Staphylococcus aureus Native Valve Infective Endocarditis: Report of 566 Episodes from the International Collaboration on Endocarditis Merged Database

Transcription

1 MAJOR ARTICLE Staphylococcus aureus Native Valve Infective Endocarditis: Report of 566 Episodes from the International Collaboration on Endocarditis Merged Database José M. Miro, 1 Ignasi Anguera, 2 Christopher H. Cabell, 7 Anita Y. Chen, 7 Judith A. Stafford, 7 G. Ralph Corey, 7 Lars Olaison, 3 Susannah Eykyn, 4 Bruno Hoen, 5 Elias Abrutyn, 8 Didier Raoult, 6 Arnold Bayer, 9 Vance G. Fowler, Jr., 7 and the International Collaboration on Endocarditis Merged Database Study Group a 1 Hospital Clinic Institut d Investigacions Biomèdiques August Pi i Sunyer, University of Barcelona, and 2 Corporació Sanitària Parc Taulí Hospital de Sabadell, Barcelona, Spain; 3 Sahlgrenska University Hospital, Göteborg, Sweden; 4 St. Thomas Hospital, London, United Kingdom; 5 Hôpital Saint-Jacques, Besançon, and 6 Unite des Rickettsies, Marseille, France; 7 Duke University, Durham, North Carolina; 8 Drexel University College of Medicine and School of Public Health, Philadelphia, Pennsylvania; 9 LA Biomedical Research Institute at Harbor University of California Los Angeles Medical Center, Torrance, California Background. Staphylococcus aureus native valve infective endocarditis (SA-NVIE) is not completely understood. The objective of this investigation was to describe the characteristics of a large, international cohort of patients with SA-NVIE. Methods. The International Collaboration on Endocarditis Merged Database (ICE-MD) is a combination of 7 existing electronic databases from 5 countries that contains data on 2212 cases of definite infective endocarditis (IE). Results. Of patients with native valve IE, 566 patients (34%) had IE due to S. aureus, and 1074 patients had IE due to pathogens other than S. aureus (non SA-NVIE). S. aureus IE were more likely to die (20% vs. 12%; P!.001), to experience an embolic event (60% vs. 31%; P!.001), or to have a central nervous system event (20% vs. 13%; P!.001) and were less likely to undergo surgery (26% vs. 39%; P!.001) than were patients with non SA-NVIE. Multivariate analysis of prognostic factors of mortality identified age (odds ratio [OR], 1.4; 95% confidence interval [CI], ), periannular abscess (OR, 2.4; 95% CI, ), heart failure (OR, 3.9; 95% CI, ), and absence of surgical therapy (OR, 2.3; 95% CI, ) as variables that were independently associated with mortality in patients with SA-NVIE. After adjusting for patient-, pathogen-, and treatment-specific characteristics by multivariate analysis, geographical region was also found to be associated with mortality in patients with SA-NVIE ( P!.001). Conclusions. S. aureus is an important and common cause of IE. The outcome of SA-NVIE is worse than that of non SA-NVIE. Several clinical parameters are independently associated with mortality for patients with SA-NVIE. The clinical characteristics and outcome of SA-NVIE vary significantly by geographic region, although the reasons for such regional variations in outcomes of SA-NVIE are unknown and are probably multifactorial. A large, prospective, multinational cohort study of patients with IE is now under way to further investigate these observations. Staphylococcus aureus infective endocarditis (SAIE) is a complication of S. aureus bacteremia in 4 clinically distinct groups: injection drug users, hospitalized patients with nosocomial infections, prosthetic valve recipients, Received 12 November 2004; accepted 23 March 2005; electronically published 6 July Reprints or correspondence: Dr. Jose M. Miro, Infectious Diseases Service, Hospital Clinic Universitari, Helios-Villarroel Bldg., Desk no. 26, Villarroel, 170, 08036, Barcelona, Spain (jmmiro@ub.edu or miro97@fundsoriano.es). Clinical Infectious Diseases 2005; 41: by the Infectious Diseases Society of America. All rights reserved /2005/ $15.00 and non injection drug users with communityacquired infective endocarditis. Although right-side SAIE (usually associated with injection drug use) typically has a favorable prognosis with medical treatment alone [1 4], left-side SAIE remains a disease with high morbidity and mortality [5 12]. Because SAIE is uncommon at any single institution, prior investigations of infective endocarditis have been limited by small sample size, Presented in part: 41st Interscience Conference on Antimicrobial Agents and Chemotherapy, American Society for Microbiology, Chicago, Illinois, 2001 (abstract L-2292). a Study group members are listed at the end of the text. S. aureus Native Valve Infective Endocarditis CID 2005:41 (15 August) 507

