Emergency clinicians routinely make decisions for SSTIs based on incomplete information; treatment guidelines remain vague regarding when antibiotics are indicated, information about local epidemiology is often incomplete, and microbiologic data for individual patients are not available in the time frame of an ED visit.
In this study, we identified a population of ED patients with presumed community-acquired SSTIs in whom S. aureus remained the most common pathogen and for whom antibiotic prescription remained high. Despite the prevalence of S. aureus as the target of therapy, antibiotic regimens varied significantly. Among patients who underwent culture and received antibiotics, discordance between the choice to treat empirically with anti-MRSA antibiotics and the presence or absence of the resistant organism in culture was high; patients were often treated narrowly for MRSA infections, or broadly for MSSA infections.
The microbiology of skin abscesses does not appear to be uniform; resistance patterns from our sample differed between children and adults. Increased resistance to TMP/SMX – among the most commonly-used antibiotics in SSTIs – was noted, particularly in MSSA isolated from children. Though the number of pediatric MSSA infections was a small proportion of the total number of patients, 20 of the 49 S. aureus cultures from children were MSSA. The implications of this finding are not immediately clear, but highlight the importance of (a) knowledge of local disease epidemiology, and (b) performance of surveillance cultures in at least some subset of ED patients treated for SSTIs. This epidemiologic surveillance is important in monitoring infections treated in the ED, and may identify emerging resistance before it becomes broadly apparent.
Importantly, differences in disease epidemiology were not reflected in the antibiogram distributed by the hospitals' microbiology laboratory. S. aureus from SSTIs had greater TMP/SMX resistance and less fluoroquinolone resistance than reflected in the antibiogram. The resistance patterns reflected in composite antibiograms may mask important differences in pathogens' behavior in SSTIs, since the antibiogram does not distinguish between pathogens isolated from blood, sputum, or other sources. The strains of S. aureus that cause purulent SSTIs differ from those that cause other invasive infections, and this may not be clear when microbiologic data is viewed in the aggregate. Laboratories should consider reporting disease-specific antibiotic resistance data, as this more granular data could drive therapeutic decision-making.
None of the demographic or clinical factors in our logistic model correlated choice of empiric anti-MRSA therapy with the presence or absence of MRSA in culture in our sample of patients who underwent culture and received antibiotics. Using prescribing behavior as a proxy for clinician beliefs, there did not appear to be specific factors interpreted by ED clinicians as being predictive of a particular pathogen's antibiotic susceptibility. However, those patients who were admitted to the hospital or who underwent I&D in the ED were more likely to receive antibiotic therapy in the ED to which the resultant cultured organism was susceptible, suggesting that those patients deemed to be more ill or to require an invasive procedure were more likely to receive broader antibiotic therapy.
Use of "double coverage" – two or more antibiotics, typically TMP-SMX plus cephalexin – was prevalent, and was likely intended to address perceived deficiencies of single-agent treatment with TMP-SMX in treating streptococci. However, cultures from the large majority of patients treated with "double coverage" yielded staphylococci alone, suggesting that empiric anti-streptococcal treatment may not be necessary. When viewed from an antibiotic stewardship perspective, "double coverage" doubles the exposure to antibiotics and may drive resistance without leading to improved therapy.
Only age group was reliably associated with use of "double coverage" in our logistic model; children were less likely to receive multiple antibiotics. Otherwise, the choice appears to be one of clinician discretion. Given that most isolates even from adult patients yielded staphylococci, and that I&D alone is sufficient therapy for most uncomplicated abscesses, use of a single antibiotic – chosen using local epidemiologic data, where available – is warranted if antibiotics are deemed necessary.
The clinician can opt not to treat uncomplicated, small purulent infections with antibiotics if adequate I&D is performed. This is increasingly supported by the evidence and in recent guidelines for treatment of CA-MRSA infections, and is not likely to decrease treatment failure or increase selective pressures toward antibiotic resistance.
We acknowledge several limitations to the current study. Its retrospective nature limits the data to that which could be collected from the medical record. However, we believe that we have adhered to high standards for retrospective ED studies.[22,23] We could not directly assess the ED clinicians' intention when choosing which SSTIs to treat with which antibiotics, and could only infer from those choices. Medical records rarely described SSTIs in detail, omitting the degree of cellulitis adjacent to an abscess. We attempt to account for this by limiting our analysis to the "accuracy" of antibiotic choices without inferring the clinicians' specific intent.
Our findings are significant in that they reflect the current state of antibiotic use and overuse. We were unable to correlate the choices of empiric antibiotics provided in the study EDs with any of the demographic or clinical variables studied. Clinicians, given the state of epidemiologic data and clinical tools available during the study period, had insufficient information to predict the susceptibility of an SSTI pathogen at the time that empiric therapy was chosen. If a clinician could (a) determine which purulent SSTIs require antibiotic treatment, and (b) estimate the narrowest effective antibiotic coverage using local disease-specific data or other tools, antibiotic overuse could be limited.
Future efforts in ED management of purulent SSTIs may focus on determining which patients benefit from antibiotic therapy, outcomes in patients treated without antibiotics, and ensuring that adequate I&D can be performed in the ED setting. PCR and other rapid-MRSA-testing technologies are becoming widely available, though these newer technologies have not yet been widely studied in the clinical setting.
BMC Emerg Med. 2013;13(26) © 2013 BioMed Central, Ltd.