尿液代謝產物和 B 型肝炎相關之肝細胞癌罹病的發生:前瞻性研究

Abstract

背景: 代謝體是人體代謝的最終產物，代謝物的改變能反應出生理異常，因此代謝物在癌症發展早期的偵測及診斷扮演重要的角色。本研究針對長期追蹤的B型肝炎帶原者，其為肝細胞癌之高危險群。透過代謝體分析平台，探討代謝分子和肝細胞癌發生的關係，藉以尋找可能的罹病風險預測標記。 材料方法: 本研究以一個包括5364名30歲以上的男性世代研究為基礎，分析355位B型肝炎帶原者及98位非帶原者的尿液檢體。第一階段研究係比對帶原者和非帶原者代謝小分子之分佈。在第二階段，針對B型肝炎帶原者進行巢式病例對照研究，納入分析170位在追蹤期間發生的肝細胞癌，對每名病例，本研究選取1-2名年齡和尿液收集時間匹配的無病世代成員做為對照個案。利用核磁共振技術探測尿液中代謝物，使用Wilcoxon rank test進行統計分析，再將和肝癌達顯著關係的代謝分子分成五分位數並用Mantel–Haenszel 檢定其變化趨勢，logistic regression model比較在校正干擾因子前後的劑量效應。 結果: 在24個辨認出來的代謝分子當中，有4個代謝分子（包含3-氨基異丁酸氨、甲尿乙內酰尿、甘胺酸及肌肝酸）和B型肝炎帶原狀態呈現顯著正向關係，但和肝細胞癌發生的關係不具統計顯著意義。另有3個代謝分子（包括苯乙酰甘氨酸、次黃嘌呤和托品酸）則是和B型肝炎帶原狀態呈正相關但和肝細胞癌的發生卻有反向的關係。檸檬酸、氧化三甲胺和牛磺酸和B型肝炎慢性感染無顯著關係，但是和肝細胞癌則有顯著相關。將這三個代謝分子切成五分位數，發現檸檬酸有正向的劑量效應，氧化三甲胺和牛磺酸雖然相較第一分位數都有下降的趨勢卻呈現U型相關。此結果在校正其他干擾因子後，並無改變。 結論: 本研究顯示檸檬酸、氧化三甲胺和牛磺酸可針對B型肝炎帶原者族群發展肝癌的過程作為生物標記。其中檸檬酸和肝癌有明確的正相關，而氧化三甲胺和牛磺酸雖然為負相關卻呈現U型的關係，則需要進一步的獨立研究。

Background: Metabolomics is the final downstream product of metabolism. Metabolomics alternations represent the aberrant of physiology, thus metabolomics plays a major role of early detection and diagnosis of cancer development. Present study focus on long term follow-up hepatitis B virus (HBV) carriers who are the high risk population of hepatocellular carcinoma (HCC). Investigating the relationship between metabolites and HCC through metabolomics analytic techniques enable to discover the potential biomarkers for predicting HCC risk. Materials and methods: Study subjects were obtained from a cohort of 5364 men aged over 30, including 355 HBV carriers and 98 non-carriers. During the first stage, we examined the difference of metabolites in HBV carriers and non-carriers group. Further, a nested case-control study design was conducted for HBV carriers, 170 HCC cases occurred during the follow-up. We matched control group for HCC among HBV carriers by age and beginning time of recruitment. Nuclear magnetic resonance (NMR) was applied to detect metabolites in urine samples. Wilcoxon rank test was used to perform univariate analysis. Those metabolites which significantly related to HCC were subsequently divided into five groups according to quintile cutoff points and Mantel–Haenszel was used to determine the trend of increasing quintile level. Finally, logistic regression model was adopted to compare the dose-response of metabolites before and after adjusting confounders. Results: A total of 24 metabolites was identified, 4 metabolites (3-aminoisobutyrate, N-methylhydantoin, glycine, creatinine) were positively correlated to HBsAg status but no statistical significance with HCC. Three metabolites (N-phenylacetylglycine, hypoxanthine, tropate) were found positively associated with HBsAg status while inversely associated with HCC. Three metabolites (citrate, taurine and trimethylamine N-oxide) were found related to HCC and were selected to further statistical analysis. According to quintile levels of three urinary metabolites, we found positive dose-response in citrate while decreasing trend was found in other two metabolites but represent a U-shape association. The result remain the same after adjustment of potential confounders. Conclusion: We found changes in metabolite profiles during transition from healthy HBV carrier status to HCC. Citrate has positive association with HCC for certain. Taurine and trimethylamine N-oxide have negative association but the U-shape relationship require a further confirmation by independent research.