2 Table 1. Main characteristics of the 7 endocarditis databases that have included data for cases of definite infective endocarditis (IE). Site of database Characteristic Philadelphia, PA [14] Durham, NC [15] Besançon, France [16] Marseille, France [17] London, England [18] Goteborg, Sweden [19] Barcelona, Spain [20] All sites combined Study period Cases of IE All Native valve 163 (86.7) 10 (70.4) 199 (78.3) 112 (69.6) 228 (78.4) 573 (85.0) 415 (84.3) 1797 (81.2) Prosthetic valve 25 (13.3) 28 (18.4) 47 (18.5) 49 (30.4) 63 (21.6) 98 (14.5) 57 (11.6) 367 (16.6) Other NA 17 (11.2) 8 (3.2) NA NA 3 (0.5) 20 (4.1) 48 (2.2) Patient characteristic Injection drug use 6 (3.2) 13 (8.6) 8 (3.2) 4 (2.5) 23 (7.9) 71 (10.5) 183 (37.2) 308 (13.9) Heart failure 52 (27.7) 45 (29.6) 111 (43.7) 64 (39.8) 73 (25.1) 246 (36.5) 190 (38.6) 781 (35.3) Surgery during index hospitalization 47 (25.0) 48 (31.6) 119 (46.9) 81 (50.3) 159 (54.6) 208 (30.9) 164 (33.3) 826 (37.4) Mortality among overall cohort a 21 (11.2) 24 (15.8) 45 (17.7) 12 (7.5) 66 (22.7) 68 (10.1) 112 (22.8) 348 (15.7) Staphylococcus aureus IE b 40 (21.3) 62 (40.8) 53 (20.9) 22 (13.7) 88 (30.2) 194 (28.8) 193 (39.2) 652 (29.5) MRSA IE c NA 25 (40.3) NA NA 12 (13.6) NA 13 (6.7) 50 (7.7) S. aureus native valve IE d 35 (87.5) 48 (77.4) 43 (81.1) 18 (81.8) 69 (78.4) 179 (92.3) 174 (90.2) 566 (86.8) Mortality a for S. aureus native valve IE 4 (11.4) 11 (22.9) 9 (20.9) 4 (22.2) 22 (31.9) 22 (12.3) 40 (23.0) 112 (19.8) NOTE. Data are no. (%) of patients, unless otherwise indicated. MRSA, methicillin-resistant S. aureus; NA, not available. a In-hospital mortality. b As a proportion of total cases of IE. c As a proportion of total cases of S. aureus IE, including both native valve MRSA IE (43 patients) and prosthetic valve MRSA IE (7 patients). d As a proportion of total cases of S. aureus IE. single-site enrollment, and/or inclusion of a wide spectrum of clinical factors influencing patient outcome. As a result of these limitations, fundamental features of S. aureus native valve infective endocarditis (SA-NVIE) remain unclear. The International Collaboration on Endocarditis Merged Database (ICE-MD) was designed to provide a large, multicenter, international resource for data on uniformly defined cases of infective endocarditis. The objective of the ICE-MD project is to evaluate regional and global aspects of infective endocarditis and to improve understanding of the clinical characteristics and outcome of a large, international cohort of patients with well-characterized infective endocarditis [13]. In this report, we defined risk factors, clinical characteristics, and prognostic factors for adverse outcomes among patients with definite SA- NVIE. PATIENTS, MATERIALS, AND METHODS Patient data. The methods used to create the ICE-MD have been described elsewhere [12 14]. In brief, the International Collaboration on Endocarditis investigators were queried to determine sites that maintained prospective databases in an electronic format. The ICE-MD was designed to allow a multinational consortium of investigators to have the ability to combine existing databases, with data collected during about prospectively identified patients with infective endocarditis. Seven sites from 5 countries (United States [15, 16], France [17, 18], the United Kingdom [19], Sweden [20], and Spain [21]) contributed data in a predesigned electronic format. The database from each center was sent to the coordinating center (Duke Clinical Research Institute, Durham, NC) for characterization and merging. Key domains, which contained core variables common to the individual databases, were determined. Standard definitions for each core variable were developed, and the merger was accomplished with use of a hierarchial variable structure that was described previously [13]. Definitions. Infective endocarditis was defined on the basis of the Duke Criteria [22]. Data entered into the merged databases that involved patients identified from 1979 to 1994 (i.e., before publication of the Duke criteria) were redefined retrospectively with use of this diagnostic schema. The variable total embolism included embolisms in any major arterial vessel, including pulmonary embolisms. CNS events were defined as CNS embolization, hemorrhage, or infection (e.g., meningitis or brain abscess). Surgical rates and mortality rates refer to inhospital surgery and mortality rates, respectively. The remaining clinical, echocardiographic, and outcome variables were defined as reported elsewhere [21 24]. Statistical analysis. Descriptive statistics for continuous variables were summarized as medians and interquartile ranges (IQRs). All sites did not collect all variables. Thus, each categorical variable was reported as the number of patients affected and as a percentage of patients for whom that variable was 508 CID 2005:41 (15 August) Miro et al.

3 Table 2. (NVIE). Clinical characteristics, surgery findings, and outcomes for patients with native valve infective endocarditis Variable non SA-NVIE (n p 1074) All (n p 566) SA-NVIE United States (n p 83) Europe (n p 483) P SA-NVIE vs. non SA-NVIE United States vs. Europe Demographic characteristic Age, median years (IQR) 60.0 (44 71) 46.0 (29 66) 58.0 (44 70) 43.0 (27 65)!.001!.001 Male sex 749/1074 (69.7) 360/566 (63.6) 43/83 (51.8) 317/483 (65.6) Diabetes mellitus 73/632 (11.6) 39/344 (11.3) 23/83 (27.7) 16/261 (6.1) 1.000!.001 Dialysis dependency 20/632 (3.2) 25/344 (7.3) 3/48 (6.3) 69/235 (29.4).006!.001 HIV infection 30/476 (6.3) 72/283 (25.4) 15/83 (18.1) 10/261 (3.8)!.001!.001 Other chronic disease 115/507 (22.7) 67/309 (21.7) 18/48 (37.5) 16/261 (6.1).795!.001 Long-term IVC 21/507 (4.1) 34/309 (11.0) 33/48 (68.8) 34/261 (13.0)!.001!.001 Acquisition in community 605/677 (89.4) 280/369 (75.9) 46/83 (55.4) 234/286 (81.8)!.001!.001 Injection drug use 59/1074 (5.5) 209/566 (36.9) 8/83 (9.6) 201/483 (41.6)!.001!.001 Congenital heart disease 123/757 (16.3) 22/387 (5.7) 6/83 (7.2) 16/304 (5.3)! Echocardiography finding! Transthoracic 305/949 (32.1) 224/531 (42.2) 15/48 (31.3) 209/483 (43.3) Transesophageal 183/949 (19.3) 81/531 (15.3) 5/48 (10.4) 76/483 (15.7) Both 393/949 (41.4) 185/531 (34.8) 28/48 (58.3) 157/483 (32.5) Vegetation location Aortic 335/1074 (31.2) 90/566 (15.9) 