Widening Health Inequality and Causal Metabolic Drivers in Global Colorectal Cancer: A Multi-Dimensional Study

Abstract

Objectives:

To describe the geographical distribution of global colorectal cancer (CRC) and the time trend of its disease burden from 1990 to 2021, and to predict its future development.

Methods:

Data were derived from the Global Burden of Disease Study 2021 (GBD 2021, IHME database), covering 204 countries and 21 GBD regions. Analyses were conducted both globally and stratified by socio-demographic index (SDI) region. Three-stage analysis was carried out around incidence, mortality and disability-adjusted life years (DALYs): trend analysis by Joinpoint regression and age-period-cohort model; decomposition of disease burden drivers by Das Gupta decomposition and data envelopment analysis; 15-year forecasting based on the Bayesian age-period-cohort analysis. In addition, the causal effect of BMI on CRC risk was assessed using the Mendelian randomization (MR) method.

Results:

In 2021, 2.19 million new CRC cases were reported globally (age-standardized incidence rate 25.61/100 000). From 1990 to 2021, the incidence rate continued to rise, while the mortality rate and DALYs declined — a divergence likely attributable to improved screening and treatment, particularly in high-SDI regions. Population aging is the primary driving factor for the increased CRC global burden (47.44%), while population growth is dominant in low SDI regions. Health inequality increased, and the slope index rose from 437 in 1990 to 575 in 2021. Projections showed that the incidence rate will increase slightly (+0.56%) by 2036, while the mortality rate (−9.5%) and DALYs (−8.1%) will continue to decline. MR analysis supported a causal association between BMI and increased CRC risk.

Conclusion:

The global CRC burden remains substantial, driven by population aging, the obesity epidemic, and widening regional disparities. Strategies should prioritize age-targeted screening in high-SDI regions, expand early detection in low-SDI areas, and integrate obesity control into national non-communicable disease prevention frameworks.

Keywords

colorectal cancer health equity body mass index global burden of disease mendelian randomization analysis

Introduction

Colorectal cancer (CRC) is a significant global health burden, ranking as the third most commonly diagnosed cancer and the second leading cause of cancer-related deaths worldwide.¹ In 2020, there were an estimated 1.9 million new cases and 935 000 deaths attributed to CRC.² The etiology of CRC is complex and multifactorial, involving an interplay of genetic, environmental, and lifestyle factors.³ Established risk factors include family history, inflammatory bowel disease, obesity, physical inactivity, smoking, excessive alcohol consumption, and diets high in red and processed meat and low in fiber.^4-6 In recent decades, the rising prevalence of obesity and the adoption of Western lifestyles in many developing countries have been implicated in the escalating burden of CRC.⁷

Despite the substantial disease burden, CRC is highly preventable and treatable if detected early.⁸ Screening programs have been shown to be effective in reducing CRC incidence and mortality by detecting and removing precancerous lesions and early-stage cancers.⁹ However, the implementation and uptake of screening vary widely across countries, largely influenced by economic development, health system capacity, and public awareness.¹⁰ Treatment for CRC typically involves surgery, chemotherapy, and radiotherapy, depending on the stage and location of the tumor.¹¹ While survival rates have improved in many high-income countries due to advances in treatment and increased access to care, the prognosis remains poor in many low- and middle-income countries due to late diagnosis and limited treatment options.¹²

Given the significant global burden and the disparities in incidence, mortality, and survival across regions, a comprehensive understanding of the epidemiology and disease burden of CRC is crucial for informing prevention and control strategies. The Global Burden of Disease (GBD) study provides a unique framework for quantifying and comparing the health loss due to CRC across countries and over time.¹³ Previous GBD analyses have highlighted the substantial and increasing burden of CRC, particularly in developing countries.¹⁴ However, the causal contribution of individual-alterable exposure factors, especially body mass index (BMI), to these population-level trends has not been fully clarified. Observational studies have consistently shown that higher BMI is associated with colorectal cancer risk, but residual confounding factors and reverse causality limit the effectiveness of causal inference in such study designs. Mendelian randomization (MR) is an epidemiological method that uses genetic variation as an instrumental variable to infer the causal relationship between exposure and outcome. Due to the random distribution of alleles during gamete formation, MR design is similar to “randomized controlled trials in nature,” which can effectively avoid reverse causality and confounding bias in traditional observational studies.¹⁵ Based on this theoretical basis, this study combines the updated GBD 2021 estimated data on the incidence, mortality, and disability-adjusted life years (DALYs) of CRC with 2-sample mixed inference (MR) analysis for BMI and colorectal cancer, so as to examine time trends, regional differences, and modifiable pathogenic factors within a single coherent framework.

Materials and Methods

Data Sources and Indicator Calculation

The data used in this study were sourced from the Global Burden of Disease (GBD) 2021 study, coordinated by the Institute for Health Metrics and Evaluation (IHME). Epidemiological data on CRC spanning 1990 to 2021 across 204 countries and territories (aggregated into 21 GBD super-regions) were downloaded from the GBD Results Tool (https://vizhub.healthdata.org/gbd-results). The collected indicators included new cases, deaths, DALYs, crude incidence, age-standardized incidence rate (ASIR), age-standardized mortality rate (ASMR), and age-standardized DALY rate (ASDR). The GBD study used the world standard population for age standardization, and all age-standardized rates were expressed per 100 000 population.

Statistical Analysis Methods

Overview of analytical framework

In order to comprehensively describe the global disease burden of CRC and its potential drivers, this study constructed an integrated multi-method analysis framework, focusing on 4 core scientific issues: (1) how the disease burden changes at the time and regional level; (2) What factors drive these changes; (3) What is the trend of disease burden in the future; (4) Whether there is a causal association between body mass index (BMI) and CRC risk. The choice of methods emphasizes complementarity rather than redundancy. Descriptive analysis, Joinpoint regression, and age-period-cohort analysis are exploratory methods that aim to identify and decompose time trends without presupposing causal mechanisms. Decomposition analysis, frontier analysis and health inequality analysis are also exploratory, but they focus more on quantifying the contribution of structural drivers such as population aging, development level and socioeconomic inequality. The prediction of disease burden based on BAPC model belongs to inferential analysis, and the existing trend is extrapolated under the probability framework. Mendelian randomization analysis is a confirmatory method, which can effectively reduce confounding bias by using genetic instrumental variables, so as to provide more robust causal inference. In general, these methods jointly construct a systematic analysis path from description to interpretation, from prediction to causal verification.

Descriptive Analysis of Disease Burden

Descriptive analysis was performed on CRC incidence and DALYs from 1990 to 2021 for the global population and 21 regions. The new cases, deaths, DALYs, ASIR, ASMR, and ASDR for each region and year were calculated, and the estimated annual percentage change (EAPC) of these indicators from 1990 to 2021 was computed. A disease burden heatmap was created to show the global distribution of CRC.

Joinpoint Regression Analysis

The Joinpoint regression model was used to analyze the trend of change in the ASIR of CRC from 1990 to 2021. The ASIR standard error was first calculated using R software, and the data were imported into Joinpoint software. The independent variable was time (year), the dependent variable was the age-standardized rate, and the grouping variable was region. The model allowed for up to 5 join points, and Monte Carlo permutation tests were used to select the best model, with a 95% confidence interval level. The results were expressed as the annual percentage change (APC).

Age-Period-Cohort Analysis

To explore the relative contributions of age, period, and cohort effects in the trend of CRC, the data from 1992 to 2021 (excluding 1990-1991 data to ensure a complete 5-year period) were divided into 5-year periods and uploaded to the APC website for analysis. The net effect of each factor on CRC incidence was calculated, and the relative risk curves were plotted for age, period, and birth cohort. The local drift percentage of disease risk between age groups was also computed.

Decomposition Analysis

Das Gupta decomposition was used to analyze the relative contribution of population, epidemiological factors, and population aging to the change in CRC incidence from 1990 to 2021. Decomposition was performed at the global level, for different gender groups, and for regions with different socio-demographic index (SDI) levels.

Frontier Analysis

Data Envelopment Analysis (DEA) was used to evaluate the relationship between the CRC disease burden and SDI. The Free Disposal Hull (FDH) model was used to fit the non-linear production frontier, and the locally weighted scatterplot smoothing (LOWESS) method was applied to generate a smooth frontier. Outlier points (ie, those falling below the frontier) were excluded. The differences between the observed values and frontier values for each region and year were calculated to assess the disease burden relative to SDI.

Health Inequality Analysis

The Slope Index of Inequality (SII) and Concentration Index (CI) were calculated for the age-standardized DALY rates of CRC. SII reflects the absolute inequality in disease burden, and CI reflects relative inequality. SII and CI were calculated for 1990 and 2021, and scatter plots were used to compare health inequality changes between the 2 years.

Disease Burden Prediction

The Bayesian Age-Period-Cohort (BAPC) model was used to predict the ASIR, ASMR, and ASDR of CRC from 2022 to 2036. The model parameters included a second-order random walk prior distribution, with 100 000 iterations in Markov Chain Monte Carlo, a burn-in period of 10 000, and a sampling interval of 10. The model considered the effects of age, period, and cohort, and combined these with predicted future period and cohort effects to calculate disease burden indicators for the next 15 years.

