Burden and control of chronic obstructive pulmonary disease

Abstract

Background: Chronic obstructive pulmonary disease (COPD) poses a high societal burden that varies between China and the West. COPD is irreversible and incurable but preventable and avoidable. COPD screening programs are not currently recommended. Smoking cessation helps prevent COPD, but approximately one-third of COPD patients are non-smokers. A better understanding of COPD could help identify high-risk people and additional targets for intervention. Objectives: To provide a comprehensive exploration of the COPD burden, risk prediction, and drivers, I compared COPD incidence and mortality rates in China and the West, developed a risk prediction model for use in Chinese older adults, and assessed the role of the interleukin (IL)-1 family and its receptors (IL-1/IL-1Rs) in COPD using Mendelian randomization (MR). Given causal inference for diseases of old age, such as COPD, is open to bias from selective survival, I conceptualized the use of control exposures to triangulate MR estimates. Methods: Based on summary statistics from the Global Burden of Disease Study 2017, I investigated the roles of risk factors (e.g., tobacco smoke and air pollution) and competing risks (e.g., ischemic heart diseases [IHD], stroke, and tracheal, bronchus, and lung cancer [TBLC]) in shaping COPD incidence and mortality rates in China and high socio-demographic index (SDI) countries (which constitutes primarily Western countries) utilizing a two-stage analytical framework. Based on Hong Kong’s Elderly Health Service (EHS) Cohort, I evaluated the performance of an existing COPD risk prediction model, and developed the context-specific EHS-COPD model for Chinese older adults in Hong Kong. Next, I used MR, which is less likely subject to confounding, to assess the role of IL-1/IL-1Rs as the targets of intervention for COPD prevention by examining their associations with airflow obstruction (AO, a surrogate for COPD) using publicly available summary genetic statistics. Finally, I conceptualized the use of control exposures in triangulating MR estimates vulnerable to selection bias when selecting samples among survivors. Results: My findings showed that the COPD incidence and mortality rates in China have been higher than in high SDI counters since the 1990s, especially with a threefold mortality rate. Tobacco smoke and air pollution were the leading risk factors for COPD among men in both settings. However, the occupational risk was more relevant to Chinese women than those in high SDI countries. Besides these risk factors, different mortality patterns due to competing risks, such as IHD, stroke, and TBLC, may partly contribute to such discrepancies. The newly developed EHS-COPD model, including time-invariant context-specific predictors and time-varying effects of age and smoking, outperformed both the original (Kotz) and the recalibrated Kotz models in predicting 5-year COPD risk. Genetically predicted higher IL-1Ra and IL-18BP were nominally associated with AO. Finally, my proposal for using control exposures provides a practical approach to triangulate MR estimates in the presence of selection bias. Conclusions: Identifying high-risk groups using the EHS-COPD model, along with suggestive interventions on IL-1Ra and IL-18BP, could help reduce the COPD burden, particularly after accounting for competing risks. Routine use of control exposures in MR can strengthen causal inference.