PRS15 Evaluating Disease and Economic Burden of Severe Asthma in India

Abstract

Objectives: To examine whether lung function is a significant predictor for incident diabetes, asthma, myocardial infarction, hypertension and all-cause mortality, using a competing risks model. Methods: The data source used was the National Health and Nutrition Examination Survey (Epidemiological follow-up). For this cohort study, participants were prospectively followed from baseline (1970) until the end of the epidemiological follow-up (1984). We measured incident events based on self-reports or hospital visit codes during the study time conditional on no prior history of the same event. Lung function was measured as forced expiratory volume in 1 second (FEV1). The covariates adjusted for the association were: gender, age, body mass index, smoking status, inactivity, marital status, alcohol consumption, chronic obstructive pulmonary disease (COPD), cancer, and Charslon's modified comorbidity score. We used the Fine-Gray subdistribution and cause-specific hazard models to quantify the hazard ratios for the association of FEV1 and the primary events of interest. Results: The final cohort included 3,018 participants (mean FEV1 value: 3.00 liter (standard deviation (SD):0.84); mean age: 45.22(SD:13.49)). We found that for one unit increase in FEV1 (liters), the subdistribution hazad ratio was 0.37(0.22,0.61), 0.20(0.10,0.40), 0.81(0.53,1.26), 0.84(0.71,0.99) and 0.86(0.62,1.18) for diabetes, asthma, heart attack, hypertension and all-cause mortality, respectively. The corresponding estimates for the cause-specific hazard ratios were 0.45(0.33,0.60), 0.43(0.31,0.60), 0.56(0.40,0.77), 0.73(0.63,0.86) and 0.66(0.51, 0.85). Conclusions: While the cause-specific hazard model indicates the instantaneous hazard ratios for the primary events of interest amongst all event-free individuals, the subdistribution hazard model provides insight into cumulative incidence function by adding to the denominator people who may have had experienced competing events. Amongst event-free individuals, FEV1 was statistically significantly associated with an increased instantaneous hazard for the primary events. Lower FEV1 values were also showed to increase the incidence of diabetes, asthma, and hypertension. Therefore, modifying lung function can potentially mitigate the risks of subsequent events.