Presented by Hanyin Zhou
Masters Candidate in Biostatistics
Plan B Adviser: Dr. Biyue Dai

The standard dose of 100 mg/kg/day of flucytosine is associated with substantial toxicity for treating Cryptococcal meningitis. In the FLOOR trial, the safety and efficacy of reduced flucytosine dose of 60 mg/kg/day was evaluated and compared with a historical control group receiving the standard dose. Since FLOOR was a non-randomized, single-arm study with a historical control, baseline and eligibility differences and changes in clinical practice may have introduced confounding and affected the estimation of the treatment effect. We applied propensity score (PS) methods to reassess two endpoints: early fungicidal activity (EFA) and 10-week rehospitalization-free survival. Inverse probability weighting (IPW), augmented IPW, PS stratification, adjusted regression, and PS matching were applied to EFA. Rehospitalization-free survival was analyzed using crude Kaplan-Meier curves, IPW hazards models, and crude and adjusted Cox models. Estimated group differences in EFA were consistently negative across methods, ranging from -0.129 (95%CI -0.249 to -0.009) to -0.085 (95%CI -0.311 to 0.023) and favoring standard-dose treatment. IPW, PS stratification, adjusted regression, and 1:3 matching produced the narrowest confidence intervals and excluded the null. No statistically significant difference in 10-week rehospitalization-free survival was found between groups. Overall, propensity score analyses supported the original conclusion that reduced-dose flucytosine was less effective in fungal clearance, and clinical outcome appeared similar between groups.