We examined prospectively the associations between dietary vitamin A intake, nondietary factors and growth in 8174 Sudanese children ages 6-72 mo who were stunted at the start of follow-up. All subjects were weighed and measured at baseline and at 6-mo intervals for 18 mo of follow-up. Dietary vitamin A intake during the prior 24 h was assessed using recall of vitamin A-containing foods at baseline and 6-mo intervals. We examined the association of dietary vitamin A intake with growth and the incidence of recovery of stunting after controlling for age, sex, breast-feeding status and socioeconomic variables. We found that carotenoid intake was associated with a greater incidence of reversal of stunting. Children in the highest quintile grew 13 mm more during the study period than children in the lowest quintile [95% confidence interval (CI): 0-25 mm] in multivariate analyses. The relative risk (RR) of recovery associated with vitamin A intake was greater in infants up to 1 y old (RR = 3.3, CI: 0.9-11.7) than in children > or =3 y of age (RR = 1.0, CI: 0.8-1. 3) (P:-value for interaction = 0.08). Diets rich in carotenoids may increase the rate of recovery from stunting in children. Dietary effects on growth might be strongest among very young children and those who have been most malnourished. Age, sex, breast-feeding status, socioeconomic status and severity of baseline stunting also were associated with reversal of stunting in this population.