Objective To study the clinical characteristics and predictors of mortality from snake bite envenomation in children.

Design Prospective observational study with a one-group cohort design.

Setting Paediatric intensive care unit of a tertiary care hospital in South India.

Subjects The study cohort consisted of 145 children (55 girls and 90 boys) <12 years of age with snake bite envenomation.

Methods Demographic and clinical details were recorded in a semistructured pro forma. Children were treated with polyvalent antisnake venom (ASV) as per WHO protocol. Details of treatment, complications and outcomes were recorded. Univariate analysis was done to identify statistical significance, and those variables found to be significant were analysed using binary logistic regression.

Results Russell's viper was the most common offending snake followed by hump-nosed pit viper. Features of haemotoxicity, neurotoxicity and combined haemotoxicity and neurotoxicity occurred in 68 (47%), 39 (26.9%) and 9 (6%) children, respectively. Acute kidney injury (AKI) occurred in 36 (25%) children. The mortality rate was 10.3%. On univariate analysis, nocturnal bites, severe leucocytosis on day 1, AKI, capillary leak syndrome and a need for more than 20 vials of ASV were significantly associated with mortality. On multivariate analysis, only severe leucocytosis on day 1 (OR 35.29; 95% CI 1.37 to 911.89) and AKI (OR 35.05 95% CI 1.74 to 706.93) were found to be independent predictors of mortality.

Conclusions This study has identified two hitherto unrecognised risk factors—severe leucocytosis on day 1 and capillary leak syndrome. These findings need to be taken into consideration when planning management strategies for snake bite envenomation in children.