Journal Logo


A Prospective Multicenter Study of Staphylococcus aureus Bacteremia

Incidence of Endocarditis, Risk Factors for Mortality, and Clinical Impact of Methicillin Resistance

Chang, Feng-Yee; MacDonald, Brent B.; Peacock, James E. Jr; Musher, Daniel M.; Triplett, Patricia; Mylotte, Joseph M.; O’Donnell, Alice; Wagener, Marilyn M.; Yu, Victor L.

Author Information
doi: 10.1097/
  • Free



Staphylococcus aureus is an increasing cause of bacteremia, both in the community and in the hospital setting. Bacteremia due to this organism poses a double threat: the bacteremia may result from a variety of deep-seated infections including endocarditis and, whatever its source, it may cause endocarditis secondarily. The goals of antibiotic therapy are to eradicate the bacteremia and prevent relapse or secondary infection on the heart valves and elsewhere. Controversy over the duration of therapy for S. aureus bacteremia has been debated for several decades.

A critical diagnostic question in patients with S. aureus bacteremia is whether endocarditis is also present. Acquisition of infection in the community, the presence of bacteremia with no apparent source (a primary bacteremia), and the occurrence of metastatic foci of infection have been thought to suggest the presence of endocarditis. However, the studies from which those predictors were derived were either small in size or retrospective in design. Since an increasing proportion of S. aureus bacteremias in both community-acquired and hospital-acquired settings is now due to methicillin-resistant S. aureus (MRSA), it is important to determine whether clinical features of endocarditis due to MRSA differ from those due to methicillin-susceptible S. aureus (MSSA).

To address those issues, we conducted a large-scale prospective, multicenter, observational study on S. aureus bacteremia. Our objectives were to determine the incidence of endocarditis in patients whose S. aureus bacteremia was community-acquired, related to hemodialysis, or hospital-acquired; to assess clinical factors that would reliably distinguish between S. aureus bacteremia and S. aureus endocarditis; to assess the emergence of MRSA as a cause of endocarditis; and to examine risk factors for mortality in patients with S. aureus endocarditis.


Study Design

From August 1994 to March 1996, 505 consecutive patients in 6 hospitals (Presbyterian University Hospital; Montefiore University Hospital; Veterans Affairs Medical Center; Pittsburgh, PA; Veterans Affairs Medical Center; Houston, TX; Wake Forest University, Baptist Medical Center; Winston-Salem, NC; Erie County Medical Center, Buffalo, NY) from whom S. aureus was isolated from blood were followed prospectively. Institutional Review Board review was performed at all hospitals. Patients with S. aureus bacteremia were identified through a daily review of blood culture results in the microbiology laboratory. Thirty-four percent (173/505) had 1 positive blood culture for S. aureus, 47% (236/505) had 2 positive blood cultures, 11% (56/505) had 3 positive blood cultures, 7% (37/505) had greater than 3 positive blood cultures, and <1% (3/505) had an unknown number of positive blood cultures. These patients were then evaluated by an investigator who participated in the design of this study. The study was observational in that decisions regarding management were made by the primary physicians in accordance with their clinical judgment. Subsequent blood cultures were obtained as they were deemed clinically indicated by the primary physicians, usually when the patients had persistent or recurrent fever on therapy. All patients were monitored for at least 6 months after onset of bacteremia by review of clinic visits after discharge, discussion with managing physicians, and/or telephone interview. Patients with endocarditis were monitored for 3 years to determine whether valvular infection had recurred.


Identification of bloodstream isolates as S. aureus, assays of antibiotic susceptibility, and determination if strains were MRSA or MSSA were done using standard techniques 54. Susceptibility to oxacillin was screened by disk diffusion using a 1 μg oxacillin disk and standard zone size criteria 54. Methicillin-resistance was verified by growth in broth supplemented with 2% sodium chloride containing 2 μg/mL oxacillin (Pos Combo Panel 4I, MicroScan R, West Sacramento, CA). The microdilution trays were inoculated with 105 cfu/mL and incubated at 35 °C for 24 hours. Following the National Committee for Clinical Laboratory Standards criteria, an isolate was considered to be resistant to oxacillin if the minimum inhibitory concentration (MIC) was greater than 2 μg/mL. Pulsed-field gel electrophoresis was performed as previously described 4,23,68.


Onset of bacteremia was defined as the date when the first positive blood culture was obtained. Definitions of community-acquired and hospital-acquired infection were in accord with the Centers for Disease Control Guidelines 22.

Endocarditis was defined according to the 1994 Duke criteria and modification 39. Cases were included that met definitions for both definite and possible endocarditis. “New” endocarditis, that is, development of endocarditis after the onset of bacteremia, was defined by fulfillment of the following 3 criteria: 1) the absence of clinical evidence of endocarditis at the time the bacteremia was first documented, 2) an interval of 1 week or more between the last positive blood culture and the earliest appearance of findings consistent with endocarditis, and 3) an identical pulsed-field gel electrophoresis pattern of the initial blood culture isolate and that obtained when endocarditis was diagnosed. These criteria are a modification of previously published criteria 15 with the addition of molecular subtyping results to identify recurring isolates. Nosocomial endocarditis was defined as endocarditis acquired more than 3 days from date of hospital admission with no clinical manifestations of endocarditis before admission. Persistent bacteremia was defined as the presence of positive blood cultures for S. aureus after ≥3 days of appropriate antistaphylococcal antibiotic therapy. Portal of entry was determined by the investigator based on clinical manifestations and positive cultures for S. aureus from extravascular sites. Catheter-related bacteremia was defined when a semiquantitative culture of the vascular catheter tip yielded greater than 15 colonies of S. aureus, and another source of bacteremia could not be identified 41. Inflammation at the catheter insertion site was not a sufficient criterion for defining the catheter as a source. Severity of illness was assessed for each patient at the onset of bacteremia using APACHE III score 32 and Pitt bacteremia score 9. The Pitt bacteremia score has been independently validated in a prospective study of 424 cases of S. aureus bacteremia 27.


