Fax: (803) 434-4795
Common polymorphisms in 5-lipoxygenase and 12-lipoxygenase genes and the risk of incident, sporadic colorectal adenoma
Article first published online: 18 JAN 2007
Copyright © 2007 American Cancer Society
Volume 109, Issue 5, pages 849–857, 1 March 2007
How to Cite
Gong, Z., Hebert, J. R., Bostick, R. M., Deng, Z., Hurley, T. G., Dixon, D. A., Nitcheva, D. and Xie, D. (2007), Common polymorphisms in 5-lipoxygenase and 12-lipoxygenase genes and the risk of incident, sporadic colorectal adenoma. Cancer, 109: 849–857. doi: 10.1002/cncr.22469
- Issue published online: 22 FEB 2007
- Article first published online: 18 JAN 2007
- Manuscript Accepted: 29 NOV 2006
- Manuscript Revised: 27 NOV 2006
- Manuscript Received: 18 SEP 2006
- Public Health Service. Grant Numbers: R01 CA66539, R01 CA-51932
- National Cancer Institute
- National Center for Research Resources. Grant Number: RR017698
- National Institutes of Health
- Department of Health and Human Services
- Georgia Cancer Coalition Distinguished Cancer Scholar
- genetic polymorphisms;
- colorectal adenoma;
- lipoxygenase pathway;
- case-control study
Lipoxygenases (LOX) are major enzymes that metabolize arachidonic acid to hydroxyl-eicosatetraenoic acids and leukotrienes, which have been implicated in inflammation and colorectal cancer risk. Polymorphisms in LOX genes may influence their function and/or expression and, thus, may modify the risk for colorectal adenoma. The authors investigated the associations of 3 polymorphisms (2 in 5-LOX, −1708 guanineadenine and 21 cytosinethymine; and 1 in 12-LOX, arginine 261 glutamine [Arg261Gln]) in LOX genes with the risk of colorectal adenoma and also explored possible interactions of these polymorphisms with several inflammation-pathway or arachidonic acid metabolism-pathway related factors with the risk of colorectal adenoma.
By using data from a community-based, case-control study of incident, sporadic colorectal adenoma that included 162 cases and 211 controls, the authors constructed multiple logistic regression models to estimate the odds ratios (OR) and 95% confidence intervals (95% CI) of colorectal adenoma after adjusting for potential confounders.
Overall, there were no significant associations of the 2 5-LOX polymorphisms with the risk of colorectal adenoma. However, there was an inverse association between the Arg261Gln polymorphism in 12-LOX and colorectal adenoma (OR, 0.63; 95% CI, 0.40–1.00). A significant interaction also was observed between the 12-LOX polymorphism (Arg261Gln) and the use of nonsteroidal anti-inflammatory drugs (Pinteraction = .02).
The current results suggested that polymorphisms of LOX genes may act independently or with other factors to affect the risk of colorectal adenoma. Further studies will be needed to confirm these findings. Cancer 2007 © 2007 American Cancer Society.
