Worldwide, colorectal cancer (CRC) is the third most commonly diagnosed cancer and the second leading cause of cancer death, with an estimated 1.9 million new cases and 0.9 million deaths in 2020 (1). There is clear evidence that 5 healthy lifestyle behaviors, namely, nonsmoking, avoiding being overweight, being physically active, consuming no or low amounts of alcohol, and having a healthy diet, reduce the risk of CRC (2,3). The number of healthy behaviors is inversely associated with CRC risk (4–6). In the European Prospective Investigation into Cancer (EPIC), when combining these 5 behaviors in a lifestyle index, the risk of CRC was 12% lower for each additional healthy behavior (4). In the Nurses' Health Study and Health Professionals Follow-up Study, it was also shown that an increasing healthy lifestyle index score, based on the same 5 behaviors, was associated with a reduced CRC risk independently from participation in endoscopic screening (6).
In planning cancer preventive strategies, researchers and policy makers should be aware not only of the impact of lifestyle behaviors on cancer risk but also how changing lifestyle in the recommended direction will affect cancer risk. Nonetheless, how lifestyle changes affect the risk of cancer, and specifically CRC, remains largely unexplored. A randomized trial showed that a successful intervention on smoking cessation and diet improvement can reduce the risk of lifestyle-related cancers in men with high risk for cancer (7). Observational studies have shown that increasing physical activity levels during adult life are associated with lower cancer mortality (8,9), and that improving the cardiorespiratory physical fitness is associated with reduced cancer risk and mortality (10). A recent observational study in Swedish women suggested that lifestyle improvements are associated with a lower risk of lifestyle-related cancers (11). Regarding the specific risk of CRC, a 2020 meta-analysis indicated that smoking cessation significantly reduces the risk of CRC after 25 years since quitting (2), and a 2015 meta-analysis showed that large weight gains from early adulthood to midlife are associated with an increased risk of CRC (12,13). Changes in other lifestyle factors related to CRC risk have not been examined.
With the aim of contributing to the limited knowledge base on lifestyle changes and risk of CRC, in this study, we investigated the effect of changes in smoking habits, body mass index (BMI), physical activity level, alcohol consumption, and a lifestyle index, which combined those 4 factors, on the subsequent incidence of CRC in the EPIC cohort.
From 1992 to 2000, 521,323 participants mostly aged from 35 to 70 years were recruited, mostly from the general population, across 23 centers in 10 European countries: Denmark, France, Germany, Greece, Italy, the Netherlands, Norway, Spain, Sweden, and the United Kingdom. The rationale, study design, and methods for EPIC have been described in detail elsewhere (14). All participants completed a lifestyle questionnaire at baseline and provided informed consent to participate in the study. Ethical approval was obtained from participating centers and the IARC Ethics Committee (reference number 20-02).
We initially excluded 28,561 participants from Greece because of administrative and data use restrictions, 24,550 participants with prevalent cancer at baseline, 9,064 with extreme energy intakes (i.e., below the first and above the 99th percentiles of the energy intake over energy requirement ratio distribution), and 3,137 without follow-up after the baseline questionnaire (Figure 1). After an average of 7 (range 2–17) years from recruitment, a second lifestyle questionnaire was administered during follow-up. Because the main exposure of interest of the current analysis was lifestyle changes, we further excluded 16,816 participants with cancer before the follow-up questionnaire and 100,828 participants for whom the follow-up lifestyle assessment questionnaire data were not available in the centralized EPIC data set in October 2020. We then excluded 3,426 participants for whom no follow-up time was available after the follow-up questionnaire date. We additionally excluded 5,900 and 11,419 participants for whom information about the 4 lifestyle factors of interest—smoking status, alcohol consumption, BMI, and physical activity—was missing at the baseline questionnaire and at the follow-up questionnaire, respectively. We finally excluded 21,757 participants for whom information of at least 1 of the 4 factors of interest was missing both at baseline and follow-up. Hence, the final analytic data set included 295,865 participants. Participants were not involved in the design, conduct, reporting, or dissemination plans of our research.
