glucose values may become normalized with as 25 Kcal/kg actual weight. Any further caloric restric-minimal amplitude of daily excursions in most patients. tion resulting in ketonuria requires an increase in caloric Episodes of hypoglycemia are often reduced with pump consumption.therapy. When they do occur, they are usually second- The presence of maternal vasculopathy should be thor-ary to errors in dose selection or failure to adhere to the oughly assessed early in pregnancy. The patient shouldrequired diet. The risk of nocturnal hypoglycemia, which be evaluated by an ophthalmologist familiar with dia-is increased in the pregnant state, necessitates that great betic retinopathy. Ophthalmologic examinations are per-care be undertaken in selecting patients for CSII. Patients formed during each trimester and repeated more often ifusing the pump who fail to exhibit normal counterregu- retinopathy is defected. Baseline renal function is estab-latory responses to hypoglycemia should probably check lished by assaying a 24-hour urine collection for cre-their glucose values at 2 to 3 a.m. to detect nocturnal atinine clearance and protein. An electrocardiogram andhypoglycemia. urine culture are also obtained. The mechanics of the CSII systems are relatively simple. Most patients with type 1 and type 2 diabetes mellitusA ﬁne-gauge butterﬂy needle device is attached by con- are followed with outpatient visits at 1-to 2-week inter-necting tubing to the pump. This cannula is reimplanted vals. At each visit, control is assessed and adjustmentsevery 2 to 3 days at a different site in the anterior abdomi- in insulin dosage are made. However, patients should benal wall. Rapid-acting insulin (usually insulin lispro) is instructed to call at any time if periods of hypoglycemiastored in the pump syringe. Infusion occurs at a basal (<50 mg/dl) or hyperglycemia (>200 mg/dl) occur. Therate, which can be ﬁxed or altered for speciﬁc time of increased risk of hypoglycemia in pregnant individualsday by a computer program. For example, the basal rate may be related to defective glucose counterregulatorycan be programmed for a lower dose at night. Preprandial hormone mechanisms.204,205 Both epinephrine and gluca-boluses can be adjusted manually before each meal and gon appear to be suppressed in pregnant diabetic womensnack. Half of the total daily insulin is usually given as the during hypoglycemia. For these reasons, patients shouldbasal rate and the remainder as premeal boluses infused test glucose levels frequently, and family members shouldbefore each meal. The largest bolus (30 to 35 percent) be instructed on the technique of glucagon injection foris administered with breakfast, followed by 25 percent the treatment of severe reactions.before dinner and 15 to 20 percent before snacks. Patients without any pancreatic reserve may haverapid elevations of blood glucose if there is pump failure Ketoacidosisor intercurrent infection. Since the advent of bufferedinsulin, insulin aggregation leading to occlusion of the With the implementation of antenatal care, programssilastic infusion tubing is uncommon. Failure of the pump stressing strict metabolic control of blood glucose levelsis associated with a steady rise in ketonemia in the non- for women requiring insulin, DKA has fortunately becomepregnant patient. a less common occurrence. Kilvert and colleagues206 It is unclear whether CSII is superior to multiple injec- reported 11 cases of ketoacidosis in 635 insulin treatedtion regimens. Coustan and colleagues200 randomized 22 pregnancies between 1971 and 1990. One fetal loss andpatients to intensive conventional therapy with multiple one spontaneous miscarriage complicated the pregnan-injections versus pump therapy. There were no differ- cies affected by DKA.ences between the two treatment groups with respect to DKA can occur in the newly diagnosed diabetic patient,outpatient mean glucose levels, glycosylated hemoglobin and the hormonal milieu of pregnancy may become thelevels, or glycemic excursions. More recently, Gabbe and background for this phenomenon. Because pregnancy iscolleagues201 reported a large retrospective cohort study a state of relative insulin resistance marked by enhancedof women who began pump therapy during gestation lipolysis and ketogenesis, DKA may develop in a preg-as compared with a group treated with multiple injec- nant woman with glucose levels barely exceeding 200 mg/tion regimens. Women using pumps, most with insulin dl (11.1 mmol/L). Thus, DKA may be diagnosed duringlispro, had fewer hypoglycemic reactions and comparable pregnancy with minimal hyperglycemia accompanied byglucose control and pregnancy outcomes.201 Studies in a fall in plasma bicarbonate and a pH value less thannonpregnant individuals employing CSII comparing new 7.30. Serum acetone is positive at a 1 : 2 dilution.insulin analogues to regular insulin indicate fewer hypo- Early recognition of signs and symptoms of DKAglycemic episodes and improved postprandial control.202 improves both maternal and fetal outcome. As in the Diet therapy is critical to successful regulation of nonpregnant state, clinical signs of volume depletionmaternal diabetes. A program consisting of three meals follow the symptoms of hyperglycemia, which includeand several snacks is employed for most patients. Dietary polydipsia and polyuria. Malaise, headache, nausea, andcomposition should be 50 to 60 percent carbohydrate, 20 vomiting are common complaints. A pregnant womanpercent protein, and 25 to 30 percent fat with less than with poor ﬂuid intake and persistent vomiting over 8 to10 percent saturated fats, up to 10 percent polyunsatu- 12 hours should be evaluated for potential DKA. A low
Chapter 37 Diabetes Mellitus Complicating Pregnancy 997 Table 37-9. Management of Diabetic Ketoacidosis hypoxia. Thus, protocols for antepartum fetal assessment During Pregnancy in pregnancies complicated by diabetes mellitus have been incorporated into the care plan for outpatient monitoring 1. Laboratory assessment during the third trimester. During this time period, when Obtain arterial blood gases to document degree of the risk of sudden intrauterine death increases, a program acidosis present; measure glucose, ketones, of fetal surveillance is initiated. Because improvement electrolytes, at 1–2 h intervals in maternal control has played a major role in reducing 2. Insulin perinatal mortality in diabetic pregnancies, antepartum Low-dose, intravenous (IV) fetal monitoring tests are now used primarily to reas- Loading dose: 0.2–0.4 units/kg sure the obstetrician and avoid unnecessary premature Maintenance: 2.0–10.0 units/hr intervention. These techniques have few false-negatives 3. Fluids results, and in a patient in whom diabetes is well con- Isotonic Sodium Chloride trolled and who exhibits no vasculopathy or signiﬁcant Total replacement in ﬁrst 12 h = 4–6 L hypertension, reassuring antepartum testing allows the 1 L in ﬁrst hr fetus to beneﬁt from further maturation in utero. 500–1,000 mL/hr for 2–4 h Maternal assessment of fetal activity serves as a simple 250 ml/h until 80% replaced screening technique in a program of fetal surveillance. 4. Glucose During the third trimester, women are instructed to Begin 5% D/NS when plasma level reaches 250 mg/dl perform daily fetal movement counting. To date, few (14 mmol/L) studies have applied this method to a large number of 5. Potassium women with diabetes mellitus. Patients with a variety If initially normal or reduced, an infusion rate up to of high-risk antepartum conditions including diabetes 15–20 mEq/h may be required; if elevated, wait until appear to have an increased incidence of alarming fetal levels decline into the normal range, then add to IV activity patterns.207 Although the false-negative rate with solution in a concentration of 20–30 mEq/L maternal monitoring of fetal activity is low (∼1 percent), 6. Bicarbonate the false-positive rate may be as high as 60 percent. Add one ampule (44 mEq) to 1 L of 0.45 NS if pH is Maternal hypoglycemia, although generally believed to <7.10 be associated with decreased fetal movement, may actu- ally stimulate fetal activity.208D/NS, dextrose in normal saline. Sadovsky and coworkers209 reported that fetal move- ment at 25 to 33 weeks’ gestation in 67 diabetic preg-serum bicarbonate level prompts an arterial blood gas nancies was lower than in controls, whereas in the ﬁnaldetermination to rule out this diagnosis. Occasionally, two months of pregnancy, activity levels were similar toDKA may present in a woman with undiagnosed diabe- the nondiabetic population. In this series, there were fourtes receiving β-mimetic agents to arrest preterm labor. cases of cessation of fetal movement. Two fetuses diedBecause of the risk of hyperglycemia and DKA in women in utero 10 and 11 hours after maternal perception thatrequiring insulin who then receive intravenous β-mimetic activity had stopped, an interval shorter than that seenmedications such as a ritodrine, magnesium sulfate has in other complications of pregnancy.become the preferred tocolytic for cases of preterm labor The nonstress test (NST) remains the preferred methodin these cases. to assess antepartum fetal well-being in the patient with Once the diagnosis of DKA is established and the patient diabetes mellitus.210 If the NST is nonreactive, a bio-is stabilized, she should be transported to a facility where physical proﬁle (BPP) or contraction stress test is thentertiary care in both perinatology and neonatology is performed (Fig. 37-13). Heart rate monitoring is begunavailable. Therapy hinges on the meticulous correction of early in the third trimester, usually by 32 weeks’ gesta-metabolic and ﬂuid abnormalities. An attempt at treat- tion. Two studies have also demonstrated an increasedment of any underlying cause for DKA, such as infection, fetal death rate within 1 week of a reactive NST inshould be instituted as well. The general management of pregnancies complicated by IDDM when compared withDKA in pregnancy is outlined in Table 37-9. Fluid resus- other high-risk gestations.211,212 If the NST is to be usedcitation and insulin infusion should be maintained even as the primary method of antepartum heart rate testing,in the face of normoglycemia until bicarbonate levels we prefer that it be done at least twice weekly oncereturn to normal, indicating that acidemia has cleared. the patient reaches 32 weeks’ gestation. In patients withDKA does represent a substantial risk for fetal com- vascular disease or poor control, in whom the incidence ofpromise. However, successful fetal resuscitation often abnormal tests and intrauterine deaths is greater, testingaccompanies correction of maternal acidosis. Therefore, is often performed earlier and more frequently.every effort should be made to correct maternal condition Doppler umbilical artery velocimetry has been pro-before intervening and delivering a preterm infant. posed as a clinical tool for antepartum fetal surveillance in pregnancies at risk for placental vascular disease. We have found that Doppler studies of the umbilical arteryANTEPARTUM FETAL EVALUATION may be predictive of fetal outcome in diabetic pregnan- cies complicated by vascular disease.213 Elevated placental Maternal diabetes may result in fetal hyperglycemia and resistance as evidenced by an increased systolic/diastolichyperinsulinemia, and thereby increase the risk for fetal ratio is associated with fetal growth restriction and
998 Section VI Pregnancy and Coexisting Disease NONSTRESS TEST Table 37-10. Antepartum Fetal Surveillance in Low-Risk (Begin 28 to 32 weeks gestation) Insulin Dependent Diabetes Mellitus* STUDY Ultrasonography at 4–6 week Yes intervals Reactive Non-reactive Maternal assessment of fetal activity, Yes daily at 28 weeks Nonstress test (NST) weekly at Yes 28 weeks Twice weekly at 34 Contraction stress test weeks or biophysical profile Contraction stress test or biophysical Yes, if elective proﬁle if NST nonreactive; L/S, delivery planned lung proﬁle prior to 39 weeks Negative Positive *Low-risk IDDM: excellent control (160–120 mg/dl), no vasculopathy (classes B, C), no stillbirth. L/S, lecithin/sphingomyelin ratio. Follow-up Possible amniocentesis Table 37-11. Antepartum Fetal Surveillance in High-Risk testing (if premature) Insulin Dependent Diabetes Mellitus* or STUDY Delivery Ultrasonography at 4–6 week intervals Yes (if term or acute disease) Maternal assessment of fetal activity, Yes daily at 28 weeks Figure 37-13. Scheme for antepartum fetal testing. Nonstress test (NST) Minimum twice weekly Contraction stress test or biophysical Yes 213 proﬁle if NST nonreactive; L/S, lungpreeclampsia in these high-risk patients. In contrast, proﬁle at 37–38 weekspatients with well-controlled diabetes without vascu-lar disease rarely demonstrate abnormal fetal umbilical *High-risk IDDM: poor control (macrosomia, hydramnios),artery waveforms. vasculopathy (classes D, F, R), prior stillbirth. Johnstone and colleagues214 have reported the largest L/S, lecithin/sphingomyelin ratio.experience with serial Doppler umbilical waveforms inpregnancies complicated by diabetes. In their study of 128 diabetes is poorly controlled, who have hypertension, orwomen, signiﬁcant abnormal ﬂow patterns were observed who have signiﬁcant vasculopathy that may be associatedin nine cases. Three of these women had nephropathy, with fetal growth restriction are at increased risk for fetaland three had preeclampsia. All of these pregnancies had compromise and probably beneﬁt most from a programnormal