Outcome (death versus survival) was recorded at 14 days, 30 days, and 60 days after the initial positive blood culture for S. aureus was obtained. Recurrence of valvular infection in patients with endocarditis was assessed at 3 months and 36 months after the initial blood culture for S. aureus.

Data Management and Analysis

Clinical and laboratory results were entered into a computer database (Prophet Systems, National Institutes of Health, Bethesda, MD). Categorical data were analyzed using a chi-square or Fisher exact test. Continuous variables were compared using the t test or the Mann-Whitney test. A step-wise regression model was used to examine the effects of multiple risk factors found to be significant by univariate analysis.


Incidence of Endocarditis

Of the 505 cases of S. aureus bacteremia, 13% (64/505) met the Duke criteria for endocarditis, including 51 with “definite” endocarditis (9 by pathologic and 42 by clinical criteria) and 13 patients with “possible” endocarditis 12. Of these 64 patients, 67% (43/64) underwent echocardiography: transthoracic only (n = 10), transesophageal only (n = 6), or both (n = 27). Fifteen percent (4/27) had no evidence of endocarditis by either transthoracic or transesophageal echocardiogram when both procedures were performed. Our study was not designed to assess sensitivity, specificity, or interobserver variability in the 6 hospitals, so no further data are provided on echocardiography.

The prevalence of endocarditis in patients with community-acquired S. aureus bacteremia was 21% (43/206), 12% (11/95) in hemodialysis patients, and 5% (10/204) in patients with hospital-acquired bacteremia (Figure 1). Sixty-nine percent (44/64) and 31% (20/64) of cases of endocarditis were due to MSSA and MRSA, respectively. Oslerian manifestations were detected in 61% (39/64) patients with endocarditis including occurrence of new murmur (54%), evidence of emboli (41%), and presence of skin lesions (22%) including petechiae (11%) and Janeway lesions (11%).

Comparison of site of acquisition, antibiotic susceptibility, and outcome for patients with Staphylococcus aureus bacteremia (n = 441) and endocarditis (n = 64). MRSA = methicillin-resistant S. aureus; MSSA = methicillin-susceptible S. aureus. Community-acquisition was significantly more likely to occur in patients with endocarditis (see Table 2, p > 0.04, 3 × 2 chi-square test). Incidence of MRSA and mortality was similar for patients with bacteremia only and for patients with endocarditis.
Risk Factor Analysis for Factors that Might Distinguish Endocarditis and Bacteremia

New Endocarditis

Seven of 64 patients met the definition of new endocarditis as a complication of S. aureus bacteremia, as defined in the Methods (Table 1). An intravascular catheter was the most likely cause of new endocarditis: 3 patients had documented catheter-related bacteremia and 2 had suspected catheter-associated bacteremia, although confirmation was not available since dialysis catheters were not removed. In 5 of 7 patients, S. aureus bacteremia relapsed 17–87 days (median, 32 d) after the discontinuation of antistaphylococcal antibiotics for treatment of the initial episode of bacteremia. The remaining 2 patients had persistent bacteremia for greater than 7 days during administration of active antistaphylococcal antibiotics. Of the 7 patients, 6 had severe arteriosclerotic heart disease or valvular disease, including 2 patients with prosthetic valves (see Table 1). All 7 patients had received therapy with vancomycin with appropriate dosing for the initial episode of bacteremia, and all received less than 4 weeks of therapy (see Table 1). In our patients with S. aureus bacteremia, 12 had prosthetic valves; 17% (2/12) developed “new endocarditis” as defined in the Methods.

Seven Patients with New Endocarditis*

In a more focused analysis of the impact of duration of therapy on development of new endocarditis, of 250 nosocomial bacteremias, new endocarditis developed in 4% (4/101) who received 2 weeks or less of antistaphylococcal therapy, 4% (3/74) who received 15–27 days of therapy, and 0% (0/33) who received 4 weeks or greater of therapy. (Forty-two patients who died before 14 days after onset of S. aureus bacteremia were excluded because these patients would have had less opportunity to develop endocarditis.) These differences were not statistically significant (p = 0.5, chi-square for trend). In order to detect a difference of 4% at an alpha of 0.05 and power of 80%, 476 patients would have been required in a randomized trial.

Differentiation of S. Aureus Endocarditis from S. Aureus Bacteremia

Based on univariate analyses, valvular heart disease (native valve disease, prosthetic valve disease), prior endocarditis, intravenous drug use, community acquisition, and absence of a recognized portal of entry at the time of presentation were significant risk factors for endocarditis (Table 2). Nonwhite race was significantly associated with endocarditis. When race was subclassified, the frequency of endocarditis in each racial group was as follows: white 10% (35/336), black 16% (25/152), Hispanic 15% (2/13), and Asian 50% (2/4). Persistent bacteremia during the course of therapy was also significantly associated with endocarditis (27% (17/64), p < 0.0001, see Table 2).

Factors demonstrating an association with endocarditis by univariate analyses were then evaluated in a multivariate logistic regression model. Statistically significant associations were found with the following: native valve disease (odds ratio [OR] = 4.5; 95% confidence intervals [CI] = 2.0–9.9; p = .002), the presence of a prosthetic valve (OR = 10.5; 95% CI = 2.5–43.7; p = 0.0012), persistent bacteremia (OR = 7.4; 95% CI = 3.3–16.6; p = .0001), intravenous drug use (OR = 3.2; 95% CI = 1.2–8.6; p = .02), an unidentifiable portal of entry (OR = 3.3; 95% CI = 1.3–6.6; p = .008), a history of prior endocarditis (OR = 10.0; 95% CI = 2.0–50.0; p = .005), community acquisition (OR = 2.9; 95% CI = 1.4–4.9; p = .004), and nonwhite race (OR = 2.5; 95% CI = 1.2–5.3; p = .014).