Mendelian Randomization Analysis

A 2-sample Mendelian randomization (MR) analysis was conducted to evaluate the causal relationship between BMI and CRC risk. A 2-sample Mendelian randomization (MR) analysis was conducted to evaluate the causal relationship between body mass index (BMI) and CRC risk, based on 3 core assumptions: (1) the instrumental variables (IVs) are strongly associated with BMI; (2) the IVs are independent of confounders; and (3) the IVs influence CRC only through BMI. Summary-level genetic association data for BMI were obtained from publicly available GWAS (European ancestry), and those for CRC were obtained from the FinnGen biobank; as only de-identified summary statistics were used, no additional ethical approval was required. Single nucleotide polymorphisms (SNPs) associated with BMI at genome-wide significance (P < 5×10-8) and mutually independent (r2 < 0.001, clumping window = 10 000 kb) were selected as IVs. After excluding palindromic SNPs with intermediate allele frequencies and harmonizing effect alleles between the exposure and outcome datasets, 422 SNPs were retained. Instrument strength was evaluated by the F-statistic, with F > 10 considered sufficient to minimize weak instrument bias. The inverse-variance weighted (IVW) method was used as the primary analysis, with the weighted median, MR-Egger regression, simple mode, and weighted mode methods as complementary approaches relying on different pleiotropy assumptions. Causal estimates were expressed as odds ratios (OR) with 95% confidence intervals per 1-SD increase in genetically predicted BMI. Horizontal pleiotropy was assessed using the MR-Egger intercept test and MR- PRESSO; between-SNP heterogeneity was quantified using Cochran’s Q statistic; and leave- 1-out analysis was performed to identify potentially influential SNPs. MR analyses were implemented using the TwoSampleMR R package.

All data analysis and plotting were performed using R 4.2.0, and the BAPC model used the BAPC package. Statistical significance was set at P < .05. The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement [Supplemental FILE 1].

Results

Temporal Trends in Colorectal Cancer Burden

In 2021, there were 2 194 143 new cases of CRC globally (95% uncertainty interval [UI]: 2,001,271.82-2,359,390.09), with an age-standardized incidence rate (ASIR) of 25.61 per 100 000 (95% UI: 23.32-27.52). Regionally, East Asia had the highest number of new cases (684 927, 95% UI: 559,522.76-823,301.33), while Oceania had the lowest (482 cases, 95% UI: 409.98-560.14). The ASIR showed significant regional differences, with High-income Asia Pacific having the highest rate (44.89/100 000, 95% UI: 40.20-47.85) and South Asia having the lowest (5.65/100 000, 95% UI: 5.08-6.30). From 1990 to 2021, the estimated annual percentage change (EAPC) in ASIR was highest in South Asia, at 2.05% (95% UI: 1.99 to 2.11), while Southeast Asia saw the largest decrease (EAPC = −0.80%, 95% UI: −0.93 to −0.67; Table 1). Global CRC-related deaths totaled 1 044 072 (95% UI: 950,187.61-1,120,169.34), with an age-standardized mortality rate (ASMR) of 12.40 per 100 000 (95% UI: 11.24-13.31). Regionally, East Asia had the highest number of deaths (287 880, 95% UI: 235,559.37-343,280.49), while Oceania had the lowest (382, 95% UI: 324.23-445.96). The ASMR was highest in Central Europe (22.585/100 000, 95% UI: 20.81-24.28) and lowest in South Asia (4.63/100 000, 95% UI: 4.17-5.16). In Tropical Latin America, the ASMR showed the largest decrease (EAPC = −1.92%, 95% UI: −2.01 to −1.82), while Western Sub-Saharan Africa saw the highest increase (EAPC = 0.97%, 95% UI: 0.70 to 1.25; Table 1).

Table 1.

Global and 21 GBD Regions Colorectal Cancer Incidence, Deaths, DALYs, ASIR, ASMR, and ASDR in 2021 and Their Estimated Annual Percentage Change from 1990 to 2021.

Location	Incidence number (95% UI)	ASIR (95% UI)	ASIR EAPC 1990-2021 (95%UI)	Deaths number (95% UI)	ASMR (95% UI)	ASMR EAPC 1990-2021 (95% UI)	DALYs number (95% UI)	ASDR (95% UI)	ASDR EAPC 1990-2021 (95% UI)
Global	2,194,143.250 (2,001,271.824-2,359,390.088)	25.607 (23.322-27.516)	1.426 (1.321 to 1.531)	1,044,072.211 (950,187.610-1,120,169.335)	12.398 (11.241-13.306)	0.463 (0.354 to 0.573)	24,401,100.180 (22,689,368.549-26,161,517.726)	283.242 (263.114-303.326)	0.385 (0.273 to 0.498)
Andean Latin America	8452.057 (6698.577-10,526.053)	14.389 (11.395-17.891)	0.980 (0.813 to 1.147)	5777.836 (4597.210-7047.450)	10.004 (7.975-12.183)	−0.026 (−0.152 to 0.100)	136,183.213 (108,534.692-167,981.611)	226.761 (180.593-279.799)	−0.122 (-0.248 to 0.004)
Australasia	23,280.743 (20,501.569-26,412.689)	43.968 (38.907-49.552)	0.236 (0.058 to 0.415)	8275.772 (7178.893-9408.206)	14.618 (12.769-16.519)	0.060 (−0.091 to 0.212)	166,740.101 (147,523.691-186,890.825)	327.189 (290.904-364.679)	0.024 (−0.123 to 0.171)
Caribbean	18,481.225 (16,074.151-20,935.797)	34.332 (29.870-38.855)	1.264 (1.187-1.341)	7881.536 (6866.692-8970.236)	14.573 (12.717-16.595)	0.274 (0.234 to 0.314)	179,751.924 (155,874.184-205,420.066)	335.019 (290.413-383.111)	0.325 (0.281 to 0.370)
Central Asia	8892.968 (7941.559-9809.245)	10.823 (9.692-11.904)	−0.183 (−0.273 to 0.092)	6146.247 (5492.549-6771.900)	7.867 (7.055-8.655)	−0.369 (−0.474 to −0.264)	169,952.250 (151,537.430-187,707.997)	196.551 (175.594-216.681)	−0.386 (−0.480 to −0.292)
Central Europe	85,866.578 (79,290.327-92,786.951)	38.817 (35.706-41.958)	0.153 (0.115 to 0.191)	51,843.486 (47,752.159-55,681.287)	*22.585 (20.808-24.275)	−0.806 (−0.838 to −0.773)	1,087,562.425 (1,003,626.421-1,168,378.102)	*506.482 (467.975-544.568)	−0.833 (−0.868 to −0.798)
Central Latin America	44,551.535 (39,665.469- 49,720.643)	17.745 (15.755-19.812)	1.424 (1.306 to 1.541)	22,934.275 (20,341.655-25,521.139)	9.302 (8.259-10.351)	0.661 (0.569 to 0.754)	594,116.791 (529,274.148-662,367.068)	231.951 (206.641-258.569)	0.773 (0.674 to 0.872)
Central Sub-Saharan Africa	4206.451 (3206.030-5573.585)	7.869 (6.094-10.535)	−0.149 (-0.356 to 0.059)	3686.773 (2802.235-4932.004)	7.453 (5.766-10.144)	−0.495 (−0.646 to −0.344)	109,622.047 (82,688.064-146,971.080)	177.792 (135.169-238.587)	−0.878 (−0.999 to −0.758)
East Asia	*684,927.152 (559,522.761-823,301.329)	31.602 (25.896-37.848)	−0.104 (-0.219 to 0.011)	*287,880.009 (235,559.365-343,280.486)	13.782 (11.325-16.346)	−0.213 (−0.309 to −0.116)	*7,148,995.324 (5,822,935.265-8,561,078.939)	334.505 (274.007-399.929)	−0.578 (-0.681 to −0.474)
Eastern Europe	113,252.178 (104,414.531-122,488.524)	32.106 (29.592-34.730)	1.313 (1.151 to 1.475)	64,373.247 (59,115.498-69,822.450)	18.055 (16.583-19.563)	0.208 (0.056 to 0.360)	1,465,104.148 (1,343,980.891-1,600,585.462)	424.543 (389.695-463.673)	−0.013 (−0.148 to 0.122)
Eastern Sub-Saharan Africa	17,951.902 (15,710.425-20,779.075)	11.232 (9.843-12.769)	0.730 (0.556 to 0.905)	15,967.479 (13,939.888-18,320.333)	10.762 (9.407-12.151)	−0.097 (−0.267 to 0.073)	444,252.946 (385,799.783-525,458.316)	242.129 (211.299-279.078)	−0.043 (−0.191 to 0.106)
High-income Asia Pacific	207,276.974 (179,498.139-223,331.926)	*44.890 (40.200-47.850)	0.332 (0.228 to 0.437)	80,690.589 (67,266.650-88,001.374)	14.995 (13.033-16.094)	−0.735 (−0.781 to −0.689)	1,441,629.999 (1,272,495.981-1,549,758.695)	331.482 (303.848- 352.473)	−0.930 (−0.979 to −0.881)
High-income North America	244,681.446 (226,550.009-256,376.909)	38.750 (36.135-40.479)	−0.111 (-0.264 to 0.043)	85,865.380 (77,871.228-90,928.137)	12.955 (11.892-13.652)	−1.403 (−1.450 to −1.356)	1,907,928.368 (1,788,110.923-2,003,823.632)	316.043 (298.523-330.840)	−1.410 (-1.464 to −1.356)
North Africa and Middle East	66,086.551 (58,129.632-74,936.328)	14.429 (12.670-16.349)	0.460 (0.325 to 0.595)	37,394.507 (32,758.786-42,274.127)	8.945 (7.853-10.080)	0.021 (−0.087 to 0.130)	1,012,651.597 (886,198.884-1,154,502.813)	209.038 (183.159-237.276)	−0.098 (−0.206 to 0.010)
Oceania	*481.600 (409.984-560.140)	6.444 (5.549-7.419)	0.624 (0.508 to 0.739)	*381.662 (324.226-445.960)	5.584 (4.798-6.416)	−0.190 (−0.306 to −0.075)	*11,642.972 (9763.997-13,750.355)	136.647 (116.102-159.447)	−0.479 (−0.613 to −0.345)
South Asia	85,115.134 (76,613.731-95,247.802)	*5.647 (5.080-6.298)	*2.050 (1.989 to 2.112)	66,942.714 (60,196.675-74,844.400)	*4.626 (4.165-5.159)	0.857 (0.776 to 0.939)	1,908,668.138 (1,711,982.005-2,155,055.236)	*120.375 (108.304-135.299)	*1.106 (1.029 to 1.183)
Southeast Asia	116,941.670 (101,259.791-132,256.456)	17.699 (15.407-19.894)	*−0.798 (−0.927 to −0.668)	79,419.596 (68,449.689-89,290.459)	12.744 (11.115-14.264)	−1.628 (−1.692 to −1.565)	2,166,650.375 (1,868,880.141-2,456,349.669)	313.372 (270.746-353.540)	−1.361 (-1.416 to −1.305)
Southern Latin America	24,746.824 (21,921.766-27,561.451)	28.316 (25.100-31.528)	1.747 (1.654 to 1.840)	16,117.019 (14,307.191-18,001.796)	18.125 (16.116-20.247)	−0.444 (−0.503 to −0.386)	349,581.126 (311,369.419- 391,893.391)	407.990 (363.051-457.564)	−0.587 (−0.663 to −0.510)
Southern Sub-Saharan Africa	7623.406 (6875.232-8472.177)	13.449 (12.157-14.825)	0.777 (0.706 to 0.849)	6129.623 (5550.193-6785.914)	11.465 (10.368-12.610)	0.618 (0.548 to 0.687)	166,061.782 (149,959.402-186,753.431)	270.635 (244.941-301.851)	0.461 (0.398 to 0.524)
Tropical Latin America	44,244.626 (40,859.427-47,143.628)	17.169 (15.817-18.314)	−0.579 (−0.714 to −0.444)	29,414.564 (26,980.718-31,338.903)	11.578 (10.589-12.342)	*−1.917 (−2.010 to −1.823)	744,957.014 (694,956.435-786,849.085)	286.216 (266.669-302.446)	*−2.063 (−2.169 to −1.957)
Western Europe	375,461.835 (337,713.278-401,850.236)	40.540 (37.170-43.067)	1.449 (1.397 to 1.502)	156,637.467 (136,856.634-169,938.671)	15.112 (13.529-16.253)	0.736 (0.668 to 0.804)	2,912,355.199 (2,641,021.363-3,112,518.769)	326.815 (301.667-347.226)	0.612 (0.549 to 0.675)
Western Sub-Saharan Africa	11,620.394 (9673.144-13,673.399)	6.285 (5.362-7.297)	1.292 (1.062 to 1.522)	10,312.429 (8668.115-12,099.874)	5.980 (5.121-6.882)	*0.974 (0.698 to 1.251)	276,692.441 (224,567.806-328,426.256)	132.051 (110.390-155.480)	1.064 (0.776 to 1.354)