Four lifestyle factors were investigated: smoking status, alcohol consumption, BMI, and physical activity. For each factor, scores ranging from 0 to 4 were assigned to increasingly healthy categories of behavior (see Supplementary Figure 1, Supplementary Digital Content, https://links.lww.com/AJG/C740). The “healthiest” behaviors were never smoking (never smoked = 4 points, smoke cessation >10 years = 3, smoke cessation ≤10 years = 2, current smoking ≤15 cigarettes/d = 1, current smoking>15 cigarettes/d = 0), low consumption of alcohol (<6.0 g/d = 4 points, 6.0–11.9 = 3, 12.0–23.9 = 2, 24.0–59.9 = 1, ≥60 = 0), top quintile of physical activity based on recreational and household metabolic equivalent of task units (MET) (5th quintile = 4 points, 4th quintile = 3, 3rd quintile = 2, 2nd quintile = 1, and 1st quintile = 0), and low BMI (<22 = 4 points, 22–23.9 = 3, 24–25.9 = 2, 26–29.9 = 1, and ≥30 = 0). A healthy lifestyle index (HLI) was obtained by summing the scores of each lifestyle factor, thus ranging from 0 to 16. Changes in the HLI from the baseline questionnaire to the follow-up questionnaire were our main exposure of interest.
Information on diet was available only at baseline and was therefore not included in the HLI for the current analysis. Intakes of 6 dietary factors—namely, cereal fiber, red and processed meat, the ratio of polyunsaturated to saturated fat, margarine, glycemic load, and fruits and vegetables—were combined only at baseline in a diet score (15), which was used as an adjustment variable in all analyses.
Cases of CRC were identified through population cancer registries in Denmark, Italy, the Netherlands, Norway, Spain, Sweden, and the United Kingdom. A combination of methods was used, including health insurance records, contacts with cancer and pathology registries, and active follow-up of EPIC participants and their next of kin in France and Germany.
CRC cases were defined as carcinomas with topography codes C18, C19, and C20 according to the 10th Revision of the International Statistical Classification of Diseases, Injuries and Causes of Death. In addition to CRC, we also examined associations for the following subsites: proximal colon (C18.0-C18.5), distal colon (C18.6-C18.7), and rectum (C19-C20). When analyzing colorectum subsites, CRC coded as C18.8 (overlapping more than 1 subsite) and C18.9 (unspecified subsite) were censored.
Categorical variables were summarized as frequencies and percentages, and continuous variables as means, medians, SDs, and interquartile ranges. In forest plots, HLI and HLI changes were summarized as mean values. In survival analyses, participants were followed from return of follow-up questionnaire until any first cancer, excluding nonmelanoma skin cancer, death, emigration, or end of follow-up, whichever came first. Kaplan-Meier survival curves were constructed, separately by tertiles of HLI at baseline and stratified by HLI at follow-up. Multivariable Cox proportional hazards regression models, using participants' age as the underlying time scale, were used to estimate hazard ratios (HR) and the corresponding 95% confidence intervals (CI). We used 2 decimals for the CI, but we reported 3 decimals in some cases to show full statistical significance. The models were stratified by study center, age at recruitment rounded to 1 year, and sex, and adjusted for the highest education level achieved (none or primary; technical, professional, or secondary; university or higher; and missing), diet score at baseline, and the calendar date of follow-up questionnaire. The models with lifestyle changes as the main exposure were additionally adjusted for the continuous HLI score at baseline.
To estimate the association between lifestyle changes and risk of CRC, we used the difference between the HLI score at follow-up and the HLI score at baseline both as a continuous variable (model 1) and a categorical variable (according to 7 groups: ≤−3, −2, −1, 0, 1, 2, and ≥3; model 2). In an additional model, we estimated the associations between the changes in the HLI's individual 4 components (mutually adjusted; Model 3) and CRC risk. Similar models were conducted stratified by sex, age (≤55 or >55 years at baseline), and time between questionnaires (≤median or > median value). We investigated heterogeneity of the estimates between the strata using the Cochran Q test. HR were presented by the colorectal subsite. For example, when proximal colon cancer was the outcome of interest, the observations of the participants who had a diagnosis of distal colon cancer and rectum cancer were censored at the date of diagnosis.
