Retrospective analysis of risk factors for left-sided vs right-sided colon cancer in a southern Chinese Han population

Abstract

BACKGROUND

Colorectal cancer (CRC) is the most prevalent malignancy of the digestive system. Elucidating its etiological characteristics, risk factors, and clinical heterogeneity is pivotal to optimizing screening strategies and implementing personalized prevention. While molecular and clinical disparities between left-sided CRC (L-CRC) and right-sided CRC (R-CRC) are well-recognized, systematic investigations into their distinct risk factor profiles remain limited. The literature predominantly relies on western cohorts, with a paucity of large-scale studies on Chinese populations. Therefore, this study aimed to contribute to the insufficient data regarding risk factors for L-CRC and R-CRC among the Han Chinese population in southern China.

AIM

To identify independent risk factors distinguishing L-CRC from R-CRC and characterize the distribution of clinical tumor stages between two subtypes.

METHODS

We conducted a retrospective analysis of colon cancer cases diagnosed over five years at Sun Yat-sen Memorial Hospital. The cohort included Han Chinese patients, native to Guangdong Province, aged 40-65 years, with confirmed colon cancer. Patients were stratified into L-CRC, R-CRC and control groups. Demographic, metabolic, and lifestyle variables were collected. Statistical analyses were performed using Welch’s t-test, χ² test, Mann-Whitney U test, and multivariate logistic regression.

RESULTS

Our analysis revealed significant disparities between L-CRC and R-CRC regarding demographic, metabolic, and lifestyle-related risk profiles. L-CRC was independently associated with elevated body mass index, high low-density lipoprotein levels, hepatitis B virus infection, a family history of CRC or polyps, and smoking. In contrast, R-CRC showed independent associations with male sex, hypertension, and elevated triglyceride levels. Furthermore, R-CRC cases were significantly more likely to present at TNM stage IV than L-CRC cases.

CONCLUSION

Our findings underscore the importance of incorporating tumor location into risk assessment models and provide evidence characterizing the differential clinical features of L-CRC and R-CRC within the Han Chinese population.

Key Words: Colorectal cancer; Left colorectal cancer; Right colorectal cancer; Risk factors; Han Chinese; Guangdong Province; Tumor staging; Personalized screening

Core Tip: This is a retrospective study that identifies distinct risk profiles for left-sided and right-sided colorectal cancer within the southern Chinese Han population. The findings reveal that screening strategies in specific location, informed by metabolic and lifestyle factors, are crucial for improving early detection and prognosis, especially in this underrepresented cohort.

INTRODUCTION

Colorectal cancer (CRC) is one of the most prevalent malignant tumor of the digestive system in the world, characterized by considerable morbidity and mortality. Due to the rapid westernization of lifestyles and an accelerating aging population, the incidence of CRC has continuously increased in China. CRC is threatening national well-being, thus becoming a critical public health challenge. Elucidating the clinical heterogeneity of CRC, etiological characteristics and risk factor profiles are imperative to optimize screening strategies and advance precision prevention and treatment.

The colon is anatomically and embryologically divided into distinct segments with divergent origins. Derived from the midgut, the right-sided colon is composed by the cecum, ascending colon, hepatic flexure, and the proximal two-thirds of the transverse colon. Conversely, originating from the hindgut, the left-sided colon comprises the distal one-third of the transverse colon, splenic flexure, descending colon, sigmoid colon, and rectum[1,2]. These anatomical distinctions, have significant implications for prognosis and therapeutic responsiveness, are underpinned by molecular signatures, pathological, and unique clinical[3].

Right-sided CRC (R-CRC) are often found among older women and frequently present with iron-deficiency anemia in clinic[4]. Additionally, they are strongly associated with some molecule, such as microsatellite instability, BRAF mutations, and the CpG island methylator phenotype[5]. In contrast, left-sided CRC (L-CRC) are more common among men, characterized by chromosomal instability, and typically manifest with obstructive symptoms due to luminal narrowing[6]. Notably, L-CRC tumors exhibit higher response rates to anti-EGFR therapies than the right-sided part[7-10].

