糖尿病增加結核病復發之風險：以人群為基礎之嵌入型病例對照研究

Abstract

背景 過去的研究顯示糖尿病會增加結核病（TB）的風險，和不利的結核病治療結果，例如:結核病的治療失敗和治療期間死亡; 特別是在發展中國家其糖尿病的盛行率不斷增加，將可能造成結核病與糖尿病共病流行的風險。目前的研究還不清楚糖尿病是否也增加了已完成治療結核病的病人再次復發的風險。 方法 以台灣2006-2007年所有確診的肺結核世代完成治療後分析結核病的復發率。接著以同一追蹤世代進行以人群為基礎的嵌入型回溯性病例對照研究，來探討糖尿病與結核病復發的相關性。研究世代是以在2006-2007年台灣所有確診的肺結核通報病例，在完成治療後，追蹤直到2010年12月31日。復發的病例定義為須細菌學或病理確認結核病復發並且通報至全國結核病資料庫。對照組的選擇採取發生密度抽樣，亦即在復發病人通報至結核病通報系統的同日，從研究世代中選取未復發的個案作為對照組，病例與對照個案數使用1:1的比例，並且以相同的結核病完成治療後的追蹤時間作為配對條件。我們自結核病通報系統收集人口學特徵，並且進行回溯性病歷回顧，收集在前次結核病治療期間的臨床資訊。我們用條件式邏輯斯回歸分析來衡量糖尿病和結核病復發之間的相關性。 結果 截至2010年12月31日共計有305位的復發病人, 佔21939位追蹤世代的1.4％。結核病復發的發生率487.6/100,000 (434.4-545.5/100,000)。在排除了經病歷回顧後不符合復發定義，和欲分析變項資料不明個案之後，一共有251位病例和251位對照個案進行最後的相關性分析。有糖尿病的結核病病人相較於沒有糖尿病病人有較高的結核病復發風險（勝算比 1.74，95％信賴區間：1.18 - 2.59）。在調整了性別、年齡、是否為原住民、是否飲酒、吸煙、癌症、末期腎臟疾病、同時合併肺外結核病、是否具有開洞病灶、是否結核病治療為非標準治療處方、以及公衛人員直接觀察病人服藥是否超過60％的前次結核病治療時間，糖尿病仍然增加結核病的復發風險（調整後勝算比1.84, 95％信賴區間：1.14-2.98）。 結論 糖尿病增加結核病復發的風險。而血糖的控制與結核病復發的相關性研究仍需要進一步研究，以了解在結核病治療期間改善血糖控制，是否可能降低結核病復發的風險。

Background Diabetes mellitus increases the risk of developing tuberculosis (TB), and is associated with unfavorable treatment outcomes, including treatment failure and death. It remains unclear whether diabetes also increases the risk of TB recurrence among those who had completed treatment. Methods We analyzed the national data from TB registry to estimate the incidence of recurrent TB recurrence rate among the 2006-2007 cohort of incident TB cases in Taiwan. Under the same cohort, we conducted a nested case-control study to investigate the association between DM status during previous anti-TB treatment and subsequent recurrence of TB. The study population, all confirmed pulmonary TB cases reported to Taiwan TB registry during 2006 to 2007 with completing treatment outcome, was followed till December 31, 2010. Recurrent cases were defined as illnesses in patients with recurrent bacteriological or pathological confirmation of TB after having completed previous anti-TB treatment. For case-control analysis, controls were selected from the study population with 1:1 ratio to cases using incidence-density sampling and matched on time since treatment completion. Demographic and clinical characteristics of previous TB episodes were collected through retrospective medical charts review and TB registration database. We used conditional logistic regression analysis to measure the association between diabetes and recurrent TB. Results There were 305 (1.4%) cases of recurrent TB among the 21,939 pulmonary TB patients enrolled; two hundred and fifty one matched case-control pairs were analyzed after excluding missing information. The estimated incidence of recurrent TB among the 2006-2007 cohort was 487.6/100,000 person year (434.4-545.5/100,000). The presence of diabetes during previous TB treatment increased the risk of TB recurrence (crude OR= 1.74, 95% CI: 1.18 - 2.59). The association remained unchanged after adjusting for gender, indigenous population, alcohol consumption, smoking, cancer, end-stage of renal disease, coexisting of extra-pulmonary lesion, initial cavitation, suboptimal regimen, and directly observed therapy over 60% of treatment duration (adjusted OR= 1.84, 95% CI: 1.14-2.98). Conclusions The presence of diabetes mellitus was independently associated with increased risk of developing recurrent TB. To prevent recurrent TB among this population, further studies are needed to understand whether improved glycemic control during anti-TB treatment decreases the risk of recurrence.