10/83 (12.1) 80/483 (16.6)! Mitral 364/1074 (33.9) 145/566 (25.6) 19/83 (22.9) 126/483 (26.1)! Tricuspic 54/1074 (5.0) 177/566 (31.3) 13/83 (15.7) 164/483 (34.0)!.001!.001 Pulmonic 12/581 (2.1) 6/335 (1.8) 0/60 (0.0) 6/275 (2.2) Intracardiac abscess 94/1074 (8.8) 39/566 (6.9) 5/83 (6.0) 34/483 (7.0) Outcome Pulmonary embolism 37/929 (4.0) 138/426 (32.4) 9/70 (12.9) 129/356 (36.2)!.001!.001 CNS event 124/931 (13.3) 86/417 (20.6) 12/60 (20.0) 74/357 (20.7)! Total embolism 330/1074 (30.7) 343/566 (60.6) 35/83 (42.2) 308/483 (63.8)!.001!.001 Heart failure 401/1073 (37.4) 183/560 (32.7) 25/83 (30.1) 158/477 (33.1) Surgery at index episode 416/1074 (38.7) 148/566 (26.2) 16/83 (19.3) 132/483 (27.3)! In-hospital death 130/1062 (12.2) 112/562 (19.9) 15/83 (18.1) 97/479 (20.3)! NOTE. Data are no. of patients with finding/no. of patients with data (%), unless otherwise indicated. IVC, intravascular catheter; SA, Staphylococcus aureus. available. The Wilcoxon rank sum test and Fisher s exact test were used to evaluate group differences for continuous and categorical variables, respectively. Descriptive univariate analysis of mortality-related factors was performed. For univariate analysis, sites for which data were not available for the variable of interest were excluded. For the multivariable analysis, only those variables that were available from all sites were entered into the model. P values of!.05 were considered to be statistically significant. All analyses were performed using SAS software, versions 6.12 and 8.2 (SAS Institute). RESULTS Data for a total of 2212 cases from the 20-year study period were collected (table 1). Native valve infective endocarditis (NVIE) was present in 81.2% of patients, and prosthetic valve infective endocarditis was present in 16.6%. Definite infective endocarditis due to S. aureus was present in 652 patients (29.5%), and SA-NVIE was present in 566 patients (25.6%). These 566 cases represent 353 cases in patients described in a previous report [14], 209 cases associated with injection drug use, and 4 cases in patients aged!18 years. SA-NVIE were younger and more likely to have comorbid conditions or to have engaged in injection drug use than were patients with non SA-NVIE (table 2). SA-NVIE were significantly less likely to have undergone valvular surgery (26.2% vs. 38.7%; P!.001) and were more likely to die during hospitalization (19.9% vs. 12.2%; P!.001) than were patients with non SA-NVIE. US patients with SA-NVIE were older, had a higher frequency of comorbid conditions, and had a lower rate of injection drug S. aureus Native Valve Infective Endocarditis CID 2005:41 (15 August) 509

4 Table 3. Clinical characteristics, surgery findings, and outcomes for patients with methicillin-susceptible Staphylococcus aureus (MSSA) native valve infective endocarditis (NVIE) and those with methicillin-resistant S. aureus (MRSA) NVIE. Variable MSSA NVIE (n p 248) MRSA NVIE (n p 43) P Demographic characteristic Age, median years (IQR) 33.0 (25 58) 60.0 (44 73)!.001 Male sex 164/248 (66.1) 21/43 (48.8).039 Year of diagnosis!.001 Before /248 (40.7) 4/43 (9.3) 1990 or after 147/248 (59.3) 39/43 (90.7) Diabetes mellitus 18/248 (7.3) 12/43 (27.9)!.001 HIV infection 70/187 (37.4) 2/35 (5.7)!.001 Dialysis dependency 18/248 (7.3) 7/43 (16.3).072 Long-term IVC 15/248 (6.1) 17/43 (39.5)!.001 Other chronic disease 37/248 (14.9) 20/43 (46.5)!.001 Acquisition in community 208/248 (83.9) 8/43 (18.6)!.001 Injection drug use 136/248 (54.8) 6/43 (14.0)!.001 Congenital heart disease 13/248 (5.2) 2/43 (4.7) Vegetation location Aortic 43/248 (17.3) 3/43 (7.0).112 Mitral 45/248 (18.2) 17/43 (39.5).004 Tricuspid 106/248 (42.7) 10/43 (23.3).018 Pulmonic 2/248 (0.8) 0/43 (0.0) Periannular abscess 20/248 (8.1) 2/43 (4.7).753 Outcome Pulmonary embolism 117/239 (49.0) 7/43 (16.3)!.001 CNS event 49/240 (20.4) 8/39 (20.5) Total embolization 158/248 (63.7) 19/43 (44.2).018 Heart failure 57/243 (23.5) 11/42 (26.2).700 Surgery at index episode 61/248 (24.6) 11/43 (25.6).851 In-hospital death 57/246 (23.2) 16/43 (37.2).058 NOTE. Data are no. of patients with finding/no. of patients with data (%), unless otherwise indicated. Data were obtained from 3 sites (2 sites in Europe and 1 site in the United States). IVC, intravascular catheter. use than did European patients with SA-NVIE. Rates of periannular abscess formation, heart failure, and surgical therapy were similar between these 2 geographic regions. Methicillin-resistant S. aureus (MRSA) endocarditis occurred in 14.8% of patients from centers where antimicrobial susceptibility testing was performed (table 3). Mitral valve involvement predominated in patents with NVIE due to MRSA, and tricuspid infection predominated in patients with NVIE due to methicillin-susceptible S. aureus (MSSA). Rates of surgical therapy were similar among patients with NVIE due to MSSA and those with NVIE due to MRSA, although the mortality rate was higher for patients with NVIE due to MRSA (23.2% vs. 37.2%; P p.058). Right-side NVIE occurred in younger patients and in injection drug users, and it occurred less frequently during the last decade of the study ( ) (table 4). Right-side NVIE was associated with fewer comorbid conditions (e.g., diabetes mellitus and chronic diseases) than was left-side NVIE. Surgical treatment was required more frequently for patients with leftside NVIE (39.7% vs. 12.4%; P!.001), and the in-hospital mortality rate was also higher for patients with left-side