In the specific risk-factor spectra, despite the widespread recognition of the clinicopathological and molecular dichotomies mentioned above, systematic investigations that distinguish L-CRC from R-CRC remain limited. Literature in being focus on Western cohorts, but largely ignores large-scale data specific to the Chinese population. It is not quite the correct approach to direct extrapolation of Western findings to China, due to profound differences in comorbidity landscapes, dietary patterns, lifestyle habits, and genetic backgrounds.

For the China’s vast geographical expanse, there is significant heterogeneity between the north and south, in diet, climate, and disease epidemiology. For example, a developed region in southern China, Guangdong Province, is characterized by alongside a high prevalence of hepatitis B, a diet rich in seafood and relatively low in fat. This sharply contrasts with the diets of northern people, which are typically more love in red meat and lipids. There is not a specific research focusing on the differential risk factors for L-CRC and R-CRC among the Han Chinese population in southern China in current.

There was a narrow focus on isolated factors in previous studies, often lacking a comprehensive, multidimensional assessment that integrates metabolic parameters, comorbidities, demographics, lifestyle, and family history. Few studies have rigorously controlled for confounders and evaluated the independent effects of these variables in systematic. So the site-specific associations of certain risk factors with CRC, such as lipid profiles, history of infection, medical and family history, and lifestyle factors, remain inadequately elucidated. Although associations between these factors and overall CRC risk have been reported[11-15], it remains ambiguous in whether their effects differ across tumor subsites and what mechanisms underlie these differences.

To address these gaps, this study leveraged existing data from a large tertiary care hospital in southern China, to run the first large-scale risk-factor comparison in a southern Chinese Han cohort. We collected the data of 2009 patients with pathologically confirmed CRC to systematically characterize differences between L-CRC (n = 1132) and R-CRC (n = 877) across domains, including demographics, lifestyle factors, metabolic indices, comorbidities, family history, and clinical staging. Using univariate and multivariate logistic regression analyses, we aimed to identify independent risk factors specific to each subsite and comparatively analyze their respective risk profiles. We also examined the distribution of clinical stages across the two groups to provide robust epidemiological evidence and thereby refine region-specific prevention and early detection strategies.

MATERIALS AND METHODS

Study design and participants

This retrospective study was conducted at Sun Yat-sen Memorial Hospital, Sun Yat-sen University, and included patients diagnosed with CRC over six years (from January 2020 to December 2025). All statistical analyses were performed under the supervision of a qualified biomedical statistician. A total of 2580 medical records were initially screened on March 1, 2025; 2009 met the eligibility criteria and were included in the final analysis. Among these, 1132 patients’ diagnoses were classified as L-CRC and 877 as R-CRC. A further 600 non-cancer patients were randomly selected as the control group. The detailed case selection process is illustrated in Figure 1.

Figure 1 Study flowchart for the comparative analysis of left-sided colorectal cancer and right-sided colorectal cancer. The study began with the collection of data from 2580 colorectal cancer (CRC) patients, of whom 2009 met the inclusion or exclusion criteria and were included in the analysis [1132 left-sided CRC (L-CRC), 877 right-sided CRC (R-CRC) and 600 control]. Clinical data, including demographics, laboratory tests [triglyceride (TG), low-density lipoprotein (LDL)], medical and family history, and lifestyle factors (smoking, alcohol), were collected. Logistic regression analysis was conducted to identify major factors associated with L-CRC vs R-CRC, using age, TG, LDL, past history, family history, smoking, alcohol, and tumor stage as input variables. The analysis included both univariable and multivariable models to estimate adjusted ORs and 95%CIs. Comparative analysis was performed using t-tests for continuous variables (age, TG, LDL), χ² tests for binary variables (past medical history, family history, smoking, alcohol), and Mann-Whitney U tests for ordinal variables (tumor stage). L-CRC: Left-sided colorectal cancer; R-CRC: Right-sided colorectal cancer; BMI: Body mass index.

Inclusion criteria: (1) Age between 40 years and 65 years; (2) Histopathologically confirmed diagnosis of colon cancer; (3) Of Guangdong Province origin; and (4) Han Chinese in ethnicity.

Exclusion criteria: (1) Neuroendocrine tumors of the colon; (2) Colon invasion by tumors originating from adjacent organs (e.g., liver, ovary, pancreas, and cervix); and (3) Presence of more than three comorbidities.