MRSA Versus MSSA Endocarditis

Sixty-nine percent (44/64) of cases of endocarditis were due to MSSA and 31% (20/64) were due to MRSA (see Figure 1, Table 3), similar to the proportion of all bacteremias that were MSSA (65%) or MRSA (35%) (see Figure 1 and Table 2). As anticipated, MSSA endocarditis was more likely to be community-acquired and to occur in intravenous drug users and in persons who had previously had endocarditis. MRSA endocarditis was more likely to be acquired in the hospital or to occur in patients with renal insufficiency and those on hemodialysis (see Table 3). Patients with MRSA endocarditis were also older than those with MSSA endocarditis (62.3 versus 51.4 yr, respectively), presumably reflecting the preponderance of nosocomial infection. Underlying diseases in each group were comparable and the severity of illness at the time bacteremia was documented (APACHE III score or Pitt Bacteremia Study) was not significantly different (see Table 3).

Characteristics of Patients with MRSA versus MSSA Endocarditis (Persistent Bacteremia and Mortality Were Significantly Higher in MRSA Compared with Mssa Endocarditis.)

Patients with MRSA endocarditis were significantly more likely to experience persistent bacteremia than those with MSSA endocarditis by univariate (p = 0.0001 in Table 3) and multivariate (p = 0.002) analyses. Risk factors entered into multivariate logistic regression analyses to assess differences between MRSA and MSSA endocarditis were those significant by univariate analysis (see Table 3). Only renal insufficiency as defined by creatinine >2 mg/dL at the time of presentation (p = 0.036) occurred significantly more often in MRSA endocarditis by both univariate and multivariate analyses. Prior antibiotic therapy within 2 weeks of the first positive blood culture and nosocomial acquisition were not significant risk factors (see Table 3). The following parameters were associated with persistent bacteremia by univariate analyses: hemodialysis, creatinine >2 mg/dL, total APACHE III score, intravenous drug use, and MRSA (data not shown). By multivariate analyses in a logistic regression model, only MRSA remained statistically significantly associated with persistent bacteremia (p = 0.001).

Mortality in Patients With Endocarditis

The mortality rate was 31% (20/64) and 34% (22/64) at the 30-day (see Figure 1) and 60-day follow-ups, respectively. Numerous logistic regression models were evaluated; the parameters entered into the model were those significant by univariate analyses when assessing mortality at 30 days (Table 4) including age, MRSA as the bacteremic isolate, severity of illness (APACHE III or Pitt bacteremia score), and persistent bacteremia. Acquisition in the community was also inserted into the model because 5 studies suggested it as a risk factor for mortality 7,14,16,20,66. APACHE III score and persistent bacteremia were significantly associated with mortality in all models, while MRSA bacteremia and presence of atrioventricular block on electrocardiogram were either significant (p < 0.05) or approached significance (p = 0.08) in various models (data not shown).

Factors Associated With Mortality (30 days) in 64 Patients with Endocarditis*

Of the 64 patients with endocarditis, only 6 patients underwent resection of the affected valve and all survived; in contrast, 50% (29/58) of patients who were treated with medical therapy only survived (p = 0.02). When 5 patients who died within 7 days of their first positive blood culture were excluded from the analyses, the association remained statistically significant (p = 0.03). No significant difference in severity of illness (APACHE score, Pitt bacteremia score) was found between those patients undergoing surgery versus those receiving medical therapy alone. When the 13 patients with possible endocarditis were excluded and the data reanalyzed, the results were essentially identical (data not shown).


To our knowledge, the current study is the largest prospective multicenter study on S. aureus endocarditis. Its strengths include the size and diversity of a multicenter study. Unlike the studies from an individual hospital conducted over a period of many years 18,57, our study was performed in less than 2 years in 6 hospitals. The epidemiology of S. aureus bacteremia is highly dependent on the patient population served by the facility, including the proportion of dialysis patients, intravenous drug users, and immunosuppressed hosts. Results from a multihospital study should be more generalizable than those from 1 hospital.

The clinical differentiation of S. aureus bacteremia from infective endocarditis has long been a vexing problem for clinicians and has obvious therapeutic and prognostic implications. In the absence of typical Oslerian manifestations (such as changing murmur, splenomegaly, embolic lesions), the clinical diagnosis of infective endocarditis among patients with S. aureus bacteremia can be challenging. Nolan and Beaty 57 reported 3 useful bedside criteria for predicting the presence of infective endocarditis in patients with S. aureus bacteremia: community-acquisition of bacteremia, unapparent primary focus, and metastatic foci. However, the latter 2 criteria may not be highly predictive risk factors for endocarditis. Several studies have documented the presence of S. aureus endocarditis in patients who have a detectable primary focus of S. aureus infection, usually from skin, soft tissue, or bone 10,47. An increased incidence of endocarditis has been observed in populations with a large proportion of intravenous drug users 5,66 or with underlying valvular heart disease 76, or prior endocarditis 3,38,73.

Our study confirmed that valvular heart disease, intravenous drug use 66, community acquisition 5,10, prior endocarditis, and unknown source of infection are risk factors for endocarditis (see Table 2). Unexpectedly, nonwhite (black, Hispanic, Asian) race was also found to be a significant risk factor for endocarditis by both univariate and multivariate analyses, although the reason for this racial predilection is unclear.

