Reported are our findings from a case-control study of the association between prolonged breast-feeding and clinical malnutrition in an urban setting in West Africa. The cases were children aged 12-36 months who had been hospitalized with a diagnosis of clinical malnutrition. Children of a similar age who lived in neighbouring courtyards were recruited as controls. For 152 case-control pairs in which both children were receiving solid foods, non-breast-feeding was associated with an increased risk of clinical malnutrition (crude odds ratio = 2.37; 95% confidence interval = 1.24, 4.55). This association remained statistically significant after controlling for various potentially confounding variables (P = 0.03). Our findings suggest that either prolonged breast-feeding may offer substantial protection against clinical malnutrition in the study population or malnutrition leads mothers to stop breast-feeding. These results are inconsistent with those of a number of workers who have reported that prolonged breast-feeding is associated with an increased risk of malnutrition. This inconsistency might have arisen because of differences in the definition of malnutrition used or because of variations in the quantity and quality of weaning foods available in different settings. We found no evidence to support the hypothesis that prolonged breast-feeding may be detrimental to children.