Several viral infections have been reported to result in more severe disease in pregnant than non-pregnant women, but the relative risks have not been well characterised. This has now been done for Lassa fever in a prospective study of 68 pregnant and 79 non-pregnant women who were admitted to hospital in Sierra Leone with confirmed Lassa fever. Lassa fever was the main cause of maternal mortality in the hospital, accounting for 25% of maternal deaths. Twelve of 40 patients in the third trimester died, compared with two of 28 in the first two trimesters and 10 of 79 non-pregnant women. The odds ratio for death in the third trimester compared with the first two trimesters was 5.57 (95% confidence intervals 1.02 to 30.26). The condition of the mother improved rapidly after evacuation of the uterus, whether by spontaneous abortion, evacuation of retained products of conception, or normal delivery; 10 of 26 women without uterine evacuation died, but only four of 39 women with evacuation died (p = 0.0016). The odds ratio for death with pregnancy intact was 5.47 (95% confidence interval 1.35 to 22.16). Fetal and neonatal loss was 87%. The risk of death from Lassa fever in the third trimester is significantly higher than that in the first two trimesters and higher than that for non-pregnant women, but evacuation of the uterus can significantly improve the mother's chance of survival.