The control group’s selection criteria: (1) Age between 40 years and 65 years; (2) All control subjects were rigorously confirmed to be free of any neoplastic disease; (3) Of Guangdong Province origin; and (4) Han Chinese in ethnicity.

Data collection

Clinical and laboratory data were extracted from electronic medical records. The following variables were collected for each patient: Sex, age, smoking history, alcohol consumption history, history of diabetes mellitus (DM), hepatitis B virus (HBV) infection, hypertension, family history of colon cancer, family history of colonic polyps, tumor stage (I-IV), body mass index (BMI; kg/m²), triglycerides (TGs; mmol/L), low-density lipoprotein (LDL; mg/dL).

Missing data handling

In some patient records, there were missing data for certain variables, including BMI (1.2%), TG (3.2%), LDL (4.1%), and smoking history (2.3%). The proportion of missing data for all variables was less than 5%. Given the relatively low missing rate, a complete case analysis was performed, excluding patients with any missing values for variables included in the multivariable model. To assess the robustness of the results, sensitivity analyses were conducted using multiple imputations by chained equations with 20 imputed datasets, incorporating all variables in the imputation model. The results from the complete case analysis and multiple imputation were consistent, indicating that missing data did not substantially bias the findings.

Ethical approval

This study was conducted in accordance with the principles of the Declaration of Helsinki. The study was reviewed and approved by the Institutional Review Board Institutional Review Board Committee at Sun Yat-Sen Memorial Hospital, Sun Yat-Sen University (Approved No. SYSKY-2026-377-01). Owing to the study’s retrospective nature, the review board waived the requirement for informed consent. Patient data were anonymized and de-identified before analysis to ensure confidentiality.

Statistical analyses

For continuous variables (age, BMI), Welch’s t-test was applied to account for unequal variances between the two groups. Results are presented as mean ± SD. For binary categorical variables (sex, history of DM, HBV infection, hypertension, smoking, alcohol consumption, hypertriglyceridemia (TG ≥ 2.3 mmol/L)[16,17], High LDL cholesterol (LDL ≥ 190 mg/dL)[18], family history of colon cancer, and family history of colonic polyps), Pearson’s χ² test was used. The n (%) are reported. For the ordinal tumor stage variable (stages I-IV), the Mann-Whitney U test was used to assess differences in distribution between the L-CRC and R-CRC groups. To identify independent factors associated with tumor location, a binary logistic regression analysis was performed.

To ensure the inclusion of potentially important confounding factors, possible variables were selected based on two criteria: (1) Variables that showed a significant association with tumor location in univariate analysis at a threshold of P < 0.05; and (2) Variables considered clinically relevant, regardless of statistical significance, including age, sex, and family history. A liberal threshold (P < 0.05) was used in univariate screening to avoid omitting potentially important confounders.

All candidate variables were entered simultaneously using the enter method (forced entry) to ensure full adjustment for potential confounders. Stepwise selection methods were not employed, to avoid overfitting and maintain interpretability, given the relatively large sample size (n = 2009) and the number of candidate variables (n = 12).

Before model fitting, we assessed the collinearity among continuous variables (BMI, age), using the variance inflation factor. All variance inflation factor values were below 2.5, indicating no substantial multicollinearity. The goodness-of-fit of the logistic regression models was assessed using the Hosmer-Lemeshow test, with a P value > 0.05 indicating adequate model fit. Model discrimination was evaluated using the area under the receiver operating characteristic curve, with 95%CIs calculated by bootstrap resampling (1000 replicates).

Results are presented as adjusted odds ratios with 95%CIs. A two-sided P value < 0.05 was considered statistically significant.

All statistical analyses were conducted using SPSS (version 26.0; IBM Corp., Armonk, NY, United States).

RESULTS

A total of 2580 cases were initially collected from January 2020 to December 2025, and 2009 patients were ultimately included in the final analysis (L-CRC group, n = 1132; R-CRC group, n = 877). Comparative analysis revealed that there were no significant differences between the two groups regarding hypertension, alcohol consumption, or family history of colon polyps. However, statistically significant differences were observed in age, sex, smoking status, family history of CRC, HBV status, DM, body weight, blood glucose, TG, and LDL levels (Table 1; all P < 0.0001). Furthermore, distinct different distributions in tumor staging were observed between two groups.

