Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Jan 31.
Published in final edited form as: J Prim Prev. 2015 Feb;36(1):21–31. doi: 10.1007/s10935-014-0372-1

A Healthy Lifestyle Index Is Associated With Reduced Risk of Colorectal Adenomatous Polyps Among Non-Users of Non-Steroidal Anti-Inflammatory Drugs

Fred K Tabung 1,2, Susan E Steck 1,2,3, James B Burch 1,2,4, Chin-Fu Chen 5, Hongmei Zhang 6, Thomas G Hurley 1,2,3, Philip Cavicchia 2, Melannie Alexander 1,2,4, Nitin Shivappa 1,2, Kim E Creek 7, Stephen C Lloyd 8, James R Hebert 1,2,3
PMCID: PMC4289087  NIHMSID: NIHMS636870  PMID: 25331980

Abstract

In a Columbia, South Carolina-based case-control study, we developed a healthy lifestyle index from five modifiable lifestyle factors (smoking, alcohol intake, physical activity, diet, and body mass index), and examined the association between this lifestyle index and the risk of colorectal adenomatous polyps (adenoma). Participants were recruited from a local endoscopy center and completed questionnaires related to lifestyle behaviors prior to colonoscopy. We scored responses on each of five lifestyle factors as unhealthy (0 point) or healthy (1 point) based on current evidence and recommendations. We added the five scores to produce a combined lifestyle index for each participant ranging from 0 (least healthy) to 5 (healthiest), which was dichotomized into unhealthy (0–2) and healthy (3–5) lifestyle scores. We used logistic regression to calculate odds ratios (OR) and 95% confidence intervals (CI) for adenoma with adjustment for multiple covariates. We identified 47 adenoma cases and 91 controls. In the main analyses, there was a statistically nonsignificant inverse association between the dichotomous (OR 0.54; 95% CI 0.22, 1.29) and continuous (OR 0.75; 95%CI 0.51, 1.10) lifestyle index and adenoma. Odds of adenoma were significantly modified by the use of non-steroidal anti-inflammatory drugs (NSAIDs) (pinteraction=0.04). For participants who reported no use of NSAIDs, those in the healthy lifestyle category had a 72% reduction in odds of adenoma as compared to those in the unhealthy category (OR 0.28; 95%CI 0.08, 0.98), whereas a one-unit increase in the index significantly reduced odds of adenoma by 53% (OR 0.47; 95% CI 0.26, 0.88). Although these findings should be interpreted cautiously given our small sample size, our results suggest that higher scores from this index are associated with reduced odds of adenomas, especially in nonusers of NSAIDs. Lifestyle interventions are required to test this approach as a strategy to prevent colorectal adenomatous polyps.

Keywords: Colorectal adenomatous polyps, Healthy lifestyle index, NSAIDs

INTRODUCTION

Colorectal cancer is the third most common cancer and second leading cause of cancer mortality among men and women in the United States (Siegel et al., 2013). Adenomatous polyps (adenomas) are precursor lesions to colorectal cancer (Burnett-Hartman et al., 2012; Sillars-Hardebol et al., 2012; Winawer, 1999). The adenoma-carcinoma sequence is a series of well characterized steps leading to colorectal cancer (Rouillier et al., 2005; Triantafillidis et al., 2009; Vogelstein & Kinzler, 1993). Although, several modifiable risk factors have been implicated in causing colorectal cancer, they are typically studied independently (Wei et al., 2009). Cigarette or tobacco smoking has consistently been associated with incident colorectal adenomas (Zisman, 2006) and colorectal cancer (Anderson et al., 2011; Botteri et al., 2008). Several studies have supported a positive association between alcohol intake and colorectal cancer risk (Bagnardi et al., 2001; Cho et al., 2004; Moskal et al., 2007). Expert panels have found sufficient evidence to link overweight, obesity and lack of physical activity to increased colon cancer incidence (Howard et al., 2008; IARC, 2002; WCRF/AICR, 2011). Red and processed meat have been consistently linked to increased risk for colorectal cancer (Ryan-Harshman & Aldoori, 2007; Santarelli et al., 2008), although the association with fat intake is less consistent (Flood et al., 2003; Lin et al., 2004). Results from studies examining the effect of fruit and vegetable intake on colorectal cancer risk have been mixed (Koushik et al., 2007; Riboli & Norat, 2003). Dietary advice for cancer prevention often includes a reduction of red meat and total dietary fat consumption and an increase in the intake of vegetables, fruit and fiber from various sources (USDA & DHHS, 2010).

A few studies (Chan & Giovannucci, 2010; Driver et al., 2007; Kirkegaard et al., 2010) have examined the combined effects of some of these risk factors in relation to colorectal adenoma or colorectal cancer, with the suggestion that a multi-factor lifestyle approach may be more informative in the design of preventive strategies for the disease. Kirkegaard and colleagues, suggested that a lifestyle index based on achievable national and international, public health recommendations would be a practical tool for counselling people on the effect of living in accordance with the recommendations to reduce the risk of certain diseases (Kirkegaard et al., 2010).

In order to study the relationship between colorectal adenomas and healthy lifestyles we created a scoring system consisting of five potentially modifiable lifestyle factors including cigarette smoking, alcohol intake, physical activity, diet (fruit/vegetable and fat intakes), and body mass index (BMI). We ranked study participants in a South Carolina-based pilot study according to their adherence to this scoring system, and examined associations of the combined lifestyle index with colorectal adenomatous polyp occurrence to test the hypothesis that a healthier lifestyle index score would be associated with lower odds of adenoma.

