失眠與第二型糖尿病風險之觀察性研究

Insomnia and the Risk of Type 2 Diabetes Mellitus: An Observation Study

林嘉玲(Chia-Ling Lin)

指導教授 ： 張媚

國立臺灣大學/醫學院/護理學研究所/博士(2017年)

https://doi.org/10.6342/NTU201702699

摘要

目的: 本論文研究利用全民健保資料庫(Taiwan’s National Health Insurance Database, NHIRD)探討失眠研究族群第二型糖尿病(Type 2 Diabetes Mellitus, T2DM)之發生風險及累積效應，並與無失眠的參考族群進行各項特性的比較；另外，探討失眠族群使用不同類型安眠藥者第二型糖尿病發生風險之差異。
方法: 本研究資料來源為台灣全民健保資料庫2005年100萬人承保抽樣歸人檔。研究分析2000至2010年之申報資料；採回溯性世代研究設計(retrospective cohort study)，先排除1999年12月31日之前已有失眠或第二型糖尿病病史，定義2000年1月1日至2005年12月31日期間新診斷的94,535位失眠個案為研究族群，後續追蹤至2010年12月31日，計算其糖尿病發生率。另外，採隨機抽樣法選取同時段無失眠診斷的個案，配對性別、年齡及診斷時間，以2倍的人數189,070位作為參考族群，比較兩個族群發生糖尿病的風險差異。另外，將安眠藥使用分成BZD類、Non-BZD類、zolpidem及Mixed 組分析其與第二型糖尿病發生風險相關性。利用Cox proportional-hazards regression model 分析兩個族群、四組安眠藥使用者第二型糖尿病發生風險，呈現hazard ratios (HR) and 95% confidence intervals (CI)
結果: 失眠研究族群T2DM的發生率明顯高於參考族群(78.78 vs 39.45, per 1,000 person-years)，失眠族群T2DM的發生風險為參考族群的1.51倍 (95% CI 1.48-1.54)，達顯著差異。每增加一年失眠暴露時間，其T2DM發生風險增加1.2%。若將暴露時間分層，以暴露1年為參考組，暴露時間為2-4年、5-8年及>8年，其發生T2DM風險分別是參考族群的1.29倍(95% CI 1.25-1.33)、1.76倍(95% CI 1.73-1.79)、1.82倍(95% CI 1.77-1.87)。若以無失眠無使用安眠藥為參考族群，僅用BZDs、僅用Non-BZD、僅用zolpidem、Mixed組相對於參考族群，發生T2DM相對風險分別是1.77倍(95% CI 1.42-1.98)、1.65倍(95% CI1.27-1.87)、1.60倍(95% CI 1.10-1.77)、2.69倍(95% CI 1.51-3.43)；其中以同時使用zolpidem及使用BZDs者，其T2DM發生風險為無使用zolpidem且無使用BZD者1.62倍( 95% CI 1.29–1.87)。
結論: 失眠族群T2DM的發生風險相較於參考族群高，失眠暴露時間愈久，T2DM發生風險愈高。各種安眠藥使用者中，使用一種以上安眠藥者對T2DM發生風險的影響程度最大，其中同時使用zolpidem及BZDs者對T2DM發生風險具有加成協同(additively synergistic)。本研究結果可提供臨床照護者重新審視長期失眠與安眠藥使用與T2DM風險相關性，將有助於T2DM照護策略的擬定。

關鍵字

失眠 ； 安眠藥 ； 第二型糖尿病 ； 回溯性世代研究 ； 全民健保資料庫

並列摘要

Aim: The present study investigated the risk of type 2 diabetes mellitus (T2DM) in patients with and without insomnia and explored the risk of T2DM in patients with hypnotic use through the secondary analysis of the data from the Taiwan National Health Insurance Database.
Methods: The present retrospective study used the data from the Longitudinal Health Insurance Database of the Taiwan National Health Research Institute, which consists of all original claims data of 1 million beneficiaries randomly selected from the original registry for beneficiaries in 2005. We excluded patients with a history of insomnia or T2DM before December 31, 1999. The study group included 94,535 patients who received a diagnosis of insomnia between January 1, 2000, and December 31, 2005, and had the onset of T2DM before December 31, 2010. The incidence of T2DM in the study group was estimated. The reference group of 189,070 individuals was obtained by randomly selecting insured individuals without insomnia and T2DM and by two-fold frequency matching by sex, age, and index year. According to the different hypnotics used, the study group was divided into four subgroups: benzodiazepine (BZD), non-BZD, zolpidem, and mixed hypnotic users and the risk of T2DM was analyzed in each subgroup. We conducted a Cox proportional hazard regression analysis to estimate the effects of insomnia and hypnotic use on the risk of T2DM.
Results: During the follow-up period, the incidence of T2DM in the insomnia group was significantly higher than that in the reference group (78.78 vs 39.45, per 1,000 person-years). Overall, the insomnia group had a higher risk of T2DM than the reference group (adjusted hazard ratio [HR], 1.51; 95% confidence interval [CI], 1.48–1.54). The risk of T2DM increased by 1.2% with each additional year of insomnia duration. The risk of T2DM increased 1.29 (95% CI, 1.25–1.33), 1.76 (95% CI, 1.73–1.79), and 1.82 (95% CI, 1.77–1.87) times for an insomnia duration of 2–4, 5–8, and >8 years, respectively, compared with an insomnia duration of up to 1 year. Furthermore, the risk of T2DM in BZD, non-BZD, zolpidem, and mixed hypnotic users increased 1.77 (95% CI, 1.42–1.98), 1.65 (95% CI, 1.27–1.87), 1.60 (95% CI, 1.10–1.77), and 2.69 (95% CI, 1.51–3.43) times, respectively, compared with non-hypnotic users. In addition, patients using both zolpidem and BZDs had a higher risk of T2DM than those not using zolpidem or BZDs (adjusted HR, 1.62, 95% CI, 1.29–1.87).
Conclusion: Compared with the reference group, the insomnia group had a higher risk of T2DM, and the longer the insomnia duration, the higher the risk of T2DM. Patients using more than one type of hypnotics had a higher risk of T2DM. Furthermore, T2DM exerted a synergistic effect in patients using both zolpidem and BZDs compared with those without hypnotic use. The present results may help clinicians re-examine the relationship among chronic insomnia, hypnotic use, and the risk of T2DM and contribute to T2DM management.

並列關鍵字

Insomnia ； Hypnotics ； Type 2 diabetes mellitus ； Retrospective cohort study

參考文獻

Lee, H.H., Chen, Y.C., Chen, J.J., Lo, S.H., Guo, Y.L., & Hu, H.Y. (2017). Insomnia and the Risk of Atrial Fibrillation: A Population-Based Cohort Study. Acta Cardiologica Sinica, 33(2), 165.

楊雅婷、王敬蓉、葉玲瓏、許怡欣(2015)．運用健保研究資料庫進行之第二型糖尿病研究趨勢分析：以PubMed資料庫之期刊文獻為例．數據分析，10 (3)，

台北市：國立陽明大學公共衛生研究所。

沈茂庭、劉家慧（2009）．健保局對安眠藥管制藥品的管理與努力．醫療品質雜

行政院衛生署管制藥品管理局。