長期環境空氣污染物與青光眼、乾眼症和慢性結膜炎的相關性：一項臺灣代表性的世代研究

Abstract

長期接觸環境中的空氣污染物與慢性健康問題息息相關。眼睛因為是直接暴露於空氣污染物的器官，其易受傷害性值得特別關注。青光眼是在全球僅次於白內障的第二大失明原因，已成為重要的公共衛生挑戰。而雖然乾眼症與慢性結膜炎等眼表疾病較少導致失明，但嚴重時卻會對日常生活帶來困擾。先前已有研究探討這三種眼部疾病與空氣污染之間的關聯，但結果仍不一致，且在臺灣鮮少有研究聚焦於長期空氣污染的暴露與使用世代研究設計。 本研究使用來自台灣生物資料庫的資料，收錄20歲以上、具完全行為能力的台灣社區民眾。我們將此資料與健保資料庫連結，包括門急診與住院明細，以 ICD 編碼判定上述幾種眼科疾病診斷。在空氣污染的暴露上，估算參與者在開始參與研究前的十年間，六種主要環境空氣污染物（PM2.5、PM10、NO2、SO2、CO 和 O3）的累積暴露量，並進行標準化後分為三分位數。統計方法採用Cox 比例風險模型，在調整潛在干擾因子後，檢視不同污染暴露三分位數與疾病風險的相對危險性。 在排除已罹患該疾病或因重要資訊缺漏而不符資格的參與者後，Cohort 1（分析青光眼風險）包含 130,784 人，其中 1,525 人發生青光眼。Cohort 2（乾眼症）包含 124,590 人，其中 3,767 人發生乾眼症。Cohort 3（慢性結膜炎）包含 100,651 人，其中 6,606 人發生慢性結膜炎。多變量Cox迴歸分析顯示，NO2與CO與青光眼風險呈正相關，而O3則呈負相關。六種污染物皆與乾眼症與慢性結膜炎的風險呈顯著正相關。在交互作用檢測中，我們發現SO2在鄉村與郊區地區對青光眼風險的影響較強。而在乾眼症方面，除了O3外，其餘污染物在非都市地區的影響亦較強烈；而除NO2和CO外，其他污染物在這些地區對慢性結膜炎風險的影響相似。我們也使用限制立方樣條分析探索可能的非線性趨勢，結果顯示出不同污染物對於不同眼科疾病有明顯不同的模式。 我們的研究結果顯示，與交通相關的空氣污染物可能會增加青光眼風險，而乾眼症與慢性結膜炎等眼表疾病對空氣污染暴露特別敏感。值得注意的是，居住於鄉村與郊區的參與者受到空氣污染物的影響較都市居民更大。此外，我們探討了非線性的暴露與疾病風險之間關係，觀察到多樣的特殊模式，對於未來研究的研究設計與政策制定可提供參考。這些結果凸顯未來研究進一步確定特定污染物提高疾病風險的閾值之重要性。

Long-term exposure to ambient air pollution has been associated with chronic health damage. The eyes, organs directly exposed to air pollutants, warrant special attention regarding their vulnerability. Glaucoma is the second leading cause of blindness globally after cataracts, resulting in a significant public health challenge. Additionally, ocular surface diseases, including dry eye disease and chronic conjunctivitis, while rarely causing blindness, significantly impact daily life. Although previous studies have highlighted the association between these eye conditions and air pollution, their findings remain inconsistent. Research on long-term exposure and cohort study designs remains limited in Taiwan. Taiwan Biobank collects data from Taiwanese community residents aged 20 years or older with full legal capacity. We utilized data from participants enrolled between 2008 and 2020, with linkage to the National Health Insurance Research Database, which includes outpatient/emergency and inpatient claims, to determine disease diagnoses based on specific ICD codes. Cumulative exposure to six major ambient air pollutants (PM2.5, PM10, NO2, SO2, CO, and O3) over a decade preceding enrollment was estimated, standardized, and categorized into tertiles. Cox proportional hazards regression was used to examine the hazard ratios for disease risk associated with air pollution exposure in tertiles, adjusting for potential confounders. After excluding ineligible participants, Cohort 1 for the outcome of glaucoma included 130,784 individuals, among whom 1,525 developed glaucoma. Cohort 2 for the outcome of dry eye disease included 124,590 individuals, with 3,767 cases of dry eye disease identified. Cohort 3 for the outcome of chronic conjunctivitis included 100,651 individuals, with 6,606 cases. Multivariable Cox regression analysis revealed that, in Cohort 1, NO2 and CO were positively associated with the risk of glaucoma, while O3 was inversely associated. In Cohorts 2 and 3, all six pollutants showed significant associations with the elevated risks of dry eye disease and chronic conjunctivitis. In the detection of interactions, we found that SO₂ had a significantly stronger effect on glaucoma risk in rural and suburban areas. Moreover, all pollutants except O₃ showed remarkably stronger effects on dry eye risk in these areas, and all pollutants except NO₂ and CO exhibited similar effects on chronic conjunctivitis risk in these regions. We also used restricted cubic splines to explore potential non-linear trends, revealing distinct patterns. Our findings suggest that traffic-related air pollutants may increase the risk of glaucoma, while ocular surface diseases (i.e., dry eye disease and chronic conjunctivitis) are particularly sensitive to air pollution exposure. We also found that participants living in rural and suburban areas experienced stronger effects from air pollutants compared to those residing in urban areas. In addition, we explored non-linear relationships and observed a variety of patterns, offering further insights into advanced study design and policy-making in the future. These results warrant the need for future studies to determine the thresholds of specific pollutants that increase the risk of these diseases.