outcomes. Importantly, several cases of fetal dis- of antepartum fetal surveillance.217tress as deﬁned by abnormal biophysical testing were Ultrasound is a valuable tool in evaluating fetalaccompanied by normal Doppler studies. Therefore, it growth, estimating fetal weight, and detecting hydram-appears that undue reliance should not be placed on nios and malformations. A determination of maternalnormal waveform values in the diabetic pregnancy.214 serum alpha-fetoprotein (MSAFP) at 16 weeks’ gestation It is important to include not only the results of ante- is often employed in association with a detailed ultra-partum fetal testing but to weigh all of the clinical fea- sound study during the midtrimester in an attempt totures involving mother and fetus before a decision is detect neural tube defects and other anomalies. Normalmade to intervene for suspected fetal distress, especially values of MSAFP for diabetic women are lower than inif this decision may result in a preterm delivery (Tables the nondiabetic population.218 A lower threshold for the37-10 and 37-11). In reviewing nine series involving 993 upper limit of normal, 1.5 multiples of the median, thusdiabetic patients, an abnormal test of fetal condition led may be preferable in pregnancies complicated by diabetesto delivery 5 percent of the time.215 It appears that out- mellitus in order to help detect spina biﬁda and otherpatient testing protocols work well in diabetic patients major malformations that are increased in this popula-requiring insulin. Whether such testing is required for all tion. A comprehensive ultrasound examination includingwomen with diabetes mellitus remains controversial.216 fetal echocardiography is performed at 20 to 22 weeks’In a study of 114 women with well-controlled IDDM, gestation for the investigation of possible cardiac anoma-10 patients were delivered for abnormal fetal testing. lies. Using such an approach, Greene and Benacerraf 219Eight of these 10 women had nephropathy or hyperten- detected 18 of 32 malformations in a series of 432 diabeticsion. Nephropathy or hypertension was associated with pregnancies. The speciﬁcity was in excess of 99 percent,intervention for fetal well-being in 8 of 20 women with and the negative predictive value was 97 percent. Spinathese risk factors in comparison with 2 of 94 without biﬁda was identiﬁed in all cases; however, ventricularthese complications. Thus, it appears that women whose septal defects, limb abnormalities, and facial clefts were
Chapter 37 Diabetes Mellitus Complicating Pregnancy 999missed. A review of the prenatal diagnosis experience in controls. These authors reported that the onset of PG289 women with IDDM in The Ohio State University production was delayed in GDM from 35.9 ± 1.1 weeksDiabetes in Pregnancy Program revealed 29 anomalies in to 37.3 ± 1.0 weeks and to 38.7 ± 0.9 weeks in preges-which 12 were cardiac, 14 were noncardiac, and 3 were tational diabetic pregnancies.223 In this study, delayedcombined.87 Twelve of ﬁfteen (80 percent) cardiac and appearance of PG was not associated with level of glyce-10 of 17 (59 percent) noncardiac lesions were identiﬁed mic control. It follows that the clinician must be familiarprenatally. When considering cardiac defects alone, we with the laboratory analysis of amniotic ﬂuid in his orcould not identify a glycosylated hemoglobin cutoff for her institution and the neonatal outcome associated withthese anomalies. Therefore, we believe detailed cardiac various tests of fetal lung maturity.imaging should be offered to all patients with type 1 When antepartum testing suggests fetal compromise,and type 2 diabetes mellitus to assist in the detection of delivery must be considered. If amniotic ﬂuid analysiscardiac lesions, especially those of the great vessels and yields a mature test result, delivery should be accom-cardiac septum. plished promptly. In the presence of presumed lung Ultrasound examinations should be performed during immaturity, the decision to proceed with delivery shouldthe third trimester to assess fetal growth. The detection be based on conﬁrmation of deteriorating fetal conditionof fetal macrosomia, the leading risk factor for shoul- by several abnormal tests. For example, if the NST asder dystocia, is important in the selection of patients well as the BPP indicates fetal compromise, delivery iswho are best delivered by cesarean section. An increased indicated. Finally, there are several maternal indicationsrate of cephalopelvic disproportion and shoulder dysto- for delivery including signiﬁcant preeclampsia, worsen-cia accompanied by signiﬁcant risk of traumatic birth ing renal function, or deteriorating vision secondary toinjury and asphyxia have been consistently associated proliferative retinopathy. If a patient reaches term gesta-with the vaginal delivery of large infants. The risk of tion with a mature fetal lung proﬁle and is at signiﬁcantsuch complications rises exponentially when birth weight risk for intrauterine demise because of poor control or aexceeds 4 kg and is greater for the fetus of a diabetic history of a prior stillbirth, delivery is planned.mother when compared with a fetus with similar weight Choosing the route of delivery for the diabetic patientwhose mother does not have diabetes.220 Sonographic remains controversial. Cesarean delivery rates as highmeasurements of the fetal abdominal circumference have as 50 percent are common in series of pregestationalproved most helpful in predicting fetal macrosomia.221 diabetic women. This ﬁgure is likely to represent theThe abdomen is likely to be large because of increased practice trends of most U.S. obstetricians and perinatolo-glycogen deposition in the fetal liver and subcutaneous gists.210 Delivery by cesarean section usually is favoredfat deposition. Using serial sonographic examinations, when fetal distress has been suggested by antepartumaccelerated abdominal growth can often be identiﬁed by heart rate monitoring.32 weeks’ gestation.222 The increased rate of shoulder dystocia and brachial plexus injury in the offspring of diabetic women has prompted adoption of early induction strategies as wellTIMING AND MODE OF DELIVERY as selection of patients for cesarean section based on ultrasound estimation of fetal size. Such approaches are Delivery should be delayed until fetal maturation has limited by the relative inaccuracy of ultrasound predic-taken place, provided that the patient’s diabetes is well tion of birth weight. Despite the limitations, Kjos andcontrolled and antepartum surveillance remains normal. colleagues demonstrated that induction at 38 weeks in aIn our practice, elective induction of labor is often planned population of women with GDM was associated with aat 38 1/2 to 40 weeks’ gestation in well-controlled patients lower frequency of large-for-gestational age infants andwithout vascular disease. Patients with vascular disease shoulder dystocia without an increased rate of cesareanare delivered before term only if hypertension worsens, delivery.224 This is in contrast to studies of induction inif signiﬁcant fetal growth restriction is present, or if bio- nondiabetic women in which suspected macrosomia isphysical testing mandates early delivery. Before elective apparently associated with an increased rate of cesareandelivery prior to 39 weeks’ gestation, an amniocentesis delivery. In a sophisticated decision tree analysis of costmay be performed to document fetal pulmonary maturity. effectiveness, Rouse and colleagues found that whereasThere is much evidence that tests of fetal lung maturity elective cesarean delivery for macrosomia to prevent per-have the same predictive value in diabetic pregnancies as manent brachial plexus injury was prohibitively expen-they do in the normal population.125 sive in the non-diabetic woman at a cost of several million The presence of the acidic phospholipid phosphati- dollars per permanent brachial plexus injury prevented,dylglycerol (PG) is the ﬁnal marker of fetal pulmonary 489 cesarean deliveries at a cost/avoided birth injury ofmaturation. Several authors have suggested that fetal $880,000 per case for those diabetic pregnancies with anhyperinsulinemia may be associated with delayed appear- estimated fetal size greater than 4,000 g seemed to be atance of PG and an increased incidence of RDS. Landon least tenable.225et al.109 have correlated the appearance of PG in amniotic Acker et al.226 have reported that the overall risk forﬂuid with maternal glycemic control during gestation. shoulder dystocia in the macrosomic IDM is greater thanRDS may occur in the IDM with a mature L/S ratio or for the large, normal infant. In their series of diabeticfetal lung maturity index but absent PG. Moore and women, the risk for shoulder dystocia with a fetal weightcolleagues223 compared PG production in amniotic ﬂuid greater than 4,000 g was approximately 30 percent.specimens from 295 diabetic women and 590 matched Somewhat less impressive, yet signiﬁcantly greater fre-
1000 Section VI Pregnancy and Coexisting Disease Table 37-12. Rate of Shoulder Dystocia Related to Birth Table 37-13. Insulin Management During Labor Weight and Diabetic Status and Delivery BIRTH WEIGHT WITHOUT DIABETES WITH DIABETES Usual dose of intermediate-acting insulin is given at (G) (%) (%) bedtime. Morning dose of insulin is withheld <4,000 0.1–1.1 0.6–3.7 Intravenous infusion of normal saline is begun 4,000–4,449 1.1–10.0 4.9–23.1 Once active labor begins or glucose levels fall below ≥4,500 4.1–22.6 20.0–50.0 70 mg/dl, the infusion is changed from saline to 5%Adapted from ACOG Practice Patterns No. 7. October 1997. dextrose and delivered at a rate of 2.5 mg/kg/min. Glucose levels are checked hourly using a portable reﬂectance meter allowing for adjustment in the infusion rate.quencies of shoulder dystocia for delivery of macrosomic Regular (short-acting) insulin is administered byIDMs versus non-IDMs have been reported by Nesbitt intravenous infusion if glucose levels exceed 140 mg/dl.and colleagues (Table 37-12).220 At present, ACOG rec-ommends consideration of cesarean delivery in diabetic From Jovanovic L, Peterson CM: Management of the pregnant,women when estimated fetal weight exceeds 4,500 g.227 insulin-dependent diabetic woman. Diabetes Care 3:63, 1980.Our approach is to consider elective cesarean when theestimated weight is 4,000 to 4,500 g, taking into consid-eration obstetric history and clinical pelvimetry. Despite glucose control and allows the neonatal team to prepareattempts to select patients with obvious fetal macroso- for the care of the newborn. The patient is given nothingmia for elective cesarean delivery, arrest of dilatation or by mouth, and her usual morning insulin dose is with-descent despite adequate labor should alert the physician held. If her surgery is not performed early in the day, oneto the possibility of cephalopelvic disproportion. About third to one half of the patient’s intermediate-acting dose25 percent of macrosomic infants (>4,000 g) delivered of insulin may be administered. Regional anesthesia isafter a prolonged second stage have shoulder dystocia.228 employed because an awake patient permits earlier detec-It follows that cesarean delivery should be considered in tion of hypoglycemia. Following surgery, glucose levelsa patient who demonstrates signiﬁcant protracted labor are monitored every 2 hours and an intravenous solutionor failure of descent. of 5 percent dextrose is administered. Following delivery, insulin requirements are usually signiﬁcantly lower than were pregnancy or prepregnancyGLUCOREGULATION DURING LABOR needs. The objective of tight control used in the antepar-AND DELIVERY tum period is relaxed for the ﬁrst 24 to 48 hours. Patients delivered vaginally, who are able to eat a regular diet, Because neonatal hypoglycemia is in part related to are given one third to one half of their end of pregnancymaternal glucose levels during labor, it is important dose of neutral protamine Hagedorn (NPH) insulin andto maintain maternal plasma glucose levels within the short-acting or rapid-acting insulin the morning of thephysiologic normal range. The patient is given nothing ﬁrst postpartum day. Frequent glucose determinationsby mouth after midnight of the evening before induc- are used to guide insulin dosage. Most patients are stabi-tion or elective cesarean delivery. The usual bedtime lized on this regimen within a few days after delivery.dose of insulin is administered, or for women receiving Women with diabetes are encouraged to breast-feed. Thepump therapy, the infusion is continued overnight. Upon additional 500 kcal required daily are given as approxi-arrival to the labor and delivery department, early in the mately 100 g of carbohydrate and 20 g of protein.230 Themorning, the patient’s capillary glucose level is assessed insulin dose may be somewhat lower in lactating diabeticwith a bedside reﬂectance meter. Continuous infusion women. Hypoglycemia appears to be common in the ﬁrstof both insulin and glucose are then administered based week following delivery and immediately after nursing.on maternal glucose levels (Table 37-13). Ten units ofshort acting insulin may be added to 1,000 ml of solu-tion containing 5 percent dextrose. An infusion rate of MANAGEMENT OF THE PATIENT WITH100 to 125 ml/h (1 unit/h), in most cases, will result in GESTATIONAL DIABETESgood glucose control. Insulin may also be infused from asyringe pump at a dose of 0.25 to 2.0 units/h and adjusted The mainstay of treatment of GDM is nutritional coun-to maintain normal glucose values. Glucose levels are seling and dietary intervention. The optimal diet shouldrecorded hourly, and the infusion rate is adjusted accord- provide caloric and nutrient needs to sustain pregnancyingly. Well-controlled patients are often euglycemic once without resulting in signiﬁcant postprandial hyperglyce-active labor begins and then require glucose at an infusion mia.231 Women with GDM generally do not need hos-rate of 2.5 mg/kg/min.229 It may be necessary to increase pitalization for dietary instruction and management.the insulin infusion during the second stage of labor in Once the diagnosis is established, patients are begun onresponse to hyperglycemia associated with increased cat- a dietary program of 2,000 to 2,500 kcal daily.231 Thisecholamine secretion. represents approximately 35 kcal/kg of present preg- When cesarean delivery is to be performed, it should be nancy weight. Jovanovic-Peterson and Peterson232 havescheduled for early morning. This simpliﬁes intrapartum noted that such a diet composed of 50 to 60 percent
