Epidemiology, Treatment, and Outcomes of                   Nosocomial Bacteremic Staphylococcus                   aureus P...
Epidemiology, Treatment, andOutcomes of Nosocomial BacteremicStaphylococcus aureus Pneumonia*C. Andrew DeRyke, PharmD; Tho...
one review3 of three major studies examining the               University Human Investigation Committee. This investigatio...
Data Collection                                                        antimicrobial therapy. To prevent bias, investigato...
model. The univariate predictors had to represent 10% of the             tive Staphylococcus sp. The most common concom-co...
Table 1—Characteristics of MSSA vs MRSA                      median length of time needed to switch to optimal            ...
Figure 3. Outcomes of NBSAP based on MSSA vs MRSA pneumonia.end of antimicrobial therapy for any treatment reg-           ...
Figure 4. Outcomes of NBSAP based on early vs delayed treatment.bacteremic staphylococcal pneumonia. The distribu-        ...
were not statistically significant, and an additional          not predict clinical success or decreased mortality.evaluat...
tients who will experience the greatest benefit from                       linezolid with vancomycin. Intensive Care Med 2...
Epidemiology, Treatment, and Outcomes of Nosocomial Bacteremic                 Staphylococcus aureus Pneumonia* C. Andrew ...
Upcoming SlideShare
Loading in …5

Epidemiology treatment and_outcomes_of_sa_nosocomial_pneumonia_chest_2005-1


Published on

Guias Norteamericanas para neumonia nosocomial.

Published in: Health & Medicine
1 Like
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Epidemiology treatment and_outcomes_of_sa_nosocomial_pneumonia_chest_2005-1

  1. 1. Epidemiology, Treatment, and Outcomes of Nosocomial Bacteremic Staphylococcus aureus Pneumonia* C. Andrew DeRyke, Thomas P. Lodise, Jr., Michael J. Rybak and Peggy S. McKinnon Chest 2005;128;1414-1422 DOI 10.1378/chest.128.3.1414 The online version of this article, along with updated information and services can be found online on the World Wide Web at: http://chestjournal.chestpubs.org/content/128/3/1414.full.html Chest is the official journal of the American College of Chest Physicians. It has been published monthly since 1935. Copyright2005by the American College of Chest Physicians, 3300 Dundee Road, Northbrook, IL 60062. All rights reserved. No part of this article or PDF may be reproduced or distributed without the prior written permission of the copyright holder. (http://chestjournal.chestpubs.org/site/misc/reprints.xhtml) ISSN:0012-3692Downloaded from chestjournal.chestpubs.org at 97185 Health Organization on April 17, 2011 © 2005 American College of Chest Physicians
  2. 2. Epidemiology, Treatment, andOutcomes of Nosocomial BacteremicStaphylococcus aureus Pneumonia*C. Andrew DeRyke, PharmD; Thomas P. Lodise, Jr., PharmD;Michael J. Rybak, PharmD, MS; and Peggy S. McKinnon, PharmD Objective: To describe outcomes associated with nosocomial bacteremic Staphylococcus aureus pneumonia (NBSAP) and to determine whether delay in adequate antimicrobial treatment is a risk factor for negative clinical and microbiological outcomes. Design: Retrospective cohort analysis. Setting: This study was conducted at Detroit Receiving Hospital and University Health Center, which is a 279-bed, level 1 trauma center in Detroit, MI. Patients: All episodes of NBSAP identified from January 1, 1999, to April 30, 2004. Results: Of 206 patients identified over a 5-year period with positive blood and respiratory cultures for S aureus, 60 patients met strict clinical, radiographic, and microbiological criteria for NBSAP. The overall mean ( SD) characteristics include the following: age, 55.5 15.0 years; acute physiology and chronic health evaluation II score, 20 (range, 3 to 41); ICU at onset, 93.3%; mechanical ventilation, 83.3%; length of stay (LOS) prior to NBSAP, 9 days (range, 2 to 81 days); methicillin-resistant S aureus (MRSA) rate, 70%; and all-cause hospital and infection-related mortality (IRM), 55.5% and 40.0%, respectively. Overall, S aureus pneumonia developed late in the patient’s hospital stay in ICU patients previously receiving mechanical ventilation and was associated with high crude mortality and IRM rates. No significant difference existed with respect to mortality or infection-related LOS between patients who had received early appropriate antibiotic therapy vs those who had received delayed appropriate antibiotic therapy at the onset of pneumonia or in patients with methicillin-sensitive S aureus pneumonia vs those with MRSA pneumonia. Conclusion: IRM from NBSAP is high, and standard therapies evaluated at the time of this study resulted in poor clinical outcomes. Delayed therapy was not found to be a predictor of adverse outcomes; however, this lack of ability to detect a difference may be a product of small sample size. These findings suggest that newer agents with enhanced clinical activity in NBSAP are needed and that these should be evaluated in a real-world setting, where outcomes of the most ill patients can be assessed. (CHEST 2005; 128:1414 –1422) Key words: bacteremia; clinical outcomes; critical care; cross-infection; hospital mortality; ICU; infection; staphylo- coccal infections; staphylococcal pneumonia; ventilator-associated pneumonia Abbreviations: APACHE acute physiology and chronic health evaluation; CART classification and regression tree analysis; IR-LOS infection-related length of stay; IRM infection-related mortality; NBSAP nosocomial bactere- mic Staphylococcus aureus pneumonia; MRSA methicillin-resistant Staphylococcus aureus; MSSA methicillin- sensitive Staphylococcus aureus; VAP ventilator-associated pneumoniaN osocomial pneumonia is infection and second most common hospital currently the is the monia ranges from 7.8 to 68.0%, and is influenced by the duration of hospital and ICU stay, the specificleading cause of death from hospital-acquired infec- diagnostic method used for pathogen detection, andtions.1 The incidence of acquiring nosocomial pneu- the patient population studied.1 The rate of nosoco- mial pneumonia secondary to Staphylococcus aureus*From the Anti-Infective Research Laboratory, Wayne State has increased steadily over the past 2 decades.2 InUniversity, Detroit, MI (Drs. Rybak and McKinnon); the De-partment of Pharmacy, Barnes Jewish Hospital, St. Louis, MO Reproduction of this article is prohibited without written permission(Dr. McKinnon); the Center for Anti-Infective Research and from the American College of Chest Physicians (www.chestjournal.Development, Hartford Hospital, Hartford, CT (Dr. DeRyke); org/misc/reprints.shtml).and Albany College of Pharmacy, Albany, NY (Dr. Lodise). Correspondence to: Peggy S. McKinnon, PharmD, Barnes-JewishManuscript received December 23, 2004; revision accepted Hospital, Mailstop 90 –52-411, 216 South Kings Hwy, St. Louis,January 26, 2005. MO 63110; e-mail: psm9154@bjc.org1414 Clinical Investigations Downloaded from chestjournal.chestpubs.org at 97185 Health Organization on April 17, 2011 © 2005 American College of Chest Physicians