In the United States, colorectal cancer is the third most common cancer and the second leading cause of cancer death.1 Epidemiological and experimental evidence indicates that both genetic and environmental factors are important determinants of colorectal cancer risk.2, 3
Lipid metabolism, especially in relation to arachidonic acid (AA)-metabolizing pathways, plays critical roles in chronic inflammation and colorectal carcinogenesis.4, 5 The lipoxygenase (LOX) and cyclooxygenase (COX) pathways are 2 major AA-metabolizing pathways.6 Various AA metabolites derived from both LOX and COX pathways are modulators of inflammation and have been shown to affect carcinogenesis in both in vitro and in vivo in experimental models.7–9 Considerable evidence has been accumulated suggesting that the inducible form of COX, COX-2, plays an important role in tumor development.10 In contrast, the role of the LOX pathway in colorectal cancer is less clear. The LOXs metabolize AA to biologically active metabolites, including hydroperoxy-eicosatetraenoic acids (HPETEs) and hydroxyl-eicosatetraenoic acids (HETEs), as well as leukotrienes (LTs).11
5-LOX is the key enzyme in the biosynthesis of LTs, which are known proinflammatory mediators.12 It has been demonstrated in several studies that LTs inhibit apoptosis, stimulate colonic cell proliferation, and are procarcinogenic.13–15 In addition, inhibitors of LT biosynthesis had therapeutic potential in a variety of diseases with known links to inflammation, including some cancers.16–1812-LOX can catalyze the stereo-specific oxygenation of AA to form 12-HPETE and 12-HETE,19 and it is overexpressed in a variety of tumors including breast, colorectal, and prostate.20–22 Others found that 12-HETE enhanced carcinogenesis through up-regulating tumor cell-adhesion molecules23; stimulating cell proliferation, angiogenesis, and tumor spread22; and inhibiting apoptosis.2412-LOX inhibitors can block cell cycle progression and induce apoptosis in cancer cells.25 Genetic variation in the 5-LOX and 12-LOX genes may alter enzyme expression levels or biochemical function and, consequently, may have an impact on the synthesis of bioactive lipid products. Two genetic polymorphisms of 5-LOX, including a mutation (−1708 guanine to adenine [GA]) in the negative regulatory region and a mutation (21 cytosine to thymine [CT]) within exon 1, have been identified.26 Recently, a polymorphism (G to A change) in the 12-LOX gene, consisting of an arginine (Arg) to a glutamine (Gln) substitution at amino acid 261 (Arg261Gln), was discovered. This polymorphism reportedly is associated with bipolar disorder.27
Among environmental factors, diet appears to play an important role in the risk for colorectal cancer. Generally, a diet high in red meat and fat is associated with increased risk, whereas high calcium, vegetable, fruit, fiber, and folate intakes have been associated with decreased risk of colorectal cancer.3 Decreased risk of colorectal cancer also has been observed among individuals who take aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) regularly.28 A physically active lifestyle and maintenance of normal body weight also may be important in the primary prevention of colorectal cancer.29, 30 The mechanisms by which fat intake and obesity may increase risk and by which NSAID use and physical activity may decrease risk for colorectal neoplasms may involve modulation of inflammation and AA pathways. Thus, polymorphisms in LOX genes may interact with these factors to affect the risk of colorectal adenoma.
The objectives of the current study were to examine associations of the 3 LOX gene polymorphisms with risk of colorectal adenoma and to explore possible interactions of the polymorphisms with several inflammation pathway- or AA pathway-related risk factors in relation to risk of adenoma. Finally, we investigated whether associations differed by adenoma characteristics.
MATERIALS AND METHODS
Study Design and Study Population
The design and population characteristics of this community- and colonoscopy-based case-control study have been described previously.31 Briefly, participants were recruited through community gastroenterology practices in Winston-Salem and Charlotte, North Carolina. All patients who were scheduled for a colonoscopy between 1994 and 1997 were screened for specific study eligibility criteria and then recruited before colonoscopy. Eligibility criteria included age from 30 years to 74 years, no previous adenoma, no individual history of cancer (except nonmelanoma skin cancer), no known genetic syndromes associated with predisposition to colonic neoplasia, no history of ulcerative colitis or Crohn disease, and resident of either of the 2 North Carolina metropolitan areas. The study was approved by the Institutional Review Board of Wake Forest University. Informed consent was obtained from each participant prior to data collection.
Cases were identified as eligible colonoscopy patients who had study pathologist-confirmed, incident adenomatous polyps, and controls were identified as patients without adenomatous polyps. Polyps were removed at colonoscopy and were examined by a study pathologist using criteria from the National Polyp Study.32 Data on polyp location, number, size, shape, histologic type, and degree of dysplasia were collected.
Among all 4 clinical sites, initially, 669 participants were identified as eligible for the study. Of these, 617 participants were contacted, and 417 participants (62.3% of all eligible participants) signed consent forms and participated in the study. Seventeen of those 417 participants subsequently were determined ineligible for the study; thus, 400 potential participants were available for genotypic analysis. There was sufficient DNA from 373 of those 400 participants to enable genotyping.
