Native valve endocarditis due to coagulase-negative staphylococci

Native Valve Endocarditis due to Coagulase-Negative Staphylococci Clinical and Microbiologic Features

GREGORY

M. CAPUTO,

M.D.*

Boston, Massachusetts GORDON Richmond,

L. ARCHER,

M.D.

Virginia

STEPHEN B. CALDERWOOD, M.D. MARK J. DINUBILE, M.D.7 ADOLF W. KARCHMER, M.D. Boston, Massachusetts

From the Infectious Disease Sections, Department of Medicine, the New England Deaconess Hospital, the Massachusetts General Hospital, and Harvard Medical School, Boston, Massachusetts, and the Infectious Disease Division, Department of Medicine, Medical College of Virginia Hospital, Virginia Commonwealth University, Richmond, Virginia. Requests for reprints should be addressed to Dr. Adolf W. Karchmer, New England Deaconess Hospital, 185 Pilgrim Road, Boston, Massachusetts 02215. Manuscript submitted December 1, 1988, and accepted April 28, 1987. *Current address: Medical Center of Delaware, Newark, Delaware 19718. tCurrent address: Division of Infectious Disease, Cooper Hospital-University Medical Center, 1 Cooper Plaza, Camden, New Jersey 08103.

Twenty-one patients with native valve endocarditis caused by coagulase-negative staphylococci were studied; 14 had pre-existing valvular or congenital heart disease. Although commonly subacute in presentation, complications of endocarditis were frequent: arterial emboli in five patients, new electrocardiographic conduction system abnormalities in nine, congestive heart failure in eight, annular or myocardial abscesses in five, and disruption of valve leaflets in three. Cures were achieved in 10 of 12 patients treated medically and seven of nine treated surgically. In microbiologic studies of 16 coagulase-negative staphylococci from patients with endocarditis, only eight were identified as Staphylococcus epidermidis. All isolates were susceptible to vancomycin. Antibiotic resistance (methicillin, four isolates; gentamicin, two isolates; rifampin, one isolate) was usually associated with nosocomial acquisition of endocarditis. Rather than representing contamination, coagulase-negative staphylococci in blood cultures may indicate lifethreatening endocarditis. However, with careful attention to the selection of antibiotics for therapy and to the occurrence of heart failure due to intracardiac complications, treatment of this form of endocarditis is frequently successful. Organisms must always be tested for cryptic resistance to beta-lactam antibiotics. Valve replacement may be required frequently. Coagulase-negative staphylococci are the most common organisms isolated from patients with prosthetic valve endocarditis [ 1,2]. In contrast, these organisms rarely cause native valve endocarditis, accounting for only 1 to 3 percent of cases [3]. Because of this rarity, the clinical and laboratory features of native valve endocarditis due to coagulase-negative staphylococci have not been well characterized. To define the clinical features of native valve endocarditis caused by coagulase-negative staphylococci, we reviewed the hospital records of 2 1 patients with this infection. In addition, we determined the species of 16 coagulasenegative staphylococci that had caused native valve endocarditis and studied their antimicrobial susceptibility.

PATIENTS AND METHODS The hospital records of 21 patients treated for coagulase-negative staphylococcal endocarditis on a native heart valve between 1965 and 1985 at either the Massachusetts General Hospital, the New England Deaconess Hospital, or the Medical College of Virginia Hospital were reviewed. Endocarditis was considered present if coagulase-negative staphylococci were isolated from two or more blood cultures and if there was either a compatible clinical syndrome (a new changing cardiac murmur, signs of peripheral embolism, or cutaneous manifestations of endocarditis) or histopathologic

October 1987

The American Journal of Medicine

Volume 83

619

COAGULASE-NEGATIVE

TABLE

I

STAPHYLOCOCCAL

ENDOCARDITIS-CAPUTO

ET AL

Laboratory Results In Patients with Endocarditis Caused by CoaguiaseNegative Staphylococci

LaboratoryResult

inoculum to Mueller-Hinton agar. The absence of growth on subculture was indicative of 99.9 percent killing and ailowed determination of the minimal bactericidal concentra-

