Objective: Microsatellite instability (MSI) is one form of genomic instability that occurs in 10% to 20% of sporadic colon tumors and almost all hereditary nonpolyposis colon cancers. However, little is known about how environmental factors (e.g., diet) may influence MSI in sporadic colon cancer.

Methods: We used data from a population-based case-control study in North Carolina (486 colon cancer cases and 1,048 controls) to examine associations of diet (total energy, macronutrients, micronutrients, and food groups) with MSI. In-person interviews elicited information on potential colon cancer risk factors, and a previously validated food frequency questionnaire adapted to include regional foods was used to assess diet over the year before diagnosis or interview date. MSI was classified as MSI-high (MSI-H) and MSI-low or microsatellite stable (MSI-L/MSS). Multivariate logistic regression models estimated energy-adjusted and non-energy-adjusted odds ratios (OR).

Results: Ten percent of the cases (n = 49) had MSI-H tumors (29% African American). The strongest associations between diet and MSI were observed in case-control comparisons: there was a robust inverse association between MSI-H status and β-carotene [OR, 0.4; 95% confidence interval (95% CI), 0.2-0.9] and positive associations with energy-adjusted refined carbohydrates (OR, 2.2; 95% CI, 0.9-5.4) and non-energy-adjusted read meat intake (OR, 2.0; 95% CI, 0.9-4.2). Compared with controls, MSI-L/MSS tumors were statistically significantly associated with energy-adjusted vitamin C, vitamin E, calcium, dietary fiber, and dark green vegetables and positively associated with total energy intake (all Ps for trend < 0.05). In case-case comparisons, no dietary factors were significantly differently related to MSI-H compared with MSI-L/MSS tumors.

Conclusion: Refined carbohydrate and red meat consumption may promote development of MSI-H tumors, whereas β-carotene may be associated with lower risk.

Colon cancer is the second leading cause of cancer death in the United States and the third most common cancer among adults, accounting for ∼15% of all cancers diagnosed annually (1). Colon tumorigenesis results from a loss of genomic stability that leads to the transformation of normal colonic epithelial cells to colon adenocarcinoma cells (2). Microsatellite instability (MSI), defined as a change in DNA microsatellite length because of insertion or deletion of nucleotides in tumor DNA when compared with normal tissue (3), represents one form of genomic instability found in almost all hereditary nonpolyposis colorectal cancers and 10% to 20% of sporadic colon carcinomas (3-5).

In addition to family history, environmental and lifestyle factors have been consistently associated with colon cancer risk (6-9). In particular, dietary factors are among the most widely accepted etiologic risk factors, and it is estimated that 12% of colon cancer are attributable to consumption of a Western-style diet (8). Given the strong associations between diet and colon cancer risk, there is considerable scientific interest in determining whether dietary factors associated with colon cancer may influence genomic instability. In addition to providing further corroboration of a hypothesized relationship, knowledge of associations between diet and molecular alterations can yield valuable insight into the possible mechanisms by which a dietary factor may contribute to malignancy.

There have been very few epidemiologic investigations of diet and MSI, and to date, results have been somewhat inconsistent. For example, Wu et al. (10) found a positive association between red meat intake and MSI in a case-only study, whereas Slattery et al. (11) observed no such association in a case-control study. Diergaarde et al. (12) reported recently positive associations of red meat and alcohol intakes and inverse associations of fruit and vegetable consumption with MSI using data from a case-control study in the Netherlands. However, to our knowledge, these relationships have not been examined in racially heterogeneous study populations.

In this report, we examined associations between various dietary factors and genomic instability in a large population-based case-control study in North Carolina that was racially diverse with a sizeable proportion of African Americans. We focused on nutrients and food groups found previously to be linked with colon cancer risk in this population from North Carolina.

Study Design and Population

The North Carolina Colon Cancer Study (NCCCS) is a population-based case-control study of colon cancer in North Carolina. Details on the NCCCS design, population, and data collection methods have been published (13). In brief, study participants were from 33 counties in the central and eastern portions of North Carolina, an area that includes rural, suburban, and urban counties with a diverse socioeconomic mix of African Americans and Whites. The study was approved by the Institutional Review Board at the University of North Carolina School of Medicine and by equivalent committees at the collaborating hospitals.

Cases were persons with a first diagnosis of histologically confirmed invasive adenocarcinoma of the colon between October 1, 1996 and September 30, 2000 identified through the rapid ascertainment system of the North Carolina Central Cancer Registry (14). Other eligibility criteria included ages 40 to 80 years at the time of diagnosis, residence in the 33-county study area in North Carolina, ability to give informed consent and complete the interview, a North Carolina driver's license or identification card if <65 years old (because controls <65 years old were sampled from driver's license rosters), and permission to contact from the primary physician. Eligible and consenting patients were generally interviewed within 5 months of surgery. White cases were undersampled to increase the proportion of non-White cases in the study population.

The noninstitutionalized population-based controls were selected from two sources: North Carolina Division of Motor Vehicle records for cases <65 years old and Center for Medicare and Medicaid for cases ≥65 years old. These listings were used to randomly select potential controls within the same 5-year age group-, sex-, and race-defined strata. Study participants were selected using a randomized recruitment approach to achieve approximate frequency matching on age, sex, and race and to enhance the proportion of African Americans (15, 16). Those identified as eligible controls were contacted in a similar fashion to the cases to schedule in-person interviews.

Completed interviews were obtained from 1,691 participants. The overall study cooperation rate [interviewed / (interviewed + refused)] was 84% for cases and 63% for controls, whereas the response rate (interviewed/eligible) was 72% for cases and 61% for controls. For both cases and controls, the cooperation and response rates were slightly higher for Whites than for African Americans (13).

Data Collection

Data were collected in person by trained nurse interviewers at the participant's home or, occasionally, at another convenient location. The questionnaire collected detailed information on several factors that might relate to colon cancer, including diet, lifestyle factors, and medical history. The referent period for the interview was the year before diagnosis (cases) or interview date (controls).

Dietary Intake

Food (nutrient and food group) intake was assessed with a modified version of the previously validated 100-item semiquantitative Block food frequency questionnaire that was adapted to capture regional dietary practices by adding 29 foods commonly consumed in North Carolina, such as cooked greens, black-eyed peas, fried shellfish, hushpuppies, grits, and cobblers (17). In the present study, participants were asked to estimate their usual frequency of consumption of various foods and typical portion sizes for the year before diagnosis (for cases) or the year preceding the interview date (for controls). Each food item had nine options for frequency (ranging from “never or less than once per month” to “two or more times per day”) and three options for portion size. The food frequency questionnaire also included adjustment questions on types of foods used in cooking and preparation techniques and questions relating to restaurant eating, consumption of low-fat foods, fortified beverages, and fats used in cooking (18). Food groups and nutrient intakes were generated by an analysis program provided by the National Cancer Institute (19).

Dietary supplement use was assessed with closed-ended questions on the use of multiple vitamins (including antioxidant combinations) and single supplements. Participants were asked specifically about duration (years), frequency (days per week), and usual dose (e.g., 100, 250, 500, 750, and 1,000 mg for vitamin C) of supplements used over the past year. Because supplements contribute a large proportion of micronutrient intakes (18), we summed across intakes from food sources and supplements to obtain total micronutrient exposure.

For these analyses, we examined only nutrients and food groups found previously to be significantly associated with colon cancer risk in this study sample (Tables 2 and 3; refs. 13, 20, 21). Participants with reported energy intakes <800 and >5,000 kcal for men and <600 and >4,000 kcal for women were excluded because their food frequency questionnaires were considered unreliable (22).

Other Participant Characteristics

Data were collected on demographic characteristics, including age, sex, education, race, smoking history, physical activity, nonsteroidal anti-inflammatory drugs (NSAID) use over the last 5 years, and first-degree family history of colon cancer. Height and weight were measured by trained staff using a standardized protocol and used to compute body mass index (BMI) as weight (in kilograms) divided by height (in meters) squared. BMI (kg/m2) was further categorized as “normal” (18.5-24.9), “overweight” (25-29.9), and “obese” (≥30; ref. 23). Cases were also asked their weight 1 and 5 years before diagnosis. Physical activity was measured in metabolic equivalent task-minutes per day for combined occupational, nonoccupational, and nonwork/weekend activities (including duration, frequency, and intensity) using a modified version of a validated 7-day physical activity recall (24, 25).

