Early identification of children at high risk of diarrhoea-associated dehydration would be of great value to health care workers in developing countries. To identify prognostic factors for life-threatening dehydration, we carried out a case-control study among under-2-year-olds in Porto Alegre, Brazil. Cases were 192 children admitted to hospital with moderate or severe dehydration, while controls were children matched to controls by neighbourhood and age, who experienced nondehydrating diarrhoea in the week preceding the interview. The following variables were significantly associated with an increased risk of dehydration, after adjustment for age and other confounding variables: absence of the father from the home; low paternal education level; young age; maternal age 25-29 years or less than 20 years; mother of mixed race; high birth order; short birth interval; low birth weight; stunting, underweight and wasting; lack of breast-feeding; presence of other under-5-year-olds in the home; families with 4-5 members; lack of antenatal care; less than three doses of diphtheria-pertussis-tetanus or poliomyelitis vaccine; previous admission to hospital; use of medicines during the fortnight prior to the episode; and living in an unclean home. The associations were particularly strong (P less than 0.001) for the child's age, birth weight and other anthropometric indicators, birth interval, and feeding mode. In terms of their sensitivity and specificity, however, these prognostic factors were not as effective as early signs and symptoms for predicting the outcome of the episode.
PIP: During the primary diarrhea season (December 1987-April 1988) in metropolitan Porto Alegre in southern Brazil, researchers compared 192 children aged less than 2 years who were admitted to a hospital with moderate or severe dehydration with 192 neighborhood- and age-matched controls who had a diarrhea episode without dehydration during the seven days before the interview. They aimed to identify factors predicting life-threatening dehydration. When the researchers controlled for age and other confounding variables, the following factors were significantly related to an increased risk of dehydration: no father in the household, low paternal education level, young age, maternal age 25-29 years or less than 20 years, mixed race mother, high birth order, short birth interval, low birth weight, stunting, underweight and wasting, non-breast milk, children aged less than 5 in the household, family size of 4-5, no prenatal care, less than three doses of diphtheria-pertussis-tetanus or poliomyelitis vaccine, previous admission to a hospital, use of medicines during the two weeks before the diarrhea episode, and living in an unsanitary household. The strongest factors associated with an increased risk of dehydration (p 0.001) included young age, low birth weight and malnutrition, short birth interval, and non-breast milk. These factors were not as effective at predicting an increased risk of dehydration as early signs and symptoms. Specifically, their sensitivities were lower than those of early signs and symptoms. Notwithstanding, these findings support current efforts towards promotion of breast feeding, prevention and treatment of malnutrition, and birth spacing since they contribute to the prevention of diarrhea-related dehydration.