Chapter 37 Diabetes Mellitus Complicating Pregnancy 1001carbohydrate will cause excessive weight gain and post- terns in women with GDM and non-GDM during anprandial hyperglycemia, and require insulin therapy in 50 oral GTT. Patients with a fasting plasma glucose lesspercent of patients. For this reason, several groups have than 95 mg/dl had signiﬁcantly greater insulin produc-studied the use of calorie restricted diets.233 Algert and tion than those with a glucose level of 95 mg/dl. To testcolleagues233 have reported that obese women with GDM whether the assignment to insulin therapy for a patientmay be managed on as little as 1,700 to 1,800 kcal/day with a fasting value of 95 mg/dl or higher was appropri-with less weight gain and no apparent reduction in fetal ate, Langer et al.242 compared rates of delivery of LGAsize. Magee et al.234 designed a study to evaluate strict infants among women with GDM grouped according tocaloric restriction as a treatment for obese subjects with their fasting plasma glucose levels and whether diet orGDM. They randomized patients to a 2,400 kcal/d diet diet and insulin were used. They found that patients withcompared with a 1,200 kcal/d group. Average glucose a fasting glucose between 96 and 105 mg/dl had a greaterlevels and fasting glucose were reduced in the hypocaloric incidence of LGA infants (28.6 percent) when receivinggroup. However, fasting glucose levels and postchallenge diet therapy alone versus obese women with GDM receiv-glucose levels were not signiﬁcantly different. Signiﬁcant ing both diet and insulin. In women with an initial fastingketonuria did develop in the calorie-restricted group, glucose between 95 and 104 mg/dl, 70 percent requiredwhich may have a detrimental effect on fetal neurologic insulin therapy to achieve optimal control.243development.235 Thus, the authors went on to study a Whereas Langer and coworkers240 have documented a1,800 kcal diet, which improved glycemia and did not relationship between maternal glycemia and macrosomiaincrease serum ketone levels.236 Similar results have been in GDM, which would establish guidelines for insulinreported by Jovanovic-Peterson and Peterson, who rec- therapy, some authors have suggested that estimationommend 30 kcal/kg present pregnant weight for normal of glycemia alone may not be sufﬁcient to optimallyweight women, 24 kcal/kg for overweight women, and prescribe insulin therapy in these cases. Buchanan and12 kcal/kg for morbidly obese women.237 These authors colleagues244 have reviewed the utility of fetal ultrasoundindicate that mild caloric restriction with modiﬁcation measurements to guide insulin therapy in women withof the carbohydrate component may be advised in obese GDM. In their study of diet-treated patients with GDM,GDM women.238 ultrasound preformed at 29 to 33 weeks was used to Once the patient with GDM is placed on an appropri- identify pregnancies with fetuses having a large abdomi-ate diet, surveillance of blood glucose levels is necessary nal circumference (>75th percentile). These patients wereto be certain that glycemic control has been established. then randomized to diet versus diet and insulin treatment.At a minimum, practitioners have performed weekly The insulin-treated group ultimately had a frequency ofassessment of fasting or postprandial glucose levels or LGA infants of 13 percent, which was far below theboth at clinic or ofﬁce visits. Some clinicians prefer to 45 percent present in the diet-treated group. The cost-have patients perform daily self–blood glucose monitor- effectiveness of this approach needs to be compared withing, which in two retrospective studies has been associ- administration of insulin to women with fasting hyper-ated with a decline in macrosomia at the expense of glycemia on diet therapy, because approximately twonearly half of all women requiring insulin therapy.239,240 A thirds of women require insulin therapy for a large fetalpractical approach may be to provide women with GDM abdominal circumference on ultrasound examination inwith a reﬂectance meter; however, if after a few weeks, this range of maternal glycemia.both fasting and postprandial measurements are within Finally, oral hypoglycemic therapy has emerged as athe normal range, the frequency of testing can be reduced suitable alternative to insulin treatment in women withand tailored accordingly. GDM.245 In 2000, Langer and colleagues reported a ran- Whereas the ACOG have recommended that fasting domized trial of 404 women receiving insulin versus gly-plasma glucose levels be maintained below 105 mg/dl buride and reported similar improvement in glycemia withand 2-hour postprandial values be less than 120 mg/dl both regimens. The frequency of macrosomia and neo-in women with GDM, thresholds of a fasting glucose natal hypoglycemia was similar in the two study groups.less than 95 mg/dl and 1-hour postprandial glucose less Only 4 percent of women failed glyburide therapy, requir-than 140 mg/dl, as well as 2-hour postprandial glucose ing a change to insulin. Cord blood analysis revealed noless than 120 mg/dl have been suggested by the Fourth detectable glyburide in exposed pregnancies. Subsequently,International Workshop Conference.241 If a patient repet- several smaller studies reported success in achieving gooditively exceeds these thresholds, then insulin therapy is glycemic control with glyburide, but with slightly highersuggested. The use of the above-mentioned cutoffs for failure rates (15 to 20 percent).225,246–248 Jacobson andinitiating insulin are based on data regarding increased colleagues249 recently reported on the implementation ofperinatal morbidity when such values are exceeded in glyburide as an alternative to insulin in a large managedwomen with preexisting diabetes. At present, there are no care organization. These authors noted a similar fre-data from controlled trials to identify ideal glycemic targets quency of LGA infants and macrosomia among 268for prevention of fetal risk for women with GDM. women treated with insulin compared with 236 receiv- Langer and colleagues242 have critically evaluated ing glyburide. In this nonrandomized study, more womenthresholds for insulin therapy in patients with GDM in the glyburide group achieved lower mean fasting andand concluded that to evaluate the effect of therapy, postprandial glucose levels compared with insulin-treatedappropriate endpoints should include fetal macrosomia subjects. Importantly, the authors noted an increased rateor LGA infants, and neonatal metabolic complications. of preeclampsia, need for neonatal phototherapy, andLanger and colleagues242 evaluated insulin secretion pat- birth injury in the “glyburide group,” all of which point
1002 Section VI Pregnancy and Coexisting Diseaseto the need for further study concerning safety.250 Along fetal lung maturity in such cases should be based onwith such studies, it is likely that other oral agents will clinical circumstances. As with antepartum fetal testing,also be evaluated for use during pregnancy. in the otherwise uncomplicated group, should elective Bung and colleagues251 conducted a prospective study induction be the standard approach for these pregnan-of the utility of exercise in the treatment of GDM. These cies complicated by GDM? Lurie and coworkers256 haveauthors randomized 41 women with GDM who man- in part addressed this issue by retrospectively examiningifested elevated fasting glucose levels and would nor- the outcomes of 124 women with GDM delivered beyondmally require insulin therapy. In the ﬁnal analysis, 17 40 weeks’ gestation compared with the same number ofwomen completed a supervised bicycle ergometry train- women with GDM delivered before their expected date ofing program compared with 17 women receiving insulin conﬁnement. Antepartum fetal surveillance was not rou-treatment. No statistical differences were observed in tinely begun until 40 weeks’ gestation. No signiﬁcant dif-weekly blood glucose determinations between study ferences in perinatal outcome, rates of cesarean delivery,groups. All fetal heart rate patterns were reactive before or shoulder dystocia were found between study groups.and after exercise. Thus, regular exercise may be an effec- A vaginal delivery rate of 75.8 percent was achieved intive treatment for GDM. Brisk walking for 30 minutes women with GDM delivering beyond 40 weeks’ ges-at least three times each week may be recommended. tation. These authors concluded that elective inductionBecause the total number of women with GDM studied before 40 weeks’ gestation should be avoided and everyin randomized trials is limited, the role of exercise as a attempt should be made to allow women with GDM,primary therapy in GDM is unknown.252 both diet and insulin treated, to proceed to spontaneous Patients with GDM who are well controlled are at labor. In contrast, a follow-up prospective study fromlow risk for an intrauterine death. For this reason, we the same institution of 96 insulin-requiring patients withdo not routinely institute antepartum fetal heart rate GDM demonstrated that induction at 38 to 39 weekstesting in uncomplicated diet-controlled GDM patients was associated with a 1.4 percent shoulder dystocia rateunless the patient has a hypertensive disorder, a history versus 10.2 percent in historic controls.257of a prior stillbirth, or suspected macrosomia.253 Women Kjos and colleagues224 conducted a prospective ran-in these categories as well as those who require insulin domized trial of active induction of labor at 38 weeks’treatment of GDM undergo twice-weekly heart rate gestation versus expectant management in a series whichtesting at 32 weeks’ gestation. Women with uncompli- included 187 insulin-requiring women with GDM. Thecated GDM do undergo fetal heart rate testing at 40 cesarean delivery rate was not signiﬁcantly differentweeks’ gestation. Using such a protocol at The Ohio State in the expectant-management group (31 percent) fromUniversity Hospital Diabetes in Pregnancy Program, only the active-induction group (25 percent). However, anthree intrauterine deaths in more than 2,000 patients increased prevalence of LGA infants (23 percent versuswith uncomplicated GDM have been observed in the 10 percent) was observed in the expectant managementlast 16 years. Thus, it appears that the third-trimester group. Moreover, the frequency of shoulder dystociastillbirth rate in these patients is no higher than that of was 3 percent in this group, with no cases reported inthe general obstetric population. A study of 389 women those undergoing induction at 38 weeks’ gestation. Thesewith GDM documented an antepartum stillbirth rate data led the authors to conclude that scheduled electiveof 7.7 per 10,000, which was not signiﬁcantly different induction be considered in insulin-requiring patients withfrom the rate of 4.8 per 1,000 observed in nondiabetic GDM because it does not increase the risk of cesareanlow-risk patients.254 In this study, because 7 percent of delivery and lowers the risk for fetal death. In patientsfetuses were delivered on the basis of a low BPP score, managed expectantly, careful monitoring of fetal growththe beneﬁt of testing all GDM pregnancies remains in should be performed because of an apparent increasingquestion. At present, without a large prospective study risk for macrosomia with advancing gestational age incomparing outcomes in monitored and nonmonitored this population.women with GDM without other risk factors, it is notpossible to determine if any beneﬁts exist to antepartumfetal surveillance in this seemingly low-risk population. COUNSELING THE DIABETIC PATIENT For pregnancies determined to be at sufﬁcient risk toundergo testing, a biweekly regimen appears to be prefer- Anomalies of the cardiac, renal, and central nervousable to weekly testing. In a series of 1,390 women with systems arise during the ﬁrst 7 weeks of gestation, a timeGDM, there were no antepartum stillbirths within 4 days when it is most unusual for patients to seek prenatal care.of a reassuring test, whereas two fetal deaths occurred Therefore, the management and counseling of womenat 28 and 36 weeks, respectively, in insulin-requiring with diabetes in the reproductive age group should beginwomen 1 week after negative testing.255 before conception. Unfortunately, it has been estimated Because many obstetricians have extrapolated the that less than 20 percent of diabetic women in the Unitedincreased risk for stillbirth in women with type 1 and States obtain prepregnancy counseling.210 Prepregnancytype 2 diabetes to those with GDM, a remarkable number counseling includes an assessment of vascular status andof these pregnancies are subject to scheduled delivery at glycemic control. Physicians who care for young womenterm. If glycemic control is suboptimal, or maternal hyper- with diabetes must be aware of the importance of suchtension or a previous stillbirth exists, such an approach counseling. At this time, the nonpregnant patient mayseems warranted. The use of amniocentesis to document learn techniques for self glucose monitoring as well as the