  3. 3. one review3 of three major studies examining the University Human Investigation Committee. This investigationetiology of ventilator-associated pneumonia (VAP), S included all of the episodes of NBSAP identified from January 1, 1999, to April 30, 2004. Nosocomial pneumonia, or hospital-aureus was the most frequently isolated Gram-posi- acquired pneumonia, was defined as pneumonia occurring 48tive organism and the second-most isolated organism h after hospital admission and excluding any infection that wasonly behind Pseudomonas aeruginosa. Most studies incubating at the time of hospital admission.17estimate that S aureus accounts for 15 to 35% of all For the purposes of this investigation, the diagnosis of NBSAPnosocomial pneumonia cases; however, the true in- was defined based on clinical, radiographic, and microbiologicalcidence depends on many factors, such as patient criteria.18 –21 Within 72 h of the first positive culture, a chestdemographics, local susceptibility patterns, and radiograph must also have been abnormal, and the patient must have had signs and symptoms consistent with nosocomial pneu-methods of diagnosis.3 monia. In order to fulfill the requirement for bacteremic pneu- monia, at least one S aureus-positive blood culture not related to For editorial comment see page 1093 another source of infection and one S aureus-positive respiratory culture must have been obtained within 72 h of each other as Although there is increased recognition of S au- well. Possible respiratory cultures included positive growth in thereus as a major pathogen causing nosocomial pneu- culture of pleural fluid, positive sputum culture/tracheal aspiratemonia, there are few studies4 –7 with descriptive data (defined as secretions from the lungs, bronchi, or trachea thatspecifically evaluating patient outcomes of S aureus contain numerous or a moderate number of neutrophils and rarepneumonia. In addition, in the last decade, evidence or few squamous epithelial cells) findings, and a positive quanti- tative culture result from minimally contaminated lower respira-has accumulated demonstrating that initial inappro- tory tract specimens (eg, BAL fluid with 10,000 cfu/mL).priate antibiotic treatment is an important indepen- Radiographic criteria for pneumonia were met if the chestdent predictor of excess mortality in patients with radiograph yielded a new or progressive and persistent infiltrate,nosocomial pneumonia.8 –11 To our knowledge, no consolidation, or cavitation. Persistence of an infiltrate wasdata exist examining the impact of delayed appropri- defined as having the infiltrate present radiographically for atate antibiotic treatment specifically for bacteremic S least 72 h. Patients were defined as symptomatic if one of theaureus pneumonia. In a retrospective study12 evalu- following were present: fever (ie, temperature 38°C or 100.4°F) or hypothermia (ie, temperature 35°C or 95°F) withating S aureus bacteremia, a delay in treatment with no other recognized cause; leukopenia (ie, WBC count, 4,000antibiotics for 44.75 h was found to be an inde- cells/ L) or leukocytosis (ie, WBC count, 10,000 cells/ L); or,pendent predictor of infection-related mortality for adults 70 years old, altered mental status with no other(IRM) [adjusted odds ratio, 3.8; 95% confidence recognized cause. Patients also had to exhibit one of the followinginterval, 1.3 to 11.0; p 0.01]. It is unknown if this signs: new onset of purulent sputum, change in the character of44.75-h breakpoint is applicable to patients with the sputum, increased respiratory secretions, or increased suc-bacteremic S aureus pneumonia. Although the im- tioning requirements; new onset or worsening of cough, dyspnea, or tachypnea (respiratory rate, 25 breaths/min); or worseningpact of methicillin resistance on the outcomes of gas exchange (eg, O2 desaturation [Pao2/fraction of inspiredpatients with S aureus bacteremia has been exten- oxygen ratio of 240], increased oxygen requirements, or in-sively evaluated, little information exists on the im- creased ventilation demand).pact of the methicillin resistance of patients with Medical charts were screened to exclude the following possiblenosocomial bacteremic S aureus pneumonia (NBSAP). alternative causes for fever and radiographic chest densities. TheFurthermore, less information exists on the impact of presence of atelectasis was defined by the complete disappear-empirical antibiotic selection on NBSAP. Over the past ance of radiographic densities within 48 h of evaluation. Conges- tive heart failure with pulmonary edema was defined as afew years, studies13–16 have suggested that vancomycin resolution of pulmonary infiltrates following diuresis. A pulmo-may not be optimal for the treatment of S aureus nary embolism was defined by the presence of at least twopneumonia, especially in the subset of patients who segmental or larger mismatched perfusion abnormalities on ahave been infected with methicillin-resistant S aureus ventilation-perfusion scan or suggestive radiographic findings on(MRSA). To evaluate the epidemiology, treatment, and pulmonary angiography and spiral CT scan.outcomes of NBSAP, a retrospective cohort analysis Patients with endovascular infections, including endocarditis,was performed. Specifically, we examined the impact of were excluded because of the potential for hematogenous spread of S aureus to the lungs, thus complicating our retrospectivemethicillin resistance, empirical therapy, and delayed diagnosis of S aureus pneumonia. Patients with endocarditis weretreatment on the outcomes of patients with NBSAP. identified by transthoracic or transesophageal echocardiography and/or documentation of the diagnosis in the medical record. Materials and Methods Study DesignStudy Population To evaluate the epidemiology, treatment, and outcomes of NBSAP, a retrospective cohort analysis was performed. Specifi- This study was conducted at Detroit Receiving Hospital and cally, we examined the impact of methicillin resistance, empiricalUniversity Health Center, which is a 279-bed, level 1 trauma therapy, and delayed treatment on the outcomes of patients withcenter in Detroit, MI, and was approved by the Wayne State NBSAP.www.chestjournal.org CHEST / 128 / 3 / SEPTEMBER, 2005 1415 Downloaded from chestjournal.chestpubs.org at 97185 Health Organization on April 17, 2011 © 2005 American College of Chest Physicians
