Abstract
Long-chain ω-3 polyunsaturated fatty acids (PUFAs) may have antineoplastic properties in the colon. The authors examined the association between intakes of different PUFAs and distal large bowel cancer in a population-based case-control study of 1,503 whites (716 cases; 787 controls) and 369 African Americans (213 cases; 156 controls) in North Carolina (2001–2006). Unconditional logistic regression was used to estimate odds ratios and 95% confidence intervals for distal large bowel cancer risk in relation to quartiles of PUFA intake. Increased consumption of long-chain ω-3 PUFAs was associated with reduced risk of distal large bowel cancer in whites (multivariable odds ratios = 0.88 (95% confidence interval (CI): 0.63, 1.22), 0.69 (95% CI: 0.49, 0.98), and 0.49 (95% CI: 0.34, 0.71) for second, third, and highest vs. lowest quartile) (Ptrend < 0.01). Intake of individual eicosapentaenoic acids and docosahexaenoic acids was inversely related to distal large bowel cancer risk, whereas the ratio of ω-6 to long-chain ω-3 PUFAs was associated with increased risk of distal large bowel cancer in whites, but not among African Americans (Pinteraction < 0.05). Study results support the hypothesis that long-chain ω-3 PUFAs have beneficial effects in colorectal carcinogenesis. Whether or not the possible benefit of long-chain ω-3 PUFAs varies by race warrants further evaluation.
Keywords: colorectal neoplasms; fatty acids, omega-3; fatty acids, unsaturated; intestine, large
Experimental and clinical data have shown that ω-3 polyunsaturated fatty acids (PUFAs) have beneficial effects in colorectal carcinogenesis, including reduced tumor growth (1, 2), suppression of angiogenesis (3), and inhibition of metastasis (4, 5). In particular, long-chain, marine ω-3 PUFAs, such as eicosapentaenoic acid and docosahexaenoic acid, are effective competitors against arachidonic acid for the composition of membrane phospholipids and cyclooxygenase (COX) enzyme, inhibiting production of proinflammatory and tumorigenic prostaglandins and leukotrienes (6, 7). Supplementation with fish oil containing eicosapentaenoic acid and docosahexaenoic acid has been shown to decrease levels of inflammatory mediators and to reduce dependency on nonsteroidal antiinflammatory drugs (NSAIDs) and other pain medications in patients with arthritis (8). In a small clinical trial, fish oil supplementation significantly reduced cell proliferation on rectal mucosa in patients with colorectal adenomas compared with placebo after 12 weeks (9).
In contrast, epidemiologic evidence has been inconsistent. Only 1 of 9 prospective studies included in an earlier systematic review reported an inverse association between ω-3 PUFA intake and colorectal cancer (10). More recently, intake of ω-3 fatty acids has been associated with lower risk of colorectal cancer in both case-control (11) and prospective (12) studies. Other studies found the inverse association only among specific subgroups (13–15). We therefore examined the association between intakes of different PUFAs and distal large bowel cancer in a population-based case-control study of whites and African Americans. We also evaluated race, NSAID use, obesity, and tumor location as potential effect measure modifiers of the relation between intake of polyunsaturated fatty acids and distal large bowel cancer. Race and tumor location were chosen on the basis of previous findings suggesting etiologic heterogeneity of colorectal cancer by race (16) and tumor location (17), and NSAID use and obesity were chosen a priori because of their shared implications in COX-associated inflammation pathways (18, 19).
MATERIALS AND METHODS
North Carolina Colon Cancer Study II
The North Carolina Colon Cancer Study II is a population-based case-control study in 33 counties in the central and eastern parts of North Carolina. To be eligible for the study, subjects had to be residents of the selected counties, aged between 40 and 80 years, African American or white race, and able to complete an interview in English. Subjects had to have a North Carolina driver's license or identification card. The cases were patients with a first diagnosis of invasive adenocarcinoma in the sigmoid, rectosigmoid, or rectum diagnosed between October 1, 2001, and May 31, 2006. A total of 1,831 eligible cases were identified through the rapid ascertainment system of the North Carolina Central Cancer Registry. The procedure for case ascertainment and its effectiveness has been described (20). Of the cases identified, 57 (3%) were excluded because of physician refusal, and 357 (20%) were found to be ineligible. Of the remaining 1,417 eligible cases, 118 (8%) could not be contacted, 242 (17%) refused to participate, and 1,057 (75%) completed an in-person interview.
Controls were randomly selected from records of the North Carolina Division of Motor Vehicles, on the basis of sampling probabilities within blocks defined by 5-year age group, sex, and race, by use of randomized recruitment (21). A total of 2,345 subjects were identified as potentially eligible controls, and 1,827 (78%) met eligibility criteria. Of these, 325 (18%) could not be contacted, 483 (26%) refused to participate, and 1,019 (56% of eligible controls) completed an in-person interview.
Of a total 2,076 subjects, dietary information was available for 2,044. We further excluded 172 subjects with implausible energy intake: if daily energy intake was less than 500 kcal (n = 4) or greater than 3,500 kcal (n = 37) in women or less than 800 kcal (n = 12) or greater than 4,000 kcal (n = 119) in men (22). The final data set for analysis included 1,872 subjects (929 cases and 943 controls). The study was approved by the Institutional Review Board at the University of North Carolina School of Medicine, and all subjects provided written informed consent.
Data collection
All subjects were interviewed in person in their home or in another convenient location by trained nurse interviewers using a standard questionnaire. The questionnaire included information regarding demographic characteristics, socioeconomic indicators (such as years of education, income level, and occupation history), medication use, medical history, family history of cancer, smoking habits, and weight 1 year prior to diagnosis (for cases) or interview (for controls). Subjects were also asked about the type and duration of various activities engaged in on typical week (working) and weekend (nonworking) days. Then the energy expenditure involved in each activity was retrieved from the “Compendium of Physical Activities.” Metabolic equivalents were calculated by multiplying the number of hours spent in these activities, the number of metabolic equivalents for the category, and body weight (23). Current weight and height were measured at the time of interview. Cancer stage (local, regional, distant, or unknown), tumor location (sigmoid, rectosigmoid, or rectum), and other diagnosis-related data were available from the Central Cancer Registry.
Dietary assessment
Dietary information was collected by using the Diet History Questionnaire, a validated food frequency questionnaire with 124 food items developed at the National Cancer Institute (24). Subjects were asked about a common unit or portion size for each food and the average consumption frequency of the food in the past year. Information about types of oil or other fats usually used for frying, sautéing, basting, or marinating was also collected. Nutrient intake was calculated by using the Diet*Calc software program provided by the National Cancer Institute (http://riskfactor.cancer.gov/DHQ/dietcalc/) (25). Briefly, an individual's nutrient intake was computed by multiplying the frequency of consumption of each food by gender-specific mean nutrient or food group values and portion size based on national dietary intake data from the 1994–1996 US Department of Agriculture's Continuing Survey of Food Intake by Individuals (CSFII).