METHODS

Study population

From October 2008 to April 2010, we recruited 143 individuals from a local endoscopy clinic in Columbia, South Carolina. Of 256 participants who expressed interest in the study, a total of 143 (56.0%) completed and returned questionnaire data and agreed to participate. We implemented a pre-consent process in the clinic, where we contacted interested individuals by phone to screen for eligibility. Eligibility criteria included upcoming scheduled colonoscopy, being 30 to 80 years of age, ability to provide a written informed consent, ability to complete the interview in English, and self-identification as either African American or European American (regardless of ethnicity). Participants completed questionnaires related to demographics, dietary intake, dietary supplement use, physical activity, and medication use. We excluded seven participants with missing information on most of the covariates. Forty-three (31.0%) of the study participants were undergoing routine screening while 95 (69.0%) were undergoing surveillance colonoscopies. The study was approved by the Institutional Review Board of the University of South Carolina.

Assessment of modifiable lifestyle factors

Prior to their clinic visit for colonoscopy, participants completed two dietary screeners and a physical activity questionnaire and brought those to the clinic visit at which time study staff checked them for completeness. Participants completed two dietary screeners to determine fruit and vegetable intake (Greene et al., 2008) and percent energy from fat (Thompson et al., 2008). Physical activity that included leisure activity, along with whether the activity took place outdoors or indoors, was assessed using a modified version of the Lifetime Total Physical Activity Questionnaire developed by Friedenreich and coworkers (Friedenreich et al., 1998), which documents the frequency, duration, and length of participation in the reported activity. We summed total duration (minutes/week) for each activity across all activities for each individual to provide an estimate of total physical activity per participant. During the on-site interview, lasting approximately 10–15 minutes, we collected information on demographic characteristics and other colorectal cancer risk factors (i.e., socioeconomic status, smoking status, and family history of cancer).

Case ascertainment

We conducted medical record abstraction to obtain clinical data, including information on the presence and number of colorectal polyps and related histological features (e.g., adenomatous, hyperplastic, or no polyp). We selected cases and controls from the same population of patients attending the endoscopy facility; and obtained information blinded as to the case status of the participant, thus avoiding one of the pitfalls of disease-differential recall in case-control studies. Cases were individuals with at least one incident; non-hereditary (sporadic) adenoma that was histologically confirmed by a pathologist. Controls were subjects who had a biopsy and were histologically confirmed as having hyperplastic polyps, or had no polyps detected during colonoscopy.

Definition of the five lifestyle factors and the combined lifestyle index

We developed a combined healthy lifestyle index for each participant based on the current epidemiological evidence on the risk factors for colorectal cancer and on current national [2008 Physical Activity Guidelines for Americans (DHHS, 2008) and 2010 Dietary Guidelines for Americans (USDA & DHHS, 2010)] and international (World Health Organization) public health guidelines. Each participant received a score of one for each of the risk factors if they were never smokers; drank ≤2 drinks/day for males and ≤1 drink/day for females; were regularly active, i.e. performing ≥150 minutes/week of moderate intensity physical activity or ≥60 minutes/week of vigorous intensity physical activity; had a “normal” BMI of <25kg/m2, [there were no underweight participants (i.e., BMI <18kg/m2)]; otherwise participants received a score of zero for each of these factors (Table 1). The diet quality score was built from two components - combined fruit and vegetable intake and percent energy from fat. Participants received a score of one if they reported ≥2.5 cups of fruits and ≥2.5 cups of vegetables per day (≥5 servings per day) for the first component and if their percent energy from fat was ≤30 for the second component; otherwise they received a score of zero. The two diet scores were combined into a diet quality score, where participants with both low fruit and vegetable and high percent energy from fat intake received a diet quality score of zero, while those with either high fruit/vegetable intake and/or low fat intake received a diet quality score of one.

Table 1.

Factors of the combined lifestyle score

Healthy lifestyle score factor Score Description Percentage
Smoking 0 Former or Current smoker 53.6
1 Never smoker 46.4
Alcohol use 0 High alcohol use: not conforming to recommended daily alcohol intake for the United States (>2 drinks/day for males and >1 drink/day for females) 17.4
1 Limited alcohol use: Conforming to recommended intake levels (=<2 drinks/day for males and =<1 drink/day for females) 82.6
Physical activity (PA) 0 Not active/less active: <150 minutes/week of moderate intensity PA or <60 minutes/week of vigorous intensity PA 78.3
1 Regularly active: >=150 minutes/week of moderate intensity PA or >=60 minutes/week of vigorous intensity PA 21.7
Diet quality╪ 0 Unhealthy diet quality: low FV¥ intake and high fat intake 52.9
1 Healthy diet quality: high FV intake or low fat intake, or both 47.1
Body mass index 0 Overweight or obese: BMI>=25 81.9
1 Normal weight: BMI 18 – <25 18.1
Open in a new tab
╪

For fruits/vegetables, ≥2.5 cups (≈ 5 servings) per day was considered adequate intake, while <30% of energy from fat was considered healthy intake,

¥

FV=fruits/vegetables

We then generated the combined healthy lifestyle index by summing the binary score for each of the five components (smoking, alcohol, physical activity, diet quality, and BMI). We dichotomized the healthy lifestyle index, which ranged from 0 (least healthy) to 5 (healthiest), into unhealthy (0–2) and healthy (3–5) categories.