The reported prevalence of infective endocarditis in patients with community-acquired S. aureus bacteremia has been quite variable, ranging from 6% to 64%5,17,18,27,31,33,35,47–49,53,66,76. In the current study, in which we precisely determined the denominator as well as the numerator, endocarditis was present in 21% (43/206) of patients with community-acquired S. aureus bacteremia. The incidence of endocarditis reported in intravenous drug users with S. aureus bacteremia has ranged from 38% to 67%5,70; in the current study, the incidence of endocarditis in intravenous drug users with S. aureus bacteremia was 35% (17/48).

Patients with hospital-acquired S. aureus bacteremia have a lower incidence of infective endocarditis (2%-17%) than those with community-acquired bacteremia 2,3,13,33,35,45,47,51,52,59,66,75. In the current study, 5% (10/204) of hospital-acquired S. aureus bacteremias proved to be secondary to endocarditis. Fowler et al 18 diagnosed endocarditis in 28% (17/61) of patients experiencing hospital-acquired S. aureus bacteremia. These results were biased by the exclusion of patients less likely to have endocarditis: 72/176 patients were excluded because of patient refusal to undergo transesophageal echocardiogram or failure of the attending physician to order the test. It is likely that those 72 patients were at lower risk for endocarditis, a point the authors also acknowledged. Therefore, the lower denominator of the remaining patients in the study by Fowler and colleagues may have led to an overestimate of the incidence of endocarditis in hospital-acquired S. aureus bacteremia.

Endocarditis may also develop as a sequela of staphylococcal bacteremia, as occurred in 7 patients in the current study (see Table 1). New endocarditis, as defined in the Methods, developed exclusively in patients with hospital-acquired bacteremia or in hemodialysis patients (see Table 1). An intravascular catheter was the portal of entry in 71% (5/7) of patients. All of the patients who developed new endocarditis had received vancomycin therapy for the initial episode of bacteremia (see Table 1). In 5 patients, a relapse of S. aureus bacteremia occurred at a median of 32 days (range, 17–87 d) following the previous episode of S. aureus bacteremia. In the remaining 2 patients, persistent bacteremia developed after 2 weeks of vancomycin. The pulsed-field gel electrophoresis patterns for the initial isolate and the recurrent isolate were identical for 6 patients; in the seventh patient, 1 of the 2 isolates was not saved. In a study by Fang et al 15, 11% of patients with a prosthetic valve developed endocarditis following an episode of bacteremia; both mitral valve location and staphylococcal etiology were significant risk factors. In the current study, the incidence of new endocarditis following S. aureus bacteremia in the subset of patients with a prosthetic valve was 17% (2/12). Thus, endocarditis must be strongly considered if S. aureus bacteremia develops in a patient with a prosthetic valve.

In a separate analysis published in this same issue 8, nafcillin proved to be superior to vancomycin for therapy of S. aureus bacteremia as measured by bacteriologic cures; persistent bacteremia occurred significantly more often in patients receiving vancomycin even when stratified by severity of illness. Fifteen percent (4/27) had endocarditis despite negative findings by both transthoracic or transesophageal echocardiography. Thus, consideration must be given to treating high-risk patients (see Table 2) with more potent antistaphylococcal therapy (for example, nafcillin rather than vancomycin) or with combination antibiotic therapy for long-term duration of at least 4 weeks, regardless of echocardiographic findings.

The current study failed to show that shorter duration of antibiotic therapy (2 weeks) was a significant risk factor for developing endocarditis, a subject that has been disputed for 2 decades and that was not answered definitively even by a Veterans Affairs Cooperative Study 29,60,61,70. The strong association of persistent bacteremia despite active antibiotic therapy with higher mortality does suggest that more intensive therapy of some type is warranted.

Fowler et al 19 showed that the source of the S. aureus in endocarditis patients was the intravascular catheter in 40% (23/59) of cases. In the current study, the intravascular catheter was the source in only 11% (7/64). This higher incidence of a catheter source for endocarditis in the Duke study is perhaps due to differing definitions of catheter infections in the 2 studies: the criteria for an intravascular catheter as the source of infection in the Duke study were less strict since Fowler used the subjective criterion of “inflammation” at the insertion site but did not use the standard objective criterion of semiquantitative growth on culture 41. Our stricter criteria may have resulted in underdiagnosis of catheter infection.

Hemodialysis patients are at notable risk for S. aureus endocarditis 25 because of the high proportion of such patients who are nasal carriers of S. aureus77, the necessity for regular percutaneous access for dialysis, and the increased calcium deposition in cardiac valves predisposing to valvular dysfunction. Our study revealed endocarditis to be associated with 12% (11/95) of S. aureus bacteremias in hemodialysis patients, an incidence identical to the Duke study 46, but higher than the 3.3% reported in a Danish study 55.

Emergence of MRSA as a common pathogen has added new challenges in management. MRSA endocarditis was responsible for only 3% (16/559) of S. aureus in 9 large series of retrospective reviews of endocarditis reported before 1998 6,14,26,43,44,65,69,71, although an extremely high rate of MRSA was reported in young intravenous drug users from Detroit, Michigan 36,37. In the current study, MRSA was the cause of 31% (20/64) of staphylococcal endocarditis, similar to data recently reported by Fowler et al 19. Persistent bacteremia occurred significantly more frequently in endocarditis caused by MRSA compared with MSSA, as assessed by multivariate (see Results) and univariate analyses (see Table 3). In our study, MRSA endocarditis was more likely to be hospital acquired (including patients on hemodialysis), while MSSA endocarditis was more likely to be community acquired (p < 0.05, see Table 3). Patients with MRSA endocarditis had a longer duration of fever after active antibiotic therapy was initiated, were significantly more likely to have persistent bacteremia, and had a higher incidence of renal insufficiency (see Table 3).