NVIE (28.6% vs. 5.9%; P!.001), compared with patients with rightside NVIE. Univariate logistic regression analyses were used to quantify the relationship between important clinical characteristics and in-hospital mortality rates (table 5). Important predictors of mortality in the study population included age (based on 10- year increments), presence of aortic or mitral valve vegetation, periannular abscess formation, heart failure, and CNS events. Characteristics associated with mortality in patients with SA- NVIE on multivariate analysis included increasing age (OR, 1.38; 95% CI, ), presence of heart failure (OR, 3.90, Table 4. Clinical characteristics, surgery findings, and outcomes for right-side and left-side Staphylococcus aureus native valve infective endocarditis (SA-NVIE) in patients with echocardiographically defined vegetations. Variable left-side SA-NVIE (n p 219) right-side SA-NVIE (n p 170) P Demographic characteristic Age, median years (IQR) 61.0 (44 70) 29.5 (24 39)!.001 Male sex 143/219 (65.3) 102/170 (60.0).292 Year of diagnosis!.001 Before /219 (10.5) 50/170 (29.4) 1990 or after 196/219 (89.5) 120/170 (70.6) Diabetes mellitus 24/110 (21.8) 5/110 (4.6)!.001 Dialysis dependency 14/110 (12.7) 5/110 (4.6).052 HIV infection 3/100 (3.0) 47/107 (43.9)!.001 Other chronic disease 36/107 (33.6) 12/110 (10.9)!.001 Long-term IVC 18/107 (16.8) 9/110 (8.2).065 Acquisition in community 77/123 (62.6) 104/120 (86.7)!.001 Injection drug use 18/219 (8.2) 131/170 (77.1)!.001 Congenital heart disease 10/135 (7.4) 2/122 (1.6).037 Echocardiography finding!.001 Transthoracic 59/216 (27.3) 106/170 (62.4) Transesophageal 42/216 (19.4) 17/170 (10.0) Both 107/216 (49.5) 40/170 (23.5) Periannular abscess 26/219 (11.9) 2/170 (1.2)!.001 Outcome Pulmonary embolism 4/160 (2.5) 91/136 (66.9)!.001 CNS event 48/163 (29.5) 10/125 (8.0)!.001 Total embolization 112/219 (51.1) 127/170 (74.7)!.001 Heart failure 98/218 (45.0) 23/167 (13.8)!.001 Surgery at index episode 87/219 (39.7) 21/170 (12.4)!.001 In-hospital death 62/217 (28.6) 10/169 (5.9)!.001 NOTE. Data are no. of patients with finding/no. of patients with data (%), unless otherwise indicated. IVC, intravascular catheter. 510 CID 2005:41 (15 August) Miro et al.

5 Table 5. Factors associated with in-hospital mortality in patients with Staphylococcus aureus native valve infective endocarditis. Variable Demographic characteristic Univariate analysis Multivariate analysis OR (95% CI) P OR (95% CI) P Year of infective endocarditis diagnosis before ( ).643 Age at entry a 1.37 ( )! ( )!.001 Male sex 0.86 ( ) ( ).777 Geographic location Philadelphia, PA 0.51 ( ) ( )!.001 Marseille, France 1.32 ( ) 1.05 ( ) Goteborg, Sweden 0.96 ( ) 0.36 ( ) Durham, NC 1.17 ( ) 0.97 ( ) Besançon, France 1.04 ( ) 0.89 ( ) Barcelona, Spain 1.20 ( ) 2.23 ( ) London, UK 1.85 ( ) 3.53 ( ) Comorbid condition Hemodialysis dependency 1.31 ( ).539 Diabetes mellitus 1.96 ( ).064 HIV infection 0.46 ( ).037 Other chronic illness 2.26 ( ).006 Long-term IVC 1.49 ( ).314 Community-acquired infection 0.90 ( ).736 Predisposing conditions Current injection drug use 0.33 ( )! ( ).927 Congenital heart disease 0.31 ( ).120 Intracardiac findings Aortic valve vegetation 2.54 ( )! ( ).050 Mitral valve vegetation 1.87 ( ) ( ).114 Tricuspid valve vegetation 0.30 ( )! ( ).296 Periannular abscess 2.87 ( ) ( ).038 Outcome Heart failure 3.58 ( )! ( )!.001 Overall systemic embolization 0.66 ( ) ( ).544 Peripheral embolization 0.83 ( ).772 Pulmonary embolization 0.24 ( )!.001 CNS event, including embolization 2.39 ( ).002 Surgery at this episode 1.14 ( ) ( ).007 NOTE. Reference for geographic location is the average overall mortality among all sites. Reference for community-acquired endocarditis is health care associated endocarditis. IVC, intravascular catheter. a Per 10-year increase in age. 95% CI, ) or periannular abscess (OR, 2.43; 95% CI, ), presence of an aortic valve vegetation (OR, 1.91; 95% CI, ), and the geographic location in which the patient was identified. Surgical therapy was associated with reduced in-hospital mortality (OR, 0.43; 95% CI, ) after adjustment for characteristics of the patients. DISCUSSION Although S. aureus is an important cause of NVIE in the modern era, its relative infrequency at any single site has limited the ability to evaluate large cohorts of patients with this condition. In addition, the findings from such single-site reports may not be generalizable to other practice settings. The current investigation is a start in the attempt to overcome the limitations of previous SA-NVIE studies, because it uses a multinational collection of data for well-characterized patients, and it contributes to our understanding of this disease with several key observations. The current investigation is, to our knowledge, the first to provide compelling evidence underscoring a regional variation S. aureus Native Valve Infective Endocarditis CID 2005:41 (15 August) 511