Because of the small proportion of participants with complete information on all 4 components of the HLI at both questionnaires, for the main analysis we used a multivariate normal missing imputation (MI) model, which included baseline and follow-up smoking status, alcohol consumption, BMI, and physical activity, and relevant covariates: study center, sex, educational level, age at follow-up questionnaire, time between questionnaires (log-transformed), diet score at baseline, CRC status, and the time to event or censorship (log-transformed). For the ordinal variables, we followed the projected distance rounding method, based on indicators. We generated 15 imputed data sets, analyzed each data set individually, and then combined the estimates using the Rubin rules (16,17). As a sensitivity analysis, we conducted a complete case analysis limited to individuals with nonmissing data for the 4 components in both questionnaires (Figure 1). Further sensitivity analyses were performed, starting the observation time 1 year and 2 years after the follow-up questionnaire, to reduce the risk of potential reverse causation caused by changes in lifestyle due to early symptoms of undiagnosed CRC.
Results with P value <0.05 were considered statistically significant. Analyses were performed using SAS software, version 9.4 (SAS Institute, Cary, NC) and R software, version 4.0.2 (R Foundation for Statistical Computing, Vienna, Austria).
Among 295,865 participants, 2,799 CRC cases were observed over a median of 7.8 years. The median time between the baseline and the follow-up questionnaire was 5.7 years (mean 7.0 years, interquartile range 5.0–9.9). Follow-up characteristics of 212,719 women and 83,146 men included in the study are reported in Table 1, overall and according to HLI score changes. The median age, BMI, physical activity MET, and alcohol consumption among women were 58.6 years, 24.5, 76.0, and 4.1 g/d, respectively, and 59.2 years, 26.4, 60.0, and 13.5 g/d, respectively, among men. There were 15.1% current smokers among women and 21.0% among men.
The mean HLI scores at baseline and at follow-up were 10.04 (SD 2.8) and 9.95 (SD 2.7) units, respectively (Figure 2). The mean HLI score change was −0.09 units (SD 2.1). The largest positive HLI change was observed in participants from Denmark, whereas the largest negative HLI change was observed in participants from Norway. Men had more favorable changes compared with women, overall and in each country. We observed larger mean HLI score changes in men with higher education (HLI change = 0.20) compared with men with lower education (0.11); we observed smaller HLI decreases in women with higher education (−0.18), compared with women with lower education (−0.29).
The association between lifestyle changes and the risk of CRC is given in Table 2. A 1-unit increase in HLI from baseline to follow-up was associated with a 3% lower risk of CRC (Model 1; HR = 0.97; 95% CI 0.95–0.997). Similar inverse associations, although not statistically significant, were observed for the risk of proximal colon cancer (HR 0.96; 95% CI 0.92–1.00), distal colon cancer (HR 0.98; 95% CI 0.94–1.03), and rectal cancer (HR 0.98; 95% CI 0.94–1.02). Compared with no change in HLI, decrements of ≥3 units were associated with higher CRC risk (HR 1.21; 95% CI 1.02–1.43), whereas increments of ≥3 units were associated with a lower risk of CRC (HR 0.88; 95% CI 0.74–1.05; Model 2). Increases in the alcohol, BMI, and physical activity scores, but not in the smoking score, showed a trend toward an inverse association with CRC risk (Model 3). Increases in the alcohol score (i.e., decreases in alcohol consumption) were significantly associated with a lower risk of CRC in participants aged 55 or younger (mean age 46 years) at baseline. Increases in the physical activity score were significantly associated with a lower risk of proximal colon cancer. Increases in the smoking score were significantly associated with a higher risk of CRC in participants who were younger at baseline and with a higher risk of distal colon cancer. We found that a 1-unit increase in HLI from baseline to follow-up was associated with a 3% lower risk of CRC in individuals with time between questionnaires ≤5.7 years (HR 0.97; 95% CI 0.95–1.00) and with time between questionnaires > 5.7 years (HR 0.97; 95% CI 0.93–1.01; p for heterogeneity 0.971).
