LAUSR

BACKGROUND Pneumonia acquired in the community is a leading cause of hospitalization and death in under-five children. Predicting mortality in children remains a challenge. There is a need of consolidated scoring system to predict mortality in under-five children in developing nations. METHODS This is a hospital-based prospective nested case-control study, conducted in a tertiary care teaching hospital of north India. Included were under-five hospitalized children due to WHO defined severe community acquired pneumonia (CAP). Those who did not survive were categorized as cases, while those who were discharged were categorized as controls. RESULTS The mortality rate among the recruited 180 hospitalized children with severe CAP was 9.4%. The mortality in under-five children was higher among infants, children who resided in rural areas and were unimmunized or partially immunized for the present age. Mortality was also statistically significantly higher among under-five children with weight for age and weight for length/height below -2Z score; SpO2 < 90% at room air at admission, cyanosis, convulsion, high C-reactive protein (CRP), blood culture positive sepsis and end point consolidation. These predictors were found to be independent risk factors for the mortality after analyzing in multivariate model while presence of wheeze and exclusive breast feeding for first six months of life were found to be protective. The receiver operating characteristic (ROC) curve for respiratory index of severity in children (RISC) score has area under curve (AUC) 0.91 while AUC of RISC score with King George's Medical University (KGMU) modification has 0.88 for prediction of mortality. At the cut-off level of 3, the sensitivity of the RISC score in predicting mortality was 94.1% while the specificity was 73.6%. However, the sensitivity of the RISC score with KGMU modification in predicting mortality at cut-off level of 3 was 88.4% with a specificity of 74.8%. CONCLUSION Various predictors for mortality under-five children are young age, malnutrition, cyanosis, high CRP, blood culture positive sepsis and end point consolidation. It is also possible to predict mortality using RISC score which comprises simple variables and can be easily used at centers of periphery. Similar accuracy had been also found through the use of an age independent modified score (RISC score with KGMU modification).Lay summaryPneumonia is a primary cause of hospitalization as well as death among the children under the age of five. A variety of severity or mortality predicting scores have been produced for adults, but such scores for children are scarce. Furthermore, their utility in developing nations has not been proven. This is a hospital-based prospective study. Included were children under five (2 to 59 months) hospitalized due to severe community acquired pneumonia (CAP) defined as per World Health Organization (WHO) and were not hospitalized in last 14 days elsewhere. Those who did not survive were classified as cases while those who were discharged were classified as controls. A total of 200 consecutively hospitalized children with severe CAP based on WHO were screened and 180 children were recruited. Among recruited children, the percentage of mortality was 9.4% while 90.6% were discharged. The mortality was higher among children younger than 12 months, those belonged to rural area and were unimmunized or partially immunized for the present age. Mortality was also higher among under-five children with severe malnutrition, anemia, SpO2 < 90% at room air at admission, cyanosis, convulsion, thrombocytopenia, high CRP, blood culture positive sepsis and end point consolidation. After assessing in a multivariate model, these predictors were determined to be independent risk factor for death, while wheezing and exclusive breast feeding throughout the first six months of life were found to be protective. The receiver operating characteristic (ROC) curve for respiratory index of severity in children (RISC) score has an area under curve (AUC) of 0.91 while AUC of RISC score with King George's Medical University (KGMU) modification was 0.88 for the prediction of death in under-five children hospitalized due to severe CAP.