Data were collected by self-administered questionnaires. Study participants were sent questionnaires before their colonoscopy visit. Questionnaires were then collected and scanned for completeness or errors by research staff at the colonoscopy visit. For each participant, information was collected on basic demographics (age, sex, education, marital status, and race), medical history, anthropometry (height, weight, and waist and hip circumferences), family history of polyps or colorectal cancer, reproductive history, and reasons for colonoscopy. Dietary information was assessed by using an adaptation of the Willett semiquantitative food frequency questionnaire (FFQ) (153 items), which was expanded to include additional vegetables, fruits, and low-fat foods.33 Physical activity information was obtained by using a modified Paffenbarger questionnaire.34 Information also was collected on smoking history, alcohol use, and regular use of aspirin and other NSAIDs. An individual was defined as a regular aspirin or NSAID user if he or she consumed these drugs at least once per week.
A blood sample was obtained from each case and control and stored at −80°C for genetic analysis. The genomic DNA was extracted from stored white blood cells. The genomic DNA pellets (50–100 μg) were dissolved in 300 μL to 800 μL of Tris-hydrochloride/ ethylenediamine tetraacetic acid buffer, of which approximately 1 μL was used for each polymerase chain reaction (PCR) reaction.
The 3 polymorphisms were detected by using the PCR-restriction fragment length polymorphism (RFLP) method. The PCR reactions were performed on a Perkin Elmer GeneAmp System 9700 according to the manufacturer's protocol. Two quality-control procedures were employed. First, laboratory personnel were blinded to which samples were from cases or controls. Second, 10% of (randomly selected) samples were repeated in each experiment to confirm concordance.
5-LOX (−1708 GA)
A 229-base-pair fragment of the 5-LOX gene was amplified using the following 2 primers: 5′-GCA CTG TAT AGC ATG TAC ATT A-3′ (forward) and 5′-CGT GAC CCA TTT TGA GTT AG-3′ (reverse). PCR was performed in a reaction mixture of 25 μL that contained standard PCR buffer, 1.0 mM MgCl2, 0.2 mM dinucleotide triphosphate (dNTP), 1 unit Taq polymerase (Gibco-Invitrogen), and 0.4 μM of each primer. The PCR conditions consisted of an initial denaturation at 94°C for 2 minutes; 35 cycles at 94°C for 30 seconds, 60°C for 30 seconds, and 72°C for 40 seconds; and a final extension at 72°C for 7 minutes. After PCR amplification, PCR products were digested with the DdeI restriction enzyme overnight and then separated by gel electrophoresis.
5-LOX (21 CT)
A 150-base-pair fragment of the 5-LOX gene was amplified using the following 2 primers: 5′-CGC CAT GCC CTC CTA CAC-3′ (forward) and 5′-CCA CGC TCG AAG TCG TTG TA-3′ (reverse). PCR was performed in a reaction mixture of 25 μL that contained standard PCR buffer, 5% dimethyl sulfoxide (DMSO), 1.0 mM MgCl2, 0.2 mM dNTP, 1 unit Taq polymerase (Gibco-Invitrogen), and 0.4 μM of each oligonucleotide primer. The reactions were heated to 94°C for 2 minutes followed by 35 cycles at 94°C for 30 seconds, 70°C for 30 seconds, and 72°C for 45 seconds. At the end, the reactions were extended for 7 minutes at 72°C. PCR products were digested with the Btg I restriction enzyme overnight and separated by gel electrophoresis.
A 187-base-pair fragment of the 12-LOX gene was amplified using the following primers: 5′-CCA GTG GGG CAG GAT GAT GAG TTG-3′ (forward) and 5′-GGA GAG AAC AGT GCC AGG CAG-3′ (reverse). PCR was performed in a reaction mixture of 25 μL that contained standard PCR buffer, 5% DMSO, 1.0 mM MgCl2, 0.2 mM dNTP, 1 unit Taq polymerase (Gibco-Invitrogen), and 0.4 μM of each oligonucleotide primer. The reactions were heated to 94°C for 2 minutes followed by 35 cycles at 94°C for 30 seconds, 62°C for 30 seconds, and 72°C for 40 seconds. Then, the reactions were extended for 7 minutes at 72°C. Finally, PCR products were digested with the Ban II restriction enzyme overnight and separated by gel electrophoresis.
