探討大腸直腸癌和血脂異常對動脈粥狀硬化相關疾病風險的關聯性

Abstract

背景： 動脈粥樣硬化疾病會因血脂的異常導致一系列與血管阻塞相關的疾病，如缺血性心臟病、缺血性腦血管疾病及周邊動脈疾病。糖尿病不僅是動脈粥樣硬化性疾病的關鍵風險因素，還顯著增加大腸直腸癌的風險，對預防和管理此類疾病具有重要意義。 目的： (一)探討在不同糖尿病狀態和嚴重程度的大腸直腸癌患者中，接受治癒性切除手術後的大腸直腸癌預後的差異 (二)檢查大腸直腸癌與後續糖尿病風險之間的關係，並進一步探討化學治療對大腸直腸癌患者糖尿病風險的影響情形 (三)利用觀察性和孟德爾隨機分派試驗方法探索脂蛋白(a)與粥狀動脈硬化疾病和全死因風險之間的非線性關係。 方法: (一) 利用世代追蹤研究，2007年至2015年間癌症登記資料庫串聯健保資料庫和死亡登記檔。在接受治癒性切除手術的I至III期大腸直腸癌患者，根據其糖尿病狀態被分為三組：無糖尿病、有糖尿病但無併發症和有糖尿病且有併發症。使用Cox回歸模型來探討糖尿病嚴重性與大腸直腸癌預後之間的關聯，包括總體存活率(OS)、無病存活率(DFS)、復發和癌症特異性存活率(CSS) (二) 利用2007年至2018年間癌症登記資料庫與健保資料庫，86,268名大腸直腸癌患者和86,268名由傾向分數配對的沒有大腸直腸癌的一般族群，探討大腸直腸癌是否增加新發糖尿病的風險。進一步分析2007年至2016年間癌症登記資料庫中的37,277名大腸直腸癌患者，以評估其化學治療暴露與糖尿病風險之間的關聯。化學治療暴露被分為四個類別：無化學治療、少於90天、90至180天以及超過180天。研究評估了這些化學治療暴露類別中糖尿病風險的差異 (三) 通過觀察性研究和孟德爾隨機分派試驗，利用2006年3月至2023年5月期間的UK Biobank，探討脂蛋白(a)與不同種族atherosclerotic cardiovascular disease (ASCVD) 風險之間的潛在線性和非線性關係。排除已確立缺血性心臟病或遺傳數據缺失的個案。參與者根據其脂蛋白(a)分為四組：低於70百分位數、70至80百分位數之間以及高於90百分位數，並作為連續變數進行分析。主要有興趣的結果為動脈粥樣硬化性心血管疾病，包括缺血性心臟病、缺血性中風和周邊動脈病的住院和死亡事件，以及全因死亡率。 結果: (一) 在對 71,747 名接受治癒性切除手術(第 I 至第 III 期)大腸直腸癌患者的分析中，依照糖尿病狀態分為三組：44,944 名無糖尿病、8,864 名有糖尿病但無併發症、5,394 名有糖尿病且有併發症。最終納入 59,202 名大腸直腸癌患者。與無糖尿病組比較，雖然「糖尿病但無併發症」組在整體存活率(OS；HR:1.05，95% CI:1.01–1.09)、無病存活率(DFS；HR:1.08，95% CI:1.04–1.12)及癌症特異存活率(CSS；HR:0.98，95% CI:0.93–1.03)表現略差，但差異並未達顯著。相較之下，「有併發症的糖尿病」組則顯著增加不良存活風險(OS：HR:1.85，95% CI:1.78–1.92；DFS：HR:1.75，95% CI:1.69–1.82；CSS：HR:1.41，95% CI:1.33–1.49)。此外，患有糖尿病之大腸直腸癌患者的復發風險也高於無糖尿病者。 (二) 本架構分析涵蓋 89,900 名第 I–IV 期大腸直腸癌患者，並與一般族群中 86,268 名非大腸直腸癌者進行 1:1 傾向分數配對。經配對後，每組皆保留 86,268 名參與者。整體結果顯示，大腸直腸癌患者罹患糖尿病的風險比配對後之一般人口高出 14%(HR:1.14；95% CI:1.09–1.20)，且在確診後第一年達到最高，之後則維持在較高的糖尿病的風險。此外，本研究亦探討 37,277 名第 II–IV 期大腸直腸癌患者接受不同化學治療時長與後續糖尿病風險的關聯。患者依照化學治療持續時間分為：未接受化學治療(15,018 例)、化學治療三個月內(12,538 例)、化學治療三至六個月(5,313 例)、以及化學治療超過六個月(4,408 例)。結果顯示，若在三年內化學治療總天數超過 180 天，後續糖尿病風險將提高約 60% 至 70%(HR:1.64；95% CI:1.07–2.49)。性別與 TNM 分期則為重要的效應修飾因子。 (三) 本研究依據脂蛋白 (a) 濃度將受試者分成四組，結果顯示：低於第 70 百分位(n = 261,486)發生 2,682 起缺血性心臟病事件及 22,220 例全因死亡；第 70–80 百分位(n = 37,318)發生 445 起缺血性心臟病事件及 3,358 例全因死亡；第 80–90 百分位(n = 37,330)發生 503 起缺血性心臟病事件及 2,999 例全因死亡；而高於第 90 百分位(n = 37,267)則觀察到最高比例之 547 起缺血性心臟病事件及 2,312 例全因死亡。Kaplan–Meier 存活曲線分析顯示，四組之缺血性心臟病事件累積發生率存在顯著差異(log–rank test < 0.001)，且脂蛋白 (a) 濃度越高，事件風險亦越高。非線性模型分析(restricted cubic splines)進一步指出脂蛋白 (a) 與缺血性心臟病風險之間具有非線性關係。為釐清此關係的因果性，本研究利用孟德爾隨機分派試驗，並使用候選基因分析(candidate SNPs approach)及UK Biobank 的關鍵 SNP(lead SNP)，運用殘差法(residual method)與雙排名法(doubly rank method)以確認脂蛋白 (a) 與心血管疾病風險之間的非線性因果關聯。 結論： (一) 在接受治癒性切除手術的大腸直腸癌患者中，糖尿病的嚴重程度與長期結果之間呈負相關，尤其於女性以及較早期階段的大腸直腸癌患者中更為顯著。 (二) 大腸直腸癌患者的糖尿病風險明顯升高，尤其長期化學治療(特別是含 capecitabine 的療程)會進一步增加糖尿病的發生機率。因此，對於大腸直腸癌患者而言，尤其在化學治療期間，密切監測血糖極為重要。 (三) 高濃度脂蛋白 (a) 顯著提升缺血性心臟病與全因死亡風險，且邊緣性呈現非線性關係；臨床上宜納入常規風險評估並及早預防。隨著針對脂蛋白 (a) 的新藥物持續研發，搭配基因檢測與精準醫療策略，未來或能更有效降低此族群的心血管事件與死亡風險。

Background Atherosclerotic disease results from abnormal blood lipid levels and encompasses various vascular obstruction–related conditions, such as ischemic heart disease, ischemic cerebrovascular disease, and peripheral arterial disease. Diabetes mellitus is not only a key risk factor for atherosclerotic conditions but also substantially raises the risk of colorectal cancer, underscoring its importance in the prevention and management of these diseases. The proposal includes three study projects: (1)Using the Taiwan Cancer Registry database to