We prospectively studied clinical outcome and nutrition absorption in male children (6-36 months of age) with persistent diarrhea (PD) and severe protein-energy malnutrition (mean z score for age, -4.2 +/- 0.8). Fifty-one children were randomly allocated to receive a combination of khitchri, a rice-lentil mixture, yogurt, and half-strength buffalo milk (group A; n = 26) or full-strength soy formula (group B; n = 25) for 14 days. The initial caloric intake (p < 0.02) and mean weight gain (p < 0.02) were greater in the group B patients. The mean stool volume and frequency were comparable between the two groups. Seven children in group A were considered clinical failures by criteria of weight loss and PD, compared to three in group B. The coefficients of absorption (COA) for protein were similar between the two groups, but group B patients had higher values of COA for energy and fat (p < 0.05) in the second week of nutritional rehabilitation. The majority (63%) of the children improved on the khichri-yogurt-milk diet but the nutritional outcome was not as good as with the soy formula. These data indicate that a traditional diet can be successfully used in the dietary management of PD and severe malnutrition but that a more optimal composition, yielding a higher success rate, should be sought.