Microsatellite Instability

DNA Extraction. DNA was extracted from formalin-fixed, paraffin-embedded tissues of each participant's colorectal tumor and surrounding noncancerous tissue. Each area to be microdissected was identified on a reference H&E-stained slide. Unstained tissue slices adjacent to the reference slide were microdissected according to areas identified on the reference slide using a surgical scalpel blade. The dissected specimen was deparaffinized in a microfuge tube with xylene, and the DNA was purified with ethanol and GeneReleaser (BioVentures, Murfeesboro, TN) according to the manufacturer's recommendations. The samples were then treated with 200 μg/mL proteinase K (Sigma, St. Louis, MO) and incubated at 55°C for 5 hours. The proteinase K was destroyed by heating the sample to 95°C for 15 minutes and the samples were immediately iced and stored for PCR analysis.

Microsatellite Analysis. For determination of MSI, we used the National Cancer Institute–recommended panel of five microsatellite markers (BAT25, BAT26, D5S346, D2S123, and D17S250) to classify tumors as MSI-high [MSI-H; associated with inactivation of DNA mismatch repair (MMR)] or MSI-low or microsatellite stable (MSI-L/MSS; not associated with DNA MMR inactivation; ref. 3). MSI-H was defined as ≥2 markers demonstrating novel alleles compared with nontumor tissue, MSI-L as 1 marker with a novel allele, and MSS as no marker with novel alleles. Because of the similar features of MSI-L and MSS tumors, these two groups were combined as non-MSI-H tumors for these analyses. One primer from each microsatellite marker pair was radio-end-labeled with 32P in a reaction containing the primer, kinase buffer, T4 polynucleotide kinase, and [γ-32P]ATP. PCR was done on the microdissected template DNA in a reaction containing 0.125 pmol each of the end-labeled and “cold” primers, 0.25 units Taq DNA polymerase, 40 mmol/L deoxynucleotide triphosphate stock solution, and final concentration of 1.5 to 2.0 mmol/L magnesium. PCR products were denatured in 95% formamide and electrophoresed on a 6% polyacrylamide gel containing 7.5 mol/L urea. The gels are then dried and exposed by autoradiography to X-ray film or in a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).

Statistical Analyses

After all exclusions, the analytic data set included 875 Whites (267 cases and 608 controls) and 659 African Americans (219 cases and 440 controls). Analyses were initially stratified by race; however, there was no evidence of effect modification, so results are presented for all participants combined. Descriptive statistics (raw means, SDs, and percentages) stratified by case (MSI-H or MSI-L/MSS) and control status were used to describe the demographic/lifestyle characteristics and dietary intakes of study participants (Tables 1 and 2). Logistic regression models were used to generate Ps to examine whether there were statistically significant differences between cases/cases and cases/controls for the participant characteristics and dietary factors under examination after adjusting for potential covariates. For Table 1, potential confounders included all the characteristics (except current BMI), whereas covariates for Table 2 were variables in Tables 1 and 2 that changed the variable coefficient of the dietary factor under examination by at least 15% when added to the model.

Table 1.

Characteristics of participants in the NCCCS by case and MSI status (n = 1,534)

Participant characteristics*Controls
Cases
Ps
(n = 1,048)MSI-H (n = 49)MSI-L/MSS (n = 437)MSI-H vs controlsMSI-H vs MSI-L/MSSMSI-L/MSS vs controls§
Tumor site (%)       
    Proximal — 78 47 — 0.0002 — 
    Distal — 16 38    
Unknown/missing — 15    
Age, y (%)       
    <55 15 20 22 0.0001 0.0002 0.0001 
    55-64 23 14 27    
    ≥65 63 65 51    
Mean ± SD 66.1 (9.5) 65.9 (10.9) 63.7 (10.2) 0.0001 0.001 0.0001 
Sex (%)       
    Males 49 41 51 0.21 0.03 0.01 
    Females 51 59 49    
Race (%)       
    White 58 71 53 0.10 0.03 0.26 
    African Americans 42 29 47    
Education (%)       
    High school or less 58 54 65 0.52 0.03 0.01 
    Some college 21 17 20    
    College graduate/advanced degree 21 29 15    
BMI, kg/m2 (%) current       
    Normal (18-24.9) 28 44 28 0.27 0.35 0.69 
    Overweight (25-29.9) 39 38 37    
    Obese (≥30) 33 18 34    
    Mean ± SD 28.5 (6.2) 25.7 (5.2) 28.5 (6.2) 0.04 0.12 0.30 
BMI, kg/m2 (%) 1 y ago       
    Normal (18.5-24.9) 27 33 20 0.94 0.64 0.09 
    Overweight (25-29.9) 40 41 40    
    Obese (≥30) 33 26 40    
    Mean ± SD 28.4 (5.8) 27.4 (5.6) 29.8 (6.3) 0.65 0.54 0.02 
Smoking history (%)       
    Never smoker 18 27 15 0.003 0.01 0.14 
    Former smoker 39 54 42    
    Current smoker 43 19 42    
Physical activity, metabolic equivalent task-min/d (%)       
    Quartile 1 26 17 24 0.47 0.51 0.71 
    Quartile 2 25 35 24    
    Quartile 3 25 19 26    
    Quartile 2 24 29 26    
    Mean ± SD (metabolic equivalent task-min/d) 2,182 (508) 2,253 (534) 2,230 (554) 0.64 0.63 0.45 
Family history of colon cancer (%)       
    Yes 10 29 20 0.0002 0.53 0.0001 
    No 90 71 80    
NSAID use over the past 5 y (%)       
    Never 21 11 0.003 0.30 0.0001 
    Occasionally 30 48 40    
    Regularly 63 31 49    
Vitamin/mineral supplement use (%)       
    Yes 52 43 41 0.16 0.26 0.008 
    No 48 57 59    
Participant characteristics*Controls
Cases
Ps
(n = 1,048)MSI-H (n = 49)MSI-L/MSS (n = 437)MSI-H vs controlsMSI-H vs MSI-L/MSSMSI-L/MSS vs controls§
Tumor site (%)       
    Proximal — 78 47 — 0.0002 — 
    Distal — 16 38    
Unknown/missing — 15    
Age, y (%)       
    <55 15 20 22 0.0001 0.0002 0.0001 
    55-64 23 14 27    
    ≥65 63 65 51    
Mean ± SD 66.1 (9.5) 65.9 (10.9) 63.7 (10.2) 0.0001 0.001 0.0001 
Sex (%)       
    Males 49 41 51 0.21 0.03 0.01 
    Females 51 59 49    
Race (%)       
    White 58 71 53 0.10 0.03 0.26 
    African Americans 42 29 47    
Education (%)       
    High school or less 58 54 65 0.52 0.03 0.01 
    Some college 21 17 20    
    College graduate/advanced degree 21 29 15    
BMI, kg/m2 (%) current       
    Normal (18-24.9) 28 44 28 0.27 0.35 0.69 
    Overweight (25-29.9) 39 38 37    
    Obese (≥30) 33 18 34    
    Mean ± SD 28.5 (6.2) 25.7 (5.2) 28.5 (6.2) 0.04 0.12 0.30 
BMI, kg/m2 (%) 1 y ago       
    Normal (18.5-24.9) 27 33 20 0.94 0.64 0.09 
    Overweight (25-29.9) 40 41 40    
    Obese (≥30) 33 26 40    
    Mean ± SD 28.4 (5.8) 27.4 (5.6) 29.8 (6.3) 0.65 0.54 0.02 
Smoking history (%)       
    Never smoker 18 27 15 0.003 0.01 0.14 
    Former smoker 39 54 42    
    Current smoker 43 19 42    
Physical activity, metabolic equivalent task-min/d (%)       
    Quartile 1 26 17 24 0.47 0.51 0.71 
    Quartile 2 25 35 24    
    Quartile 3 25 19 26    
    Quartile 2 24 29 26    
    Mean ± SD (metabolic equivalent task-min/d) 2,182 (508) 2,253 (534) 2,230 (554) 0.64 0.63 0.45 
Family history of colon cancer (%)       
    Yes 10 29 20 0.0002 0.53 0.0001 
    No 90 71 80    
NSAID use over the past 5 y (%)       
    Never 21 11 0.003 0.30 0.0001 
    Occasionally 30 48 40    
    Regularly 63 31 49    
Vitamin/mineral supplement use (%)       
    Yes 52 43 41 0.16 0.26 0.008 
    No 48 57 59    

NOTE: All data are for the reference year, which is the year before diagnosis for cases and the year before the interview for controls.