explore the prognostic variations following curative resection of CRC among patients with differing levels of diabetic control; (2) Examining the association between CRC and the risk of subsequent diabetes mellitus and investigating the impact of chemotherapy on diabetes mellitus risk in CRC patients; (3) Using the UK Biobank to explore the nonlinear relationship between lipoprotein(a) levels and the risks of ASCVD and all–cause mortality, utilizing both observational and Mendelian randomization methods. Objectives 1. To investigate differences in colorectal cancer prognosis among patients with varying diabetes status and severity following curative resection. 2. To examine the relationship between colorectal cancer and subsequent diabetes risk and further evaluate the impact of chemotherapy on diabetes risk in patients with colorectal cancer. 3. To use both observational and Mendelian randomization approaches to explore potential nonlinear associations between lipoprotein(a) and atherosclerotic cardiovascular disease (ASCVD). Methods 1. We conducted a cohort study using cancer registry data from 2007 to 2015 linked with the National Health Insurance database and death registry. Patients with stage I to III colorectal cancer who underwent curative resection were classified into three groups based on diabetes status: no diabetes, diabetes without complications, and diabetes with complications. We applied Cox proportional hazards models to analyze the relationship between diabetes severity and colorectal cancer outcomes, including overall survival (OS), disease–free survival (DFS), recurrence, and cancer–specific survival (CSS). 2. Using cancer registry data from 2007 to 2018 linked with the National Health Insurance database, we identified 86,268 patients with colorectal cancer and 86,268 propensity score–matched individuals without colorectal cancer to investigate whether colorectal cancer increases the risk of new–onset diabetes. We then further examined 37,277 patients with stage II to IV colorectal cancer (2007–2016) to assess the association between chemotherapy exposure—categorized as none, <90 days, 90–180 days, and >180 days—and diabetes risk. 3. Utilizing both observational analyses and Mendelian randomization, this study used data from the UK Biobank between March 2006 and May 2023 to investigate potential linear and nonlinear relationships between lipoprotein(a) and ASCVD risk, excluding those already diagnosed with ischemic heart disease or lacking genetic data. Participants were stratified by lipoprotein(a) levels into four groups—below the 70th percentile, 70th–80th percentile, 80th–90th percentile, and above the 90th percentile—and analyzed as a continuous variable. The primary outcomes of interest included hospital admissions and death events related to ASCVD (ischemic heart disease, ischemic stroke, and peripheral arterial disease). Results 1. Among 71,747 patients with stage I–III colorectal cancer who underwent curative resection, three groups were defined by diabetes status: 44,944 without diabetes, 8,864 with diabetes but no complications, and 5,394 with diabetes and complications, ultimately including 59,202 patients in the analysis. Compared with patients without diabetes, those with diabetes but no complications