Predictors of dengue mortality among severe dengue patients in University Malaya Medical Centre (UMMC), Kuala Lumpur, Malaysia

Abstract

Background: The number of dengue cases globally has grown rapidly since 2010 to 3.2 million cases in 2015. The case fatality rate in Malaysia has increase from below 0.2% to 0.28% in 2015. This puts a heavy clinical and financial burden on endemic countries including Malaysia. Recognizing early determinants of dengue mortality may reduce the fatality rate and the burden of the disease to countries. Methods and materials: All adult fatal dengue cases from January 2015 to March 2019 and all adult severe dengue cases available on hospital EMR from January 2018 to March 2019 in University Malaya Medical Centre (UMMC) were retrospectively included to the study. Demographic data, clinical and biochemical markers at presentation and during admission, and data on clinical outcomes were collected. Univariate analysis and multivariate logistic regression were done to determine the significant variables that can be used to predict dengue mortality by using AUROC for the specificity and sensitivity of the model. Results: There were 5246 dengue cases from January 2015 to March 2019. 18 cases were fatal. 1085 cases from January 2018 to March 2019 were reviewed on the hospital EMR. 104 had severe dengue. Univariate analysis showed 14 significant independent variables including diabetes mellitus, chronic renal disease, outbreak area, lethargy, white blood cell (WBC), serum lactate, potassium, urea and creatinine, total bilirubin, alkaline phosphatase, alanine aminotransferase, gamma-glutamyl transferase and aspartate aminotransferase. Multivariate analysis showed WBC with Adjusted Odds Ratio (AOR) = 0.644, p = 0.002 (95% CI: 0.489–0.849) and creatinine with AOR = 0.992, p = 0.015 (95% CI: 0.986–0.999) as significant predictors of mortality among severe dengue patients. Area Under Receiving Operative Characteristic (AUROC) for WBC at a cut off of 7 mmol/l and creatinine at a cut off of 131.5 mmol/l was 0.896 (p < 0.001) which indicated that they were good predictors for mortality. Conclusion: This study suggests the best prediction model to predict death among severe dengue patients is a model that includes white blood cell count and serum creatinine. These two tests are readily available in all health care settings.