Key Points
Question
Are colorectal cancer (CRC) risk factors identified in high-income populations also associated with CRC risk in adults in Nigeria?
Findings
In this case-control study of 2063 individuals in Nigeria, including 632 patients with CRC matched with 1431 controls, 5 factors significantly associated with CRC were identified: family history of cancer, adult height, body size or type, processed meat intake, and physical activity. There were no associations with diabetes, smoking, or intakes of red meat, fruit, vegetables, fiber, or alcohol.
Meaning
These findings suggest certain risk factors are shared across populations, representing a critical first step in defining CRC risk profiles to tailor prevention and early detection efforts in the region.
This case-control assesses whether 13 colorectal cancer (CRC) risk factors identified in high-income populations are associated with CRC in adults in Nigeria.
Abstract
Importance
Colorectal cancer (CRC) incidence is increasing in sub-Saharan Africa, including Nigeria, where more than half of patients die within 1 year due to late diagnosis and limited treatment. Without ready access to screening, identifying risk factors is critical. However, regional data are scarce, and it is unclear whether risk factors from high-income populations apply to Nigerians, who may have distinct clinicodemographic patterns and tumor characteristics.
Objective
To assess whether 13 CRC risk factors identified in high-income populations are associated with CRC in adults in Nigeria.
Design, Setting, and Participants
This multicenter case-control study was conducted in 6 hospitals across 3 of Nigeria’s 6 geopolitical zones, with recruitment from April 2020 to September 2024. Participants were patients newly diagnosed with CRC and cancer-free controls matched on age, sex, location, education, and recruitment site. Data were analyzed from April 2023 to November 2025.
Exposures
Family history of cancer, anthropometric factors (adult height, body mass index [BMI], and somatotype or body size over the life course), physical activity, and dietary and lifestyle factors (processed meat, red meat, fruit, vegetables, fiber, alcohol, smoking, and diabetes).
Main Outcomes and Measures
The associations of risk factors with CRC were assessed using odds ratios (ORs) and 95% CIs, estimated using multivariable unconditional logistic regression.
Results
Among 2063 participants (1103 [53.5%] male; median [IQR] age, 54 [42-64] years) enrolled, most had secondary education or higher (1756 participants [85.2%]), lived in urban areas (1817 participants [88.1%]), and were Yoruba (1524 participants [73.9%]). Cases were older than controls (median [IQR] age, 56 [45-66] vs 52 [41-63] years) but were otherwise similar. Family history of cancer (OR, 1.55; 95% CI, 1.06-2.26), adult height (OR per 5-cm increase, 1.09; 95% CI, 1.00-1.20), body size or type over the life course (tertile 3 vs tertile 1: OR, 1.88; 95% CI, 1.30-2.72), and processed meat intake (lowest vs highest tertile of intake: OR, 1.43; 95% CI, 1.07-1.91) were positively associated with CRC risk, and an inverse U-shaped association was identified between physical activity and CRC risk (OR vs first quartile of activity, quartile 2: 0.63; 95% CI, 0.44-0.88; quartile 3: 0.52; 95% CI, 0.34-0.79; quartile 4: 0.77; 95% CI, 0.58-1.02), consistent with data from high-income populations. BMI was inversely associated with CRC risk (OR vs BMI 18.5-24.9, BMI <18.5: 2.63; 95% CI, 1.79-3.86; BMI ≥30: 0.42; 95% CI, 0.28-0.63), likely reflecting prediagnostic weight loss. No associations with diabetes, smoking, or intakes of red meat, fruit, vegetables, fiber, or alcohol were observed, likely due to low prevalence or variability.
Conclusions and Relevance
This case-control study of CRC risk factors in adults in Nigeria found CRC risk patterns similar to those observed in high-income populations. These insights offer immediate opportunity to tailor education, prevention, and early detection in the region, especially as economic development leads to lifestyles more similar to high-income populations. These data provide a foundation for prospective research and investigations in other African populations.
Introduction
Colorectal cancer (CRC) is a growing global health problem, with approximately 1.9 million new diagnoses and 1 million deaths annually.1 Africa is projected to face the sharpest global surge by 2050—an estimated 140% increase in new CRC cases and a 155% increase in deaths.2 These trends are apparent in Nigeria, Africa’s most populous nation, where most patients (88%) are diagnosed at an advanced stage, have limited treatment options, and die within 1 year of diagnosis.3
Recognizing these challenges, Nigeria’s Cancer Control Plan 2023-20274 and The Lancet Oncology Commission5 emphasize the urgent need for primary cancer prevention. While populationwide screening programs have effectively reduced CRC in high-income countries (HICs),6,7 major barriers to their implementation remain in sub-Saharan Africa, including limited infrastructure, access, and affordability.5,8 In a community-based study of 346 adults in Nigeria, only 5% of participants eligible for CRC screening had ever been screened.9
To design effective prevention strategies in this resource-limited setting, it is essential to identify individuals at the highest risk. Decades of research in HICs have identified a panel of CRC risk factors, including family history of cancer, height, excess body weight, physical activity, smoking, alcohol intake, diabetes, and dietary intake of certain food groups (eg, red meat), among others.10,11,12,13,14 However, it remains unclear whether these apply to the Nigerian population, which is typically younger, more likely to present with rectal disease, and may exhibit a 2- to 3-fold higher frequency of microsatellite instability than US patients.15 Furthermore, conventional risk factors, such as smoking, alcohol use, obesity, and diabetes, are far less prevalent in Nigeria.9
We conducted a large, multicenter study of CRC risk factors in adults in Nigeria to determine whether 13 risk factors identified in HICs were associated with CRC risk in our population. Our study spanned 6 hospitals across 3 of the country’s 6 geopolitical zones and represents a critical first step toward developing targeted education, risk reduction, and screening strategies tailored to the region.
Methods
Study Design and Participants
For this case-control study, we leveraged the African Research Group for Oncology (ARGO) infrastructure to conduct a multicenter study of CRC risk factors at 6 hospitals, with participants recruited from April 2020 to September 2024 (eTable 1 in Supplement 1). This study was approved by institutional review boards at all study sites. Researchers reviewed study materials with all participants, and all participants provided written informed consent. This study is reported following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for case-control studies.
