Nonalcoholic Steatohepatitis Is Associated With a Higher Risk of Advanced Colorectal Neoplasm

Yuri Cho; Soo-Kyung Lim; Sae Kyung Joo; Dong-Hyong Jeong; Jung Ho Kim; Jeong Mo Bae; Jeong Hwan Park; Mee Soo Chang; Dong Hyeon Lee; Yong Jin Jung; Byeong Gwan Kim; Donghee Kim; Kook Lae Lee; Won Kim


Liver International. 2019;39(9):1722-1731. 

In This Article


In this biopsy-assessed prospective NAFLD cohort study, the prevalence of advanced colorectal neoplasm was strongly associated with the presence of NASH, and this association persisted after further adjustment for age, sex, and the presence of diabetes mellitus. Patients with hepatic steatosis were at higher risk of developing adenomatous colorectal polyp than were those without hepatic steatosis. Moreover, patients with biopsy-proven NASH were found to have 2.8-fold increased risk for developing advanced colorectal neoplasm than those without any histological finding of NAFLD.

Previous studies have focused on the relationship between colorectal polyp and NAFLD as assessed by non-invasive markers, abdominal ultrasonography or computed tomography.[7,9,20] The use of non-invasive diagnostic assessment of NAFLD may result in misclassification of NASH and over-diagnosis of NASH leading to biased or inflated study results. In addition, recent studies have compared the prevalence of colorectal polyp in NAFLD patients with that in healthy control subjects without consideration of the histological severity of NAFLD.[11,25] In contrast to the previous reports, our study evaluated the risk of advanced colorectal neoplasm in a large biopsy-evaluated NAFLD cohort and demonstrated that the presence of NASH was an independent risk factor for advanced colorectal neoplasm. Although liver biopsy is an invasive procedure, the confirmatory diagnosis of NASH is possible only by this invasive procedure. Moreover, our study revealed a histological evidence-based association between biopsy-proven NASH and advanced colorectal neoplasm.

Recent studies have suggested that insulin resistance, metabolic syndrome, obesity, type 2 diabetes mellitus, and dyslipidemia are closely related to a higher risk of colorectal adenomas and NAFLD is also related to such factors.[13,26] NAFLD may precede and/or promote the development of metabolic syndrome.[1] Recent studies have shown a link between metabolic syndrome and the development of advanced colonic neoplasm.[27] From this bidirectional relationship, NAFLD might be associated with advanced colonic neoplasm. The mechanism by which NAFLD causes an increased risk of advanced colonic neoplasm is not fully understood. However, NAFLD represents a condition of profound insulin resistance and a proinflammatory state. Insulin and insulin-like growth factors may promote the development of advanced colonic neoplasm through their proliferative and anti-apoptotic effects.[28]

Fibrosis is a major histological harbinger of NAFLD prognosis because most studies have shown that the stage of fibrosis influences overall- and liver-related mortality among patients with NAFLD independently of the presence or severity of other histological features.[29] In the present study, the presence of NASH was correlated with the development of advanced colorectal neoplasm. Due to the uncertainty about the mechanism by which NASH is associated with an increased risk of advanced colonic neoplasm, further studies that evaluate the pathways leading from hepatic fibrosis to advanced colonic neoplasm are needed. Some studies proposed the roles of adiponectin, interleukin-6, tumor necrosis factor-α, leptin, and pro-inflammatory cytokines as relevant predictors of colorectal neoplasm.[30,31] Dysbiosis of gut microbiota, gut microbiota-medicated inflammation, and impaired mucosal immune function have been suggested as playing important roles in the pathogenesis of NAFLD,[32] which might lead to developing advanced colorectal neoplasm. These possible mediators should be evaluated by further molecular studies in the future.

Hwang et al reported previously the first evidence of a relationship between NAFLD and an increased risk of colorectal adenomatous polyp.[20] An increased risk of NAFLD was also evident in patients with more adenomatous polyps. Untreated patients would suffer from these polyps progressing to CRC according to an adenoma-carcinoma sequence.