*

All data are in percentages, except for mean ± SD for age and BMI (current and 1 year ago).

Test for difference between MSI-H and controls after controlling for the other participant characteristics in Table 1 (not including current BMI).

Test for difference between MSI-H and MSI-L/MSS after controlling for the other participant characteristics in Table 1 (not including current BMI).

§

Test for difference between MSI-L/MSS and controls after controlling for the other participant characteristics in Table 1 (not including current BMI).

Never: no NSAID use in the past 5 years; occasionally: used at least one NSAID during the past 5 years; regularly: used at least one NSAID ≥3 times per week at some time during in the past 5 years.

Use of any vitamin or mineral supplement at least once a week over the past year.

Table 2.

Mean intakes of various dietary factors by case and MSI status among participants in the NCCCS (n = 1,534)

Dietary factorControls
Cases
P*,P*,P*,§
(n = 1,048)MSI-H (n = 49)MSI-L/MSS (n = 437)
Total energy (kcal) 1,787 (688) 2,020 (860) 1,989 (848) 0.0007 0.59 0.0001 
       
Micronutrients       
    β-Carotene (μg/d) 4,513 (4,624) 3,306 (2,220) 3,571 (2,984) 0.04 0.37 0.04 
    Vitamin C (mg/d) 230 (326) 240 (362) 168 (255) 0.99 0.24 0.05 
    Vitamin E (mg α-tocopherol equivalent/d) 86.6 (162.7) 58.0 (114.7) 55.3 (119.7) 0.25 0.78 0.01 
    Calcium (mg/d) 849 (495) 862 (513) 765 (454) 0.11 0.69 0.0005 
Macronutrients       
    Carbohydrate (g/d) 208 (83) 221 (91) 225 (95) 0.007 0.13 0.18 
    Carbohydrate (% energy) 47.1 (7.6) 44.5 (6.0) 46.1 (7.3) 0.02 0.10 0.09 
    Effective carbohydrate (g/d) 195 (79) 208 (87) 212 (90) 0.001 0.14 0.20 
    Protein (g/d) 65.4 (26.2) 72.5 (31.9) 69.9 (32.7) 0.76 0.23 0.68 
    Protein (% energy) 14.8 (2.9) 14.5 (2.4) 14.2 (2.8) 0.50 0.29 0.0001 
    Total fat (g/d) 74.5 (33.9) 88.8 (45.7) 85.1 (41.2) 0.002 0.39 0.0001 
    Total fat (% energy) 37.1 (7.2) 38.6 (5.8) 38.0 (6.5) 0.22 0.17 0.26 
    Saturated fat (g/d) 25.0 (12.3) 30.1 (16.8) 28.6 (14.7) 0.01 0.39 0.85 
    Alcohol (g/d) 51 (154) 79 (158) 94 (281) 0.56 0.40 0.008 
    Dietary fiber (g/d) 13.9 (6.0) 13.2 (5.4) 13.4 (6.0) 0.11 0.31 0.04 
Food groups       
    Fruits (servings/wk) 9.39 (6.79) 8.63 (5.79) 8.31 (6.85) 0.38 0.96 0.57 
    Vegetables (servings/wk) 16.61 (8.23) 15.87 (6.20) 14.71 (7.15) 0.12 0.85 0.008 
    Dark green, deep yellow 4.18(3.08) 3.82 (2.44) 3.41 (2.57) 0.25 0.31 0.07 
    Refined carbohydrates or cereals (servings/wk) 16.23 (6.63) 18.21 (5.99) 17.05 (7.43) 0.10 0.12 0.12 
    Red meat (servings/wk) 7.00 (4.45) 8.92 (6.56) 8.09 (4.86) 0.001 0.30 0.89 
Dietary factorControls
Cases
P*,P*,P*,§
(n = 1,048)MSI-H (n = 49)MSI-L/MSS (n = 437)
Total energy (kcal) 1,787 (688) 2,020 (860) 1,989 (848) 0.0007 0.59 0.0001 
       
Micronutrients       
    β-Carotene (μg/d) 4,513 (4,624) 3,306 (2,220) 3,571 (2,984) 0.04 0.37 0.04 
    Vitamin C (mg/d) 230 (326) 240 (362) 168 (255) 0.99 0.24 0.05 
    Vitamin E (mg α-tocopherol equivalent/d) 86.6 (162.7) 58.0 (114.7) 55.3 (119.7) 0.25 0.78 0.01 
    Calcium (mg/d) 849 (495) 862 (513) 765 (454) 0.11 0.69 0.0005 
Macronutrients       
    Carbohydrate (g/d) 208 (83) 221 (91) 225 (95) 0.007 0.13 0.18 
    Carbohydrate (% energy) 47.1 (7.6) 44.5 (6.0) 46.1 (7.3) 0.02 0.10 0.09 
    Effective carbohydrate (g/d) 195 (79) 208 (87) 212 (90) 0.001 0.14 0.20 
    Protein (g/d) 65.4 (26.2) 72.5 (31.9) 69.9 (32.7) 0.76 0.23 0.68 
    Protein (% energy) 14.8 (2.9) 14.5 (2.4) 14.2 (2.8) 0.50 0.29 0.0001 
    Total fat (g/d) 74.5 (33.9) 88.8 (45.7) 85.1 (41.2) 0.002 0.39 0.0001 
    Total fat (% energy) 37.1 (7.2) 38.6 (5.8) 38.0 (6.5) 0.22 0.17 0.26 
    Saturated fat (g/d) 25.0 (12.3) 30.1 (16.8) 28.6 (14.7) 0.01 0.39 0.85 
    Alcohol (g/d) 51 (154) 79 (158) 94 (281) 0.56 0.40 0.008 
    Dietary fiber (g/d) 13.9 (6.0) 13.2 (5.4) 13.4 (6.0) 0.11 0.31 0.04 
Food groups       
    Fruits (servings/wk) 9.39 (6.79) 8.63 (5.79) 8.31 (6.85) 0.38 0.96 0.57 
    Vegetables (servings/wk) 16.61 (8.23) 15.87 (6.20) 14.71 (7.15) 0.12 0.85 0.008 
    Dark green, deep yellow 4.18(3.08) 3.82 (2.44) 3.41 (2.57) 0.25 0.31 0.07 
    Refined carbohydrates or cereals (servings/wk) 16.23 (6.63) 18.21 (5.99) 17.05 (7.43) 0.10 0.12 0.12 
    Red meat (servings/wk) 7.00 (4.45) 8.92 (6.56) 8.09 (4.86) 0.001 0.30 0.89 

NOTE: All data are for the reference year, which is the year before diagnosis for cases and the year before the interview for controls.

*

Age, race, sex, and the offset term are included in all models. Any variable in Table 1 or Table 2 that changed the variable coefficient of the dietary variable by at least 15% when added to a model with just age, race, and sex was a candidate for inclusion in the final model. The final model was determined by backward-stepwise selection that included all such variables.

Test for difference between MSI-H and controls.

Test for difference between MSI-H and MSI-L/MSS.

§

Test for difference between MSI-L/MSS and controls.

We calculated odds ratios (OR) and 95% confidence intervals (95% CI) from unconditional logistic regression models to ascertain associations of the dietary variables (in tertiles) with MSI status. Offset terms were included in all models to correct for randomized recruitment sampling fractions (15, 16) and allow estimation of unbiased ORs. This was necessary because we conditioned recruitment on age, sex, and race in addition to disease status; thus, the ORs without the offset term will be biased compared with a traditional design in which recruitment is conditioned on disease status alone. Cut points for tertiles of nutrient intakes were determined based on the distributions among controls. All variables in Tables 1 and 2 (except current BMI) were evaluated as potential confounding factors; covariate inclusion was based on whether there was a ≥15% alteration in the variable coefficient of interest. Covariates that met this criterion were placed in the model simultaneously and a backward-stepwise regression was done to determine the final model. All ORs are reported for energy-adjusted and non-energy-adjusted nutrient intakes. The standard multivariate method was used to adjust for total energy intake; applying other energy adjustment approaches (e.g., the nutrient residual and energy partition models) did not alter the results (22). Statistical tests were two sided and Ps < 0.05 were considered statistically significant. All analyses were done using SAS 8.1 (SAS Institute, Inc., Cary, NC).