6 in the outcomes for patients with infective endocarditis. This regional variation persisted in a cohort of patients with the same disease type (NVIE) caused by the same pathogen (S. aureus), even after adjustment for a number of distinct patientspecific characteristics. Although such regional variation has been reported for other potentially lethal conditions, such as acute myocardial infarction [25], it has not been previously identified for patients with infective endocarditis. The reasons for this finding may relate, in part, to regional variations in treatment strategies (e.g., different health care access, practice variations in antimicrobial therapy, and variations in referral patterns), in patient demographic characteristics (e.g., host susceptibility to S. aureus infection), or in innate virulence attributes of the pathogen. Of note, the current prospective International Collaboration on Endocarditis cohort study should begin to address the underlying causes of such regional variations [13]. This investigation also emphasizes the increasing importance of S. aureus as a cause of NVIE in the modern era. In our cohort of patients with infective endocarditis, S. aureus was the most common cause of NVIE. By contrast, the proportion of cases of infective endocarditis due to viridans group streptococci or enterococci was significantly lower than that in studies that have encompassed cases from the last 2 decades [26, 27]. Our observations on the increasing prevalence of SA- NVIE reflect the probable convergence of 2 important trends: there is an increasing number of patients (e.g., geriatric patients and patients with transvenous pacemakers or indwelling catheters) who are at risk for endocarditis in general, and there is an increasing number of patients (e.g., patients with immunosuppression, diabetic patients, and persons undergoing hemodialysis) who are at risk for S. aureus bacteremia [8, 16, 28 30]. Given that gram-positive organisms have now overtaken gram-negative pathogens as the most common cause of septicemia in the United States [31], it is likely that rates of SA- NVIE will continue to increase. MRSA endocarditis occurred in 14.8% of patients from centers in which antimicrobial susceptibility testing was recorded, and at least one-third of SAIE cases were due to MRSA at one center since In agreement with prior observations [32, 33], patients with NVIE due to MRSA tended to have more comorbid conditions than did patients with NVIE due to MSSA. The mortality rate for MRSA-infected patients also tended to be higher, although this observation did not achieve statistical significance ( P p.054). Although some investigations have reported a similar association between MRSA infection and worse clinical outcomes [34 36], others have failed to demonstrate this association when adjustment for comorbid conditions was made [30, 31, 37, 38]. It is still a matter of debate whether the increase in mortality for MRSA-infected patients is related to increased virulence of the strain, an increase in the number of comorbid diseases, or the relative ineffectiveness of vancomycin. Given the large increase in serious community-acquired MRSA infections in the past few years [39], it is highly likely that the proportion of cases of SA-NVIE caused by MRSA will continue to increase. The findings of the current investigation also confirm the virulence of S. aureus compared with other causes of NVIE. SA-NVIE in this and prior investigations [8, 10, 11, 28, 40, 41] had a higher mortality than patients with NVIE due to other pathogens. However, this is the first investigation to demonstrate this pathogen-specific virulence across different countries and health care systems. These findings suggest that S. aureus strains causing infective endocarditis possess a similar capacity for virulence throughout diverse geographic regions of the world. The present study has several limitations. Different data collection methods were used at each site, potentially reducing the precision of certain core variables. Referral bias is also likely, because the study institutions are predominantly tertiary care referral centers. Finally, the apparent regional variation in outcome of SA-NVIE may be related to factors not specifically addressed in this investigation (e.g., varying rates of comorbid conditions, higher rates of injection drug use among individual sites, health care access, treatment practices, and potential virulence properties of bacteria). Thus, the findings of this investigation should be considered on for generation of hypotheses until they are externally validated with a separate cohort. Despite these limitations, we believe that the findings of this study underscore the global significance of S. aureus and identify several patient and environmental factors associated with mortality in SA-NVIE. A large, prospective, multinational cohort study of patients with infective endocarditis is now under way to validate these observations and to further evaluate clinical, pathogen-specific, and host-specific determinants of outcome in this devastating infection. SA-NVIE is a potentially life-threatening infection of global significance. It is primarily associated with either injection drug use or comorbid conditions. In this investigation, patients with SA-NVIE from different geographic regions had different risks of mortality. Future prospective, multinational investigations are required to validate our observations and to further evaluate clinical, pathogen-specific, and host-specific determinants of apparent differences in outcome between SA-NVIE and non SA-NVIE. MEMBERS OF THE ICE-MD STUDY GROUP J. M. Miró, A. del Río, M. A. Baraldes, M. J. Jiménez-Expósito, N. de Benito, X. Claramonte, M. E. Díaz, O. Sued, C. Manzardo, A. Moreno, J. M. Gatell, F. Marco, C. García de la María, Y. Armero, M. Almela, M. T. Jiménez de Anta, J. C. Paré, M. Azqueta, C. A. Mestres, S. Ninot, R. Cartaña, J. L. Pomar, N. Pérez, 512 CID 2005:41 (15 August) Miro et al.