  4. 4. Data Collection antimicrobial therapy. To prevent bias, investigators involved in the outcome assessments were blinded to both susceptibility data Clinical Data: Data extracted from the patient medical records and treatment data, including the time to receipt of treatment.and pharmacy database included the following: age; sex; comor-bidities present; prior antibiotic use (within 30 days prior toNBSAP); length of hospitalization before the onset of nosocomial Definitionspneumonia (total hospitalization and hospitalization in the ICU);mechanical ventilation at the onset of nosocomial pneumonia; Clinical outcomes were evaluated as successes or failures.Charlson comorbidity index score22; and severity of illness based Clinical success included clinical resolution, which was defined ason APACHE (acute physiology and chronic health evaluation) II the complete resolution of all signs and symptoms of pneumoniascores at the time of admission to the ICU.23 If the patients were (return to preinfection baseline) along with improvement, or lacknot admitted to the ICU, the APACHE II score was determined of progression, of all abnormalities on the chest radiograph, andat hospital admission. clinical improvement, which was defined in patients if a partial The following comorbid conditions were documented: diabe- resolution of clinical signs and symptoms occurred such that notes mellitus; heart failure; COPD; asthma; hepatic dysfunction; additional antimicrobial therapy was required, along with therenal failure (as indicated by the necessity for dialysis); malig- improvement or stabilization of chest radiographic findings.nancy; HIV infection; alcoholism; presence of decubitus ulcers Patients who died or whose conditions did not improve were(stage II to IV); administration of immunosuppressive drugs (ie, deemed as having clinical failure. Clinical failure was defined asreceipt of 20 mg prednisone or an equivalent corticosteroid the persistence of clinical signs and symptoms, the persistence ofper day for 14 days before the onset of nosocomial pneumonia positive culture findings, and/or a lack of resolution of infiltratesor the receipt of any neoplastic chemotherapy in the 3 months on the chest radiograph.before the onset of nosocomial pneumonia); surgery requiring Microbiological outcomes were categorized as eradication or 48 h of hospitalization in the 30 days before the onset of persistence. Microbiological eradication included documentednosocomial pneumonia; and the presence of burns on 30% of eradication, defined as the elimination of baseline pathogensthe body surface area. based on subsequent negative blood and respiratory culture Microbiological Data: Collected microbiological data included findings, and presumed eradication, which was denoted in pa-all of the positive blood or respiratory culture findings, irrespec- tients in whom presumed eradication occurred based on clinicaltive of the pathogen identified. Susceptibility testing was per- resolution, but no subsequent cultures were obtained. Microbi-formed using the microtiter-well method, and the results were ological persistence included documented persistence, which wasinterpreted according to National Committee for Clinical Labo- defined as the persistence of bacteria despite the appropriate useratory Standards guidelines24 by the clinical microbiological of antibiotic therapy based on in vitro susceptibility results, andlaboratory. presumed persistence, which was denoted in patients in whom Treatment Data: All of the antimicrobial agents administered presumed microbiological failure occurred based on clinicalto provide activity against S aureus isolates were noted. Empiric failure, but no subsequent cultures were obtained.treatment was the first antibiotic regimen provided following the Death was considered to be related to NBSAP (ie, IRM) if oneonset of infection. Treatment was considered to be appropriate or more of the following criteria were present: (1) blood and/oron the basis of the following two factors: the timing of treatment respiratory cultures were positive for S aureus at the time ofrelative to the first positive blood or respiratory culture finding; death; (2) death occurred before the resolution of signs andand the in vitro susceptibility of the blood or respiratory isolate. symptoms of the nosocomial pneumonia; (3) death occurredTiming was evaluated based on a previously described breakpoint 14 days after the onset of nosocomial pneumonia withoutof 44.75 h, which was determined by classification and regression another explanation; (4) autopsy findings indicated pneumonia astree analysis (CART) as an independent predictor of mortality in a cause of death; and (5) pneumonia was indicated as a cause ofS aureus bactermia.12 If a patient had received at least one IV death on the death certificate.antibiotic to which the S aureus blood or respiratory isolate was The calculation of IR-LOS was measured from the time whensusceptible and the antibiotic had been administered within the first positive blood or respiratory culture finding was col-44.75 h, it was considered to be appropriate early treatment. For lected until the end of antimicrobial treatment, death, or hospitalexample, an individual with MRSA bacteremia receiving vanco- discharge. The calculation of IR-LOS excluded patients who diedmycin within 44.75 h would be classified as having received early secondary to nosocomial pneumonia.appropriate treatment. A patient with MRSA initially treated with