Cases were adults (aged ≥18 years) with newly diagnosed, histologically confirmed CRC. Cancer-free control individuals (aged ≥18 years) were family or friends of patients undergoing general surgery with benign, noncancerous conditions or patients with benign conditions unrelated to CRC risk factors (eg, routine check-up) to reduce selection bias. When possible, we matched 2 controls to each case by age (±5 years), sex (male, female), education (none, primary, secondary or technical school, tertiary), geographic area (rural, urban), and recruitment hospital. Individuals with a personal history of cancer or gastrointestinal diseases or symptoms were ineligible to serve as a control.
Data Collection
We administered a comprehensive epidemiologic questionnaire developed with input from local stakeholders and adapted validated instruments (eMethods in Supplement 1). Participant tribe was collected via self-report and classified as Hausa, Ibo, Yoruba, other (eg, Ijaw, Egun, Edo), or missing and assessed because there are large sociocultural, dietary, and lifestyle variations across groups. Due to challenges in collecting prediagnostic body weight, we assessed body size using a validated somatotype scale16 at ages 10, 20, 30, 40, and 50 years and at baseline. The somatotype scale is a set of 9 pictorial body diagrams ranging from 1 (smallest) to 9 (largest)16; participants reviewed the diagrams and selected the diagram that best depicted their body type at each age. We defined tertiles calculated in male and female controls separately indicating leaner, intermediate, and larger body sizes in our population for each time point and across all ages (excluding baseline). We assessed physical activity by adapting the validated Global Physical Activity Questionnaire.17 We included a validated semiquantitative food frequency questionnaire to assess diet.18 Given that most patients present with late-stage disease,3 we asked cases to report their exposures when they were well to reduce the influence of subclinical or prediagnostic disease on associations.
Exposures and Potential Confounders
We assessed 13 risk factors that have been identified primarily in populations from HICs, including family history of any cancer (yes, no), adult height (per 5-cm increase), body mass index (BMI, calculated as weight in kilograms divided by height in meters squared; classified as <18.5, 18.5 to <25, 25 to <30, and ≥30), physician-diagnosed diabetes (yes, no), physical activity (quartiles of metabolic equivalent of task [MET] hours per week), smoking (ever, never), alcohol (ever, never), and dietary intake (red meats, fruit, vegetables, and fiber in quartiles of grams per day calculated in male and female controls separately and processed meats in tertiles of grams per week calculated in male and female controls separately, with tertiles used due to low consumption) (eMethods in Supplement 1). BMI is not an ideal measure of prediagnostic adult body size in our population because many participants did not own scales and patients had already lost weight prior to cancer diagnosis (when they were weighed). Thus, as our primary measure of prediagnostic body size, we calculated a composite somatotype score using the mean recalled somatotype across ages, excluding baseline (when many patients had already lost weight), defined continuously and in tertiles (defined in our healthy controls for each age, and across all ages [excluding baseline]). As potential confounders for adjustment, we included age, sex, education, geographic area, recruitment site, tribe, religion, occupation, use of nonsteroidal anti-inflammatory drugs, total energy intake, and a modified unweighted Global Multidimensional Poverty Index (MDPI) as a proxy of socioeconomic status (eMethods in Supplement 1).
Statistical Analysis
In descriptive analysis, we used Wilcoxon rank sum for continuous variables and χ2 and Fisher exact tests for categorical variables to examine differences by case-control status. We used multivariable unconditional logistic regression to estimate odds ratios (ORs) and 95% CIs. Minimally adjusted models included the exposure of interest and matching factors; fully adjusted models additionally included all 13 risk factors and tribe, religion, occupation, nonsteroidal anti-inflammatory drug use, total energy intake (kilocalories per day), and modified global MDPI. We imputed missing data using multiple imputation by chained equations (eMethods in Supplement 1).19
We explored whether associations differed by sex, age (<50 years vs ≥50 years), global MDPI, religion (Christian vs Muslim or other), or type of control (friend or relative of patient vs patient). We also explored associations by tumor location (colon, rectum, proximal, distal) using minimally adjusted analysis for cases with this information available.
We performed sensitivity analyses to assess the robustness of associations (eMethods in Supplement 1). We restricted analysis to participants without recent weight loss and to participants who did not think risk factors affected a person’s chance of developing cancer to evaluate potential reverse causation and recall bias, respectively. We used the Wald test to test for linear trends across ordinal variables. All tests of statistical significance were 2-sided at P < .05. All analyses were performed using R statistical software version 4.3.1 (R Project for Statistical Computing). Data were analyzed from April 2023 to November 2025.
Results
A total of 2063 individuals (1103 [53.5%] male; median [IQR] age, 54 [42-64] years) were enrolled, including 632 with CRC and 1421 controls (Table 1). Most participants had secondary education or higher (1756 [85.2%]), and lived in urban areas (1817 [88.1%]). There were 73 Hausa participants (3.5%), 197 Ibo participants (9.5%), and 1524 Yoruba participants (73.9%). Cases were older than controls (median [IQR] age, 56 [45-66] vs 52 [41-63] years) but were otherwise similar across sociodemographic characteristics. The prevalence of smoking, alcohol use, overweight or obesity, and diabetes was consistent with national data in Nigeria20,21 and lower than in the US.12,22,23
Table 1. Descriptive Characteristics of Study Population.