Statistical analyses

We described participant characteristics using frequencies of lifestyle and demographic variables by adenoma status, and used logistic regression to calculate adjusted odds ratios (OR) and 95% confidence intervals (CI) for the presence or absence of adenomatous polyps. To identify the best model in terms of model likelihood and complexity, we applied a backward elimination procedure. We used a p-value of 0.2 to eliminate covariates from the model and removed covariates having a p-value greater than 0.2 to create “reduced models”. The “reduced model” was then compared with the preceding or “full” model that contained the deleted variable, using the log-likelihood ratio test. We assessed the following covariates for confounding: age (<54, 55–64, ≥65 years), race (African American, European American), sex (male/female), educational level (up to high school, some college, college graduate or higher), family history of colorectal cancer (yes/no), marital status (yes/no), reason for colonoscopy (screening/surveillance), and regular (i.e., at least weekly) use of non-steroidal anti-inflammatory drugs (NSAIDs) (yes/no). We determined potential effect modification by family history of colorectal cancer and use of NSAIDs by generating interaction terms in the multivariable model.

We created four separate healthy lifestyle index variables representing four models. In the first model, we investigated the association of each binary lifestyle index factor with adenomatous polyps, adjusting for the other four factors (smoking, alcohol use, physical activity, diet quality, and BMI), with participants in the 0 category as the referent group. In the second and third models, we examined the association of the dichotomous and continuous healthy lifestyle index, respectively, with adenomatous polyps adjusting for covariates. In the fourth model, we investigated how the odds of adenomas changed with each additional level of the index by calculating ORs for the association of each level of the index (1 through 5) compared to the lowest level (0). We combined the two highest levels because there were only two participants in the highest level (meeting all 5 healthy lifestyle recommendations). We adjusted all four models for age, race, educational status and sex, and stratified the second and third models by NSAIDs use and additionally adjusted for NSAIDs use in the unstratified models. We considered two-sided p-values lower than 0.05 to be statistically significant and used SAS® statistical software version 9.3 (SAS Institute, Cary, NC, USA) for all analyses.

RESULTS

Among the 138 participants retained in the final analyses, 47 were diagnosed with colorectal adenomatous polyps. Participant characteristics (except age groups) did not differ significantly between adenoma cases and non-cases (Table 2). About half (49.3%) of the participants were male, 18.8% had a family history of colorectal cancer, 43.5% used NSAIDs, while 66% were classified by the lifestyle index as having an unhealthy lifestyle (overall score ≤2). Among cases, 30.0% had one adenoma while 17.0% had two or more adenomas. Table 1 shows the proportions of participants in the two categories of each of the five factors that constitute the lifestyle index: 46.4% of participants were never smokers, 82.6% had alcohol intake within the recommended limits, 21.7% followed the recommendation for regular physical activity, 47.1% followed the recommendation for fruit and vegetable intake, while 18.1% were in the normal weight category.

Table 2.

Characteristics of study participants by polyp status

Adenomatous polyp (n=47)
n(%)		 No adenomatous polyp (n=91)
n(%)		 Difference testing; p value
Combined healthy lifestyle score
Unhealthy lifestyle score 34(72.3) 57(62.6) 0.25
Healthy lifestyle score 13(27.7) 37(37.4)
Age category (years)
<54 9(19.1) 29(31.8) 0.04
55–64 17(36.2) 40(44.0)
≥65 21(44.7) 22(24.2)
Race
European Americans 37(78.7) 59(64.8) 0.09
African Americans 10(21.3) 32(35.2)
Sex
Female 20(42.6) 50(54.9) 0.17
Male 27(57.4) 41(45.1)
Years of education
Up to High School 13(27.6) 30(33.0) 0.41
some college 17(36.2) 23(25.3)
College graduate or higher 17(36.2) 38(41.7)
Family history of colorectal cancer
No 38(80.8) 74(81.3) 0.95
Yes 9(19.2) 17(18.7)
Reason for colonoscopy 0.0002
Screening 5 (10.4) 38 (41.8)
Surveillance 42 (89.6) 53 (58.2)
NSAID use
No 29(61.7) 49(53.8) 0.38
Yes 18(38.3) 42(46.2)
Marital status
No 8(17.0) 24(26.4) 0.22
Yes 39(83.0) 67(73.6)
Open in a new tab

Note. The continuous healthy lifestyle score (0–5) was dichotomized by combining 0–2 in one category and 3–5 in the second category.

Table 3 presents the main effect of each lifestyle index factor on the odds of adenoma after adjusting for the other four index factors and for age, race, sex, NSAIDs use and education. Never smoking, limited alcohol intake, regular physical activity, high intake of fruits/vegetables and low fat intake, and normal BMI all showed nonsignificant inverse associations with colorectal adenomatous polyps. OR for the associations between the healthy lifestyle index and the odds of adenoma are presented in Table 4 for all participants and stratified by NSAIDs use. When all participants were considered, there was a statistically nonsignificant inverse association between the dichotomous lifestyle index and adenoma (OR 0.54; 95% CI 0.22, 1.29), comparing participants in the unhealthy versus healthy categories, and with the continuous lifestyle index and odds of adenoma (OR 0.75; 95%CI 0.51, 1.10).

Table 3.

