利用前瞻性世代追蹤研究探討非酒精性脂肪肝臟疾病於大腸直腸腺腫發生率之作用

Abstract

背景 非酒精性脂肪肝病（NAFLD）和大腸直腸腺腫有著共同的危險因子，如肥胖，糖尿病和代謝症候群等。我們合理的推測，非酒精性脂肪肝疾病可能會增加大腸直腸腺腫的風險。然而，對於內視鏡檢查後的腫瘤發生影響仍不清楚，因此值得調查。 目的 調查NAFLD在調整其他干擾因子包括肥胖，糖尿病，代謝症候群等後，是否為經大腸鏡檢查為正常的世代其後續大腸直腸腺腫發生的危險因子。 方法 本研究是一前瞻性世代追蹤研究。於台灣北部某醫學中心，收集了接受健康檢查無疾病的世代。非酒精性脂肪肝診斷是依據超音波下脂肪肝之有無，追蹤這些世代後續大腸直腸腺腫的發生。 2004年1月至和2012年12月期間，共有53960名參與者進行了篩檢。受試者如果符合下列條件將被排除在外：（1）腹部超音波或大腸鏡檢查未在同一天進行，（2）B型肝炎或C型肝炎病毒血清標誌陽性，（3）男性每天消耗> 30克酒精和女性每天消耗> 20克，（4）存在慢性肝臟疾病，包括病毒性，毒物和自身免疫性肝病，（5）清腸準備不佳或不完整的大腸鏡檢查，（6）有大腸瘜肉或經大腸瘜肉切除病史，（7）有發炎性腸道疾病病史，或（8）大腸直腸癌病史。一共有15526名（7248名男性和8278名女性）進入最後的分析。本研究使用Poisson regression估計NAFLD對大腸直腸腺腫發生的相對危險性（Relative Risk, RR）。 結果 共有849人於後續追蹤期間被診斷出大腸直腸腺腫。大腸直腸腺腫易發生於男性，年老(>=50 歲)，有脂肪肝，糖尿病，較高的血壓和異常血清三酸甘油酯的族群。大腸直腸腺腫的發生率在NAFLD組為15.5/1000人年，正常組11.6/1000人年。累積大腸直腸腺腫的發生率在NAFLD組高於無NAFLD組。男性和女性的趨勢是相似的。非酒精性脂肪肝患者有較高的累積大腸直腸腺腫風險。調整年齡，性別和個別代謝症候群因子後，大腸直腸腺腫和NAFLD之間的關係仍然顯著，其調整後RR為1.34 (95%信賴區間(Confidence Interval, CI)為1.13-1.60)。調整代謝症候群時，NAFLD仍然是大腸直腸腺腫的一個獨立的危險因子（調整後RR = 1.27（95％CI為1.03-1.56）。 結論 有非酒精性脂肪肝的族群其大腸直腸腺腫發生率較無非酒精性脂肪肝的族群高出34%，而且不會被代謝症候群或其個別因子如中央性肥胖（通過調整WC），低HDL所干擾。

Background Non-alcoholic fatty liver disease (NAFLD) and colorectal neoplasm share common risk factors, such as obesity, diabetes, and metabolic syndrome. It is reasonable to postulate that NAFLD may increase the risk of colorectal neoplasm. Its impact on incident neoplasm after negative colonoscopy, however, remains unclear and thus worthwhile of investigation. Objectives We aimed to investigate the association between NAFLD and the risk of incident colorectal neoplasm in a negative colonoscopy cohort with adjustment for confounding factors, such as obesity, diabetes, and metabolic syndrome. Methods This study is a prospective cohort follow-up study. We collected a healthy cohort which was initially free of colorectal neoplasm with evidence from colonoscopy in the health checkup setting from a medical center in northern Taiwan, defined the baseline status of study interests, NAFLD, based on abdominal ultrasonography at the recruited day, and prospectively followed the cohort for the incidence of colorectal neoplasm. During Jan 2004 and Dec 2012, a total of 53960 participants were screened. Subjects were excluded if the following criteria were met: (1) abdominal ultrasound or colonoscopy not conducted in the same day, (2) individuals with positive serologic markers of hepatitis B or hepatitis C virus, (3) men consuming >30 g of alcohol per day and women who consuming >20 g per day, (4) subjects with history of chronic liver disease, including viral, toxic and autoimmune liver diseases, (5) poor bowel preparation or incomplete colonoscopies, (6) subjects with a history of colorectal polypectomy, (7) individuals with inflammatory bowel disease, and (8) individuals with a history of colorectal cancer. There were a total of 15526 subjects (7248 men and 8278 women) in the final analysis. We used Poisson regression model to estimate the relative risk (RR) of NAFLD for developing colorectal neoplasm with adjustment for confounding factors in the whole cohort. Results A total of 849 participants had incident colorectal neoplasm during the follow-up period. Subjects with incident colorectal neoplasm were more likely to be men, elderly age, and having NAFLD, DM, higher blood pressure and abnormal serum triglyceride level at baseline. The incidence of colorectal neoplasm in NAFLD group was 15.5/1000 person years as compared to 11.6/1000 person years in normal subject group. The cumulative colorectal neoplasm incidence for subjects with baseline NAFLD was higher than NAFLD-free subjects. The trend for male and female was similar. A higher cumulative risk of neoplasm among patient with NAFLD compared to those without. After adjustment of age, sex, and individual metabolic syndrome components, the relationships between colorectal neoplasm and NAFLD remained statistically significant (aRR=1.34, 95% CI: 1.13-1.60). When the adjustment was made for metabolic syndrome rather than its components, NAFLD was still an independent risk factor on incident colorectal neoplasm (aRR=1.27, 95% CI: 1.03–1.56). Conclusions Subjects with NAFLD had a 34% higher colorectal neoplasm incidence compared with those without NAFLD, which is independent from metabolic syndrome components such as central obesity (by adjustment of WC), and low HDL.