Obesity as a risk factor for early-onset colorectal cancer: Evidence from a nationally representative database

Abstract

BACKGROUND

Colorectal cancer (CRC) is the second leading cause of cancer-related deaths worldwide with an alarming rise in early-onset CRC (eoCRC) over the past several decades. Unlike late-onset CRC, the drivers behind eoCRC remain less clear. While certain risk factors such as obesity and smoking have demonstrated a relatively strong association with eoCRC in the literature, some studies have challenged these associations, emphasizing the need for additional studies.

AIM

To investigate the impact of various risk factors on eoCRC with a special focus on obesity.

METHODS

This cross-sectional study used de-identified data from the National Health and Nutrition Examination Survey (1999-2023), including 30321 United States adults aged 18 to 49 years. Participants with missing key variables were excluded. Standardized protocols were used to collect demographic, lifestyle, anthropometric [body mass index (BMI), body roundness index (BRI), waist circumference (WC)], and self-reported CRC data. Logistic regression and propensity score matching assessed associations between obesity-related parameters and eoCRC. Statistical analyses were performed in R and Stata, with P < 0.05 defined as significant.

RESULTS

Of 30321 participants, 48 received a diagnosis of eoCRC. Patients with eoCRC were older (mean age 39.96 years vs 34.36 years; P < 0.001) and had higher WC and BRI. None of the eoCRC patients were heavy drinkers (P = 0.006). Unadjusted models demonstrated significant associations of eoCRC with BRI quartiles, as well as BMI-defined obesity, WC, and smoking. In unadjusted models, BRI remained the strongest independent predictor; those in the highest BRI quartiles had over 10-fold greater odds of eoCRC. In fully adjusted models, BRI remained significant, but BMI- and waist-based obesity were not.

CONCLUSION

BRI is a stronger predictor of eoCRC risk compared to other obesity indices and is a superior tool for identifying young individuals at higher risk of CRC.

Key Words: Early-onset colorectal cancer; Obesity; Body roundness index; Body mass index; Waist circumference; Risk factors; Logistic regression; Propensity score matching

Core Tip: Early-onset colorectal cancer (CRC) is rising globally, yet its modifiable risk factors remain unclear. This cross-sectional study analyzed over 30000 United States adults aged 18-49 using the National Health and Nutrition Examination Survey data (1999-2023) to assess the relationship between obesity and early-onset CRC (eoCRC). Among multiple obesity indices, body roundness index (BRI) emerged as the strongest predictor of eoCRC risk—even after adjusting for body mass index, waist circumference, and lifestyle factors. BRI offers a more holistic assessment of body fat distribution, making it a potentially superior screening tool for young adults at higher risk for CRC.

INTRODUCTION

Colorectal cancer (CRC) is the second leading cause of cancer-related deaths worldwide, and the incidence of early-onset CRC (eoCRC) has been increasing at an alarming rate in the past several decades[1]. eoCRC is defined as CRC that is diagnosed prior to age 50, whereas late-onset CRC (loCRC) is defined as being diagnosed at age 50 or greater. While CRC incidence in older adults has remained relatively stable, the increasing burden among younger individuals has raised concerns and spurred extensive research into possible causal factors. While the primary drivers of loCRC are well described, the underlying causes of eoCRC remain largely unknown.

It is critical to understand the role of modifiable risk factors as prevention strategies through targeted intervention and population-based efforts may stem the rise in eoCRC. However, while certain risk factors, such as obesity and smoking, have shown a relatively consistent association with eoCRC in the literature[2-6], others (e.g., diet and physical activity) have shown conflicting results[7-11]. In fact, some studies refute previously well-established links, highlighting the need for further investigation[4,8,12-14]. Given the inconsistencies in existing evidence, further investigation is warranted to clarify the true impact of these risk factors on eoCRC.

In addition to lifestyle factors, eoCRC risk may also be influenced by demographic and genetic predispositions. Studies show a higher prevalence of eoCRC among male sex, African American, and Hispanic populations than White individuals and in certain regions of the Southern United States[15-17]. Moreover, while external risk factors contribute to sporadic eoCRC cases, their role in genetically predisposed individuals remains unclear. For example, patients with Lynch syndrome, a hereditary condition associated with a high lifetime risk of CRC, do not appear to experience an accelerated onset of CRC due to external risk factors such as obesity or diet[18,19]. This suggests that eoCRC is likely driven by an interplay between genetic susceptibility and environmental exposures, rather than a single causative factor.