Demographic and lifestyle characteristics of study participants, stratified by colon cancer and MSI status, are given in Table 1. Ten percent (49 of 486) of the cases were classified as MSI-H, of whom 29% were African American. There were 437 MSI-L/MSS cases (47% African American) and 1,048 controls (42% African American). The majority (78%) of MSI-H cases had proximal tumors, whereas MSI-L/MSS cases more often had distal tumors. MSI-H cases were older than those with MSI-L/MSS tumors but slightly younger than controls (P < 0.0001). Compared with MSI-L/MSS cases and controls, cases with MSI-H tumors were also more likely to be female, college educated, former or never smokers, have a family history of colon cancer, and never or occasional users of NSAIDs during the preceding 5 years (all Ps < 0.05). Compared with controls, MSI-L/MSS cases had a higher mean BMI in the year before diagnosis, were more likely to have a first-degree relative with colon cancer, and less likely to have used NSAIDs over the past 5 years or to have used vitamin and mineral supplements over the previous year (all Ps < 0.05).

Table 2 gives mean intakes of total energy, micronutrients, macronutrients, dietary fiber, and food groups by colon cancer and MSI status. Overall, colon cancer cases reported higher intakes of total energy and macronutrients and lower intakes of micronutrients than controls. Specifically, compared with controls, MSI-H cases reported significantly higher intakes of total energy, carbohydrate (g/d), effective carbohydrate (13), total fat (g/d), saturated fat (g/d), and red meat (servings/wk) but lower β-carotene and percentage energy from carbohydrate (all Ps < 0.05). Similarly, cases with MSI-L/MSS tumors had higher total energy, total fat (g/d), and alcohol intakes but reported significantly lower intakes of all micronutrients, dietary fiber, vegetables (servings/wk), and percentage energy from protein than controls (all Ps < 0.05). There were no statistically significant differences in reported intakes of these foods or nutrients between MSI-H and MSI-L/MSS cases.

Results of case-control and case-case comparisons examining associations (ORs and 95% CI) of dietary factors with MSI-H and MSI-L/MSS tumors are given in Table 3. Because previous analyses have shown appreciable differences in results based on whether ORs were adjusted for total energy (13, 20, 21), both sets of results are presented here. Compared with controls and regardless of energy adjustment, MSI-H tumors were not statistically significantly associated with any of the dietary factors examined, except for a strong inverse association with energy-adjusted β-carotene (OR, 0.4; 95% CI, 0.2-0.9) and a positive association with non-energy-adjusted refined carbohydrates (OR, 3.5; 95% CI, 1.5-8.2). MSI-H tumors were also associated with energy-adjusted dietary fiber intake (OR, 0.4; 95% CI, 0.2-1.0) and red meat without control for total energy (OR, 2.0; 95% CI, 0.9-4.2). When cases with MSI-L/MSS tumors were compared with controls, there were inverse associations of MSI-L/MSS with vitamin C and dark green vegetables regardless of energy adjustment and vitamin E, calcium, and dietary fiber after energy adjustment (all Ps for trend < 0.05). MSI-L/MSS tumors were also statistically significantly associated with total energy and non-energy-adjusted carbohydrate, effective carbohydrate, total fat, saturated fat, and red meat. In case-case comparisons, there was no evidence that any of the dietary factors were significantly differently related to MSI-H compared with MSI-L/MSS tumors.

Table 3.

Associations of dietary factors with MSI among participants in the NCCCS (n = 1,534)