NumberPresent/NumberTested Percent

White blood cell 9/21 43 count I 10,500 Anemia* 16121 76 Erythrocyte sedimentation 11112 92 rate L 20 mm/hour Rheumatoid factor 4112 33 Hematuria 7121 33 Red blood cell casts in urine 2121 10 Blood cultures 2 3 positive 20/21 95 100% positive 13/21 62 Vegetations by echo7/15 47 cardiogram Hematocrit less than 38 percent in men and less than 36 percent in women. l

evidence of endocarditis found at surgical or postmortem evaluation. Organisms were identified as coaguiase-negative staphylococci by the clinical microbiology laboratory of the respective hospital. Coaguiase-negative staphylococci were confirmed by Gram stain and colony morphology, ability to ferment glucose (fermentation variable for some species), production of cataiase, and inability to coagulate rabbit plasma. The antibiotic susceptibility of those staphylococci isolated from the patients who comprise the clinical review was determined in the clinical microbiology laboratory using the methodology recommended at the time, primarily the standardized disk diffusion technique [4]. Sixteen coaguiase-negative staphylococci that had caused native valve endocarditis were studied microbioiogicaliy. Six of these organisms had been isolated from patients with endocarditis who were among the 21 cases reviewed (two from each of the hospitals’ contributing cases). The remaining 10 organisms had been referred to one of us (G.L.A.) for special studies. The patients from whom these organisms had been isolated had endocarditis by our case definition but were treated at other institutions. Consequently, their complete hospital records were not available for our review. Speciation was performed according to the method of Kioos and Schieifer [5]. Peniciiiinase production was determined by the starch-iodine method following induction with 1 pg/mi of methiciiiin or penicillin. Methiciiiin resistance was evaluated by spreading 0.1 ml of an overnight culture of the organisms onto the surface of a Mueller-Hinton agar plate containing 12.5 pg of methiciliin per ml and incubating the plate at 30°C for 72 hours. Strains that failed to grow were considered susceptible to methiciliin, whereas growth was indicative of resistance to methiciliin [6]. Minimal inhibitory concentrations (MiCs) of selected antibiotics were determined by the microtiter broth dilution technique using an inocuium of 5 X IO6 colony-forming units/ml and cation-supplemented MueiierHinton broth [7]. Subcultures of microtiter wells were performed with a steel replicator that delivered 0.001 ml of

620

October

1967

The American

Journal

of Medicine

Volume

tions [7]. RESULTS Clinical

Features.

Among

our 21 patients,

there

were

17 men and four women. Their mean age was 53 years with a range of 20 to 85 years. Pre-existing valvular heart disease or a ventricular septal defect was present in 14 patients (67 percent). Only two patients acquired endocarditis while hospitalized; the remaining 19 had community acquired infection. The symptoms and physical findings of our patients were similar to those found in infective endocarditis caused by other bacteria. Fever, malaise, and unexplained weight loss were the most common symptoms. Symptoms of congestive heart failure were present in eight patients (38 percent). Temperature elevated to at least 10 1.4OF was noted in 16 patients. All patients had a heart murmur; three had splenomegaly, and five had peripheral cutaneous signs associated with endocarditis. One patient was admitted with signs of embolic stroke, and embolic central nervous system infarcts occurred during hospitalization in three other patients. Laboratory abnormalities in these patients were also similar to those noted in patients with subacute endocarditis caused by other organisms (Table I). Notably, in 20 of the 21 patients, three or more blood cultures yielded coagulase-negative staphylococci, and in 62 percent of the patients, all of the blood culture results were positive. Vegetations were demonstrated in seven of the 15 patients studied by echocardiography. Based upon the characteristics of murmurs, echocardiography, or direct examination of the heart valves, isolated aottic valve infection was present in eight patients, isolated mitral valve involvement in six patients, infection of both the mitral and aortic valves in three patients, and tricuspid infection in one patient. Endocarditis was associated with a ventricular septal defect in three patients; the actual site of infection in these three patients could not be determined. Complications of endocarditis were observed frequently. Either systemic embolization, congestive heart failure, or new conduction system abnormalities were noted in 14 of the patients (67 percent). Four patients had embolic central nervous system infarcts and one had an embolic occlusion of the popliteal artery. Valvular dysfunction associated with new or worsening congestive heart failure was noted in eight patients (38 percent). Heart block of varying degrees that had not been present on the mOSt

recent electrocardiogram (within six months of the onset of endocarditis) was noted in nine patients (43 percent). The site of infection was examined patients. Abscess in paravalvular

