In a trial of empirical antifungal therapy in persistently febrile neutropenic patients, L-AMB was shown to be comparable in overall efficacy with significantly less nephrotoxicity and infusion-related reactions when compared to AmB-d. In addition, significantly fewer proven breakthrough invasive fungal infections (IFI) occurred in the L-AMB treatment group. (

Walsh TJ, et al,
). In order to better evaluate the maximum efficacy benefit of L-AMB in a higher risk patient subset, a sub-group analysis was performed on those patients with duration of neutropenia greater than or equal to the median duration for the entire study population. The results of this sub-group analysis are presented.

Methods: From the original data set, the median duration of neutropenia from the time of treatment initiation was determined. Clinical success (defined by a composite endpoint), frequency of proven breakthrough IFI and all cause mortality were reported by treatment group for patients with duration of neutropenia greater than or equal to the median. Investigator-reported proven breakthrough IFI were reviewed and verified by an independent, blinded expert.

Results: A total of 687 patients were enrolled in the trial. Median duration of neutropenia following initiation of antifungal therapy was 7 days (range 1–50). 392 patients had neutropenia >/= 7 days, (L-AMB=205), (AmB-d=187). Clinical success: L-AMB: 128/205 (62.4%); AmB-d: 112/187 (59.9%), and all cause mortality: L-AMB: 19/205 (9.3%); AmB-d: 24/187 (12.8%) were not significantly different. The frequency of proven breakthrough IFI favored L-AMB: L-AMB: 7/205 (3.4%); AmB-d: 18/187 (9.6%), p=0.01. Organisms isolated: L-AMB: 2 Candida spp (blood); 4 Aspergillus spp. (3 pneumonia, 1 wound); 1 Mucor sp (pneumonia). AmB-d: 9 Candida spp (8 blood, 1 pneumonia); 7 Aspergillus spp (4 pneumonia, 1 sinus, 1 pericarditis, 1 wound); 1 Cryptococcus albidus (blood); 1 unclassified mould (pneumonia). The difference in emergent proven IFI for the patients with neutropenia <7d trended in favor of L-AMB, but was not statistically significant: L-AMB 3/138 (2.2%); AmB-d 8/157 (5.1%).

Conclusions: Overall clinical success and all cause mortality remained equivalent for the 2 treatment arms in the subset of patients with prolonged neutropenia. However, L-AMB significantly reduced the frequency of proven breakthrough IFI in this high-risk patient population. These data support the efficacy of L-AMB as empirical therapy, particularly for persistently neutropenic patients at higher risk for invasive fungal infections.

Author notes

Corresponding author