結核後慢性阻塞性肺病：以臺灣人群為基礎之回溯性世代研究

Abstract

背景：結核病與慢性阻塞性疾病（chronic obstructive pulmonary disease，COPD）皆為全球重大死因，且COPD可能作為結核病最主要的呼吸道後遺症，惟現有文獻普遍來自結核病高負擔國家，且多數缺乏明確的時序關係、客觀的暴露測量，限制了因果推論之可靠性。借助本土大型健康資料庫的長期追蹤，本研究旨在探討既往結核病與COPD之間的關聯性，試圖處理潛在干擾及偏差來源，補充結核後肺部疾病的流行病學證據。

方法：本研究為一項基於臺灣成年人群的回溯性世代研究。利用美兆健檢紀錄，串聯全民健保申報、結核病通報登記與死亡統計資料，並依指標年月、年齡、性別、地區、所得、教育程度及累積吸菸量進行粗化精確配對。於2006年至2022年間，共納入1,932名結核病新發個案，以及133,455名未罹患結核病的非暴露組。以Cox比例風險模型與Fine-Gray次分佈風險模型估計結核病史對COPD的效應，另以廣義線性混合模型分析肺功能重複測量數據。

結果：相較於非暴露組，結核病患者後續存在更高的COPD（23.6 vs. 3.9每千人年，p值< 0.001）及相關住院（5.1 vs. 0.5每千人年，p值< 0.001）發生率。校正吸菸、職業暴露、身體質量指數等關鍵變項後，結核病仍與隨後發生COPD（調整後風險比 = 3.00，95%信賴區間：2.60–3.46）及相關住院（調整後風險比 = 3.09，95%信賴區間：2.29–4.16）風險有關。研究族群約有1%的COPD病例可歸因於先前結核病。即使未曾吸菸或成功完成抗結核治療，仍持續面臨罹病風險。此關聯在基線時未確診氣喘，以及痰塗片或培養陽性、胸部X光呈現空洞病灶的肺結核病人中尤其明顯。本研究亦觀察到過往結核病與氣喘風險增加及肺功能下降的相關性。競爭風險模型、敏感度分析，以及陰性對照設計的一致發現，進一步強化了結果的穩健性。

解釋：在結核病發生率較低的情境下，結核病仍應被視為影響COPD發展的獨立危險因子。為減少結核病所造成的終生健康負擔，除了積極推動結核病防治工作外，建立整全的結核後照護管理模式，同樣具有重要的公共衛生與臨床意義。

Background: Tuberculosis (TB) and chronic obstructive pulmonary disease (COPD) are both leading causes of death worldwide, and COPD may represent a major respiratory sequela of TB. However, existing evidence is predominantly derived from high TB-burden areas and often lacks clear temporality and appropriate exposure measurement, thereby limiting the credibility of causal inference. Leveraging large health databases in Taiwan, this study aimed to investigate the association between TB and subsequent COPD while addressing potential confounding and sources of bias.

Methods: We conducted a retrospective cohort study among participants aged ≥ 20 years from the MJ health examination. Data were linked to Taiwan’s National Health Insurance claims, TB registry, and vital statistics. Coarsened exact matching was performed based on the index time, age, sex, region, income, education level, and cumulative smoking exposure. Between 2006 and 2022, we included 1,932 incident TB cases and 133,455 non-TB individuals. Cox proportional hazards models and Fine-Gray subdistribution hazard models were used to estimate the effect of prior TB on COPD, and generalized linear mixed models were applied to analyze repeated lung function measurements.

Findings: Compared with the unexposed group, individuals with a TB history had higher incidence rates of COPD (23.6 vs. 3.9 per 1,000 person-years, p < 0.001) and COPD-related hospitalization (5.1 vs. 0.5 per 1,000 person-years, p < 0.001). After adjustment for smoking and other key covariates, prior TB remained significantly associated with an increased risk of developing COPD (adjusted hazard ratio [aHR] = 3.00, 95% CI: 2.60–3.46) and experiencing COPD-related hospitalizations (aHR = 3.09, 95% CI: 2.29–4.16). Approximately 1% of COPD cases were attributable to previous TB exposure. The elevated risk persisted among never-smokers and patients who had been successfully treated for TB. The association was evident in participants without asthma at baseline and those with smear- or culture-positive pulmonary TB or cavitary lesions on chest X-ray. TB was also associated with potential asthma risk and impaired lung function. The consistency of findings across competing risks, sensitivity analyses, and negative control design further supports the robustness of our results.

Interpretation: Even in a relatively low TB-incidence setting, TB should be considered an independent risk factor for COPD development and clinically relevant outcomes. In addition to strengthening TB control strategies, establishing holistic post-TB care and long-term health management is imperative to reduce the lifetime disease burden among TB survivors.