An infection-related cost of hospitalization was determined fora -lactam but not receiving vancomycin within 44.75 h would be each patient. The Detroit Receiving Hospital accounting depart-considered as having received delayed treatment. In addition, we ment supplied the cost figures. For patients who lived or did notutilized CART analysis to determine whether there was a differ- die secondary to NBSAP, the cost was calculated from the onsetent breakpoint to better describe the critical time to appropriate of infection until the last day that antimicrobial therapy activeantibiotic treatment for this cohort of NBSAP patients. against S aureus was administered (ie, the IR-LOS). For those patients who died because of NBSAP, the cost was calculated from the onset of infection until the day of the patient’s death.Outcome Assessment The following primary end points were assessed: (1) IRM; (2) Statistical Analysishospital (crude) mortality; and (3) infection-related length of staylength (IR-LOS) after the onset of NBSAP. Secondary outcomes Categoric variables were compared by the Pearson 2 test orincluded the following: (1) clinical response; (2) microbiological Fisher exact test, and continuous variables were compared by theresponse; and (3) cost of hospitalization after the onset of Student t test or Mann-Whitney U test. Multivariate analysesNBSAP. were performed to determine the independent association of For the early treatment vs delayed treatment analyses, patients antibiotic resistance and delayed treatment with the clinicalwho died within 72 h of the onset of infection were excluded. outcome of interest while adjusting for confounding variables.Clinical and microbiological outcomes were assessed at the Clinical features significantly associated with the outcome byfollowing four time points: day 3; day 7; day 10; and at the end of univariate analysis were included in the explanatory multivariate1416 Clinical Investigations Downloaded from chestjournal.chestpubs.org at 97185 Health Organization on April 17, 2011 © 2005 American College of Chest Physicians
  5. 5. model. The univariate predictors had to represent 10% of the tive Staphylococcus sp. The most common concom-cohort to prevent overfitting of the multivariate model. Dichot- itant respiratory organisms were P aeruginosa (nineomous outcomes (eg, IRM and clinical status) were analyzed withstandard logistic regression. A p value of 0.05 was considered cases); A baumannii (eight cases); Klebsiella sppto be significant for two-tailed tests. A statistical software package (seven cases); Escherichia coli (six cases); Candida(SPSS, version 10.0; SPSS; Chicago, IL) was used for all of the albicans (four cases); Enterobacter spp (three cases);calculations. and nine others. The clinical and microbiological success rates at the end of treatment were 56.7% and 53.3%, respec- Results tively. Thirty-three patients (55.5%) died during hospitalization, and 24 (40.0%) died secondary toBaseline Data NBSAP (ie, IRM). Nine patients died for reasons not Of the 206 patients identified either by Interna- attributable to the pneumonia. In seven cases, treat-tional Classification of Diseases, ninth revision, codes ment for NBSAP had stopped at least 2 weeks beforefor S aureus pneumonia and/or by concomitant the patient died. Withdrawal from support occurredblood and respiratory cultures positive for S aureus, in two patients irrespective of the concomitant pneu-60 met the strict inclusion criteria as assessed for monia.NBSAP. The primary reasons for exclusion from the The relationship between APACHE II score andstudy included a lack of clinical or radiologic findings death is shown in Figure 1. A linear relationshipsupporting a pneumonia diagnosis (n 46), alter- existed demonstrating that patients who were morenate clinical diagnoses (n 43), infection that was acutely ill at the time of admission to the ICU had anot nosocomial (n 26), time correlation between greater mortality rate (R2 0.74).microbiological findings (n 25), and chart notavailable (n 6). Alternative clinical diagnoses po- Outcomes of Bacteremic MRSA vs Methicillin-tentially causing the infection isolated more than Sensitive S aureus Pneumonia MSSAPonce included the following: endocarditis (n 12);IV line-related sepsis (n 16); alternative pathogen Of the 60 patients, 42 patients (70%) were in-most likely caused infection (n 5); and pelvic fected with MRSA. A comparison of clinical featuresabscess (n 2). Overall, the mean ( SD) age was between MRSA and methicillin-sensitive S aureus55.5 15.0 years, and the median APACHE II score (MSSA) patients is presented in Table 1. By univar-was 20 (range, 3 to 41). Most patients were men iate analysis, MRSA patients were more likely than(66.7%), and were predominantly African American MSSA patients to have received prior antibiotic(56.7%) and white (38.3%). An equal number of treatment and had a longer median length of staypatients were admitted to surgery (41.7%) and med- prior to the onset of NBSAP. In addition, a signifi-icine services (41.7%), and 16.7% of the patients cant difference existed between the two groups withwere admitted to the burn unit. Twenty-three pa- respect to appropriate treatment. In the MSSAtients (38.3%) underwent surgery 1 month prior group, 72.2% of patients received appropriate ther-to the development of NBSAP. Bilateral infiltrates apy within 44.75 h of the onset of infection com-were evident on chest radiography in 75% of the pared with only 50% of patients in the MRSApatients identified. Most patients were in the ICU attime of onset of NBSAP (93.3%), and 83.3% ofpatients were receiving mechanical ventilation. Themedian length of stay prior to the onset of NBSAPwas 9 days (range, 2 to 81 days). The origin of the respiratory cultures used for themicrobiological determination of pneumonia camefrom aspirated sputum in 90% of the cases, and fromBAL fluid in 10% of cases. Of the 60 patients, 42patients (70%) were infected with MRSA. Forty-fourpatients (73.3%) had concomitant organisms (6 inblood only, 7 in blood and respiratory cultures, and31 in respiratory cultures only). The most commonconcomitant organisms in the blood were as follows:Enterococcus spp (five cases); Acinetobacter bau-mannii (three cases); Streptococcus viridans (two Figure 1. IRM grouped by APACHE II score. APACHE IIcases); Klebsiella pneumoniae (one case); Streptococ- scores grouped in values of 10. Dotted line represented linearcus pneumoniae (one case); and one cogulase-nega- regression line.www.chestjournal.org CHEST / 128 / 3 / SEPTEMBER, 2005 1417 Downloaded from chestjournal.chestpubs.org at 97185 Health Organization on April 17, 2011 © 2005 American College of Chest Physicians