Table 1 Clinical data, statistics, and comparison results of left-sided colorectal cancer and right-sided colorectal cancer groups, n (%) or mean ± SD.

	L-CRC group, n = 1132	R-CRC group, n = 877	P value
Age (year)	54.13 ± 7.5	55.1 ± 6.7	0.0023
Male sex	832 (73.5)	468 (53.4)	< 0.0001
Smoking	786 (69.3)	128 (14.6)	< 0.0001
Alcohol drinking	236 (20.8)	184 (21)	0.9422
Hypertension	905 (79.9)	657 (74.9)	0.0072
Diabetic mellitus	95 (8.4)	41 (4.7)	0.0016
Hepatitis B virus infection	214 (18.9)	104 (11.9)	< 0.0001
Family history of CRC	80 (6.8)	13 (1.5)	< 0.0001
Family history of colorectal polyps	786 (69.4)	263 (30)	< 0.0001
Tumor stage			< 0.0001
I	60 (5.3)	48 (5.5)
II	90 (8.0)	28 (3.2)
III	678 (59.9)	290 (33.1)
IV	304 (26.9)	511 (58.3)
BMI (kg/m2)	26.72 ± 2.2	23.97 ± 1.9	< 0.0001
TG (mmol/L)	2.6 ± 1.3	3.1 ± 1.5	< 0.0001
LDL (mmol/L)	3.1 ± 1.0	2.7 ± 1.0	< 0.0001

L-CRC: Left-sided colorectal cancer; R-CRC: Right-sided colorectal cancer; CRC: Colorectal cancer; BMI: Body mass index; TG: Triglyceride; LDL: Low-density lipoprotein.

Demographic characteristics and genetic background

Patients in the L-CRC group were mildly younger than those in the R-CRC group (mean age: 54.13 ± 7.5 years vs 55.1 ± 6.7 years; P = 0.0023). A markable male predominance was observed in the L-CRC group (73.5%), whereas the R-CRC group exhibited a more balanced sex distribution (53.4% male; Figure 2A-C). Regarding genetic predisposition, a family history of CRC was significantly more prevalent in the L-CRC group (6.8%) compared with the R-CRC group (1.5%; P < 0.0001; Figure 2D-G; Table 1).

Figure 2 Comparison of demographic and clinical characteristics between left-sided colorectal cancer and right-sided colorectal cancer groups. A: Violin plot showing the distribution of age (year) in left-sided colorectal cancer (L-CRC) and right-sided colorectal cancer (R-CRC) groups. The light blue violin represents the L-CRC group, and the light purple violin represents the R-CRC group. Data were analyzed using an independent samples t-test; B: Bar plot comparing the distribution of sex between L-CRC and R-CRC groups. The light blue sector indicates female, and light purple sector indicates male; C and D: Pie charts displaying the sex distribution (male vs female) in L-CRC (total = 1132) and R-CRC (total = 877) groups. The light blue sector represents males, and light purple sector represents females; E: Bar plot showing the distribution of family history of polyps in L-CRC and R-CRC groups. The light blue sector indicates Family history of polyps, and light purple sector indicates no family history of polyps; F: Bar plot showing the distribution of family history of colorectal cancer (CRC) in L-CRC and R-CRC groups. The light blue sector indicates family history of CRC, and light purple sector indicates no family history of CRC; G and H: Pie charts showing the proportion of family history of CRC in L-CRC (total = 1132) and R-CRC (total = 877) groups. The light blue sector represents family history of CRC, and light purple sector represents no family history of CRC; I and J: Pie charts showing the proportion of family history of polyps in L-CRC (total = 1132) and R-CRC (total = 877) groups. The light blue sector represents family history of polyps, and light purple sector represents no family history of polyps. All statistical analyses yielded P < 0.001. Data were analyzed using an independent samples t-test for continuous variables (A) and χ² tests for categorical variables (B-G). L-CRC: Left-sided colorectal cancer; R-CRC: Right-sided colorectal cancer; CRC: Colorectal cancer.

Lifestyle habits and metabolic profiles

The prevalence of smoking was significantly higher in the L-CRC group (69.3%) compared with the R-CRC group (14.6%). The proportion of alcohol consumers, however, had no significantly difference between two groups (L-CRC: 20.8% vs R-CRC: 21.0%; P = 0.9422).