The mortality in our study was significantly higher in MRSA endocarditis compared with MSSA endocarditis (see Table 3). This observation is consistent with the findings of other investigators 51, but is at variance with the study by Fowler et al, who found no differences in mortality between MRSA versus MSSA endocarditis 19,58,63. In a logistic model of multivariate analysis in which the age, total APACHE III score, MRSA, creatinine >2 mg/dL, and community acquisition were entered, only total APACHE III score (p = .004) was statistically significant in predicting mortality, although a trend for mortality was also seen for MRSA (p = .064).

Historically, S. aureus endocarditis has a higher mortality than endocarditis caused by most other microorganisms 42,56,65,67. S. aureus bacteremia and endocarditis have a particularly bad prognosis in the context of old age 14,17,20,24,28,30,31,47,50,52,57,64,66,69,72, rapidly fatal underlying diseases 7,24,34,40,72, and preexisting cardiovascular disease 30,31. An exception is S. aureus tricuspid valve endocarditis occurring in young intravenous drug users, which has a lower mortality 14,69. Congestive heart failure 14,65,69, central nervous system involvement 14,21,62,69, embolic complications 14,69,74, atrial fibrillation 65, atrioventricular block, and renal failure 21 have been identified previously as other risk factors for increased mortality in patients with endocarditis. Most studies have shown that hospital-acquired S. aureus endocarditis was associated with higher mortality than that seen in community-acquired disease 7,14,16,66.

We found that severity of illness (total APACHE III score and Pitt bacteremia score) and persistent bacteremia were associated with mortality in patients with endocarditis by multivariate analysis. The presence of atrioventricular block on electrocardiogram was also a significant risk factor for mortality (see Table 4). Atrioventricular block may result from extension of infection on the heart valve into the conduction system, predisposing the patient to arrhythmias and myocardial damage. Heart block on electrocardiogram has been listed as a relative indication for surgery in native valve endocarditis 1. The current study would likewise suggest that the presence of atrioventricular block should mandate consideration for surgical valve resection. Early surgery has been advocated for S. aureus endocarditis 11. In our study, patients with S. aureus endocarditis had improved outcome if they underwent valvular resection (see Results).

Although we report 1 of the largest prospective series of S. aureus endocarditis, a weakness of the current study was that many patients with S. aureus bacteremia did not undergo echocardiography, which may have led to underdiagnosis of endocarditis.

Staphylococcal endocarditis continues to be among the most studied entities in clinical medicine. Although the technical advance of transesophageal endocardiography has revolutionized the diagnosis of endocarditis, similar advances in prevention and improvements in medical therapy have not kept pace. Furthermore, it appears that endocarditis is at least as prevalent today as it was in previous years; hemodialysis, nosocomial endocarditis resulting from invasive vascular procedures, and intravenous drug abuse have all conspired to push the incidence of endocarditis higher.


S. aureus bacteremia patients at higher risk for S. aureus endocarditis include those with community-acquired bacteremia, unknown portal of entry, underlying valvular heart disease, prior endocarditis, or injection drug use. Nonwhite race, as identified in this study, was also a risk factor, although confirmation of that finding is needed. The high-risk patients with S. aureus bacteremia should be screened for concurrent endocarditis by echocardiography. Persistent bacteremia following initiation of antibiotic therapy was identified as a statistically significant independent risk factor for both endocarditis and mortality, a unique observation to our knowledge not previously reported in other series on S. aureus bacteremia. Given the strength of that observation, strong consideration should be given to repeating blood cultures in all patients with S. aureus bacteremia 3 days after initiation of antistaphylococcal therapy. If persistent bacteremia is documented, the likelihood of underlying S. aureus endocarditis should be viewed as high, even if echocardiography is unrevealing. The necessity for changing the patient’s antistaphylococcal therapy to a more “potent” regimen and/or prolonging the duration of therapy should then be addressed. Since persistent bacteremia has also been documented to occur in patients with metastatic foci of infection (that is, pulmonary emboli, septic arthritis, osteomyelitis), an investigation for occult foci should be undertaken as well.

Factors we identified as being associated with increased mortality in patients with S. aureus endocarditis were infections due to MRSA, persistent bacteremia, and atrioventricular block on electrocardiogram. In patients with S. aureus endocarditis, MRSA as the etiologic agent should be suspected in patients with hospital-acquired endocarditis and renal insufficiency. In those settings, empiric treatment with antibiotics active against MRSA would be justified. If patients with S. aureus endocarditis have any of the above risk factors for increased mortality, consideration should be given to “optimizing” the patient’s antimicrobial regimen (either more “potent” therapy or possibly combination antibiotic therapy), and to exploring the advisability of early surgical intervention, since surgery appeared to improve survival.


We thank Larry M. Baddour, Vance G. Fowler, and James J. Rahal for their critical review.


