To assess risk factors for malaria in pregnancy in Kisumu, western Kenya, we studied healthy women with an uncomplicated pregnancy of > or = 32 weeks attending the antenatal clinic in the Provincial Hospital. Between June 1996 and March 1999, malaria and human immunodeficiency virus (HIV) infection were examined in 5093 pregnant women: 20.1% of the women were parasitaemic and 24.9% were HIV-seropositive. 2502 women delivered in the hospital and a smear was obtained: the prevalence of placental malaria, maternal peripheral parasitaemia, and HIV infection was respectively 19.0%, 15.2% and 24.5%. HIV infection (risk ratio [RR] 1.58, 95% confidence interval [95% CI] 1.32-1.89), young age (< 21 years: RR 1.51, 95% CI 1.19-1.91), being a primigravidae (RR 1.41, 95% CI 1.05-1.88), a peri-urban residence (RR 1.50, 95% CI 1.21-1.88), and Luo ethnicity (RR 1.74, 95% CI 1.35-2.24) were risk factors for malaria at delivery. Use of sulfadoxine-pyrimethamine (SP), reported by 2.1% of the women, was a protective factor (RR 0.44, 95% CI 0.18-1.06). Results were similar in the third trimester. In this urban/peri-urban setting, preventing HIV infection, delaying the first pregnancy until after adolescence, and applying an effective antimalarial strategy such as intermittent therapy with SP will reduce the prevalence of malaria in pregnancy.