Choosing an antibiotic for skin infections.

Since the mid-1990s, methicillin-resistant Staphylococcus aureus (MRSA) has become the predominant pathogen responsible for suppurative skin infections (i.e., furuncles, carbuncles, and skin abscesses) in the United States and in many other countries. Emergency department visits for skin abscesses have increased during this period, as MRSA has become endemic in the community, but a similar increase in visits for cellulitis (i.e., infectious inflammation of the skin without a drainable collection of pus) has not occurred.1,2 The microbiologic characteristics of cellulitis are less well defined because material for culture is not readily available. Although S. aureus is the organism most frequently isolated from needle aspirates or punch biopsy samples in cellulitis, cultures are negative in more than 70% of such samples, and serologic evidence suggests that Streptococcus pyogenes and other beta-hemolytic streptococci cause most cases.3,4 Given the immense importance of MRSA during the past two decades, prospective clinical trial data to inform the choice of an outpatient treatment regimen for skin infections in the MRSA era are surprisingly sparse. Furthermore, despite the lack of evidence that MRSA is a major cause of cellulitis (as opposed to skin abscesses), there has been a shift toward prescribing agents active against MRSA for the treatment of cellulitis. In this issue of the Journal, Miller et al. address these knowledge gaps with a prospective, randomized, double-blind trial comparing clindamycin with trimethoprim–sulfamethoxazole (TMP-SMX) in the treatment of outpatients with skin infections.5 More than 500 adults and children were enrolled from four centers in areas in which communityacquired MRSA is endemic. Patients with multiple abscesses or a single abscess larger than 5 cm in diameter (or proportionally smaller for young children) underwent incision and drainage. Patients with smaller abscesses were not included in the analysis. Cultures were obtained from all patients with purulent lesions and from patients with cellulitis only if fluid that could be cultured was present. Not surprisingly, S. aureus was isolated from 72.7% of patients with abscesses with or without cellulitis; of these isolates, 83.0% were MRSA. Cultures were obtained from 20% of patients with cellulitis only (32 patients with cellulitis only [11.4%] had MRSA, 1 had S. pyogenes [0.4%], and 1 had group B streptococcus [0.4%]). Clinical cure rates were high and not significantly different for clindamycin and TMP-SMX — 89.5% and 88.2%, respectively — in the population that could be evaluated. No significant difference in cure rate was seen in any of the subgroup analyses. Although the trial did not have a formal noninferiority design, the results support the authors’ conclusion that there is unlikely to be a large difference in efficacy between the two treatment regimens for skin infections in outpatients. However, the design of the study obscures possible differences between the outcomes in the two major subgroups, differences that favor TMP-SMX for abscesses and clindamycin for cellulitis. There was a nonsignificant difference favoring TMP-SMX in patients who could be evaluated who had abscess only or abscess plus cellulitis, and there was a nonsignificant difference favoring clin da mycin in those who had cel-