Chapter 37 Diabetes Mellitus Complicating Pregnancy 1003 Table 37-14. Comparative Rates of Major Malformations in Offspring of Diabetic Women Receiving Preconceptional Counseling STUDY WITH PRECONCEPTIONAL COUNSELING WITHOUT PRECONCEPTIONAL COUNSELING Fuhrmann et al259 1/128 (0.8%) 22/292 (7.5%) Steel et al275 2/143 (1.4%) 10/96 (10.4%) Kitzmiller et al260 1/84 (1.2%) 12/110 (10.9%) Whillhoite et al264 1/62 (1.6%) 8/123 (6.5%)need for proper dietary management. Folic acid dietary patients with elevated glycosylated hemoglobin levels,supplementation at a dose of at least 0.4 mg daily should 13 (22 percent) malformed infants were noted. This is inbe prescribed because there is increasing evidence that contrast to a 3.4-percent incidence of major malforma-this vitamin may reduce the frequency of neural tube tions in 58 women whose glycosylated hemoglobin levelsdefects although it has not speciﬁcally been studied in were in the normal range. Overall, the risk of a majorthe diabetic population. During counseling, questions fetal anomaly may be as high as 25 percent when themay be answered regarding risk factors for complica- glycosylated hemoglobin level is several percent abovetions and the plan for general management of diabetes in normal values. Greene262 has reported that 14 of 35 preg-pregnancy. Planning for pregnancy should optimally be nancies with a glycosylated hemoglobin exceeding 12.8accomplished over several months. Glycosylated hemo- percent were complicated by major malformations. In hisglobin measurements are performed to aid in the timing series from the Joslin Clinic, the risk for major anoma-of conception. lies did not become evident until glycosylated hemoglo- A reduced rate of major congenital malformation bin values exceed 6 SD above the mean. In contrast toin patients optimally managed before conception is the investigations cited above is the DIEP (Diabetes inobserved with special diabetes clinics (Table 37-14). In Early Pregnancy) study,89 in which malformation ratesCopenhagen, the rate of malformations fell from 19.4 in IDMs were not correlated with ﬁrst trimester maternalpercent to 8.5 percent in class D and F patients who glycosylated hemoglobin levels. The authors suggestedattended a prepregnancy clinic.258 Fuhrmann et al.259 that more sensitive measures are needed to identify tera-found that intensive treatment begun before concep- togenic mechanisms or that not all malformations cantion in 307 East German diabetic women reduced the be prevented by good glycemic control. Further reviewmalformation rate to 1 percent. Nearly 90 percent of of these data, which included glycosylated hemoglobinwomen in this study maintained mean glucose levels less levels only in the early entry patients, demonstrates thatthan 100 mg/dl (5.6 mmol/L). In contrast, the incidence these women were a relatively homogeneous group withof anomalies in the offspring of 593 diabetic women respect to glycemic control; 93 percent had glycosylatedwho registered for care after 8 weeks’ gestation was 8.0 hemoglobin levels less than 7 SD below the mean, a levelpercent (47/593). Only 205 of those women had mean of control which barely increases the risk for anomaliesdaily glucose levels of less than 100 mg/dl (5.6 mmol/L). according to Greene’s data.262 Regardless of the glyco-Mills et al.97 have reported that diabetic women regis- sylated hemoglobin value obtained, all patients requiretered before pregnancy had fewer infants with anoma- a careful program of surveillance, as outlined earlier,lies when compared with late registrants (4.9 versus 9.0 to detect fetal malformations. The risk for spontane-percent). Although the incidence of 4.9 percent remains ous abortion also appears to be increased with markedhigher than that in a normal control population (2 elevations in glycosylated hemoglobin. However, forpercent), normalization of glycemia was not established diabetic women in good control, there appears to be noin the early entry group. greater likelihood of miscarriage.263 In summary, women Kitzmiller and colleagues260 studied 84 women with with type 1 and type 2 diabetes mellitus should bepregestational diabetes mellitus who were recruited for advised to achieve a HbA1c level less than 1 percentpreconception education and management during a 7- above the upper limit of the normal range before con-year period. A group of 110 pregnancies in women with ceiving to reduce the risk for a major fetal malformationIDDM presenting in the ﬁrst trimester without precon- or miscarriage.ceptional counseling served as controls in this study.One anomaly (1.2 percent) occurred in the preconcep-tion group versus 12 (10.9 percent) malformations in the CONTRACEPTIONcontrol population. Glycosylated hemoglobin levels obtained during the There is no evidence that diabetes mellitus impairsﬁrst trimester may be used to counsel diabetic women fertility. Thus, family planning is an important consid-regarding the risk for an anomalous infant. In a ret- eration for the diabetic woman. A careful history androspective study of 116 women at the Joslin Clinic, complete gynecologic examination and counseling areMiller and colleagues261 observed that elevated hemo- required before selecting a method of contraception.globin A1c concentrations early in pregnancy correlated Barrier methods continue to be a safe and inexpensivewith an increased incidence of malformations. In 58 method of birth control. The diaphragm, used correctly
1004 Section VI Pregnancy and Coexisting Diseasewith a spermicide, has a failure rate of less than 10 previous GDM. A statistically signiﬁcant, yet clinicallypercent. Because there are no inherent risks to the dia- limited deterioration in carbohydrate tolerance has beenphragm and other barrier methods, these have become reported in healthy depomedroxyprogesterone acetatethe preferred interim method of contraception for women (Depo-Provera; DMPA) users.273 As observed with otherwith diabetes mellitus. The intrauterine device may also progestins, DMPA may lower serum triglyceride andbe used by diabetic women without concerns about an HDL-C levels but not total cholesterol or LDL-C.274increased risk of infection.265 For this reason, DMPA is not recommended as a ﬁrst Combined oral contraceptives (OCs) are the most line method of contraception for women with diabetes.effective reversible method of contraception with failure The progestin-only OC would be preferred because itrates generally less than 1 percent. There is, however, does not produce signiﬁcant metabolic effects in diabeticcontinued controversy regarding their use in the diabetic women.woman. The serious side effects of pill use, includingthromboembolic disease and myocardial infarction, maybe increased in diabetic women using combined OCs.In a retrospective study, Steel and Duncan266 observed KEY POINTSﬁve cardiovascular complications in 136 diabetic womenusing primary low-dose pills. Three patients had cerebro- ❑ Pregnancy has been characterized as a diabeto-vascular accidents, one had a myocardial infarction, and genic state because of increased postprandial glucoseone an axillary vein thrombosis. In a recent retrospective levels in a late gestation.case-control study, despite diabetes increasing the riskfor cerebral thromboembolism ﬁvefold compared with ❑ Both hepatic and peripheral (tissue) insulin sensi-controls, this risk was not enhanced by use of combined tivity are reduced in normal pregnancy. As a result,oral contraceptives.267 a progressive increase in insulin secretion follows a In Steel and Duncan’s report, several women exhib- glucose challenge.ited rapid progression of retinopathy. Klein and col- ❑ In women with GDM, the hormonal milieu ofleagues268 studied OC use in a cross-sectional study of pregnancy may represent an unmasking of a sus-384 insulin dependent women and reported no associa- ceptibility to the development of type 2 diabetestion between OCs and progression of vascular compli- mellitus.cations. For physicians who prescribe low-dose OCs todiabetic women, their use should probably be restricted ❑ According to the Pedersen hypothesis, maternalto patients without serious vascular complications or hyperglycemia results in fetal hyperglycemia andadditional risk factors such as a strong family history hyperinsulinemia, resulting in excessive fetal growth.of myocardial disease or smoking. In these women, a Tight maternal glycemic control is associated with amonophasic preparation (progestin only) may be consid- reduced risk for fetal macrosomia.ered. In women receiving oral contraceptives, the lowestdose of estrogen and progesterone should be employed. ❑ Congenital malformations occur with a two- toPatients should have blood pressure monitoring after sixfold increased rate in offspring of women withthe ﬁrst cycle and quarterly with baseline and follow- pregestational diabetes compared with the normalup lipid levels as well. population. Impaired glycemic control and associ- Women using OCs may demonstrate increased resis- ated derangement in maternal metabolism appear totance to insulin as a result of a diminished concentration contribute to abnormal embryogenesis.of insulin receptors.269 Despite the fact that carbohydrate ❑ Women with class F (nephropathy) diabetesmetabolism may be affected by the progestin component have an increased risk for preeclampsia and pretermof the pill, disturbances in diabetic control are actually delivery that correlates with their degree of renaluncommon with its use. In Steel and Duncan’s study,270 impairment.81 percent of patients using the pill did not require achange in insulin dose. Triphasic OCs may also be used ❑ Diabetes retinopathy may worsen during preg-safely in former GDM women without other risk factors. nancy, yet for women optimally treated with laserSkouby et al.271 have demonstrated that normal glucose photocoagulation before pregnancy, signiﬁcant dete-tolerance and lipid levels can be expected in non-obese rioration of vision is uncommon.former GDM women followed after 6 months of therapy.Kjos and colleagues272 performed a prospective random- ❑ Screening for GDM is generally performed betweenized study of 230 women with recent GDM. OC users 24 and 28 weeks’ gestation. Screening strategieswere randomized to low-dose norethindrone or levonorg- include universal screening or limiting screening toestrel preparations in combination with ethinyl estradiol. women over age 25 with risk factors for developingThe rate of subsequent diabetes in OC users was 15 to adult-onset diabetes mellitus.20 percent, after one year follow-up. This rate was not ❑ Treatment of women with type 1 and type 2 dia-signiﬁcantly different from non-OC users (17 percent). betes mellitus during pregnancy requires intensiveImportantly, no adverse effects on total cholesterol, LDL, therapy consisting of frequent self–blood glucoseHDL, or triglycerides were found with OC use. monitoring and aggressive insulin dosing by multiple At present, there is little information available concern- injections or CSII (insulin pump).ing long-acting progestins in women with diabetes or
