A high intermittent dose regimen (group A: 10 mg kg(-1) on day 1, 5 mg kg(-1) on days 3 and 6) was compared with standard dosing (group B: 3 mg kg(-1) per day for 14 days) of liposomal amphotericin B (LAB) for empirical treatment of persistent febrile neutropenia. A total cumulative dose of 1275 mg (group A) and 2800 mg (group B) was administered. Infusion-related adverse drug events, mainly rigors/chills, occurred more frequently with group A (11/45, 24 % infusions) than with group B (12/201, 6 % infusions) (P=0.002), which extended the mean infusion time by 20 min (P=0.001). Creatinine levels were similar in the two regimens: the A : B ratio of the area under the curve for creatinine (AUC(CREATININE)) for days 2-7 was 1.09 (P=0.27) and for days 2-14 was 1.05 (P=0.51). Rises in creatinine were mild (clinical toxicity criteria 1) in all patients with elevations. Hypokalaemia tended to be less severe in group A with a lower proportion of hypokalaemic days [57/143 (39 %) vs 80/137 (58 %), P=0.21], a higher AUC(POTASSIUM) (A : B ratio of 1.06, P=0.12), a lower proportion of patients with hypokalaemia at the end of study (10 vs 61 %, P=0.01) and fewer potassium-supplemented days [12/210 (6 %) vs 41/210 (19.5 %), P<0.1]. There were mildly elevated median levels of serum bilirubin, alanine aminotransferase, aspartate aminotransferase and alkaline phosphatase, which were similar for the two regimens and were usually associated with other co-existing co-morbid conditions. The AUC for these enzymes was also similar in the two groups. No patient had discontinuation of the study drug due to toxicity. Composite success was identical for each regimen (11/15 patients, 73 %). Three of the fifteen patients in group B and none in group A developed invasive fungal infections (IFIs). Beta-D-Glucan levels were similar in both groups for patients without an IFI [AUC(GLUCAN) of 362 and 683 (P=0.36) for groups A and B, respectively]. The rate of defervescence was similar for each regimen (P=0.75). This feasibility study suggests that a short intermittent high-dose course of 10/5/5 mg LAB kg(-1) on days 1, 3 and 6 may be as safe and effective as a standard 14 day course of 3 mg kg(-1) per day, with drug-acquisition cost savings and reduced drug exposure. A larger study is indicated for confirmation of this.