7 J. Ramírez, and T. Ribalta (Barcelona, Spain); P. Stolley (Baltimore, MD); B. Hoen, C. Selton-Suty, T. Doco-Lecompte, F. Duchêne, N. Khayat, Y. Bernard, and C. Chirouze (Besançon and/or Nancy, France); G. R. Corey, D. J. Sexton, V. G. Fowler, Jr., C. W. Woods, A. Wang, G. E. Peterson, J. G. Jollis, D. J. Anderson, R. Singh, C. H. Cabell, D. Glower, A. Chen, and J. Stafford (Durham, NC); L. Olaison and the Swedish Society of Infectious Diseases Quality Assurance Study Group for Endocarditis (Goteborg, Sweden); A. Thalme (Stockholm, Sweden); S. Eykyn (London, UK); D. Raoult, G. Habib, J. P. Casalta, K. Barrau, and P. E. Fournier (Marseille, France); and E. Abrutyn, B. L. Strom, J. A. Berlin, J. L. Kinman, R. S. Feldman, M. E. Levison, O. M. Korzeniowski, and D. Kaye (Philadelphia, PA). Acknowledgments Financial support. National Institutes of Health (AI [to V.G.F.] and HL70861 [to C.H.C.]), the Red Española de Investigación en Patología Infecciosa (V-2003-REDC14A-O), and the Fundación Privada Máximo Soriano Jiménez for the grant supporting the Hospital Clínic Endocarditis Database (to J.M.M.). J.M.M. received a research grant from the Institut d Investigacions Biomèdiques August Pi i Sunyer. Potential conflicts of interest. V.G.F. has received research grants from Cubist, Inhibitex, Nabi, National Institutes of Health, Theravance, Merck, Ortho-McNeil, and Vicuron; has received speaking honoraria from Pfizer, Cubist, and Aventis; and has consulted for Merck, Nabi, Inhibitex, Elusys, Cubist, Vicuron, and GlaxoSmithKine. All other authors: no conflicts. References 1. Hecht SR, Berger M. Right-sided endocarditis in intravenous drug users: prognostic features in 102 episodes. Ann Intern Med 1992; 117: Torres-Tortosa M, Gonzalez-Serrano M, Perez-Guzman E, et al. Infective endocarditis in heroin addicts in the province of Cadiz: a multicenter study of 150 cases. Med Clin (Barc) 1992; 98: Ribera E, Miro JM, Cortes E, et al. Influence of human immunodeficiency virus I infection and degree of immunosuppression in the clinical characteristics and outcome of infective endocarditis in intravenous drug users. Arch Intern Med 1998; 158: Miro JM, Del Rio A, Mestres CA. Infective endocarditis in intravenous drug abusers and HIV-1 infected patients. Infect Dis Clin North Am 2002; 16: Espersen F, Frimodt-Moller N. Staphylococcus aureus endocarditis: a review of 119 cases. Arch Intern Med 1986; 146: Sanabria TJ, Alpert JS, Golberg R, Pape LA, Cheeseman SH. Increasing frequency of staphylococcal infective endocarditis: experience at a university hospital, 1981 through Arch Intern Med 1990; 150: Karchmer AW. Staphylococcal endocarditis. In: Kaye D, ed. Infective endocarditis. 2nd ed. New York: Raven Press, 1992: Watanakunakorn C. Staphylococcus aureus endocarditis at a commmunity teaching hospital, 1980 to 1991: an analysis of 106 cases. Arch Intern Med 1994; 154: Roder BL, Wandall DA, Espersen F, Frimodt-Moller N, Skinhoj P, Rosdahl VT. Neurologic manifestations in Staphylococcus aureus endocarditis: a review of 260 bacteremic cases in nondrug addicts. Am J Med 1997; 102: Roder BL, Wandall DA, Frimodt-Moller N, Espersen F, Skinhoj P, Rosdahl VT. Clinical features of Staphylococcus aureus endocarditis: a 10-year experience in Denmark. Arch Intern Med 1999; 159: Jensen AG, Wachmann CH, Espersen F, Scheibel J, Skinhoj P, Frimodt- Moller N. Treatment and outcome of Staphylococcus aureus bacteremia: a prospective study of 278 cases. Arch Intern Med 2002; 162: Chirouze C, Cabell CH, Fowler VG Jr, et al. Prognostic factors in 61 cases of Staphylococcus aureus prosthetic valve endocarditis from the International Collaboration on Endocarditis Merged Database. Clin Infect Dis 2004; 38: Cabell CH, Abrutyn E. Progress toward a global understanding of infective endocarditis: early lessons from the International Collaboration on Endocarditis investigation. Infect Dis Clin North Am 2002;16: Chu VH, Cabell CH, Abrutyn E, et al. Native valve endocarditis due to coagulase-negative staphylococci: report of 99 episodes from the International Collaboration on Endocarditis Merged Database. Clin Infect Dis 2004; 39: Berlin JA, Abrutyn E, Strom BL, et al. Incidence of infective endocarditis in the Delaware Valley, Am J Cardiol 1995; 76: Cabell CH, Jollis JG, Peterson GE, et al. Changing patient characteristics and the effect on mortality in endocarditis. Arch Intern Med 2002; 162: Hoen B, Beguinot I, Rabaud C, et al. The Duke criteria for diagnosing infective endocarditis are specific: analysis of 100 patients with acute fever or fever of unknown origin. Clin Infect Dis 1996; 23: Di Salvo G, Habib G, Pergola V, et al. Echocardiography predicts embolic events in infective endocarditis. J Am Coll Cardiol 2001; 37: Lamas CC, Eykyn SJ. Bicuspid aortic valve a silent danger: analysis of 50 cases of infective endocarditis. Clin Infect Dis 2000; 30: Olaison L, Schadewitz K. Enterococcal endocarditis in Sweden, : can shorter therapy with aminoglycosides be used? Swedish Society of Infectious Diseases Quality Assurance Study Group for Endocarditis. Clin Infect Dis 2002; 34: Losa JE, Miro JM, Del Rio A, et al. Infective endocarditis not related to intravenous drug abuse in HIV-1 infected patients: report of eight cases and review of the literature. Clin Microbiol Infect 2003; 9: Durack DT, Lukes AS, Bright DK. New criteria for diagnosis of infective endocarditis: utilization of specific echocardiographic findings. Duke Endocarditis Service. Am