Age-adjusted and fully-adjusted odds ratios (95% confidence intervals) for each factor of the combined lifestyle score

Lifestyle index Factor cases / controls Factor category Age- adjusted OR 95%CI Fully- adjusted OR§ 95%CI
Smoking score 31/43 Current/former smoker 1.00 referent 1.00 referent
16/48 Never smoker 0.48 0.23–1.02 0.65 0.28–1.47
Alcohol score 10/24 High intake 1.00 referent 1.00 referent
37/77 Limited intake 0.54 0.21–1.39 0.74 0.26–2.14
Physical activity score 37/71 No/less activity 1.00 referent 1.00 referent
10/20 Regular activity 0.86 0.35–2.09 0.81 0.29–2.26
Diet score 25/48 Low FV¥ intake and high fat intake 1.00 referent 1.00 referent
22/43 High FV intake or low fat intake or both 0.93 0.45–1.93 0.90 0.39–2.05
BMI¥ score 39/74 Overweight/obese 1.00 referent 1.00 referent
8/17 Normal weight 0.79 0.31–2.07 0.64 0.22–1.86
Open in a new tab
§

In the model, each factor was adjusted for the other four factors and additionally for age, sex, educational status, NSAID use, and race.

¥

FV=fruits/vegetables; BMI = Body Mass Index

Table 4.

Odds ratios (95% confidence intervals) for categorical and continuous healthy lifestyle index stratified by NSAIDs use

All participants No NSAIDs use NSAIDs use
OR* 95%CI OR 95%CI OR 95%CI
Dichotomous healthy lifestyle index Unhealthy lifestyle score 1.00 referent 1.00 referent 1.00 referent
Healthy lifestyle score 0.54 0.22–1.29 0.28 0.08–0.98 1.30 0.35–4.91
Continuous healthy lifestyle index¥ 0.75 0.52–1.10 0.47 0.26–0.88 1.14 0.64–2.05
Open in a new tab

Note. All models were adjusted for age, sex, educational status, race and reason for colonoscopy. NSAIDs-unstratified model was additionally adjusted for NSAIDs use.

¥

The continuous healthy lifestyle score (0–5) was dichotomized by combining 0–2 in one category and 3–5 in the second category.

*

OR=odds ratio (95% confidence interval) of having an adenoma vs. no, normal or hyperplastic polyps.

The association between the index and odds of adenoma varied significantly by NSAIDs use (p value for interaction, pinteraction=0.04), but not by family history of colorectal cancer (pinteraction=0.30). When models for the dichotomous and continuous healthy lifestyle indices were stratified by NSAIDs use, we found that among participants who reported no use of NSAIDs, those in the healthy lifestyle index category (3–5) had a 72% reduced odds of adenoma compared to those in the unhealthy lifestyle category (0–2) (OR 0.28; 95% CI 0.08, 0.98), whereas a one-unit increase in the index (continuous) significantly reduced odds of adenoma by 53% (OR 0.47; 95% CI 0.26, 0.88). For participants who reported using NSAIDs, there was no association of the healthy lifestyle index with adenomas (see Table 4).

There was no association between the lifestyle index and adenoma when each category of the healthy lifestyle index was compared to the referent of 0 (unhealthy): OR and 95%CI were as follows for each level of the healthy lifestyle index compared to 0: index=1: OR 1.43; 95%CI 0.24, 8.65; index=2: OR 0.71; 95%CI 0.14, 3.53; index=3: OR 0.43; 95%CI 0.08, 2.38; index=4/5: OR 0.63; 95%CI 0.08, 4.98.

DISCUSSION

In this case-control study we found that having a higher score on the healthy lifestyle index compared with a lower score was associated with reduced odds of colorectal adenomatous polyps only in non-users of NSAIDs, whereas no association was observed among those who reported using NSAIDs regularly. Findings from the main analysis, though suggestive of inverse associations between a healthy lifestyle index and odds of colorectal adenoma, were not statistically significant.

While the overall association between the healthy lifestyle index and colorectal adenomatous polyps was not statistically significant, the statistically significant association among non-users of NSAIDs was consistent with findings from previous studies (Driver et al., 2007; Fu et al., 2012; Hartman et al., 2005; Kirkegaard et al., 2010; Odegaard et al., 2013; Platz et al., 2000; Wei et al., 2009). The comparability of these studies is, however, limited due to the adoption of different combinations, cut-points and weighting of lifestyle factors. Odegaard and coworkers combined six factors (including all five factors in the present study plus sleep habits) and examined their association with colorectal cancer in a Chinese population. They found that higher index scores were associated with a decreased risk of developing colon (but not rectal) cancer in Chinese men and women (Odegaard et al., 2013). The difference by anatomic subsite is similar to results of other studies we have conducted on risk of colorectal cancer (Cavicchia et al., 2013). In a systematic comparison of six risk factors (cigarette smoking, obesity, no regular use of NSAIDs, high intake of red meat, low intake of fiber, and low intake of calcium) by type of colorectal polyp in the Tennessee Colorectal Polyp Study, Fu and colleagues found that the risk of polyps increased progressively with an increasing number of adverse lifestyle factors. Polyps considered in three separate groups were adenoma only, hyperplastic polyps and synchronous hyperplastic & adenoma (Fu et al., 2012). In a Danish study, Kirkegaard and colleagues investigated the associations of a 5-factor lifestyle index (based on smoking, alcohol intake, physical activity, diet [dietary fiber, energy percentage from fat, red and processed meat, and fruits and vegetables] and waist circumference) and the risk of colorectal cancer. After adjustment for potential confounders, they found that each additional point achieved on the lifestyle index was associated with a lower risk of colorectal cancer (Kirkegaard et al., 2010). Another study generated a combined lifestyle index using the same five factors as in our study, but examined the association of the lifestyle index with risk of pancreatic cancer. The study found that participants with the highest score (5 points), compared to those with the lowest score (0 point), had a 58% reduced risk of pancreatic cancer (Jiao et al., 2009).