To label the top-ranked (maximum) values of each parameter across 21 GBD geographical regions. It is applied to mark the largest number of new colorectal cancer cases, deaths and DALYs, the highest age-standardized rates (ASIR, ASMR, ASDR), and the most extreme estimated annual percentage change (EAPC).

The global disability-adjusted life years (DALYs) caused by CRC were 24 401 100 (95% UI: 22,689,368.55-26,161,517.73), with an age-standardized DALY rate (ASDR) of 283.24 per 100 000 (95% UI: 263.11-303.33). East Asia had the highest number of DALYs (7 148 995, 95% UI: 5,822,935.27-8,561,078.94), while Oceania had the lowest (11,642.97, 95% UI: 9763.99-13,750.36). The ASDR was highest in Central Europe (506.48/100 000, 95% UI: 467.98-544.57) and lowest in South Asia (120.38/100 000, 95% UI: 108.30-135.30). The most significant decline in ASDR occurred in Tropical Latin America (EAPC = −2.06%, 95% UI: − 2.17 to −1.96), while South Asia saw the largest increase (EAPC = 1.11%, 95% UI: 1.03 to 1.18; Table 1).

Joinpoint Regression Analysis

Using the Joinpoint regression model, this study analyzed the temporal trends in the age-standardized incidence rate (ASIR), age-standardized mortality rate (ASMR), and age-standardized disability-adjusted life year rate (ASDR) of CRC from 1990 to 2021.

The ASIR analysis revealed 3 significant inflection points (1995, 2001, 2007), showing a fluctuating upward trend (Figure 1a). Specifically, from 1990 to 1995, the ASIR increased rapidly (APC = 0.856%, P < .05), then plateaued from 1996 to 2001 (APC = −0.093%), accelerated again from 2002 to 2007 (APC = 0.308%, P < .05), and remained relatively stable from 2008 to 2021 (APC = −0.007%). Gender-stratified analysis revealed notable differences, with a transient increase in female ASIR from 1990 to 1994 (APC = 0.613%, P < .05), followed by a slight decline from 1995 to 1998 (APC = −0.298%) and further reduction from 1999 to 2001 (APC = −0.682%), which was still non-significant. The decline slowed between 2002 and 2007 (APC = −0.181%, P < .05), then dropped more sharply from 2008 to 2012 (APC = −0.688%, P < .05), stabilizing again from 2013 to 2021 (APC = − 0.108%). In contrast, male ASIR showed a continuous upward trend, with rapid growth from 1990 to 1995 (APC = 1.101%, P < .05), followed by slower growth from 1996 to 2001 (APC = 0.301%, P < .05), an expansion of the increase from 2002 to 2009 (APC = 0.693%, P < .05), and a significant deceleration from 2010 to 2021 (APC = 0.187%, P < .05).

Figure 1.

Joinpoint regression analysis of global age-standardized colorectal cancer rates from 1990 to 2021: incidence (a), mortality (b), and disability-adjusted life years (DALYs; c).

ASMR analysis identified 5 significant inflection points (1993, 1996, 2002, 2006, 2013), with an overall declining trend (Figure 1b). From 1990 to 1993, the ASMR declined slowly (APC = −0.191%), then the decline significantly accelerated from 1994 to 1996 (APC = −1.002%, P < .05). The rate of decrease slowed slightly from 1997 to 2002 (APC = −0.586%, P < .05), then the decline became steeper again from 2003 to 2006 (APC = −1.350%, P < .05). From 2007 to 2013, the decline slowed (APC = −0.820%, P < .05), continuing at a stable rate from 2014 to 2021 (APC = −0.538%, P < .05). Gender-stratified analysis revealed that female ASMR declined significantly throughout the period, with the largest drop occurring from 2003 to 2006 (APC = −1.760%, P < .05). Male ASMR showed a more complex pattern, with insignificant declines during 1990 to 1993 (APC = −0.057%, P > .05) and 2007 to 2009 (APC = −0.225%, P > .05), but with statistically significant reductions in all other periods, particularly from 2004 to 2006 (APC = −1.306%, P < .05).

The ASDR analysis showed a similar trend with ASMR, identifying 5 inflection points (1993, 1996, 2002, 2006, 2013; Figure 1c). The overall trend from 1990 to 1993 showed a slow decrease (APC = −0.170%, P > .05), which accelerated significantly from 2003 to 2006 (APC = −1.392%, P < .05) and slowed down again from 2014 to 2021 (APC = −0.475%, P < .05). Gender differences were significant, with female ASDR showing a steady decline, with the steepest decline between 2003 and 2006 (APC = −1.877%, P < .05). In contrast, male ASDR showed a slight increase in the early period (1990-1993, APC = 0.049%, P > .05), followed by significant declines in subsequent periods, particularly from 2004 to 2006 (APC = −1.155%, P < .05).