All statistical analyses were conducted using SAS version 9.1.35 Baseline characteristics of cases and controls were calculated and compared by using analyses of covariance for continuous variables and chi-square tests for categorical variables. In the control group, each polymorphism was tested to ensure that it fit Hardy-Weinberg equilibrium based on chi-square tests.
For the main effect model, multiple logistic regression models were used to test hypotheses for an association between genotypes and colorectal adenoma while controlling for potential confounders. The odds ratio (OR) and 95% confidence interval (CI) for the association between each genotype and disease were calculated by using the more common homozygous allele among controls as the reference group. Several risk factors were examined as possible confounders of the genotype and colorectal adenoma associations. Among these were age, sex, race, body mass index (BMI) (weight [kg]/height [m2]), waist-to-hip ratio (WHR), family history of colorectal cancer in first-degree relatives (FHCC), smoking, alcohol consumption, NSAID use, total physical activity, total dietary intake of fat, energy intake, and total dietary intakes of fish, red meat, fruit, and vegetables. Criteria for inclusion of a covariate in the final model included 1) biologic plausibility, 2) whether it fit the model at P ≤ .10, and 3) whether it altered the OR for the primary exposure variable, the genotype, by ≥10%. Age, sex, FHCC, smoking status, total physical activity, and current NSAID use were included in the final model.
Statistical tests for interactions were performed by fitting an interaction term for each genotype-effect modifier combination in the final model. The multiplicative joint effects of genotypes and exposures were calculated by including 3-level dummy variables in the models.36 The dummy variables were constructed by using the most common homozygous genotype and the lowest level of exposure as the referent. In these analyses, specific nutrient intakes were dichotomized into low or high based on the sex-specific medians from the distributions of intakes among controls. For analyses that examined whether the effects of these polymorphisms varied by different adenomas characteristics (eg, size <1 cm or ≥1 cm), polytomous logistic regression was used to model ORs for each group of cases versus controls.37
Various demographic- and health-related characteristics of cases and controls are shown in Table 1. On average, cases were older, more likely to be men, and more likely to be smokers or drinkers. Controls were more likely to have a first-degree relative with colorectal cancer. Other factors were balanced fairly well between cases and controls.
|Variable||Cases (n = 162)||Controls (n = 211)||P†|
|Demographic and major risk factors|
|Age ± SD, y||58.5 ± 8.6||56 ± 9.9||.02|
|White race, %||87.7||90||.48|
|College education, %||18.5||24.2||.31|
|Family history of colon cancer, %‡||19.9||36.5||.0005|
|Currently smoke cigarettes, %||30.3||19||.01|
|Currently drink alcohol, %||54.4||43.1||.02|
|Aspirin or NSAID use, %||47.5||54||.21|
|Total physical activity, METs/d||50 ± 15.7||51.3 ± 13.8||.25|
|Dietary intake± SD|
|Total energy, kcal/d||2023 ± (760)||1968 ± (720)||.88|
|Total fat, g/d||71 ± (33)||65 ± (30)||.49|
|Total polyunsaturated fat, g/d||14 ± (5.9)||13 ± (6.6)||.77|
|Total red meat, servings/wk||4.6 ± (4.4)||4.0 ± (3.3)||.16|
|Total fish intake, servings/wk||2.2 ± (1.5)||1.9 ± (1.7)||.12|
|Total fruits and vegetables, servings/d||6.4 ± (3.5)||6.4 ± (3.8)||.95|
|Body measurement± SD|
|Waist-to-hip ratio||0.94 ± 0.12||0.88 ± 0.10||.14|
|Body mass index, kg/m2||27 ± 5.9||27 ± 5.3||.87|
Adjusted associations between genotypes and colorectal adenomas are shown in Table 2. All polymorphisms were in Hardy-Weinberg equilibrium (all P > .05) in cases and controls. Overall, there was no significant association between either of the 2 5-LOX polymorphisms and colorectal adenoma. However, the Arg261Gln polymorphism in 12-LOX was associated with a decreased risk of colorectal adenoma (Ptrend = .02). Individuals with ≥1 Gln variant were at 37% lower risk (OR, 0.63) of colorectal adenoma compared with individuals who had the Arg/Arg genotype (wild type) after controlling for covariates. Associations between the 3 polymorphisms and the risk of colorectal adenoma according to adenoma characteristics also were examined. There were no significant differences for the effect of genotypes on different adenoma subgroups (data not shown).