Among these factors, obesity has emerged as a particularly compelling risk factor, given its increasing prevalence[20,21] and its well-established role in the pathogenesis of loCRC[22]. The underlying mechanisms linking obesity to colorectal carcinogenesis are multifactorial. One key pathway involves hyperinsulinemia, which occurs with or without metabolic syndrome. Chronic insulin resistance leads to persistently elevated insulin levels, a condition known to promote tumorigenesis in loCRC[23-25]. Additionally, obesity is associated with gut microbiome alterations[26,27], resulting in the release of pro-inflammatory microbial molecules that compromise intestinal barrier integrity[28]. Furthermore, obesity-driven epigenetic modifications and microbial metabolite changes may contribute to colonic epithelial DNA damage, thereby increasing cancer risk[29,30]. Given the growing body of evidence linking obesity to eoCRC, our study aims to further explore this association and contribute additional insights to the existing literature. While obesity is increasingly recognized as a significant risk factor, inconsistencies in the literature highlight the need for further investigation. By contributing new evidence, this study seeks to clarify the role of obesity in eoCRC to better understand its precise role in early-onset-disease.

MATERIALS AND METHODS

Study participants and research design

This cross-sectional study used de-identified publicly available data from the National Health and Nutrition Examination Survey (NHANES) database, a nationally representative survey conducted in the United States to assess health and nutritional status. Survey cycles from 1999 to 2023 were analyzed. Per NHANES protocols, participants were given written informed consent. Individuals aged 18 to 49 were included, and those who had incomplete information on essential variables pertinent to the study [height, body mass index (BMI), or cancer diagnosis] were excluded. After applying inclusion and exclusion criteria, a total of 30321 participants were included in the final analysis (Figure 1).

Figure 1 Flowchart of the study participants. BMI: Body mass index; CRC: Colorectal cancer.

Data collection and measurements

NHANES collects extensive health-related information through structured interviews, physical examinations, and laboratory tests. The dataset includes demographic variables (e.g., age, sex, race/ethnicity), lifestyle factors (e.g., smoking status, alcohol consumption), anthropometric measurements [e.g., weight, BMI, waist circumference (WC)], and medical history, including self-reported CRC diagnosis. For this analysis, only participants aged 18 to 49 years were included, and individuals diagnosed with CRC were considered only if the diagnosis occurred within this age range. A trained team of researchers collected participant data through standardized NHANES protocols. Anthropometric measurements were obtained by trained health technicians at mobile examination centers, and biological samples were collected and processed at designated laboratories following standardized procedures to ensure data consistency and accuracy.

Definition of variables

The presence of cancer as a diagnosis was determined based on self-reported medical history. We have excluded all participants who skipped this question. Participants who reported a diagnosis of CRC were also asked to provide the age at diagnosis, which allowed the identification of early-onset cases (diagnosed between ages 18 and 49). Demographic and lifestyle variables included age, gender, race, smoking, and alcohol consumption status, and were derived from NHANES questionnaires. Heavy drinkers and non-heavy drinkers were defined based on the National Institute on Alcohol Abuse and Alcoholism definition of heavy drinking[31]. Smoking status was classified using Centers for Disease Control and Prevention criteria as current smokers and never smokers[32]. Obesity was classified with BMI following World Health Organization criteria in which people were classified as overweight (BMI ≥ 25 kg/m²) and obese (BMI ≥ 30 kg/m²)[33]. For WC, participants were classified by the following cut-off points: Normal weight (men ≤ 102 cm; women ≤ 88 cm) and obese (men > 102 cm; women > 88 cm), according to the Journal of American Medical Association Internal Medicine criteria[34]. The body roundness index (BRI), was calculated as follows: BRI = 364.2 - 365.5 × √[1 - (WC/2π)² / (0.5 × height)²], and BRI values were then divided into quartiles (Q1-Q4) to assess associations with eoCRC[35].

Statistical analysis

Baseline characteristics were analyzed using the unweighted “CreateTableOne” function. Categorical variables were treated as factors and summarized as frequencies and percentages, while continuous variables were presented as mean ± SD. Comparisons between groups were performed using Student’s t-test or Wilcoxon rank-sum test for continuous variables and χ² or Fisher’s exact test for categorical variables.