Intakes of 19 different fatty acids were estimated including 7 PUFAs: linoleic acid (18:2), α-linolenic acid (18:3), stearidonic acid (18:4), arachidonic acid (20:4), eicosapentaenoic acid (20:5), docosapentaenoic acid (22:5), and docosahexaenoic acid (22:6). Total ω-6 PUFA intake was calculated as the sum of linoleic acid and arachidonic acid, and the total ω-3 PUFA intake was the sum of α-linolenic acid, stearidonic acid, eicosapentaenoic acid, docosapentaenoic acid, and docosahexaenoic acid. Long-chain ω-3 PUFA intake was computed to reflect the sum of eicosapentaenoic acid, docosapentaenoic acid, and docosahexaenoic acid. Our nutrient analysis did not allow us to examine food source-specific nutrient intake. In the Western diet, however, approximately 70% of dietary long-chain ω-3 fatty acids come from fish and other seafood, whereas cereal-based dishes and oils are the major dietary source of short-chain ω-3 fatty acids (26). These food items are well covered by the food frequency questionnaire used in this study.
Statistical analysis
Selected characteristics were compared between cases and controls. Subjects were divided into 4 groups based on the distribution of each PUFA intake among controls. Unconditional logistic regression was used to estimate odds ratios and 95% confidence intervals for the association between intakes of different PUFAs and distal large bowel cancer. Linear trends across quartiles of PUFA intake were assessed by modeling the categorical variable as a continuous variable.
All logistic regression models included age (40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, or 75–79 years), sex, race, daily energy intake during the past year (quintiles based on the distribution among controls), and an offset term to adjust for sampling probability. A randomized recruitment method was used for selection of cases and controls to achieve oversampling of cases in certain categories and their probability-matched controls. Because this method systematically introduces a bias in the estimated odds ratios, the odds ratios should be corrected by including a fixed offset term corresponding to the sampling categories which, in this case, were of age, sex, race, and case-control status (21). Results without the inclusion of the offset term, however, were largely the same.
A fully adjusted model was constructed including covariates that are potentially associated with colorectal cancer or polyunsaturated fatty acid intake as follows: years of education (<12, 12–15, or ≥16 years), annual household income (<$15,000, $15,000–$24,999, $25,000–$49,999, or ≥$50,000), smoking status (current, past, or never smoker) at the time of diagnosis (for cases) or interview (for controls), family history of colorectal cancer (having had at least 1 first-degree relative with colorectal cancer or not), NSAID use in the past 5 years (ever or never), hormone replacement therapy use for at least 1 year in women, body mass index in the past year (<25, 25–29.9, or ≥30 kg/m2), endoscopic screening use in the past 10 years (yes or no), and intakes of total fat, red meat, fruit and vegetables, and vitamin E during the past year (quintiles based on the distribution among controls). Vitamin E was evaluated as a potential covariate on the basis of findings from experimental studies showing that long-chain ω-3 ceased to exert its antitumorigenic effects after concurrent addition of vitamin E (7). Endoscopic screening use, years of education, body mass index, and intakes of red meat and vitamin E resulted in more than 10% changes in the beta coefficients for highest quartile of long-chain ω-3 PUFA intake in relation to distal large bowel cancer when covariates were removed; therefore, they were included in the final multivariable model in addition to age, sex, race, daily energy intake, and an offset term.
Likelihood ratio tests comparing models with and without multiplicative interaction terms were used to assess potential effect measure modification of the association between PUFA intakes and distal large bowel cancer by race, NSAID use, and self-reported body mass index. Subsite-specific effect estimates were modeled by using multinomial logistic regression models. All statistical tests were 2 sided and had an α level of 0.05. STATA, version 10.0, software (StataCorp LP, College Station, Texas) was used for all statistical analyses.
RESULTS
Among 929 cases and 943 controls with adequate dietary information, the mean age was 62 years and 64 years, respectively (Table 1). Compared with controls, cases were more likely to be African Americans and less likely to take NSAIDs and to receive endoscopic screening. Cases also tended to be overweight or obese and had higher intakes of polyunsaturated fat, total fat, and total energy, whereas intakes of fruit and vegetables and long-chain ω-3 PUFAs were lower in cases compared with controls. Interestingly, cases were more physically active than controls, but the positive association disappeared after the minimal adjustment for age, sex, and race (data not shown).
Table 1.
Characteristics of Cases and Controls, North Carolina Colon Cancer Study II, 2001–2006
| Cases (n = 929) |
Controls (n = 943) |
P Value | |||||
| No. | % | Mean (SD) | No. | % | Mean (SD) | ||
| Age, years | <0.01 | ||||||
| 40–49 | 119 | 12.8 | 93 | 9.9 | |||
| 50–59 | 249 | 26.8 | 216 | 22.9 | |||
| 60–69 | 287 | 30.9 | 300 | 31.8 | |||
| 70–79 | 274 | 29.5 | 334 | 35.4 | |||
| Mean years | 62 | 64 | |||||
| Females | 441 | 44.2 | 400 | 42.4 | 0.43 | ||
| African Americans | 213 | 22.9 | 156 | 16.5 | <0.01 | ||
| Years of education | <0.01 | ||||||
| <12 | 174 | 19.2 | 109 | 11.8 | |||
| 12–15 | 481 | 53.1 | 460 | 49.8 | |||
| ≥16 | 251 | 27.7 | 354 | 38.4 | |||
| Annual household income | 0.04 | ||||||
| <$15,000 | 157 | 18.4 | 124 | 14.3 | |||
| $15,000–$24,999 | 145 | 17 | 135 | 15.6 | |||
| $25,000–$49,999 | 211 | 24.7 | 212 | 24.5 | |||
| ≥$50,000 | 341 | 39.9 | 396 | 45.7 | |||
| Smoking status | 0.36 | ||||||
| Never | 340 | 37.6 | 352 | 38.4 | |||
| Past | 417 | 46 | 437 | 47.7 | |||
| Current | 148 | 16.4 | 128 | 14 | |||
| Body mass index 1 year prior | <0.01 | ||||||
| Normal | 198 | 22.7 | 271 | 30.2 | |||
| Overweight | 324 | 37.2 | 359 | 40 | |||
| Obese | 350 | 40.1 | 268 | 29.8 | |||
| Physical activity, MET-minutes/day | <0.01 | ||||||
| <1,849 | 249 | 28.6 | 217 | 24.5 | |||
| 1,849–<1,962 | 188 | 21.6 | 222 | 25.1 | |||
| 1,962–<2,190 | 176 | 20.2 | 221 | 25 | |||
| ≥2,190 | 257 | 29.5 | 225 | 25.4 | |||
| Use of NSAIDs | 689 | 76.1 | 771 | 83.6 | <0.01 | ||
| HRT use in females | 147 | 35.9 | 190 | 49.2 | <0.01 | ||
| Family history of colorectal cancer | 120 | 18.4 | 97 | 14 | 0.03 | ||
| Endoscopic screening | 202 | 22.4 | 583 | 63.4 | <0.01 | ||
| Dietary intakes | |||||||
| Calories, kcal/day | 2,180.5 (752.8) | 2,056.8 (694.9) | <0.01 | ||||
| Fat, g/day | 87.3 (36.3) | 82.2 (35.2) | <0.01 | ||||
| Red meat, g/day | 73.8 (52.8) | 67.8 (46.5) | <0.01 | ||||
| Fruits/vegetables, servings/day | 2.6 (1.5) | 2.8 (1.6) | <0.01 | ||||
| Vitamin E, mg/day | 9.7 (4.5) | 9.9 (4.6) | 0.29 | ||||
| PUFAs, g/day | 20.2 (8.9) | 19.3 (8.6) | 0.03 | ||||
| ω-6 PUFAs, g/day | 18.1 (8.1) | 17.3 (7.8) | 0.03 | ||||
| Arachidonic acid, mg/day | 120.0 (61.9) | 115.0 (60.1) | 0.08 | ||||
| ω-3 PUFAs, g/day | 1.9 (0.8) | 1.9 (0.8) | 0.03 | ||||
| Short-chain ω-3 PUFAs, g/day | 1.8 (0.8) | 1.7 (0.8) | <0.01 | ||||
| α-Linolenic acid, g/day | 1.8 (0.8) | 1.7 (0.8) | <0.01 | ||||
| Long-chain ω-3 PUFAs, mg/day | 124.6 (112.2) | 139.6 (121.6) | <0.01 | ||||
| Eicosapentaenoic acid, mg/day | 35.1 (37.7) | 40.8 (41.1) | <0.01 | ||||
| Docosahexaenoic acid, mg/day | 73.2 (61.9) | 81.6 (67.6) | <0.01 | ||||
| ω-6:ω-3 PUFA ratio | 9.5 (2.1) | 9.5 (2.2) | 0.96 | ||||
| ω-6:long-chain ω-3 PUFA ratio | 258.3 (272.8) | 223.4 (249.4) | <0.01 | ||||
Abbreviations: HRT, hormone replacement therapy; MET, metabolic equivalent; NSAID, nonsteroidal antiinflammatory drug; PUFA, polyunsaturated fatty acid.
