Metabolic factors associated with colorectal neoplasms: a single center retrospective cross-sectional analysis of colonoscopy patients.

Background
Colorectal cancer (CRCs) is the leading cause of cancer-related deaths worldwide and the second and third most commonly diagnosed cancers in women and men, respectively [1]. The main strategy against CRCs is early detection using colonoscopy and the treatment of precancerous lesions or cancer at an early stage [2–6]. Thus, risk stratification for colonoscopy surveillance is crucial for the prevention of colorectal cancer.
Several previous studies have identified the risk factors for colorectal neoplasms, most of which are associated with genetic and metabolic factors [7–13]. Regarding the latter, obesity, type 2 diabetes mellitus (DM), or non-alcoholic fatty liver disease (NAFLD) are the risk factors for CRCs according to the previous reports [7–10]. Moreover, in our previous study, a wide waist circumference, hypertension, and DM were associated with the incidence of advanced colorectal neoplasms, especially late-onset neoplasms, and these factors were more associated with adenoma rather than sessile serrated lesions [11, 12]. However, these factors have rarely been elucidated using detailed information, such as laboratory data, abdominal ultrasonography (AUS), fibro-scan data, or abdominal computed tomography (CT) scan data.
To address these issues, a single-center retrospective study was conducted to identify the risk factors for colorectal neoplasms, particularly metabolic factors.

Methods

Study design, setting, and patients
All patient data were retrospectively extracted from the endoscopic databases of the Institute for Adult Diseases of the Asahi Life Foundation.
This retrospective study used the opt-out method. This study was approved by the Institutional Review Board of the Institute for Adult Diseases, Asahi Life Foundation (Registration no. 14901, date of approval: May 10, 2023), and conformed to the provisions of the Declaration of Helsinki (as revised in Fortaleza, Brazil, October 2013).
This study was designed and reported in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines to ensure transparency and completeness in reporting observational research.

Inclusion and exclusion criteria
Patients who underwent colonoscopy, laboratory data analysis, or abdominal ultrasonography between 2001 and 2021 were evaluated. The majority of patients undergo treatment for DM, hypertension, and dyslipidemia in the hospital. Among them, colonoscopies are conducted for those who test positive for fecal occult blood or exhibit abdominal symptoms. Patients who underwent repeat colonoscopy during the study period, those with missing data, those with a history of CRCs prior to the study period, those with inflammatory bowel disease, or those with polyposis disease were excluded. Finally, a total of 1,916 patients were included in the analysis (Fig. 1).

Variables and outcomes
The following clinical factors were evaluated as variables partly related to the risk of developing colorectal neoplasms: age, sex, body mass index, waist circumference (≥ 85 or < 85 cm), family history of CRC within the second degree of consanguinity, laboratory data at the first medical check-up reflecting the treatment naïve state (aspartate aminotransferase, alanine aminotransferase [ALT], alkaline phosphatase [ALP], ɤGTP, albumin, platelet [PLT], total-bilirubin, creatinine [Cre], low-density lipoprotein cholesterol, How-density lipoprotein cholesterol [HDL], triglyceride, HbA1c, and uric acid [UA]), fatty liver scoring (Fib4-index, NAFLD fibrosis score, AST/ALT ratio [AAR], aspartate aminotransferase to platelet ratio index [APRI]), and BARD score calculated in line with the previous studies [14]), the findings of AUS (fatty liver, gallbladder polyp, and aortic calcification), comorbidities (prescribed medications for hypertension, DM, dyslipidemia, and hyperuricemia), medication use (aspirin, metformin, febuxostat, and statin), fibroscan data, and fat CT data (visceral fat area, subcutaneous fat area, total fat area, and visceral fat index).
The primary outcome measure was the incidence of endoscopically diagnosed and pathologically confirmed colorectal neoplasms. Colorectal neoplasms include adenomas, sessile serrated lesions, and adenocarcinomas including intramucosal carcinomas. In addition, the factors associated with right and left colorectal neoplasms were analyzed separately. The right neoplasms included those in the cecum, ascending colon, and transverse colon, whereas the remaining neoplasms included those in the descending colon, sigmoid colon, and rectum.