Unweighted univariate analysis was performed to compare baseline characteristics among participants. Categorical variables were expressed as percentages. The analysis included key demographic and clinical variables: Age, gender, race/ethnicity, obesity status (based on BMI and WC), BRI quartiles, smoking status, and heavy alcohol consumption. To assess the relationship between eoCRC and obesity, separate unweighted unadjusted logistic regression models were performed, evaluating obesity status based on BMI, WC, and BRI quartile independently. A multivariable logistic regression unweighted analysis was performed to examine the association between BRI quartiles and the risk of eoCRC. The lowest BRI quartile was used as the reference group. The model adjusted for obesity status (based on BMI and WC), race/ethnicity, age, smoking status, and heavy alcohol consumption. Reference groups were defined as follows: White non-Hispanic individuals for race, non-smokers for smoking status, normal-weight individuals for BMI and WC, and non-heavy drinkers for alcohol consumption. All statistical analyses were performed using R Studio software (version 2023.12.1 + 402). Results with a P < 0.05 were considered statistically significant.

To balance covariates between eoCRC cases and controls, propensity score matching was performed using a 1:4 nearest neighbor approach with a caliper of 0.1, based on age, gender, and ethnicity. After matching, we first conducted an unadjusted logistic regression to estimate the crude associations between obesity-related measures and eoCRC risk. Next, we adjusted for smoking status and heavy drinking - two well-established lifestyle risk factors for CRC. Finally, we estimated a fully adjusted model, incorporating all three obesity-related metrics (BRI quartiles, BMI-based obesity, and WC-based obesity) to assess their independent contributions. Odds ratios (OR) and 95% confidence intervals (CI) were reported for each predictor. Statistical significance was defined at P < 0.05. All additional analyses were conducted in Stata.

To address potential small-sample bias, we performed Firth’s penalized logistic regression as a sensitivity analysis. This approach provides bias-reduced maximum likelihood estimates, particularly suited for rare events and small sample sizes. The model included BRI quartiles, smoking status, and heavy drinking. This was followed by a post hoc power analysis

RESULTS

Baseline characteristics

The study consisted of 30321 participants, 48 (0.16%) of whom were diagnosed with eoCRC. Of these 48 patients, 28 were within the target age range of 18-49 at the time of the survey. The other 20 patients were above the target age range at the time of the survey but responded that they had been within the target age range at the time of diagnosis. The mean age was 34.37 years (SD = 8.61) in the overall cohort, while the mean age of eoCRC patients was 39.96 years (SD = 7.61, P < 0.001). These findings are shown in Table 1. Gender distribution was similar between both groups (P = 0.751). A significantly higher proportion of eoCRC patients were smokers (P < 0.001). Interestingly, none of the eoCRC patients were categorized as heavy drinkers (P = 0.006).

Table 1 Baseline characteristics of study participants.

	Overall	Non eoCRC	eoCRC	P value
Number (n)	30321	30273	48
Age (year)	34.37 ± 8.61	34.36 ± 8.61	39.96 ± 7.61	< 0.001
Gender	16170 (53.3)	16146 (53.3)	24 (50.0)	0.751
Smoker	11945 (39.4)	11912 (39.3)	33 (68.8)	< 0.001
Heavy drinker	4493 (18.8)	4493 (18.8)	0 (0.0)	0.006
BRI quartile				0.002
Q1 (lowest)	10176 (34.7)	10171 (34.7)	5 (11.1)
Q2	7364 (25.1)	7350 (25.1)	14 (31.1)
Q3	5972 (20.4)	5963 (20.4)	9 (20.0)
Q4 (highest)	5827 (19.9)	5810 (19.8)	17 (37.8)
Waist circumference (cm)	96.44 ± 16.86	96.43 ± 16.86	105.73 ± 13.97	< 0.001
	14369 (49.0)	14337 (48.9)	32 (71.1)	0.005
BMI				0.028
Normal weight	9893 (32.6)	9886 (32.7)	7 (14.6)
Overweight	9519 (31.4)	9500 (31.4)	19 (39.6)
Obese	10909 (36.0)	10887 (36.0)	22 (45.8)
Height (cm)	168.21 ± 9.99	168.21 ± 9.99	169.22 ± 9.36	0.484
Weight (kg)	82.05 ± 22.61	82.04 ± 22.62	88.12 ± 18.09	0.063
BRI	5.06 ± 2.35	5.05 ± 2.35	6.19 ± 2.23	0.001
BMI	28.92 ± 7.30	28.92 ± 7.30	30.84 ± 6.43	0.068

Categorical variables were treated as factors and summarized as frequencies and percentages, while continuous variables were presented as mean ± SD or median (IQR), as appropriate. A comparison between groups were performed using Student’s t-test or Wilcoxon rank-sum test for continuous variables and χ² or Fisher’s exact test for categorical variables. P < 0.05 were considered as significant value. eoCRC: Early-onset colorectal cancer; BRI: Body roundness index; BMI: Body mass index.