| Characteristic | Individuals, No. (%) | P valuea | ||
|---|---|---|---|---|
| Case (n = 632) | Control (n = 1431) | Overall (N = 2063) | ||
| Age, median (IQR), y | 56 (45-66) | 52 (41-63) | 54 (42-64) | <.001 |
| Sex | ||||
| Male | 338 (53.5) | 765 (53.5) | 1103 (53.5) | >.99 |
| Female | 294 (46.5) | 666 (46.5) | 960 (46.5) | |
| Education | ||||
| No education | 29 (4.6) | 59 (4.1) | 88 (4.3) | .88 |
| Primary school or less | 66 (10.4) | 144 (10.1) | 210 (10.2) | |
| Secondary school or vocational | 188 (29.7) | 422 (29.5) | 610 (29.6) | |
| University or higher | 341 (54.0) | 805 (56.3) | 1146 (55.6) | |
| Missing | 8 | 1 | 9 | NA |
| Tribe | ||||
| Hausa | 19 (3.0) | 54 (3.8) | 73 (3.5) | .69 |
| Ibo | 65 (10.3) | 132 (9.2) | 197 (9.5) | |
| Yoruba | 463 (73.3) | 1061 (74.1) | 1524 (73.9) | |
| Other | 72 (11.4) | 175 (12.2) | 247 (12.0) | |
| Missing | 13 | 9 | 22 | NA |
| Geographic area | ||||
| Rural | 66 (10.4) | 169 (11.8) | 235 (11.4) | .40 |
| Urban | 559 (88.4) | 1258 (87.9) | 1817 (88.1) | |
| Missing | 7 | 4 | 11 | NA |
| Religion | ||||
| Christian | 452 (71.5) | 996 (69.6) | 1448 (70.2) | .34 |
| Muslim | 166 (26.3) | 423 (29.6) | 589 (28.6) | |
| Other | 1 (0.2) | 2 (0.1) | 3 (0.1) | |
| Missing | 13 | 10 | 23 | NA |
| Occupation | ||||
| Other | 168 (26.6) | 353 (24.7) | 521 (25.3) | .04 |
| Retired or pensioner | 102 (16.1) | 195 (13.6) | 297 (14.4) | |
| Civil servant | 104 (16.5) | 293 (20.5) | 397 (19.2) | |
| Trader | 122 (19.3) | 256 (17.9) | 378 (18.3) | |
| Self-employed | 111 (17.6) | 309 (21.6) | 420 (20.4) | |
| Missing | 25 | 25 | 50 | |
| NSAID use | ||||
| Any | 367 (58.1) | 782 (54.6) | 1149 (55.7) | .22 |
| Missing | 69 | 173 | 242 | NA |
| MDPI | ||||
| Low | 178 (28.2) | 378 (26.4) | 556 (27.0) | .04 |
| Medium | 184 (29.1) | 371 (25.9) | 555 (26.9) | |
| High | 148 (23.4) | 407 (28.4) | 555 (26.9) | |
| Missing | 122 | 275 | 397 | NA |
| Family history of any cancer | 62 (9.8) | 92 (6.4) | 154 (7.5) | .007 |
| Height, median (IQR), cm | 169 (163-173) | 168 (162-171) | 168 (162-172) | .004 |
| Missing | 110 | 220 | 330 | NA |
| BMI | ||||
| 18.5-24.9 | 110 (17.4) | 334 (23.3) | 444 (21.5) | .001 |
| <18.5 | 347 (54.9) | 674 (47.1) | 1021 (49.5) | |
| 25-29.9 | 35 (5.5) | 108 (7.5) | 143 (6.9) | |
| ≥30 | 15 (2.4) | 41 (2.9) | 56 (2.7) | |
| Median (IQR) | 22.6 (19.6-25.9) | 24.1 (21.5-27.3) | 23.7 (21.0-27.0) | <.001 |
| Missing | 125 | 274 | 399 | NA |
| Composite somatotype | ||||
| Leaner (0-2.3) | 84 (13.3) | 220 (15.4) | 304 (14.7) | .14 |
| Intermediate (>2.3-3.0) | 196 (31.0) | 457 (31.9) | 653 (31.7) | |
| Larger (>3.0) | 249 (39.4) | 497 (34.7) | 746 (36.2) | |
| Missing | 103 | 257 | 360 | NA |
| Physical activity quartiles (MET h/wk) | ||||
| 1 (5.2) | 206 (32.6) | 323 (22.6) | 529 (25.6) | <.001 |
| 2 (15.7) | 117 (18.5) | 318 (22.2) | 435 (21.1) | |
| 3 (30.4) | 113 (17.9) | 319 (22.3) | 432 (20.9) | |
| 4 (53.1) | 138 (21.8) | 321 (22.4) | 459 (22.2) | |
| Missing | 58 | 150 | 208 | NA |
| Smoking status | ||||
| Never | 574 (90.8) | 1341 (93.7) | 1915 (92.8) | .54 |
| Ever | 41 (6.5) | 85 (5.9) | 126 (6.1) | |
| Missing | 17 | 5 | 22 | NA |
| Alcohol intake | ||||
| Never | 463 (73.3) | 1058 (73.9) | 1521 (73.7) | .47 |
| Ever | 146 (23.1) | 362 (25.3) | 508 (24.6) | |
| Missing | 23 | 11 | 34 | NA |
| Diabetes | ||||
| Yes | 44 (7.0) | 82 (5.7) | 126 (6.1) | .23 |
| Missing | 16 | 8 | 24 | NA |
| Tertiles of processed meat intake (median g/wk)b | ||||
| 1 (0) | 184 (29.1) | 433 (30.3) | 617 (29.9) | .004 |
| 2 (0-1.4) | 163 (25.8) | 448 (31.3) | 611 (29.6) | |
| 3 (>1.4) | 231 (36.6) | 426 (29.8) | 657 (31.8) | |
| Missing | 54 | 124 | 178 | NA |
| Quartiles of red meat intake (median g/wk) | ||||
| 1 (0-239) | 147 (23.3) | 315 (22.0) | 462 (22.4) | .96 |
| 2 (240-510) | 137 (21.7) | 305 (21.3) | 442 (21.4) | |
| 3 (511-1014) | 129 (20.4) | 298 (20.8) | 427 (20.7) | |
| 4 (>1014) | 134 (21.2) | 304 (21.2) | 438 (21.2) | |
| Missing | 85 | 209 | 294 | NA |
| Quartiles of fruit intake (median g/wk) | ||||
| 1 (0-1192) | 118 (18.7) | 349 (24.4) | 467 (22.6) | <.001 |
| 2 (1193-4862) | 109 (17.2) | 301 (21.0) | 410 (19.9) | |
| 3 (4863-7259) | 166 (26.3) | 290 (20.3) | 456 (22.1) | |
| 4 (>7259) | 166 (26.3) | 286 (20.0) | 452 (21.9) | |
| Missing | 73 | 205 | 278 | NA |
| Quartiles of vegetable intake (median g/wk) | ||||
| 1 (0-1143) | 141 (22.3) | 363 (25.4) | 504 (24.4) | .42 |
| 2 (1144-2447) | 143 (22.6) | 310 (21.7) | 453 (22.0) | |
| 3 (2448-3725) | 125 (19.8) | 315 (22.0) | 440 (21.3) | |
| 4 (>3725) | 146 (23.1) | 312 (21.8) | 458 (22.2) | |
| Missing | 77 | 131 | 208 | NA |
| Quartiles of fiber intake (median g/wk) | ||||
| 1 (0-189) | 101 (16.0) | 270 (18.9) | 371 (18.0) | .17 |
| 2 (190-259) | 153 (24.2) | 333 (23.3) | 486 (23.6) | |
| 3 (260-324) | 144 (22.8) | 332 (23.2) | 476 (23.1) | |
| 4 (>324) | 146 (23.1) | 280 (19.6) | 426 (20.6) | |
| Missing | 88 | 216 | 304 | NA |
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); MDPI, Multidimensional Poverty Index; MET, metabolic equivalent task; NA, not applicable; NSAID, nonsteroidal anti-inflammatory drug.