Chapter 37 Diabetes Mellitus Complicating Pregnancy 1005 13. DeFronzo RA, Tobin JD, Andres R: Glucose clamp technique: ❑ The cornerstone of treatment for GDM is dietary A method for quantifying insulin secretion and resistance. Am J therapy. Insulin or glyburide are reserved for indi- Physiol 237:E214, 1979. viduals who manifest signiﬁcant fasting hyperglyce- 14. Catalano PM, Tyzbir ED, Wolfe RR, et al: Carbohydrate metabo- mia or postprandial glucose elevations despite dietary lism during pregnancy in control subjects and women with gesta- tional diabetes. Am J Physiol 264:E60, 1993. intervention. 15. Catalano PM, Huston L, Amini SB, Kalhan SC: Longitudinal changes in glucose metabolism during pregnancy in obese women ❑ Antepartum fetal assessment for women with with normal glucose tolerance and gestational diabetes. Am J both pregestational diabetes or GDM is based on Obstet Gynecol 180:903, 1999. the degree of risk believed to be present in each 16. Spellacy WN, Goetz FC, Greenberg BZ, et al: Plasma insulin case. Glycemic control, prior obstetric history, and normal “early” pregnancy. Obstet Gynecol 25:862, 1965. 17. Burt RL: Peripheral utilization of glucose in pregnancy. III Insulin the presence of vascular disease or hypertension are intolerance. Obstet Gynecol 2:558, 1956. important considerations. 18. Fisher PM, Sutherland HW, Bewsher PD: The insulin response to glucose infusion in normal human pregnancy. Diabetologia ❑ Delivery should be delayed until fetal maturation 19:15, 1980. has occurred, provided that diabetes is well con- 19. Buchanan TZ, Metzger BE, Freinkel N, et al: Insulin sensitivity and trolled and fetal surveillance remains normal. The β-cell responsiveness to glucose during late pregnancy in lean and mode of delivery for the suspected large fetus remains moderately obese women with normal glucose tolerance or mild gestational diabetes. Am J Obstet Gynecol 162:1008, 1990. controversial. In cases of suspected macrosomia, a 20. Ryan EA, O’Sullivan MJ, Skyler JS: Insulin action during preg- low threshold for cesarean delivery has been recom- nancy. Studies with the euglycemic clamp technique. Diabetes mended to prevent a traumatic birth. 34:380, 1985. 21. Hay WW, Sparks JW, Wilkening RB, et al: Partition of maternal ❑ Women with type 1 and type 2 diabetes melli- glucose production between conceptus and maternal tissue in tus should seek prepregnancy consultation. Efforts sheep. Am J Physiol 234:E347, 1983. 22. Marconi AM, Paolini C, Buscaglia M, et al: The impact of gesta- to improve glycemic control before conception have tional age and fetal growth on the maternal-fetal glucose concen- been associated with a signiﬁcant reduction in the tration difference. Obstet Gynecol 87:937, 1996. rate of congenital malformations in the offspring of 23. Kirwan JP, Hauguel-de Mouzon S, Lepercq J, et al: TNFα is such women. a predictor of insulin resistance in human pregnancy. Diabetes 51:2207, 2002. 24. Barros LF, Yudilevich DL, Jarvis SM, et al: Quantitation and immunolocalization of glucose transporters in the human pla- centa. Placenta 16:623, 1995. 25. Jansson T, Wennergren M, Illsley NP: Glucose transporter expres-REFERENCES sion and distribution in the human placenta throughout gestation and in intrauterine growth retardation. J Clin Endocrinol Metab 1. Weiss PAM, Hoffman H: Intensiﬁed conventional insulin 77:1554, 1993. therapy for the pregnant diabetic patient. Obstet Gynecol 64:629, 26. Hauguel-de Mouzon S, Challier J, Kacemi A, et al: The GLUT3 1984. glucose transporter isoform is differentially expressed with the 2. Catalano PM, Tyzbir ED, Roman NM, et al: Longitudinal changes human placental cell types. J Clin Endocrinol Metab 82:2689, in insulin release and insulin resistance in non-obese pregnant 1997. women. Am J Obstet Gynecol 165:1667, 1991. 27. Xing A, Cauzac M, Challier J, et al: Unexpected expression of 3. Bellman O, Hartman E: Inﬂuence of pregnancy on the kinetics of GLUT 4 glucose transporter in villous stromal cells of human insulin. Am J Obstet Gynecol 122:829, 1975. placenta. J Clin Endocrinol Metab 83:4097, 1999. 4. Lind T, Bell S, Gilmore E: Insulin disappearance rate in pregnant 28. White P: Pregnancy complicating diabetes. Am J Med 7:609, and non-pregnant women and in non-pregnant women given 1949. GHRH. Eur J Clin Invest 7:47, 1977. 29. Ellard S, Beards F, Allen LIS, et al: A high prevalence of the glu- 5. Burt RL, Davidson IWF: Insulin half-life and utilization in normal cokinase mutation in gestational diabetic subjects selected by dual pregnancy. Obstet Gynecol 43:161, 1974. criteria. Diabetalogia 43:250, 2000. 6. Goodner CJ, Freinkel N: Carbohydrate metabolism in pregnancy: 30. Stoffel M, Bell MK, Blackburn CC, et al: Identiﬁcation of gluco- The degradation of insulin by extracts of maternal and fetal struc- kinase mutations in subjects with gestational diabetes mellitus. tures in the pregnant rat. Endocrinology 65:957, 1959. Diabetes 42:937, 1993. 7. Catalano PM, Drago NM, Amini SB: Longitudinal changes in 31. Diamond MP, Reece EA, Caprios L, et al: Impairment of counter pancreatic b cell function and metabolic clearance rate of insulin regulatory hormone responses to hypoglycemia in pregnant in pregnant women with normal and abnormal glucose tolerance. women with insulin-dependent diabetes mellitus. Am J Obstet Diabetes Care 21:403, 1998. Gynecol 166:70, 1992. 8. Kalhan SC, D’Angelo LJ, Savin SM, et al: Glucose production in 32. Rosenn BM, Miodovnik M, Khoury JC, et al: Counter regulatory pregnant women at term gestation: Sources of glucose for human hormonal responses to hypoglycemic during pregnancy. Obstet fetus. J Clin Invest 63:388, 1979. Gynecol 87:568, 1996. 9. Cowett RA, Susa JB, Kahn CB, et al: Glucose kinetics in nondia- 33. McManus RM, Ryan EA: Insulin requirements in insulin-depen- betic and diabetic women during the third trimester of pregnancy. dent and insulin-requiring GDM women during the ﬁnal month Am J Obstet Gynecol 146:773, 1983. of pregnancy. Diabetes Care 15:1323, 1992.10. Catalano PM, Tyzbir ED, Wolfe RR, et al: Longitudinal changes 34. Schmitz O, Klebe J, Moller, J, et al: In vivo insulin action in type in basal hepatic glucose production and suppression during 1 (insulin-dependent) diabetic pregnant women as assessed by the insulin infusion in normal pregnant women. Am J Obstet Gynecol insulin clamp technique. J Clin Endocrinol Metab 61:877, 1985. 167:913, 1992. 35. Jansson T, Powell TL: Glucose transport and GLUT1 expression11. Sivan E, Chen X, Hombo CJ, et al: Longitudinal study of car- are upregulated in placentas from pregnancies complicated by bohydrate metabolism in healthy obese women. Diabetes Care severe diabetes (abstract). Placenta 18:A30, 1997. 20:1470, 1997. 36. Sims EAH, Calles-Escadon J: Classiﬁcation of diabetes: A fresh12. Pacini G, Bergman RN: MINMOD: A computer program to cal- look for the 1990s? Diabetes Care 13:1123, 1990. culate insulin sensitivity and pancreatic responsitivity from the 37. Steel JM, Irvine WJ, Clark BJ: The signiﬁcance of pancreatic islet frequently sampled intravenous glucose tolerance test. Comput cell antibody and abnormal glucose tolerance during pregnancy. Methods Programs Biomed 23:113, 986. J Clin Lab Immunol 4:83, 1980.
1006 Section VI Pregnancy and Coexisting Disease38. Ginsberg-Fellner F, Mark EM, Nechemias C, et al: Autoantibodies 60. Cetin I, Nobile de Santis MS, Taricco E, et al: Maternal and fetal to islet cells: Comparison of methods. (Letter.). Lancet 2:1218, amino acid concentrations in normal pregnancies and in preg- 1982. nancies with gestational diabetes mellitus. Am J Obstet Gynecol39. Catalano PM, Tyzbir ED, Sims EAH: Incidence and signiﬁcance of 192:610, 2005. islet cell antibodies in women with previous gestational diabetes 61. Ogburn PL, Goldstein M, Walker J, Stonestreet BS: Prolonged mellitus. Diabetes Care 13:478, 1990. hyperinsulinemia reduces plasma fatty acid levels in the major lipid40. Ward WK, Johnson CLW, Beard JC, et al: Abnormalities of islet groups in fetal sheep. Am J Obstet Gynecol 161:728, 1989. β-cell function, insulin action and fat distribution in women with 62. Kliegman R, Gross T, Morton S, Dunnington R: Intrauterine histories of gestational diabetes: Relationship to obesity. J Clin growth and post natal fasting metabolism in infants of obese Endocrinol Metab 61:1039, 1985. mothers. J Pediatr 104:601, 1984.41. Catalano PM, Bernstein IM, Wolfe RR, et al: Subclinical abnor- 63. Darmady JM, Postle AD: Lipid metabolism in pregnancy. Br J malities of glucose metabolism in subjects with previous gesta- Obstet Gynaecol 82:211, 1982. tional diabetes. Am J Obstet Gynecol 166:1255, 1986. 64. Knopp RH, Chapman M, Bergeline RO, et al: Relationship of42. Ryan EA, Imes S, Liu D, et al: Defects in insulin secretion and lipoprotein lipids to mild fasting hyperglycemia and diabetes in action in women with a history of gestational diabetes. Diabetes pregnancy. Diabetes Care 3:416, 1980. 44:506, 1995. 65. Montelongo A, Lasuncion MA, Pallardo LF, et al: Longitudinal43. Yen SCC, Tsai CC, Vela P: Gestational diabetogenesis: Quantita- study of plasma lipoproteins and hormones during pregnancy in tive analysis of glucose-insulin interrelationship between normal normal and diabetic women. Diabetes 41:1651, 1992. pregnancy and pregnancy with gestational diabetes. Am J Obstet 66. Koukkou E, Watts GF, Lowy C: Serum lipid, lipoprotein and Gynecol 111:792, 1971. apolipoprotein changes in gestational diabetes mellitus: a cross-44. Fisher PM, Sutherland HW, Bewsher PD: The insulin response sectional and prospective study. J Clin Pathol 49:634, 1996. to glucose infusion in gestational diabetes. Diabetologia 19:14, 67. Catalano PM, Nizielski SE, Shao J, et al: Down regulation of IRS- 1980. 1 and PPARgamma in obese women with gestational diabetes:45. Xiang AH, Peters RH, Trigo E, et al: Multiple metabolic defects Relationship to free fatty acids during pregnancy. Am J Physiol during late pregnancy in women at high risk for type 2 diabetes. Endocrinol 282:E522–33, 2002. Diabetes 48:848, 1999. 68. Catalano PM, Hollenbeck C: Energy requirements in pregnancy:46. Catalano PM, Roman-Drago N, Amini SB, Sims EAH: Longitu- A review. Obstet Gynecol Surv 47:368, 1992. dinal changes in body composition and energy balance in lean 69. Goldberg GR, Prentice AM, Coward WA, et al: Longitudinal women with normal and abnormal glucose tolerance during preg- assessment of energy expenditure in pregnancy by the doubly nancy. Am J Obstet Gynecol 179:156, 1998. labels water method. Am J Clin Nutr 57:494, 1993.47. Bergman RN, Philips LS, Cobelli C: Physiologic evaluation of 70. Forsum E, Sadurskis A, Wager J: Resting metabolic rate and Boyd factors controlling glucose disposition in man. Measurement of composition of healthy Swedish women during pregnancy. Am J insulin sensitivity and β-cell sensitivity from the response to intra- Clin Nutr 47:94, 1988. venous glucose. J Clin Invest 68:1457, 1981. 71. Lawrence M, Lawrence F, Coward WA, et al: Energy require-48. Buchanan TA: Pancreatic β-cell defects in gestational diabetes: ments of pregnancy in the Gambia. Lancet ii:1072, 1987. Implications for the pathogenesis and prevention of type 2 diabe- 72. Forsum E, Kabir N, Sadurskis A: Westerp: Total energy expendi- tes. J Clin Endocrinol Metab 86:898, 2001. ture of healthy Swedish women during pregnancy and lactation.49. Garvey WT, Maianu L, Hancock JA, et al: Gene expression of Am J Cline Nutr 56:334, 1992. GLUT4 in skeletal muscle from insulin-resistance patients with 73. Butte NF, Wong WW, Treuth MS, et al: Energy requirements obesity, IGT, GDM, and NIDDM. Diabetes 41:465, 1992. during pregnancy ased on otal energy expenditure and energy50. Friedman JE, Ishizuka T, Shao J, et al: Impaired glucose transport disposition. Am J Clin Nutr 79:1078, 2004. and insulin receptor tyrosine phosphorylation in skeletal muscle 74. King JC, Butte NF, Bronstein MN, et al: Energy metabolism from obese women with gestational diabetes. Diabetes 48:1807, during pregnancy: Inﬂuence of maternal energy status. Am J Cline 1999. Nutr 59:4395, 1994.51. Hytten FE, Leitch I: The gross composition of the components of 75. Prentice AM, Poppitt SD, Goldberg CR, et al: Energy balance in weight gain. In Landis EM, Pappenheimer JR (eds): The Physiol- pregnancy and lactation. In Allen L, King J, Lonnerdal B (eds): ogy of Human Pregnancy, 2nd ed. London, Blackwell Scientiﬁc, Nutrient Regulation During Pregnancy, Lactation and Infant 1971, p 371. Growth, New York, Plenum Press, 1994, p 11.52. Metzger BD, Unger RH, Freinkel N: Carbohydrate metabolism 76. Swinburn BA, Myomba BC, Saad MF, et al: Insulin resistance in pregnancy. XIV. Relationships between circulation glucagon, associated with lower rates of weight gain in PIMA Indians. J insulin, glucose and amino acids in response to a “mixed meal” Clin Invest 88:168, 1991. in late pregnancy. Metabolism 26:151, 1977. 77. Catalano PM, Roma NM, Tyzbir ED, et al: Weight gain in women53. Freinkel N, Metzger BE, Nitzan M, et al: “Accelerated starvation” with gestational diabetes. Obstet Gynecol 81:523, 1993. and mechanisms for the conservation of maternal nitrogen during 78. Okereke NC, Huston-Presley L, Amini SB, et al: Longitudinal pregnancy. Israel J Med Sci 8:426, 1972. changes in energy expenditure and body composition in obese54. Kalkhoff RK, Kandaraki E, Morrow PG, et al: Relationship women with normal and impaired glucose tolerance. Am J Physiol between neonatal birth weight and maternal plasma amino acids Endocrinol Metab 287:E472, 2004. proﬁles in lean and obese nondiabetic women with type 1 diabetic 79. Catalano PM, Thomas A, Huston-Presley L, et al: Increased fetal pregnant women. Metabolism 37:234, 1988. adiposity: A very sensitive marker of abnormal in utero develop-55. Ogata ES: The small for gestational age neonate. In Consell RM ment. Am J Obstet Gynecol 189:1698, 2003. (ed): Principles of Perinatal-Neonatal Metabolism, 2nd ed. New 80. Landon MB, Gabbe SG: Fetal surveillance in the pregnancy com- York, Springer-Verlag, 1998, p 1097. plicated by diabetes mellitus. Clin Perinatol 20:549, 1993.56. Duggleby SC, Jackson AA: Protein, amino acid and nitrogen 81. Centers for Disease Control and Prevention: Perinatal mortality metabolism during pregnancy: How might the mother meet the and congenital malformations in infants born to women with needs of her fetus? Curr Opin Clin Nutr Metab Care 5:503, insulin dependent diabetes. MMWR Morb Mortal Wkly Rep 2002. 39:363, 1990.57. Duggleby SC, Jackson AA: Relationship of maternal protein turn- 82. Salversen DR, Brudenell MJ, Nicholaides KH: Fetal polycythemia over and lean body mass during pregnancy and birthweight. Clin and thrombocytopenia in pregnancies complicated by maternal Sci (Lond) 101:65, 2001. diabetes. Am J Obstet Gynecol 166:1987, 1992.58. Kalhan SC, Rossi KQ, Gruca LL, et al: Relation between transami- 83. Madsen H: Fetal oxygenation in diabetic pregnancy. Dan Med nation of brached-chain amino acids and urea synthesis: Evidence Bull 33:64, 1986. from human pregnancy. Am J Physiol 275:E423, 1998. 84. Nylund L, Lunell NO, Lewander R, et al: Uteroplacental blood59. Catalano P, Drago N, Highman T, et al: Longitudinal changes in ﬂow in diabetic pregnancy: Measurements with indium 113m amino acid insulin sensitivity during pregnancy. J Soc Gynecol and a computer linked gamma camera. Am J Obstet Gynecol Invest 145:131A, 1996. 144:298, 1982.