Overall, despite an upward trend in ASIR, both ASMR and ASDR continued to decline, indicating substantial progress in the prevention, diagnosis, and treatment of CRC. Gender differences were also evident, with females showing greater improvements across various disease burden indicators, especially in incidence, where females showed a decreasing trend while males exhibited a sustained increase, resulting in an expanding gender gap.

Age-Period-Cohort Effects on CRC Incidence and Mortality

The age-period-cohort analysis of CRC incidence and mortality revealed multidimensional changes in the disease’s epidemiological characteristics. The age effect showed that both the incidence and mortality risks increased with age, but with different patterns (Figure 2a and e). The incidence risk started to rise at the age of 17.5 years (0.285/100 000), accelerated, peaking at 92.5 years (346.885/100 000) and then slightly declined. The mortality risk followed a similar starting point (0.262/100 000 at 17.5 years), continuing a relatively stable increase until 97.5 years (289.216/100 000), without significant decline. This pattern discrepancy may reflect the more severe prognostic challenges faced by older patients and the influence of comorbidities, suggesting that even when incidence rates stabilize or decrease in the elderly, mortality remains high.

Figure 2.

Age-period-cohort effect analysis of global colorectal cancer incidence and mortality from 1990 to 2021: (a and e) age effect analysis results. Horizontal axis represents age, vertical axis represents relative risk, (b and f) period effect analysis results. Horizontal axis represents period, vertical axis represents relative risk, with 2004 as the reference period, (c and g) cohort effect analysis results. Horizontal axis represents birth cohort, vertical axis represents relative risk, with 1957 (c) and 1947 (g) as reference cohorts and (d and h) local drift percentages of colorectal cancer risk across different age groups. Horizontal axis represents age group, vertical axis represents local drift percentage, which represents the percentage change in disease risk for each age group compared to the previous age group.

The period effect analysis showed contrasting trends (Figure 2b and f). Incidence risk increased gradually from 1994.5 (RR = 0.973) to 2019.5 (RR = 1.016), while mortality risk decreased from 1994.5 (RR = 1.1) to 2019.5 (RR = 0.881). This “incidence-mortality decoupling” phenomenon highlights the dual effects of “early detection” and “effective treatment” in modern healthcare systems for cancer prevention and control.

The cohort effect analysis further demonstrated the historical evolution of disease burden. Incidence risk showed a “rise-then-fall” nonlinear pattern in more recent birth cohorts, with the 1987 cohort reaching a peak (RR = 1.201), while mortality risk showed a steady decline across birth cohorts, from RR = 1.381 for the 1897 cohort to RR = 0.499 for the 2002 cohort, a decrease of nearly 64%. This marked difference suggests that, although recent cohorts still face relatively high incidence risks, their mortality risks have been substantially controlled, reflecting the cumulative effects of modern prevention and treatment technologies.

Decomposition Analysis of the Disease Burden of Colorectal Cancer

The decomposition analysis of incidence burden from 1990 to 2021 showed a rising global burden of CRC, with population aging contributing the most (47.44%), followed by population growth (44.85%), and epidemiological factors (7.71%). A marked difference was observed between men and women in the drivers of disease burden growth (Figure 3a). In men, the increase in incidence burden was driven by 3 factors: aging (44.28%), population growth (39.21%), and epidemiological factors (16.52%). In contrast, women’s disease burden growth showed a different pattern: aging (54.40%) and population growth (54.00%) had nearly equal contributions, but epidemiological factors had a unique suppressive effect (−8.40%). Different SDI levels exhibited significant regional heterogeneity in the growth of the incidence burden (Figure 3b). In regions with a Middle or higher SDI, the increase in incidence burden was mainly driven by population aging, especially in high-SDI regions, where aging accounted for 72.53%. In contrast, in low-SDI regions, the growth of the burden was almost entirely driven by population growth (100.13%).

Figure 3.

Decomposition analysis of changes in colorectal cancer incidence (a and b) and mortality (c and d) burden from 1990 to 2021.

For mortality burden, the global increase in CRC mortality burden from 1990 to 2021 was driven primarily by population growth (82.06%), followed by aging (58.35%), with epidemiological factors showing a significant suppressive effect (−40.41%). In men, the increase in mortality burden was driven by population growth (69.65%) and changes in the age structure (53.47%), with a relatively weak suppressive effect of epidemiological factors (−23.12%). In women, the increase in mortality burden showed stronger effects of population growth (102.26%) and aging (69.25%), while epidemiological factors had a more significant suppressive effect (−71.51%; Figure 3c). The growth of mortality burden in different SDI regions exhibited a clear shifting pattern (Figure 3d). In low-SDI regions, population growth drove nearly all of the increase (95.82%), with the influence of aging and epidemiological factors being weak. As SDI levels increased, the effect of aging gradually became stronger (from 4.11% in low-SDI regions to 122.95% in high-SDI regions), while the epidemiological factors shifted from positive contributions (16.02% in low-to-middle SDI regions) to a significant suppressive effect (−102.24% in high-SDI regions).

Relationship Between Colorectal Cancer Disease Burden and Socio-Demographic Development Levels

To explore the relationship between CRC disease burden and socio-demographic development levels, this study used Data Envelopment Analysis (DEA) to assess the relationship between the SDI and age-standardized DALY rates of CRC for 204 countries and regions. The DEA frontier line represented the theoretically achievable lowest DALY rate at a given SDI level, providing an objective benchmark for assessing the efficiency of CRC prevention and control across countries.

The frontier analysis using DEA revealed a significant non-linear relationship between CRC DALY rates and SDI (Figure 4a). Specifically, in regions with an SDI below 0.2, the DALY rate decreased significantly as SDI increased. Once SDI exceeded 0.2, the DALY rate stabilized. A 2021 cross-sectional analysis (Figure 4b) further revealed that the actual DALY rates in several countries were significantly higher than the theoretical optimal values corresponding to their SDI levels, indicating considerable room for improvement in disease prevention and control. Countries such as Hungary, Bulgaria, Uruguay, and Greenland had substantial discrepancies between their actual DALY rates and the theoretical optimal values. These findings suggest that even at the same level of socio-economic development, countries may have noticeable differences in CRC prevention and control effectiveness, potentially due to variations in healthcare system organization, targeted prevention strategies, and resource allocation.

Figure 4.

Relationship between socio-demographic index and age-standardized DALY rates for colorectal cancer from 1990 to 2021: (a) scatter plot of socio-demographic index versus age-standardized DALY rates for colorectal cancer across 204 countries from 1990 to 2021. The black solid line represents the fitted frontier production function curve, indicating the minimum DALY rate at a given SDI level. Different colored points represent different years and (b) trends in age-standardized DALY rates for colorectal cancer in selected countries in 2021. The black solid line represents the fitted frontier production function curve, and the colored point-lines represent DALY rate changes for individual countries. Black text indicates the 15 points with the largest distance differences; blue text indicates the 5 countries with the smallest distance differences among low SDI countries; red text indicates the 5 countries with the largest distance differences among high SDI countries.

Health Inequality Analysis

To assess the health equity of the CRC disease burden, this study employed the Slope Index of Inequality (SII) and Concentration Index (CI) to analyze the distribution of age-standardized DALY rates of CRC in 1990 and 2021. SII reflects the absolute level of inequality in disease burden, while CI reflects relative inequality.

The health inequality analysis showed significant differences in CRC DALY rates across countries with varying socio-economic levels. The SII value increased from 437.05 in 1990 to 575.34 in 2021, reflecting a 138.29-point increase, indicating a significant worsening of absolute inequality (Figure 5a). The consistently positive SII reflects the clear socio-economic gradient in CRC DALY rates, where countries with higher SDI typically experience higher DALY rates than countries with lower SDI. Over time, the SII value showed a general upward trend, with a slow increase from 1990 to 1995, an accelerated rise from 1996 to 2005, steady growth from 2006 to 2015, peaking at 609.11 in 2016 to 2019, followed by a sharp decline in 2020 to 2021 (Figure 5b).

Figure 5.

Health inequality analysis of age-standardized DALY rates for colorectal cancer from 1990 to 2021: (a) trend chart of the slope index for age-standardized DALY rates for colorectal cancer from 1990 to 2021. The slope index measures the absolute inequality of colorectal cancer burden, while the concentration index measures relative inequality; larger absolute values of these indices indicate greater inequality, (b) scatter plot of the slope index for age-standardized DALY rates for colorectal cancer from 1990 to 2021, (c) inequality curve for age-standardized DALY rates for colorectal cancer, with the horizontal axis representing cumulative population share and the vertical axis representing cumulative DALY rate share. The dashed line is the reference line; under complete equality, each population group’s share of incidence would equal its population share. Greater deviation from the reference line indicates greater inequality and (d) scatter plot of the concentration index for age-standardized DALY rates for colorectal cancer from 1990 to 2021.