Unadjusted Logistic regression

The unadjusted logistic regression models shown in Table 2 evaluated the association between key predictors (BRI quartiles, obesity status, smoking, and heavy drinking) and eoCRC risk, without adjusting for other confounders. There was a significant association between increasing BRI quartiles and eoCRC risk, with the highest quartile showing a nearly 10-fold increase in odds. Obesity and overweight - as defined by BMI - were also strong risk factors for eoCRC, showing a 3.75-4.26-fold increase in odds compared to normal weight individuals. WC alone was also significantly associated with eoCRC risk. As for lifestyle factors, smoking significantly increased eoCRC risk by approximately 2.4 times, while heavy drinking was not significantly associated.

Table 2 Summary of unadjusted Logistic regression results.

	OR	Std. Error	P value	95%CI (lower)	95%CI (upper)
BRI quartile 2	14.23	4.49	< 0.011	2.64	1.42
BRI quartile 3	17.73	5.03	< 0.012	3.23	1.42
BRI quartile 4	31.84	10.12	< 0.001	3.22	5.92
Obesity (BMI)	12.55	4.77	< 0.002	2.36	1.81
Overweight (BMI)	10.99	4.03	< 0.006	2.06	1.48
Obese (waist)	7.86	3.62	0.001	1.43	1.67

We compared body roundness index quartiles to waist and body mass index. P < 0.05 were considered as significant value. OR: Odds ratio; 95%CI: 95% confidence interval; BRI: Body roundness index; BMI: Body mass index.

Adjusted Logistic regression

After adjusting for smoking and heavy drinking, the statistically significant association between higher BRI quartiles and eoCRC risk remained. This is shown in Table 3 in which the highest quartile (Q4) has 10-fold higher eoCRC risk compared to Q1 (OR = 10.12, 95%CI: 3.22-31.84). BMI-based overweight and obesity also remained significant predictors of eoCRC (OR = 4.03 and 4.77, respectively). The risk for overweight individuals is nearly as high as for obese individuals, suggesting that even moderate weight gain contributes to eoCRC risk. WC alone remained a significant predictor of eoCRC risk (OR = 3.62, P = 0.001).

Table 3 Summary of adjusted Logistic regression results.

Variable	OR	Std. Error	P value	95%CI (lower)	95%CI (upper)
BRI quartile 2	4.1	2.29	< 0.012	1.37	12.26
BRI quartile 3	4.36	2.64	< 0.015	1.33	14.31
BRI quartile 4	9.74	5.52	< 0.0001	3.2	29.6
Obesity (BMI)	4.26	2.02	< 0.002	1.68	10.8
Overweight (BMI)	3.75	1.81	< 0.006	1.46	9.67
Obese (waist)	1.14	0.37	0.002	0.41	1.86
Smoker	2.43	0.85	< 0.011	1.22	4.83
Not heavy drinker	0.67	0.28	0.341	0.29	1.53

We compared body roundness index quartiles to waist and body mass index while adjusting for smoking and heavy drinking. P < 0.05 were considered as significant value. OR: Odds ratio; 95%CI: 95% confidence interval; BRI: Body roundness index; BMI: Body mass index.

Adjusted Logistic regression

Lastly, three adjusted logistic regression models were compared to evaluate the association between BRI quartiles, BMI-based obesity, and WC-based obesity with eoCRC risk, while adjusting for smoking and heavy drinking. This is shown in Table 4 below.

Table 4 Independent contributions of body roundness index, body mass index, and waist circumference to early-onset colorectal cancer risk.

Variable	BRI + BMI model		BRI + waist model		BRI + BMI + waist model
	OR	P value	OR	P value	OR	P value
BRI quartile 2	4.61	0.106	4.02	0.024	3.95	0.158
BRI quartile 3	6.47	0.097	3.92	0.082	4.76	0.199
BRI quartile 4	16.04	0.018	7.7	0.007	11.7	0.053
Obese (BMI)	0.62	0.64			0.6	0.615
Overweight (BMI)	1.05	0.96			1.07	0.936
Obese (waist)			1.34	0.608	1.46	0.506

We assessed the significant of body roundness index across 3 different models (model 1 body roundness index + body mass index; model 2 body roundness index + waist; and model 3 body roundness index + body mass index + waist). P < 0.05 were considered as significant value. BRI: Body roundness index; BMI: Body mass index.