Dietary factor*Median intake/d in controlsMSI-H vs controls, OR (95% CI)
MSI-L/MSS vs controls, OR (95% CI)
MSI-H vs MSI-L/MSS, OR (95% CI)
Energy-adjustedNot adjusted for energy§Energy-adjustedNot adjusted for energy§Energy-adjustedNot adjusted for energy§
Total energy (kcal)        
    Tertile 1 1,166 — 1.0 — 1.0 — 1.0 
    Tertile 2 1,744 — 0.7 (0.3-1.6) — 1.1 (0.8-1.5) — 0.9 (0.4-2.0) 
    Tertile 3 2,446 — 1.3 (0.4-4.4) — 2.5 (1.7-3.5) — 1.2 (0.5-2.8) 
P for linear trend  — 0.93 — 0.0001 — 0.85 
Micronutrients        
β-Carotene        
    Tertile 1 (reference) 1,517 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 3,224 0.6 (0.3-1.2) 0.6 (0.3-1.3) 0.8 (0.6-1.1) 0.8 (0.6-1.1) 0.7 (0.3-1.6) 0.7 (0.3-1.6) 
    Tertile 3 6,383 0.4 (0.2-0.9) 0.4 (0.2-0.9) 0.7 (0.5-1.0) 0.8 (0.6-1.1) 0.7 (0.3-1.7) 0.7 (0.3-1.7) 
P for linear trend  0.02 0.03 0.06 0.17 0.38 0.36 
Vitamin C        
    Tertile 1 (reference) 57.7 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 115.5 0.9 (0.4-1.9) 0.9 (0.4-1.9) 0.9 (0.7-1.2) 0.9 (0.7-1.2) 1.1 (0.5-2.4) 1.2 (0.5-2.5) 
    Tertile 3 261.8 0.8 (0.4-1.8) 0.8 (0.3-1.7) 0.7 (0.5-1.0) 0.7 (0.5-0.9) 1.4 (0.6-3.0) 1.5 (0.7-3.1) 
P for linear trend  0.62 0.55 0.04 0.04 0.48 0.38 
Vitamin E        
    Tertile 1 (reference) 6.3 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 12.8 0.8 (0.4-1.7) 1.1 (0.5-2.2) 0.8 (0.6-1.1) 1.0 (0.8-1.4) 1.0 (0.5-2.1) 1.1 (0.6-2.3) 
    Tertile 3 168.5 0.5 (0.2-1.0) 0.6 (0.3-1.3) 0.6 (0.5-0.8) 0.8 (0.6-1.0) 0.7 (0.3-1.6) 0.8 (0.4-1.8) 
P for linear trend  0.10 0.35 0.005 0.19 0.51 0.76 
Calcium        
    Tertile 1 (reference) 410.9 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 694.0 1.1 (0.5-2.4) 1.2 (0.6-2.7) 0.6 (0.5-0.9) 0.9 (0.7-1.2) 2.0 (0.9-4.4) 2.1 (1.0-4.5) 
    Tertile 3 1,299.7 0.5 (0.2-1.3) 0.6 (0.3-1.5) 0.5 (0.4-0.7) 0.8 (0.6-1.1) 1.3 (0.5-3.4) 1.5 (0.7-3.3) 
P for linear trend  0.32 0.57 0.0001 0.23 0.33 0.17 
Macronutrients        
Carbohydrate (g/d)        
    Tertile 1 134.2 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 201.9 0.8 (0.4-1.8) 1.1 (0.5-2.5) 0.9 (0.6-1.2) 1.2 (0.9-1.6) 1.0 (0.4-2.3) 1.1 (0.5-2.5) 
    Tertile 3 284.4 0.4 (0.1-1.5) 1.1 (0.4-2.9) 1.0 (0.7-1.6) 2.1 (1.5-2.9) 0.5 (0.1-1.6) 0.6 (0.2-1.7) 
P for linear trend  0.24 0.77 0.96 0.0002 0.34 0.50 
E-carbohydrate (g/d)        
    Tertile 1 (reference) 124.7 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 188.0 1.0 (0.4-2.2) 1.3 (0.6-2.9) 0.9 (0.7-1.3) 1.2 (0.9-1.7) 1.1 (0.5-2.5) 1.1 (0.5-2.6) 
    Tertile 3 264.8 0.5 (0.1-1.7) 1.2 (0.5-3.2) 1.0 (0.6-1.6) 2.1 (1.5-2.9) 0.6 (0.2-1.9) 0.7 (0.3-1.9) 
P for linear trend  0.37 0.56 0.92 0.0002 0.50 0.67 
Protein (g/d)        
    Tertile 1 (reference) 41.9 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 62.7 0.8 (0.4-1.8) 1.2 (0.5-2.6) 0.8 (0.6-1.1) 1.1 (0.8-1.5) 0.8 (0.3-1.9) 0.8 (0.3-1.8) 
    Tertile 3 89.2 0.9 (0.3-2.8) 2.5 (1.0-6.1) 0.7 (0.4-1.0) 1.4 (1.0-2.0) 1.3 (0.4-3.9) 1.4 (0.6-3.0) 
P for linear trend  0.83 0.11 0.06 0.06 0.85 0.72 
Total fat (g/d)        
    Tertile 1 (reference) 44.3 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 71.0 0.6 (0.2-1.4) 0.8 (0.4-1.8) 0.8 (0.5-1.0) 1.0 (0.7-1.3) 0.7 (0.3-1.7) 0.8 (0.3-1.8) 
    Tertile 3 108.8 0.9 (0.3-2.7) 2.0 (1.0-4.4) 1.1 (0.7-1.8) 2.1 (1.6-2.9) 1.0 (0.3-2.9) 1.3 (0.6-2.9) 
P for linear trend  0.58 0.25 0.93 0.0005 0.80 0.81 
Saturated fat (g/d)        
    Tertile 1 (reference) 14.1 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 23.5 0.6 (0.2-1.3) 0.8 (0.3-1.7) 0.7 (0.5-1.0) 0.9 (0.7 1.3) 0.6 (0.2-1.4) 0.6 (0.3-1.5) 
    Tertile 3 37.3 0.6 (0.2-1.8) 1.5 (0.7-3.1) 1.0 (0.6-1.5) 2.0 (1.4-2.7) 0.7 (0.2-2.0) 1.0 (0.4-2.2) 
P for linear trend  0.26 0.63 0.35 0.002 0.33 0.68 
Alcohol (g/d)**        
    Nondrinkers (reference) 1.0 1.0 1.0 1.0 1.0 1.0 
    Drinkers 84.0 1.2 (0.6-2.3) 1.2 (0.6-2.3) 0.8 (0.6-1.1) 0.9 (0.7-1.1) 1.2 (0.6-2.4) 1.2 (0.6-2.5) 
P for linear trend  0.55 0.54 0.20 0.33 0.66 0.62 
Dietary fiber (g/d)        
    Tertile 1 (reference) 8.1 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 13.0 0.7 (0.3-1.6) 0.8 (0.4-1.7) 0.9 (0.6-1.2) 1.1 (0.8-1.5) 0.9 (0.4-2.0) 0.9 (0.4-2.0) 
    Tertile 3 19.0 0.4 (0.2-1.0) 0.5 (0.2-1.2) 0.6 (0.4-0.9) 1.0 (0.7-1.5) 0.7 (0.2-2.1) 0.7 (0.3-2.0) 
P for linear trend  0.08 0.15 0.01 0.76 0.55 0.53 
Food groups        
Fruits/juices (servings/wk)        
    Tertile 1 (reference) 2.8 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 7.0 1.1 (0.5-2.3) 1.0 (0.5-2.3) 0.9 (0.7-1.3) 0.9 (0.7-1.2) 1.3 (0.5-3.0) 1.3 (0.6-3.1) 
    Tertile 3 14.0 1.9 (0.8-4.3) 1.6 (0.7-3.7) 0.8 (0.6-1.1) 0.8 (0.6-1.1) 1.5 (0.6-3.4) 1.6 (0.7-3.6) 
P for linear trend  0.24 0.43 0.31 0.29 0.38 0.30 
Vegetables (servings/wk)        
    Tertile 1 (reference) 8.4 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 14.7 1.0 (0.5-2.2) 1.2 (0.6-2.6) 0.8 (0.6-1.1) 0.9 (0.7-1.2) 1.3 (0.6-2.9) 1.3 (0.6-2.9) 
    Tertile 3 23.8 0.7 (0.3-1.7) 1.0 (0.4-2.1) 0.7 (0.5-1.0) 0.9 (0.6-1.2) 0.9 (0.4-2.3) 1.1 (0.4-2.5) 
P for linear trend  0.61 0.88 0.05 0.38 0.91 0.74 
Dark green deep yellow vegetables (servings/wk)        
    Tertile 1 (reference) 1.4 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 3.5 1.0 (0.5-2.2) 1.0 (0.5-2.2) 0.9 (0.7-1.1) 0.8 (0.6-1.1) 1.4 (0.7-3.0) 1.4 (0.7-3.0) 
    Tertile 3 7.0 0.8 (0.4-1.8) 0.8 (0.4-1.8) 0.6 (0.4-0.9) 0.5 (0.4-0.8) 1.8 (0.8-4.2) 2.0 (0.9-4.5) 
P for linear trend  0.72 0.78 0.01 0.001 0.14 0.11 
Refined carbohydrates or cereals (servings/wk)        
    Tertile 1 (reference) 9.8 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 16.1 1.8 (0.8-4.1) 2.1 (0.9-4.8) 1.0 (0.7-1.4) 1.0 (0.8-1.4) 1.5 (0.6-3.4) 1.5 (0.7-3.4) 
    Tertile 3 23.1 2.2 (0.9-5.4) 3.5 (1.5-8.2) 1.2 (0.9-1.7) 1.2 (0.8-1.6) 1.6 (0.7-4.0) 1.8 (0.8-3.9) 
P for linear trend  0.09 0.007 0.33 0.48 0.27 0.17 
Red meat (servings/wk)        
    Tertile 1 (reference) 2.8 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 7.0 1.5 (0.7-3.1) 1.7 (0.8-3.6) 1.2 (0.9-1.6) 1.4 (1.0-1.8) 1.1 (0.5-2.5) 1.2 (0.5-2.6) 
    Tertile 3 11.2 1.3 (0.6-3.0) 2.0 (0.9-4.2) 0.9 (0.7-1.3) 1.5 (1.1-2.0) 0.9 (0.4-2.3) 1.1 (0.5-2.5) 
P for linear trend  0.42 0.07 0.90 0.003 0.94 0.74 