J Med 1994; 96: Martinez E, Miro JM, Almirante B, et al. Effect of penicillin resistance of Streptococcus pneumoniae on the presentation, prognosis, and treatment of pneumococcal endocarditis in adults. Clin Infect Dis 2002; 35: Anguera I, Quaglio G, Miro JM, et al. Aortocardiac fistulas complicating infective endocarditis. Am J Cardiol 2001; 87: Giugliano RP, Llevadot J, Wilcox RG, et al. Geographic variation in patient and hospital characteristics, management, and clinical outcomes in ST-elevation myocardial infarction treated with fibrinolysis: results from InTIME-II. InTIME (Intravenous npa for Treatment of Infarcting Myocardium Early) II Investigators. Eur Heart J 2001; 22: Delahaye F, Goulet V, Lacassin F, et al. Characteristics of infective endocarditis in France in 1991: a 1-year survey. Eur Heart J 1995; 16: Garvey GJ, Neu HC. Infective endocarditis an evolving disease: a review of endocarditis at the Columbia-Presbyterian Medical Center, Medicine (Baltimore) 1978; 57: Ellis ME, Rhydderch D, Zwaan F, Guy ML, Baillie F. High incidence of line-related infection and mechanical failure of an antiseptic-impregnated central venous catheter in highly immunocompromised patients. Scand J Infect Dis 1996; 28: McCarthy JT, Steckelberg J. Infective endocarditis in patients receiving long-term hemodialysis. Mayo Clin Proc 2000; 75: Fowler VG Jr, Sanders LL, Kong LK, et al. Infective endocarditis due to Staphylococcus aureus: 59 prospectively identified cases with followup. Clin Infect Dis 1999; 28: Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of sepsis in the United States from 1979 through N Engl J Med 2003; 348: Habarth S, Rutschmann O, Sudre P, Pittet D. Impact of methicillin S. aureus Native Valve Infective Endocarditis CID 2005:41 (15 August) 513

8 resistance on the outcome of patients with bacteremia caused by Staphylococcus aureus. Arch Intern Med 1998; 158: Soriano A, Martinez JA, Mensa J, et al. Pathogenic significance of methicillin resistance for patients with Staphylococcus aureus bacteremia. Clin Infect Dis 2000; 30: Romero-Vivas J, Rubio M, Fernandez C, Picazo JJ. Mortality associated with nosocomial bacteremia due to methicillin-resistant Staphylococcus aureus. Clin Infect Dis 1995; 21: Mekontso-Dessap A, Kirsch M, Brun-Buisson C, Loisance D. Poststernotomy mediastinitis due to Staphylococcus aureus: comparison of methicillin-resistant and methicillin-susceptible cases. Clin Infect Dis 2001; 32: Cosgrove SE, Sakoulas G, Perencevich EN, Schwaber MJ, Karchmer AW, Carmeli Y. Comparison of mortality associated with methicillinresistant and methicillin-susceptible Staphylococcus aureus bacteremia: a meta-analysis. Clin Infect Dis 2003; 1: Mylotte JM, Aeschlimann JR, Rotella DL. Staphylococcus aureus bacteremia: factors predicting hospital mortality. Infect Control Hosp Epidemiol 1996; 17: Wisplinghoff H, Seifert H, Coimbra M, Wenzel RP, Edmond MB. Systemic inflammatory response syndrome in adult patients with nosocomial bloodstream infection due to Staphylococcus aureus. Clin Infect Dis 2001; 33: Diekema DJ, Pfaller MA, Schmitz FJ, et al. Survey of infections due to Staphylococcus species: frequency of occurrence and antimicrobial susceptibility of isolates collected in the United States, Canada, Latin America, Europe, and the Western Pacific region for the SENTRY Antimicrobial Surveillance Program, SENTRY Participants Group. Clin Infect Dis 2001; 32(Suppl 2):S Watanakunakorn C, Burkert T. Infective endocarditis at a large community teaching hospital, : a review of 210 episodes. Medicine (Baltimore) 1993; 72: Hasbun R, Vikram HR, Barakat LA, Buenconsejo J, Quagliarello VJ. Complicated left-sided native valve endocarditis in adults: risk classification for mortality. JAMA 2003; 289: CID 2005:41 (15 August) Miro et al.

9 ERRATUM In an article in the 15 August 2005 issue of the journal (Miro JM, Anguera I, Cabell CH, et al. Staphylococcus aureus native valve infective endocarditis: report of 566 episodes from the International Collaboration on Endocarditis Merged Database. Clin Infect Dis 2005; 41:507 14), errors appeared in the abstract and in table 1. In the Results section of the abstract, the percentage listed after 566 patients was incorrect and has been deleted. Also, the percentages of patients with S. aureus infective endocarditis who experienced embolic and central nervous system events should be 61% and 21%, respectively (not 60% and 20%, respectively). Finally, the 95% CI for periannular abscess should be (not ). The corrected section of the abstract appears below. The authors regret these errors. Results. Of patients with native valve IE, 566 patients had IE due to S. aureus, and 1074 patients had IE due to pathogens other than S. aureus (non SA-NVIE). S. aureus IE were more likely to die (20% vs. 12%; P!.001), to experience an embolic event (61% vs. 31%; P!.001), or to have a central nervous system event (21% vs. 13%; P!.001) and were less likely to undergo surgery (26% vs. 39%; P!.001) than were patients with non SA-NVIE. Multivariate analysis of prognostic factors of mortality identified age (odds ratio [OR], 1.4; 95% confidence interval [CI], ), periannular abscess (OR, 2.4; 95% CI, ), heart failure (OR, 3.9; 95% CI, ), and absence of surgical therapy (OR, 2.3; 95% CI, ) as variables that were independently associated with mortality in patients with SA-NVIE. After adjusting for patient-, pathogen-, and treatment-specific characteristics by multivariate