Pairwise association of key lifestyle factors and risk of colorectal cancer: a prospective pooled multicohort study

E Roos, Karri Seppä, Olli Pietiläinen, Heidi Ryynänen, Sanna Heikkinen, Johan G. Eriksson, Tommi Härkänen, Pekka Jousilahti, Paul Knekt, Seppo Koskinen, Maarit A. Laaksonen, Satu Männistö, Teemu Roos, Ossi Rahkonen, Nea Malila, Janne Pitkäniemi, The Metca Study GroupCancer 2022 · DOI

TL;DR

This study investigated how combinations of four key lifestyle factors (smoking, alcohol consumption, physical inactivity, and BMI) influence colorectal cancer (CRC) risk. It found that while no statistically significant interactions were detected after multiple testing corrections, certain pairwise combinations of these factors, particularly in men, were associated with a significantly elevated risk of CRC compared to individuals with only one or none of these risk factors.

What Problem Does This Paper Address?

Colorectal cancer (CRC) is a major global health concern, and while several lifestyle factors are known to increase its risk, most previous research has examined these factors in isolation or assumed they act independently. This approach overlooks the potential for combined effects and interactions, where the risk associated with two factors together might be greater or lesser than the sum of their individual risks. Understanding these synergistic or antagonistic relationships is crucial for developing more effective prevention strategies. This paper aims to fill this gap by systematically examining the pairwise associations and interactions of four prominent lifestyle factors with CRC risk in a large, prospective Finnish cohort.

Methods

The researchers utilized a pooled dataset from seven Finnish health behavior surveys conducted between 1972 and 2015, creating the METCA consortium. This large cohort comprised over 171,000 individuals. Key lifestyle factors – smoking, alcohol consumption, physical inactivity, and body mass index (BMI) – were harmonized and categorized into exposed/unexposed groups. CRC incidence was then modeled using Poisson regression, incorporating both main effects and pairwise interaction terms for the lifestyle factors. The models were fitted separately for men and women, and hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated. Statistical significance of interactions was assessed using likelihood ratio tests, with p-values adjusted for multiple comparisons using the Benjamini-Hochberg method.

Datasets

The study used data pooled from seven Finnish population-based health behavior surveys:

  • The National FINRISK Study (1972-2015)
  • The Adult Health, Wellbeing, and Services Studies (ATH1: 2010-2011, ATH2: 2012-2015)
  • The Health 2000 Survey (H2000)
  • The Finnish Mobile Clinic Health Survey (FMCF: 1972-1977)
  • The Mini-Finland Health Survey (MFH: 1978-1980)
  • The Helsinki Health Study (HHS: 2000-2002)
  • The Helsinki Birth Cohort Study (HBCS)

Cancer diagnoses were obtained through linkage with the Finnish Cancer Registry. No new datasets were introduced.

Key Insights and Ideas

  • Synergy in Risk: While formal statistical interactions weren't significant after correction, the data strongly suggests that combining risk factors amplifies CRC risk beyond their individual contributions. For instance, male smokers with overweight had a notably higher risk than expected from each factor alone.
  • Gender Differences: The study highlights significant differences in how lifestyle factors impact CRC risk between men and women. Men generally showed stronger associations, with combinations of adverse factors consistently increasing their risk.
  • Smoking's Potency in Men: Smoking appears to be a particularly potent risk factor for CRC in men, especially when combined with overweight. The lack of a significant association in women might be due to shorter exposure durations in the cohort.
  • Alcohol and Overweight Interaction Nuance: Interestingly, women showed a negative interaction between alcohol use and overweight, suggesting their combined effect was less than expected, though this also didn't reach statistical significance after corrections. This contrasts with the general trend of amplified risk with combined factors.
  • Protective Potential: The findings reinforce that adopting healthy lifestyle habits – specifically avoiding smoking, maintaining a healthy BMI (<25 kg/m²), and abstaining from alcohol – can substantially reduce CRC risk.

Key Results

  • The highest pairwise CRC risk was observed in male smokers who were overweight (HR 1.75, 95% CI 1.36–2.26).
  • In women, the highest pairwise risk was found in those who were overweight and consumed alcohol (HR 1.45, 95% CI 1.14-1.85).
  • Overall, men exhibited stronger associations between lifestyle factors and CRC risk than women.
  • For men, being overweight or a smoker, when combined with any other adverse lifestyle factor, increased CRC risk.
  • For women, elevated CRC risks were seen in those who were physically inactive and consumed alcohol or were overweight.
  • After adjusting for multiple comparisons, no statistically significant interactions were detected between any pairs of lifestyle factors.

Limitations and Future Work

  • Self-Reported Data: A primary limitation is that exposure data (smoking, alcohol, weight) were largely self-reported and collected at a single time point, which can introduce recall bias and affect accuracy, especially for factors like alcohol consumption and weight.
  • Dichotomization: Categorizing lifestyle factors into binary (exposed/unexposed) groups might underestimate the true HRs for CRC, as it simplifies complex behaviors.
  • Varied Survey Phrasing: Harmonizing data from different surveys over several decades presented challenges due to variations in question wording, potentially affecting exposure prevalence and thus the results.
  • Physical Activity Measure: The study only included leisure-time physical activity, omitting work-related activity, which could influence overall physical activity levels and CRC risk.
  • Future Work: The authors suggest that future research could explore the effects of combinations of more than two lifestyle factors, as well as investigate the role of diet, which was not included in this analysis. Further research into the specific mechanisms underlying the observed gender differences in lifestyle factor associations would also be valuable.

Summary updated

Abstract

This study strengthens the evidence of overweight, smoking, and alcohol consumption as CRC risk factors. Substantial protective benefits in CRC risk can be achieved by preventing smoking, maintaining BMI to <25 kg/m<sup>2</sup> and not consuming alcohol.
BibTeX
@inproceedings{roos2022pairwise,
  author = {E Roos and Karri Seppä and Olli Pietiläinen and Heidi Ryynänen and Sanna Heikkinen and Johan G. Eriksson and Tommi Härkänen and Pekka Jousilahti and Paul Knekt and Seppo Koskinen and Maarit A. Laaksonen and Satu Männistö and Teemu Roos and Ossi Rahkonen and Nea Malila and Janne Pitkäniemi and The METCA Study Group},
  title = {{Pairwise association of key lifestyle factors and risk of colorectal cancer: a prospective pooled multicohort study}},
  booktitle = {Cancer},
  year = {2022},
  doi = {10.1002/cnr2.1612},
}
Figure