63

in seven of these nine tissue was found in

COAGULASE-NEGATIVE

three patients and extensive disruption of two cusps of the aortic valve without overt paravalvular abscess was noted in another. Vegetations without evidence of invasion were found in the remaining three patients. Two patients had glomerulonephritis. Although serum creatinine concentrations remained normal in these patients, red blood cell casts were found in their urine. The invasive and destructive potential of coagulasenegative staphylococci was demonstrated by the pathologic findings noted at surgery or postmortem examination. Infection extended beyond the valve leaflet in five patients, causing annular abscess in two and a myocardial abscess in three. Infection was confined to the leaflet in the other five patients in whom the site of infection was examined. However, in two of these patients, the valve leaflet was perforated and in a third two cusps of the aortic valve were destroyed. Antibiotic therapy based upon the results of routine susceptibility tests resulted in a cure in 17 (8 1 percent) of 2 1 patients (Table II). Relapse of infection did not occur. For 13 of the survivors, duration of follow-up ranged from five months to seven years; one patient was lost to followup after discharge. In three other survivors, follow-up was abbreviated by death or cardiac surgery within two months after discharge. In these three patients, there was no evidence of endocarditis at surgery or postmortem examination. Methicillin-resistant coagulase-negative staphylococci were recovered from the two patients with nosocomially acquired endocarditis (one of the organisms was included in our microbiologic studies) and one patient with community acquired endocarditis. The coagulasenegative staphylococci isolated from the remaining 18 patients with community acquired endocarditis appeared susceptible to penicillinase-resistant penicillins and cephalosporins. Although there were no significant differences in outcome based on therapy, the results of various therapeutic efforts are of interest. The duration of antibiotic therapy in patients in whom endocarditis was cured averaged 4 1.9 days (median 42 days, range 27 to 59 days). A betalactam agent was used as the primary antibiotic in 16 patients; seven of these patients also received an aminoglycoside and/or rifampin for at least IO days. Infection was cured in 12 of the 16 patients. The beta-lactam antibiotic was penicillin G in six, a semisynthetic penicillinase-resistant penicillin in five (two died), cephalothin in one (one died), and in sequence a cephalosporin and a semisynthetic penicillinase-resistant penicillin in four (one died). Among the five patients receiving vancomycin as the primary antibiotic, two patients treated with vancomytin (except for brief periods of beta-lactam antibiotic therapy) had cures, as did three patients who received vancomycin in combination with gentamicin. Ten patients receiving three or more days of either gentamicin and/or rifampin as a supplement to the primary antibiotic were

October

TAf3LE If

STAPHYLOCOCCAL

ENDOCARDITIS-CAPLJTO

ET AL

utcome of ~nd~~arditis Caused by Coagulase-Negative Staphylococci

Type of Therapy

Number Cured/Number

Primary antibiotic Beta-lactam Vancomycin Antibiotic regimen Single agent Combination therapy* Valve replacement surgery Surgical intervention No surgery * Primary antibiotic side and/or rifampin

Percent

Treated

12/16 515

75 100

7111 10/10

64 100

719

78 83

10112

therapy supplemented with for three or more days.