Studies have shown that long-term use of low-dose aspirin or non-aspirin NSAIDs prevents the occurrence of colorectal adenomas (Baron et al., 2003; Bertagnolli et al., 2006; García-Rodríguez & Huerta-Alvarez, 2000; Sandler et al., 1998). In the present study we observed a significant interaction between our healthy lifestyle index and NSAIDs use, where an inverse association between the lifestyle index and colorectal adenomas was observed among nonusers of NSAIDs only. None of the studies cited previously on the combination of lifestyle factors and risk of colorectal cancer or colorectal adenomas stratified models by NSAIDs use. Some studies investigating the effect of diet or physical activity (components of the lifestyle index) on the risk of colorectal adenoma have observed significant effect modification by regular NSAIDs use (Hartman et al., 2005; Hauret et al., 2004). Hartman and colleagues (2005) found that a low-fat, high-fiber diet rich in fruits and vegetables reduced the risk of colorectal adenoma recurrence among nonusers of NSAIDs, while Hauret and coworkers (2004) observed that NSAIDs modified the effect of physical activity on incident sporadic colorectal adenoma, with inverse effects restricted to nonusers of NSAIDs. Our results are consistent with the two studies discussed above.

It is possible that the beneficial effect of NSAIDs is so high as to mask any potential beneficial effect of healthy lifestyle factors, which would explain our findings. Another interpretation suggested by Hauret et al (2004) is that the anti-inflammatory effect of NSAIDs on the colonic epithelium is so strong as to render inconsequential the relative contribution of physical activity, and thus a physical activity-adenoma association would not be observed among NSAID users, yet would be strong among nonusers of NSAIDs (Hauret et al., (2004). This would mean that the potential for protective effects of lifestyle modification on the risk of colorectal adenoma is limited overall, and once that level was achieved, no further protective effects would be observed (Hartman et al., 2005). In any event, it is important to remain mindful that an important advantage of lifestyle modification over regular NSAIDs use is that prudent dietary modification does not have the potential side effects associated with the regular use of NSAIDs (Rennie et al., 2003) and is associated with numerous other benefits beyond preventing colorectal cancer (Hebert et al., 1999).

The selection of the five factors used to generate the healthy lifestyle index was based on public health recommendations and current epidemiological evidence. We used BMI (kg/m2) to estimate overweight and obesity but other studies have used waist circumference (WC) with the argument that it captures abdominal fat better than BMI (Kirkegaard et al., 2010). Though WC and BMI are not interchangeable they are usually highly correlated (Vazquez et al., 2007). For example, a meta-analysis of nine prospective British studies revealed no major differences in the discriminatory capabilities of models with BMI, WC or waist-to-hip ratio (WHR) for cardiovascular or total mortality outcomes (Czernichow et al., 2011). Another meta-analysis of 32 studies used random effects models to examine the association of BMI, WC and WHR with risk of diabetes, and found that the three obesity indicators have similar associations with incident diabetes (Vazquez et al., 2007).

We defined diet quality based only on fruit and vegetable intake and fat intake because these were the dietary data that were collected in our study. These two dietary factors may be surrogate markers for specific dietary patterns. A previous study generated a modified lifestyle index with the dietary factor represented by only fruit and vegetable intake, as compared to the original lifestyle index with the diet factor composed of red and processed meat, fruit and vegetables, and percent energy from fat (Kirkegaard et al., 2010). The results in regards to the modified lifestyle index were not materially different from those of the original lifestyle index, which indicates the potential to use the index without a comprehensive diet assessment (Kirkegaard et al., 2010). However, a more comprehensive dietary assessment (such as by multiple 24-hour recalls or food frequency questionnaire) would have allowed for the addition of other variables in the diet quality score, such as red or cooked meat intake, which have been associated with colorectal cancer in other studies (WCRF/AICR, 2011). For the smoking factor, we combined former and current smokers in one category due to the current evidence that risk for smoking-related conditions may persist for up to 25 years after quitting smoking (Gong et al., 2012).

A potential limitation to our study was that all lifestyle index factors as well as data on all other covariates were self-reported. Self-reported measurements are likely to have some degree of misclassification. In this case, the misclassification would likely be non-differential leading to potential attenuation of odds ratios, because all study participants completed questionnaires prior to colonoscopy and knowledge of their adenoma status. We dichotomized covariates, including the healthy lifestyle index, which could have led to loss of information or statistical power and the potential introduction of residual confounding, but results from models constructed with the index as continuous or dichotomized were all statistically significant. Categorizing the healthy lifestyle index also provided an opportunity to directly compare participants who would be considered exposed/not exposed to healthy lifestyle behaviors. Our small sample size limited both the power of the study in most of the models and our ability to stratify by NSAIDs use in the model with the lifestyle index treated as a categorical variable with five categories. The effect estimates in our study also had wide confidence intervals and we cannot rule out that our statistically significant findings may be attributed to chance. Lastly, our study design was cross-sectional and included a large proportion of surveillance (69%) compared with screening (31%) colonoscopies. It is possible that some surveillance colonoscopy participants may have recently changed their lifestyle behaviors as a result of previous adenoma diagnosis. A change in lifestyle behaviors towards more healthy behaviors such as engaging in more physical activity, eating healthier diets, reducing alcohol intake, or losing weight among those who were more likely to have an adenoma (e.g., surveillance colonoscopy patients) would decrease the odds of a protective lifestyle association (i.e., drive the association towards the null). Indeed, adjusting for reason for colonoscopy further strengthened the protective association of the healthy lifestyle index with adenoma, especially in non-users of NSAIDs, suggesting that the potential for a recent change in diet due to symptoms is not a likely explanation for our findings.