  6. 6. Table 1—Characteristics of MSSA vs MRSA median length of time needed to switch to optimal Pneumonia therapy for the eight patients who had been empir- MSSA MRSA ically treated with vancomycin was 71.5 h (range, 16 Characteristics (n 18) (n 42) p Value to 123 h). Of the 42 MRSA patients, 24 (57.1%)Age, yr 55.2 17.6 58.4 13.9 0.45 received empiric vancomycin treatment, 8 (19.0%)APACHE II score 20.5 (5–41) 20.0 (3–36) 0.72 received -lactam agents, 6 (14.3%) received clinda-Charlson index score 2 (0–6) 2 (0–7) 0.53 mycin, 3 (7.1%) received trimethoprim/sulfamethox-Diabetes 3 (16.7) 15 (35.7) 0.14 azole, 1 was started on therapy with levofloxacin; andDecubitus ulcers 0 7 (16.7) 0.09 1 patient died before receiving appropriate antibioticPrior antibiotics 3 (16.7) 30 (71.4) 0.01Vent at NBSAP onset 13 (72.2) 37 (88.1) 0.13 treatment. Once susceptibility data were available,LOS prior to NBSAP onset, d 4 (2–31) 11.5 (4–81) 0.03 40 patients (95.2%) received vancomycin for primaryICU LOS prior to NBSAP 3 (0–31) 9 (0–58) 0.059 antimicrobial treatment of pneumonia. Two patients onset, d were never appropriately treated with antibiotics.Time to adequate therapy, h 24 (8–51) 44 (1–149) 0.013 Nine patients (50.0%) in the MSSA subgroup were*Values given as mean SD, No. of patients (%), or median (range), treated with combination therapy compared with unless otherwise indicated. LOS length of stay; Vent receiving seven patients (16.7%) in the MRSA subgroup. The mechanical ventilation. median duration of combination therapy was 3 days in both groups (MSSA group range, 1 to 10 days; MRSA group range, 1 to 8 days). Eleven vancomycinsubgroup (p 0.01) [Fig 2]. Furthermore, MSSA trough levels were collected in a total of nine patientspneumonia patients received appropriate treatment with a median value of 17.4 g/mL (range, 9.2 toin significantly less time than did MRSA pneumonia 26.7 g/mL).patients (24 h [range, 8 to 51 h] vs 44 h [range, 1 to The comparison values of hospital mortality, IRM,149 h], respectively; p 0.013). No difference was and IR-LOS between MSSA and MRSA patients aredetected in the duration of therapy between the two displayed in Figure 3. No significant differences ingroups (MSSA group, 10 days [range, 2.5 to 25.0 these primary end points were observed between thedays]; MRSA group, 11 days [range, 1.0 to 21.5 groups. Similarly, no significant differences in hos-days]). pital mortality, IRM, and IR-LOS were noted for the Of the 18 MSSA patients, 8 patients (44.4%) were different empiric antibiotic regimens, stratified byempirically treated with vancomycin, 7 patients methicillin susceptibility. Multivariate analyses were(38.9%) were empirically treated with a -lactam, 1 performed, and the associations between methicillinpatient was empirically treated with clindamycin, 1 susceptibility and outcomes were identical to thosepatient was empirically treated with levofloxacin, and from the univariate analyses.1 patient died prior to identification of the causative IRM was higher in the empiric vancomycin grouporganism and subsequent appropriate antibiotic than in the empiric -lactam group; however, thistherapy. Once sensitivity reports were obtained, was not statistically significant. For the 18 patients77.8% of the MSSA patients received a -lactam infected with MSSA, the hospital mortality and IRMagent as the primary antimicrobial treatment. The rates in the patients receiving empiric -lactams were 57.1% and 28.6%, respectively, compared with 62.5% and 50.0%, respectively, in the empiric van- comycin group. For the 42 patients infected with MRSA, the hospital mortality and IRM rates in patients receiving empiric vancomycin were 50.0% and 45.8%, respectively, compared with 62.5% and 25.0%, respectively, in the empiric -lactam group. These findings, however, were not statistically signif- icant, and an additional evaluation of this subset revealed that patients who had been empirically treated with -lactams tended to be younger (me- dian age, 52 vs 56 years, respectively) and to have lower APACHE II scores (median, 17 vs 21, respec- tively). Overall clinical success was achieved in 59.5% of MRSA patients compared with 50.0% of MSSAFigure 2. Receipt of adequate treatment within the 44.75-h patients. No differences in clinical or microbiologicalbreakpoint established for S aureus bacteremia. success were determined at days 3, 7, or 10, or at the1418 Clinical Investigations Downloaded from chestjournal.chestpubs.org at 97185 Health Organization on April 17, 2011 © 2005 American College of Chest Physicians
  7. 7. Figure 3. Outcomes of NBSAP based on MSSA vs MRSA pneumonia.end of antimicrobial therapy for any treatment reg- no significant difference in the rate of response wasimen, stratified by methicillin susceptibility. Simi- seen at days 3, 7, and 10, and at the end oflarly, because 83.3% of patients were receiving me- antimicrobial therapy or the first day to clinicalchanical ventilation at the time of onset, no improvement (5 days [range, 2 to 14 days] vs 5 daysdifference in clinical response was observed in this [range, 2 to 11 days], respectively). In the CARTVAP subset of patients compared with the whole analysis that was performed to identify a specificcohort. Of note, the overall clinical cure rate in “time to therapy” for the bacteremic S aureus pneu-patients receiving vancomycin was 56.3%; 58.3% in monia cohort, no additional time breakpoint in ap-the subset of MRSA patients. propriate antibiotic treatment was found that pro- duced an increased probability of IRM.Outcomes in Delayed vs Early Appropriate The evaluation of the appropriateness of therapyTherapy for concomitant pathogens was also assessed. Alter- native pathogens were appropriately treated 85% of Five patients were excluded in the appropriate the time based on the susceptibility profile. Thetherapy analyses secondary to death within 72 h of clinical success was 59.5% for patients treated ap-the onset of infection. Of