Predictors of mortality in children hospitalized with dysentery in Harare, Zimbabwe

Objectives: To document the pattern of complications and identify risk factors for subsequent mortality in a hospitalized paediatric population during a Shigella dysenteriae type 1 epidemic.

Design: Hospital based prognostic study.

Setting: Paediatric wards in Harare and Parirenyatwa tertiary referral hospitals.

Subjects: All children aged one month to 12 years admitted to the wards with a history of bloody diarrhoea between January 1993 to June 30 1994 were included in the study (n = 312).

Main outcome measures: Patterns of complications and predictors of mortality.

Results: The peak age group of the patients was 12 to 23 months. Ninety five children died giving an overall case fatality rate (CFR) of 30.4% (95% CI, 25.3 to 35.6). One third (n = 107) had haemolytic uraemic syndrome (HUS), among whom the CFR was 43%. Other complications included altered consciousness (35%), seizures (12%), coma (4.5%), severe dehydration (14%), abdominal distention (17%), rectal prolapse (5%) and ileus (6%). In a multivariate analysis, temperature < 36 degrees C (OR 2.12; 95% CI, 1.33 to 3.39), severe dehydration (OR, 1.70; 95% CI, 1.15 to 2.53), hyperkalemia (> 5.5 mmol/L) (OR, 1.74; 95% CI, 1.01 to 1.97), hyponatremia (< 120 mmol/L) (OR, 1.57; 95% CI, 1.17 to 2.11), urea (> 8 mmol/L) (OR, 1.74; 95% CI, 1.29 to 2.36), and abdominal distention (OR, 1.67; 95% CI, 1.16 to 2.41) were found to be predictors of mortality.

Conclusion: Some of the major clinical and laboratory features allowing the early identification of children at increased risk of dying during dysentery outbreaks have been delineated. The high fatality associated with complicated dysentery highlights the need for preventive strategies which are sustainable in the region.