Staphylococcus aureus: The new adventures of a legendary pathogen
ABSTRACTNosocomial infections with strains of methicillin-resistant Staphylococcus aureus (MRSA) began to emerge in the 1960s, are increasing in frequency, and tend to have worse outcomes than infections due to methicillin-susceptible S aureus. Community-associated MRSA infections emerged in the 1990s. Community-associated MRSA strains have up to now been epidemiologically and bacteriologically distinct from hospital-associated MRSA strains, but in a new twist, MRSA strains that have so far been only community-associated are invading the hospital. Another worrisome trend is increasing resistance to vancomycin (Vancocin).
KEY POINTS
- Community-associated MRSA infections tend to affect patients younger than those who traditionally get hospital-associated MRSA infections. Most of these infections are of the skin and soft tissues, but this pathogen can also affect deeper tissues, and bacteremia and necrotizing pneumonia have been reported.
- For patients with skin and soft-tissue infections due to MRSA, incision and drainage rather than antibiotic therapy is often the key intervention.
- Vancomycin has been our stalwart for treating MRSA infections for more than 40 years, but it is not working as well as it used to, at least in certain situations. Vancomycin should not be used to treat infections due to methicillin-susceptible S aureus.
- Needed are better understanding of the factors that influence persistent S aureus bacteremia, well-controlled, prospective studies, and continued antibiotic development.
IN UNEXPLAINED BACTEREMIA, LOOK FOR ENDOCARDITIS
In blood cultures from patients with bacteremia, S aureus is never a contaminant. Even if just one blood culture is positive for S aureus, believe that S aureus is the culprit.
Reports in the 1950s suggested that at least half of patients who had S aureus bacteremia had endocarditis,25 leading to recommendations that all patients with S aureus bacteremia without an obvious primary source of infection should be evaluated for endocarditis. Subsequent estimates were lower, in the range of 15% to 25%.26,27 However, throughout the world S aureus endocarditis continues to have a very high mortality rate: at least a third of patients die.28
Clinical criteria (community acquisition, no primary focus, and metastatic sequelae) were developed to try to predict the risk of endocarditis in bacteremic patients.26 However, these criteria did not work very well. The clinical definition of endocarditis has evolved. The criteria of von Reyn et al29 from 1981 did not use echocardiography as part of the definition, but the 1994 Duke criteria,30 which were refined31 in 2000, use both clinical and echocardiographic parameters.
Stratton et al32 performed transthoracic echocardiography in 14 patients with bacteremia and found 1 patient with cryptic tricuspid infective endocarditis. Bayer et al33 subsequently reported that of 72 patients with bacteremia, 6 (18%) of those who had no clinical findings suggestive of infectious endocarditis had findings on echocardiography that led to changes in their regimen. Adding echocardiography to three clinical risk factors increased the sensitivity of diagnosing endocarditis from 70% to 85% with a specificity of 100% and predictive value of 96%.
The Duke criteria call for transesophageal echocardiography, which is not feasible in some patients, eg, those with cirrhosis and esophageal varices.
S aureus endocarditis has changed over the years as our patient population has changed, and MRSA endocarditis tends to hit some of our most vulnerable patients. In a study by Miro et al34 in 2005, MRSA was the leading pathogen in patients who were diagnosed with S aureus endocarditis in 1990 or later. We will only see these numbers go up. Patients with diabetes tend to have more MRSA, and of diabetic patients with MRSA endocarditis, 30% to 40% die in the hospital.
Indications for surgery
Certain conditions are indications for surgery among patients with endocarditis, and no antibiotic will cure the endocarditis if the patient has one of these conditions, eg:
- Persistent bacteremia during antibiotic therapy
- Recurrent emboli
- Heart failure that cannot be controlled
- Perivalvular or myocardial abscesses
- Large vegetations
- Early prosthetic valve infection
- Certain arrhythmias.
How long should S aureus bacteremia be treated?
In cases of bacteremia in which endocarditis has been ruled out and removable foci of infection (eg, intravascular catheters) have been removed, some evidence indicates that treatment for 2 weeks would be as effective as the 4 to 6 weeks that we would use for endocarditis or other severe or invasive infections.35 The issue is controversial. If the patient has had frequent hospitalizations or a chronic medical condition I would hesitate to treat for less than 4 weeks, even if the infection appears to be associated with a removable focus.
Treatment of endocarditis
In the guidelines for treatment of endocarditis from the American Heart Association and Infectious Diseases Society of America,36 all the recommendations are relatively old and many of them are somewhat empiric—they are not based on evidence from randomized clinical trials. Rather, they are best opinions based on clinical experience and some observational studies over the years.
For MSSA. In cases of native-valve endocarditis, oxacillin (Bactocill), nafcillin (Unipen), or another semisynthetic beta-lactam antibiotic is recommended. For penicillin-allergic patients, we have other options, such as cefazolin (Ancef, Kefzol).
Combination therapy is frequently recommended for native valve endocarditis as well as for prosthetic valve endocarditis, with either rifampin or gentamicin along with a primary agent. There is some evidence that one can clear staphylococcal bacteremia a day or two more quickly by use of combination therapy with nafcillin plus an aminoglycoside than with nafcillin alone.37,38 For MSSA-associated endocarditis, vancomycin does not work as well as beta-lactam antibiotics.39,40
Korzeniowski and Sande37 and Chambers et al38 reported that the mean duration of bacteremia was 3.4 days for patients treated with nafcillin alone and 2.9 days for those treated with nafcillin plus an aminoglycoside. These studies led to consideration of a short course of gentamicin to clear the bacteremia quickly.
With MRSA, bacteremia often requires a week or more to clear. Levine et al21 reported a study in 42 patients, mostly injection-drug users, with right-sided native-valve endocarditis. The median duration of bacteremia was 7 days in patients who received vancomycin alone vs 9 days in those who received vancomycin plus rifampin; however, some patients were bacteremic for up to 27 days. Fever persisted for a median of 7 days, probably partly due to septic pulmonary emboli. Three patients died, and three required valve replacement.