Long-chain ω-3 PUFA intake was inversely related to distal large bowel cancer in a dose-response manner (Ptrend < 0.01) (Table 2). After adjustment for potential confounding variables, the inverse association remained significant for the highest quartile of long-chain ω-3 fatty acid intake. The multivariable odds ratios were 0.96 (95% confidence interval (CI): 0.71, 1.31) for the second quartile, 0.75 (95% CI: 0.54, 1.04) for the third quartile, and 0.61 (95% CI: 0.44, 0.86) for the fourth quartile in comparison with the lowest quartile, respectively. Further adjustment for vitamin D, which is rare in the diet but is most abundantly found in fatty fish as is long-chain ω-3 PUFA (27), did not change the estimated association between long-chain ω-3 PUFA and distal large bowel cancer. We also analyzed the data using the residual method for energy adjustment, but the estimates remained largely unchanged (data not shown). Intakes of individual long-chain ω-3 PUFAs, eicosapentaenoic acid, and docosahexaenoic acid were also similarly inversely associated with distal large bowel cancer risk in a dose-dependent manner (Ptrend < 0.01), whereas the ratio of ω-6 to long-chain ω-3 PUFA intake was associated with increased risk of distal large bowel cancer (for the highest vs. the lowest quartile: the multivariable odds ratio (OR) = 1.42 (95% CI: 1.04, 1.95)). On the other hand, intakes of total PUFAs, total ω-6 PUFAs, and short-chain ω-3 PUFAs were not related to distal large bowel cancer risk.
Table 2.
Association Between Quartiles of Polyunsaturated Fatty Acid Intake and Distal Large Bowel Cancer Risk, North Carolina Colon Cancer Study II, 2001–2006
| No. of Cases | No. of Controls | Model 1a |
Model 2b |
|||
| Odds Ratio | 95% Confidence Interval | Odds Ratio | 95% Confidence Interval | |||
| Total PUFAs, g/day | ||||||
| <13.1 | 196 | 235 | 1.00 | Referent | 1.00 | Referent |
| 13.1–<18.1 | 226 | 235 | 1.14 | 0.84, 1.54 | 1.16 | 0.81, 1.67 |
| 18.1–<24.4 | 242 | 238 | 1.07 | 0.76, 1.49 | 1.14 | 0.74, 1.76 |
| ≥24.4 | 265 | 235 | 0.93 | 0.63, 1.37 | 0.99 | 0.59, 1.67 |
| Ptrend | 0.63 | 0.95 | ||||
| Total ω-6 PUFAs, g/day | ||||||
| <11.5 | 196 | 232 | 1.00 | Referent | 1.00 | Referent |
| 11.5–<16.3 | 232 | 240 | 1.13 | 0.84, 1.53 | 1.09 | 0.76, 1.56 |
| 16.3–<22.0 | 242 | 238 | 1.02 | 0.73, 1.43 | 1.07 | 0.69, 1.65 |
| ≥22.0 | 259 | 233 | 0.88 | 0.59, 1.30 | 0.87 | 0.52, 1.47 |
| Ptrend | 0.42 | 0.61 | ||||
| Arachidonic acid, g/day | ||||||
| <0.07 | 154 | 198 | 1.00 | Referent | 1.00 | Referent |
| 0.07–<0.1 | 211 | 215 | 1.22 | 0.91, 1.65 | 1.17 | 0.82, 1.68 |
| 0.1–<0.15 | 297 | 272 | 1.27 | 0.93, 1.72 | 1.09 | 0.74, 1.59 |
| ≥0.15 | 267 | 258 | 0.93 | 0.66, 1.31 | 0.83 | 0.53, 1.31 |
| Ptrend | 0.67 | 0.36 | ||||
| Total ω-3 PUFAs, g/day | ||||||
| <1.27 | 198 | 233 | 1.00 | Referent | 1.00 | Referent |
| 1.27–<1.72 | 218 | 238 | 1.07 | 0.80, 1.43 | 1.10 | 0.78, 1.55 |
| 1.72–<2.31 | 247 | 235 | 1.08 | 0.79, 1.50 | 1.14 | 0.76, 1.67 |
| ≥2.31 | 266 | 237 | 0.95 | 0.66, 1.37 | 0.96 | 0.61, 1.51 |
| Ptrend | 0.80 | 0.86 | ||||
| Short-chain ω-3 PUFAs, g/day | ||||||
| <1.16 | 194 | 230 | 1.00 | Referent | 1.00 | Referent |
| 1.16–<1.57 | 194 | 236 | 0.97 | 0.72, 1.30 | 0.99 | 0.70, 1.41 |
| 1.57–<2.14 | 265 | 241 | 1.20 | 0.87, 1.64 | 1.23 | 0.84, 1.81 |
| ≥2.14 | 276 | 236 | 1.05 | 0.73, 1.52 | 1.04 | 0.87, 1.62 |
| Ptrend | 0.52 | 0.67 | ||||
| α-Linolenic acid, g/day | ||||||
| <1.16 | 198 | 233 | 1.00 | Referent | 1.00 | Referent |
| 1.16–<1.57 | 194 | 237 | 0.95 | 0.71, 1.28 | 0.98 | 0.69, 1.38 |
| 1.57–<2.13 | 257 | 235 | 1.18 | 0.86, 1.62 | 1.20 | 0.81, 1.76 |
| ≥2.13 | 280 | 238 | 1.05 | 0.73, 1.51 | 1.03 | 0.66, 1.60 |
| Ptrend | 0.53 | 0.72 | ||||
| Long-chain ω-3 PUFAs, g/day | ||||||
| <0.06 | 198 | 233 | 1.00 | Referent | 1.00 | Referent |
| 0.06–<0.11 | 194 | 237 | 0.79 | 0.61, 1.03 | 0.96 | 0.71, 1.31 |
| 0.11–<0.18 | 257 | 235 | 0.63 | 0.48, 0.83 | 0.75 | 0.54, 1.04 |
| ≥0.18 | 280 | 238 | 0.51 | 0.39, 0.68 | 0.61 | 0.44, 0.86 |
| Ptrend | <0.01 | <0.01 | ||||
| Eicosapentaenoic acid, g/day | ||||||
| <0.01 | 319 | 257 | 1.00 | Referent | 1.00 | Referent |
| 0.01–<0.03 | 266 | 279 | 0.72 | 0.56, 0.92 | 0.89 | 0.67, 1.17 |
| 0.03–<0.05 | 177 | 192 | 0.65 | 0.50, 0.86 | 0.77 | 0.55, 1.06 |
| ≥0.05 | 167 | 215 | 0.54 | 0.41, 0.71 | 0.65 | 0.47, 0.91 |
| Ptrend | <0.01 | <0.01 | ||||
| Docosahexaenoic acid, g/day | ||||||
| <0.04 | 263 | 216 | 1.00 | Referent | 1.00 | Referent |
| 0.04–<0.07 | 259 | 250 | 0.76 | 0.59, 0.99 | 0.90 | 0.66, 1.22 |
| 0.07–<0.11 | 212 | 236 | 0.62 | 0.47, 0.82 | 0.75 | 0.55, 1.03 |
| ≥0.11 | 195 | 241 | 0.49 | 0.37, 0.66 | 0.58 | 0.41, 0.81 |