Among participants with a baseline score of HLI ≤9 (bottom tertile), those with a follow-up score of HLI ≥12 (top tertile) had a lower risk of CRC (Figure 3a; HR 0.77; 95% CI 0.59–1.00) than those with a follow-up score of HLI ≤9. The crude CRC incidence rates in the 2 groups were 134 and 162 per 100,000 person-years, respectively. Among participants with a baseline score of HLI ≥12, those with a follow-up score of HLI ≤9 had a higher risk of CRC (Figure 3c; HR 1.34; 95% CI 1.02–1.75) compared with those with a follow-up score of HLI ≥12. The crude CRC incidence rates in the 2 groups were 119 and 86 per 100,000 person-years, respectively.
Compared with participants in the MI analysis, participants in the complete case analysis were younger (mean age 55.3 vs 58.3 years) and included a larger proportion of men (34.7% vs 28.1%) at the follow-up questionnaire. In the complete case analysis, the associations were generally stronger than those in the MI analysis (see Supplementary Table 1, Supplementary Digital Content, https://links.lww.com/AJG/C742). For example, the HR for CRC for a 1-unit increase in the HLI change was 0.95 (95% CI 0.92–0.99). Significant associations were observed also in women, in younger individuals, and for the proximal subsite. The complete case analysis for the mean HLI changes stratified by sex and country is reported in Supplementary Figure 2 (see Supplementary Digital Content, https://links.lww.com/AJG/C741). The results did not change substantially when the first year and the first 2 years of follow-up were excluded from the analysis (data not shown).
In this analysis conducted in the EPIC cohort, lifestyle behaviors were assessed twice, at baseline and once during follow-up. We showed that lifestyle changes across several years between these 2 time points were associated with the subsequent risk of CRC. Specifically, each unit increment in the HLI score (i.e., toward a healthier lifestyle) was associated with a significant 3% lower risk of CRC, after adjustment for baseline HLI. When the HLI score was analyzed in tertiles, improvement from an unfavorable lifestyle (score 0–9) to a favorable one (score 12–16) was associated with a 23% lower risk of CRC, compared with no change. On the other hand, a decline from a favorable lifestyle (score 12–16) to an unfavorable one (score 0–9) was associated with a 34% higher risk of CRC, compared with no change.
Similar associations between HLI changes and CRC risk were observed in men and women, in different age groups, and for specific cancer sites, although the associations were only statistically significant in men. The complete case analysis showed generally stronger results than the MI analysis, with a statistically significant association between an increase in HLI score and a lower risk of CRC in the whole population, in women, in individuals of 55 years or younger at baseline, and for the proximal subsite. To further evaluate the beneficial effect of an increase in HLI score and the detrimental effect of a decrease in HLI score, we divided the population according to tertiles of baseline HLI score. Notably, changing from an unfavorable to a favorable lifestyle was inversely associated with the risk of CRC, whereas changing from a favorable to an unfavorable lifestyle was positively associated with the risk of CRC.
Changes in the BMI score from baseline to follow-up showed a trend toward an association with CRC risk. In 2 previous EPIC studies, body weight gain from age 20–50 years was associated with an increased risk of CRC (14), whereas weight changes after age 50 were not (18). In a 2015 meta-analysis, which included those 2 EPIC studies, the authors found that large body weight gains from early adulthood to midlife were associated with an increased risk of CRC, whereas no association was found for large body weight gains from midlife to late in life, or for moderate body weight gains or weight loss at any age (12). Similar results indicating that body weight gains in early adulthood, but not late adulthood, were positively associated with CRC risk were found in the Nurses' Health Study and Health Professionals Follow-up (19). A 2019 study among 81,388 individuals, aged 55–74 years, did not show a clear association between body weight change and the risk of CRC (20). This evidence suggests that there might be an effect of body weight gain in early adulthood, but not in older adulthood, on the risk of CRC. In this study, we did not find an interaction between age at baseline and body weight gain, possibly due to the relatively high mean age in our population at follow-up.
