We investigated the gastrointestinal handling and post-absorptive metabolic handling of [1,1,1-13C]tripalmitin and [1-13C]glycocholate during recovery from severe childhood malnutrition. Eight children were studied on three occasions: at admission (phase 1), during rapid catch-up growth (phase 2) and when weight-for-height had reached 90 % of the reference (phase 3). Breath samples were obtained over a 24 h period and stools were collected over 3 d following the administration of each tracer. At admission, the lipid content of stool expressed as a percentage of ingested lipid was 6 (range 0.7-28.9) but less variation was shown between children at phase 2 (3.3 (range 0.9-4.1)) and phase 3 (1.4 (range 0.4-2.5)). The excretion of 13C in stool varied markedly between children at admission (11.1 (sd 5.4) % administered dose) and during rehabilitation (phase 2, 15.4 (sd 16.5) % administered dose; phase 3, 6.2 (sd 10.2) % administered dose). About 5 % of the absorbed label was recovered on breath at each stage (% absorbed dose; phase 1, 5.1 (sd 6.0); phase 2, 5.2 (sd 3.1); phase 3, 6.4 (sd 6.6)). None of the children exhibited significant bile salt malabsorption as a consequence of small intestinal overgrowth. Of the 13C measured in stool, more label was recovered in fatty acids than triacylglycerols during each of the three phases and this was interpreted to reflect a failure to absorb the products of digestion. The results show that not all the children had problems associated with the digestion and absorption of 13C-labelled tripalmitin in severe malnutrition and during recovery, which was not reflected in gross lipid balance across the gastrointestinal tract. Absorbed lipid was more likely to be deposited as adipose tissue than to satisfy the immediate needs for energy.