Dietary factor*Median intake/d in controlsMSI-H vs controls, OR (95% CI)
MSI-L/MSS vs controls, OR (95% CI)
MSI-H vs MSI-L/MSS, OR (95% CI)
Energy-adjustedNot adjusted for energy§Energy-adjustedNot adjusted for energy§Energy-adjustedNot adjusted for energy§
Total energy (kcal)        
    Tertile 1 1,166 — 1.0 — 1.0 — 1.0 
    Tertile 2 1,744 — 0.7 (0.3-1.6) — 1.1 (0.8-1.5) — 0.9 (0.4-2.0) 
    Tertile 3 2,446 — 1.3 (0.4-4.4) — 2.5 (1.7-3.5) — 1.2 (0.5-2.8) 
P for linear trend  — 0.93 — 0.0001 — 0.85 
Micronutrients        
β-Carotene        
    Tertile 1 (reference) 1,517 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 3,224 0.6 (0.3-1.2) 0.6 (0.3-1.3) 0.8 (0.6-1.1) 0.8 (0.6-1.1) 0.7 (0.3-1.6) 0.7 (0.3-1.6) 
    Tertile 3 6,383 0.4 (0.2-0.9) 0.4 (0.2-0.9) 0.7 (0.5-1.0) 0.8 (0.6-1.1) 0.7 (0.3-1.7) 0.7 (0.3-1.7) 
P for linear trend  0.02 0.03 0.06 0.17 0.38 0.36 
Vitamin C        
    Tertile 1 (reference) 57.7 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 115.5 0.9 (0.4-1.9) 0.9 (0.4-1.9) 0.9 (0.7-1.2) 0.9 (0.7-1.2) 1.1 (0.5-2.4) 1.2 (0.5-2.5) 
    Tertile 3 261.8 0.8 (0.4-1.8) 0.8 (0.3-1.7) 0.7 (0.5-1.0) 0.7 (0.5-0.9) 1.4 (0.6-3.0) 1.5 (0.7-3.1) 
P for linear trend  0.62 0.55 0.04 0.04 0.48 0.38 
Vitamin E        
    Tertile 1 (reference) 6.3 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 12.8 0.8 (0.4-1.7) 1.1 (0.5-2.2) 0.8 (0.6-1.1) 1.0 (0.8-1.4) 1.0 (0.5-2.1) 1.1 (0.6-2.3) 
    Tertile 3 168.5 0.5 (0.2-1.0) 0.6 (0.3-1.3) 0.6 (0.5-0.8) 0.8 (0.6-1.0) 0.7 (0.3-1.6) 0.8 (0.4-1.8) 
P for linear trend  0.10 0.35 0.005 0.19 0.51 0.76 
Calcium        
    Tertile 1 (reference) 410.9 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 694.0 1.1 (0.5-2.4) 1.2 (0.6-2.7) 0.6 (0.5-0.9) 0.9 (0.7-1.2) 2.0 (0.9-4.4) 2.1 (1.0-4.5) 
    Tertile 3 1,299.7 0.5 (0.2-1.3) 0.6 (0.3-1.5) 0.5 (0.4-0.7) 0.8 (0.6-1.1) 1.3 (0.5-3.4) 1.5 (0.7-3.3) 
P for linear trend  0.32 0.57 0.0001 0.23 0.33 0.17 
Macronutrients        
Carbohydrate (g/d)        
    Tertile 1 134.2 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 201.9 0.8 (0.4-1.8) 1.1 (0.5-2.5) 0.9 (0.6-1.2) 1.2 (0.9-1.6) 1.0 (0.4-2.3) 1.1 (0.5-2.5) 
    Tertile 3 284.4 0.4 (0.1-1.5) 1.1 (0.4-2.9) 1.0 (0.7-1.6) 2.1 (1.5-2.9) 0.5 (0.1-1.6) 0.6 (0.2-1.7) 
P for linear trend  0.24 0.77 0.96 0.0002 0.34 0.50 
E-carbohydrate (g/d)        
    Tertile 1 (reference) 124.7 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 188.0 1.0 (0.4-2.2) 1.3 (0.6-2.9) 0.9 (0.7-1.3) 1.2 (0.9-1.7) 1.1 (0.5-2.5) 1.1 (0.5-2.6) 
    Tertile 3 264.8 0.5 (0.1-1.7) 1.2 (0.5-3.2) 1.0 (0.6-1.6) 2.1 (1.5-2.9) 0.6 (0.2-1.9) 0.7 (0.3-1.9) 
P for linear trend  0.37 0.56 0.92 0.0002 0.50 0.67 
Protein (g/d)        
    Tertile 1 (reference) 41.9 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 62.7 0.8 (0.4-1.8) 1.2 (0.5-2.6) 0.8 (0.6-1.1) 1.1 (0.8-1.5) 0.8 (0.3-1.9) 0.8 (0.3-1.8) 
    Tertile 3 89.2 0.9 (0.3-2.8) 2.5 (1.0-6.1) 0.7 (0.4-1.0) 1.4 (1.0-2.0) 1.3 (0.4-3.9) 1.4 (0.6-3.0) 
P for linear trend  0.83 0.11 0.06 0.06 0.85 0.72 
Total fat (g/d)        
    Tertile 1 (reference) 44.3 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 71.0 0.6 (0.2-1.4) 0.8 (0.4-1.8) 0.8 (0.5-1.0) 1.0 (0.7-1.3) 0.7 (0.3-1.7) 0.8 (0.3-1.8) 
    Tertile 3 108.8 0.9 (0.3-2.7) 2.0 (1.0-4.4) 1.1 (0.7-1.8) 2.1 (1.6-2.9) 1.0 (0.3-2.9) 1.3 (0.6-2.9) 
P for linear trend  0.58 0.25 0.93 0.0005 0.80 0.81 
Saturated fat (g/d)        
    Tertile 1 (reference) 14.1 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 23.5 0.6 (0.2-1.3) 0.8 (0.3-1.7) 0.7 (0.5-1.0) 0.9 (0.7 1.3) 0.6 (0.2-1.4) 0.6 (0.3-1.5) 
    Tertile 3 37.3 0.6 (0.2-1.8) 1.5 (0.7-3.1) 1.0 (0.6-1.5) 2.0 (1.4-2.7) 0.7 (0.2-2.0) 1.0 (0.4-2.2) 
P for linear trend  0.26 0.63 0.35 0.002 0.33 0.68 
Alcohol (g/d)**        
    Nondrinkers (reference) 1.0 1.0 1.0 1.0 1.0 1.0 
    Drinkers 84.0 1.2 (0.6-2.3) 1.2 (0.6-2.3) 0.8 (0.6-1.1) 0.9 (0.7-1.1) 1.2 (0.6-2.4) 1.2 (0.6-2.5) 
P for linear trend  0.55 0.54 0.20 0.33 0.66 0.62 
Dietary fiber (g/d)        
    Tertile 1 (reference) 8.1 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 13.0 0.7 (0.3-1.6) 0.8 (0.4-1.7) 0.9 (0.6-1.2) 1.1 (0.8-1.5) 0.9 (0.4-2.0) 0.9 (0.4-2.0) 
    Tertile 3 19.0 0.4 (0.2-1.0) 0.5 (0.2-1.2) 0.6 (0.4-0.9) 1.0 (0.7-1.5) 0.7 (0.2-2.1) 0.7 (0.3-2.0) 
P for linear trend  0.08 0.15 0.01 0.76 0.55 0.53 
Food groups        
Fruits/juices (servings/wk)        
    Tertile 1 (reference) 2.8 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 7.0 1.1 (0.5-2.3) 1.0 (0.5-2.3) 0.9 (0.7-1.3) 0.9 (0.7-1.2) 1.3 (0.5-3.0) 1.3 (0.6-3.1) 
    Tertile 3 14.0 1.9 (0.8-4.3) 1.6 (0.7-3.7) 0.8 (0.6-1.1) 0.8 (0.6-1.1) 1.5 (0.6-3.4) 1.6 (0.7-3.6) 
P for linear trend  0.24 0.43 0.31 0.29 0.38 0.30 
Vegetables (servings/wk)        
    Tertile 1 (reference) 8.4 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 14.7 1.0 (0.5-2.2) 1.2 (0.6-2.6) 0.8 (0.6-1.1) 0.9 (0.7-1.2) 1.3 (0.6-2.9) 1.3 (0.6-2.9) 
    Tertile 3 23.8 0.7 (0.3-1.7) 1.0 (0.4-2.1) 0.7 (0.5-1.0) 0.9 (0.6-1.2) 0.9 (0.4-2.3) 1.1 (0.4-2.5) 
P for linear trend  0.61 0.88 0.05 0.38 0.91 0.74 
Dark green deep yellow vegetables (servings/wk)        
    Tertile 1 (reference) 1.4 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 3.5 1.0 (0.5-2.2) 1.0 (0.5-2.2) 0.9 (0.7-1.1) 0.8 (0.6-1.1) 1.4 (0.7-3.0) 1.4 (0.7-3.0) 
    Tertile 3 7.0 0.8 (0.4-1.8) 0.8 (0.4-1.8) 0.6 (0.4-0.9) 0.5 (0.4-0.8) 1.8 (0.8-4.2) 2.0 (0.9-4.5) 
P for linear trend  0.72 0.78 0.01 0.001 0.14 0.11 
Refined carbohydrates or cereals (servings/wk)        
    Tertile 1 (reference) 9.8 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 16.1 1.8 (0.8-4.1) 2.1 (0.9-4.8) 1.0 (0.7-1.4) 1.0 (0.8-1.4) 1.5 (0.6-3.4) 1.5 (0.7-3.4) 
    Tertile 3 23.1 2.2 (0.9-5.4) 3.5 (1.5-8.2) 1.2 (0.9-1.7) 1.2 (0.8-1.6) 1.6 (0.7-4.0) 1.8 (0.8-3.9) 
P for linear trend  0.09 0.007 0.33 0.48 0.27 0.17 
Red meat (servings/wk)        
    Tertile 1 (reference) 2.8 1.0 1.0 1.0 1.0 1.0 1.0 
    Tertile 2 7.0 1.5 (0.7-3.1) 1.7 (0.8-3.6) 1.2 (0.9-1.6) 1.4 (1.0-1.8) 1.1 (0.5-2.5) 1.2 (0.5-2.6) 
    Tertile 3 11.2 1.3 (0.6-3.0) 2.0 (0.9-4.2) 0.9 (0.7-1.3) 1.5 (1.1-2.0) 0.9 (0.4-2.3) 1.1 (0.5-2.5) 
P for linear trend  0.42 0.07 0.90 0.003 0.94 0.74 

