Predictors of mortality in a cohort of adult Sri Lankans

Abstract

INTRODUCTION & OBJECTIVES: There is limited knowledge on early predictors of mortality among Sri Lankans. We investigated predictors of mortality in the Ragama Health Study cohort 10 years after recruitment. METHODS: The study population (35-64-year-olds selected by age-stratified random sampling from the Ragama Medical Officer of Health area) was initially screened in 2007 for socio-demographic, lifestyle, anthropometric and clinical risk factors with informed written consent. Their vital status was assessed in 2017 and the causes of death of the deceased were retrieved from death certificates available with the next-of-kin or declaration of death records from the hospital where the death occurred. RESULTS: Of the 2986 enrolled in 2007, 169 [Males: 107 (63.3%)] had died over the 10-year period, giving an overall mortality rate of 5.7% (95% CI: 5.0 % - 6.4%). The number of deaths due to cardiovascular and cancer related causes were 73/169 and 31/169, respectively. On multivariate analysis using Cox's proportional hazards model, advancing age, male sex, lower educational level, unsafe alcohol consumption and the presence of metabolic syndrome at baseline were independent predictors of all-cause mortality. Advancing age, male sex, unsafe alcohol consumption and the presence of metabolic syndrome at baseline were independent predictors of cardiovascular mortality, while advancing age and unsafe alcohol consumption were the only independent predictors of cancer-related mortality. CONCLUSION: Unsafe alcohol consumption and the presence of metabolic syndrome were important modifiable risks for mortality in this population. Addressing these risks in early adulthood will help to improve longevity.