Statistical analysis
Continuous variables were expressed as means with 95% standard deviation, whereas categorical variables were expressed as numbers and frequencies (%). The ORs and 95% confidence intervals (CIs) for colorectal neoplasms were calculated using multivariate logistic regression models adjusted for age > 60 years and male sex for clinical factors. A subgroup analysis was performed for patients with fatty liver diagnosed using AUS, fibro-scan, or abdominal CT. A P–value of < 0.05 was considered significant. All statistical analyses were performed using SAS software (ver. 9.4; SAS Institute, Cary, NC, USA).

Results

Patient characteristics
The data from 1,916 patients, including 781 patients with neoplasms and 54 patients with adenocarcinomas, were analyzed. Patient characteristics are presented in Table 1. Male sex, age, waist circumference, laboratory data, including albumin, Cre, HDL, and HbA1c, NAFLD fibrosis score, and AUS findings of fatty liver were significantly associated with the incidence of any neoplasm. Age, HDL laboratory data, and NAFLD fibrosis scores were significantly associated with the incidence of adenocarcinomas.

Factors associated with right and left neoplasms
Factors associated with right and left neoplasms are shown in Table 2. Among the 1,916 patients, 513 and 560 patients developed right and left neoplasms, respectively. Male sex, age, waist circumference, and laboratory data, including PLT, HDL, and HbA1c levels, NAFLD fibrosis score, and AUS findings of fatty liver were significantly associated with the incidence of right neoplasms, whereas male sex, age, waist circumference, and laboratory data of HDL, and AUS findings of fatty liver were significantly associated with the incidence of left neoplasms.

Multivariate analysis of risk factors for colorectal neoplasms
The results of the multivariate analysis for the risk factors of any, right, and left colorectal neoplasms are shown in Fig. 2.

Male sex (aOR:1.54), age > 60 years (aOR:1.29), waist > 85 cm (aOR:1.28), and fatty liver (aOR:1.31) were significantly associated with higher odds of developing any neoplasms. Male sex (aOR:1.41), age > 60 years (aOR:1.35), waist > 85 cm (aOR:1.49), and fatty liver (aOR:1.40) were significantly associated with higher odds of developing right neoplasms, whereas male sex (aOR:1.74) and fatty liver (aOR:1.27 were significantly associated with a higher odds of developing any neoplasms.

Factors associated with colorectal neoplasms in patients with fatty liver
Factors associated with colorectal neoplasms in patients with fatty liver disease are presented in Table 3. Among 871 patients, 383 patients developed neoplasms and 28 patients developed adenocarcinomas. Male sex, age, and laboratory data, including albumin and UA were significantly associated with the incidence of any neoplasm, whereas age were significantly associated with the incidence of adenocarcinomas.

Factors associated with colorectal neoplasms regarding fibro-scan data and fat-CT scan data
The factors associated with the development of colorectal neoplasms in patients with fibro-scan and fat-CT data are shown in Table 4. Among the 559 patients with fibro-scan data, 235 and 15 patients developed neoplasms and adenocarcinoma, respectively. Neither VCTE nor FibroScan scores were significantly associated with the incidence of neoplasms or adenocarcinomas. Among 869 patients with abdominal CT scan data, 342 patients developed neoplasms and 27 patients developed adenocarcinomas. The visceral fat area and visceral fat index were significantly associated with the incidence of any neoplasm, whereas subcutaneous fat and total fat area were not.