The CI analysis exhibited a relatively stable pattern. The CI value changed slightly from −0.375 in 1990 to −0.371 in 2021 (Figure 5c), with a small change of 0.004. The negative CI values indicated that the CRC disease burden is primarily concentrated in low-SDI regions, but the small decrease in the absolute value of CI over time suggests that the relative inequality has not seen substantial improvement. The CI scatter plot (Figure 5d) revealed a “V” shaped distribution of CI from 1990 to 2021. The CI values fluctuated slightly between −0.375 and −0.373 from 1990 to 1995, gradually declined to a minimum of −0.395 from 1996 to 2007, and then slowly increased from 2008 onward, reaching −0.371 by 2021, indicating limited change in relative inequality.

Prediction of Colorectal Cancer Burden

The BAPC model predictions show different trends in the age-standardized incidence rate (ASIR), age-standardized mortality rate (ASMR), and age-standardized DALY rate (ASDR) for CRC over the next 15 years (Figure 6). Global CRC ASIR is expected to rise slowly (Figure 6a). The incidence increased from 24.206 per 100 000 in 1990 to 25.776 per 100 000 in 2021, a 6.485% increase. The model predicts that this upward trend will continue, with an increase from 26.308 per 100 000 in 2022 to 26.455 per 100 000 in 2036, a small increase of 0.559% during the forecast period, indicating that the disease burden will still increase. Female incidence is expected to decline steadily to 19.322 per 100 000 by 2036 (Figure 6b), a 4.842% decrease compared to 2022, while male incidence is expected to continue rising, reaching 35.02 per 100 000 by 2036, a 5.52% increase compared to 2022 (Figure 6c).

Figure 6.

Prediction of global age-standardized incidence (a, b and c), mortality (d, e and f), and DALY (g, h and i) rates for colorectal cancer by gender from 2022 to 2036 based on BAPC model.

Global CRC ASMR is predicted to continue its downward trend (Figure 6d). The mortality rate decreased from 15.702 per 100 000 in 1990 to 12.518 per 100 000 in 2021, a 20.277% decrease. The model predicts this trend will continue, with a further reduction from 12.588 per 100 000 in 2022 to 11.394 per 100 000 in 2036, a 9.485% decrease during the forecast period. Both male and female ASMR are predicted to decline, but at different rates. Female ASMR is expected to fall to 8.843 per 100 000 by 2036, a decrease of 11.827% from 2022, while male ASMR is expected to reach 14.548 per 100 000 by 2036, a smaller decrease of 6.960% (Figure 6e and f).

Global CRC ASDR is also predicted to continue decreasing (Figure 6g). The DALY rate decreased from 359.035 per 100 000 in 1990 to 284.683 per 100 000 in 2021, a 20.709% reduction. The model predicts a further decline from 288.322 per 100 000 in 2022 to 265.077 per 100 000 in 2036, an 8.063% decrease. By gender, female ASDR is expected to decrease to 202.382 per 100 000 by 2036, a 10.265% decrease compared to 2022, while male ASDR is predicted to decrease to 343.397 per 100 000, a 4.730% decrease (Figure 6h and i).

Overall, these predictions reflect a complex pattern of CRC disease burden changes. The continued rise in incidence, coupled with the decline in mortality and DALY rates, indicates that despite the increasing number of new cases, progress in diagnosis and treatment is improving patient survival. Gender differences are particularly notable, with males facing a higher incidence risk and a more pronounced upward trend, while females not only have a lower incidence but also experience a greater reduction in mortality and DALY rates. This difference may be attributed to lifestyle, dietary habits, screening participation, and biological factors, among other factors.

Regional Distribution of Colorectal Cancer Disease Burden

Comparing the global distribution patterns of CRC disease burden in 1990 and 2021, the study found that there have been significant changes in the geographical distribution of the disease over the past 32 years, though some patterns have remained relatively stable.

In terms of crude incidence rate (Figure 7a and b), high-income regions, primarily in Western Europe, had the highest rates in both 1990 and 2021 (greater than 70/100 000), including countries like Monaco, the Netherlands, and San Marino. On the other hand, sub-Saharan African countries, such as Gambia, Mozambique, and Côte d’Ivoire, maintained very low rates (less than 2/100 000). Notably, Japan joined the group of countries with the highest incidence rates in 2021, reflecting the increasing disease burden in some high-income countries in Asia. Age-standardized incidence rates (Figure 7c and d) were slightly lower than crude rates, but the regional distribution remained similar, with a clear socio-economic gradient. Western developed countries, such as the Netherlands, Monaco, and Bermuda, maintained high rates (greater than 60/100 000), while African nations and regions in the South Pacific, such as Gambia, Mozambique, and Papua New Guinea, had lower rates (less than 5/100 000). This distribution reflects the true disease risk across regions after excluding the influence of population age structure.

Figure 7.

Global distribution of crude incidence rate and age-standardized incidence rate of colorectal cancer in 1990 and 2021: (a) crude incidence rate by country in 1990, (b) crude incidence rate by country in 2021, (c) age-standardized incidence rate by country in 1990 and (d) age-standardized incidence rate by country in 2021.

In terms of mortality burden, crude mortality rates (Figure 8a and b) displayed similar geographical differences, with European countries like Monaco, the Czech Republic, Germany, Bulgaria, and Croatia maintaining high rates (greater than 40/100 000) in both 1990 and 2021, while African countries maintained lower rates (less than 2/100 000). However, age-standardized mortality rates (Figure 8c and d) showed some different trends. For example, Greenland, the Czech Republic, and Hungary had the highest rates in 1990 (greater than 30/100 000), while in 2021, countries like Uruguay, Hungary, and Bulgaria had the highest rates (greater than 25/100 000). Some Western high-income countries, such as the United States and Australia, showed a decline in their rankings, indicating progress in disease treatment. The lowest mortality rates remained concentrated in certain African and Asian regions (less than 4/100 000).

Figure 8.

Global distribution of crude mortality rate and age-standardized mortality rate of colorectal cancer in 1990 and 2021: (a) crude mortality rate by country in 1990, (b) crude mortality rate by country in 2021, (c) age-standardized mortality rate by country in 1990 and (d) age-standardized mortality rate by country in 2021. Darker areas represent countries with higher age-standardized mortality rates, while lighter areas represent countries with lower age-standardized mortality rates.

Overall, from 1990 to 2021, CRC incidence showed a clear socio-economic gradient: high-income countries, especially in Europe and North America, consistently maintained high incidence levels, while low-income countries, particularly those in sub-Saharan Africa, continued to report lower incidence rates. Notably, although high-income countries had higher incidence rates, their age-standardized mortality rates (ASMR) gradually decreased, reflecting improvements in early diagnosis and treatment.

To further explore the relationship between SDI and CRC disease burden, the study analyzed the relationship between SDI and ASIR and ASMR across different countries and regions (Figure 9a-d). At the regional level, the data from 1990 to 2021 revealed that Western Europe, High-income Asia Pacific, and Australasia consistently maintained higher ASIR levels, while South Asia, Central Sub-Saharan Africa, and Western Sub-Saharan Africa remained at lower levels. A notable observation was that Eastern Europe had higher ASIR levels than expected based on its SDI level, suggesting the presence of regional-specific high-risk factors. The fitted curve showed a typical “√”-shaped pattern, with relatively lower ASIR in the low-SDI range (less than 0.5), a sharp increase in the middle SDI range (0.5-0.7), and a peak followed by stabilization or a slight decrease in the high-SDI range (greater than 0.7). When analyzing at the national level (Figure 9b), this relationship was even more pronounced, with countries such as the Netherlands, Monaco, and Bermuda exhibiting significantly higher ASIR than countries with lower SDI, such as Gambia, Papua New Guinea, and Mozambique. Even among high-SDI countries, there were significant differences, such as the ASIR of the Netherlands (69.803/100 000), which was nearly 2.5 times that of Switzerland (28.108/100 000), both of which are high-SDI countries.

Figure 9.

Trends in age-standardized incidence and mortality rates of colorectal cancer across different GBD regions and countries from 1990 to 2021. (a) scatter plot showing the relationship between SDI and incidence rate trends in 21 GBD regions from 1990 to 2021, (b) scatter plot showing the relationship between SDI and incidence rate trends in 204 countries in 2021, (c) scatter plot showing the relationship between SDI and mortality rate trends in 21 GBD regions from 1990 to 2021 and (d) scatter plot showing the relationship between SDI and mortality rate trends in 204 countries in 2021.

The relationship between SDI and ASMR exhibited a different pattern from ASIR (Figure 9c and d). At the regional level (Figure 9c), regions with high SDI, such as High-income North America and Western Europe, had relatively low ASMR. The fitted curve showed an “N”-shaped relationship: in the low-SDI range (less than 0.4), mortality rates increased as SDI increased; in the middle-SDI range (0.4-0.6), they peaked; and in the high-SDI range (greater than 0.6), ASMR declined significantly. At the national level (Figure 9d), this non-linear relationship became clearer. Countries with medium to high SDI, such as Uruguay, Hungary, and Bulgaria, had higher ASMR than most high-SDI countries, suggesting that despite their higher development levels, these countries may still have deficiencies in prevention and treatment strategies. Conversely, low-SDI countries, such as Gambia and Bangladesh, had low ASMR despite limited healthcare resources, likely due to the relatively low overall disease burden.

