Methods: This was an analysis of pooled data from two published studies of adults with MRSA bacteremia (Dilworth et al, Antimicrob Agents Chemother. 2014;58(1):102-9; Casapao et al, Pharmacotherapy. 2017;37(11):1347-1356). All patients received intravenous VAN for ≥ 72 hours. COMBO patients received an intravenous BL for ≥ 48 hours with VAN, started within 24 hours of VAN. The remaining patients comprised the VAN group. The primary outcome was PB (≥ 5 days). The impact of BL class on PB was assessed. Acute kidney injury (AKI, serum creatinine increase from baseline by 0.5mg/dL or 50%) was examined as a secondary outcome. Demographics were compared between groups. Multivariable logistic regression models compared PB between COMBO and VAN.
Results: In total, 156 patients were included (VAN=66; COMBO=90). The groups were similar except COMBO patients were more likely to have a pulmonary bacteremia source (12.2% v. 1.5%, p=0.014) and a higher median (IQR) vancomycin serum level (mg/L, 17.8 (13.9, 23.6) v. 15.7 (11.3, 20.6); p=0.039). PB was less common in COMBO (26.7% v. 43.9%, p=0.027). In a multivariable model COMBO was inversely associated with PB (adjusted odds ratio [aOR], 95% confidence intervals [CI], 0.48, 0.24-0.95). AKI was more common in COMBO (18.9% v. 7.6%, p=0.062). PB and AKI rates by BL class are shown in the table below, with VAN listed for reference:
|Variable, n (%)||
|PB||0 (0)||8 (32)||15 (26.8)||0.191||29 (43.9)|
|AKI||1 (12.5)||4 (16)||12 (21.4)||0.749||5 (7.6)|
Conclusion: COMBO reduced the likelihood of PB but had a higher AKI rate. There were no significant differences in PB by BL class. Clinically, COMBO may reduce PB rates and prevent overuse of salvage antibiotic therapy. BL choice for COMBO warrants further investigation.
T. J. Dilworth,
O. M. Ibrahim, None
D. M. Jacobs, None
D. R. Bowers, None
N. D. Beyda, None
R. C. Mercier, None