We found that decreasing alcohol consumption was associated with a lower risk of CRC, especially in younger individuals, when adjusted for the other components in the HLI and possible confounders. To the best of our knowledge, no previous studies reported evidence on this association. Our findings suggest that preventive measures for reducing alcohol consumption should target people at a young age more forcefully.
We found that increasing levels of physical activity were associated with a lower risk of CRC, specifically proximal cancer. Moreover, in the complete case analysis, we found statistically significant associations between changes in the physical activity level and risk of CRC in the overall population and in younger individuals. Like for alcohol reduction, our results suggest that it might be important to promote physical activity early in life.
We observed that increases in the smoking score, equivalent to reducing smoking levels, were associated with an increased risk of CRC, particularly in the younger individuals and for distal cancer. This may be a result of reverse causation, whereby participants who quit smoking or reduced the number of cigarettes may have experienced early symptoms of CRC. Changes in tobacco exposure were associated with a risk of CRC even after exclusion of the first 2 years of follow-up. Notably, changes in smoking habits performed a marginal influence on the association between HLI and CRC in our study because only a small proportion of the population changed their smoking habits in this study.
Our study shows country-specific and sex-specific differences in HLI score changes. In general, countries with the highest HLI score at baseline (e.g., Norway and United Kingdom) showed a decrease in HLI score at follow-up, while countries with the lowest HLI score at baseline (e.g., Denmark and Sweden) had a higher HLI score at follow-up. In addition, countries with the highest mean age at baseline (e.g., Denmark, Sweden and France) had the most favorable HLI score changes. Unlike baseline age and HLI score, educational levels did not explain the differences in the country-specific HLI changes. Within each country that recruited both men and women, women had a higher HLI score at baseline than men, and women showed lower increases or higher decreases in HLI score at follow-up compared with men. This may have occurred because men, starting with a lower HLI score than women, have more room to improve their lifestyle. Country-specific differences in the questionnaires and their updates at follow-up may also explain the difference between countries. An increase in smoking among Norwegian women has been of concern in recent decades (21) and may explain the worsening in the HLI score in the Norwegian population. However, studies on time trends have shown that the lifestyle of European populations, both in men and women, has generally improved during the last decades (22,23). Altogether, it is difficult to disentangle the contributions of these different changes in score components to our results.
To the best of our knowledge, this is the first study to report an association between multifactorial lifestyle changes and the risk of CRC. According to our results, changing lifestyle habits in adult life is significantly associated with the risk of CRC. If confirmed by other studies, this observation may provide strong evidence to design intervention studies for CRC prevention targeting middle-aged adults, and other research on preventive strategies, which is urgently needed given the scale of the CRC burden (24). An important novel result of our study is that lifestyle changes can affect CRC risk in both directions: improving adherence to a healthy lifestyle was inversely associated with CRC risk, while worsening adherence was positively associated with CRC risk. This is a clear message that practicing clinicians and gastroenterologists could give to their patients and to CRC screening participants to improve CRC prevention. The large sample size and the prospective multicountry and multicenter design of the EPIC cohort are major strengths of the study, including that the results were consistent across different analytical strategies. For the sake of consistency, a scoring system that was used previously in EPIC publications was used in this study. Although specific components, for example, smoking or obesity, might weigh more in the computation of the HLI, this approach has the advantage of ensuring comparability across studies and according to different cancer and other disease outcomes.