Wilcoxon rank sum test; Pearson χ2 test; Fisher exact test.
Tertiles of processed meat intake were used due to low consumption.
In fully adjusted models, 5 of 13 risk factors showed statistically significant CRC associations consistent with prior data in HIC populations (Figure 1 and Table 2). These included positive associations with nonmodifiable factors in adulthood: a family history of cancer (OR, 1.55; 95% CI, 1.06-2.26) and adult height (OR per 5-cm increase, 1.09; 95% CI, 1.00-1.20).
Figure 1. Forest Plots of Associations of Nonmodifiable, Metabolic, Lifestyle, and Dietary Factors With Colorectal Cancer Risk.
Odds ratios (ORs) were adjusted for all matching factors (age, education, geographic area, recruitment hospital, and sex), all hypothesized risk factors, and nonsteroid anti-inflammatory drug use (ever/never), modified global multidimensional poverty index, occupation, tribe, religion, and total energy intake. Composite somatotype tertile (T) 1 indicates leaner; T3, larger. BMI indicates body mass index (calculated as weight in kilograms divided by height in meters squared; collected at diagnosis for cases). Height P for trend = .049. Metabolic equivalent of tasks (MET) quartile (Q) 1 indicates 5.2 h/wk; Q2, 15.7 h/wk; Q3, 30.4 h/wk; Q4, 53.1 h/wk. Processed meat T 1 indicates 0 g/wk; T3, >1.4 g/wk. Red meat Q1 indicates 0-239 g/wk; Q4, >1014. Fruit Q1 indicates 0-1192 g/wk; Q4, >7259 g/wk. Vegetable Q1 indicates 0-1143g/wk; Q4, >3725 g/wk. Fiber Q1 indicates 0-189 g/wk; Q4, >325.
Table 2. Association Between Hypothesized Risk Factors and Colorectal Cancer in Adults in Nigeria.
| Characteristic | OR (95% CI) | |
|---|---|---|
| Minimally adjusted | Fully adjusteda | |
| Nonmodifiable factors | ||
| Family history of any cancer | ||
| No | 1 [Reference] | 1 [Reference] |
| Yes | 1.45 (1.02-2.05) | 1.55 (1.06-2.26) |
| Height, per 5-cm increase | 1.15 (1.06-1.25) | 1.09 (1.00-1.20)b |
| Metabolic factors | ||
| Composite somatotype tertilesc | ||
| Leaner | 1 [Reference] | 1 [Reference] |
| Intermediate | 0.96 (0.70-1.32) | 1.17 (0.83-1.64) |
| Larger | 1.18 (0.87-1.59) | 1.88 (1.30-2.72) |
| Composite somatotype, per 1-unit increase | 1.19 (1.09-1.30) | 1.46 (1.30-1.63) |
| BMI | ||
| 18.5-24.9 | 1 [Reference] | 1 [Reference] |
| <18.5 | 2.19 (1.54-3.11) | 2.63 (1.79-3.86) |
| 25-29.9 | 0.74 (0.57-0.94) | 0.60 (0.46-0.79) |
| ≥30 | 0.62 (0.44-0.89) | 0.42 (0.28-0.63) |
| P for trend | <.001 | <.001 |
| Diabetes | ||
| No | 1 [Reference] | 1 [Reference] |
| Yes | 1.09 (0.74-1.60) | 1.03 (0.67-1.56) |
| Lifestyle factors | ||
| Physical activity quartiles (median MET h/wk) | ||
| 1 (5.2) | 1 [Reference] | 1 [Reference] |
| 2 (15.7) | 0.60 (0.44-0.82) | 0.63 (0.44-0.88) |
| 3 (30.4) | 0.46 (0.31-0.68) | 0.52 (0.34-0.79) |
| 4 (53.1) | 0.67 (0.53-0.86) | 0.77 (0.58-1.02) |
| P value for trend | <.001 | .14 |
| Smoking status | ||
| Never | 1 [Reference] | 1 [Reference] |
| Ever | 1.09 (0.73-1.63) | 1.33 (0.84-2.09) |
| Dietary factors | ||
| Tertiles of processed meat intake (median g/wk)d | ||
| 1 (0) | 1 [Reference] | 1 [Reference] |
| 2 (0-1.4) | 0.84 (0.65-1.08) | 0.74 (0.55-1.03) |
| 3 (>1.4) | 1.27 (1.00-1.61) | 1.43 (1.07-1.91) |
| P value for trend | .04 | .03 |
| Quartiles of red meat intake (median g/wk) | ||
| 1 (0-239) | 1 [Reference] | 1 [Reference] |
| 2 (240-510) | 0.98 (0.74-1.30) | 0.96 (0.69-1.32) |
| 3 (511-1014) | 0.97 (0.73-1.29) | 0.81 (0.59-1.13) |
| 4 (>1014) | 1.00 (0.76-1.33) | 0.79 (0.55-1.12) |
| P value for trend | .99 | .18 |
| Quartiles of fruit intake (median g/wk) | ||
| 1 (0-1192) | 1 [Reference] | 1 [Reference] |
| 2 (1193-4862) | 0.98 (0.74-1.30) | 0.97 (0.67-1.41) |
| 3 (4863-7259) | 1.54 (1.16-2.04) | 1.61 (1.04-2.49) |
| 4 (>7259) | 1.33 (1.00-1.77) | 1.27 (0.78-2.07) |
| P value for trend | <.001 | .15 |
| Quartiles of vegetable intake (median g/wk) | ||
| 1 (0-1143) | 1 [Reference] | 1 [Reference] |
| 2 (1144-2447) | 1.13 (0.86-1.49) | 1.05 (0.74-1.47) |
| 3 (2448-3725) | 0.98 (0.74-1.31) | 0.94 (0.64-1.38) |
| 4 (>3725) | 1.15 (0.87-1.52) | 1.06 (0.70-1.59) |
| P value for trend | .53 | .94 |
| Quartiles of fiber intake (median g/wk) | ||