In terms of metabolic indices, the L-CRC group demonstrated a significantly higher BMI than the R-CRC group (26.72 ± 2.2 vs 23.97 ± 1.9; P < 0.0001). The prevalence of high LDL cholesterol (LDL > 190 mg/dL) was significantly higher in the L-CRC group (63%) compared with the R-CRC group (52%). Conversely, the proportion of hypertriglyceridemia (TG > 2.3 mmol/L), did not differ significantly between two groups (L-CRC: 60% vs R-CRC: 53.5%; P = 0.5590; Figure 3).

Figure 3 Comparison of metabolic indicators and lifestyle factors between left-sided colorectal cancer and right-sided colorectal cancer groups. A: Comparison of body mass index between the left-sided colorectal cancer (L-CRC) and right-sided colorectal cancer (R-CRC) groups. Data are presented as mean ± SD. Statistical analysis was performed using an independent samples t-test; B-E: Pie charts illustrating the proportions of patients in the L-CRC (n = 1132) and R-CRC (n = 877) groups based on smoking status (smoking vs un-smoking; B), drinking status (drinking vs un-drinking; C), low-density lipoprotein (LDL) levels (LDL ≥ 190 mg/dL vs LDL < 190 mg/dL; D), and hypertriglyceridemia status (hypertriglyceridemia vs no hypertriglyceridemia; E); F-I: Bar charts depicting the absolute number of patients in each group corresponding to the characteristics shown in panels B-E, respectively. Statistical significance for categorical variables (panels B-E) was determined using the χ² test. All statistical tests yielded P < 0.001. L-CRC: Left-sided colorectal cancer; R-CRC: Right-sided colorectal cancer; LDL: Low-density lipoprotein.

Comorbidities and infectious background

The comorbidities burden differed significantly between the two groups. The prevalence of DM was higher in the L-CRC group than in the R-CRC group (8.4% vs 4.7%; P = 0.0016). Similarly, HBV infection was significantly more common in the L-CRC group than in the R-CRC group (18.9% vs 11.9%; P < 0.0001). Hypertension was also more prevalent in the L-CRC group than in the R-CRC group (79.9% vs 74.9%; P = 0.0072; Figure 4).

Figure 4 Comparison of comorbidities between left-sided colorectal cancer and right-sided colorectal cancer groups. A: Pie charts showing the proportion of diabetes mellitus (DM) in left-sided colorectal cancer (L-CRC; total = 1132) and right-sided colorectal cancer (R-CRC; total = 877) groups; B: Bar plot comparing the number of cases with or without DM between the L-CRC and R-CRC groups. Data were analyzed using a χ² test; C: Pie charts showing the proportion of hepatitis B virus (HBV) infection in L-CRC (total = 1132) and R-CRC (total = 877) groups; D: Bar plot comparing the number of cases with or without HBV infection between the L-CRC and R-CRC groups. Data were analyzed using a χ² test; E: Pie charts showing the proportion of hypertension in L-CRC (total = 1132) and R-CRC (total = 877) groups; F: Bar plot comparing the number of cases with or without hypertension between the L-CRC and R-CRC groups. Data were analyzed using a χ² test. All data are presented as percentages or case numbers. P < 0.001 for all comparisons. L-CRC: Left-sided colorectal cancer; R-CRC: Right-sided colorectal cancer; DM: Diabetes mellitus; HBV: Hepatitis B virus.

Tumor staging distribution

The analysis of tumor staging at diagnosis revealed a distinct different distribution pattern: The L-CRC group was predominantly comprised of patients with stage III disease (59.9%), whereas the R-CRC group had more patients presenting with stage IV disease (33.1%; Table 1).

Logistic regression analysis of risk factors

Specifically, higher BMI, smoking, HBV infection, family history of CRC, family history of colon polyps, and high LDL cholesterol were identified as independent risk factors for L-CRC. In contrast, male gender, hypertension, and hypertriglyceridemia were identified as independent risk factors associated with R-CRC (Tables 2, 3 and 4).

Table 2 Independent risk factors of left-sided colorectal cancer and right-sided colorectal cancer.