| Ptrend | <0.01 | <0.01 | ||||
| ω-6:ω-3 PUFA ratio | ||||||
| <8.1 | 214 | 239 | 1.00 | Referent | 1.00 | Referent |
| 8.1–<9.2 | 259 | 238 | 1.17 | 0.90, 1.52 | 1.13 | 0.84, 1.53 |
| 9.2–<10.5 | 246 | 236 | 1.07 | 0.82, 1.39 | 1.00 | 0.74, 1.36 |
| ≥10.5 | 210 | 230 | 0.98 | 0.75, 1.29 | 1.08 | 0.78, 1.49 |
| Ptrend | <0.01 | 0.86 | ||||
| ω-6:long-chain ω-3 PUFA ratio | ||||||
| <90.3 | 171 | 235 | 1.00 | Referent | 1.00 | Referent |
| 90.3–<151.7 | 221 | 236 | 1.16 | 0.87, 1.53 | 1.05 | 0.76, 1.45 |
| 151.7–<254.7 | 227 | 236 | 1.27 | 0.96, 1.68 | 1.20 | 0.87, 1.65 |
| ≥254.7 | 310 | 236 | 1.75 | 1.34, 2.30 | 1.42 | 1.04, 1.95 |
| Ptrend | <0.01 | 0.02 | ||||
Abbreviation: PUFA, polyunsaturated fatty acid.
Adjusted for age (40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–79 years), race, sex, daily energy intake (<1,554.6, 1,554.6–<1,991.0, 1,991.0–<2,595.4, or ≥2,595.4 kcal/day), and offset term.
Additionally adjusted for years of education (<12, 12–15, ≥16 years), endoscopic screening in the past 10 years (yes or no), red meat intake (<34.7, 34.7–<57.1, 57.1–<95.8, ≥95.8 g/day), vitamin E intake (<6.6, 6.6–<9.2, 9.2–<12.4, ≥12.4 mg/day), and body mass index 1 year prior (<25, 25–29.9, ≥30 kg/m2).
When we stratified by race, the inverse association between higher intake of long-chain ω-3 PUFA and distal large bowel cancer risk was limited to whites (Pinteraction = 0.02) (Table 3). Among African Americans, the intake of long-chain ω-3 PUFAs was positively associated with distal large bowel cancer (for the highest vs. the lowest quartile: OR = 1.79 (95% CI: 0.85, 3.77)), although the trend was not significant (Ptrend = 0.49). Further adjustment for consumption of saturated fatty acid and proportion of fried fish out of total fish intake did not reverse the inverse association in African Americans (data not shown). We also evaluated NSAID use and obesity (body mass index, ≥30 kg/m2) as potential effect modifiers but found similar inverse associations across the strata. There was no evidence of interaction with race, NSAID use, or obesity for other types of PUFAs (data not shown).
Table 3.
Association Between Quartiles of Long-Chain ω-3 Polyunsaturated Fatty Acid Intake and Distal Large Bowel Cancer Risk According to Race, Nonsteroidal Antiinflammatory Drug Use, and Obesity, North Carolina Colon Cancer Study II, 2001–2006
| First Quartile (<0.06 g/day) |
Second Quartile (0.06–<0.11 g/day) |
Third Quartile (0.11–<0.18 g/day) |
Fourth Quartile (≥0.18 g/day) |
Ptrend | |||||
| Odds Ratioa | 95% Confidence Interval | Odds Ratioa | 95% Confidence Interval | Odds Ratioa | 95% Confidence Interval | Odds Ratioa | 95% Confidence Interval | ||
| Race | |||||||||
| Whites | 1.00 | Referent | 0.88 | 0.63, 1.22 | 0.69 | 0.49, 0.98 | 0.49 | 0.34, 0.71 | <0.01 |
| African Americans | 1.00 | Referent | 1.86 | 0.86, 4 | 1.42 | 0.66, 3.05 | 1.79 | 0.85, 3.77 | 0.49 |
| Pinteraction = 0.02 | |||||||||
| NSAID use in the past 5 years | |||||||||
| Yes | 1.00 | Referent | 0.96 | 0.68, 1.35 | 0.79 | 0.55, 1.13 | 0.65 | 0.45, 0.93 | 0.01 |
| No | 1.00 | Referent | 1 | 0.51, 1.96 | 0.66 | 0.31, 1.35 | 0.51 | 0.26, 1 | 0.05 |
| Pinteraction = 0.86 | |||||||||
| Obesity | |||||||||
| Body mass index ≥30 | 1.00 | Referent | 1.09 | 0.65, 1.86 | 0.72 | 0.42, 1.25 | 0.67 | 0.39, 1.17 | 0.17 |
| Body mass index <30 | 1.00 | Referent | 0.9 | 0.62, 1.3 | 0.77 | 0.53, 1.14 | 0.59 | 0.4, 0.87 | <0.01 |
| Pinteraction = 0.85 | |||||||||
Abbreviation: NSAID, nonsteroidal antiinflammatory drug.
Adjusted for age (40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–79 years), race, sex, daily energy intake (<1,554.6, 1,554.6–<1,991.0, 1,991.0–<2,595.4, ≥2,595.4 kcal/day), offset term, years of education (<12, 12–15, ≥16 years), endoscopic screening in the past 10 years (yes or no), red meat intake (<34.7, 34.7–<57.1, 57.1–<95.8, ≥95.8 g/day), vitamin E intake (<6.6, 6.6–<9.2, 9.2–<12.4, ≥12.4 mg/day), and body mass index 1 year prior (<25, 25–29.9, ≥30 kg/m2).
The relations with PUFAs did not vary by anatomic location of tumor (Table 4). Increased consumption of long-chain ω-3 PUFAs was inversely related to reduced risk of sigmoid, rectosigmoid, and rectal cancer in a dose-dependent manner (all Ptrend < 0.05). The positive association with the ratio of ω-6 to long-chain ω-3 PUFAs was also similarly observed for all subtypes of distal large bowel cancer (for the highest vs. the lowest quartile: multivariable ORs = 1.39 (95% CI: 0.95, 2.03) for sigmoid cancer, 1.75 (95% CI: 0.95, 3.22) for rectosigmoid cancer, and 1.34 (95% CI: 0.89, 2.02) for rectal cancer, respectively).
