RESUMO
INTRODUCTION: To promote judicious prescribing of methicillin-resistant Staphylococcus aureus (MRSA)-active therapy for skin and soft tissue infections (SSTI), we previously developed an MRSA risk assessment tool. The objective of this study was to validate this risk assessment tool internationally. METHODS: A multicenter, prospective cohort study of adults with purulent SSTI was performed at seven international sites from July 2016 to March 2018. Patient MRSA risk scores were computed as follows: MRSA infection/colonization history (2 points); previous hospitalization, previous antibiotics, chronic kidney disease, intravenous drug use, human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS), diabetes with obesity (1 point each). Predictive performance of MRSA surveillance percentage, MRSA risk score, and estimated MRSA probability (surveillance percentage adjusted by risk score) were quantified using the area under the receiver operating characteristic curves (aROC) and compared. Performance characteristics of different risk score thresholds across varying baseline MRSA prevalence were examined. RESULTS: Two hundred three patients were included. Common SSTI were wounds (28.6%), abscess (25.1%), and cellulitis with abscess (20.7%). Patients with higher risk scores were more likely to have MRSA (P < 0.001). The MRSA risk score aROC (95%CI) [0.748 (0.678-0.819)] was significantly greater than MRSA surveillance percentage [0.646 (0.569-0.722)] (P = 0.016). Estimated MRSA probability aROC [0.781 (0.716-0.845)] was significantly greater than surveillance percentage (P < 0.001) but not the risk score (P = 0.192). The estimated negative predictive value (NPV) of an MRSA score ≥ 1 (i.e., a score of 0) was greater than 90% when MRSA prevalence was 30% or less. CONCLUSION: The MRSA risk score and estimated MRSA probability were significantly more predictive of MRSA compared with surveillance percentage. An MRSA risk score of zero had high predictive value and could help avoid unnecessary empiric MRSA coverage in low-acuity patients. Further study, including impact of such risk assessment tools on prescribing patterns and outcomes are required before implementation.
RESUMO
The objective of this study was to determine if real-world ceftaroline treatment in adults hospitalized for acute bacterial skin and skin structure infections (ABSSSI) is associated with decreased infection-related length of stay (LOSinf) compared to that with vancomycin. This was a retrospective, multicenter, cohort study from 2012 to 2017. Cox proportional hazard regression, propensity score matching, and inverse probability of treatment weighting (IPTW) were used to determine the independent effect of treatment group on LOSinf The patients were adults hospitalized with ABSSSI and treated with ceftaroline or vancomycin for ≥72 h within 120 h of diagnosis at four academic medical centers and two community hospitals in Arizona, Florida, Michigan, and West Virginia. A total of 724 patients were included (325 ceftaroline treated and 399 vancomycin treated). In general, ceftaroline-treated patients had characteristics consistent with a higher risk of poor outcomes. The unadjusted median LOSinf values were 5 (interquartile range [IQR], 3 to 7) days and 6 (IQR, 4 to 8) days in the vancomycin and ceftaroline groups, respectively (hazard ratio [HR], 0.866; 95% confidence interval [CI], 0.747 to 1.002). The Cox proportional hazard model (adjusted HR [aHR], 0.891; 95% CI, 0.748 to 1.060), propensity score-matched (aHR, 0.955; 95% CI, 0.786 to 1.159), and IPTW (aHR, 0.918; 95% CI, 0.793 to 1.063) analyses demonstrated no significant difference in LOSinf between groups. Patients treated with ceftaroline were significantly more likely to meet criteria for discharge readiness at day 3 in unadjusted and adjusted analyses. Although discharge readiness at day 3 was higher in ceftaroline-treated patients, LOSinf values were similar between treatment groups. Clinical and nonclinical factors were associated with LOSinf.