Our study has some limitations. We acknowledge that the lack of data on diet collected during follow-up may have led to inadequately adjusted risk estimates and residual confounding. For example, if improvements in diet were associated with both improvements in the HLI score and a decreased CRC risk, then we might have overestimated the association between HLI score and CRC risk. The collection and harmonization of dietary data at follow-up is currently ongoing in EPIC. Furthermore, socioeconomic status affects both lifestyle and CRC risk, and the use of only educational level as a proxy for socioeconomic status may also have led to residual confounding. EPIC participants might not be representative of the general population due to healthy cohort effects, and this warrants cautious interpretation of our findings. However, we can speculate that our findings on the benefit of adopting healthy choices during adulthood might have a larger impact on CRC risk in the general population, characterized by less healthy profiles. Moreover, the HLI score may be too simplistic, assuming equal associations between each lifestyle factor and CRC risk. The HLI score may therefore not accurately capture the complex relationship between lifestyle habits and risk of CRC. However, the main aim of this study was to investigate the association between changes in a multifactorial index summarizing information on major lifestyle factors and the risk of CRC.
This large European cohort study used longitudinal data to show for the first time that changes in lifestyle habits in adult life are associated with the risk of CRC. Favorable changes were associated with a reduced risk of CRC, whereas unfavorable changes were associated with an increased risk of CRC.
CONFLICTS OF INTEREST
Guarantor of the article: Pietro Ferrari, PhD.
Specific author contributions: E.B.: Conceptualization: equal; formal analysis: lead; methodology: lead; supervision: lead; writing-original draft: lead. G.P.: Formal analysis: supporting; methodology: Lead; visualization: lead; writing—original draft: equal; writing-review and editing: supporting. P.B.: Conceptualization: equal; writing-original draft: equal; writing-review and editing: equal. V.B.: Formal analysis: supporting; methodology: supporting; writing-review and editing: supporting. S.L.F.C.: Conceptualization: supporting; methodology: supporting; writing-review and editing: supporting. T.M.S.: Conceptualization: supporting; writing-review and editing: supporting. G.H.: Conceptualization: equal; supervision: supporting; writing-review and editing: supporting. C.C.D.: Writing-review and editing: supporting. C.S.A.: Writing-review and editing: supporting. A.T.: Writing-review and editing: supporting. A.K.E.: Writing-review and editing: and supporting. G.S.: Writing-review and editing: supporting. A.P.-C.: Writing-review and editing: supporting. J.-M.H.: Writing-review and editing: supporting. P.J.: Writing-review and editing: supporting. S.H.: Writing-review and editing: supporting. B.S.: Writing-review and editing: supporting. A.B.: Writing-review and editing: supporting. E.M.M.: Writing-review and editing: supporting. J.W.G.D.: Writing-review and editing: supporting. M.B.S.: Writing-review and editing: supporting. B.B.-d.-M.: Writing-review and editing: supporting. M.-J.S.: Writing-review and editing: supporting. A.J.C.: Writing-review and editing: supporting. K.K.T.: Writing-review and editing: supporting. M.S.D.M.: Writing-review and editing: supporting. R.K.: Writing-review and editing: supporting. V.K.: Writing-review and editing: supporting. J.A.R.: Writing-review and editing: supporting. N.L.: Writing-review and editing: supporting. G.S.: Writing-review and editing: supporting. P.A.: Writing-review and editing: supporting. P.C.: Writing-review and editing: supporting. C.S.: Writing-review and editing: supporting. M.G.: Writing-review and editing: supporting. M.T.: Writing-review and editing: supporting. H.F.: Writing-review and editing: supporting. V.V.: Writing-review and editing: equal. E.W.: Writing-review and editing: supporting. E.R.: Writing-review and editing: supporting. M.J.G.: Conceptualization: equal; supervision: equal; writing-original draft: supporting; writing-review and editing: lead. M.J.: Conceptualization: equal; supervision: equal; writing-review and editing: lead. P.F.: Conceptualization: lead; data curation: lead; formal analysis: equal; funding acquisition: equal; methodology: equal; project administration: lead; supervision: lead; writing-review and editing: lead.