Table 4.
Tumor-specific Odds Ratios and 95% Confidence Intervals According to Quartiles of Polyunsaturated Fatty Acid Intake, North Carolina Colon Cancer Study II, 2001–2006
| Sigmoid |
Rectosigmoid |
Rectum |
P Valuea | ||||
| Odds Ratiob | 95% Confidence Interval | Odds Ratiob | 95% Confidence Interval | Odds Ratiob | 95% Confidence Interval | ||
| Total PUFAs, g/day | 0.95 | ||||||
| <13.1 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | |
| 13.1–<18.1 | 0.92 | 0.59, 1.43 | 1.25 | 0.63, 2.50 | 1.50 | 0.93, 2.44 | |
| 18.1–<24.4 | 1.05 | 0.62, 1.78 | 1.70 | 0.76, 3.81 | 1.06 | 0.59, 1.93 | |
| ≥24.4 | 1.02 | 0.54, 1.91 | 0.89 | 0.33, 2.35 | 1.04 | 0.52, 2.09 | |
| Ptrend | 0.86 | 0.82 | 0.82 | ||||
| Total ω-6 PUFAs, g/day | 0.81 | ||||||
| <11.5 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | |
| 11.5–<16.3 | 0.84 | 0.54, 1.30 | 1.22 | 0.62, 2.41 | 1.43 | 0.88, 2.31 | |
| 16.3–<22.0 | 1.01 | 0.60, 1.71 | 1.53 | 0.69, 3.43 | 0.97 | 0.53, 1.76 | |
| ≥22.0 | 0.82 | 0.44, 1.55 | 0.75 | 0.28, 1.98 | 1.01 | 0.51, 2.03 | |
| Ptrend | 0.70 | 0.56 | 0.79 | ||||
| Total ω-3 PUFAs, g/day | 0.64 | ||||||
| <1.27 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | |
| 1.27–<1.72 | 1.16 | 0.77, 1.74 | 0.83 | 0.42, 1.64 | 1.40 | 0.72, 1.81 | |
| 1.72–<2.31 | 0.90 | 0.56, 1.45 | 1.74 | 0.84, 3.59 | 1.25 | 0.74, 2.10 | |
| ≥2.31 | 0.83 | 0.48, 1.43 | 1.03 | 0.45, 2.37 | 1.13 | 0.62, 2.04 | |
| Ptrend | 0.37 | 0.72 | 0.68 | ||||
| Short-chain ω-3 PUFAs, g/day | 0.34 | ||||||
| <1.16 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | |
| 1.16–<1.57 | 0.91 | 0.60, 1.38 | 0.84 | 0.43, 1.64 | 1.22 | 0.76, 1.96 | |
| 1.57–<2.14 | 0.99 | 0.62, 1.58 | 1.39 | 0.68, 2.85 | 1.57 | 0.93, 2.64 | |
| ≥2.14 | 0.87 | 0.50, 1.49 | 0.91 | 0.40, 2.09 | 1.39 | 0.76, 2.52 | |
| Ptrend | 0.72 | 0.93 | 0.238 | ||||
| Long-chain ω-3 PUFAs, g/day | 0.84 | ||||||
| <0.06 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | |
| 0.06–<0.11 | 0.88 | 0.61, 1.26 | 0.81 | 0.48, 1.37 | 1.17 | 0.78, 1.74 | |
| 0.11–<0.18 | 0.68 | 0.46, 1.00 | 0.57 | 0.32, 1.02 | 0.97 | 0.64, 1.47 | |
| ≥0.18 | 0.61 | 0.41, 0.92 | 0.55 | 0.31, 1.00 | 0.67 | 0.43, 1.05 | |
| Ptrend | <0.01 | 0.03 | 0.05 | ||||
| ω-6:ω-3 PUFA ratio | 0.18 | ||||||
| <8.1 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | |
| 8.1–<9.2 | 1.16 | 0.80, 1.67 | 1.43 | 0.83, 2.46 | 1.00 | 0.68, 1.48 | |
| 9.2–<10.5 | 0.99 | 0.68, 1.45 | 0.96 | 0.54, 1.71 | 1.02 | 0.69, 1.51 | |
| ≥10.5 | 1.23 | 0.83, 1.83 | 1.35 | 0.75, 2.43 | 0.83 | 0.54, 1.28 | |
| Ptrend | 0.48 | 0.66 | 0.49 | ||||
| ω-6:long-chain ω-3 PUFA ratio | 0.72 | ||||||
| <90.3 | 1.00 | Referent | 1.00 | Referent | 1.00 | Referent | |
| 90.3–<151.7 | 1.09 | 0.74, 1.60 | 1.23 | 0.65, 2.35 | 0.95 | 0.62, 1.45 | |
| 151.7–<254.7 | 1.06 | 0.71, 1.56 | 2.01 | 1.10, 3.68 | 1.10 | 0.72, 1.68 | |
| ≥254.7 | 1.39 | 0.95, 2.03 | 1.75 | 0.95, 3.22 | 1.34 | 0.89, 2.02 | |
| Ptrend | 0.10 | 0.03 | 0.09 | ||||
Abbreviation: PUFA, polyunsaturated fatty acid.
Wald test P value for heterogeneity of the beta coefficient for highest quartile across tumor location.
Adjusted for age (40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–79 years), race, sex, daily energy intake (<1,554.6, 1,554.6–<1,991.0, 1,991.0–<2,595.4, ≥2,595.4 kcal/day), offset term, years of education (<12, 12–15, ≥16 years), endoscopic screening in the past 10 years (yes or no), red meat intake (<34.7, 34.7–<57.1, 57.1–<95.8, ≥95.8 g/day), vitamin E intake (<6.6, 6.6–<9.2, 9.2–<12.4, ≥12.4 mg/day), and body mass index 1 year prior (<25, 25–29.9, ≥30 kg/m2).
DISCUSSION
In this population-based case-control study, we assessed the relation between different PUFA intakes and the risk of distal large bowel cancer. We observed a reduced risk of distal large bowel cancer in relation to higher intake of long-chain ω-3 PUFAs and individual eicosapentaenoic acid and docosahexaenoic acid. We also observed a significantly increased risk of distal large bowel cancer with a high ratio of ω-6 to long-chain ω-3 PUFAs. The associations were limited to whites and not observed among African Americans. However, there was no evidence of effect measure modification by NSAID use, obesity, and tumor location.
Previous epidemiologic studies investigating the relation between ω-3 intake and colorectal cancer risk are limited and inconsistent. A pooled analysis of 3 prospective studies of the association between ω-3 fatty acid intake and colorectal cancer risk did not reach statistical significance (28), and only 1 of 9 prospective studies included in an earlier systematic review found a significant inverse relation (10). More recently, marine ω-3 PUFAs and intakes of total ω-3 PUFAs and individual eicosapentaenoic acid and docosahexaenoic acid were significantly associated with reduced risk of colorectal cancer in a prospective cohort of men (12) and in a Scottish case-control study (11). However, in other studies, the inverse association was limited to men (29), distal colon cancer (13), or nonadvanced colorectal cancer (15), or no association was found (14).