the remaining 55 patients, propriately for both NBSAP and the alternative24 patients (43.6%) did not receive appropriate pathogen. This result was similar to the clinicalantibiotic treatment within 44.75 h of the onset of success rate observed for the entire cohort (56.7%).infection, and 31 patients (56.4%) received appro- Complete, infection-related cost data were avail-priate antibiotic treatment within 44.75 h of the able for 22 of the 60 patients. The greatest reason foronset of infection. There were no significant differ- increased cost was length of hospitalization. In theences between the two groups (delayed vs early) with patients who lived, the median total cost was $35,072respect to the APACHE II score at time of admission (range, $19,764 to $312,511) compared with $22,098to the ICU (20.5 [range, 5 to 41] vs 18 [range, 4 to (range, $1,218 to $66,351) in the patients who died.40], respectively; p 0.6) and Charlson comorbidity No differences in cost were evident in patients basedindex score (2 [range, 0 to 6] vs 2 [range, 0 to 6], on methicillin susceptibility or early empiric treat-respectively; p 0.9). The median time to the start ment.of appropriate treatment was 68.5 h (range, 45 to149 h) in the delayed-treatment group and 25 h(range, 4 to 44 h) in the early-treatment group. Only Discussion16 patients received appropriate antibiotic treatmentwithin 24 h of the onset of infection The previous literature specifically focusing on S No differences existed with respect to the primary aureus nosocomial pneumonia is limited. Our reviewoutcomes based on the receipt of delayed vs early of 60 patients is one of the largest reports of aappropriate antibiotic therapy (Fig 4). In addition, real-world experience describing the outcomes ofwww.chestjournal.org CHEST / 128 / 3 / SEPTEMBER, 2005 1419 Downloaded from chestjournal.chestpubs.org at 97185 Health Organization on April 17, 2011 © 2005 American College of Chest Physicians
  8. 8. Figure 4. Outcomes of NBSAP based on early vs delayed treatment.bacteremic staphylococcal pneumonia. The distribu- pneumonia patients. In the 55 patients with bacteri-tion of pathogens responsible for nosocomial pneu- ologically evaluable S aureus pneumonia, the overallmonia differs depending on factors such as the clinical success rate in vancomycin-treated patientslength of hospital admission before onset of the was 50.9% compared with 44.4% in the 18 patientsdisease, admission to the ICU, and duration of infected with MRSA. Stevens et al14 specificallymechanical ventilation.25 S aureus pneumonia typi- evaluated therapy with linezolid vs therapy withcally develops in patients who have had a longer vancomycin for the treatment of MRSA infections.length of hospital stay before onset of disease, have Data from 29 patients with MRSA pneumonia re-received mechanical ventilation for 5 days (ie, vealed that a clinical cure was achieved in 75.0% oflate-onset VAP), and have been exposed to previous patients treated with either agent. In a study analyz-antimicrobial therapy. A prediction model26 has ing data from two previous double-blind studies ofbeen developed to determine which characteristics patients with S aureus nosocomial pneumonia, Wun-predict for the development of MRSA in patients derink et al15 found that there was no difference inwith S aureus bacteremia. The greatest risk factor clinical cure between the use of linezolid and van-was previous antibiotic exposure with an odds ratio comycin among all of the S aureus nosocomialof 9.2 (95% confidence interval, 4.8 to 17.9). Rello pneumonia patients (51.5% vs 43.4%, respectively;and Diaz3 also found previous antimicrobial therapy p 0.182). However, in the subset of 133 MRSAto be a risk factor for MRSA pneumonia as well. Our nosocomial pneumonia patients, a significant differ-findings for this cohort of 60 patients concur with ence was found (59% vs 35.5%, respectively;this description. The median length of hospital stay p 0.009). Kollef et al16 examined clinical cures inbefore the first positive culture finding was 9 days. patients who specifically had VAP and found signif-Most of our patients (93.3%) were in the ICU icant differences in favor of the empiric use ofreceiving mechanical ventilation (83.3%) at the onset linezolid over vancomycin in both the S aureus VAPof disease, and 55% of patients received antibiotic group (48.9% vs 35.2%, respectively; p 0.06) andtherapy prior to the onset of pneumonia. Further- the MRSA VAP group (62.2% vs 21.2%, respectively;more, 71.4% of patients who developed MRSA p 0.001). Our findings were similar to these re-pneumonia had received previous antibiotic therapy ported values. Overall clinical success was achievedwithin 30 days prior to the onset of nosocomial in 56.7% of all patients and in 59.5% of the subset ofpneumonia. This was significantly different from the MRSA patients compared with 50.0% of the subset16.7% of MSSA pneumonia patients (p 0.01). of MSSA patients. Highlighting the empiric use of Clinical success has been reported in S aureus vancomycin, the overall clinical success rate wasnosocomial pneumonia in a number of trials, mostly 56.3%, with 58.3% cure rates in the MRSA patientcomparing vancomycin with the newer agents quinu- population. Among patients with MRSA, the IRMpristin/dalfopristin and linezolid. Fagon et al27 com- rate was higher in patients receiving empiric vanco-pared the efficacy of therapy with quinupristin/ mycin (45.8%) compared with that in patients receiv-dalfopristin and vancomycin in Gram-positive ing empiric -lactam agents (25.0%). These findings1420 Clinical Investigations Downloaded from chestjournal.chestpubs.org at 97185 Health Organization on April 17, 2011 © 2005 American College of Chest Physicians
