Study objective: To compare meropenem with ciprofloxacin for treatment of gram-negative bacilli sepsis in penicillin-allergic patients in the intensive care unit (ICU) to determine if increased anaphylaxis risk with meropenem precluded its use when weighed against risks of inactive therapy with ciprofloxacin.
Design: A decision model constructed from probability distributions from the literature and data from a local ICU antibiogram. A probabilistic sensitivity analysis was performed to evaluate uncertainty in variable estimates by using one-way analyses, two-way analyses, and Monte Carlo simulation.
Measurements and main results: Microbiologic activity of treatment, anaphylaxis according to treatment regimen, curability of infection according to patient morbidity status, risk of superinfection, and recovery from gram-negative bacilli sepsis were the variables modeled. Effectiveness was defined by long-term survival and was modeled as life-years (LYs) and quality-adjusted life-years (QALYs) gained according to treatment group. Base case results were the incremental differences between the average effectiveness of each strategy calculated from the Monte Carlo simulation. Mean LYs and QALYs gained with meropenem were 9.9 (95% confidence interval [CI] 8.9-10.8) and 5.9 (95% CI 4.8-6.7), respectively. Mean LYs and QALYs gained with ciprofloxacin were 8.9 (95% CI 8.1-9.6) and 5.3 (95% CI 4.4-6.3), respectively. The incremental difference in effectiveness-or average benefit expected by selecting meropenem over ciprofloxacin-was 1.0 LY (95% CI 0.3-1.7 LYs) and 0.6 QALY (95% CI 0.2-1.1 QALYs) favoring meropenem.
Conclusion: Use of empiric meropenem over ciprofloxacin may be justified in patients in the ICU who are allergic to penicillin.