| 1 (0-189) | 1 [Reference] | 1 [Reference] |
| 2 (190-259) | 1.30 (0.98-1.72) | 1.14 (0.78-1.66) |
| 3 (260-324) | 1.21 (0.91-1.61) | 0.91 (0.59-1.42) |
| 4 (>324) | 1.38 (1.05-1.82) | 1.00 (0.60-1.68) |
| P value for trend | .04 | .72 |
| Alcohol intake | ||
| Never | 1 [Reference] | 1 [Reference] |
| Ever | 0.91 (0.72-1.15) | 0.79 (0.60-1.05) |
Abbreviations: NA, not applicable; OR, odds ratio.
Fully adjusted models were adjusted for all matching factors (age, education, geographic area, recruitment hospital, and sex), for all other covariates in the table, nonsteroidal anti-inflammatory drug use (ever/never), modified global multidimensional poverty index, occupation, tribe, religion, and total energy intake.
The P value for height was <.05 (P = .049); however, the CI includes 1.00 due to rounding.
Composite somatotype was included as a continuous variable in primary analysis and was replaced by tertiles of composite somatotype in a separate analysis.
We used tertiles of processed meat intake due to low consumption.
We observed a positive association with average body size or type over the life course (composite somatotype, larger vs leaner: OR, 1.88; 95% CI, 1.30-2.72; OR per 1-unit increase, 1.46; 95% CI, 1.30-1.63). In contrast, we observed an unexpected inverse association with BMI. Compared to participants with BMI 18.5 to 24.9, those with BMI less than 18.5 were more likely to have CRC (OR, 2.63; 95% CI, 1.79-3.86) and those with BMI 25 to 29.9 (OR, 0.60; 95% CI, 0.46, 0.79) or 30 or higher (OR, 0.42; 95% CI, 0.28-0.63) were less likely to have CRC. However, 88% of CRCs diagnosed in Nigerians are stage III or IV,3 and the inverse association with BMI was attenuated and no longer statistically significant after excluding individuals reporting weight loss (−5 kg) in the past 2 years (eTable 4 in Supplement 1). Consistent with BMI associations, compared with participants with lean body types, participants with intermediate (OR, 0.47; 95% CI, 0.35-0.63) and larger (OR, 0.30; 95% CI, 0.22-0.42) (P for trend < .001) somatotypes at baseline had lower risk of CRC (Figure 2; eTable 2 in Supplement 1). In contrast, for somatotypes at other ages, participants with the largest (vs leaner) body sizes had higher risk (Figure 2; eTable 2 in Supplement 1). We observed no association between physician-diagnosed diabetes and CRC (OR, 1.03; 95% CI, 0.67-1.56).
Figure 2. Forest Plot of Associations Between Somatotype at Different Ages and Colorectal Cancer Risk.
Odds ratios (ORs) were adjusted for all matching factors (age, education, geographic area, recruitment hospital, and sex), all hypothesized risk factors, and nonsteroid anti-inflammatory drug use (ever/never), modified global multidimensional poverty index, occupation, tribe, religion, and total energy intake. P for all trends < .05.
Compared with participants in the first quartile of physical activity, participants in the second (OR, 0.63; 95% CI, 0.44-0.88) and third (OR, 0.52; 95% CI, 0.34-0.79) quartiles had lower risk, while there was no statistically significant association for the fourth quartile (OR, 0.77; 95% CI, 0.58-1.02). The trend was not statistically significant given the nonlinear pattern (P for trend = .14). We did not observe an association between ever (vs never) smoking and CRC (OR, 1.33; 95% CI, 0.84-2.09).
Despite the low intake of processed meats in our population, we observed a positive association with CRC. Compared with participants in the first tertile (0 g/wk), there was no significant association for participants in the second tertile (median 0-1.4 g/wk: OR, 0.74; 95% CI, 0.55-1.03), but there was an association for participants in the third tertile (median >1.4 g/wk: OR, 1.43; 95% CI, 1.07-1.91) (P for trend = .03). We did not observe associations with the remaining dietary factors (red meat, fruit, vegetable, fiber, and alcohol).
We observed evidence of differences in associations by age for BMI, fiber intake, and composite somatotype (eTable 3 in Supplement 1). We observed no differences in associations by sex, global MDPI, or religion, although statistical power was limited.