Chapter 37 Diabetes Mellitus Complicating Pregnancy 1007 85. Kitzmiller JL, Phillippe M, von Oeyen P, et al: Hyperglycemia as compared with normal control subjects. Am J Med 72:921, hypoxia, and fetal acidosis in Rhesus monkeys (abstract). Pre- 1981. sented in 28th Annual Meeting of The Society for Gynecologic 109. Landon MB, Gabbe SG, Piana R, et al: Neonatal morbidity in Investigation, St. Louis, MO, March, 1981. pregnancy complicated by diabetes mellitus predictive value of 86. Phillips AF, Dubin JW, Matty PJ, et al: Arterial hypoxemia and maternal glycemic proﬁles. Am J Obstet Gynecol 156:1089, hyperinsulinemia in the chronically hyperglycemia fetal lamb. 1987. Pediatr Res 16:653, 1982. 110. Jovanovic-Peterson L, Peterson CM, Reed CF, et al: Mater- 87. Albert TJ, Landon MB, Wheller JJ, et al: Prenatal detection of nal postprandial glucose levels and infant birthweight: The fetal anomalies in pregnancies complicated by insulin-dependent Diabetes in Early Pregnancy Study. Am J Obstet Gynecol 164:103, diabetes mellitus. Am J Obstet Gynecol (in press). 1991. 88. Simpson JL, Elias S, Martin O, et al: Diabetes in pregnancy, 111. Catalano PM, Tyzbir ED, Allen SR, et al: Evaluation of fetal Northwestern University Series (1977–1981). I. Prospective study growth by estimation of body composition. Obstet Gynecol of anomalies in offspring of mothers with diabetes mellitus. Am J 79:46, 1992. Obstet Gynecol 146:263, 1983. 112. Catalano PM, Drago NM, Amini SB: Factors affecting fetal growth 89. Mills JL, Knopp RH, Simpson JP, et al: Lack of relation of and body composition. Am J Obstet Gynecol 172:1459, 1995. increased malformation rates in infants of diabetic mothers to 113. Catalano PM, Drago NM, Amini SB: Maternal carbohydrate glycemic control during organogenesis. N Engl J Med 318:671, metabolism and its relationship to fetal growth and body compo- 1988. sition. Am J Obstet Gynecol 172:1464, 1995. 90. Eriksson U: The pathogenesis of congenital malformations in dia- 114. Caruso A, Paradisi G, Ferrazzani S, et al: Effect of maternal carbo- betic pregnancy. Diabetes Metab Rev 11:63, 1995. hydrate metabolism in fetal growth. Obstet Gynecol 8, 1998. 91. Reece EA, Hobbins JC: Diabetic embryopathy: pathogenesis, 115. Silverman BL, Rizzo TA, Cho NH, Metzger BE: Long-term effects prenatal diagnosis and prevention. Obstet Gynecol Surv 41:325, of the intrauterine environment. Diabetes, 21:142, 1998. 1986. 116. Pettit DJ, Nelson RG, Saad MF, et al: Diabetes and obesity in the 92. Koppe J, Smoremberg-School M: Diabetes, congenital malforma- offspring of Pima Indian women with diabetes during pregnancy. tions and HLA types. In Listen E, Band H, Frus-Hansen B (eds): Diabetes Care 16:310, 1993. Intensive Care in the Newborn, Vol. 4. Newark, Masson Publish- 117. Taylor R, Lee C, Kyne-Grzebalski D, et al: Clinical outcomes ing, 1983, pp 15. of pregnancy in women with type I diabetes. Obstet Gynecol 93. Simpson JL, Mills J, Ober C, et al: DR3+ and DR4+ diabetes 99:537, 2002. women have increased risk for anomalies. Presented at the 37th 118. Hertel J, Anderson GE, Brandt NJ, et al: Metabolic events in Annual Meeting of the Society for Gynecologic Investigation. infants of diabetic mothers during ﬁrst 24 hours after birth. Acta Abstract 3901, St. Louis, Missouri, 1990. Paediart Scand 71:19, 1982. 94. Freinkel N, Lewis NJ, Akazawa S, et al: The honeybee syndrome: 119. Bourbon JR, Farrell PM: Fetal lung development in the diabetic Implication of the teratogenicity of mannose in rat-embryo culture. pregnancy. Pediatr Res 19:253, 1985. N Engl J Med 310:223, 1984. 120. Smith BT, Giroud CJP, Robert M, Avery ME: Insulin antagonism 95. Goldman AS, Baker L, Piddington R, et al: Hyperglycemia-induced of cortisol action on lecithin synthesis by cultures of fetal lung teratogenesis is mediated by a functional deﬁciency of arachidonic cells. J Pediatr 87:953, 1975. acid. Proc Natl Acad Sci 82:8227, 1985. 121. Smith BT: Pulmonary surfactant during fetal development and 96. Pinter E, Reece EA: Arachidonic acid prevents hyperglycemia- neonatal adaptation: Hormonal control. In Robertson B, Van associated yolk sac damage and embryopathy. Am J Obstet Golde LMB, Batenburg JJ (eds): Pulmonary Surfactant. Amster- Gynecol 166:691, 1986. dam, Elsevier, 1985, p 357. 97. Pinter E, Reece EA, Leranth CZ, et al: Yolk sac failure in 122. Post M, Barsoumian A, Smith BT: The cellular mechanisms of embryopathy due to hyperglycemia ultrastructural analysis of glucocorticoid acceleration of fetal lung maturation. J Biol Chem yolk sac differentiation associated with embryopathy in rat 261:2179, 1986. conceptuses under hyperglycemic conditions. Teratology 33:73, 123. Carlson KS, Smith BT, Post M: Insulin acts on the ﬁbroblast 1986. to inhibit glucocorticoid stimulation of lung maturation. J Appl 98. Eriksson NJ: Protection by free oxygen radical scavenging enzymes Physiol 57:1577, 1984. against glucose-induced embryonic malformations in vitro. Diabe- 124. Dudley DKL, Black DM: Reliability of lecithin/sphingomyelin tologia 34:325, 1991. ratios in diabetic pregnancy. Obstet Gynecol 66:521, 1985. 99. Spellacy WN, Miller S, Winegar A, Peterson PQ: Macrosomia- 125. Kjos SL, Walther F: Prevalence and etiology of respiratory distress maternal characteristics and infant complications. Obstet Gynecol in infants of diabetic mothers: predictive value of lung maturation 66:185, 1985. tests. Am J Obstet Gynecol 163:898, 1990.100. Reiher H, Fuhrmann K, Noack S, et al: Age-dependent insulin 126. Mimouni F, Miodovnik M, Whittset J, et al: Respiratory distress secretion of the endocrine pancreas in vitro from fetuses of dia- syndrome in infants of diabetic mothers in the 1980s: no direct betic and nondiabetic patients. Diabetes Care 6:446, 1983. adverse effect of maternal diabetes with modern management.101. Fallucca F, Garguilo P, Troili F, et al: Amniotic ﬂuid insulin, C- Obstet Gynecol 69:191, 1987. peptide concentrations and fetal morbidity in infants of diabetic 127. Cordero L, Treuer SH, Landon MB, Gabbe SG: Management of mothers. Am J Obstet Gynecol 153:534, 1985. infants of diabetic mother. Arch Pediatr Adolesc Med 152:249,102. Krew MA, Kehl RJ, Thomas A, Catalano PM: Relationship of 1998. amniotic ﬂuid C-peptide levels to neonatal body composition. 128. Mimouni F, Miodovnik M, Tsang RC, et al: Decreased amni- Obstet Gynecol 84:96, 1994. otic ﬂuid magnesium concentration in diabetic pregnancy. Obstet103. Brans YW, Shannon DL, Hunter MA, et al: Maternal diabetes and Gynecol 69:12, 1987. neonatal macrosomia, II Neonatal anthropometric measurements. 129. Widness JA, Cowett RM, Coustan DR, et al: Neonatal morbidi- Early Hum Dev 8:297, 1983. ties in infants of mothers with glucose intolerance in pregnancy.104. Modanlou HD, Komatsu G, Dorchester W, et al: Large-for- Diabetes 34:61, 1985. gestational age neonates: Anthropometric reasons for shoulder 130. Ylinen K, Raivio K, Teramo K: Haemoglobin A1c predicts the dystocia. Obstet Gynecol 60:417, 1982. perinatal outcome in insulin-dependent diabetic pregnancies. Br105. Gabbe SG, Mestman JH, Freeman RK, et al: Management and J Obstet Gynaecol 88:961, 1981. outcome of pregnancy in diabetes mellitus, class B-R. Am J Obstet 131. Stevenson DK, Bartoletti AL, Ostrander CR, Johnson JD: Pul- Gyencol 129:723, 1977. monary excretion of carbon monoxide in the human infants as106. Kitzmiller JL, Gloherty JP: Diabetic pregnancy and perinatal mor- an index of bilirubin production. II. Infants of diabetic mothers. bidity. Am J Obstet Gynecol 131:560, 1978. J Pediatr 94:956, 1979.107. Roversi GD, Gargiulo M: A new approach to the treatment of 132. Shannon K, Davis JC, Kitzmiller JL, et al: Erythropoiesis in infants diabetic pregnant women. Am J Obstet Gynecol 135:567, 1979. of diabetic mothers. Pediatr Res 30:161, 1986.108. Jovanovic L, Druzin M, Peterson CM: Effect of euglycemia on 133. White P: Pregnancy complicating diabetes. Am J Med 7:609, the outcome of pregnancy in insulin-dependent diabetic women 1949.
