Trough serum concentrations (Cmin) of gentamicin were followed during up to 96h of treatment in 44 neonates (17 preterm and 27 term), treated with intramuscular gentamicin 2.5 ± 0.3 mg/kg (mean ± SD) twice daily, a dosage that was not changed during the follow-up period. Relationships with patients' gestational age, postnatal age, postconceptional age and bodyweight were analysed to identify circumstances in which gentamicin should be monitored.Gentamicin Cminvalues after 24h correlated better with neonate's postconceptional age (r = −0.42) or gestational age (r = −0.37) than with postnatal age or bodyweight. Correlations with postconceptional age and gestational age improved after 96h (r = −0.71 and r = −0.67, respectively).From 24 to 96h Cminincreased from 1.5 to 2 mg/L (p < 0.001) in the preterm neonates and from 1.5 to 2.5 mg/L (p < 0.01) in those preterm neonates ⩽32 weeks of gestational age, while differences between neonates ⩽3 days and >3 days of postnatal age were nonsignificant. The Cminat 24h was potentially toxic (>2 mg/L) in 9% of the neonates (12% of preterm and 7% of term neonates). At 96h, the percentage of neonates with toxic Cminvalues increased to 25% (65% of all preterm neonates and 100% of preterm neonates ⩽32 weeks of gestational age), whereas in term neonates it decreased to 0%.In conclusion, in preterm neonates ⩽32 weeks of gestational age a dosage of 2.5 mg/kg every 24h should be used, and gentamicin concentrations should be monitored. However, in term neonates >7 days of postnatal age a dosage of 3.5 mg/kg twice daily should be recommended.