  9. 9. were not statistically significant, and an additional not predict clinical success or decreased mortality.evaluation of this subset of patients revealed that Another reason may be in part attributable to thepatients empirically treated with -lactam agents slow activity or lack of efficacy of vancomycin, suchtended to be younger and to have lower APACHE II that even the early administration of this agent isscores, partially explaining this finding. Because associated with poor clinical outcomes.31,32 80% of patients were already receiving mechani- The greatest limitation of our study was the smallcal ventilation at the onset of NBSAP, these percent- number of patients (n 60) who met our strictages did not change after analyzing this subset of inclusion criteria. The main reason that so fewpatients. patients were available for inclusion over the 5 years Patients included in this analysis were critically ill was the requirement for patients to have concomi-and had numerous underlying chronic comorbidities, tant S aureus bacteremia, as well as pneumonia. Thisas evidenced by a median APACHE II score of 20 requirement was essential because of the retrospec-(range, 3 to 41). Consequently, the mortality rate was tive design of our investigation and the high rate ofhigh in our study, with an associated crude mortality tracheal aspirates used as the respiratory source ofrate of 55.5% and an IRM rate of 40.0%. Thesevalues also correlate well with those from previous this pathogen. We wanted to be confident that theliterature6,7,28 –30 in which the mortality rate attribut- patients who were involved in this analysis developedable to S aureus has ranged between 28% and 50%. bacteremia secondary to nosocomial pneumonia andThe mortality rate among nosocomial acquired bac- not by hematogenous spread of S aureus, as canteremic pneumonia patients was found to be 52.5% occur in other conditions, such as endocarditis,in a previous study.5 which was one of the exclusion criteria. The most surprising outcomes from our results Additional limitations were inherent to the retro-were a lack of difference with respect to overall spective design of the study. Characterizing NBSAPhospital mortality rate, IRM rate, and IR-LOS based was difficult, although stringent inclusion criteriaon the receipt of early vs delayed antibiotic therapy. were used. Many patients who likely had nosocomialNumerous reports8 –11 have been published stating S aureus pneumonia were excluded from the analysisthat mortality significantly increases in patients who because of the lack of a concomitant positive blooddevelop pneumonia if empiric antibiotic therapies culture finding. Also, numerous comorbidities con-are not started at the time of clinical suspicion, founded the accurate assessment of IRM. To ac-before cultures have been collected and sensitivity count for this, an independent reviewer with nohas been reported. Our lack of ability to detect a knowledge of the culture and sensitivity data or ofdifference may be a product of a small sample size the choice and timing of antibiotic therapy wasleading to a potential type II error. In a previous designated to determine the primary outcome basedstudy12 examining S aureus bacteremia, a delay in on clinical data recorded throughout the course oftherapy of 44.75 h was found to be associated with the patient’s hospital stay.a 3.8 times higher rate of IRM. The 44.75-h break- In conclusion, this investigation includes one ofpoint for appropriate therapy, however, was not the largest collections of patients to have bacteremicpredictive in this series of pneumonia patients. In nosocomial pneumonia that was specifically attribut-addition, an additional breakpoint of time to appro- able to S aureus. S aureus pneumonia developed latepriate therapy was not identified using CART. Some in the patient’s hospital stay and in ICU patients whopotential reasons for no differences being detected had previously received mechanical ventilation, andare the low number of patients (n 60) and the high was associated with a high crude mortality and IRMmortality rate (55%) associated with this series. In rates. No significant differences were detected withaddition, only 16 patients received appropriate anti- respect to crude mortality, IRM, or IR-LOS betweenbiotic therapy within 24 h of the onset of infection. It patients who developed MSSA vs those who devel-may be difficult to ascertain the impact of delayed oped MRSA pneumonia or depending on whethertherapy given the high rate of delayed treatment. the patient had received early vs delayed appropriate Of interest, as demonstrated in Figure 1, a linear antibiotic therapy. The small sample size may haverelationship existed with respect to IRM and increas- precluded our ability to detect a difference in theseing APACHE II score. Because of the requirement groups. Patients who were empirically treated withof concomitant bacteremia, this subset of pneumonia vancomycin for MSSA and MRSA pneumonia expe-patients likely represents those patients who are rienced a high mortality rate. These findings suggestmost severely ill with rapid disease progression. The that newer agents with enhanced clinical activity insevere illness exhibited in these patients may be the NBSAP are needed and that these should be evalu-main factor determining mortality and may serve as ated in a real-world setting, where outcomes of thethe reason why appropriate antibiotic treatment did sickest patients can be assessed. The group of pa-www.chestjournal.org CHEST / 128 / 3 / SEPTEMBER, 2005 1421 Downloaded from chestjournal.chestpubs.org at 97185 Health Organization on April 17, 2011 © 2005 American College of Chest Physicians
  10. 10. tients who will experience the greatest benefit from linezolid with vancomycin. Intensive Care Med 2004; 30:these newer agents remains to be determined. 388 –394 17 American Thoracic Society. Hospital-acquired pneumonia in adults: diagnosis, assessment of severity, initial antimicrobial therapy, and preventive strategies: a consensus statement; References American Thoracic Society, November 1995. Am J Respir 1 Hoffken G, Niederman MS. Nosocomial pneumonia: the Crit Care Med 1996; 153:1711–1725 importance of a de-escalating strategy for antibiotic treatment 18 Micek ST, Ward S, Fraser VJ, et al. A randomized controlled of pneumonia in the ICU. Chest 2002; 122:2183–2196 trial of an antibiotic discontinuation policy for clinically 2 Lynch JP III. Hospital-acquired pneumonia: risk factors, suspected ventilator-associated pneumonia. Chest 2004; 125: microbiology, and treatment. Chest 2001; 119:373S–384S 1791–1799 3 Rello J, Diaz E. Pneumonia in the intensive care unit. Crit 19 Rubinstein E, Cammarata S, Oliphant T, et al. Linezolid Care Med 2003; 31:2544 –2551 (PNU-100766) versus vancomycin in the treatment of hospi- 4 Gonzalez C, Rubio M, Romero-Vivas J, et al. Bacteremic talized patients with nosocomial