Impact of High BMI on Colorectal Cancer

To comprehensively assess the impact of high BMI on CRC, the study performed a spatiotemporal analysis of data from global and 21 GBD regions (Figure 10). The results showed that high BMI has become an important risk factor for CRC mortality and DALYs, with its attributable proportion increasing significantly over the past 32 years.

Figure 10.

Contribution of high body mass index (BMI) to the burden of colorectal cancer. (a) proportion (%) of colorectal cancer deaths and DALYs attributable to high BMI in 2021 globally and across 21 GBD regions and (b) trends in the proportion (%) of colorectal cancer deaths and DALYs attributable to high BMI globally from 1990 to 2021.

In 2021, high BMI was responsible for 9.5% of CRC-related deaths globally and 9.7% of DALYs (Figure 10a). Significant regional differences were observed. In terms of mortality, North America (High-income) had the highest attributable proportion (15.0%), followed by North Africa and the Middle East (14.7%) and Eastern Europe (14.0%). South Asia and Eastern Sub-Saharan Africa had the lowest attributable proportions (4.4% for both), with Southeast Asia also showing relatively low proportions (4.8%).

From a time trend perspective (Figure 10b), the attributable proportions of high BMI to CRC mortality and DALYs steadily increased from 1990 to 2021. The mortality attributable proportion rose from 7.32% in 1990 to 9.46% in 2021, an increase of 29.2%, while the DALYs attributable proportion increased from 7.15% to 9.65%, a 35.0% rise. This consistent upward trend highlights the growing impact of high BMI as a risk factor for CRC.

Mendelian Randomization Analysis for BMI and Colorectal Cancer Risk

A total of 422 SNPs significantly associated with BMI and independent from 1 another were selected as instrumental variables from GWAS databases and Finnish biobank data. The results showed a significant positive association between genetically predicted BMI and CRC risk (OR = 1.51, 95% CI: 1.01-2.24, P = .043) using the IVW method (Figure 11). This indicates that for every 1-unit increase in BMI, the risk of CRC increases by approximately 51%. The weighted median method yielded a similar effect estimate (OR = 1.68, 95% CI: 0.86-3.26), but it was not statistically significant (P = .126). MR-Egger regression suggested a potential positive association (OR = 1.91, 95% CI: 0.65-5.63), though it was not statistically supported (P = .240). The 2 mode methods showed different effect estimates, with the weighted mode method yielding estimates closer to the IVW method (OR = 1.80, 95% CI: 0.49-6.65), while the simple mode method estimated an effect closer to zero (OR = 1.03, 95% CI: 0.10-10.13). The MR-Egger intercept did not show significant directional pleiotropy (P > .05), suggesting that the instrumental variables likely do not suffer from systematic pleiotropy bias. Despite variations in statistical significance across different sensitivity methods (weighted median, mode methods), the direction of effect was consistent with the main analysis, strengthening confidence in the causal inference.

Figure 11.

Forest plot of the association between BMI and colorectal cancer risk revealed by Mendelian randomization analysis.

Discussion

The global burden of CRC has exhibited a complex and evolving landscape from 1990 to 2021. Our comprehensive research reveals that while ASIR of CRC increased by 6.485% during this period, the ASMR and ASDR showed significant declines of 20.277% and 20.709%, respectively. This trend is consistent with previous study.^16-18 This divergence between incidence and mortality trends highlights both the challenges and successes in CRC control worldwide. Notably, high-SDI regions, such as North America and Western Europe, have seen CRC burden primarily driven by population aging. In contrast, low-SDI regions, like sub-Saharan Africa, are more influenced by population growth. This striking difference is largely due to the rapid adoption of Western lifestyles in economically transitioning regions, which often involves dietary changes and reduced physical activity.¹⁹ We have observed a clear gender difference, with an increase in incidence in men and a decrease in incidence in women. In terms of health inequalities, the increase in SII is 138.29, which underscores the need for equitable prevention and control efforts worldwide.

The rising global incidence of CRC can be attributed to multiple interconnected factors.²⁰ Higher adherence to a Western dietary pattern, particularly increased intake of refined grains and sugar-sweetened beverages, correlated with greater risks of CRC recurrence, disease progression, and mortality in RFS and DFS analyses.²¹ According to the results from the World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR), physical inactivity and weight gain are associated with an increased risk of CRC.²² Our Mendelian randomization analysis provides robust evidence supporting the causal relationship between elevated BMI and increased CRC risk (OR = 1.51), aligning with previous observational studies. This metabolic risk factor has been particularly pronounced in rapidly developing regions such as East Asia, where CRC incidence has risen sharply.²³ However, in high-income countries, widespread screening programs have paradoxically contributed to rising incidence rates through increased detection of early-stage and precancerous lesions.²⁴ In countries and regions with medium to low SDI, the lack of widespread access to screening technologies such as endoscopy results in insufficient screening coverage and missed diagnoses.²⁵ Our decomposition analysis shows that these gains brought by economic development have been partially offset by demographic changes, with population aging alone accounting for 47.44% of the increase in incidence burden. This spotlights the ongoing challenge posed by shifting population structures, particularly in developed countries where aging populations are driving absolute increases in CRC cases despite stable or declining age-specific rates.²⁶

The encouraging declines in CRC mortality and DALY rates observed globally reflect substantial progress in cancer control, though the benefits have not been equally distributed across regions. Therapeutic advances, including the development of more effective chemotherapy regimens and targeted therapies, have significantly improved survival outcomes, particularly in high-SDI countries.²⁷ A population-based study from Taiwan showed that CRC patients who underwent minimally invasive surgery (such as laparoscopic surgery) had a lower 5-year mortality rate compared to those who underwent traditional open surgery (HR 0.60, 95% CI 0.58-0.62; P < .001).²⁸ Simultaneously, the expansion of fecal occult blood testing (FOBT) and colonoscopy screening programs has enabled early identification and management of CRC, contributing to the mortality reductions we observed.^29,30 In recent years, the continuous optimization of public health policies in various countries has significantly enhanced the capacity for early detection and treatment of CRC.³¹ For example, in 2018, the American Cancer Society (ACS) lowered the recommended screening age from 50 to 45 years old. This measure aims to detect CRC at an earlier stage, especially in addressing the differences in incidence rates among various racial subgroups.³²

Analysis of regional patterns reveals striking discrepancies in CRC burden that reflect underlying differences in risk factor profiles and healthcare system capacity across regions. The exceptionally high ASIR in High-income Asia Pacific (44.89/100 000) compared to South Asia (5.65/100 000), which may be due to the rapid changes in lifestyle in economically transitioning regions that have dramatically impacted cancer epidemiology.³³ The persistence of traditional dietary patterns and higher levels of physical activity in South Asia may provide some protection against CRC development, while the adoption of Western lifestyles in East Asia appears to be driving increased risk.³⁴ These regional variations underscore the need for tailored prevention strategies that account for local epidemiological transitions and cultural contexts.

Age-period-cohort analysis provides valuable insights into the temporal evolution of CRC risk across generations. The strong age effect we observed, with risk increasing exponentially until peaking in the eighth decade of life, aligns with the multistep accumulation of mutations in intestinal stem cells and immune senescence-related microenvironment alterations.^35,36 The period effect analysis shows a marked surge in incidence rates during the 1990s, corresponding to the implementation of initial screening programs with limited sensitivity.³⁷ Early FOBT demonstrated a concerning 62% false-negative rate, creating a potentially dangerous false sense of security in CRC diagnosis.³⁸ Oppositely, the observed mortality decline post-2000 likely stems from groundbreaking advances in FOLFOX chemotherapy regimens and molecular targeted therapies.^39,40 Cohort comparisons illustrate significant generational differences, with higher mortality rates among cohorts born before 1920 due to delayed diagnosis resulting from limited medical resources. In contrast, cohorts born after 1950 benefited from advances in screening technologies and increased health awareness.⁴¹ Most concerning is that local drift analysis identified the fastest incidence growth in the 40 to 60 age group, likely driven by the younger onset of metabolic syndrome, thus highlighting the need for targeted early screening interventions in specific populations as recommended by the latest guidelines.^42,43