analysis, geographical region was also found to be associated with mortality in patients with SA-NVIE ( ). P!.001 For table 1, the references listed for the 7 databases are incorrect. The correct references are as follows: Philadelphia, PA [15] (not [14]); Durham, NC [16] (not [15]); Besançon, France [17] (not [16]); Marseille, France [18] (not [17]); London, England [19] (not [18]); Goteborg, Sweden [20] (not [19]); and Barcelona, Spain [21] (not [20]). The journal and the authors regret these errors. Furthermore, the number of patients with native valve infective endocarditis for Durham, NC, should be 107 (not 10). The journal regrets this error. The corrected version of table 1 appears on the next page (p. 1076). In addition, the authors would like to emend table 2. For the column labeled non SA-NVIE ( ), n p 1074 it should be noted that the data refer to cases that involve known pathogens. The corrected version of table 2 appears after table 1 (on p. 1077), with the added information shown in footnote a. Clinical Infectious Diseases 2005; 41: by the Infectious Diseases Society of America. All rights reserved /2005/ $15.00 ERRATUM CID 2005:41 (1 October) 1075

10 Table 1. Main characteristics of the 7 endocarditis databases that have included data for cases of definite infective endocarditis (IE). Site of database Characteristic Philadelphia, PA [15] Durham, NC [16] Besançon, France [17] Marseille, France [18] London, England [19] Goteborg, Sweden [20] Barcelona, Spain [21] All sites combined Study period Cases of IE All Native valve 163 (86.7) 107 (70.4) 199 (78.3) 112 (69.6) 228 (78.4) 573 (85.0) 415 (84.3) 1797 (81.2) Prosthetic valve 25 (13.3) 28 (18.4) 47 (18.5) 49 (30.4) 63 (21.6) 98 (14.5) 57 (11.6) 367 (16.6) Other NA 17 (11.2) 8 (3.2) NA NA 3 (0.5) 20 (4.1) 48 (2.2) Patient characteristic Injection drug use 6 (3.2) 13 (8.6) 8 (3.2) 4 (2.5) 23 (7.9) 71 (10.5) 183 (37.2) 308 (13.9) Heart failure 52 (27.7) 45 (29.6) 111 (43.7) 64 (39.8) 73 (25.1) 246 (36.5) 190 (38.6) 781 (35.3) Surgery during index hospitalization 47 (25.0) 48 (31.6) 119 (46.9) 81 (50.3) 159 (54.6) 208 (30.9) 164 (33.3) 826 (37.4) Mortality among overall cohort a 21 (11.2) 24 (15.8) 45 (17.7) 12 (7.5) 66 (22.7) 68 (10.1) 112 (22.8) 348 (15.7) Staphylococcus aureus IE b 40 (21.3) 62 (40.8) 53 (20.9) 22 (13.7) 88 (30.2) 194 (28.8) 193 (39.2) 652 (29.5) MRSA IE c NA 25 (40.3) NA NA 12 (13.6) NA 13 (6.7) 50 (7.7) S. aureus native valve IE d 35 (87.5) 48 (77.4) 43 (81.1) 18 (81.8) 69 (78.4) 179 (92.3) 174 (90.2) 566 (86.8) Mortality a for S. aureus native valve IE 4 (11.4) 11 (22.9) 9 (20.9) 4 (22.2) 22 (31.9) 22 (12.3) 40 (23.0) 112 (19.8) NOTE. Data are no. (%) of patients, unless otherwise indicated. MRSA, methicillin-resistant S. aureus; NA, not available. a In-hospital mortality. b As a proportion of total cases of IE. c As a proportion of total cases of S. aureus IE, including both native valve MRSA IE (43 patients) and prosthetic valve MRSA IE (7 patients). d As a proportion of total cases of S. aureus IE CID 2005:41 (1 October) ERRATUM

11 Table 2. Clinical characteristics, surgery findings, and outcomes for patients with native valve infective endocarditis (NVIE). Variable Demographic characteristic non SA-NVIE a (n p 1074) All (n p 566) SA-NVIE United States (n p 83) Europe (n p 483) P SA-NVIE vs. non SA-NVIE United States vs. Europe Age, median years (IQR) 60.0 (44 71) 46.0 (29 66) 58.0 (44 70) 43.0 (27 65)!.001!.001 Male sex 749/1074 (69.7) 360/566 (63.6) 43/83 (51.8) 317/483 (65.6) Diabetes mellitus 73/632 (11.6) 39/344 (11.3) 23/83 (27.7) 16/261 (6.1) 1.000!.001 Dialysis dependency 20/632 (3.2) 25/344 (7.3) 3/48 (6.3) 69/235 (29.4).006!.001 HIV infection 30/476 (6.3) 72/283 (25.4) 15/83 (18.1) 10/261 (3.8)!.001!.001 Other chronic disease 115/507 (22.7) 67/309 (21.7) 18/48 (37.5) 16/261 (6.1).795!.001 Long-term IVC 21/507 (4.1) 34/309 (11.0) 33/48 (68.8) 34/261 (13.0)!.001!.001 Acquisition in community 605/677 (89.4) 280/369 (75.9) 46/83 (55.4) 234/286 (81.8)!.001!.001 Injection drug use 59/1074 (5.5) 209/566 (36.9) 8/83 (9.6) 201/483 (41.6)!.001!.001 Congenital heart disease 123/757 (16.3) 22/387 (5.7) 6/83 (7.2) 16/304 (5.3)! Echocardiography finding! Transthoracic 305/949 (32.1) 224/531 (42.2) 15/48 (31.3) 209/483 (43.3) Transesophageal 183/949 (19.3) 81/531 (15.3) 5/48 (10.4) 76/483 (15.7) Both 393/949 (41.4) 185/531 (34.8) 28/48 (58.3) 157/483 (32.5) Vegetation location Aortic 335/1074 (31.2) 90/566 (15.9) 10/83 (12.1) 80/483 (16.6)! Mitral 364/1074 (33.9) 145/566 (25.6) 19/83 (22.9) 126/483 (26.1)! Tricuspic 54/1074 (5.0) 177/566 (31.3) 13/83 (15.7) 164/483 (34.0)!.001!.001 Pulmonic 12/581 (2.1) 6/335 (1.8) 0/60 (0.0) 6/275 (2.2) Intracardiac abscess 94/1074 (8.8) 39/566 (6.9) 5/83 (6.0) 34/483 (7.0) Outcome Pulmonary embolism 37/929 (4.0) 138/426 (32.4) 9/70 (12.9) 129/356 (36.2)!.001!.001 CNS event 124/931 (13.3) 86/417 (20.6) 12/60 (20.0) 74/357 (20.7)! Total embolism 330/1074 (30.7) 343/566 (60.6) 35/83 (42.2) 308/483 (63.8)!.001!.001 Heart failure 401/1073 (37.4) 183/560 (32.7) 25/83 (30.1) 158/477 (33.1) Surgery at index episode 416/1074 (38.7) 148/566 (26.2) 16/83 (19.3) 132/483 (27.3)! In-hospital death 130/1062 (12.2) 112/562 (19.9) 15/83 (18.1) 97/479 (20.3)! NOTE. Data are no. of patients with finding/no. of patients with data (%), unless otherwise indicated. IVC, intravascular catheter; SA, Staphylococcus aureus. a Data are for cases that involve known pathogens. ERRATUM CID 2005:41 (1 October) 1077