Nutritional blindness and diarrhoea in Bangladesh.

Nutritional blindness due to vitamin A deficiency needlessly blinds 30000 preschool-age children in Bangladesh each year. The sight of over 100000 children is threatened by corneal lesions. These figures come from a national study by Helen Keller International (HKI), the Institute of Public Health and Nutrition (IPHN), and the Programme for the Prevention of Blindness. The International Centre for Diarrhoea1 Disease Research in Bangladesh (ICDDR,B) collaborated in data analysis. Over 22000 children under 6 years were examined by experienced ophthalmologists in 100 sites throughout rural Bangladesh and in the worst possible urban slums. A seminar on the findings was held in Dhaka in May 1983. Several related publications are available (Cohen et al. 1983a,b; IPHN and HKI, 1983). In summary, prevalence of rural nightblindness for children was 3.6%, Bitot spot 0.9% and active corneal lesions (X2/X3) 10.2 per 10000 (95% confidence limits 6.5-15.9 per 10000). It is encouraging to see the ICDDR,B addressing the association between diarrhoea1 disease and xerophthalmia. Unfortunately, the article by Khan et al. (1984) appears to contain serious errors. Methodological errors. The sample frame is biased, being based on self-reporting of eye complaints. Such bias may explain how, for example, over ten times fewer children under 6 years were found to have any of the eye signs or symptoms of xerophthalmia, as compared with the national study which systematically interviewed and examined each child in every sample cluster household. Still, it is difficult to understand why not a single blind child was reported from an all-ages population of 182000, as compared with a rural preschool-age binocular blindness rate of 6.4 per 10000 (95% confidence limits 3.7-11.2 per 10000) in the IPHN and HKI (1983) study. Analytical errors. The World Health Organization’s classification of xerophthalmia, in which corneal erosions, ulceration and keratomalacia are classified as X3A/X3B, has not been understood. The test only refers to corneal xerosis. More seriously, it is almost meaningless to give all-age rates in the summary for a disease whose most important effects predominantly involve young children. A similar comparison would be the use of crude (all-ages) rates for stroke. The dominance of xerophthalmia in young children is brought out in Fig. 1 of the paper by Khan et al. (1984) but here, as in other instances, prevalence (cases existing at the time of the survey) is completely confused with incidence (new cases). Errors in conclusions. It is difficult to equate the low prevalence reported for xerophthalmia by Khan et al. (1984) with serum vitamin A levels for 20% of the under 6 year-old population of below lOOpg/l, measured by the ICDDR,B in the same population in about the same year (Brown et al. 1980). Any attempt to do so should probably start with the high potency vitamin A capsule coverage in the study area for which no figure is given. We agree that much needs to be known about the association between diarrhoea and eye lesions threatening sight. The danger is that if health-policy makers were to follow the unrepresentative conclusions of Khan et al. (1984), which tend to diminish the scale of nutritional eye disease, even more Bangladeshi children risk blindness.