To assess the association of weight change prior to CRC diagnosis with CRC risk, we restricted to individuals who reported no weight change in the 2 years before diagnosis (or enrollment date for controls). Estimates for family history of cancer, adult height, composite somatotype, and physical activity were generally similar; however, estimates for BMI and processed meat intake were attenuated and not statistically significant (eTable 4 in Supplement 1). Prior to study participation, only 465 individuals (22.5%) were aware of CRC; in this subgroup, approximately half (257 individuals [55.3%]) believed risk factors contribute to this disease. To assess the potential influence of recall bias, we restricted analyses to the subgroup of participants reporting no prior knowledge of cancer or CRC risk factors, in whom we anticipated minimal recall bias (eTable 5 in Supplement 1). While most interpretations remained unchanged, the association with family history of cancer was attenuated and not statistically significant (OR, 1.43; 95% CI, 0.87-2.34). Finally, results were not materially different when stratified by control source (family member or friend vs control patient), although estimates were imprecise due to limited statistical power.
Discussion
The findings of our large, multicenter case-control study suggest important similarities in CRC risk profiles between Nigerian and HIC populations. Our findings reinforce the importance of nonmodifiable factors, including family history of cancer and adult height, across populations.24 Our data also confirm the roles of modifiable risk factors, including body size or type, physical activity, and processed meat intake in our population. In contrast, smoking, alcohol intake, diabetes, and certain traditional dietary risk factors (red meat, fruit, vegetables, fiber) were not associated with CRC in this population, possibly due to their lower prevalence or variability. Our study is a crucial first step in identifying the risk factor profile for CRC in the region as countries undergo rapid economic development shifting lifestyles toward those of high-income populations. Research is needed to confirm our results in other West African and sub-Saharan African populations.
We observed a significant positive association of body size or type throughout life with CRC risk, consistent with findings from US studies.16 These results support weight maintenance as a potential prevention target in Nigeria, particularly as the prevalence of obesity continues to increase due to rapid economic growth and a shift toward a Westernized lifestyle.5 We observed a strong and unexpected inverse association with BMI, likely due to weight loss experienced by individuals before diagnosis, given most patients in our study had stage III or IV CRC. While there are few studies that have examined this association in sub-Saharan Africa,24,25,26 a robust case-control study of CRC in Zimbabwe24 did not assess the association with obesity because of similar difficulties estimating prediagnostic BMI in patients with CRC. This highlights the importance for regional studies to consider alternate measures to BMI. While we did not observe an association with diabetes,10 diabetes prevalence was low in our population, and in sub-Saharan Africa, it is estimated that more than 70% of individuals with diabetes are undiagnosed.27 Therefore, we had low statistical power to detect the modest effect estimates typically observed with diabetes in HICs.10 Future examination of diabetes and CRC in the region is warranted.
We observed a U-shaped association between physical activity and CRC risk. Although this association has not previously been examined in Nigeria, to our knowledge, physical activity was inversely associated with CRC in a 2024 meta-analysis of African studies.28 Our population reported notably higher levels of physical activity than traditionally observed in HICs,29 consistent with other studies in Africa,30 which may be attributed to high-intensity regional activities (eg, farming). This may also help explain the observed U-shaped association. A large meta-analysis of Australian, European, and US cohorts31 reported the inverse association with vigorous-intensity activity for CRC plateaued at higher levels of activity. We observed no association with smoking, which may be due to limited statistical power and modest effect estimates typically observed in HICs.32
For dietary factors, we observed a positive association with the highest (vs lowest) tertile of processed meat intake and CRC; this aligns with findings across Asia, Australia, Europe, and the US.33 However, low processed meat consumption limited contrasts across intake categories, which may explain no association comparing the second to the first tertile. Unlike studies in HICs,33 we found no associations with intakes of red meat, fruit, vegetables, or fiber. Similarly, a 2024 meta-analysis of CRC risk factors in Africa showed no associations with red meat or vegetable intake.28 While leafy vegetable consumption is inversely associated with CRC, the lack of an association in our study may reflect a diet that is high in the consumption of starchy vegetables (eg, yams, cassava), which may increase metabolic dysregulation and CRC risk.10,18 In HICs, associations with dietary intake typically show modest effect estimate sizes.33 The low variation in intake for red meat, fruit, vegetables, and fiber may have precluded us from detecting similar effect estimate sizes. While red meat was consumed in similar quantities to studies in HICs,33 the lack of association may be due to differences in the types of meats (eg, goat) or preparation methods (eg, soups or stews). Lastly, we did not observe an association with alcohol intake, a risk factor for CRC in HICs,10 although the prevalence of alcohol consumption was low and limited our ability to detect the modest effect estimates typically observed in HICs.33 These data highlight the need for region-specific studies of dietary intake.
These findings provide insights to inform the development of effective and resource-efficient interventions aimed at addressing the increasing burden of CRC and improving patient outcomes in Nigeria. Tailored programs focused on educating individuals about risk factors can play a key role in this effort. Prior to study participation, only 22% of individuals were aware of CRC; in this subgroup, approximately half (55.3%) believed risk factors contribute to this disease.
In our 2021 survey in Osun State, Nigeria, we found that 52% of community members had interacted with a primary health care practitioner in the past year9 and more than 70% of participants had access to radio and television, representing an opportunity for communitywide education initiatives. We also found very low CRC screening uptake (5%) in eligible community members.9 Constrained health care resources and socioeconomic barriers present serious challenges to populationwide screening in individuals with average risk of CRC in Nigeria.8,9 Data in HICs show epidemiologic risk factors can stratify individuals by their CRC risk.10 Associated risk factors in this study can be included in prediction modeling to identify individuals in Nigeria at high risk of CRC who may benefit from targeted screening.