pneumonia: a randomized, pneumonia due to Staphylococcus aureus: a comparison of double-blind, multicenter study. Clin Infect Dis 2001; 32: disease caused by methicillin-resistant and methicillin-sus- 402– 412 ceptible organisms. Clin Infect Dis 1999; 29:1171–1177 20 Beck KD, Gastmeier P. Clinical or epidemiologic diagnosis of 5 Gonzalez C, Rubio M, Romero-Vivas J, et al. Staphylococcus nosocomial pneumonia: is there any difference? Am J Infect aureus bacteremic pneumonia: differences between commu- Control 2003; 31:331–335 nity and nosocomial acquisition. Int J Infect Dis 2003; 21 Garner JS, Jarvis WR, Emori TG, et al. CDC definitions for 7:102–108 nosocomial infections, 1988. Am J Infect Control 1988; 6 Rello J, Torres A, Ricart M, et al. Ventilator-associated 16:128 –140 pneumonia by Staphylococcus aureus: comparison of methi- 22 Charlson ME, Pompei P, Ales KL, et al. A new method of cillin-resistant and methicillin-sensitive episodes. Am J Respir classifying prognostic comorbidity in longitudinal studies: Crit Care Med 1994; 150:1545–1549 development and validation. J Chronic Dis 1987; 40:373–383 7 Combes A, Luyt CE, Fagon JY, et al. Impact of methicillin 23 Knaus WA, Draper EA, Wagner DP, et al. APACHE II: a resistance on outcome of Staphylococcus aureus ventilator- severity of disease classification system. Crit Care Med 1985; associated pneumonia. Am J Respir Crit Care Med 2004; 13:818 – 829 170:786 –792 24 National Committee for Clinical Laboratory Standards. 8 Luna CM, Vujacich P, Niederman MS, et al. Impact of BAL Methods for dilution antimicrobial susceptibility tests for data on the therapy and outcome of ventilator-associated bacteria that grow aerobically: approved standard M7–A6. pneumonia. Chest 1997; 111:676 – 685 Wayne, PA: National Committee for Clinical Laboratory 9 Rello J, Gallego M, Mariscal D, et al. The value of routine Standards, 2004 microbial investigation in ventilator-associated pneumonia. 25 Chastre J, Fagon JY. Ventilator-associated pneumonia. Am J Am J Respir Crit Care Med 1997; 156:196 –200 Respir Crit Care Med 2002; 165:867–90310 Kollef MH, Ward S. The influence of mini-BAL cultures on 26 Lodise TP Jr, McKinnon PS, Rybak M. Prediction model to patient outcomes: implications for the antibiotic management identify patients with Staphylococcus aureus bacteremia at of ventilator-associated pneumonia. Chest 1998; 113:412– 420 risk for methicillin resistance. Infect Control Hosp Epidemiol11 Alvarez-Lerma F. Modification of empiric antibiotic treat- 2003; 24:655– 661 ment in patients with pneumonia acquired in the intensive 27 Fagon J, Patrick H, Haas DW, et al. Treatment of Gram- care unit: ICU-Acquired Pneumonia Study Group. Intensive positive nosocomial pneumonia: prospective randomized Care Med 1996; 22:387–394 comparison of quinupristin/dalfopristin versus vancomycin;12 Lodise TP, McKinnon PS, Swiderski L, et al. Outcomes Nosocomial Pneumonia Group. Am J Respir Crit Care Med analysis of delayed antibiotic treatment for hospital-acquired 2000; 161:753–762 Staphylococcus aureus bacteremia. Clin Infect Dis 2003; 28 Kaye MG, Fox MJ, Bartlett JG, et al. The clinical spectrum of 36:1418 –1423 Staphylococcus aureus pulmonary infection. Chest 1990;13 Sandiumenge A, Diaz E, Bodi M, et al. Therapy of ventilator- 97:788 –792 associated pneumonia: a patient-based approach based on the 29 Musher DM, Lamm N, Darouiche RO, et al. The current ten rules of “The Tarragona Strategy.” Intensive Care Med spectrum of Staphylococcus aureus infection in a tertiary care 2003; 29:876 – 883 hospital. Medicine (Baltimore) 1994; 73:186 –20814 Stevens DL, Herr D, Lampiris H, et al. Linezolid versus 30 Iwahara T, Ichiyama S, Nada T, et al. Clinical and epidemi- vancomycin for the treatment of methicillin-resistant Staph- ologic investigations of nosocomial pulmonary infections ylococcus aureus infections. Clin Infect Dis 2002; 34:1481– caused by methicillin-resistant Staphylococcus aureus. Chest 1490 1994; 105:826 – 83115 Wunderink RG, Rello J, Cammarata SK, et al. Linezolid vs 31 Levine DP, Fromm BS, Reddy BR. Slow response to vanco- vancomycin: analysis of two double-blind studies of patients mycin or vancomycin plus rifampin in methicillin-resistant with methicillin-resistant Staphylococcus aureus nosocomial Staphylococcus aureus endocarditis. Ann Intern Med 1991; pneumonia. Chest 2003; 124:1789 –1797 115:674 – 68016 Kollef MH, Rello J, Cammarata SK, et al. Clinical cure and 32 Small PM, Chambers HF. Vancomycin for Staphylococcus survival in Gram-positive ventilator-associated pneumonia: aureus endocarditis in intravenous drug users. Antimicrob retrospective analysis of two double-blind studies comparing Agents Chemother 1990; 34:1227–12311422 Clinical Investigations Downloaded from chestjournal.chestpubs.org at 97185 Health Organization on April 17, 2011 © 2005 American College of Chest Physicians
  11. 11. Epidemiology, Treatment, and Outcomes of Nosocomial Bacteremic Staphylococcus aureus Pneumonia* C. Andrew DeRyke, Thomas P. Lodise, Jr., Michael J. Rybak and Peggy S. McKinnon Chest 2005;128; 1414-1422 DOI 10.1378/chest.128.3.1414 This information is current as of April 17, 2011Updated Information & ServicesUpdated Information and services can be found at:http://chestjournal.chestpubs.org/content/128/3/1414.full.htmlReferencesThis article cites 31 articles, 17 of which can be accessed free at:http://chestjournal.chestpubs.org/content/128/3/1414.full.html#ref-list-1Cited BysThis article has been cited by 4 HighWire-hosted articles:http://chestjournal.chestpubs.org/content/128/3/1414.full.html#related-urlsPermissions & LicensingInformation about reproducing this article in parts (figures, tables) or in its entirety can befound online at:http://www.chestpubs.org/site/misc/reprints.xhtmlReprintsInformation about ordering reprints can be found online:http://www.chestpubs.org/site/misc/reprints.xhtmlCitation AlertsReceive free e-mail alerts when new articles cite this article. To sign up, select the"Services" link to the right of the online article.Images in PowerPoint formatFigures that appear in CHEST articles can be downloaded for teaching purposes inPowerPoint slide format. See any online figure for directions. Downloaded from chestjournal.chestpubs.org at 97185 Health Organization on April 17, 2011 © 2005 American College of Chest Physicians