Primary Therapy for Invasive Aspergillosis With Triazoles or L-AmB: A Multicenter Retrospective Study

Objective. To compare clinical outcomes of patients treated with liposomal amphotericin B (L-AmB) versus mold-active triazoles as primary treatment for invasive aspergillosis (IA). Methods. Retrospective study of adult patients treated with either L-AmB or triazoles for proven or probable IA at 2 academic hospitals over a 10-year period. The primary endpoint was all-cause 90-day mortality from IA diagnosis. Landmark trial emulation at day 7 postdiagnosis was used to compare initial triazole versus L-AmB for IA. Confounding by indication was addressed using inverse probability of treatment weighting (IPTW) with stabilized weights, and treatment effects were estimated using Cox regression with both IPTW and covariate adjustment. Results. Overall, 401 patients were included. Median age 65 (interquartile range 56–74) years, 60.8% male. Main predisposing conditions were: hematologic malignancy 151 (37.7%), severe respiratory viral infection 120 (29.9%), and chronic steroid treatment 64 (16%). Overall, 105 (26.2%) patients received L-AmB and 296 (73.8%) triazoles as initial therapy. Patients on L-AmB were more likely to have therapy changed (63.8% vs 17.2%, P < .001) for switching to oral triazoles (48, 71.6%), while the main reason for changing triazoles was adverse events (23, 45.1%). Overall 90-day survival rates were similar between triazole (58.8%; 95% confidence interval [CI], 53.4–64.7) and L-AmB (53.3%, 44.6–63.8) groups (P = .3). IPTW-weighted Kaplan-Meier survival curves from day 7 landmark demonstrated an adjusted hazard ratio of 1.43 (95% CI, 0.87–2.33; P = .61). Conclusions. Primary L-AmB therapy was well tolerated and associated with similar survival rates as triazoles. Further studies are needed to investigate the impact of primary L-AmB on IA patient outcomes.