Discussion
This retrospective study found that male sex, old age, wide waist circumference, fatty liver, and serum levels of several metabolic indicators, including, Albumin, Cre, HDL, and HbA1c, and NAFLD fibrosis score were associated with a higher risk of neoplasms in the colorectum, whereas old age, serum levels of HDL, and NAFLD fibrosis score were associated with a higher risk of colorectal adenocarcinoma. The results for all neoplasms were similar to those for right and left neoplasms. Moreover, in patients with fatty liver, male sex, old age, and serum levels of albumin and UA were associated with the incidence of any neoplasm in the colorectum, whereas old age was similar to those of colorectal adenocarcinoma. Abdominal CT scan data showed that the visceral fat area and visceral fat index were significantly associated with the incidence of any neoplasm, whereas subcutaneous fat and total fat areas were not.
Consistent with the findings of several previous studies [9, 15], fatty liver is an independent risk factor for colorectal neoplasms. Moreover, the severity of liver fibrosis is associated with the prevalence of colorectal neoplasms, although the fibro-scan results did not attain statistical significance due to the limited number of patients. According to previous studies, fatty liver affects colorectal carcinogenesis through enhanced insulin resistance, chronic inflammation with various signal pathways, such as IL-6 and TNFα, adipocytokines, or alteration of gut microbiota [16–18]. Otherwise, fatty liver may simply reflect the gained visceral fat, which has been shown to be a risk factor for CRCs [19]. Further studies are required to elucidate the relationship between fatty liver and the development of colorectal neoplasms.
Regarding the differences between right and left neoplasms, right neoplasms are more associated with the serum level of HbA1c and fatty liver, whereas left neoplasms are associated with male sex in particular. These results may reflect the different pathogeneses of right and left neoplasms, such as different precancerous lesions (adenoma vs. sessile serrated lesion), different gut microbiota, or different developmental origins [13, 20, 21].
In the current study, waist circumstance was more strongly associated with colorectal neoplasms than BMI, consistent with our previous study [11]. This might be due to the fact that Japanese individuals are likely to gain more visceral fat than subcutaneous fat compared with Westerners [19, 22] and visceral fat was also associated with the incidence of colorectal neoplasms, compared with subcutaneous fat according to the result of the abdominal CT scan analysis. A potential mechanism involves the induction of a protumorigenic status by proinflammatory cytokines and adipokines secreted by visceral adipocytes. Chronic inflammation fosters carcinogenesis through various mechanisms, such as promoting cancer cell proliferation and angiogenesis [23, 24]. The waist is reported to reflect visceral fat rather than subcutaneous fat [25]; thus, Japanese patients with longer waists may have a higher risk of colorectal neoplasms, even when their BMI is relatively low.
A high serum level of uric acid was associated with the incidence of any neoplasm in the fatty liver subgroup, which is consistent with previous reports [26–28]. Therefore, uric acid may directly affect colorectal carcinogenesis. According to previous studies, uric acid, as a pro-oxidant, plays a partial role in tumorigenesis by entering normal cells and enhancing tumor cell proliferation, migration, and survival [29]. Indeed, the prevention of colorectal cancers has been reported through the treatment of hyperuricemia [30], although the current study did not find an association between the use of febuxostat and the incidence of colorectal neoplasms. Further studies are required to fully understand the mechanism of action for colorectal carcinogenesis.
In terms of dyslipidemia, a low serum level of HDL was more associated with both neoplasms and adenocarcinoma than high serum levels of LDL or TG. This might be because the high LDL or TG level might be treated with statins or other lipid-lowering agents, which are chemopreventive against CRCs [1, 13]. Otherwise, reduced serum HDL levels might be associated with aberrantly expressed cholesterol metabolism genes in colorectal cancer [31].
In this study, comorbidities and medication use, which have been reported to affect colorectal carcinogenesis, were not associated with the prevalence of colorectal neoplasms. This discrepancy might be due to the fact that most patients at the current hospital tended to have metabolic diseases, as many as 60% of patients treated with DM medications, and the effect of comorbidities or drug use might be underestimated.
Several previous studies have shown an association between a family history of CRCs and the incidence of colorectal tumor development [5–8, 11]. However, family history was not a risk factor for neoplasms or adenocarcinoma of the colorectum in our study. One reasonable explanation might be that our cohort included many elderly patients with several metabolic factors; additionally, the impact of congenital factors might have been underestimated, in line with our previous study [11].
Our study has several strengths. We evaluated the association between colorectal neoplasms and various clinical factors, including waist circumference, body mass index (BMI), and laboratory data, which are generally difficult to obtain from endoscopic databases.

Limitations of the study
Our study has several limitations. First, it was a retrospective study. Second, information on the indications for colonoscopy and the risk factors including genetic profiles and lifestyle, such as alcohol consumption and smoking, was not obtained from the database. Third, the medical information at the other hospitals was limited. Fourth, the generalizability of our results may be limited, as this was a single-center study conducted in a population with a high prevalence of lifestyle-related diseases. Finally, the current study lacks detailed histopathological information on colorectal neoplasms, such as sessile serrated lesions, villous, or tubulovillous adenomas.

Conclusion
Male sex, old age, wide waist circumference, fatty liver, and visceral fat are risk factors for colorectal neoplasms. Colonoscopy is recommended for elderly men, particularly those with these metabolic factors.

Supplementary Information