Strengths of the present study include our detailed assessment of exposure. Compared to studies that have investigated single lifestyle factors in relation to risk of colorectal adenoma, our study takes the next step by examining the effect of a combined healthy lifestyle index, which captures the influence of multiple health behaviors, on the occurrence of adenomatous polyps. Another strength is that all data were collected before study participants knew of their polyp status, thus obviating the problem of diet-related reporting and information bias.

In conclusion, our findings suggest that a higher score on the combined healthy lifestyle index generated on the basis of recommendations for five lifestyle factors (smoking, alcohol intake, physical activity, diet quality, and BMI) is associated with reduced odds for colorectal adenomas in nonusers of NSAIDs. This study supports the evidence that lifestyle modification is important for the prevention of colorectal adenomas which are precursors of colorectal cancer, and adds to the body of evidence on the beneficial effects of combined lifestyle factors on risk of colorectal adenomatous polyps. Lifestyle interventions are required to test this as a strategy to prevent colorectal adenomatous polyps.

Acknowledgments

This work was supported by a University of South Carolina Research Opportunity Award (PI: SE. Steck) and by a supplemental grant from the National Cancer Institute (NCI) (U01 CA114601-03S5) as part of the South Carolina Cancer Disparities Community Network (SCCDCN, PI: JR Hébert; Co-Project Leaders: JB Burch and SE Steck) and the NCI-funded SCCDCN-II (U54 CA153461-01). Dr. Tabung and Dr. Hebert were supported by National Institutes of Health grants numbers F31CA177255 and K05CA136975, respectively. Dr. Steck was supported by the USC Center for Colon Cancer Research (COBRE 5P20RR017698). Melannie Alexander was supported by the University of South Carolina Behavioral-Biomedical Interface Program, funded in part by training grant T32-5R18CE001240 from the National Institute of General Medical Sciences. We are grateful to the South Carolina Medical Endoscopy Center (SCMEC) for providing their facility for participant recruitment and data collection and to our study participants for volunteering time to participate in the study.