Our frontier analysis revealed that while high SDI is generally associated with lower mortality rates through better healthcare access, the relationship with incidence is more nuanced.⁴⁴ The characteristic “√-shaped” pattern of CRC incidence across SDI levels demonstrates those nations in the upper-middle SDI range (0.5-0.8) experience elevated risk due to Westernized lifestyle transitions during economic development, before achieving effective prevention and early detection capabilities at very high SDI levels (>0.8).⁴⁵ The “N-shaped” mortality curve emerges through distinct mechanisms across SDI strata: the insufficient diagnostic capacity leads to underdiagnosis and an artificially low mortality rate, middle-SDI zones exhibit mortality peaks where healthcare infrastructure lags behind rapidly escalating risk factors, while high-SDI populations demonstrate declining rates through timely diagnostic interventions and early therapeutic management.^46,47 This carries significant policy implications, suggesting that rapidly transitioning middle-SDI nations should prioritize preventive interventions. A modeling study evaluating long-term benefits and cost-effectiveness of various CRC screening strategies in China (2020-2060) demonstrated that implementing a stepwise screening approach combining fecal immunochemical testing (FIT) with diagnostic colonoscopy could effectively reduce CRC disease burden while conserving endoscopy capacity and costs.⁴⁸ In high-SDI regions, optimizing comprehensive care for elderly CRC patients is pivotal for enhancing both survival outcomes and quality of life. Implementing a multidimensional approach encompassing comprehensive geriatric assessment (CGA), personalized treatment protocols, laparoscopic surgical techniques, and geriatric liaison services has demonstrated significant improvements in clinical efficacy and patient-reported outcomes for this vulnerable population.^49,50

This study pioneers the combined application of Das Gupta decomposition, DEA frontier analysis, and BAPC modeling to comprehensively dissect the drivers of CRC burden across demographic, region, and socioeconomic dimensions. Our health equity lens quantifies global inequalities through SII and CI, generating policy-relevant advice for advancing WHO’s “Universal Health Coverage” agenda.⁵¹ Furthermore, we strengthen causal inference by incorporating Mendelian randomization, which provides Class III evidence for the BMI-CRC association, substantiating obesity control as a public health priority. However, certain limitations should be acknowledged. Data quality exhibits significant regional heterogeneity, with disease burden assessments in low-SDI countries remaining potentially underestimated due to underdeveloped national cancer registry systems.⁵² While our analysis treated CRC as 1 category, significant epidemiological distinctions exist between colon and rectal cancer subtypes that warrant careful consideration. For example, Alexander et al. found that the correlation between red meat intake and rectal cancer was slightly stronger than that with colon cancer.⁵³ Moreover, although Mendelian randomization reduces confounding, potential pleiotropy of genetic variants may affect causal estimates, necessitating further validation with larger sample sizes of GWAS data.

Conclusions

In conclusion, our findings paint a picture of both progress and persistent challenges in global CRC control. While therapeutic and screening advances have driven impressive mortality reductions, rising incidence and growing health inequalities remind us that much work remains. Future efforts should focus on regionally tailored strategies: targeted intensification of CRC screening for elderly populations in high-SDI regions, widespread early screening in low-to-middle SDI regions, and global collaboration to reduce health inequalities. Additionally, as obesity is a modifiable risk factor, controlling obesity rates should be incorporated as a core objective in national Non-Communicable Disease (NCD) prevention strategies prevention and control. The insights presented in this study provide valuable epidemiological evidence and actionable measures to support global CRC prevention and control efforts.

Supplemental Material

sj-docx-1-cix-10.1177_11769351261454666 – Supplemental material for Widening Health Inequality and Causal Metabolic Drivers in Global Colorectal Cancer: A Multi-Dimensional Study

Supplemental material, sj-docx-1-cix-10.1177_11769351261454666 for Widening Health Inequality and Causal Metabolic Drivers in Global Colorectal Cancer: A Multi-Dimensional Study by Song Gao, Mingying Peng, Susu Han, Yufei Tang, Juan Ren, Yi Cen, Fenggang Hou, Lianyu Chen and Xiaoling Yin in Cancer Informatics

Footnotes

Acknowledgements

We gratefully acknowledge the GBD 2021 collaborators for their invaluable contributions, which were instrumental in the successful completion of this study.

Abbreviations

APC: annual percentage change; ASDR: Age-Standardized DALY Rate; ASIR: Age-Standardized Incidence Rate; ASMR: Age-Standardized Mortality Rate; BAPC: Bayesian Age-Period-Cohort; BMI: Body Mass Index; CI: Concentration Index; CRC: Colorectal Cancer; DALYs: Disability-Adjusted Life Years; DEA: Data Envelopment Analysis; EAPC: Estimated Annual Percentage Change; FDH: Free Disposal Hull; FOBT: Fecal Occult Blood Test; GBD: Global Burden of Disease; GWAS: Genome-Wide Association Study; IHME: Institute for Health Metrics and Evaluation; IVW: Inverse Variance Weighted; LOWESS: Locally Weighted Scatterplot Smoothing; MR: Mendelian Randomization; NCD: Non-Communicable Disease; SDI: Socio-Demographic Index; SII: Slope Index of Inequality; SNP: Single Nucleotide Polymorphism; UI: Uncertainty Interval.

ORCID iDs

Song Gao

Mingying Peng

Ethical Considerations

As the GBD study did not involve any personal or sensitive information, no ethical approval was necessary for the execution of this study.

Author Contributions

SG and FH contributed to the study design, data collection, interpretation, and analysis, and drafted and revised the manuscript. MP contributed equally to the research, with SH and YL responsible for data collection and initial analysis. YL responsible for literature review and study validation. LC provided strategic guidance, supervised the research, and contributed to interpreting results. SG, MP and SH were the main writers of the manuscript. All authors contributed to the article and approved the submitted version.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by National Natural Science Foundation of China (grants 81873122); Fudan University DIGAOJIAN Project (No. DGF601020-1); Pudong New District Traditional Chinese Medicine Weak Discipline Development Project (PDZY-2020-0406).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

All GBD 2021 data used in this study are publicly available via the IHME GBD Results Tool (). No individual-level data were used or generated in this study. Analytical code is available from the corresponding author upon reasonable request.*

Supplemental Material

Supplemental material for this article is available online.

Use of Artificial Intelligence

Large language model tools (eg, Doubao) were used solely for language clarity during manuscript preparation. No scientific data were generated, analyzed, or modified using AI tools. The authors take full responsibility for the integrity and accuracy of all reported results.

References

Sung

Ferlay

Siegel

, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209-249. doi:10.3322/caac.21660

Morgan

Arnold

Gini

, et al. Global burden of colorectal cancer in 2020 and 2040: incidence and mortality estimates from GLOBOCAN. Gut. 2023;72(2):338-344. doi:10.1136/gutjnl-2022-327736

Kuipers

Grady

Lieberman

, et al. Colorectal cancer. Nat Rev Dis Primers. 2015;1:15065. doi:10.1038/nrdp.2015.65

Johnson

Wei

Ensor

, et al. Meta-analyses of colorectal cancer risk factors. Cancer Causes Control: CCCI. 2013;24(6):1207-1222.

Huxley

Ansary-Moghaddam

Clifton

Czernichow

Parr

Woodward

The impact of dietary and lifestyle risk factors on risk of colorectal cancer: a quantitative overview of the epidemiological evidence. Int J Cancer. 2009;125(1):171-180. doi:10.1002/ijc.24343

Thanikachalam

Khan

Colorectal cancer and nutrition. Nutr. 2019;11(1):164. doi:10.3390/nu11010164

Bardou

Barkun

Martel

Obesity and colorectal cancer. Gut. 2013;62(6):933-947. doi:10.1136/gutjnl-2013-304701

Schreuders

Ruco

Rabeneck

, et al. Colorectal cancer screening: a global overview of existing programmes. Gut. 2015;64(10):1637-1649. doi:10.1136/gutjnl-2014-309086

Shaukat

Kahi

Burke

Rabeneck

Sauer

Rex

DK.

ACG clinical guidelines: colorectal cancer screening 2021. Am J Gastroenterol. 2021;116(3):458-479. doi:10.14309/ajg.0000000000001122

10.

Schliemann

Ramanathan

Matovu

, et al. The implementation of colorectal cancer screening interventions in low-and middle-income countries: a scoping review. BMC Cancer. 2021;21(1):1125. doi:10.1186/s12885-021-08809-1

11.

Van Cutsem

Cervantes

Adam

, et al. ESMO consensus guidelines for the management of patients with metastatic colorectal cancer. Ann Oncol Off J Eur Soc -med Oncol. 2016;27(8):1386-1422. doi:10.1093/annonc/mdw235

12.

Arnold

Sierra

Laversanne

Soerjomataram

Jemal

Bray

Global patterns and trends in colorectal cancer incidence and mortality. Gut. 2017;66(4):683-691. doi:10.1136/gutjnl-2015-310912

13.

Vos

Lim

Abbafati

, et al. Global burden of 369 diseases and injuries in 204 countries and territories, 1990-2019: a systematic analysis for the global burden of disease study 2019. Lancet. 2020;396(10258):1204-1222. doi:10.1016/s01406736(20)30925-9

14.

Fitzmaurice

Abate

Abbasi

, et al. Global, regional, and national cancer incidence, mortality, years of life lost, years lived with disability, and disability-adjusted life-years for 29 cancer groups, 1990 to 2017: a systematic analysis for the global burden of disease study. JAMA Oncol. 2019;5(12):1749-1768. doi:10.1001/jamaoncol.2019.2996

15.