an aminoglyco-

considered to have received combination therapy. Supplemental gentamicin was administered to one patient for only three days, whereas in the remaining nine patients, the average duration of treatment with additional antibiotics was 31 days (range 10 to 52 days). All 10 patients treated with combination therapy had cures. Four patients received a penicillin plus an aminoglycoside, three were given vancomycin plus gentamicin, two received oxacillin plus gentamicin and rifampin, and one was treated with penicillin G plus rifampin. Seven of nine patients undergoing valve replacement during active endocarditis had cures, whereas 10 of the 12 treated medically survived. New and progressive congestive heart failure was the indication for surgery in eight patients, whereas the combination of chronic congestive heart failure, severe aortic and mitral valve regurgitation, and new onset first-degree heart block prompted surgery in the ninth. Surgery was performed after one day of antibiotic therapy in one patient and after nine days of therapy in another; cures were achieved in both. The average duration of preoperative antimicrobial therapy in the remaining seven patients was 22.7 days (range 14 to 37 days). Culture and Gram stain of cardiac valves removed surgically were performed in eight cases. Five valves gave negative results by culture but contained gram-positive cocci on smear; two valves (after 18 and 24 days of antibiotic therapy) yielded coagulase-negative staphylococci on culture as well as having organisms visible on smear, and one specimen (after 17 days of antibiotic therapy) was negative by smear but the culture yielded coagulase-negative staphylococci. The average duration of antibiotic therapy after surgery among patients who survived was 30 days (range 17 to 43 days). Microbiologic Studies. The species of the coagulasenegative staphylococci that caused endocarditis in the 2 1 patients who were the subjects of the clinical review were not routinely determined. However, the species were

1987

The American

Journal

of Medicine

Volume

83

621

COAGULASE-NEGATIVE

TABLE

III

STAPHYLOCOCCAL

Antibiotic

ENDOCARDITIS-CAPUTO

Susceptibilities

Antibiotic

ET AL

of 16 Coagulase-Negative

Staphylococci

MIC50”

Gentamicin Methicillin Nafcillin Vancomycin Rifampin Teichoplanin Cefazolin Cefamandole Cephalothin Clindamycin

0.05 6.3 3.1 6.3 0.025 3.1 3.1 1.6 1.6 3.1

* MIC is the minimal inhibitory concentration tested, respectively. t Parentheses contain minimal bactericidal designated as the same in parentheses.

Endocardltis

fW90* 0.1 25 6.3 12.5 0.1 6.3 25 2.5 6.3 > 100

(O.l)? (same) (same) (same) (same) (same) (same) (same) (same) (0.8) expressed

Causing

as pglml;

concentrations

(MBC)

MI&

determined for the 16 organisms studied in vitro. Eight of these organisms were identified as Staphylococcus epidermidis, three as Staphylococcus warneri, two as Staphylococcus cohnii, and one each as Staphylococcus saprophyticus, Staphylococcus hemolyticus, and Staphylococcus hominis. Among the 16 patients from whom isolates were obtained for study, three had nosocomially acquired endocarditis; two of these three isolates were S. epidermidis and one was S. hominis. The range of MlCs and the MI& and MI&,, of the isolates indicate the high degree of susceptibility of these organisms to rifampin, gentamicin, vancomycin, and teichoplanin (Table III). One isolate from a nosocomially acquired infection was resistant to rifampin (MIC greater than 6.3 pg/ml) and to gentamicin (MIC 12.5 pg/ml), and relatively resistant to vancomycin (MIC 12.5 ,ug/ml). Another isolate, also from nosocomially acquired endocarditis, was resistant to gentamicin (MIC 12.5 pg/ml) and relatively resistant to teichoplanin (MIC 12.5 pug/ml). Eleven strains produced penicillinase. Only three isolates were methicillin-resistant; these were the three organisms recovered from patients with nosocomially acquired endocarditis. These three organisms skewed the MI& for nafcillin, cephalothin, and cefamandole to higher values. These three strains had high MlCs for beta-lactam antibiotics, whereas the other 13 strains were susceptible to these antibiotics. The MIC& (range) of the 13 methicillin-susceptible organisms was 1.6 pg/ml(O.4 to 6.3 pgl ml) for cephalothin, 3.1 pg/ml (0.8 to 6.3 pg/ml) for cefamandole, and 6.3 pg/ml (0.4 to 6.3 pg/ml) for nafcillin.