NOTE: All data are for the reference year, which is the year before diagnosis for cases and the year before the interview for controls.

*

Cutoffs based on intakes among controls.

For columns 3-5, the reference groups are controls, MSI-L/MSS, and controls, respectively.

Age, sex, race, the offset term, and total energy were included in all energy-adjusted logistic regression models. All the other variables in Tables 1 and 2, except current BMI, were evaluated as potential confounders. A variable was included in the final model if it changed the variable coefficient of interest by least 15%. The set of confounders in the models varied for each dietary factor.

§

Age, sex, race, and the offset term were included in all the non-energy-adjusted logistic regression models. All the other variables in Tables 1 and 2, except total energy and current BMI, were evaluated as potential confounders. A variable was included in the final model if it changed the variable coefficient of interest by least 15%. The set of confounders in the models varied for each dietary factor.

Intakes are from food sources (food frequency questionnaire) and dietary supplements.

Effective carbohydrates defined as [total carbohydrate (g/d)] − [total fiber (g/d)].

**

Two-thirds of participants did not report consuming alcohol.

MSI is one of the molecular mechanisms leading to genomic instability in colon carcinogenesis. Few studies have examined the possible associations of dietary factors with MSI. Using data from participants enrolled in the NCCCS, we compared intakes of total energy, macronutrients, micronutrients, and food groups in sporadic colon cancer cases and population-based controls in relation to MSI status. According to our results, consumption of refined carbohydrates (cereals) and red meat may promote the development of MSI-H tumors whereas β-carotene was inversely associated with MSI-H. In case-case comparisons, no dietary factors were significantly differently related to MSI-H compared with MSI-L/MSS tumors. We also observed significant inverse associations between MSI-L/MSS and energy-adjusted vitamin C, vitamin E, calcium, dietary fiber, and dark green vegetables and positive associations with total energy and non-energy-adjusted intakes of red meat and most macronutrients.

In MSI-H versus control comparisons, β-carotene was significantly inversely associated with MSI-H. Epidemiologic studies suggest that high consumption of fruits and vegetables, particularly those high in carotenoids, may play a role in preventing colon cancer. In previous analysis in our study sample, β-carotene was significantly associated with a reduction in colon cancer risk (20). There are few studies that have examined β-carotene in relation to MSI and colon cancer. Similar to our findings, Slattery et al. (11) observed a trend toward a protective effect of β-carotene on MSI-H, but the results were not statistically significant. Although the mechanisms by which carotenoids may lower cancer risk are not entirely understood, possibilities include their antioxidant activities, inhibition of cell proliferation, induction of detoxifying enzymes (26), stimulation of gap junction communication, and enhancement of immune function (27). β-Carotene may also inhibit growth and promote apoptosis in colon cancer through the redox regulation of nuclear factor-κB activity (28).

Our finding of a positive association between refined carbohydrates and MSI-H is in agreement with a report by Slattery et al. (11), whose data were suggestive of an association between high refined grains and increased likelihood of having MSI-positive colonic tumors. High consumption of refined carbohydrates and factors related to insulin resistance have been associated with colon cancer. Epidemiologic studies that have shown associations of glycemic index and load, glucose intolerance, and insulin with colon cancer (29, 30) suggest a role for refined carbohydrates in the etiology of colon cancer. The mechanism by which refined carbohydrates influence MSI in colon cancer is unknown. One possibility is that glucose may induce changes in gene expression. For example, in a recent study, Fornoni et al. observed that exposure of mesangial cells to high glucose concentrations in vitro induced MSI. Thus, dietary sugars, such as refined glucose, may create conditions to produce significant structural and biological alterations that might include genomic instability within DNA (31).

High intake of red meat was positively associated with MSI-H tumors compared with controls, suggesting that red meat intake may contribute to microsatellite unstable tumors in the colon. Similar observations have been reported by some but not all studies. Wu et al. found a positive association between intake of well-done red meat and MSI-H in a case-only study but did not observe a significant association with red meat overall (10). Slattery et al. observed no associations between red meat intake and MSI status in a population-based case-control study of colon cancer (11). Other studies have reported positive associations between red meat consumption and MSI-L/MSS but not MSI-H colon tumors (12). Red meat intake has also been positively associated with MSI-H in gastric cancer (32). Reasons for the dissimilarities between our study and others may relate to differences in meat cooking practices and/or differences in composition of study populations: more specifically, our study population was racially diverse with a sizeable number of African Americans, whereas most studies in the literature consist predominantly of Caucasians.

The underlying mechanism between red meat intake and MSI-H is unclear; however, there are several possible explanations for the observed associations. Red meat prepared at high temperatures produces heterocyclic amines, which are bulk-adduct-forming agents that have been associated with increased cancer risk. Colon cells may also be more susceptible to DNA damage due to reduced MMR activity (33-35). In vitro cell studies suggest that dietary factors may modulate the relationship between MMR and cancer risk. In one study, murine cells that were deficient in MMR activity were more susceptible to DNA damage and likely to escape apoptosis following exposure to chronic oxidative stress (33, 36). Thus, in susceptible individuals with decreased MMR function, the genes involved in the MSI-H pathway may be more susceptible to mutagens in red meat that are produced during cooking. Another possible explanation for the association between red meat and MSI-H tumors relates to dietary heme in red meat. Heme has been proposed as the link between red meat intake and increased risk of colon cancer because of its cytotoxic and proliferative effects on colonic cells. The increased proliferation may lead to an elevated frequency of spontaneous mutations in target genes that may be repaired less efficiently in susceptible individuals (32).

Compared with controls, we observed significant associations between MSI-L/MSS and several nutrients as noted above. Reasons why different dietary factors may be associated with MSI-L/MSS versus MSI-H tumors are not clear. We also note that in case-case comparisons many of the dietary factors that were associated with MSI-H were also associated with MSI-L/MSS, but no dietary variables were significantly differently related to MSI-H compared with MSI-L/MSS tumors, which may be due to the small number of participants with MSI-H tumors (n = 49). Alternatively, this finding may suggest that factors other than those related to diet might be involved in the pathology of microsatellite unstable tumors.

Our findings did not suggest racial differences in the associations of dietary factors with MSI. This was somewhat surprising because the available evidence suggests that MSI may be more frequent among African Americans compared with Whites and the elevated risk of MSI tumors among African Americans has been postulated to be related to dietary differences (37). Possibly, the small number of MSI-H cases precluded our ability to detect statistically significant differences by race. However, there were no differences between African Americans and Whites for comparisons between controls and cases with MSI-L/MSS tumors, for which we had larger numbers. Nonetheless, given that there are racial differences in colon cancer risk and diet-colon cancer associations (1, 13, 20, 21, 38), the influence of diet on MSI in colon cancer merits further exploration in larger, racially diverse study samples.

The strengths of our study include a population-based study design, a racially diverse population with a large number of African Americans, rapid case ascertainment, and detailed exposure information obtained through in-person interviews. As with other case-control studies, selection and differential bias are some potential limitations. Specifically, selection bias is a concern due to the lower participation rates of controls, and differential recall between cases and controls is a possibility because exposure information was collected after diagnosis of the disease; in particular, cases may recall dietary exposures differently from controls because of the presence of their illness and/or symptoms. Nonetheless, we made every effort to reduce bias; for example, we limited most of our exposure information to 1 year before diagnosis for cases or the year before the interview for controls and exposure information was collected as soon as possible after diagnosis. Our response rates were also comparable with previous population-based case-control studies (11, 32, 39, 40).

