Introduction: To investigate whether the El Nino phenomenon and ambient temperature had an effect on the epidemiology of childhood diarrhoea, we analysed data on daily number of admissions of children with diarrhoea to the Oral Rehydration Unit of the Instituto de Salud del Nino in Lima, Peru, between January, 1993, and November, 1998. Methods: We obtained daily data on hospital admissions from the Oral Rehydration Unit, and meteororological data from the Peruvian Weather Service, and used time-series linear regression models to assess the effects of the 1997-98 El Nino event on admissions for diarrhoea. Findings: 57,331 children under 10 years old were admitted to the unit during the study. During the 1997-98 El Nino episode, mean ambient temperature in Lima increased up to 5°C above normal, and the number of daily admissions for diarrhoea increased to 200% of the previous rate. 6225 excess admissions were attributable to El Nino, and these cost US$277,000. During the period before the El Nino episode, admissions for diarrhoea increased by 8% per 1°C increase in mean ambient temperature. The effects of El Nino and ambient temperature on the number of admissions for diarrhoea were greatest during the winter months. Interpretation: El Nino had an effect on hospital admissions greater than that explained by the regular seasonal variability in ambient temperature. The excess increase in ambient temperature was the main environmental variable affecting admissions. If our findings are reproducible in other regions, diarrhoeal diseases may increase by millions of cases worldwide with each degree of increase in ambient temperature above normal.
Bibliographical noteFunding Information:
This study was supported by a National Research Service Award of the National Institutes of Child Health and Development (F31-HD08488), awarded to William Checkley. Partial support was provided by USAID through an Applied Research on Child Health grant awarded to AB PRISMA, an EPA Cooperative Agreement grant (CR-823130) awarded to the Johns Hopkins School of Public Health, an ITRED/NIH Fogarty Training Grant, and by the charitable RG-ER foundation for the advancement of climate research. We thank D Burke, MD Chestnut, and J Friedland for helpful comments, P Arkin and C Ropelewski for expert advice on time-series analysis, L Kelley for editorial assistance, M Varela and S Carrera for data entry, and J B Phu and D Sara for technical support.