Whereas there is much information concerning the effects of vitamin A status on response to infectious challenge, the effects of infection or trauma on vitamin A metabolism and status are less well documented. These relationships need to be understood to optimize clinical and public health programs to improve vitamin A status and health of children in less-developed countries. We measured acute changes in retinol and retinol-binding protein in 57 young South African children hospitalized following respiratory epithelial damage resulting from accidental ingestion of kerosene. In addition, vitamin A status, as measured by the modified relative dose response test, of these children 3 mo later was compared with that of neighborhood control children to determine whether their illness had depleted retinol stores. Plasma retinol was already significantly below control levels when children were admitted [geometric mean (95% CI): 0.57 micromol/L (0.48-0.67) compared with 1.15 micromol/L (1.02-1.30) for controls] and decreased further the following morning [0.38 micromol/L (0.31-0.46)]. Significant differences in retinol-binding protein were not detected until the next morning [5.99 mg/L (4.70-7.63) compared with 14.0 mg/L (11.8-16.6) for controls] and were not as large as the relative differences in retinol. This dissociation between changes in retinol and its binding protein suggests that there may be increased retinol uptake by certain tissues during the acute phase response. The proportion of case children (37/46, 80%) with inadequate liver retinol stores 3 mo after the illness was slightly, but not significantly (chi2 = 2.16, P = 0.14), greater than the proportion of control children (28/42, 67%). Acute respiratory illness therefore did not further deplete retinol stores in this population in which stores were already frequently inadequate.

Birth to Ten is a longitudinal birth cohort study which began in April 1990 in the Johannesburg/Soweto area of Transvaal, South Africa. In this paper, the reason for the initiation of the study and its location in the current sociopolitical context is discussed. The health status of South Africa children in terms of infant mortality, morbidity (notifiable diseases) and nutritional status is described and mention is made of measures of psychological health and the importance of environmental pollution on health. The existing health service infrastructure in the study area is described and the fragmentation of health services between races, between preventive and curative services and on a geographical basis is highlighted. The study objectives, design, population, inclusion and exclusion criteria, methods of measurement and logistics are described.

Birth to Ten is a longitudinal birth cohort study which began in April 1990 in the Johannesburg/Soweto area of Transvaal, South Africa. In this paper, the reason for the initiation of the study and its location in the current sociopolitical context is discussed. The health status of South African children in terms of infant mortality, morbidity (notifiable diseases) and nutritional status is described and mention is made of measures of psychological health and the importance of environmental pollution on health. The existing health service infrastructure in the study area is described and the fragmentation of health services between races, between preventive and curative services and on a geographical basis is highlighted. The study objectives, design, population, inclusion and exclusion criteria, methods of measurement and logistics are described.

Kerosene pneumonitis is usually self-limiting and secondary infection is rare. Long-term studies in developed countries have produced conflicting results about complications. The position in developing countries, where children are exposed to adverse environmental and nutritional factors, is unknown. The aim of the present work was to determine whether there is an increase in respiratory or other illnesses following kerosene pneumonitis and whether these changes could be related to the severity of the initial lung damage. Fifty-seven children with clinical signs of pneumonitis were examined on admission and after overnight observation. Clinical signs were assessed for their usefulness for predicting severity. Cases and matched neighbourhood controls were seen every 2 weeks for 3 months. The time to predict most reliably the severity of short-term ill effects was 12-24 hours after the initial insult. There was no significant difference in respiratory symptoms during the 3-month follow-up in cases compared with controls. However, mild diarrhoea and fever were reported significantly more often in cases than in controls. Morbidity after clinical recovery was not shown to be a problem, irrespective of the severity of the acute event.