|Genotype||Cases||Controls||OR (95% CI)|
|5-LOX −1708 GA|
|GA and AA||42||26.1||57||27.4||0.96 (0.59–1.55)*|
|GA and AA||0.99 (0.59–1.64)†|
|5-LOX 21 CT|
|CT and TT||52||32.3||55||26.1||1.31 (0.82–2.09)*|
|CT and TT||1.35 (0.83–2.21)†|
|Arg/Gln and Gln/Gln||88||54.3||140||66.3||0.62 (0.40–0.96)*|
|Arg/Gln and Gln/Gln||0.63 (0.40–1.00)†|
Adjusted associations of the 5-LOX genotypes (−1708 GA and 21 CT) with the risk of colorectal adenomas according to inflammation pathway- and AA pathway-related factors are shown in Tables 3 and 4, respectively. There were no statistically significant interactions of genotypes with these factors (all Pinteraction ≥ .22). However, the homozygous wild genotypes for both polymorphisms were associated with a reduced risk of colorectal adenoma among individuals who took NSAIDs regularly or who were more active physically, but no such associations were observed among individuals who had the variant allele. In addition, having ≥1 variant allele (CT or TT) in the 5-LOX (21 CT) polymorphism was associated with an increased risk of colorectal adenoma among individuals who also had a large WHR, but no such association was observed among individuals who had the CC genotype.
|GG||GA and AA|
|No. of cases/ controls||OR (95% CI)||No. of cases/controls||OR (95% CI)|
|Yes||22/46||0.51 (0.27–0.96)||6/14||0.80 (0.28–2.27)|
|Total fat intake|
|High||53/78||0.77 (0.45–1.30)||22/29||1.06 (0.52–2.14)|
|Total polyunsaturated fat|
|High||63/75||1.10 (0.65–1.86)||21/28||1.25 (0.61–2.57)|
|Total physical activity|
|High||48/84||0.49 (0.29–0.83)||22/27||0.77 (0.38–1.55)|
|High||69/72||1.43 (0.84–2.43)||26/27||1.69 (0.84–3.42)|
|Body mass index|
|High||53/76||0.82 (0.48–1.40)||22/27||1.06 (0.52–2.15)|
|CC||CT and TT|
|No. of cases/ controls||OR (95% CI)||No. of cases/ controls||OR (95% CI)|
|Yes||17/50||0.42 (0.22–0.82)||10/11||1.14 (0.43–3.01)|
|Total fat intake|
|High||50/76||0.86 (0.51–1.47)||26/31||1.27 (0.64–2.50)|
|Total polyunsaturated fat|
|High||57/71||1.22 (0.71–2.09)||28/32||1.52 (0.78–2.97)|
|Total physical activity|
|High||48/84||0.55 (0.32–0.94)||21/29||0.82 (0.40–1.67)|
|High||63/77||1.39 (0.82–2.38)||32/24||2.17 (1.08–4.35)|
|Body mass index|
|High||52/79||0.90 (0.52–1.54)||24/26||1.26 (0.62–2.56)|
Adjusted associations of 12-LOX (Arg261Gln) genotypes and colorectal adenoma according to inflammation pathway- and AA pathway-related factors are shown in Table 5. There was a significant interaction between 12-LOX (Arg261Gln) genotypes and NSAID use (Pinteraction = .02). The combination of a Gln variant allele and regularly taking NSAIDs was associated with a 71% reduction in the risk of colorectal adenoma (OR, 0.29). The association between WHR and adenoma tended to be stronger among individuals who had the Arg/Arg genotype than among individuals who had a Gln variant allele.