Factor	L-CRC	R-CRC
Male	-	+
Smoking	+	+
Hypertension	-	+
Elevated TG	-	+
Elevated LDL	+	-
High BMI	+	-
HBV infection	+	-
Family history of CRC	+	-
Family history of polyps	+	-

L-CRC: Left-sided colorectal cancer; R-CRC: Right-sided colorectal cancer; CRC: Colorectal cancer; BMI: Body mass index; TG: Triglyceride; LDL: Low-density lipoprotein; HBV: Hepatitis B virus.

Table 3 The ORs of influencing factors between left-sided colorectal cancer and control, analyzed by logistic regression.

ORs	Variable	Estimate	95%CI (profile likelihood)	|Z|	P value
β1	Age	1.199	1.100-1.310	4.073	< 0.0001
β2	BMI	1.325	0.7148-2.481	0.8872	0.3750
β3	Male[1]	1.458	0.7772-2.771	1.165	0.2440
β4	Smoking[1]	2.034	0.6574-6.688	1.204	0.2286
β5	Drinking[1]	2.234	1.028-4.907	2.020	0.0434
β6	DM[1]	49.20	26.61-95.61	11.97	< 0.0001
β7	HBV[1]	1.095	0.6693-1.790	0.3630	0.7166
β8	Hypertension[1]	0.09002	0.03404-0.2213	5.052	< 0.0001
β9	Family history of CRC[1]	1.304	0.9099-1.871	1.445	0.1484
β10	Family history of polyps[1]	0.03006	0.01040-0.08200	6.661	< 0.0001
β11	TG	0.3745	0.1559-0.9105	2.185	0.0289
β12	LDL	1.227	1.195-1.262	14.61	< 0.0001

CRC: Colorectal cancer; BMI: Body mass index; TG: Triglyceride; LDL: Low-density lipoprotein; HBV: Hepatitis B virus; DM: Diabetes mellitus.

Table 4 The ORs of influencing factors between right-sided colorectal cancer and control, analyzed by logistic regression.

ORs	Variable	Estimate	95%CI (profile likelihood)	|Z|	P value
β1	Age	0.5913	0.5142-0.6738	7.640	< 0.0001
β2	BMI	1.891	0.8374-4.438	1.505	0.1324
β3	Male[1]	3.672	1.708-8.173	3.268	0.0011
β4	Smoking[1]	0.1550	0.02957-0.7704	2.266	0.0234
β5	Drinking[1]	60.59	22.44-197.6	7.475	< 0.0001
β6	DM[1]	9.927e-005	0.00002135-0.0003921	12.46	< 0.0001
β7	HBV[1]	0.1010	0.04727-0.2061	6.123	< 0.0001
β8	Hypertension[1]	518.7	94.94-3370	6.918	< 0.0001
β9	CRC family history[1]	1.706	0.9574-3.083	1.795	0.0726
β10	Polyp family history[1]	0.0001669	0.00002886-0.0008014	10.30	< 0.0001
β11	TG	0.0001938	0.00002737-0.001135	9.045	< 0.0001
β12	LDL	1.216	1.172-1.265	10.13	< 0.0001

CRC: Colorectal cancer; BMI: Body mass index; TG: Triglyceride; LDL: Low-density lipoprotein; HBV: Hepatitis B virus; DM: Diabetes mellitus.

DISCUSSION

This study aimed to identify independent risk factors among the Han Chinese population of Guangdong Province, differentiating L-CRC from R-CRC. The further aim is to inform evidence based preventive measures by characterize the distribution of clinical tumor stages between left and R-CRC. The retrospective large case design employed in this study represents a robust approach for clarifying the distinct etiology of left and R-CRC. The substantial sample size (n = 2009) and noncancer control group in this study not only identified the specific positioning risk factors, but also enable strict control of confounding variables with professional biostatistical oversight. This approach provides highly reliable evidence for distinguishing the pathogenesis of right and L-CRC within the southern Chinese population.

This analysis also revealed significant disparities in clinical presentation and laboratory profiles between left and R-CRC.

Our results indicated that the age of L-CRC was significantly younger (mean age: 54.13 ± 7.5 years) and the sex of L-CRC was male in predominant (73.5%), whereas the R-CRC showed a more average sex distribution (65.2% female). This distribution may be driven by variations in gut microbiota, sex hormone levels, and sex exposure to carcinogens[19-21].

