Therapeutic drug monitoring of high-dosage intravenous flucloxacillin and mortality in patients with Staphylococcus aureus bacteremia: a retrospective cohort study using inverse probability weighting

Abstract Purpose Despite the increasing use of TDM, data on its effect on relevant patient outcomes for anti-staphylococcal penicillins (ASPs) remain scarce. Methods We performed a retrospective study using data from two Dutch cohorts of patients with Staphylococcus aureus bacteremia (SAB). Patients with MSSA bacteremia who started with the equivalent of flucloxacillin 12 g/24 h within seven days after the first positive blood culture were included. We compared the effects of a TDM-based dosing strategy versus a fixed-dosage strategy during high-dosage flucloxacillin therapy. Primary outcome was 90-day all-cause mortality. Secondary outcomes included incidence of antibiotic-associated adverse events and antibiotic consumption. We fitted a marginal structural model using inverse probability weighting to adjust for confounding (i.e. age, Charlson Comorbidity Index, persistent bacteremia, dominant focus of infection and ICU admission), time-dependent bias and exposure-confounder feedback. Results We included 285 patients of whom 132 were treated with a TDM-based dosing strategy and 153 with a fixed-dosage strategy. TDM-based dosing was on average associated with a 2.30 gram (95% CI 1.90–2.71) lower daily flucloxacillin dosage than a fixed-dosage regimen. In the TDM group 90-day all-cause mortality was 27% and in the fixed-dosage group 39%. In adjusted analyses, TDM-based dosing was not significantly associated with reduced mortality (HR 0.69, 0.46–1.05) nor with AKI (HR 1.41, 0.89–2.23), DILI (HR 1.79, 0.42–7.68) or hypokalemia (HR 0.90, 0.54–1.48). Conclusion TDM in patients with SAB treated with high-dosage ASPs is associated with lower antibiotic consumption without a statistically significant difference in mortality. We found no evidence of a decrease in antibiotic-associated adverse events.