1008 Section VI Pregnancy and Coexisting Disease134. Summary and Recommendations of the Second International 160. Kitzmiller JL, Gavin LA, Gin GD, et al: Managing diabetes and Workshop-Conference on Gestational Diabetes, Diabetes 34:123, pregnancy. Curr Probl Obstet Gynecol Fertil 11:113, 1988. 1985. 161. Moloney JBM, Drury MI: The effect of pregnancy on the135. Summary and Recommendations of the Third Int’l Workshop natural course of diabetic retinopathy. Am J Ophthalmol 93:745, Conference, Diabetes 40:197, 1991. 1982.136. Gabbe SG, Mestman JH, Freeman RK, et al: Management and 162. Phelps RL, Sakol P, Metzger BE, et al: Changes in diabetic reti- outcome of Class A diabetes mellitus. Am J Obstet Gynecol nopathy during pregnancy, correlations with regulation of hyper- 127:465, 1977. glycemia. Arch Ophthalmol 104:1806, 1986.137. Selby JV, Fitzsimmons SC, Newman JM, et al: The natural history 163. Chew EY, Mills JL, Metzger BE, et al: Metabolic control and and epidemiology of diabetic nephropathy. JAMA 263:1954, progression of retinopathy. The diabetes in early pregnancy study. 1990. Diabetes Care 18:631, 1995.138. Rossing P: Promotion, prediction and prevention of progression 164. Sinclair SH, Nesler C, Foxman B, et al: Macular edema and preg- of nephropathy in type 1 diabetes mellitus [Review]. Diabet Med nancy in insulin dependent diabetes. Am J Ophthalmol 97:154, 15:900, 1998. 1984.139. Parving H-H, Hovind P, Rossing K, et al: Evolving strategies 165. Gordon MC, Landon MB, Boyle J, et al: Myocardial infarction for renoprotection: diabetic nephropathy [Review]. Curr Opin during pregnancy in a patient with Class R/F diabetes mellitus: Nephrol Hypertens 10:515, 2001. A case report and review of literature on Class H IDDM. Obstet140. Lip GYH, Churchill D, Beevers M, et al: Angiotensin-converting Gynecol Surv 51:437, 1996. enzyme inhibitors in early pregnancy. Lancet 350:1446, 1997. 166. Hare JW: Maternal complications. In Hare JW (ed): Diabetes141. Burrows RF, Burrows EA: Assessing the teratogenic potential of complicating pregnancy. The Joslin Clinic Method. New York, angiotensin-converting enzyme inhibitors in pregnancy. Aust NZ Alan R Liss, 1989, p 96. J Obstet Gynaecol 38:306, 1998. 167. Stephenson MJ: Screening for gestational diabetes mellitus: a criti-142. Cox RM, Anderson JM, Cox P: Defective embryogenesis with cal review. J Fam Prac 37:277, 1993. angiotensin II receptor antagonists in pregnany. Br J Obstet Gyn- 168. Solomon CG, Willett WC, Carey VJ, et al: A prospective study aecol 110:1038, 2003. of pregravid determinants of gestational diabetes mellitus. JAMA143. Hod M, van Dijk DJ, Karp M, et al: Diabetic nephropathy and 278:1078, 1997. pregnancy: the effect of ACE inhibitors prior to pregnancy on 169. O’Sullivan JB: Body weight and subsequent diabetes mellitus. maternal outcome. Nephrol Dial Transplant 10:2328, 1995. JAMA 248:949, 1982.144. Combs CA, Rosenn B, Kitzmiller JL, et al: Early-pregnancy 170. Dornhorst A, Rossi M: Risk and prevention of type 2 diabetes in proteinuria in diabetes related to preeclampsia. Obstet Gynecol women with gestational diabetes. Diabetes Care 21:B43, 1998. 82:802, 1993. 171. Kjos SL, Peters RK, Xiang A, et al: Predicting future diabetes in145. Ekbom P, Damn P, Feldt-Rasmussen B, et al: Pregnancy outcome Latino women with gestational diabetes. Utility of early postpar- in type 1 diabetic women with microalbuminuria. Diabetes Care tum glucose tolerance testing. Diabetes 44:586, 1995. 24:1739, 2001. 172. Coustan DR, Nelson C, Carpenter NW, et al: Maternal age and146. Gordon M, Landon MB, Samuels P, et al: Perinatal outcome screening for gestational diabetes: a population based study. and long-term follow-up associated with modern management of Obstet Gynecol 73:557, 1989. diabetic nephropathy (Class F). Obstet Gynecol 87:401, 1996. 173. Metzger BE, Coustan DR: and the Organizing Committee:147. Kitzmiller JL, Brown ER, Phillippe M, et al: Diabetic nephropathy Summary and recommendations of the Fourth International and perinatal outcome. Am J Obstet Gynecol 141:741, 1981. Workshop-Conference on Gestational Diabetes Mellitus. Diabetes148. Reece EA, Coustan DR, Hayslett JP, et al: Diabetic nephropathy: Care 21:B161, 1998. Pregnancy performance and fetomaternal outcome. Am J Obstet 174. Danilenko-Dixon DR, VanWinter JT, Nelson RL, et al: Universal Gynecol 159:56, 1988. versus selective gestational diabetes screening: Application of the149. Kitzmiller JL: Diabetic nephropathy. In Reece EA, Coustan DR, 1997 American Diabetes Association recommendations. Am J Gabbe SG (eds): Diabetes in pregnancy. Philadelphia, Lippincott Obstet Gynecol 181:79, 1999. Williams Wilkins, 2004, p 383. 175. ACOG Practice Bulletin Number 30. Gestational Diabetes,150. Rosenn BM, Miodovnik M, Khoury JC, et al: Outcome of preg- September 2001. nancy in women with diabetic nephropathy. Am J Obstet Gynecol 176. Periodic health examination, 1992 update: 1. Screening for gesta- 176:S631, 1997. tional diabetes mellitus. Can Med Assoc J 147:435, 1992.151. Rossing K, Jacobsen P, Hommel E, et al: Pregnancy and progres- 177. Brody SC, Harris R, Lohr K: Screening for gestational diabetes: sion of diabetic nephropathy. Diabetologia 45:36, 2002. A summary of the evidence for the U.S. Preventive Services Task152. Armenti VT, McGrory CH, Cater J, et al: The national trans- Force. Obstet Gynecol 101:380, 2003. plantatin registry: comparision between pregnancy outcomes 178. Naylor CD: Diagnosing gestational diabetes mellitus: Is the gold in diabetic cyclosporine-treated female kidney recipients and standard valid? Diabetes Care 12:565, 1989. CyA-treated female pancreas-kidney recipients. Transplant Proc 179. Sermer M, Naylor CD, Gore DJ, et al: Impact of increasing car- 29:669, 1997. bohydrate intolerance on maternal-fetal outcomes in 3637 women153. Ogburn PL Jr, Kitzmiller JL, Hare JW, et al: Pregnancy follow- without gestational diabetes. Am J Obstet Gynecol 173:146, ing renal transplantation in Class T diabetes mellitus. JAMA 1995. 255:911, 1986. 180. Sacks DA, Greenspoon JS, Abu-Fadil S, et al: Toward universal154. McCrory CH, Grosheck MA, Sollinger HW, et al: Pregnancy criteria for gestational diabetes: The 75-gram glucose tolerance outcomes in female pancreas-kidney transplants. Transplant Proc test in pregnancy. Am J Obstet Gynecol 172:607, 1995. 31:652, 1999. 181. Langer O, Yogev Y, Most O, Xenakis EM: Gestational diabetes:155. Barrou BM, Gruessner AC, Sutherland DE, et al: Pregnancy after The consequences of not treating. Am J Obstet Gynecol 192:989, pancreas transplantation in the cyclosporine era: report from 2005. the International Pancreas Transplant Registry. Transplantation 182. Crowther CA, Hiller JE, Moss JR, et al: Effect of treatment of 65:524, 1998. gestational diabetes mellitus on pregnancy outcomes. N Engl J156. Carstensen LL, Frost-Lansen K, Fulgeberg S, Nerup J: Does preg- Med 352:2477, 2005. nancy inﬂuence the prognosis of uncomplicated insulin-dependent 183. HAPO Study cooperative Research Group: The Hyperglycemia diabetes? Diabetes Care 5:1, 1982. and Adverse Pregnancy Outcome (HAPO) Study. Int J Gynaecol157. Klein BEK, Moss SE, Klein R: Effect of pregnancy on the progres- Obstet 78:69, 2002. sion of diabetic retinopathy. Diabetes Care 13:34, 1990. 184. Landon MB, Thom E, Spong CY, et al: A planned randomized158. Rosenn B, Miodovnik KM, Kranias G, et al: Progression of clinical trial of treatment for mild gestational diabetes. J. Matern diabetic retinopathy in pregnancy: association with hypertension Fetal Neonatal Med 11:226, 2002. in pregnancy. Am J Obstet Gynecol 166:1214, 1992. 185. Greene MF, Solomon CG: Gestational diabetes mellitus—time to159. Horvat M, Maclear H, Goldberg L, Crock CW: Diabetic retinopa- treat. N Engl J Med 352:24, 2005. thy in pregnancy: A 12 year prospective study. Br J Ophthalmol 186. Coustan DR, Widness JA, Carpenter NW, et al: Should the 64:398, 1980. ﬁfty-gram, one-hour plasma glucose screening test be adminis-
Chapter 37 Diabetes Mellitus Complicating Pregnancy 1009 tered in the fasting or fed state? Am J Obstet Gynecol 154:1031, 211. Barret JM, Salyer SL, Boehm FH: The non-stress test: An 1986. evaluation of 1000 patients. Am J Obstet Gynecol 141:153,187. Sermer M, Naylor CD, Gare DJ, et al: Impact of time since 1981. last meal on the gestational glucose challenge test. Am J Obstet 212. Miller JM, Horger EO: Antepartum heart rate testing in diabetic Gynecol 171:607, 1994. pregnancy. J Repro Med 30:515, 1985.188. Landon MB: Gestational diabetes mellitus: Screening and diagno- 213. Landon MB, Gabbe SG, Bruner JP, Ludmir J: Doppler umbilical sis. Laboratory Medicine 21:527, 1990. artery velocimetry in pregnancy complicated by insulin dependent189. Carpenter MW, Coustan DR: Criteria for screening tests of gesta- diabetes mellitus. Obstet Gynecol 73:961, 1989. tional diabetes. Am J Obstet Gynecol 144:768, 1982. 214. Johnstone FD, Steel JM, Haddad NG, et al: Doppler umbilical190. Bobrowski RA, Bottoms SF, Michallef JA, et al: Is the 50-gram artery ﬂow velocity waveforms in diabetic pregnancy. Br J Obstet glucose screening test ever diagnostic? J Matern Fetal Med 5:317, Gynecol 99:135, 1992. 1996. 215. Landon MB, Gabbe SG: Fetal surveillance in the pregnancy com-191. Sacks DA, Abu-Fadil S, Greenspoon J, et al: Do the current stan- plicated by diabetes mellitus. Clin Perinatol 20:549, 1993. dards for glucose tolerance testing in pregnancy represent a valid 216. Landon MB, Langer O, Gabbe SG, et al: Fetal surveillance in conversion of O’Sullivan’s original criteria? Am J Obstet Gynecol pregnancies complicated by insulin dependent diabetes mellitus. 161:638, 1989. Am J Obstet Gynecol 167:617, 1992.192. National Institutes of Health Diabetes Data Group: Classiﬁcation 217. Landon MB, Vickers S: Fetal surveillance in pregnancy compli- and diagnosis of diabetes mellitus and other categories of glucose cated by diabetes mellitus: is it necessary? J Matern Fetal Neonatal intolerance. Diabetes 20:139, 1979. Med 12:413, 2002.193. Naylor CD, Sermer M, Chen E, et al: Cesarean delivery in relation 218. Milunsky A, Alpert E, Kitzmiller JL, et al: Prenatal diagnosis of to birth weight and gestational glucose intolerance: Pathophysiol- neural tube defects VIII. The importance of serum alpha-fetopro- ogy or practice style? Toronto Tri-Hospital Gestational Diabetes tein screening in diabetic pregnant women. Am J Obstet Gynecol Investigators. JAMA 275:1165, 1996. 142:1030, 1982.194. Landon MB, Gabbe SG: Insulin treatment of the pregnant patient 219. Greene MF, Benacerraf B: Prenatal diagnosis in diabetic gravi- with diabetes mellitus. In Reece EA, Coustan DR, Gabbe SG (ed): das: utility of ultrasound and MSAFP screening. Obstet Gynecol Diabetes Mellitus in Women. Philadelphia, Lippincott, Williams, 77:420, 1991. and Wilkins. 2004. 220. Nesbitt TS, Gilbert WM, Herrchen B: Shoulder dystocia and asso-195. Kitzmiller JL, Main EK, Ward B, et al: Insulin lispro and the ciated risk factors with macrosomic infants born in California. development of proliferative retinopathy during pregnancy. Dia- Am J Obstet Gynecol 179:476, 1998. betes Care 22:873, 1999. 