Davey Smith

Hemani

. Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet. 2014;23:R89-R98.

16.

Skrivankova

Richmond

Woolf

BAR

, et al. Strengthening the reporting of observational studies in epidemiology using mendelian randomization: the STROBE-MR statement. JAMA. 2021;326:1614-1621.

17.

Skrivankova

Richmond

Woolf

BAR

, et al. Strengthening the reporting of observational studies in epidemiology using mendelian randomization (STROBE-MR): a Korean translation of explanation and elaboration. BMJ. 2025;48:e68.

18.

Zhou

Yang

Zhao

, et al. Evolving landscape of colorectal cancer: global and regional burden, risk factor dynamics, and future scenarios (the global burden of disease 1990-2050). Ageing Res Rev. 2025;104:102666. doi:10.1016/j.arr.2025.102666

19.

Popkin

SW.

The nutrition transition to a stage of high obesity and noncommunicable disease prevalence dominated by ultra-processed foods is not inevitable. Obes Rev. 2022;23(1):e13366. doi:10.1111/obr.13366

20.

Shi

Wang

Chen

, et al. The association of life’s essential 8 scores trajectory patterns with the risk of all cancer types. Sci Rep. 2025;15(1):9600. doi:10.1038/s41598-025-94009-x

21.

Fretwell

Louca

Cohoon

, et al. Still too little evidence: the role of diet in colorectal cancer survivorship - a systematic review. Crit Rev Food Sci Nutr. 2025;65(16):3173-3185. doi:10.1080/10408398.2024.2360068

22.

Bray

Ferlay

Soerjomataram

Siegel

Torre

Jemal

Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2018;68(6):394-424. doi:10.3322/caac.21492

23.

Sun

Zheng

Metabolic diseases in the East Asian populations. Nat Rev Gastroenterol Hepatol. 2025;22:500-516. doi:10.1038/s41575-025-01058-8

24.

Xia

Zhang

Zhuang

Cost effectiveness of colorectal cancer screening strategies in middle- and high-income countries: a systematic review. 2025;40(3):584-598. doi:10.1111/jgh.16882

25.

Rozani

Lykoudis

PM.

Overcoming geographical and socioeconomic limitations in colorectal cancer screening. World J Gastrointest Oncol. 2024;16(5):1683-1689. doi:10.4251/wjgo.v16.i5.1683

26.

O’Hanlon

. Cancer and older people. A research collection and commentary on the theme. Age Ageing. 2025;54(5):afaf113.

27.

Shaham

Vij

Tripathi

MK.

Advances in targeted and chemotherapeutic strategies for colorectal cancer: current insights and future directions. Biomedicines. 2025;13(3):642. doi:10.3390/biomedicines13030642

28.

Yeh

Lai

Teng

Huang

Liu

CJ.

The impact of surgical volume on outcomes in newly diagnosed colorectal cancer patients receiving definitive surgeries. Sci Rep. 2024;14(1):8227. doi:10.1038/s41598-024-55959-w

29.

Chowdhury

Hone

KGMS

Prévost

, et al. Optimizing fecal occult blood test (FOBT) colorectal cancer screening using gut bacteriome as a biomarker. Clin Colorectal Cancer. 2024;23(1):22-34.e2. doi:10.1016/j.clcc.2023.10.004

30.

Castells

Quintero

Bujanda

, et al. Effect of invitation to colonoscopy versus faecal immunochemical test screening on colorectal cancer mortality (COLONPREV): a pragmatic, randomised, controlled, non-inferiority trial. Lancet. 2025;405(10486):1231-1239. doi:10.1016/s0140-6736(25)00145-x

31.

Bacchus

McKee

Gwede

Cogle

CR.

Implementation of cancer plans in the United States: a review. Healthcare. 2021;9(3):291. doi:10.3390/healthcare9030291

32.

Glaser

Crabtree-Ide

McNulty

, et al. Improving guideline-recommended colorectal cancer screening in a federally qualified health center (FQHC): implementing a patient navigation and practice facilitation intervention to promote health equity. Int J Environ Res Public Health. 2024;21(2):126.

33.

Chu

Lin

Croker

, et al. Dietary-lifestyle patterns and colorectal cancer risk: Global Cancer Update Programme (CUP Global) systematic literature review. Am J Clin Nutr. 2025;121(5):986-998. doi:10.1016/j.ajcnut.2025.01.014

34.

Ashari

Abd Rashid

Mohd Razif

Yeong Yeh

Jan Mohamed

HJ.

Diet is linked to colorectal cancer risk among Asian adults: a scoping review. Malays J Med Sci. 2023;30(3):8-31. doi:10.21315/mjms2023.30.3.2

35.

Zhu

Chen

, et al. Nutrient-sensing alteration leads to age-associated distortion of intestinal stem cell differentiating direction. Nat Commun. 2024;15(1):9243. doi:10.1038/s41467-024-53675-7

36.

Lin

Zhang

, et al. Age, sex, and specific gene mutations affect the effects of immune checkpoint inhibitors in colorectal cancer. Pharmacol Res. 2020;159:105028. doi:10.1016/j.phrs.2020.105028

37.

Austoker

Cancer prevention in primary care screening for colorectal cancer. BMJ. 1994;(6951)):382-386. doi:10.1136/bmj.309.6951.382

38.

Schnell

Aranha

Sontag

, et al. Fecal occult blood testing: a false sense of security? Surgery. 1994;116(4):798-802.

39.

André

Boni

Navarro

, et al. Improved overall survival with oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment in stage II or III colon cancer in the MOSAIC trial. J Clin Oncol. 2009;27(19):3109-3116. doi:10.1200/jco.2008.20.6771

40.

Van Cutsem

Köhne

Hitre

, et al. Cetuximab and chemotherapy as initial treatment for metastatic colorectal cancer. N Engl J Med. 2009;360(14):1408-1417. doi:10.1056/NEJMoa0805019

41.

Kwan

ASH

Uwishema

Mshaymesh

, et al. Advances in the diagnosis of colorectal cancer: the application of molecular biomarkers and imaging techniques: a literature review. Ann Med Surg. 2025;87(1):192-203. doi:10.1097/ms9.0000000000002830

42.

Chen

Zheng

Zong

, et al. Metabolic syndrome, metabolic comorbid conditions and risk of early-onset colorectal cancer. Gut. 2021;70(6):1147-1154. doi:10.1136/gutjnl-2020321661

43.

Davidson

Barry

Mangione

, et al. Screening for colorectal cancer: US preventive services task force recommendation statement. JAMA. 2021;325(19):1965-1977. doi:10.1001/jama.2021.6238

44.

Klimeck

Heisser

Hoffmeister

Brenner

Colorectal cancer: a health and economic problem. Best Pract Res Clin Gastroenterol. 2023;66:101839.

45.

Walker

Jani

Liu

, et al. Does a “Western Lifestyle” confer a higher burden of colorectal cancer? A comparison of EU15+ countries versus global trends between 1990 and 2019. Cancers. 2024;16(12):2277. doi:10.3390/cancers16122277

46.

Konduru

Dahia

Szabo

Barreto

SG.

Evolving dynamics of colorectal cancer in high socio-demographic regions. Cancer Control. 2025;32:10732748251321672.

47.

Khan

Lengyel

CG.

Challenges in the management of colorectal cancer in low-and middle-income countries. Cancer Treat Res Commun. 2023;35:100705. doi:10.1016/j.ctarc.2023.100705

48.

Luo

Yan

, et al. Evaluation of long-term benefits and cost-effectiveness of nation-wide colorectal cancer screening strategies in China in 2020-2060: a modelling analysis. Lancet Reg Health West Pac. 2024;51:101172.

49.

O’Donnell

CDJ

Hubbard

Jin

. Updates on the management of colorectal cancer in older adults. Cancers. 2024;16(10):1820. doi:10.3390/cancers16101820

50.

González-Senac

Mayordomo-Cava

Macías-Valle

, et al. Colorectal cancer in elderly patients with surgical indication: state of the art, current management, role of frailty and benefits of a geriatric liaison. Int J Environ Res Public Health. 2021;18(11):6072. doi:10.3390/ijerph18116072

51.

Foster

Benjamin

Parwani

Perehudoff

Universal health coverage and social protection: evolution and future opportunities for global health law and equity. J Law Med Ethics. 2025;53(S1):35-39.

52.

Tripathee

MacLennan

Poobalan

Omar

Guntupalli

AM.

The role of hospital-based cancer registries (HBCRs) as information systems in the delivery of evidence-based integrated cancer care: a scoping review. Health Syst. 2024;13(3):177-191. doi:10.1080/20476965.2023.2216749

53.

Alexander

Weed

Miller

Mohamed

MA.

Red meat and colorectal cancer: a quantitative update on the state of the epidemiologic science. J Am Coll Nutr. 2015;34(6):521-543. doi:10.1080/07315724.2014.992553

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.04 MB

0.00 MB