Tuberculosis Incidence and Survival Analysis in Diabetes Cohort-Using Secondary Database

Current literature about the diabetes and tuberculosis co-infection is limited to studies on diabetes among tuberculosis patients, or hospital-based studies on specific populations. Therefore, the objectives of this study are to estimate the incidence and mortality of a diabetes cohort with tuberculosis co-infection in Taiwan using the National Health Insurance database and the Central Communicable disease notification system of the Center of Disease Control(CDC), and to compare survival between diabetics with tuberculosis and those without, and define the high risk group in these co-morbid population.e followed the diabetes cohort sample (n=154347) for five years starting in 2001, traced diagnosis records of diabetes patients (ICD-9-CM code 010~018) from the National Health Insurance database with the tuberculosis records from the Central Communicable Disease notification system of CDC. We also followed diabetics with tuberculosis to track new tuberculosis cases per year and to confirm death status using the death registration system. Survival analysis using the life table method and the log-rank test was conducted to compare the survival curves, Cox Proportional Hazard Model for testing the hazard ratios among the potential risk factors, and included time-dependent covariate to adjust for the Cox’s model.ncidence of co-morbidity with tuberculosis in the five-year diabetes cohort is 3.39 per thousand persons on average, and 1.91 per thousand persons after adjusting for age, which is 2.9 times higher than the incidence of tuberculosis (0.67 per thousand persons) of the general Taiwan population. Tuberculosis incidence in male diabetics is approximately 2.3 times that of female diabetics, and 7 times higher than that of the Taiwan population. About 34.6%~46.6% of the diabetes cohort are tuberculosis smear positives every year between 2001 and 2005. Mortality rates for the diabetes cohort from 2001 to 2005 are 158.90, 189.66, 182.24, 180.85, and 222.22 per thousand persons per year, respectively. The hazard ratio for male diabetics co-morbidity with tuberculosis is 2.39 times that of female diabetics, and diabetes patients of age 65 and over have the highest risk of tuberculosis co-morbidity with diabetes (HR: 7.31, p=0.005). After controlling for age, gender, and living area, diabetics with tuberculosis have a higher risk of death compared to diabetics without tuberculosis co-infection (HR: 2.36, p<0.0001).sing secondary database, our research indicates that diabetes patients are almost 3 times more likely to be co-infected with tuberculosis, after adjusting for age, gender, and living area. The results of this study will provide the critical information needed to create strategy in tuberculosis control in the future.