Few case-control studies have examined possible risk factors for diarr
hoeal deaths in under-age-five children in the developing countries. W
e analysed data from the surveillance system of our diarrhoea treatmen
t centre/hospital for the period 1990-94 on 928 children less than 5 y
ears of age. In univariate analysis, II factors were significantly ass
ociated with death: lack of breastfeeding, severe malnutrition, compli
cated diarrhoea, pneumonia, xerophthalmia, duration of diarrhoea 7-14
days, moderate or severe dehydration, recent history of measles, Shige
lla flexneri infection, maternal illiteracy, and very low household in
come. Rotavirus diarrhoea was negatively associated with fatal outcome
. In the assessment of severe malnutrition, weight-for-height measurem
ent discriminated mortality risk better than weight-for-age or height-
for-age indices. Only two factors retained their significance, severe
malnutrition and non-breastfeeding in the multivariate analysis with a
djusted odds ratio (95% confidence interval) of 84.2 (9.1, 775.9) and
4.2 (1.3, 13.2) respectively.