Strengths and Limitations
Our study has several strengths. We examined the risk factor profile of CRC in a large population recruited from 6 hospitals spanning 3 of Nigeria’s 6 geopolitical zones. We created a tailored questionnaire, incorporating validated tools. We carefully designed our study to address the inherent limitations of its case-control design, although some limitations remain. To reduce the influence of preclinical or undiagnosed disease on reported exposures, we asked cases to respond based on when they felt well. However, we could only collect body weight at diagnosis. Sensitivity analysis showed that the inverse association of CRC with BMI was attenuated when analyzing individuals who reported no significant weight changes in the 2 years before enrollment. This is unsurprising, given that 88% of CRC diagnoses in Nigerians are late stage (stage III or IV), and these findings are likely driven by prediagnostic weight loss.3 Despite concerns about differential recall in cases and controls, analysis restricted to individuals who did not attribute cancer to risk factors (where we expected minimal recall bias) yielded consistent results, except for family history of cancer, which was less prevalent and no longer statistically significant, suggesting that participants with a family history of cancer may be more aware of cancer risk factors due to their relative’s diagnosis. Moreover, prior to their participation, only 22% of our population reported awareness of CRC, with no significant differences between cases and controls in their knowledge of CRC or the belief that risk factors contribute to this cancer. While hospital-based controls may have introduced selection bias, controls were visitors accompanying patients without cancer and patients undergoing routine medical care unrelated to CRC or any other cancer. Although measurement errors may have influenced our findings, we leveraged subject matter experts, established questionnaires, and validated instruments when creating our questionnaire to help minimize this potential bias. The prevalence of risk factors among our control individuals aligned with available national data,21 and we anticipate a similar extent of errors in cases and controls. To reduce residual or unmeasured confounding, we collected and adjusted for a large panel of factors. Furthermore, education levels in our participants were higher than the general Nigerian population, potentially limiting generalizability; however, this may enhance internal validity by reducing confounding related to socioeconomic status.
Conclusions
In this case-control study of factors associated with CRC in adults in Nigeria, we confirmed the relevance of several risk factors identified in HICs. Our data emphasize the importance of considering these factors in Nigeria’s cancer control strategies, providing a foundation for educational programs and evidence-based risk-reduction strategies. Studies of cervical cancer screening in sub-Saharan Africa have demonstrated that peer health education, community health educators, and mobile health messaging significantly improved screening uptake.34 Future studies can assess the utility of applying significant risk factors in our study to identify individuals with high risk of CRC who may benefit most from targeted efforts to increase CRC screening uptake. Our findings provide immediate insights and support future prospective studies in Nigeria and other West African populations. Studies are needed to explore potential associations with novel regional risk factors.
eMethods.
eTable 1. Descriptive characteristics of study population by recruitment hospital
eTable 2. Associations between tertiles of somatotype at different ages with risk of colorectal cancer in Nigerian adults
eTable 3. Association between hypothesized risk factors and colorectal cancer in Nigerian adults by age
eTable 4. Association between hypothesized risk factors and colorectal cancer in Nigerian adults, restricted to participants reporting no weight loss in the last 2 years
eTable 5. Association between hypothesized risk factors and colorectal cancer in Nigerian adults restricted to participants with no knowledge of cancer risk factors
Data Sharing Statement
References
- 1.Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024;74(3):229-263. doi: 10.3322/caac.21834 [DOI] [PubMed] [Google Scholar]
- 2.International Agency for Research on Cancer . Cancer tomorrow: predictions of the future cancer incidence and mortality burden worldwide up until 2050. Accessed June 25, 2026. https://gco.iarc.who.int/tomorrow/en
- 3.Sharma A, Alatise OI, Adisa AO, et al. Treatment of colorectal cancer in Sub-Saharan Africa: Results from a prospective Nigerian hospital registry. J Surg Oncol. 2020;121(2):342-349. doi: 10.1002/jso.25768 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Federal Ministry of Health . National Strategic Cancer Control Plan 2023-2027. National Institute for Cancer Research and Treatment; 2023. [Google Scholar]
- 5.Ngwa W, Addai BW, Adewole I, et al. Cancer in sub-Saharan Africa: a Lancet Oncology Commission. Lancet Oncol. 2022;23(6):e251-e312. doi: 10.1016/S1470-2045(21)00720-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Cancer Research UK . Bowel cancer statistics. Accessed June 25, 2026. https://www.cancerresearchuk.org/health-professional/cancer-statistics/statistics-by-cancer-type/bowel-cancer#heading-Four
- 7.National Cancer Institute . Colon and rectum (including appendix) recent trends in SEER age-adjusted incidence rates, 2000-2023. Accessed June 25, 2026. https://seer.cancer.gov/statistics-network/explorer/application.html?site=20&data_type=1&graph_type=2&compareBy=sex&chk_sex_3=3&chk_sex_2=2&rate_type=2&race=1&age_range=1&stage=101&advopt_precision=1&advopt_show_ci=on&hdn_view=0&advopt_show_apc=on&advopt_display=2#resultsRegion0.
- 8.Alatise OI, Dare AJ, Akinyemi PA, et al. ; African Research for Oncology Colorectal Screening Early Detection Group . Colorectal cancer screening with fecal immunochemical testing: a community-based, cross-sectional study in average-risk individuals in Nigeria. Lancet Glob Health. 2022;10(7):e1012-e1022. doi: 10.1016/S2214-109X(22)00121-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sharma A, Alatise OI, O’Connell K, et al. Healthcare utilisation, cancer screening and potential barriers to accessing cancer care in rural South West Nigeria: a cross-sectional study. BMJ Open. 2021;11(7):e040352. doi: 10.1136/bmjopen-2020-040352 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Jeon J, Du M, Schoen RE, et al. ; Colorectal Transdisciplinary Study and Genetics and Epidemiology of Colorectal Cancer Consortium . Determining risk of colorectal cancer and starting age of screening based on lifestyle, environmental, and genetic factors. Gastroenterology. 2018;154(8):2152-2164.e19. doi: 10.1053/j.gastro.2018.02.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ananthakrishnan AN, Du M, Berndt SI, et al. Red meat intake, NAT2, and risk of colorectal cancer: a pooled analysis of 11 studies. Cancer Epidemiol Biomarkers Prev. 2015;24(1):198-205. doi: 10.1158/1055-9965.EPI-14-0897 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Botteri E, Iodice S, Bagnardi V, Raimondi S, Lowenfels AB, Maisonneuve P. Smoking and colorectal cancer: a meta-analysis. JAMA. 2008;300(23):2765-2778. doi: 10.1001/jama.2008.839 [DOI] [PubMed] [Google Scholar]