1. Alsip SG, Blackstone EH, Kirklin JW, Cobbs CG. Indications for cardiac surgery in patients with active infective endocarditis. Am J Med. 78: 138–142, 1985.
2. Arnow PM, Quimosing EM, Beach M. Consequences of intravascular catheter sepsis. Clin Infect Dis. 16: 778–84, 1993.
3. Baddour LM. Twelve-year review of recurrent native-valve infective endocarditis: A disease of the modern antibiotic era. Rev Infect Dis. 10: 1163–1170, 1988.
4. Bannerman TL, Hancock GA, Tenover FC, Miller JM. Pulsed-field gel electrophoresis as a replacement for bacteriophage typing of Staphylococcus aureus. J Clin Microbiol. 33: 551–555, 1995.
5. Bayer AS, Lam K, Ginzton L, Norman DC, Chiu CY, Ward JI. Staphylococcus aureus bacteremia: Clinical, serologic, and echocardiographic findings in patients with and without endocarditis. Arch Intern Med. 147: 457–462, 1987.
6. Benn M, Hagelskjaer LH, Tvede M. Infective endocarditis, 1984 through 1993: A clinical and microbiological survey. J Intern Med. 242: 15–22, 1997.
7. Bryan CS, Kirkhart B, Brenner ER. Staphylococcal bacteremia: Current patterns in nonuniversity hospitals. South Med J. 77: 693–696, 1984.
8. Chang FY, Peacock JE Jr, Musher DM, Triplett P, MacDonald BB, Mylotte JM, O’Donnell A, Wagener MM, Yu VL. Staphylococcus aureus bacteremia: Recurrence and the impact of antibiotic treatment in a prospective multicenter study. Medicine (Baltimore). 82: 333–339, 2003.
9. Chow JW, Yu VL. Combination antibiotic therapy versus monotherapy for gram-negative bacteremia: A commentary. Int J Antimicrob Agents. 11: 7–12, 1999.
10. Cooper G, Platt R. Staphylococcus aureus bacteremia in diabetic patients: Endocarditis and morality. Am J Med. 73: 658–62, 1982.
11. Dismukes WE, Karchmer AW, Buckley MJ, Austen WG, Swartz MN. Prosthetic valve endocarditis: Analysis of 38 cases. Circulation. 48: 365–377, 1973.
12. Durack DT, Lukes AS, Bright DK. New criteria for diagnosis of infective endocarditis: Utilization of specific echocardiographic findings. Am J Med. 96: 200–59, 1994.
13. Ehni WF, Reller LB. Short-course therapy for catheter-associated Staphylococcus aureus bacteremia. Arch Intern Med. 149: 1157–1161, 1989.
14. Espersen F, Frimodt-Moller N. Staphylococcus aureus endocarditis. A review of 119 cases. Arch Intern Med. 146: 1118–1121, 1986.
15. Fang GD, Keys TF, Gentry GO, Harris AA, Rivera N, Getz K, Fuchs PC, Gustafson M, Wong ES, Goetz A, Wagener MM, Yu VL. Prosthetic valve endocarditis resulting from nosocomial bacteremia: A prospective multicenter study. Ann Intern Med. 119: 560–567, 1993.
16. Fernandez-Guerrero ML, Verdejo C, Azofra J, de Gorgolas M. Hospital-acquired infectious endocarditis not associated with cardiac surgery: An emerging problem. Clin Infect Dis. 20: 16–23, 1995.
17. Finklestein R, Sobel JD, Nagler A, Merzbach D. Staphylococcus aureus bacteremia and endocarditis: Comparison of nosocomial and community-acquired infection. J Med. 15: 193–211, 1984.
18. Fowler VG, Li J, Corey GR, Bokey J, Marr KA, Gopal AK, Kong LK, Gottlieb G, Donavan CL, Sexton DJ, Ryan T. Role of echocardiography in evaluation of patients with Staphylococcus aureus bacteremia: Experience in 103 patients. J Am Coll Cardiol. 30: 1072–1078, 1997.
19. Fowler VG Jr, Sanders LL, Kong LK, McClelland RS, Gottlieb GS, Li J, Ryan T, Sexton DJ, Roussakis G, Harrell LJ, Corey GR. Infective endocarditis due to Staphylococcus aureus: 59 prospectively identified cases with follow-up. Clin Infect Dis. 28: 106–114, 1999.
20. Fridmodt-Moller N, Espersen F, Rosdahl VT. Antibiotic treatment of Staphylococcus aureus endocarditis: A review of 119 cases. Acta Med Scand. 222: 175–182, 1987.
21. Gagliardi JP, Nettles RE, McCarty DE, Sanders LL, Corey GR, Sexton DJ. Native valve endocarditis in elderly and younger adult patients: Comparison of clinical features and outcomes with use of the Duke criteria and the Duke endocarditis database. Clin Infect Dis. 26: 1165–1168, 1998.
22. Garner JS, Jarvis W, Emori TG, Horan TC, Hughes JM. Centers for Disease Control definitions for nosocomial infections. Am J Infect Control. 16: 128–140, 1988.
23. Goering RV, Winters MA. Rapid method for epidemiological evaluation of gram-positive cocci by field inversion gel electrophoresis. J Clin Microbiol. 30: 577–580, 1992.
24. Grandsen WR, Eykyn SJ, Phillips I. Staphylococcus aureus bacteremia: 400 episodes in St. Thomas’s Hospital. Br Med J. 288: 300–303, 1984.
25. Hanslik T, Flahault A, Vaillant JN, Boulard JC, Moulonguet-Doleris L, Prinseau J, Balgin A. High risk of severe endocarditis in patients on chronic dialysis. Nephrol Dial Transplant. 12: 1301–1302, 1997.
26. Hecht SR, Berger M. Right-sided endocarditis in intravenous drug users: Prognostic features in 102 episodes. Ann Intern Med. 117: 560–566, 1992.
27. Hill PC, Birch M, Chambers S, Drinkovic D, Ellis-Pegler RB, Everts R, Murdoch D, Pottumarthy S, Roberts SA, Swager C, Taylor SL, Thomas MG, Wong CG, Morris AJ. Prospective study of 424 cases of Staphylococcus aureus bactaeremia: Determination of factors affecting incidence and mortality. Intern Med J. 31: 97–103, 2001.
28. Jensen AG, Espersen F, Skinhoj P, Rosdahl VT, Frimodt-Moller N. Staphylococcus aureus meningitis: A review of 104 nationwide, consecutive cases. Arch Intern Med. 153: 1902–1908, 1993.
29. Jernigan JA, Farr BM. Short-course therapy of catheter-related Staphylococcus aureus bacteremia: A meta-analysis. Ann Intern Med. 119: 304–11, 1993.
30. Jessen Q, Rosendal K, Bulow P, Farber V, Eriksen KR. Changing staphylococci and staphylococcal infections. N Engl J Med. 281: 627–635, 1969.
31. Julander I. Unfavourable prognostic factors in Staphylococcus aureus septicemia and endocarditis. Scand J Infect Dis. 17: 179–187, 1985.
32. Knaus WA, Wagner DP, Draper EA, Zimmerman JE, Bergner M, Bastos PG, Sirio CA, Murphy DJ, Lotring T, Damiano A, Harrell FE, Jr. APACHE III prognostic system: Risk prediction of hospital mortality for critically ill hospitalized adults. Chest. 100: 1619–1636, 1991.
33. Knudsen AM, Rosdahl VT, Espersen F, Frimodt-Moller N, Skinhoj P, Bentzon MW. Catheter-related Staphylococcus aureus infections. J Hosp Infect. 23: 123–131, 1993.
34. Kuikka A, Valtonen VV. Improved outcome of Staphylococcus aureus bacteremia. Infect Dis Clin Pract. 3: 282–287, 1994.
35. Lautenschlager S, Herzong C, Zimmerli W. Course and outcome of bacteremia due to Staphylococcus aureus: Evaluation of different clinical case definitions. Clin Infect Dis. 16: 567–573, 1993.
36. Levine DP, Cushing RD, Jui J, Brown WJ. Community-acquired methicillin-resistant Staphylococcus aureus endocarditis in the Detroit Medical Center. Ann Intern Med. 97: 330–338, 1982.
37. Levine DP, Fromm BS, Reddy BR. Slow response to vancomycin or vancomycin plus rifampin in methicillin-resistant Staphylococcus aureus endocarditis. Ann Intern Med. 115: 674–680, 1991.