(0.2) (50) (12.5) (same) (same) (same)

0.013-12.5 (same) 1.6->I00 (same) 0.4->lOO (0.8->lOO) 1.6-12.5 (same) 0.013->6.3 (same) 0.2-12.5 (same) 0.8->I00 (same) 0.8-50 (0.8-100) 0.4-50 (0.4->lOO) 0.2->lOO (0.4->lOO)

(same) (same) (same) (same)

and MICgo indicate

expressed

Rancle”

as pg/ml.

values

for 50 and 90 percent

If the MIC

and MBC

are

of the organisms

identical,

the MBC

is

Coagulase-negative staphylococci are the most common bacteria to contaminate blood cultures [8]. Therefore, physicians frequently must decide whether a coagulasenegative staphylococcus isolated from a blood culture represents infection or contamination. The isolation of

coagulase-negative staphylococci from multiple blood cultures obtained from a patient with valvular or congenital heart disease suggests, in the absence of an intravascular foreign body, the diagnosis of endocarditis. The presumption of infection is further strengthened when the isolates from separate cultures are shown to represent a unique strain [9-l I]. Determination of species, antibiogram analysis, and phage typing have been evaluated as methods of identifying unique organisms. However, because of shared characteristics, none of these tests when used alone is sufficient to characterize a specific organism [12]. Identical plasmid profiles, as determined by agarose gel electrophoresis of lysed organisms, can accurately identify unique coagulase-negative staphylococci [ IO,1 11. When the results of only a few blood cultures are positive and the diagnosis of coagulase-negative staphylococcal native valve endocarditis is not certain, demonstrating that the blood culture isolates are identical may strengthen the diagnostic impression. In our patients, as well as in previous reports, native valve endocarditis caused by coagulase-negative staphylococci was generally subacute in presentation [9,13151. The clinical and laboratory findings were similar to those found in subacute endocarditis caused by other bacteria. Prior valvular disease or congenital heart disease was noted in 67 percent of our patients and was common among patients in other series [9,13,14]. Our study indicates that some patients with native valve endocarditis due to coagulase-negative staphylococci have invasive and destructive infection that results in major complications, including congestive heart failure as a consequence of valvular damage or heart block due to disruption of the conduction system. In 38 percent of our patients, new or worsening heart failure developed. In four previous studies, in 11 of 43 evaluable patients (26 percent), congestive heart failure developed during the course of endocarditis [9,13- 151. Among our 10 patients in whom the site of infection was examined, five were noted to have annular or myocardial abscesses and three

622

83

COMMENTS

October

1987

The American

Journal

of Medicine

Volume

COAGULASE-NEGATIVE

had perforation or disruption of valve leaflets. Similarly, Tuazon and Miller [I 31 noted valve perforation and annular abscess in one of their nine episodes of endocarditis. Geraci et al [ 141 described valve perforation in two of their 15 patients. In nine of our patients (43 percent), new or changing conduction system abnormalities developed. The potentially destructive nature of these organisms is thus confirmed. The differences between coagulase-negative staphylococci causing infection of prosthetic and native cardiac valves deserve comment. Prosthetic valve endocarditis is most freqeuntly caused by hospital-acquired methicillinresistant S. epidermidis [1,6]. In contrast, our data both from the clinical review and the microbiologic study indicate that most cases of native valve endocarditis are community acquired and that the infecting strains are commonly susceptible to methicillin. In total, between the clinical review and the microbiologic study, there were five infections caused by methicillin-resistant strains among 31 unique episodes of native valve endocarditis due to coagulase-negative staphylococci (organisms from six episodes reviewed clinically were studied microbiologically; however, these episodes are counted only once). Four episodes of endocarditis were acquired nosocomially and all were caused by methicillin-resistant organisms. The association of antibiotic susceptibility and the place wherein endocarditis was acquired is not absolute. One of our 19 cases of community acquired endocarditis was caused by a methicillin-resistant strain. Lockman et al [ 161 also reported a community acquired episode of native valve endocarditis that was caused by a methicillin-resistant S. epidermidis strain. Consequently, all coagulase-negative staphylococci causing endocarditis must be rigorously tested to establish their susceptibility to beta-la&m antibiotics [6,7]. The distribution of species of coagulase-negative staphylococci that cause native and prosthetic valve endocarditis also differs. Only eight of the 16 coagulasenegative staphylococci causing native valve endocarditis were identifed as S. epidermidis. S. saprophyticus, a notably unusual cause of infection outside the urinary tract, caused one case of endocarditis. Several reports have noted species other than S. epidermidis as causes of native valve endocarditis [ 17,181. In contrast, of 55 coagulase-negative staphylococci causing prosthetic valve endocarditis, 53 were S. epidermidis [6]. Penicillinase production was detected in 11 of the 16 strains we studied. This contrasts sharply with the high frequency of penicillin-susceptible organisms from native valve endocarditis in two previous series [ 14,151. These investigators had not, however, attempted to induce penicillinase production prior to testing; hence, susceptibility to penicillin may have been overestimated. Although increased gentamicin resistance among nosocomial isolates of coagulase-negative staphylococci has been noted