BRI + BMI model

Higher BRI quartiles are associated with increased eoCRC risk, with Q4 being the strongest predictor. The highest quartile (Q4) shows a significantly higher eoCRC risk compared to Q1 (OR = 16.04, P = 0.018). Q2 and Q3 show a trend but do not reach statistical significance. BMI-based obesity does not appear to be a significant predictor of eoCRC risk when adjusted for BRI and other factors. This suggests that body roundness (BRI) may be a stronger predictor than BMI alone.

BRI + WC model

Higher BRI quartiles are associated with increased eoCRC risk, with Q4 being the strongest predictor. The highest quartile (Q4) shows a significantly higher eoCRC risk compared to Q1 (OR = 7.70, P = 0.007). WC is not significantly associated with eoCRC risk. This suggests that body roundness (BRI) may be a stronger predictor than WC alone.

Full model (BRI + BMI + WC)

In the full model, BRI Q4 retained marginal significance (OR = 11.70, P = 0.053), while BMI-based and waist-based obesity remained non-significant. BRI quartiles remain the strongest predictor of eoCRC, even after adjusting for both BMI and WC.

Sensitivity analysis

Given the limited number of eoCRC cases and potential issues related to small-sample bias, we conducted additional sensitivity analyses to assess the robustness of our findings. First, we implemented Firth’s penalized logistic regression which provides bias-reduced estimates and is well-suited for rare events and small sample sizes. This method confirmed the strong association between higher BRI quartiles and increased eoCRC risk, with statistically significant results consistent with our primary analysis. Individuals in the highest quartile showed over 9 times the odds of eoCRC compared to the lowest quartile (OR = 9.07, P < 0.001) as seen Supplementary Table 1.

Additionally, we performed a post hoc power analysis which demonstrated that our matched sample (n = 199) had a > 99% power to detect large effect sizes (OR ≥ 9, as observed for BRI Q4), but limited power to detect more modest associations (OR approximately 2-3) as seen in Supplementary Table 2. These findings suggest that the non-significant associations observed for mid-range BRI quartiles and other obesity measures may reflect limited statistical power rather than a true lack of effect.