|Arg/Arg||Arg/Gln and Gln/Gln|
|No. of cases/ controls||OR (95% CI)||No. of cases/ controls||OR (95% CI)|
|Yes||15/11||1.38 (0.56–3.41)||13/50||0.29 (0.13–0.62)|
|Total fat intake|
|High||33/39||0.68 (0.33–1.38)||43/68||0.55 (0.29–1.07)|
|Total polyunsaturated fat|
|High||38/40||0.96 (0.47–1.95)||47/63||0.72 (0.37–1.40)|
|Total physical activity|
|High||33/43||0.47 (0.23–0.94)||37/70||0.37 (0.19–0.71)|
|High||49/35||1.87 (0.90–3.90)||46/66||1.04 (0.52–2.07)|
|Body mass index|
|High||34/35||0.87 (0.42–1.79)||42/70||0.58 (0.30–1.11)|
In this study, we observed that the polymorphism (Arg261Gln) in the 12-LOX gene was associated with a decreased risk of colorectal adenoma, whereas no significant association was observed for either of the 2 polymorphisms (−1708 GA and 21 CT) in the 5-LOX gene. We also observed a significant interaction between the 12-LOX Arg261Gln polymorphism and NSAID use, with particularly reduced risk for those individuals with a Gln variant allele who regularly took NSAIDs.
Our results suggested that the Arg261Gln polymorphism in 12-LOX was associated with an approximate 40% reduction in the risk of colorectal adenoma. We observed that metabolites of AA from the 12-LOX pathway, 12-HETE, enhanced carcinogenesis through stimulating cell growth, angiogenesis, tumor invasion, and metastasis.21 Recent evidence from experimental studies also indicated that inhibiting 12-LOX activity significantly reduced epithelial cancer cell proliferation and induced apoptosis.19, 38 The Arg261Gln polymorphism is located in the lipoxygenase domain, which is a highly conserved region necessary for LOX activity.27 This variant amino acid may lower adenoma risk by reducing enzyme activity or active AA metabolites. Functional studies will be necessary to evaluate the enzyme activity of this polymorphism in the future.
In single, hospital-based, case-control study by Goodman et al, those authors observed no such association between the Arg261Gln polymorphism and the risk of colon cancer.39 It is possible that the polymorphism is associated with colorectal adenoma only, although a mechanism for such an effect is difficult to formulate. Another potential issue is that controls in our study were defined by full colonoscopic examination, which can eliminate misclassification bias. In addition, Goodman et al did not include any other putative risk factors for colon cancer as covariates in their analysis, such as diet, physical activity, and NSAID use. Thus, current research remains too limited to draw conclusions regarding the role of this polymorphism in the risk of colorectal neoplasm. Future studies will be needed to confirm our finding that the Arg261Gln polymorphism in the 12-LOX gene is associated with a decreased risk of colorectal adenoma.
It is noteworthy that we also observed a significant interaction between 12-LOX genotype and NSAID use. Epidemiologic studies consistently have associated NSAID use with a decreased risk of colorectal adenoma and cancer.40 Consistent with this finding, NSAID use was associated with a decreased risk of colorectal adenoma in the current study. It has been postulated that the mechanisms by which NSAIDs lead to decreased colorectal carcinogenesis occur through inhibiting the COX-2 pathway; however, findings from a recent study of the chemopreventive activity of NSAID derivatives that no longer have COX-inhibitory activity suggest that there are other chemical targets.41COX-independent mechanisms also are suggested by the finding that some NSAIDs inhibit proliferation and induce cell death in cells that do not express COX-1 and COX-2.42 There is little evidence indicating an association between NSAID use and the LOX pathway. However, we observed a statistically significant interaction between this polymorphism and NSAID use, suggesting that mechanisms other than COX inhibition may be contributing to the anti-inflammatory and anticarcinogenic effect of NSAIDs on colorectal adenoma risk. Alternatively, the Gln261 mutation of the 12-LOX gene may alter its effective use of cellular AA, thus leaving more AA for the COX pathway and allowing the effect of NSAIDs to be more apparent.