- 13.Fedirko V, Tramacere I, Bagnardi V, et al. Alcohol drinking and colorectal cancer risk: an overall and dose-response meta-analysis of published studies. Ann Oncol. 2011;22(9):1958-1972. doi: 10.1093/annonc/mdq653 [DOI] [PubMed] [Google Scholar]
- 14.Johnson CM, Wei C, Ensor JE, et al. Meta-analyses of colorectal cancer risk factors. Cancer Causes Control. 2013;24(6):1207-1222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Alatise OI, Knapp GC, Sharma A, et al. Molecular and phenotypic profiling of colorectal cancer patients in West Africa reveals biological insights. Nat Commun. 2021;12(1):6821. doi: 10.1038/s41467-021-27106-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Zhang X, Wu K, Giovannucci EL, et al. Early life body fatness and risk of colorectal cancer in U.S. women and men—results from two large cohort studies. Cancer Epidemiol Biomarkers Prev. 2015;24(4):690-697. doi: 10.1158/1055-9965.EPI-14-0909-T [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Ainsworth BE, Haskell WL, Herrmann SD, et al. 2011 Compendium of Physical Activities: a second update of codes and MET values. Med Sci Sports Exerc. 2011;43(8):1575-1581. doi: 10.1249/MSS.0b013e31821ece12 [DOI] [PubMed] [Google Scholar]
- 18.Samson ML, Peeri NC, Alatise OI, et al. Validating a semi-quantitative food frequency questionnaire to assess regional diet in a study of cancer in South West Nigeria. Cancer Causes Control. 2023;34(6):495-503. doi: 10.1007/s10552-023-01684-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.van Buuren S, Groothuis-Oudshoorn K. mice: multivariate imputation by chained equations in R. J Stat Softw. 2011;45(3):1-67. doi: 10.18637/jss.v045.i03 [DOI] [Google Scholar]
- 20.National Population Commission, Federal Republic of Nigeria . Nigeria Demographic and Health Survey 2013. ICF International; 2014. [Google Scholar]
- 21.The World Bank . General household survey, panel 2018-2019, wave 4. Updated October 2021. Accessed June 25, 2026. https://microdata.worldbank.org/index.php/catalog/3557/study-description#metadata-questionnaires
- 22.Centers for Disease Control and Prevention . Health, United States—Data Finder. Accessed June 25, 2026. https://www.cdc.gov/nchs/hus/data-finder.htm?year=2019&table=Table%20026.
- 23.Liang PS, Chen TY, Giovannucci E. Cigarette smoking and colorectal cancer incidence and mortality: systematic review and meta-analysis. Int J Cancer. 2009;124(10):2406-2415. doi: 10.1002/ijc.24191 [DOI] [PubMed] [Google Scholar]
- 24.Katsidzira L, Gangaidzo IT, Makunike-Mutasa R, et al. A case-control study of risk factors for colorectal cancer in an African population. Eur J Cancer Prev. 2019;28(3):145-150. doi: 10.1097/CEJ.0000000000000439 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Wismayer R, Kiwanuka J, Wabinga H, Odida M. Risk factors for colorectal adenocarcinoma in an indigenous population in East Africa. Cancer Manag Res. 2022;14:2657-2669. doi: 10.2147/CMAR.S381479 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.El-Moselhy EA, Abdel-Halim MM, Eid AMME, et al. Colorectal cancer risk factors: a multi-center, case-control study in Egypt. Clin Epidemiol Glob Health. 2025;33:102017. doi: 10.1016/j.cegh.2025.102017 [DOI] [Google Scholar]
- 27.International Diabetes Federation . Africa: diabetes regional report 2000-2050. Accessed June 25, 2026. https://diabetesatlas.org/data-by-location/region/africa/
- 28.Tazinkeng NN, Pearlstein EF, Manda-Mapalo M, et al. Incidence and risk factors for colorectal cancer in Africa: a systematic review and meta-analysis. BMC Gastroenterol. 2024;24(1):303. doi: 10.1186/s12876-024-03385-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Guthold R, Stevens GA, Riley LM, Bull FC. Worldwide trends in insufficient physical activity from 2001 to 2016: a pooled analysis of 358 population-based surveys with 1.9 million participants. Lancet Glob Health. 2018;6(10):e1077-e1086. doi: 10.1016/S2214-109X(18)30357-7 [DOI] [PubMed] [Google Scholar]
- 30.Guthold R, Louazani SA, Riley LM, et al. Physical activity in 22 African countries: results from the World Health Organization STEPwise approach to chronic disease risk factor surveillance. Am J Prev Med. 2011;41(1):52-60. doi: 10.1016/j.amepre.2011.03.008 [DOI] [PubMed] [Google Scholar]
- 31.Matthews CE, Moore SC, Arem H, et al. Amount and intensity of leisure-time physical activity and lower cancer risk. J Clin Oncol. 2020;38(7):686-697. doi: 10.1200/JCO.19.02407 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Gong J, Hutter C, Baron JA, et al. A pooled analysis of smoking and colorectal cancer: timing of exposure and interactions with environmental factors. Cancer Epidemiol Biomarkers Prev. 2012;21(11):1974-1985. doi: 10.1158/1055-9965.EPI-12-0692 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Veettil SK, Wong TY, Loo YS, et al. Role of diet in colorectal cancer incidence: umbrella review of meta-analyses of prospective observational studies. JAMA Netw Open. 2021;4(2):e2037341. doi: 10.1001/jamanetworkopen.2020.37341 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Lott BE, Trejo MJ, Baum C, et al. Interventions to increase uptake of cervical screening in sub-Saharan Africa: a scoping review using the integrated behavioral model. BMC Public Health. 2020;20(1):654. doi: 10.1186/s12889-020-08777-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods.
eTable 1. Descriptive characteristics of study population by recruitment hospital
eTable 2. Associations between tertiles of somatotype at different ages with risk of colorectal cancer in Nigerian adults
eTable 3. Association between hypothesized risk factors and colorectal cancer in Nigerian adults by age
eTable 4. Association between hypothesized risk factors and colorectal cancer in Nigerian adults, restricted to participants reporting no weight loss in the last 2 years
eTable 5. Association between hypothesized risk factors and colorectal cancer in Nigerian adults restricted to participants with no knowledge of cancer risk factors
Data Sharing Statement