Contrary to the inconsistent epidemiologic findings, there is a wealth of experimental data demonstrating antiinflammatory and anticarcinogenic effects of ω-3 PUFAs (7). Diets high in ω-3 PUFAs have been shown to be effective in reducing tumor growth (1, 2) and metastases (4, 5) mainly through effects on the cyclooxygenase pathway (7). High intake of ω-3 PUFAs can replace arachidonic acid in membrane phospholipids (30) and compete for COX enzyme to form eicosapentaenoic acid-derived eicosanoids in favor of arachidonic acid-derived eicosanoids (7). Excess free arachidonic acid can perturb the homeostasis of normal eicosanoids (31), and arachidonic acid-derived eicosanoids such as prostaglandin E2 and leukotriene B4 are precursors for a number of key mediators of inflammation and cell proliferation (6, 30). Moreover, ω-3 PUFAs compete with ω-6 PUFAs for desaturation and elongation enzymes, decreasing the extent of conversion from linoleic acid to arachidonic acid (7). Emerging evidence also suggests COX-independent pathways of ω-3 PUFAs in tumor suppression (7, 32): Tumor formation was inhibited with ω-3 PUFA-enriched diet in athymic nude mice with and without cells injected with COX-1 or COX-2 (32).
α-Linolenic acid is the major ω-3 PUFA in Western diets; its consumption is about 10 times higher than that of eicosapentaenoic acid or docosahexaenoic acid (26). Consistent with previous findings (11, 33), our results did not show the inverse association between α-linolenic acid and distal large bowel cancer. It may not be surprising that the inverse association with colorectal cancer was limited to long-chain ω-3 PUFAs. Although α-linolenic acid can also compete with ω-6 PUFAs, its potency has been estimated to be approximately one fifth that of eicosapentaenoic acid or docosahexaenoic acid (7). Moreover, whereas α-linolenic acid can be converted to eicosapentaenoic acid or docosahexaenoic acid in humans, the extent of conversion is very inefficient: as low as 0.2% to eicosapentaenoic acid and 0.05% to docosahexaenoic acid (34). It has been shown that only eicosapentaenoic acid and docosahexaenoic acid, but not α-linolenic acid, reduced tumorigenesis in ApcMin/+ mice (35).
In our study, the ratio of ω-6 to long-chain ω-3 PUFAs was associated with an increased risk of distal large bowel cancer, but intake of ω-6 PUFAs per se was not related to cancer risk. Increased serum levels of total ω-6 PUFAs have been positively associated with the risk of having colorectal adenomas (36), but most previous studies did not find a positive association with high ω-6 PUFA intake (11, 14, 15, 37). Population-wide consumption of linoleic acid-rich foods is high; a slight variation in linoleic acid intake may not affect tissue arachidonic acid concentrations (7).
We found an unexpected positive association between long-chain ω-3 PUFA intake and distal large bowel cancer risk in African Americans. It has been documented that African Americans consume more fried foods than do whites (38). Thus, we explored whether racial differences in preferred preparation methods of fish may have influenced the association between long-chain ω-3 fatty acid intake and distal large bowel cancer risk. In fact, the African Americans in our study consumed fried fish more frequently than whites did. However, the inverse association remained unchanged after further adjustment for the relative proportion of fried fish out of total fish intake and intake of saturated fatty acids. Moreover, higher intakes of the 3 major food items for fish consumption, such as canned tuna, fish sticks/fried fish, and nonfried fish and seafood, were all inversely associated with distal large bowel cancer risk in whites. On the other hand, ω-3 PUFA intake assessed from the questionnaire may represent the bioavailability of specific fatty acids more poorly in African Americans compared with whites. In a small study conducted in the same catchment area as this study, correlations of food frequency questionnaire estimates of both eicosapentaenoic acid and docosahexaenoic acid with concentrations in erythrocytes and adipose tissue were substantially lower in African Americans than in Caucasians (39). Finally, such statistical heterogeneity of the association may indicate different carcinogenic processes in African Americans and whites. Although we did not find studies of African Americans to compare our findings, we previously observed inverse association with NSAIDs only among whites (40). Given that 2 potentially chemopreventive agents sharing COX-2 and inflammation pathways were inversely associated with colorectal cancer risk only among whites, it is possible that the inflammation pathways may be less important in colorectal carcinogenesis among African Americans. However, caution is required in such interpretation because the findings may have been due to chance.
Limitations of the current study include the retrospective nature of data collection. Although our food frequency questionnaire was designed to query habitual diet before the diagnosis of disease, prediagnosis symptoms may have altered the diet of cases, or diagnosis of cancer itself may have influenced past diet on recall (22). However, it is unlikely that either prediagnosis symptoms or diagnosis of cancer would lead cases to avoid fish consumption, which is the major source of long-chain ω-3 PUFAs (26).
Intake of polyunsaturated fatty acids in our study was estimated from a single measure of usual dietary intake derived from a food frequency questionnaire and, thus, measurement errors inherent in such dietary assessment apply to our study. Particularly, fat is often used in food preparation and hard to recognize and quantify from the food frequency questionnaire, and it may be underreported in overweight or obese individuals because of social implications (41). However, among the different types of fatty acids, ω-3 PUFAs appear to be least subject to these issues as shown by higher correlation between food frequency questionnaire estimates and biomarker as high as 0.8 for docosahexaenoic acid compared with 0.3 for monounsaturated fats (41).
Long-chain ω-3 PUFA supplements contain high amounts of eicosapentaenoic acid and docosahexaenoic acid (42). In our study, 21 (1.1%) and 118 (6.3%) participants reported taking cod liver oil or fish oil supplements more than once per week, respectively. Excluding the supplement users from the analyses did not change the association between long-chain ω-3 PUFA and distal large bowel cancer. Similarly, the results remained unchanged when the supplement users were added to the highest quartile of dietary intakes of long-chain ω-3 PUFA.
In this study, we examined 11 groupings of polyunsaturated fatty acids that are categorized on the basis of their biochemical and nutritional characteristics. This may raise concern over multiple comparisons. However, we did not adjust for multiple testing, because Bonferroni correction and other traditional statistical procedures that assume independence of hypotheses do not protect against type 1 error associated with multiple testing of such highly correlated hypotheses (43). Instead, we interpreted individual associations in the context of epidemiologic and experimental studies as well as our own data, as previously recommended (44, 45). Residual confounding remains a potential limitation to our study. Intake of long-chain ω-3 PUFAs may represent a generally healthy diet and lifestyle. In fact, endoscopic screening, obesity, and intakes of meat and vitamin E were important confounders in our analyses. However, adjustment for these variables still did not fully diminish the association with long-chain ω-3 PUFA intake. The possibility of selection bias should also be considered. The participation rate among controls in our study was 56%, which does not appear to be unusual in contemporary population-based case-control studies (46). Participating individuals may represent groups of higher socioeconomic status and healthier lifestyles than the target population (47). However, in our study, the association between long-chain ω-3 PUFA and distal large bowel cancer was relatively strong, with an estimated odds ratio of 0.61 for the comparison of the highest versus the lowest quartiles, and therefore, it is unlikely that this finding is accounted for entirely by selection bias.
In conclusion, the results of this study support the hypothesis that increasing consumption of long-chain ω-3 PUFAs may be beneficial in preventing distal large bowel cancer. Whether the possible benefit from this dietary modification varies by race warrants further evaluation.