38. Levison ME, Kaye D, Mandell GL, Hook EW. Characteristics of patients with multiple episodes of bacterial endocarditis. JAMA. 211: 1355–1357, 1970.
39. Li JS, Sexton DJ, Mick N, Nettles R, Fowler VG Jr, Ryan T, Bashore T, Corey GR. Proposed modifications to the Duke Criteria for the diagnosis of infective endocarditis. Clin Infect Dis. 30: 633–638, 2000.
40. Libman H, Arbeit RD. Complications associated with Staphylococcus aureus. Arch Intern Med. 144: 541–545, 1984.
41. Maki DG, Weise CE, Sarafin HW. A semi-quantitative culture method for identifying intravenous catheter-related infection. N Engl J Med. 296: 1305–1309, 1977.
42. Manford M, Matharu J, Farrington K. Infective endocarditis in a district general hospital. J R Soc Med. 85: 262–266, 1992.
43. Manolis AS, Melita H. Echocardiographic and clinical correlates in drug addicts with infective endocarditis. Implications in vegetation size. Arch Intern Med. 2461–2465, 1988.
44. Mansur AJ, Grinberg M, da Luz PL, Bellotti G. The complications of infective endocarditis: A reappraisal in the 1980s. Arch Intern Med. 152: 2428–2432, 1992.
45. Manzella JP, Kellogg JA, Clark JK. Quantitative colony counts in patients with Staphylococcus aureus bacteremia. J Infect Dis. 155: 1347–1348, 1987.
46. Marr KA, Kong L, Fowler VG, Gopal A, Sexton DJ, Conlon PJ, Corey GR. Incidence and outcome of Staphylococcus aureus bacteremia in hemodialysis. Kidney Int. 54: 1684–1689, 1998.
47. Mirimanoff RO, Glauser MP. Endocarditis during Staphylococcus aureus septicemia in a population of non-drug addicts. Arch Intern Med. 142: 1311–1313, 1982.
48. Mortara LA, Bayer AS. Staphylococcus aureus bacteremia and endocarditis. Infect Dis Clin North Am. 7: 53–68, 1993.
49. Musher DM, Lamm N, Darouiche RO, Young EJ, Hamill RJ, Landon GC. The current spectrum of Staphylococcus aureus infection in a tertiary care hospital. Medicine (Baltimore). 73: 186–208, 1994.
50. Musher DM, McKenzie SO. Infections due to Staphylococcus aureus. Medicine (Baltimore). 56: 383–409, 1977.
51. Mylotte JM, Beam TR Jr, Allen JC. Staphylococcus aureus bacteremia: A prospective study. South Med J. 76: 1131–1135, 1983.
52. Mylotte JM, McDermott C, Spooner JA. Prospective study of 114 consecutive episodes of Staphylococcus aureus bacteremia. Rev Infect Dis. 9: 891–907, 1987.
53. Mylotte JM, Tayara A. S. aureus bacteremia: Prediction of 30 day mortality in a large cohort. Clin Infect Dis. 3: 1170–1174, 2000.
54. National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial disc susceptibility tests. NCCLS document M2-A4, Villanova, PA: NCCLS, 1990.
55. Nielsen J, Kolmos HJJ, Espersen F. Staphylococcus aureus bacteremia among patients undergoing dialysis—focus on dialysis catheter-related cases. Nephrol Dial Transplant. 13: 139–145, 1998.
56. Nissen H, Nielsen PF, Frederiksen M, Helleberg C, Nielsen JS. Native valve infective endocarditis in the general population: A 10-year survey of the clinical picture during the 1980s. Eur Heart J. 13: 872–877, 1992.
57. Nolan CM, Beaty HN. Staphylococcus aureus bacteremia: Current clinical patterns. Am J Med. 60: 495–500, 1976.
58. Pujol M, Pena C, Pallares R, Ayats J, Ariza J, Gudiol F. Risk factors for nosocomial bacteremia due to methicillin-resistant Staphylococcus aureus. Eur J Clin Microbiol Infect Dis. 13: 96–102, 1994.
59. Raad II, Sabbagh MF. Optimal duration of therapy for catheter-related Staphylococcus aureus bacteremia: A study of 55 cases and review. Clin Infect Dis. 14: 75–82, 1992.
60. Rahal JJ. Short-term therapy for intravascular catheter-related Staphylococcus aureus bacteremia. Clin Infect Dis. 133: 1946–1947, 2002.
61. Rahal JJ Jr, Chan YK, Johnson G. Relationship of staphylococcal tolerance, teichoic acid antibody, and serum bactericidal activity to therapeutic outcome in Staphylococcus aureus bacteremia. Am J Med. 81: 43–52, 1986.
62. 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. 102: 379–386, 1997.
63. Romero-Vivas J, Rubio M, Fernandez C, Picazo JJ. Mortality associated with nosocomial bacteremia due to methicillin-resistant Staphylococcus aureus. Clin Infect Dis. 21: 1417–1423, 1995.
64. Rosendal K, Jessen OFV, Bentzon MW. Frequency, phage types and antibiotic resistance of Staphylococcus aureus isolated from blood cultures in Denmark, 1975–1981. Scand J Infect Dis. 41: 19–29, 1983.
65. Sanabria TJ, Alpert JS, Goldberg R, Pape LA, Cheeseman SH. Increasing frequency of staphylococcal infective endocarditis: Experience at a university hospital. Arch Intern Med. 150: 1305–1309, 1990.
66. Shah M, Watanakunkorn C. Changing patterns of Staphylococcus aureus bacteremia. Am J Med Sci. 278: 115–121, 1979.
67. Skehan JD, Murray M, Mills PG. Infective endocarditis: Incidence and mortality in the North East Thames Region. Br Heart J. 59: 62–68, 1988.
68. Tenover FC, Arbeit RD, Goering RV, Mickelsen PA, Murray BE, Persing DH, Swaminathan B. Interpreting chromosomal DNA restriction patterns produced by pulsed-field electrophoresis: Criteria for bacterial strain typing. J Clin Microbiol. 33: 2233–2239, 1995.
69. Watanakunakorn C. Staphylococcus aureus endocarditis at a community teaching hospital, 1980 to 1991. An analysis of 106 cases. Arch Intern Med. 154: 2330–2335, 1994.
70. Watanakunakorn C, Baird IM. Prognostic factors in Staphylococcus aureus endocarditis and results of therapy with a penicillin and gentamicin. Am J Med Sci. 273: 133–139, 1977.
71. Watanakunakorn C, Burkert T. Infective endocarditis at a large community teaching hospital, 1980–1990. A review of 210 episodes. Medicine (Baltimore). 72: 90–102, 1993.
72. Watanakunakorn C, Tan JS, Phair JP. Some salient features of Staphylococcus aureus endocarditis. Am J Med. 54: 473–478, 1973.
73. Welton DE, Young JB, Gentry WO, Raizner AE, Alexander JK, Chahine RA, Miller RR. Recurrent infective endocarditis. Analysis of predisposing factors and clinical features. Am J Med. 66: 932–938, 1979.
74. Werner GS, Schulz R, Fuchs JB, Andreas S, Prange H, Ruschewski W, Kreuzer H. Infective endocarditis in the elderly in the era of transesophageal echocardiography: Clinical features and prognosis compared with younger patients. Am J Med. 100: 90–97, 1996.
75. Whimbey E, Kiehn TE, Brannon P, Benezra D, Armstrong D. Clinical significance of colony counts in immunocompromised patients with Staphylococcus aureus bacteremia. J Infect Dis. 155: 1328–1330, 1987.
76. Wilson R, Hamburger M. Fifteen years’ experience with staphylococcus septicemia in a large city hospital: Analysis of fifty-five cases in the Cincinnati General Hospital 1940 to 1954. Am J Med. 22: 437–457, 1957.
77. Yu VL, Goetz A, Wagener MM, Smith PB, Rihs JD, Hanchett J, Zuravleff JJ. Staphylococcus aureus nasal carriage and infection in patients on hemodialysis: Efficacy of antibiotic prophylaxis. N Engl J Med. 315: 91–96, 1986.