DISCUSSION

Unlike BMI, BRI considers an individual’s body shape and more clearly represents central adiposity due to incorporation of WC relative to height. Thomas et al[35] constructed BRI based on a geometric model that showed better association with total body fat percentage and visceral adipose tissue volume than with BMI and WC alone. Central obesity, which is reflected in higher BRI values, appears to be the strongest contributor to insulin resistance. In contrast, individuals classified as obese based on BMI (> 30) but with a more evenly distributed body fat pattern tend to experience lower levels of insulin resistance, reduced secondary hyperinsulinemia, and consequently, a lower risk of obesity-induced carcinogenesis[36]. This is further supported by our findings, where higher BRI quartiles were consistently associated with an increased risk of eoCRC (Table 3). A similar cross-sectional study conducted recently in 2023 was able to provide convincing evidence of a significant positive association between elevated BRI and CRC risk[37]. However, that study covered data from 1999 to 2020, while our study is revised up to 2023, with approximately 4000 more patients analyzed. Moreover, our study excluded individuals who did not meet criteria for early onset, while the aforementioned study included participants diagnosed at any age. Given the concerning trend of rising CRC incidence among younger populations while rates remain stable or even decline in older adults (> 49 years old)[38], it is crucial to investigate risk factors unique to early-onset cases. Since visceral fat is a key factor in metabolic derangements, BRI has been suggested for use as a new screening marker of metabolic syndrome—a cluster of cardiometabolic risk factors consisting of abdominal obesity, dyslipidemia, hypertension, and insulin resistance. In a large Brazilian sample presented by do Prado et al[39], BRI cutoff points were found to be highly sensitive and specific for the diagnosis of metabolic syndrome, and the association of BRI with metabolic syndrome was stronger than that of BMI and WC after adjusting for age, sex, physical activity, and smoking status. Similarly, Fahami et al[40] found that participants in the highest quartile of BRI score were significantly more likely to have features of metabolic syndrome, including hyperglycemia, hypertension, hypertriglyceridemia, and lower levels of high-density lipoprotein compared with individuals in the lowest quartile. These results are consistent with other studies that have reported BRI might be a better representative of visceral adiposity that could lead to metabolic syndrome by inducing insulin resistance, systemic low-grade inflammation, and unfavorable lipid profiles[41]. Visceral adiposity is now well recognized as a core mediator of systemic metabolic derangement. Increased visceral fat leads to a state of chronic low-grade inflammation mediated by the secretion of pro-inflammatory cytokines including TNF-α, IL-6, and MCP-1, resulting in insulin resistance and endothelial dysfunction[41]. This inflammatory microenvironment is conducive to carcinogenesis by increasing cellular proliferation, reducing apoptosis, inducing angiogenesis, and causing DNA damage through production of reactive oxygen species. Adipose-derived inflammatory cytokines can turn on oncogenic signaling pathways, such as NF-κB and STAT3, thereby facilitating tumorigenesis[41]. In addition, central obesity changes composition of the intestinal microbiota by increasing the levels of pro-inflammatory bacteria (e.g., Enterobacteriaceae) and decreasing beneficial short chain fatty acid producing bacteria such as Faecalibacterium prausnitzii[42]. These changes in the microbiome result in impaired mucosal barrier function, bacterial translocation, endotoxemia, and chronic host immune response activation, which establish local and systemic conditions that promote colorectal tumorigenesis[42]. Given these biological links of visceral adiposity and metabolic dysregulation to carcinogenesis, we aimed to directly examine the relationship of BRI and risk of eoCRC within a large, nationally representative cohort. Using NHANES data from 1999 to 2023, this study aimed to evaluate the association between obesity and eoCRC. The association between increasing BRI quartiles and eoCRC risk was clear and consistent. Those in the highest quartile had significantly higher odds of eoCRC than those in the lowest quartile, even after controlling for other obesity metrics (BMI and WC) and known lifestyle risk factors (e.g., smoking and heavy alcohol use). While all two-dimensional body fat indices (BMI and WC) were associated with increased risk in unadjusted and partially adjusted models, significant associations between the two-dimensional indices and risk disappeared when BRI was introduced to full multivariable models, suggesting BRI may be a more holistic measure of body fat distribution and shape. BRI is therefore a potentially better measure for evaluating obesity-related cancer risk among younger adults. BRI remained predictive even after controlling confounding factors, including smoking and alcohol use. Notably, only BRI retained statistical significance (P = 0.053) in the full model that accounted for BRI, BMI, and WC. These observations are consistent with the current literature, which suggests that BMI alone has limited ability to assess health outcomes and supports consideration of other estimates, such as body shape indices, in the cancer setting. The role of obesity in carcinogenesis is likely multifactorial, with secondary hyperinsulinemia-induced carcinogenesis being one of the most well-supported mechanisms in the literature[23,25]. Besides obesity, factors such as alcohol consumption, smoking, diet, and physical activity levels have all been implicated in the increasing incidence of eoCRC[2,4,7,10]. However, the literature presents conflicting findings regarding the actual impact of these factors on eoCRC development. Therefore, adjusting for all potential confounders was essential to ensure the accuracy of our results, and even after these adjustments, our findings remained strongly positive and statistically significant. The study’s primary strength is the use of propensity score matching to reduce the effects of confounding demographic factors and alleviate the case-control imbalance, as there were only 48 eoCRC individuals identified within a cohort of over 30000 individuals. By matching cases and controls by age, gender, and ethnicity, we reduced the potential bias that demographic factors may have introduced and improved the internal validity of the study. Furthermore, the potential effect of small sample size was offset by our sensitivity analyses which supported the robustness of the observed relationship between BRI and eoCRC risk in our propensity score-matched study design. Nevertheless, certain limitations must be acknowledged. First, due to the cross-sectional design, causal inferences cannot be made. Additionally, self-reported cancer diagnosis and age at diagnosis may introduce recall bias. Finally, residual confounding from unmeasured factors—such as diet, physical activity, and genetic predisposition—could have influenced our results.

CONCLUSION

Our study demonstrated that BRI was a stronger predictor of eoCRC risk than other obesity indices. BRI therefore has considerable potential as a screening tool for identifying young individuals at higher risk of CRC. Further research is warranted to clarify the role of body composition in early carcinogenesis and to determine whether BRI could inform screening recommendations or early interventions in at-risk populations.

ACKNOWLEDGEMENTS

The Statistical Core for Health and Medicine (SCHM) at Heritage College of Osteopathic Medicine in Ohio University provided statistical assistance.