A family history of CRC was more prevalent in the L-CRC group, although no difference was observed for family history of polyps. It can be inferred that genetic susceptibility plays a more hinge role in the pathogenesis of L-CRC. While hereditary syndromes, such as Lynch syndrome, typically manifest in the right colon, the stronger familial aggregation observed in scattered L-CRC implies the potential involvement of unidentified low-penetrance susceptibility variants or shared environmental exposures specific to the left colon[22].

Regarding lifestyle, smoking was significantly more prevalent in the L-CRC, whereas alcohol consumption showed no differences in the univariate analysis. Smoking is a perfect risk factor for CRC. However, its pronounced effect on the left colon may be attributed to the direct action of carcinogens metabolized in the bowel on the distal mucosa[23-25].

The L-CRC presented a unique profile in metabolic, includes higher basic mass index and low density lipoprotein levels, but lower triglyceride levels. Despite a higher prevalence of DM in the L-CRC, the lipid spectrum displayed an atypical pattern of high LDL and low TG. This phenotype suggests that the L-CRC pathogenesis is more closely linked to specific branches of insulin resistance and lipid metabolism dysregulation, than to classic hyperlipidemia[26].

A notable finding was the significantly higher proportion of HBV infection in the L-CRC group. While hepatic V virus is primarily associated with hepatocellular carcinoma, emerging evidence suggests that it influences CRC development through immune modulation or alterations in the gut microbiome[27,28]. The higher HBV detection rate in the L-CRC group may reflect differences in past infection history or immune background; this warrants further investigation into viral load and liver function.

Additionally, the prevalence of DM was significantly higher in the L-CRC group. DM is an independent risk factor for CRC. Its stronger association with L-CRC in the study supports the hypothesis that hyperinsulinemia may preferentially promote proliferation in left colonic epithelium, potentially due to receptor specific in site expression profiles[29,30].

The distinct lipid profile of L-CRC (elevated LDL and reduced TG) has potential diagnostic value. Previous studies indicate that tumor cells reprogram lipid metabolism in a specific site manner. L-CRC may rely more heavily on LDL as a cholesterol source for membrane synthesis, whereas R-CRC may depend on endogenous TG metabolism. Furthermore, the observed differences in glucose levels corroborate the distinct metabolic backgrounds of the two tumor subsites[31,32].

Multivariate logistic regression, adjusted for confounders, identified independent risk factors for each subsite (Tables 3, 4 and 5). Basic mass index, alcohol consumption, HBV infection, family history of CRC, family history of colon polyps, and LDL were identified as independent risk factors for L-CRC. Although the univariate analysis showed no difference in alcohol consumption, multivariate adjustment revealed it as an independent risk factor for L-CRC. This suggests that alcohol acts synergistically with metabolic or genetic factors. Acetaldehyde, a metabolite of alcohol, is locally mutagenic and may reach higher concentrations in the distal colon. The sex of male was also confirmed as an independent factor, consistent with the male advantage epidemiology of L-CRC. Basic mass index was confirmed as an independent risk factor for L-CRC. While some studies link obesity more strongly to R-CRC, our data suggest that in this population, high basic mass index significantly elevated L-CRC risk. Elevated LDL was also an independent risk factor, whereas TG did not show an independent effect. This implies that lipid management, particularly LDL control, may be crucial for preventing L-CRC. HBV infection was confirmed as an independent risk factor for L-CRC. This is a novel finding at the multivariate level, suggesting that HBV history should be considered when identifying the individuals in high risk for L-CRC. Both CRC and polyps family history were independent risk factors, reinforcing the strong genetic susceptibility of the left colon.

Table 5 ORs of influencing factors between left-sided colorectal cancer and right-sided colorectal cancer analyzed by logistic regression.