221. Tamura RK, Shabbagha RE, Depp R, et al: Diabetic macrosomia:196. Buchbinder A, Miodovnik M, McElvy S, et al: Is insulin lispro accuracy of third trimester ultrasound. Obstet Gynecol 67:828, a culprit in the progression of diabetic retinopathy during preg- 1986. nancy? Am J Obstet Gynecol 182:S79, 2000. 222. Landon MB, Mintz MG, Gabbe SG: Sonographic evaluation of197. Loukovaara S, Immonen I, Teramo KA, Kaaja R: Progression of fetal abdominal growth: predictor of the large-for-gestational age retinopathy during pregnancy in type 1 diabetic women treated infant in pregnancies. Am J Obstet Gynecol 160:115, 1989 with insulin lispro. Diab Care 26:1193, 2003. 223. Moore TR: A comparison of amniotic ﬂuid pulmonary phospho-198. Pettitt DJ, Kolaczynski JW, Ospina P, et al: Comparision of an lipids in normal and diabetic pregnancy. Am J Obstet Gynecol insulin analog, insulin aspart and regular human insulin with no 186:641, 2002. insulin in gestational diabetes mellitus. Diabetes Care 26:183, 224. Kjos S, Henry O, Montoro M, et al: Insulin-requiring diabetes 2003. in pregnancy: A randomized trial of active induction of labor199. Devlin JT, Hothersall L, Wilkis JL: Use of insulin glargine during and expectant management. Am J Obstet Gynecol 169:611, pregnancy in type 1 diabetic women. Diabetes Care 25:1095, 1993. 2002. 225. Rouse DJ, Owen J, Goldenberg RL, et al: The effectiveness and200. Coustan DR, Reece EA, Sherwin RS, et al: A randomized clinical costs of elective cesarean delivery for fetal macrosomia diagnosed trial of the insulin pump versus intensive conventional therapy in by ultrasound. JAMA 276:1480, 1996. diabetic pregnancy. JAMA 255:631, 1986. 226. Acker DB, Sachs BP, Friedman EA: Risk factors for shoulder201. Gabbe SG, Holing E, Temple P: Beneﬁts, risks, costs, and patient dystocia. Obstet Gynecol 6:762, 1985. satisfaction associated with insulin pump therapy for the preg- 227. ACOG practice bulletin: Shoulder dystocia. Number 40, nancy complicated by type 1 diabetes mellitus. Am J Obstet November 2002. Gynecol 182:1283, 2000. 228. Benedetti TJ, Gabbe SG: Shoulder dystocia: A complication of202. Bode BW, Weinstein R, Bell D, et al: Comparision of insulin fetal macrosomia and prolonged second stage of labor with mid- aspart with buffered regular insulin and insulin lispro of conti- pelvic delivery. Obstet Gynecol 52:526, 1978. nous subcutaneous insulin infusion: a randomized study in type 229. Jovanovic L, Peterson CM: Management of the pregnant, insulin- 1 diabetes. Diabetes Care 25:439, 2002. dependent diabetic woman. Diabetes Care 3:63, 1980.203. American Diabetes Association: Gestational diabetes mellitus 230. Hollingsworth DR, Ney DM: Dietary management of diabetes position statement. Diabetes Care 25:S94, 2002. during pregnancy. In Reece EA, Coustan DR (eds): Diabetes Mel-204. Diamond MP, Reece EA, Caprio S, et al: Impairment of coun- litus in Pregnancy: Principles and Practice. New York, Churchill terregulatory hormone responses to hypoglycemia in pregnant Livingstone, 1988, p 285. women with insulin-dependent diabetes mellitus. Am J Obstet 231. Mumford MI, Jovanovic-Peterson L, Peterson CM: Alternative Gynecol 166:70, 1992. therapies for the management of gestational diabetes. Clin Peri-205. Rosenn BM, Miodovnik M, Khourty JC, et al: Courterregulatory natol 20:619, 1993. hormonal responses to hypoglycemia during pregnancy. Obstet 232. Jovanovic-Peterson L, Peterson CM: Nutritional management of Gynecol 87:568, 1996. the obese gestational diabetic pregnant women. J Am Coll Nutr206. Kilvert JA, Nicholson HO, Wright AD: Ketoacidosis in diabetic 11:246, 1992. pregnancy. Diabetic Medicine 10:278, 1993. 233. Algert S, Shragg P, Hollingsworth DR: Moderate caloric restric-207. Rayburn WF, McKean HE: Maternal perception of fetal tion in obese women with gestational diabetes. Obstet Gynecol movement and perinatal outcome. Obstet Gynecol 56:161, 65:487, 1985. 1980. 234. Magee MS, Knopp RH, Benedetti TJ: Metabolic effects of208. Holden KP, Jovanovic L, Druzin M, et al: Increased fetal activity 1200 kcal diet in obese pregnant women with gestational diabe- with low maternal blood glucose levels in pregnancies complicated tes. Diabetes 39:324, 1990. by diabetes. Am J Perinatol 1:161, 1984. 235. Rizzo T, Metzger BE, Burns WJ, et al: Correlations between209. Sadovsky E, Brjejinski A, Mor-Yosef S, et al: Fetal activity in antepartum maternal metabolism and intelligence of offspring. diabetic pregnancy. J Fetal Med 3:1, 1983. N Engl J Med 325:911, 1991.210. Landon MB, Gabbe SG, Sachs L: Management of diabetes mel- 236. Knopp RH, Magee MS, Raisys V, et al: Hypocaloric diets and litus and pregnancy: A survey of obstetricians and maternal-fetal ketogenesis in the management of obese gestational diabetic specialists. Obstet Gynecol 75:635, 1990. women. J Am Coll Nutr 10:649, 1991.
1010 Section VI Pregnancy and Coexisting Disease237. Peterson CM, Jovanovic-Peterson L: Percentage of carbohydrate 257. Lurie S, Insler V, Hagay Z: Induction of labor at 38 to 39 weeks of and glycemia response to breakfast, lunch, and dinner in women gestation reduces the incidence of shoulder dystocia in gestational with gestational diabetes. Diabetes 40:172, 1991. diabetic patient Class A2. Am J Perinatol 13:293, 1996.238. Jovanovic L: American Diabetes Association’s Fourth Interna- 258. Molsted-Pedersen L: Pregnancy and diabetes, a survey. Acta tional Workshop Conference on Gestational Diabetes Mellitus: Endocrinol 238:13, 1980. Summary and Discussion. Therapeutic Interventions. Diabetes 259. Fuhrmann K, Reiher H, Semmler K, et al: Prevention of congenital Care 2):B131, 1998. malformations in infants of insulin-dependent diabetic mothers.239. Goldberg J, Franklin B, Lasser L, et al: Gestational diabetes: Diabetes Care 6:219, 1983. Impact of home glucose monitoring on neonatal birth weight. 260. Kitzmiller JL, Gavin LA, Gin GD, et al: Preconception manage- Am J Obstet Gynecol 154:546, 1986. ment of diabetes continued through early pregnancy prevents240. Langer O, Rodriguez DA, Xenakis EMJ, et al: Intensiﬁed versus the excess frequency of major congenital anomalies in infants of conventional management of gestational diabetes. Am J Obstet diabetic mothers. JAMA 265:731, 1991. Gynecol 170:1036, 1994. 261. Miller E, Hare JW, Cloherty JP, et al: Elevated maternal HbA1241. Metzger BE, Coustan DR: Summary and recommendations of in early pregnancy and major congenital anomalies in infants of the Fourth International Workshop Conference on Gestational diabetic mothers. N Engl J Med 304:1331, 1981. Diabetes Mellitus. Diabetes Care 21:B161, 1998. 262. Greene MF: Prevention and diagnosis of congenital anomalies in242. Langer O, Brustman L, Anyaegbunam A, et al: Glycemic control diabetic pregnancies. Clin Perinatol 20:533, 1993. in gestational diabetes mellitus - how tight is tight enough; small 263. Mills J, Simpson JL, Drisoll SG, et al: Incidence of spontaneous for gestational age versus large for gestational age? Am J Obstet abortion among normal and insulin-dependent diabetic women Gynecol 161:645, 1989. whose pregnancies were identiﬁed within 21 days of conception.243. Langer O, Berkus M, Brustman L, et al: Rationale for insulin N Engl J Med 319:1617, 1988. management in gestational diabetes mellitus. Diabetes 40:186, 264. Whillhoite MB, Bennert HW, Palomaki GE, et al: The impact of 1991. preconception counseling on pregnancy outcomes. The experi-244. Buchanan TA, Kjos S, Schafer U, et al: Utility of fetal measurements ence of the Maine Diabetes in Pregnancy Program. Diabetes Care in the management of GDM. Diabetes Care 21:B99, 1998. 16:450, 1993.245. Langer O, Conway DL, Berkus MD, et al: A comparison of 265. Kjos SL, Ballagh SA, LaCour M, et al: The copper T380A intra- glyburide and insulin in women with gestational diabetes mellitus. uterine device in women with type II diabetes mellitus. Obstet N Engl J Med 343:1134, 2000. Gynecol 84:1006, 1994.246. Conway DL, Gonzales O, Skiver D: Use of glyburide for the 266. Steel JM, Duncan LJP: Serious complications of oral contra- treatment of gestational diabetes: the San Antonio experience. ceptives in insulin-dependent diabetes. Contraception 17:291, J Matern Fetal Neonatal Med 15:51, 2004. 1978.247. Kremer CJ, Duff P: Glyburide for the treatment of gestational 267. Lidegard O: Oral contraceptives, pregnancy, and the risk of cere- diabetes. Am J Obstet Gynecol 190:1438, 2004. bral thromboembolism: The inﬂuence of diabetes, hypertension,248. Chmait R, Dinise T, Moore T: Prospective observational study migraine and previous thrombotic disease. Br J Obstet Gynecol to establish predictors of glyburide success in women with gesta- 102:153, 1995. tional diabetes mellitus. J Perinatol 24:617, 2004. 268. Klein BEK, Moss SE, Klein R: Oral contraceptives in women with249. Jacobson GF, Ramos GA, Ching JY, et al: Comparison of glybu- diabetes. Diabetes Care 13:895, 1990. ride and insulin for the management of gestational diabetes in a 269. DePiaro R, Forte F, Bertoli A, et al: Changes in insulin recep- large managed care organization. Am J Obstet Gynecol 193:118, tors during oral contraception. J Clin Endocrinol Metab 52:29, 2005. 1981.250. Durnwald C, Landon MB: Glyburide: The new alternative 270. Steel JM, Duncan LJP: The effect of oral contraceptives on insulin for treating gestational diabetes? Am J Obstet Gynecol 193:1, requirements in diabetes. Br J Fam Plan 3:77, 1978. 2005. 271. Skouby S, Kuhl C, Molsted-Pederson L, et al: Triphasic oral con-251. Bung P, Artal R, Khodiguian N, Kjos S: Exercise in gestational traception: Metabolic effects in normal women and those with diabetes: an optional therapeutic approach? Diabetes 40:182, previous gestational diabetes. Am J Obstet Gynecol 163:495, 1991. 1985.252. Avery MD, Leon AS, Kopher RA: Effects of a partially home- 272. Kjos SL, Shoupe D, Dougan S, et al: Effect of low-dose oral con- based exercise program for women with gestational diabetes. traceptives on carbohydrate and lipid metabolism in women with Obstet Gynecol 89:10, 1997. recent gestational diabetes: results of a controlled randomized253. Landon MB, Gabbe SG: Antepartum fetal surveillance in gesta- prospective study. Am J Obstet Gynecol 163:1822, 1990. tional diabetes mellitus. Diabetes 34:50, 1985. 273. Liew DFM, Ng CSA, Yong YM, et al: Long term effects of254. Girz BA, Divon MY, Merkatz IR: Sudden fetal death in women Depo-Provera on carbohydrate and lipid metabolism. Contracep- with well controlled, intensively monitored gestational diabetes. tion 31:51, 1985. J Perinatol 12:229, 1992. 274. DeSlypere JP, Thiery N, Vermeulen A: Effect of long-term hor-255. Kjos S, Leung A, Henry OA, et al: Antepartum surveillance in monal contraception on plasma lipids. Contraception 31:633, diabetic pregnancies: predictors of fetal distress in labor. Am J 1985. Obstet Gynecol 173:1532, 1995. 275. Steel JM, Duncan LJP: The effect of oral contraceptives on insulin256. Lurie S, Matzkel A, Weissman A, et al: Outcome of pregnancy in requirements in diabetes. Br J Fam Plan 3:77, 1978. Class A1 and A2 gestational diabetic patients delivered beyond 276. Grenfel A, Brudnell JM, Doddridge MC, Watkins PJ: Pregnancy in 40 weeks gestation. Am J Perinatal 9:484, 1992. diabetic women who have proteinuria. Q J Med 59:379, 1986.