Predictors of In-Patient Mortality of Severe Acute Malnutrition of Hospitalised Children in a Tertiary Facility in Southern Nigeria

Abstract

Introduction Severe acute malnutrition (SAM) remains a public health concern in developing countries. Children with SAM are nine times more likely to die compared with those that are well-nourished. Most studies on SAM in Nigeria focus on disease burden without evaluating risk factors that may be useful as interventions in reducing mortality. This study evaluated predictors of mortality and outcome among hospitalised children with SAM in Southern Nigeria. Methods Children with SAM admitted into the paediatric medical ward of the University of Calabar Teaching Hospital between September 2017 and November 2019 were studied prospectively. A multivariable logistic regression was used to identify factors that independently predicted mortality with a p-value <0.05 considered significant. Results One hundred children were studied. The mean age was 14.28 ± 14.04 months, of which 89% were less than two years of age. Oedematous and non-oedematous SAM were found in 18.5% and 81.5%, respectively. Co-morbidities included tuberculosis (TB) (13.0%), HIV (12.0%), and HIV/TB co-infection (3.0%). Clinical presentation included fever (21.7%), anaemia (19.9%), diarrhoea (19.1%), skin changes (8.7%), and shock (1.8%). The mean duration of hospital stay was 11.48 ± 6.87 days. Nine of the children were discharged against medical advice and were excluded from further analysis. About 92.3% were discharged for follow-up and 7.7% died. After multivariable regression, the predictors of mortality were shock (p=0.037, adjusted odds ratio (aOR): 17.51, 95% confidence interval (95% CI): 1.19-258.77) and skin changes (p=0.035, aOR: 9.81, 95% CI: 1.18-81.46). Conclusion The presence of shock and skin changes are independently associated with mortality in hospitalised children with SAM. Prompt referral of children with SAM and more so with complications of shock and skin changes is hereby advocated to reduce mortality.