To determine factors that affect outcome in neonates with culture-proven sepsis, the charts of all neonates with culture-proven sepsis admitted to the University Hospital of the West Indies between January 1995 and December 2000 were reviewed retrospectively. Neonates who survived without developing any complications (favourable outcome group) were compared with those who died and/or developed severe complications during the course of treatment (poor outcome group). Chi-square tests were done to determine factors associated with poor outcome; univariate and multivariate logistic regression analyses were also performed. One hundred and thirty-five neonates had culture-proven sepsis, of which 89 (66%) were term infants and 46 (34%) were preterm. Male to female ratio was 1.6:1. One hundred and twenty-six (93%) survived and 9 (7%) died. Case fatality rates were higher for premature infants (15%) than for term infants (2%). Twenty-four (18%) of the neonates with culture proven sepsis had a poor outcome. Gram negative organisms accounted for 19 (70%) of the cases with poor outcome. Prematurity (p < 0.001), very low birthweight (p < 0.001) and female gender (p < 0.05) were factors associated with poor outcome. Strategies aimed at decreasing morbidity and mortality in neonates with sepsis must include measures that will decrease the incidence of prematurity and low birthweight.