Traditional gentamicin dosing every 8-24 h depending on age and weight in neonates does not provide the ideal concentration-time profile to both optimize the concentration-dependent killing by aminoglycosides and minimize toxicity. Fifty-three neonates were audited prospectively while receiving gentamicin 2.5 mg/kg every 8-24 h, aiming for peak concentrations (Cmax) of 6-10 mg/L and trough concentrations (Cmin) <2 mg/L. After the first dose, the mean (+/- s.d.) Cmax was 5.5 +/- 0.7 mg/L with sub-therapeutic concentrations (<6 mg/L) in 62% of patients, while the mean Cmin was >2 mg/L in 15% of the neonates. After the third dose the Cmax was 7.5 +/- 1.5 mg/L, with 17% <6 mg/L, whereas the mean Cmin was 2.2 +/- 1.1 mg/L with 49% of values >2 mg/L. An extended interval dosing method (24, 36 and 48 h) for infant weights of 0.75-5 kg was developed by simulation, and audited prospectively in 51 neonates. Prospective analysis of the extended interval dosing method showed a mean Cmax after the first dose of 13.1 +/- 3.6 mg/L, while the mean Cmin was 0.7 +/- 0.6 mg/L. Seventy-eight per cent had Cmax of >10 mg/L after the first dose. The mean area under the concentration versus time curve AUC0-24 was 93 mg*h/L (target = 100 mg*h/L). The extended interval dosing achieved higher Cmax values while ensuring that overall exposure per 24 h was acceptable. Prospective testing of the method demonstrated good predictive ability.