Cited By

This article has been cited 5 time(s).

Staphylococcus aureus Bacteremia: Recurrence and the Impact of Antibiotic Treatment in a Prospective Multicenter Study
Chang, F; Peacock, JE; Musher, DM; Triplett, P; MacDonald, BB; Mylotte, JM; O’Donnell, A; Wagener, MM; Yu, VL
Medicine, 82(5): 333-339.
PDF (121) | CrossRef
The American Journal of the Medical Sciences
Impact of Methicillin Resistance on Clinical Features and Outcomes of Infective Endocarditis Due to Staphylococcus aureus
Hsu, R; Chu, S
The American Journal of the Medical Sciences, 328(3): 150-155.

PDF (146)
Endocarditis Caused by Staphylococcus aureus: A Reappraisal of the Epidemiologic, Clinical, and Pathologic Manifestations With Analysis of Factors Determining Outcome
Fernández Guerrero, ML; González López, JJ; Goyenechea, A; Fraile, J; de Górgolas, M
Medicine, 88(1): 1-22.
PDF (1339) | CrossRef
The American Journal of the Medical Sciences
Staphylococcus aureus Endocarditis: The Grady Memorial Hospital Experience
Sherwood, M; Smith, D; Crisel, R; Veledar, E; Lerakis, S
The American Journal of the Medical Sciences, 331(2): 84-87.

PDF (110)
Infectious Diseases Consultation Lowers Mortality From Staphylococcus aureus Bacteremia
Lahey, T; Shah, R; Gittzus, J; Schwartzman, J; Kirkland, K
Medicine, 88(5): 263-267.
PDF (126) | CrossRef
© 2003 Lippincott Williams & Wilkins, Inc.