Financial support: The study was supported by the grant LIBERTY (AAP SHS-E-SP 2020, PI: P Ferrari) from the French Institut National du Cancer (INCa). The coordination of EPIC is financially supported by International Agency for Research on Cancer (IARC) and also by the Department of Epidemiology and Biostatistics, School of Public Health, Imperial College London, which has additional infrastructure support provided by the NIHR Imperial Biomedical Research Centre (BRC). The national cohorts are supported by Danish Cancer Society (Denmark); Ligue Contre le Cancer, Institut Gustave Roussy, Mutuelle Générale de l’Education Nationale, and Institut National de la Santé et de la Recherche Médicale (INSERM) (France); German Cancer Aid, German Cancer Research Center (DKFZ), German Institute of Human Nutrition Potsdam-Rehbruecke (DIfE), and Federal Ministry of Education and Research (BMBF) (Germany); Associazione Italiana per la Ricerca sul Cancro-AIRC-Italy, Compagnia di SanPaolo and National Research Council (Italy); Dutch Ministry of Public Health, Welfare and Sports (VWS), Netherlands Cancer Registry (NKR), LK Research Funds, Dutch Prevention Funds, Dutch ZON (Zorg Onderzoek Nederland), World Cancer Research Fund (WCRF), and Statistics Netherlands (The Netherlands); Health Research Fund (FIS)—nstituto de Salud Carlos III (ISCIII), Regional Governments of Andalucía, Asturias, Basque Country, Murcia and Navarra, and the Catalan Institute of Oncology (ICO) (Spain); Swedish Cancer Society and Swedish Research Council and County Councils of Skåne and Västerbotten (Sweden); and Cancer Research UK (14136 to EPIC-Norfolk; C8221/A29017 to EPIC-Oxford) and Medical Research Council (1000143 to EPIC-Norfolk; MR/M012190/1 to EPIC-Oxford) (United Kingdom).
Potential competing interests: None to report.
IARC disclaimer: Where authors are identified as personnel of the International Agency for Research on Cancer/World Health Organization, the authors alone are responsible for the views expressed in this article and they do not necessarily represent the decisions, policy or views of the International Agency for Research on Cancer/World Health Organization.
WHAT IS KNOWN
- ✓ Leading a healthy lifestyle reduces the risk of colorectal cancer (CRC).
- ✓ There is limited knowledge on the impact of lifestyle changes on the risk of CRC.
WHAT IS NEW HERE
- ✓ We measured changes in lifestyle among 300,000 participants in the European Prospective Investigation into Cancer cohort, using a baseline and a follow-up questionnaire.
- ✓ Improving adherence to a healthy lifestyle was inversely associated with CRC risk.
- ✓ Worsening adherence was positively associated with CRC risk.
- ✓ These results justify recommendations for healthy lifestyle changes and healthy lifestyle maintenance for CRC prevention.
1. Sung H, Ferlay J, Siegel RL, et al. Global cancer Statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clinicians 2021;71(3):209–49.
2. Botteri E, Borroni E, Sloan EK, et al. Smoking and colorectal cancer risk, overall and by molecular subtypes: A meta-analysis. Am J Gastroenterol 2020;115(12):1940–9.
3. World Cancer Research Fund/American Institute for Cancer Research. Diet, Nutrition, Physical Activity and Cancer: a Global Perspective. Continuous Update Project Expert Report 2018 (www.dietandcancerreport.org
). Accessed May 15, 2022.
4. Aleksandrova K, Pischon T, Jenab M, et al. Combined impact of healthy lifestyle factors on colorectal cancer: A large European cohort study. BMC Med 2014;12(1):168.
5. Kirkegaard H, Johnsen NF, Christensen J, et al. Association of adherence to lifestyle recommendations and risk of colorectal cancer: A prospective Danish cohort study. BMJ 2010;341:c5504.
6. Wang K, Ma W, Wu K, et al. Healthy lifestyle, endoscopic screening, and colorectal cancer incidence and mortality in the United States: A nationwide cohort study. Plos Med 2021;18(2):e1003522.