A more recent study that assessed NAFLD severity by non-invasive tools, including Fibrosis-4 index and NAFLD fibrosis score, also revealed an association between NAFLD severity and colorectal neoplasm; however, the diagnostic accuracy of these non-invasive methods is questionable.[7,25] Furthermore, some of patients with NAFLD identified by these non-invasive tools indeed may have been misdiagnosed with NAFLD because they were not diagnosed by liver biopsy, which is a gold standard of the diagnosis of NAFLD.[33,34] Therefore, the diagnostic accuracy of non-invasive tools in those studies was potentially limited. Objective detection of NAFLD would be necessary for confirmation of those findings.

In the present study, NASH was associated with a higher risk of advanced colorectal neoplasm and these results suggested the benefit of CRC screening in NAFLD patients. Several features of our approach differed from those of other studies. First, our study included only biopsy-proven NAFLD patients. Using a reference tool for diagnosis of NAFLD, we could more accurately evaluate the evidence-based relationship of NAFLD and advanced colorectal neoplasm than non-invasive tool-based studies. Second, in contrast to the previous studies that also used liver biopsy as a diagnostic tool, our study was able to characterize in greater detail the histological features of patients with NAFLD by grading histological findings, including lobular inflammation, hepatocellular ballooning, and fibrosis severity. By considering detailed histological features in our analysis, we could obtain more complex information about the association between the histological severity and characteristics of NAFLD and the development of advanced colorectal neoplasm. Third, we have adjusted for the potential confounders which may affect the development of colorectal adenomatous polyps, such as age, sex, and metabolic risk factors. Thus, we could better describe the relationship between the histological severity of NAFLD and the risk of advanced colorectal neoplasm.

The limitation of this study is that the detection of colorectal polyps is prone to intra- as well as inter-observer variation. Despite our experienced examiners who performed colonoscopy, intra- and inter-observer variations could affect the detection of colorectal polyps. Although our initial cohort included more than 750 biopsy-proven NAFLD patients, only 476 patients underwent colonoscopy, which might produce a selection bias in our study. To overcome this limitation, we are currently building another multicenter, prospective cohort in Korea and will soon perform an extended study for external validation. Second, given the known association between diabetes mellitus and colorectal adenomas,[35,36] the presence of diabetes mellitus might act as a confounding factor. To minimize the confounding effect, we have performed multivariate analyses for developing of adenomatous colorectal polyp or advanced colorectal neoplasm including diabetes mellitus as a covariate. Since this study is a cross-sectional study, proving whether NAFLD by itself is a predictor of colorectal neoplasm may be difficult. The mechanism linking NAFLD to colorectal neoplasm is not yet completely understood. NAFLD represents a condition of insulin resistance and pro-inflammatory state. Insulin or insulin-like growth factors may promote the development of CRC. In the present study, we have adjusted for the presence of diabetes mellitus to prove that the presence of NASH is an independent risk factor for developing advanced colorectal neoplasm. Given the high prevalence of NAFLD, performing a screening colonoscopy in all patients with NAFLD may not be feasible due to limited resources. However, for the patients with biopsy-proven NASH, the need for screening colonoscopy is more compelling due to a higher risk of advanced colorectal neoplasm.

In conclusion, NASH may be an independent risk factor for advanced colorectal neoplasm. Understanding the sequential progression from colorectal adenoma to CRC according to the histological spectrum of NAFLD and recommendations to perform screening colonoscopy in patients with NASH are important and useful messages for clinicians. Therefore, further studies are needed to better understand the pathophysiology of NAFLD associated with advanced colorectal neoplasm, the benefit of early screening of advanced colorectal neoplasm in NASH patients, the effect of genetic traits on the development of NASH and advanced colorectal neoplasm, and the impact of NAFLD treatment on the modulation of the risk of advanced colorectal neoplasm.