showed a marginally worse prognosis (OS: HR:1.05, 95% CI:1.01–1.09; DFS: HR:1.08, 95% CI:1.04–1.12; CSS: HR:0.98, 95% CI:0.93–1.03) without reaching statistical significance. By contrast, diabetes with complications was associated with a markedly higher risk of unfavorable outcomes (OS: HR:1.85, 95% CI:1.78–1.92; DFS: HR:1.75, 95% CI:1.69–1.82; CSS: HR:1.41, 95% CI:1.33–1.49). Colorectal cancer patients with diabetes also experienced a higher recurrence risk than those without diabetes. 2. In the analysis involving 89,900 patients with stage I–IV colorectal cancer matched 1:1 with 86,268 individuals from the general population, each group retained 86,268 participants after propensity score matching. Overall, colorectal cancer conferred a 14% higher risk of diabetes compared with matched individuals (HR:1.14, 95% CI:1.09–1.20), peaking within the first-year post–diagnosis and remaining elevated thereafter. Further investigation among 37,277 patients with stage II–IV colorectal cancer revealed that those who underwent more than 180 days of chemotherapy within three years were 60–70% more likely to develop diabetes (HR:1.64, 95% CI:1.07–2.49). Sex and TNM stage emerged as important effect modifiers. 3. Based on lipoprotein(a) levels, participants were allocated to four groups: below the 70th percentile (n=261,486; 2,682 ischemic heart disease events and 22,220 all–cause deaths), 70th–80th percentile (n=37,318; 445 ischemic heart disease events and 3,358 all–cause deaths), 80th–90th percentile (n=37,330; 503 ischemic heart disease events and 2,999 all–cause deaths), and above the 90th percentile (n=37,267; 547 ischemic heart disease events and 2,312 all–cause deaths). Kaplan–Meier curves indicated significant differences in the cumulative incidence of ischemic heart disease across groups (log–rank test <0.001), with higher lipoprotein(a) levels associated with greater risk. Nonlinear modeling (restricted cubic splines) suggested a nonlinear association between lipoprotein(a) and ischemic heart disease risk. To clarify causality, the study employed Mendelian randomization combined with a candidate SNPs approach and lead SNPs from the UK Biobank, utilizing the residual method and doubly rank method to confirm the nonlinear causal relationship between lipoprotein(a) and cardiovascular disease risk. Conclusion 1.Among patients with stage I–III colorectal cancer undergoing curative resection, diabetes severity was adversely associated with long–term outcomes, especially in women and those with earlier–stage disease. 2.Colorectal cancer patients face a significantly higher risk of developing diabetes, which is further elevated by prolonged chemotherapy, particularly regimens involving capecitabine. Close glycemic monitoring is therefore crucial for these patients, especially during chemotherapy. 3.Elevated lipoprotein(a) levels substantially increase the risks of ischemic heart disease, showing a marginally nonlinear pattern. Incorporating lipoprotein(a) into routine risk assessment and initiating early preventive measures are advisable. As new therapeutic agents targeting lipoprotein(a) continue to emerge, integrating genetic testing and precision medicine may more effectively reduce cardiovascular risk in this high–risk population.