Footnotes

CONFLICT OF INTEREST

The authors declare no conflict of interest

References

  1. Anderson JC, Moezardalan K, Messina CR, Latreille M, Shaw RD. Smoking and the association of advanced colorectal neoplasia in an asymptomatic average risk population: Analysis of exposure and anatomical location in men and women. Digestive Diseases and Sciences. 2011;56(12):3616–3623. doi: 10.1007/s10620-011-1814-8. [DOI] [PubMed] [Google Scholar]
  2. Bagnardi V, Blangiardo M, Vecchia CL, Corrao G. A meta-analysis of alcohol drinking and cancer risk. British Journal of Cancer. 2001;85(11):1700–1705. doi: 10.1054/bjoc.2001.2140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Baron JA, Cole BF, Sandler RS, Haile RW, Ahnen D, Bresalier R, et al. A randomized trial of aspirin to prevent colorectal adenomas. New England Journal of Medicine. 2003;348(10):891–899. doi: 10.1056/NEJMoa021735. [DOI] [PubMed] [Google Scholar]
  4. Bertagnolli MM, Eagle CJ, Zauber AG, Redston M, Scott DS, Kim K, et al. Celecoxib for the prevention of sporadic colorectal adenomas. New England Journal of Medicine. 2006;355(9):873–884. doi: 10.1056/NEJMoa061355. [DOI] [PubMed] [Google Scholar]
  5. Botteri E, Iodice S, Bagnardi V, Raimondi S, Lowenfels AB, Maisonneuve P. Smoking and colorectal cancer: A meta-analysis. Journal of the American Medical Association. 2008;300(23):2765–2778. doi: 10.1001/jama.2008.839. [DOI] [PubMed] [Google Scholar]
  6. Burnett-Hartman AN, Newcomb PA, Phipps AI, Passarelli MN, Grady WM, Upton MP, et al. Colorectal endoscopy, advanced adenomas, and sessile serrated polyps: Implications for proximal colon cancer. American Journal of Gastroenterology. 2012;107(8):1213–1219. doi: 10.1038/ajg.2012.167. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cavicchia P, Adams SA, Steck SE, Hussey JR, Liu J, Daguise VG, et al. Racial disparities in colorectal cancer incidence by type 2 diabetes mellitus status. Cancer Causes & Control. 2013;24(2):277–285. doi: 10.1007/s10552-012-0095-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chan AT, Giovannucci EL. Primary prevention of colorectal cancer. Gastroenterology. 2010;138(6):2029–2043. doi: 10.1053/j.gastro.2010.01.057. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cho E, Smith-Warner SA, Ritz J, van den Brandt PA, Colditz GA, Folsom AR, et al. Alcohol intake and colorectal cancer: A pooled analysis of 8 cohort studies. Annals of Internal Medicine. 2004;140(8):603–613. doi: 10.7326/0003-4819-140-8-200404200-00007. [DOI] [PubMed] [Google Scholar]
  10. Czernichow S, Kengne AP, Stamatakis E, Hamer M, Batty GD. Body mass index, waist circumference and waist–hip ratio: Which is the better discriminator of cardiovascular disease mortality risk? Evidence from an individual-participant: meta-Analysis of 82,864 participants from nine cohort studies. Obesity Reviews. 2011;12(9):680–687. doi: 10.1111/j.1467-789X.2011.00879.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. DHHS. [Accessed 4/8/2013];2008 Physical activity guidelines for Americans. 2008 http://www.health.gov/paguidelines/
  12. Driver JA, Gaziano JM, Gelber RP, Lee IM, Buring JE, Kurth T. Development of a risk score for colorectal cancer in men. The American journal of medicine. 2007;120(3):257–263. doi: 10.1016/j.amjmed.2006.05.055. [DOI] [PubMed] [Google Scholar]
  13. Flood A, Velie EM, Sinha R, Chaterjee N, Lacey JV, Jr, Schairer C, et al. Meat, fat, and their subtypes as risk factors for colorectal cancer in a prospective cohort of women. American Journal of Epidemiology. 2003;158(1):59–68. doi: 10.1093/aje/kwg099. [DOI] [PubMed] [Google Scholar]
  14. Friedenreich CM, Courneya KS, Bryant HE. The lifetime total physical activity questionnaire: Development and reliability. Medicine and Science in Sports and Exercise. 1998;30:266–274. doi: 10.1097/00005768-199802000-00015. [DOI] [PubMed] [Google Scholar]
  15. Fu Z, Shrubsole MJ, Smalley WE, Wu H, Chen Z, Shyr Y, et al. Lifestyle factors and their combined impact on the risk of colorectal polyps. American Journal of Epidemiology. 2012;176(9):766–776. doi: 10.1093/aje/kws157. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. García-Rodríguez LA, Huerta-Alvarez C. Reduced incidence of colorectal adenoma among long-term users of nonsteroidal anti-inflammatory drugs: A pooled analysis of published studies and a new population-based study. Epidemiology. 2000;11(4):376–381. doi: 10.1097/00001648-200007000-00003. [DOI] [PubMed] [Google Scholar]
  17. Gong J, Hutter C, Baron JA, Berndt S, Caan B, Campbell PT, et al. A pooled analysis of smoking and colorectal cancer: Timing of exposure and interactions with environmental factors. Cancer Epidemiology Biomarkers & Prevention. 2012;21(11):1974–1985. doi: 10.1158/1055-9965.EPI-12-0692. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Greene GW, Resnicow K, Thompson FE, Peterson KE, Hurley TG, Hebert JR, et al. Correspondence of the NCI fruit and vegetable screener to repeat 24-H recalls and serum carotenoids in behavioral intervention trials. Journal of Nutrition. 2008;138:200S–204S. doi: 10.1093/jn/138.1.200S. [DOI] [PubMed] [Google Scholar]
  19. Hartman TJ, Yu B, Albert PS, Slattery ML, Paskett E, Kikendall JW, et al. Does nonsteroidal anti-inflammatory drug use modify the effect of a low-fat, high-fiber diet on recurrence of colorectal adenomas? Cancer Epidemiology, Biomarkers & Prevention. 2005;14(10):2359–2365. doi: 10.1158/1055-9965.EPI-05-0333. [DOI] [PubMed] [Google Scholar]
  20. Hauret KG, Bostick RM, Matthews CE, Hussey JR, Fina MF, Geisinger KR, et al. Physical activity and reduced risk of incident sporadic colorectal adenomas: Observational support for mechanisms involving energy balance and inflammation modulation. American Journal of Epidemiology. 2004;159(10):983–992. doi: 10.1093/aje/kwh130. [DOI] [PubMed] [Google Scholar]
  21. Hebert J, Ebbeling CB, Ockene IS, Ma Y, Rider L, Merriam PA, et al. A dietitian-delivered group nutrition program leads to reductions in dietary fat, serum cholesterol, and body weight: Findings from the Worcester area trial for counseling in hyperlipidemia (WATCH) Journla of the American Dietetic Association. 1999;99:544–552. doi: 10.1016/s0002-8223(99)00136-4. [DOI] [PubMed] [Google Scholar]