Epidemiologic studies have implicated high-fat diets in increasing the risk of colorectal cancer.43 Polyunsaturated fatty acids (PUFAs), in particular AA, have been linked to colorectal carcinogenesis. AA and linoleic acid are the most abundant PUFAs in the Western diet.44 In the current study, no significant associations were observed between total fat or total PUFAs and colorectal adenoma. Also, no significant interactions were detected. These findings may be explained in part by the use of the FFQ to assess dietary intake of foods and nutrients. It has been demonstrated that this self-reported measurement method results in biases, which may result in further complicating the picture.45
We did not observe a significant association between 5-LOX polymorphisms and colorectal adenoma. A recently published study also indicated that there was no association between the −1708 GA polymorphism and colorectal adenoma.46 The 2 5-LOX polymorphisms result in synonymous changes. It is possible that the 2 polymorphisms do not change the function of 5-LOX or that the magnitude of changes were insufficient to be detected in the current study with its relatively small sample size. It is worth noting that a consistent pattern was observed when we examined interactions between these 2 polymorphisms and physical activity or NSAID use. In addition, a significantly increased risk was observed among individuals with a variant allele who also had a high WHR. Thus, the effect of an individual polymorphism on phenotype may be small, but it may interact with other factors to increase predisposition to disease.
This study had several strengths. Self-administered questionnaire data were collected prior to diagnosis, thus minimizing the probability of recall bias. Detailed information on other important risk factors (eg, NSAID use, physical activity, dietary factors, and smoking) for colorectal adenoma was collected, allowing us to examine the effects of genetic factors while controlling for other covariates. Data on adenoma characteristics were collected carefully, which allowed us to assess genotypes and adenoma risk according to different characteristics of the adenomas. Finally, because of the high prevalence of adenoma in the general population, cases and controls were defined accurately by full colonoscopic examination to minimize bias resulting from misclassification of adenoma status.
This study also had several limitations. First, although the colonoscopy-based design minimized possible misclassification bias, the study population may not have been fully representative of the general population, because such individuals may be at higher risk for colorectal cancer. However, this would tend to attenuate associations. Second, small numbers in certain subgroups limited the power to detect possible interactions. Third, colorectal adenomas appear to share the same spectrum of risk factors that are observed with colon cancer, but not all adenomas progress to cancer. In counterpoint, an advantage of studying adenomas is that they are asymptomatic, which reduces the likelihood of recall bias to some extent. Studies of adenomas may help to identify factors that are important in the etiology of colorectal cancers. Moreover, like in many epidemiologic studies, dietary intake measurements relied on the FFQ, which is a self-reported instrument with known limitations.47, 48 Finally, even though our key findings were plausible biologically and our risk estimates were statistically significant, there were no functional studies that tested the enzyme activity of the 12-LOX Arg261Gln polymorphism. Therefore, future studies will be necessary to confirm the intriguing association raised by this study.
Although further studies will be needed to determine definitely whether these polymorphisms modify the risk of colorectal neoplasms, the current results provide some evidence that understanding these genes may help to provide the basis for lifestyle or chemoprevention interventions for individuals who are at increased risk for colorectal adenoma and cancer.
- 1American Cancer Society. Cancer Facts and Figures, 2005. Atlanta, Ga: American Cancer Society; 2005.
- 2Cancer Incidence in Five Continents. Lyon: International Agency for Research on Cancer; 1992..
- 35SAS Institute Inc. SAS Language: Reference, Version 9.1. Cary, NC: SAS Institute; 2004.
- 37An Introduction to Categorical Data Analysis. New York: John Wiley & Sons, Inc.; 1996..