October

STAPHYLOCOCCAL

ENDOCARMTIS-CAPUTO

ET AL

[ 191, 88 percent of our strains were susceptible to gentamicin. High-level resistance to vancomycin among coagulase-negative staphylococci is rare [20]. However, the MI&, of vancomycin for our strains was relatively high, although within the achievable range for serum concentrations. Our patients with native valve endocarditis caused by coagulase-negative staphylococci were treated with a variety of regimens. Because of the small number of patients treated with a specific regimen, it is not possible to establish the most effective treatment. However, several important observations can be made. First, the mortality rate for coagulase-negative staphylococcal native valve endocarditis was relatively low (19 percent) for our patients. Although Tuazon and Miller [ 131 noted a 33 percent mortality rate in their small series, Geraci et al [ 141 and Williams et al [9] reported mortality rates of 13 and 20 percent, respectively. Second, our data suggested that a beta-lactam agent (usually penicillin G or a semisynthetic penicillinase-resistant penicillin contingent upon the susceptibility of the organism) will be effective as the primary antibiotic in most cases of community acquired native valve endocarditis caused by methicillin-susceptible coagulase-negative staphylococci. Other investigators have reported the efficacy of beta-lactam antibiotics in this setting [9,13]. The efficacy of beta-lactam antibiotics in patients with native valve endocarditis contrasts sharply with the poor results reported when coagulasenegative staphylococcal prosthetic valve endocarditis is treated with similar antimicrobials. In a recent study of patients with prosthetic valve endocarditis caused by methicillin-resistant coagulase-negative staphylococci, cures were achieved in only 50 percent of patients treated with a beta-la&m agent used alone or in combination with an aminoglycoside and/or rifampin [6]. In contrast, when patients with this type of prosthetic endocarditis were treated with vancomycin in combination with rifampin and/or an aminoglycoside, cure rates of 80 percent were achieved [6,2-l]. The divergent results noted for beta-lactam antibiotic therapy of coagulase-negative staphylococcal endocarditis on native valves versus that on prosthetic valves are consistent with the finding that most coagulase-negative staphylococci causing native valve infection are methicillin susceptible, whereas most of those causing prosthetic valve infection are methicillin resistant. The experience reported for the treatment of prosthetic valve endocarditis [6] and the inefficacy of beta-lactam antibiotic therapy for experimental endocarditis due to methicillin-resistant coagulase-negative staphylococci [22] argue strongly that vancomycin be used to treat native valve endocarditis caused by methicillin-resistant coagulase-negative staphylococci. Finally, our data suggest that combination therapy might be more effective than single-drug therapy. When our data are combined with those of previous series, the same trend,

1987

The American

Journal

of Medicine

Volume

83

623

COAGULASE-NEGATIVE

TABLE IV

STAPHYLOCOCCAL

ENDOCARDITIS-CAPUTO

ET AL

Single-Agent versus Combinatlon Therapy for Natlve Valve Endocardltis Caused by Coagulase-Negative Staphylococci Study Current Series

