Impaired bioavailability of vitamin A in adults and children with persistent diarrhoea in Zambia P. KELLY*  , J. MUSUKU à , E. KAFWEMBE§, G. LIBBY*, I. ZULU  , J. MURPHY ± & M. J. G. FARTHING* *Department of Adult and Paediatric Gastroenterology, St Bartholomew's and Royal London School of Medicine and Dentistry, London, UK; Departments of  Medicine and àPaediatrics, University of Zambia School of Medicine, Lusaka, Zambia; §Tropical Diseases Research Centre, Ndola Central Hospital, Ndola, Zambia; and ± Institute of Human Nutrition, University of Southampton, Southampton, UK Accepted for publication 9 March 2001 INTRODUCTION Vitamin A is an important micronutrient comprising several retinoids: retinol, retinoic acid, and retinol esters. Its biological activities encompass effects on epithelial integrity and gene transcription, and there is considerable evidence that vitamin A supplementa- tion has important bene®ts for children in the tropics. 1±3 The Zambian population is at high risk of vitamin A de®ciency as demonstrated by low serum retinol concentrations in 20% of adults living in Ndola (E. Kafwembe, unpubl. obs.). The HIV epidemic has had a severe impact on the Zambian population, SUMMARY Background: We have previously demonstrated a strong relationship between low serum retinol concentration and mortality in Zambian AIDS patients with diarrhoea, but were unable to detect any bene®t from oral micronutrient supplementation. Aim: To test the hypothesis that this is related to impaired availability of vitamin A, we analysed serum retinol concentration changes over 6 h following oral mega-dose therapy (60, 120 or 180 mg retinol). Methods: Twenty-four men without diarrhoea, 15 adults with persistent diarrhoea and 11 children (six girls, ®ve boys) with persistent diarrhoea were studied. Results: Men with persistent diarrhoea had lower baseline serum retinol concentrations (median 0.39 lmol/L, interquartile range 0.21±0.56) than con- trols (median 1.16 lmol/L, interquartile range 0.84±1.47; P  0.0003). After 60 mg retinol, the rise in serum retinol in HIV seropositive controls (median 0.63 lmol/L, interquartile range 0.35±0.77) did not differ signi®cantly from that observed in HIV seronega- tive controls (median 0.35 lmol/L, interquartile range ± 0.04±0.56; P  0.20). Increasing the dose to 120 mg or 180 mg retinol did not enhance the increase in serum retinol concentration. The increase in serum retinol was less in adults with persistent diarrhoea (median 0.25 lmol/L, interquartile range 0.04±0.35) and in children (median 0.11 lmol/L, interquartile range 0.04±0.46) than in men without diarrhoea (median 0.44 lmol/L, interquartile range 0.26±0.74; P  0.03). Adults and children with diarrhoea had greater losses of retinol in urine over a 24-h period than controls, but less than 1% of the ingested dose was excreted. Conclusions: These results suggest that persistent diar- rhoea in this population is associated with reduced bioavailability of retinol. Further work is required to determine the metabolic fate of therapeutic doses of retinol and to determine appropriate replacement strategies for HIV infected individuals. Correspondence to: Dr P. Kelly, Department of Adult and Paediatric Gastroenterology, St Bartholomew's and The Royal London School of Medicine and Dentistry, Turner Street, London E1 2AD, UK. E-mail: mpkelly@mds.qmw.ac.uk Aliment Pharmacol Ther 2001; 15: 973±979. Ó 2001 Blackwell Science Ltd 973