7. Botteri E, de Lange T, Tonstad S, Berstad P. Exploring the effect of a lifestyle intervention on cancer risk: 43-year follow-up of the randomized oslo diet and antismoking study. J Intern Med 2018;284(3):282–91.
8. Mok A, Khaw KT, Luben R, et al. Physical activity trajectories and mortality: Population based cohort study. BMJ 2019;365:l2323.
9. Saint-Maurice PF, Coughlan D, Kelly SP, et al. Association of leisure-time physical activity across the adult life course with all-cause and cause-specific mortality. JAMA Netw Open 2019;2(3):e190355.
10. Robsahm TE, Heir T, Sandvik L, et al. Changes in midlife fitness, body mass index, and smoking influence cancer incidence and mortality: A prospective cohort study in men. Cancer Med 2019;8(10):4875–82.
11. Botteri E, Berstad P, Sandin S, Weiderpass E. Lifestyle changes and risk of cancer: Experience from the Swedish women's lifestyle and health cohort study. Acta Oncologica 2021;60(7):827–34.
12. Karahalios A, English DR, Simpson JA. Weight change and risk of colorectal cancer: A systematic review and meta-analysis. Am J Epidemiol 2015;181(11):832–45.
13. Riboli E, Hunt KJ, Slimani N, et al. European prospective investigation into cancer and nutrition (EPIC): Study populations and data collection. Public Health Nutr 2002;5(6B):1113–24.
14. Aleksandrova K, Pischon T, Buijsse B, et al. Adult weight change and risk of colorectal cancer in the European Prospective Investigation into Cancer and Nutrition. Eur J Cancer 2013;49(16):3526–36.
15. McKenzie F, Ferrari P, Freisling H, et al. Healthy lifestyle and risk of breast cancer among postmenopausal women in the European Prospective Investigation into Cancer and Nutrition cohort study. Int J Cancer 2015;136(11):2640–8.
16. Lee KJ, Galati JC, Simpson JA, Carlin JB. Comparison of methods for imputing ordinal data using multivariate normal imputation: A case study of non-linear effects in a large cohort study. Stat Med 2012;31(30):4164–74.
17. Rubin DB, Schenker N. Multiple imputation in health-are databases: An overview and some applications. Stat Med 1991;10(4):585–98.
18. Steins Bisschop CN, van Gils CH, Emaus MJ, et al. Weight change later in life and colon and rectal cancer risk in participants in the EPIC-PANACEA study. Am J Clin Nutr 2014;99(1):139–47.
19. Song M, Hu FB, Spiegelman D, et al. Adulthood weight change and risk of colorectal cancer in the Nurses' health study and health professionals follow-up study. Cancer Prev Res 2015;8(7):620–7.
20. Li JB, Luo S, Wong MCS, et al. Longitudinal associations between BMI change and the risks of colorectal cancer incidence, cancer-relate and all-cause mortality among 81, 388 older adults : BMI change and the risks of colorectal cancer incidence and mortality. BMC Cancer 2019;19(1):1082.
21. Hansen MS, Licaj I, Braaten T, et al. The fraction of lung cancer attributable to smoking in the Norwegian Women and Cancer (NOWAC) Study. Br J Cancer 2021;124(3):658–62.
22. Finger JD, Busch MA, Heidemann C, et al. Time trends in healthy lifestyle among adults in Germany: Results from three national health interview and examination surveys between 1990 and 2011. Plos One 2019;14(9):e0222218.
23. Hansen H, Johnsen NF, Molsted S. Time trends in leisure time physical activity, smoking, alcohol consumption and body mass index in Danish adults with and without COPD. BMC Pulm Med 2016;16(1):110.
24. Lawler M, Alsina D, Adams RA, et al. Critical research gaps and recommendations to inform research prioritisation for more effective prevention and improved outcomes in colorectal cancer. Gut 2018;67(1):179–93.