Odds ratios	Variable	Estimate	95%CI (profile likelihood)	|Z|	P value
β1	Age	0.9912	0.9707-1.012	0.8289	0.4072
β2	BMI	1.937	1.763-2.142	13.33	< 0.0001
β3	Male[1]	0.06851	0.01992-0.1776	4.919	< 0.0001
β4	Smoking[1]	136.3	52.46-470.5	8.988	< 0.0001
β5	Drinking[1]	0.3043	0.2071-0.4441	6.118	< 0.0001
β6	DM[1]	0.08025	0.03281-0.1930	5.586	< 0.0001
β7	HBV[1]	2.281	1.284-4.098	2.788	0.0053
β8	Hypertension[1]	0.2831	0.1645-0.4846	4.581	< 0.0001
β9	CRC family history[1]	1.277	0.4858-3.544	0.4839	0.6284
β10	Polyp family history[1]	1.837	1.199-2.819	2.789	0.0053
β11	TG	0.6445	0.5608-0.7374	6.292	< 0.0001
β12	LDL	2.078	1.709-2.538	7.257	< 0.0001
β13	Stage[3]	0.6550	0.3226-1.293	1.196	0.2318
β14	Stage[4]	0.1322	0.06448-0.2621	5.663	< 0.0001
β15	Stage[1]	0.3928	0.1562-0.9749	2.003	0.0452

CRC: Colorectal cancer; BMI: Body mass index; TG: Triglyceride; LDL: Low-density lipoprotein; HBV: Hepatitis B virus; DM: Diabetes mellitus.

Men, hypertension, and elevated TG were identified as independent risk factors for R-CRC. The baseline comparison (Table 1) was unadjusted. However, the multivariate model was adjusted for age, basic mass index, sex, and other factors, revealing a significant independent effect for R-CRC (Tables 4 and 5). Hypertension was uniquely associated with R-CRC. This may be mediated by chronic inflammation, endothelial dysfunction, or hemodynamic sensitivity of the right colon supplied by the superior mesenteric artery. Elevated TG was an independent risk factor for R-CRC, contrasting with the LDL association in L-CRC. It is inferred that R-CRC may be associated with endogenous TG metabolism. While R-CRC had a higher proportion of women in the univariate analysis, men emerged as a risk factor in multivariate analysis. This indicates that the women predominance is confounded by age or other metabolic factors, and that men are intrinsically at higher risk when other variables are controlled.

Our analysis showed that R-CRC was more likely to present at stage IV (Table 5). This phenomenon is widely reported and can be attributed to the following factors. The wider lumen and liquid content of the right colon can lead to nonspecific symptoms like anemia and fatigue, rather than obstruction, delaying diagnosis.

Despite the clear contributions, several limitations must be considered. First, this was a single center retrospective study conducted in Guangdong Province, southern China. The specific dietary habits and high HBV endemicity of this region may limit the generalizability of the findings to northern China or other ethnic groups. Second, the study population was exclusively Han Chinese. Third, we lacked molecular data, which prevented us from assessing correlations between risk factors and molecular subtypes. Forth, the restriction to patients aged 40-65 years was intended to minimize confounding from extreme age and align with Chinese screening guidelines. However, this limits the generalizability of our findings to the older population (> 65 years). Fifth, by excluding patients with complex multimorbidity, we aimed to reduce heterogeneity; however, this limits the generalizability of our findings to populations with multiple comorbid conditions. Sixth, the retrospective design may lead to selection bias and residual confounding. Despite our best efforts to adjust for known covariates, unmeasured confounding factors, such as dietary details or antibiotic usage history, may still affect the results. The lack of systematic records of CRC screening history prevents us from evaluating the potential interference of previous colonoscopy and tumor marker screenings on diagnoses. Although studies suggest an association between inflammatory bowel disease (IBD) and colon cancer[33], we observed no malignant transformation among IBD cases collected over the same period. This may reflect improved disease management or insufficient follow-up duration, warranting longer-term observation. The findings of this study should be considered exploratory hypotheses, that need to be validated in future prospective, multicenter cohort studies and functional experiments. Finally, exposure data like smoking and alcohol were qualitative, lacking quantification of dosage and duration.

CONCLUSION

In this large cohort of southern Chinese Han patients, we demonstrated that L-CRC and R-CRC differed significantly in terms of demographic, metabolic, and lifestyle-related risks. L-CRC was independently associated with high BMI, elevated LDL, HBV infection, family history of CRC or polyps, and smoking. R-CRC was independently associated with male sex, hypertension, and elevated TG. Furthermore, R-CRC was more frequently diagnosed at stage IV. This is the first study to report HBV as an independent risk factor specifically for L-CRC in this population. These findings underscore the importance of considering tumor location in risk assessment and suggest that, based on individual risk profiles, such as prioritizing colonoscopy for those with metabolic syndrome or HBV history, risk stratification may be informed but require prospective validation.