In summary, our findings indicate that refined carbohydrates and red meat consumption may promote the development of MSI-H colon cancer, whereas β-carotene may be associated with a lower risk of MSI-H tumors. It is possible that differences between diet and colon cancer risk by phenotype may in part explain inconsistencies across studies of diet and cancer. Identification of the underlying mechanisms by which dietary factors affect tumor genomic stability might prove useful in the treatment of colon cancer as well as in the incorporation of dietary factors into colon cancer prevention strategies.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
American Cancer Society. Cancer facts and figures. Atlanta (GA): American Cancer Society; 2003.
2
Grady WM. Genomic instability and colon cancer.
Cancer Metastasis Rev
2004 Jan-Jun
;
23
:
11
–27.
3
Boland CR, Thibodeau SN, Hamilton SR, et al. A National Cancer Institute workshop on microsatellite instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer.
Cancer Res
1998
;
58
:
5248
–57.
4
Thibodeau SN, Bren G, Schaid D. Microsatellite instability in cancer of the proximal colon.
Science (Washington DC)
1993
;
260
:
816
–9.
5
Aaltonen LA, Peltomäki P, Leach F, et al. Clues to the pathogenesis of familial colorectal cancer.
Science (Washington DC)
1993
;
260
:
812
–6.
6
Potter JD. Colorectal cancer: molecules and populations.
J Natl Cancer Inst
1999
;
91
:
916
–32.
7
Food, nutrition, and the prevention of cancer: a global perspective. World Cancer Research Fund/American Institute for Cancer Research; 1997.
8
Slattery ML. Diet, lifestyle, and colon cancer.
Semin Gastrointest Dis
2000
;
11
:
142
–6.
9
Lichtenstein P, Holm NV, Verkasalo PK, et al. Environmental and heritable factors in the causation of cancer—analyses of cohorts of twins from Sweden, Denmark, and Finland.
N Engl J Med
2000
;
13
:
78
–85.
10
Wu AH, Shibata D, Yu MC, Lai MY, Ross RK. Dietary heterocyclic amines and microsatellite instability in colon adenocarcinomas.
Carcinogenesis
2001
;
22
:
1681
–4.
11
Slattery ML, Anderson K, Curtin K, Ma KN, Schaffer D, Samowitz W. Dietary intake and microsatellite instability in colon tumors.
Int J Cancer
2001
;
93
:
601
–7.
12
Diergaarde B, Braam H, van Muijen GN, Ligtenberg MJ, Kok FJ, Kampman E. Dietary factors and microsatellite instability in sporadic colon carcinomas.
Cancer Epidemiol Biomarkers Prev
2003
;
12
:
1130
–6.
13
Satia-Abouta J, Galanko JA, Potter JD, Ammerman A, Martin CF, Sandler RS. Associations of total energy and macronutrients with colon cancer risk in African Americans and Whites: results from the North Carolina Colon Cancer Study.
Am J Epidemiol
2003
;
158
:
951
–62.
14
Aldrich TE, Vann D, Moorman PG, Newman B. Rapid reporting of cancer incidence in a population-based study of breast cancer: one constructive use of a central cancer registry.
Breast Cancer Res Treat 1995
1995 Jul
;
35
:
61
–4.
15
Wacholder S, Weinberg CR. Flexible maximum likelihood methods for assessing joint effects in case-control studies with complex sampling.
Biometrics
1994
;
50
:
350
–7.
16
Weinberg CR, Sandler DP. Randomized recruitment in case-control studies.
Am J Epidemiol
1991
;
134
:
421
–32.
17
Gerber AM, James SA, Ammerman AS, et al. Socioeconomic status and electrolyte intake in black adults: the Pitt County Study.
Am J Public Health
1991
;
81
:
1608
–12.
18
Patterson RE, Kristal AR, Tinker LF, Carter RA, Bolton MP, Agurs-Collins T. Measurement characteristics of the Women's Health Initiative food frequency questionnaire.
Ann Epidemiol
1999
;
9
:
178
–87.
19
Block G, Hartman AM, Dresser CM, Carroll MD, Gannon J, Gardner L. A data-based approach to diet questionnaire design and testing.
Am J Epidemiol
1986
;
124
:
453
–69.
20
Satia-Abouta J, Galanko JA, Martin CF, Potter JD, Ammerman A, Sandler RS. Associations of micronutrients with colon cancer risk in African Americans and whites: results from the North Carolina Colon Cancer Study.
Cancer Epidemiol Biomarkers Prev
2003
;
12
:
747
–54.
21
Satia-Abouta J, Galanko JA, Martin CF, Ammerman A, Sandler RS. Food groups and colon cancer risk in African-Americans and Caucasians.
Int J Cancer
2004
;
109
:
728
–36.
22
Willett W. Nutritional epidemiology. 2nd ed. New York (NY): Oxford University Press; 1998.
23
Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: executive summary. Expert Panel on the Identification, Evaluation, and Treatment of Overweight in Adults.
Am J Clin Nutr
1998
;
68
:
899
–917.
24
Blair SN, Haskell WL, Ho P, et al. Assessment of habitual physical activity by a seven-day recall in a community survey and controlled experiments.
Am J Epidemiol
1985
;
122
:
794
–804.
25
Richardson MT, Ainsworth BE, Jacobs DR, Leon AS. Validation of the Stanford 7-day recall to assess habitual physical activity.
Am J Epidemiol
2001
;
11
:
145
–53.
26
Cooper DA, Eldridge AL, Peters JC. Dietary carotenoids and certain cancers, heart disease, and age-related macular degeneration: a review of recent research.
Nutr Rev
1999
;
57
:
201
–14.
27
Zhang LX, Cooney RV, Bertram JS. Carotenoids enhance gap junctional communication and inhibit lipid peroxidation in C3H/10T1/2 cells: relationship to their cancer chemopreventive action.
Carcinogenesis
1991
;
12
:
2109
–14.
28
Palozza P, Serini S, Torsello A, et al. β-Carotene regulates NF-κB DNA-binding activity by a redox mechanism in human leukemia and colon adenocarcinoma cells.
J Nutr
2003
;
133
:
381
–8.
29
Giovannucci E. Obesity, gender, and colon cancer. Gut 2002;51:147.
30
Franceschi S, Dal Maso L, Augustin L, et al. Dietary glycemic load and colorectal cancer risk.
Ann Oncol
2001
;
12
:
173
–8.
31
Fornoni A, Lenz O, Striker LJ, Striker GE. Glucose induces clonal selection and reversible dinucleotide repeat expansion in mesangial cells isolated from glomerulosclerosis-prone mice.
Diabetes
2003
;
52
:
2594
–602.
32
Palli D, Russo A, Ottini L, et al. Red meat, family history, and increased risk of gastric cancer with microsatellite instability.
Cancer Res
2001
;
61
:
5415
–9.
33
Buermeyer AB, Wilson-Van Patten C, Baker SM, Liskay RM. The human MLH1 cDNA complements DNA mismatch repair defects in Mlh1-deficient mouse embryonic fibroblasts.
Cancer Res
1999
;
59
:
538
–41.
34
Chang CL, Marra G, Chauhan DP, et al. Oxidative stress inactivates the DNA mismatch repair system.
Am J Physiol Cell Physiol
2002
;
283
:
C148
–54.
35
Gasche C, Chang CL, Rhees J, Goel A, Boland CR. Oxidative stress increases frameshift mutations in human colorectal cancer cells.
Cancer Res
2001
;
61
:
7444
–8.
36
DeWeese TL, Shipman JM, Larrier NA, et al. Mouse embryonic stem cells carrying one or two defective Msh2 alleles respond abnormally to oxidative stress inflicted by low-level radiation.
Proc Natl Acad Sci U S A
1998
;
95
:
11915
–20.
37
Ashktorab H, Smoot DT, Carethers JM, et al. High incidence of microsatellite instability in colorectal cancer from African Americans.
Clin Cancer Res
2003
;
9
:
1112
–7.
38
Carethers JM. Racial and ethnic factors in the genetic pathogenesis of colorectal cancer.
J Assoc Acad Minor Phys
1999
;
10
:
59
–67.
39
Moorman PG, Newman B, Millikan RC, Tse CK, Sandler DP. Participation rates in a case-control study: the impact of age, race, and race of interviewer.
Ann Epidemiol
1999
;
9
:
188
–95.
40
Slattery ML, Edwards SL, Caan BJ, Kerber RA, Potter JD. Response rates among control subjects in case-control studies.
Ann Epidemiol
1995
;
5
:
245
–9.