[14]

[13]

[15]

Total (percent)

Single agent Number cured Number treated Combination Number cured Number treated

7 11

9 10

3 6

2 2

21 (72)” 29

10 10

4 4

3 3

6 8

23 (92)” 25

* p > 0.05

with

(chi-square

Yates’

correction).

although not statistically significant, is noted (Table IV). Combination therapy, particularly when the supplemental antibiotic is an aminoglycoside, is often associated with an increased incidence of drug toxicity as compared with therapy with a single drug [23]. Therefore, it may be wise to reserve combination therapy for infection caused by more resistant organisms or for patients with a suboptimal response to initial single-drug treatment. The frequent and high degree of susceptibility to betala&am antibiotics among coagulase-negative staphylococci causing native valve endocarditis and the efficacy of beta-lactam therapy for these infections that we and others [13-l 51 have noted, suggest that endocarditis caused by a methicillin-susceptible strain can be treated with a semisynthetic penicillinase-resistant penicillin. However, when contrasted with the efficacy of vancomytin, the failure of therapy with semisynthetic penicillinaseresistant penicillins and cephalosporins for prosthetic valve endocarditis [6] and experimental endocarditis [22] caused by methicillin-resistant coagulase-negative

staphylococci is a stern reminder of the importance of precise antibiotic susceptibility testing for these organisms. It is important to note that routine susceptibility tests of coagulase-negative staphylococci, as pet-formed in clinical laboratories, often overestimate the susceptibility of these bacteria to beta-lactam antibiotics, particularly cephalosporins [7,24]. If methicillin-resistant coagulasenegative staphylococci are misidentified as methicillin susceptible and a beta-lactam antibiotic is utilized, inadequate antibiotic therapy is likely [6]. In nine of our patients, complications of endocarditis developed that prompted valve replacement surgery during active infection. The frequency of invasive and destructive infection noted in our series, as well as in those reported by others [9,13-151, suggests that surgery will be an important adjunct to antibiotic therapy in many patients. Although coagulase-negative staphylococci isolated from blood cultures are frequently contaminants, this series, as well as several others [9,13- 151, documents the potential for these organisms to cause native valve endocarditis. In contrast to endocarditis casued by Staphylococcus aureus, endocarditis caused by coagulase-negative staphylococci is frequently cured. To achieve an optimal outcome when treating patients with coagulasenegative staphylococcal endocarditis, careful attention must be directed not only to the antibiotic susceptibility of the infecting organism but also to the early detection of intracardiac complications that require surgical intervention. ACKNOWLEDGMENT We wish to thank Wesley E. Kloos for determining that three coagulase-negative staphylococci were S. warneri and that one was S. saprophyticus. We also thank Ruth Colman and Robin Staub for clerical assistance.

REFERENCES 1.

2.

3.

4.

5.

624

Calderwood SB, Swinski LA, Waternaux CM, Karchmer AW, Buckley MJ: Risk factors for the development of prosthetic valve endocarditis. Circulation 1985; 72: 31-37. Mayer KH, Schoenbaum SC: Evaluation and management of prosthetic valve endocarditis. Prog Cardiovasc Dis 1982; 25: 43-54. Kaye D: Infecting microorganisms. in: Kaye D, ed. Infective endocarditis. Baltimore: University Park Press, 1976; 4354. Matsen JM, Barry AL: Susceptibility testing: diffusion test procedures. In: Lenette EH, Truant JP, eds. Manual of clinical microbiology. Washington: American Society of Microbiology, 1974; 410-417. Kloos WE, Schleifer KH: Simplified scheme for routine identification of human Staphylococcus species. J Clin Microbiol 1975; 1: 82-88.

October

1967

The American

Journal

of Medicine

Volume

6.

7.

8.

9.