Acknowledgments
Author affiliations: Epidemiology Branch, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, North Carolina (Sangmi Kim, Dale P. Sandler); and Center for Gastrointestinal Biology and Disease, University of North Carolina, Chapel Hill, North Carolina (Joseph Galanko, Christopher Martin, Robert S. Sandler).
This research was supported, in part, by grants from the National Institutes of Health (P30 DK34987 and R01 CA66635) and by the Intramural Program of the National Institutes of Health, National Institute of Environmental Health Sciences (Z01 ES04400509).
The authors thank Dr. Aimee D'Aloisio and Dr. Honglei Chen for their valuable comments on the earlier version of this manuscript.
This work was presented at the American Association for Cancer Research's Frontiers in Cancer Prevention Research Conference, Houston, Texas, December 6–9, 2009.
Conflict of interest: none declared.
Glossary
Abbreviations
- CI
confidence interval
- COX
cyclooxygenase
- NSAID
nonsteroidal antiinflammatory drug
- OR
odds ratio
- PUFA
polyunsaturated fatty acid
References
- 1.Singh J, Hamid R, Reddy BS. Dietary fat and colon cancer: modulating effect of types and amount of dietary fat on ras-p21 function during promotion and progression stages of colon cancer. Cancer Res. 1997;57(2):253–258. [PubMed] [Google Scholar]
- 2.Latham P, Lund EK, Johnson IT. Dietary n-3 PUFA increases the apoptotic response to 1,2-dimethylhydrazine, reduces mitosis and suppresses the induction of carcinogenesis in the rat colon. Carcinogenesis. 1999;20(4):645–650. doi: 10.1093/carcin/20.4.645. [DOI] [PubMed] [Google Scholar]
- 3.Calviello G, Di Nicuolo F, Gragnoli S, et al. n-3 PUFAs reduce VEGF expression in human colon cancer cells modulating the COX-2/PGE2 induced ERK-1 and -2 and HIF-1-α induction pathway. Carcinogenesis. 2004;25(12):2303–2310. doi: 10.1093/carcin/bgh265. [DOI] [PubMed] [Google Scholar]
- 4.Iigo M, Nakagawa T, Ishikawa C, et al. Inhibitory effects of docosahexaenoic acid on colon carcinoma 26 metastasis to the lung. Br J Cancer. 1997;75(5):650–655. doi: 10.1038/bjc.1997.116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kontogiannea M, Gupta A, Ntanios F, et al. Omega-3 fatty acids decrease endothelial adhesion of human colorectal carcinoma cells. J Surg Res. 2000;92(2):201–205. doi: 10.1006/jsre.2000.5930. [DOI] [PubMed] [Google Scholar]
- 6.James MJ, Gibson RA, Cleland LG. Dietary polyunsaturated fatty acids and inflammatory mediator production. Am J Clin Nutr. 2000;71(1 suppl):343S–348S. doi: 10.1093/ajcn/71.1.343s. [DOI] [PubMed] [Google Scholar]
- 7.Larsson SC, Kumlin M, Ingelman-Sundberg M, et al. Dietary long-chain n-3 fatty acids for the prevention of cancer: a review of potential mechanisms. Am J Clin Nutr. 2004;79(6):935–945. doi: 10.1093/ajcn/79.6.935. [DOI] [PubMed] [Google Scholar]
- 8.Kremer JM. n-3 Fatty acid supplements in rheumatoid arthritis. Am J Clin Nutr. 2000;71(suppl):349S–351S. doi: 10.1093/ajcn/71.1.349s. [DOI] [PubMed] [Google Scholar]
- 9.Anti M, Marra G, Armelao F, et al. Effect of omega-3 fatty acids on rectal mucosal cell proliferation in subjects at risk for colon cancer. Gastroenterology. 1992;103(3):883–891. doi: 10.1016/0016-5085(92)90021-p. [DOI] [PubMed] [Google Scholar]
- 10.MacLean CH, Newberry SJ, Mojica WA, et al. Effects of omega-3 fatty acids on cancer risk: a systematic review. JAMA. 2006;295(4):403–415. doi: 10.1001/jama.295.4.403. [DOI] [PubMed] [Google Scholar]
- 11.Theodoratou E, McNeill G, Cetnarskyj R, et al. Dietary fatty acids and colorectal cancer: a case-control study. Am J Epidemiol. 2007;166(2):181–195. doi: 10.1093/aje/kwm063. [DOI] [PubMed] [Google Scholar]
- 12.Hall MN, Chavarro JE, Lee IM, et al. A 22-year prospective study of fish, n-3 fatty acid intake, and colorectal cancer risk in men. Cancer Epidemiol Biomarkers Prev. 2008;17(5):1136–1143. doi: 10.1158/1055-9965.EPI-07-2803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Kimura Y, Kono S, Toyomura K, et al. Meat, fish and fat intake in relation to subsite-specific risk of colorectal cancer: the Fukuoka Colorectal Cancer Study. Cancer Sci. 2007;98(4):590–597. doi: 10.1111/j.1349-7006.2007.00425.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Daniel CR, McCullough ML, Patel RC, et al. Dietary intake of omega-6 and omega-3 fatty acids and risk of colorectal cancer in a prospective cohort of U.S. men and women. Cancer Epidemiol Biomarkers Prev. 2009;18(2):516–525. doi: 10.1158/1055-9965.EPI-08-0750. [DOI] [PubMed] [Google Scholar]
- 15.Butler LM, Wang R, Koh WP, et al. Marine n-3 and saturated fatty acids in relation to risk of colorectal cancer in Singapore Chinese: a prospective study. Int J Cancer. 2009;124(3):678–686. doi: 10.1002/ijc.23950. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Nelson RL, Dollear T, Freels S, et al. The relation of age, race, and gender to the subsite location of colorectal carcinoma. Cancer. 1997;80(2):193–197. doi: 10.1002/(sici)1097-0142(19970715)80:2<193::aid-cncr4>3.0.co;2-v. [DOI] [PubMed] [Google Scholar]
- 17.Gervaz P, Bucher P, Morel P. Two colons-two cancers: paradigm shift and clinical implications. J Surg Oncol. 2004;88(4):261–266. doi: 10.1002/jso.20156. [DOI] [PubMed] [Google Scholar]
- 18.Zha S, Yegnasubramanian V, Nelson WG, et al. Cyclooxygenases in cancer: progress and perspective. Cancer Lett. 2004;215(1):1–20. doi: 10.1016/j.canlet.2004.06.014. [DOI] [PubMed] [Google Scholar]
- 19.Trayhurn P, Wood IS. Adipokines: inflammation and the pleiotropic role of white adipose tissue. Br J Nutr. 2004;92(3):347–355. doi: 10.1079/bjn20041213. [DOI] [PubMed] [Google Scholar]
- 20.Aldrich TE, Vann D, Moorman PG, et al. 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;35(1):61–64. doi: 10.1007/BF00694746. [DOI] [PubMed] [Google Scholar]
- 21.Weinberg CR, Sandler DP. Randomized recruitment in case-control studies. Am J Epidemiol. 1991;134(4):421–432. doi: 10.1093/oxfordjournals.aje.a116104. [DOI] [PubMed] [Google Scholar]