  22. Howard RA, Freedman DM, Park Y, Hollenbeck A, Schatzkin A, Leitzmann MF. Physical activity, sedentary behavior, and the risk of colon and rectal cancer in the NIH-AARP diet and health study. Cancer Causes & Control. 2008;19(9):939–953. doi: 10.1007/s10552-008-9159-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. IARC. IARC Handbooks of Cancer Prevention. IARC Press; Lyon, France: 2002. Weight control and physical activity; p. 6. [Google Scholar]
  24. Jiao L, Mitrou PN, Reedy J, Graubard IB, Hollenbeck RA, Schatzkin A, et al. A combined healthy lifestyle score and risk of pancreatic cancer in a large cohort study. Archives of Internal Medicine. 2009;169(8):764–770. doi: 10.1001/archinternmed.2009.46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kirkegaard H, Johnsen NF, Christensen J, Frederiksen K, Overvad K, Tjønneland A. Association of adherence to lifestyle recommendations and risk of colorectal cancer: A prospective Danish cohort study. British Medical Journal. 2010:341. doi: 10.1136/bmj.c5504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Koushik A, Hunter DJ, Spiegelman D, Beeson WL, van den Brandt PA, Buring JE, et al. Fruits, vegetables, and colon cancer risk in a pooled analysis of 14 cohort studies. Journal of the National Cancer Institute. 2007;99(19):1471–1483. doi: 10.1093/jnci/djm155. [DOI] [PubMed] [Google Scholar]
  27. Lin J, Zhang SM, Cook NR, Lee IM, Buring JE. Dietary fat and fatty acids and risk of colorectal cancer in comen. American Journal of Epidemiology. 2004;160(10):1011–1022. doi: 10.1093/aje/kwh319. [DOI] [PubMed] [Google Scholar]
  28. Moskal A, Norat T, Ferrari P, Riboli E. Alcohol intake and colorectal cancer risk: A dose–response meta-analysis of published cohort studies. International Journal of Cancer. 2007;120(3):664–671. doi: 10.1002/ijc.22299. [DOI] [PubMed] [Google Scholar]
  29. Odegaard AO, Koh WP, Yuan JM. Combined lifestyle factors and risk of incident colorectal cancer in a Chinese population. Cancer Prevention Research. 2013;6(4):360–367. doi: 10.1158/1940-6207.CAPR-12-0384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Platz EA, Willett WC, Colditz GA, Rimm EB, Spiegelman D, Giovannucci E. Proportion of colon cancer risk that might be preventable in a cohort of middle-aged US men. Cancer Causes & Control. 2000;11(7):579–588. doi: 10.1023/a:1008999232442. [DOI] [PubMed] [Google Scholar]
  31. Rennie KL, Hughes J, Lang R, Jebb SA. Nutritional management of rheumatoid arthritis: A review of the evidence. Journal of Human Nutrition and Dietetics. 2003;16(2):97–109. doi: 10.1046/j.1365-277x.2003.00423.x. [DOI] [PubMed] [Google Scholar]
  32. Riboli E, Norat T. Epidemiologic evidence of the protective effect of fuit and vegetables on cancer risk. The American Journal of Clinical Nutrition. 2003;78(3):559S–569S. doi: 10.1093/ajcn/78.3.559S. [DOI] [PubMed] [Google Scholar]
  33. Rouillier P, Senesse P, Cottet V, Valléau A, Faivre J, Boutron-Ruault MC. Dietary patterns and the adenomacarcinoma sequence of colorectal cancer. European Journal of Nutrition. 2005;44(5):311–318. doi: 10.1007/s00394-004-0525-8. [DOI] [PubMed] [Google Scholar]
  34. Ryan-Harshman M, Aldoori W. Diet and colorectal cancer. Canadian Family Physician. 2007;53(11):1913–1920. [PMC free article] [PubMed] [Google Scholar]
  35. Sandler RS, Galanko JC, Murray SC, Helm JF, Woosley JT. Aspirin and nonsteroidal anti-inflammatory agents and risk for colorectal adenomas. Gastroenterology. 1998;114(3):441–447. doi: 10.1016/s0016-5085(98)70526-8. [DOI] [PubMed] [Google Scholar]
  36. Santarelli RL, Pierre F, Corpet DE. Processed meat and colorectal cancer: A review of epidemiologic and experimental evidence. Nutrition and Cancer. 2008;60(2):131–144. doi: 10.1080/01635580701684872. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA: A Cancer Journal for Clinicians. 2013;63(1):11–30. doi: 10.3322/caac.21166. [DOI] [PubMed] [Google Scholar]
  38. Sillars-Hardebol AH, Carvalho B, van Engeland M, Fijneman RJ, Meijer GA. The adenoma hunt in colorectal cancer screening: Defining the target. The Journal of Pathology. 2012;226(1):1–6. doi: 10.1002/path.3012. [DOI] [PubMed] [Google Scholar]
  39. Thompson FE, Midthune D, Williams GC, Yaroch AL, Hurley TG, Resnicow K, et al. Evaluation of a short dietary assessment instrument for percentage energy from fat in an intervention study. Journal of Nutrition. 2008;138:193S–199S. doi: 10.1093/jn/138.1.193S. [DOI] [PubMed] [Google Scholar]
  40. Triantafillidis JK, Nasioulas G, Kosmidis PA. Colorectal cancer and inflammatory bowel disease: Epidemiology, risk factors, mechanisms of carcinogenesis and prevention strategies. Anticancer Research. 2009;29(7):2727–2737. [PubMed] [Google Scholar]
  41. USDA, & DHHS. 2010 Dietary guidelines for Americans. 2010. [Accessed 4/8/2013 2013]. [Google Scholar]
  42. Vazquez G, Duval S, Jacobs DR, Jr, Silventoinen K. Comparison of body mass index, waist circumference, and waist/hip ratio in predicting incident diabetes: A meta-analysis. Epidemiologic Reviews. 2007;29(1):115–128. doi: 10.1093/epirev/mxm008. [DOI] [PubMed] [Google Scholar]
  43. Vogelstein B, Kinzler KW. The Multistep nature of cancer. Trends in Genetics. 1993;9(4):138–141. doi: 10.1016/0168-9525(93)90209-z. [DOI] [PubMed] [Google Scholar]
  44. Wei EK, Colditz GA, Giovannucci EL, Fuchs CS, Rosner BA. Cumulative risk of colon cancer up to age 70 years by risk factor status using data from the Nurses’ Health Study. American Journal of Epidemiology. 2009;170(7):863–872. doi: 10.1093/aje/kwp210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Winawer SJ. Natural history of colorectal cancer. American Journal of Medicine. 1999;106(1A):3S–6S. 50S–51S. doi: 10.1016/s0002-9343(98)00338-6. [DOI] [PubMed] [Google Scholar]
  46. World Cancer Research Fund / American Institute for Cancer Research. Continuous update project report summary. Food, nutrition, physical activity, and the prevention of colorectal cancer. 2011 http://www.wcrf.org/int/research-we-fund/continuous-update-project-cup.
  47. Zisman Al, Nickolov A, Brand RE, Gorchow A, Roy HK. Associations between the age at diagnosis and location of colorectal cancer and the use of alcohol and tobacco: Implications for screening. Archives of Internal Medicine. 2006;166(6):629–634. doi: 10.1001/archinte.166.6.629. [DOI] [PubMed] [Google Scholar]

RESOURCES