63

Karchmer AW, Archer GL, Dismukes WE: Staphylococcus epidermidis causing prosthetic valve endocarditis: microbiologic and clinical observations as guides to therapy. Ann Intern Med 1983; 98: 447-455. Archer GL: Antimicrobial susceptibility and selection of resistance among Staphylococcus epidermidis isolates recovered from patients with infections of indwelling foreign devices. Antimicrob Agents Chemother 1978; 14: 353-359. Weinstein MP, Reller LB, Murphy JR, Lichtenstein KA: The clinical significance of positive blood cultures: a comprehensive analysis of 500 episodes of bacteremia and fungemia in adults: laboratory and epidemiologic observations. Rev Infect Dis 1983; 5: 35-53. Williams DN, Peterson PK, Verhoef J, Laverdiere M, Sabath LD: Endocarditis caused by coagulase-negative staphylo-

COAGULASE-NEGATIVE

10.

11.

12.

13.

14.

15.

16.

17.

cocci. Infection 1979; 7: 5-9. Archer GL, Vishniavsky N, Stiver HG: Plasmid pattern analysis of Staphylococcus epidermidis isolates from patients with prosthetic valve endocarditis. Infect lmmun 1982; 35: 627-632. Archer GL, Karchmer AW, Vishniavsky N, Johnson JL: Plasmid-pattern analysis for the differentiation of infecting from noninfecting Staphylococcus epidermidis. J Infect Dis 1984; 149: 913-920. Christensen GD, Parisi JT, Bisno AL, Simpson WA, Beachey EH: Characterization of clinically significant strains of coagulase-negative staphylococci. J Glin Microbial 1983; 18: 258-269. Tuazon CU, Miller H: Clinical and microbiologic aspects of serious infections caused by Staphylococcus epidermidis. Stand J Infect Dis 1983; 15: 347-360. Geraci JE, Hanson KC, Giuliani ER: Endocarditis caused by coagulase-negative staphylococci. Mayo Clin Proc 1968; 43: 420-434. Quinn EL, Cox F: Staphylococcus albus (epidermidis) endocarditis: report of sixteen cases seen between 1953 and 1962. Antimicrob Agents Chemo 1963; 3: 635-642. Lockman DS, Chew WH, Rissing JP: Staphylococcus epidermidis endocarditis: a community-acquired methicillin-resistant isolate. Southern Med J 1983; 76: 1203-1204. Dan M, Marien GJR, Goldsand G: Endocarditis caused by Staphylococcus warneri on a normal aortic valve following vasectomy. Can Med Assoc J 1984; 131: 211-213.

October

18.

19.

20.

2 1.

22.

23.

24.

1987

STAPHYLOCOCCAL

ENDOCARDITIS-CAPUTO

ET AL

Baddour LM, Phillips TN, Bisno AL: Coagulase-negative staphylococcal endocarditis. Arch Intern Med 1986; 146: 119-121. Archer GL, Armstrong BC: Alteration of staphylococcal flora in cardiac surgery patients receiving antibiotic prophylaxis. J Infect Dis 1983; 147: 642-649. Schwalbe RS, Stapleton JT, Gilligan PH: Emergence of vancomycin resistance in coagulase-negative staphylococci. N Engl J Med 1987; 316: 927-931. Karchmer AW, Archer GL, Endocarditis Study Group: Methicillin resistant Staphylococcus epidermidis prosthetic valve endocarditis: a therapeutic trial (abstr 476). Abstracts of the 24th Conference on Antimicrobial Agents and Chemotherapy, American Society for Microbiology, 1984. Vazquez GJ, Archer GL: Antibiotic therapy of experimental Staphylococcus epidermidis endocarditis. Antimicrob Agents Chemother 1980; 17: 280-285. Korzeniowski 0, Sande MA, The National Collaborative Endocarditis Study Group: Combination antimicrobial therapy for Staphylococcus aureus endocarditis in patients addicted to parenteral drugs and in nonaddicts: a prospective study. Ann Intern Med 1982; 97: 496-502. Lowy FD, Chang DS, Aning V, Williams S, Szilagyi G: Reliability of in vitro susceptibility tests for detecting coagulase-negative staphylococcal resistance to penicillinaseresistant semisynthetic penicillins. J Clin Microbial 1983; 18: 1122-l 126.

The American

Journal

of Medicine

Volume

83

625