- 22.Willett W. Nutritional Epidemiology. New York, NY: Oxford University Press; 1998. [Google Scholar]
- 23.Ainsworth BE, Haskell WL, Leon AS, et al. Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc. 1993;25(1):71–80. doi: 10.1249/00005768-199301000-00011. [DOI] [PubMed] [Google Scholar]
- 24.Diet History Questionnaire. Bethesda, MD: National Cancer Institute, National Institutes of Health; 2009. ( http://riskfactor.cancer.gov/DHQ) [Google Scholar]
- 25.Subar AF, Midthune D, Kulldorff M, et al. Evaluation of alternative approaches to assign nutrient values to food groups in food frequency questionnaires. Am J Epidemiol. 2000;152(3):279–286. doi: 10.1093/aje/152.3.279. [DOI] [PubMed] [Google Scholar]
- 26.Meyer BJ, Mann NJ, Lewis JL, et al. Dietary intakes and food sources of omega-6 and omega-3 polyunsaturated fatty acids. Lipids. 2003;38(4):391–398. doi: 10.1007/s11745-003-1074-0. [DOI] [PubMed] [Google Scholar]
- 27.Chen TC, Chimeh F, Lu Z, et al. Factors that influence the cutaneous synthesis and dietary sources of vitamin D. Arch Biochem Biophys. 2007;460(2):213–217. doi: 10.1016/j.abb.2006.12.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Geelen A, Schouten JM, Kamphuis C, et al. Fish consumption, n-3 fatty acids, and colorectal cancer: a meta-analysis of prospective cohort studies. Am J Epidemiol. 2007;166(10):1116–1125. doi: 10.1093/aje/kwm197. [DOI] [PubMed] [Google Scholar]
- 29.Kojima M, Wakai K, Tokudome S, et al. Serum levels of polyunsaturated fatty acids and risk of colorectal cancer: a prospective study. Am J Epidemiol. 2005;161(5):462–471. doi: 10.1093/aje/kwi066. [DOI] [PubMed] [Google Scholar]
- 30.Bagga D, Wang L, Farias-Eisner R, et al. Differential effects of prostaglandin derived from omega-6 and omega-3 polyunsaturated fatty acids on COX-2 expression and IL-6 secretion. Proc Natl Acad Sci U S A. 2003;100(4):1751–1756. doi: 10.1073/pnas.0334211100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Kinsella JE, Lokesh B, Stone RA. Dietary n-3 polyunsaturated fatty acids and amelioration of cardiovascular disease: possible mechanisms. Am J Clin Nutr. 1990;52(1):1–28. doi: 10.1093/ajcn/52.1.1. [DOI] [PubMed] [Google Scholar]
- 32.Boudreau MD, Sohn KH, Rhee SH, et al. Suppression of tumor cell growth both in nude mice and in culture by n-3 polyunsaturated fatty acids: mediation through cyclooxygenase-independent pathways. Cancer Res. 2001;61(4):1386–1391. [PubMed] [Google Scholar]
- 33.Kuriki K, Wakai K, Hirose K, et al. Risk of colorectal cancer is linked to erythrocyte compositions of fatty acids as biomarkers for dietary intakes of fish, fat, and fatty acids. Cancer Epidemiol Biomarkers Prev. 2006;15(10):1791–1798. doi: 10.1158/1055-9965.EPI-06-0180. [DOI] [PubMed] [Google Scholar]
- 34.Berquin IM, Edwards IJ, Chen YQ. Multi-targeted therapy of cancer by omega-3 fatty acids. Cancer Lett. 2008;269(2):363–377. doi: 10.1016/j.canlet.2008.03.044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Petrik MB, McEntee MF, Johnson BT, et al. Highly unsaturated (n-3) fatty acids, but not α-linolenic, conjugated linoleic or γ-linolenic acids, reduce tumorigenesis in ApcMin/+ mice. J Nutr. 2000;130(10):2434–2443. doi: 10.1093/jn/130.10.2434. [DOI] [PubMed] [Google Scholar]
- 36.Pot GK, Geelen A, van Heijningen EM, et al. Opposing associations of serum n-3 and n-6 polyunsaturated fatty acids with colorectal adenoma risk: an endoscopy-based case-control study. Int J Cancer. 2008;123(8):1974–1977. doi: 10.1002/ijc.23729. [DOI] [PubMed] [Google Scholar]
- 37.Hall MN, Campos H, Li H, et al. Blood levels of long-chain polyunsaturated fatty acids, aspirin, and the risk of colorectal cancer. Cancer Epidemiol Biomarkers Prev. 2007;16(2):314–321. doi: 10.1158/1055-9965.EPI-06-0346. [DOI] [PubMed] [Google Scholar]
- 38.Patterson BH, Harlan LC, Block G, et al. Food choices of whites, blacks, and Hispanics: data from the 1987 National Health Interview Survey. Nutr Cancer. 1995;23(2):105–119. doi: 10.1080/01635589509514367. [DOI] [PubMed] [Google Scholar]
- 39.Godley PA, Campbell MK, Miller C, et al. Correlation between biomarkers of omega-3 fatty acid consumption and questionnaire data in African American and Caucasian United States males with and without prostatic carcinoma. Cancer Epidemiol Biomarkers Prev. 1996;5(2):115–119. [PubMed] [Google Scholar]
- 40.Kim S, Martin C, Galanko J, et al. Use of nonsteroidal antiinflammatory drugs and distal large bowel cancer in whites and African Americans. Am J Epidemiol. 2008;168(11):1292–1300. doi: 10.1093/aje/kwn255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Arab L. Biomarkers of fat and fatty acid intake. J Nutr. 2003;133(suppl 3):925S–932S. doi: 10.1093/jn/133.3.925S. [DOI] [PubMed] [Google Scholar]
- 42.Kris-Etherton PM, Taylor DS, Yu-Poth S, et al. Polyunsaturated fatty acids in the food chain in the United States. Am J Clin Nutr. 2000;71(1 suppl):179S–188S. doi: 10.1093/ajcn/71.1.179S. [DOI] [PubMed] [Google Scholar]
- 43.Bender R, Lange S. Adjusting for multiple testing—when and how? J Clin Epidemiol. 2001;54(4):343–349. doi: 10.1016/s0895-4356(00)00314-0. [DOI] [PubMed] [Google Scholar]
- 44.Savitz DA, Olshan AF. Multiple comparisons and related issues in the interpretation of epidemiologic data. Am J Epidemiol. 1995;142(9):904–908. doi: 10.1093/oxfordjournals.aje.a117737. [DOI] [PubMed] [Google Scholar]
- 45.Rothman KJ. No adjustments are needed for multiple comparisons. Epidemiology. 1990;1(1):43–46. [PubMed] [Google Scholar]
- 46.Morton LM, Cahill J, Hartge P. Reporting participation in epidemiologic studies: a survey of practice. Am J Epidemiol. 2006;163(3):197–203. doi: 10.1093/aje/kwj036. [DOI] [PubMed] [Google Scholar]
- 47.Goldberg M, Chastang JF, Leclerc A, et al. Socioeconomic, demographic, occupational, and health factors associated with participation in a long-term epidemiologic survey: a prospective study of the French GAZEL cohort and its target population. Am J Epidemiol. 2001;154(4):373–384. doi: 10.1093/aje/154.4.373. [DOI] [PubMed] [Google Scholar]