Résumé : The effect of maternal HIV infection on birthweight was estimated. In the prenatal clinic of the Centre Hospitalier de Kigali, HIV screening was proposed to women with a gestational age (GA) of less than 28 weeks. HIV-infected (HIV+) and uninfected (HIV-) women were recruited, when they consented. At inclusion, socioeconomic, obstetrical data, and body weight were collected, a clinical examination was performed, and tests for sexually transmitted diseases (STDs) and malaria were performed. Two prenatal visits were made, at 28-32 and 32-36 weeks, with clinical data and weight measurement. At delivery, birthweight, body length, and head circumference of the infant were documented. At inclusion and at the second follow-up visit, HIV+ women (N = 177) weighed less than HIV- women (N = 194) (p = 0.004). Mean birthweight in infants born to HIV+ women was 2947 g (SD = 429) and 3104 g (SD = 461) in infants born to HIV- women (p = 0.001). Frequencies of low birthweight (LBW, weight < 2500 g), prematurity (GA < 37 weeks, according to Finnström score at birth), and intrauterine growth retardation (defined by LBW and GA > or = 37 weeks) were higher in infants born to HIV+ women than to HIV- women (p = 0.009, 0.01, and 0.053, respectively). In multivariate logistic regression, the association between maternal HIV infection and LBW disappeared (p = 0.61), while low GA (p = 0.01) and low last prenatal weight (p = 0.01) were independant risk factors of LBW. LBW in infants born to HIV+ women could be partly attributable to impaired maternal weight. These results underline the need for nutritional surveillance and dietary counselling, hoping to improve the prognosis of pregnancy in HIV+ women, regardless of other therapeutic interventions.