Reevaluating the impact of physical inactivity on cancer risk: Methodological limitations and considerations

We read with interest the article by Lynch et al.¹ estimating cancer incidence attributable to physical inactivity in the United States. Although we appreciate this important contribution to understanding the cancer prevention potential of physical activity, several methodological limitations warrant consideration.

The study’s reliance on self-reported leisure-time physical activity data from the 2005 National Health Interview Survey likely introduces significant measurement error. Self-reported physical activity is known to substantially overestimate actual activity levels compared to objective measures. Moreover, the exclusion of occupational, transportation, and household physical activity may have led to significant underestimation of total physical activity, particularly among lower socioeconomic groups, where leisure-time activity may be limited but other forms of physical activity are substantial.

The authors’ assumption of a uniform 10-year latency period between physical activity assessment and cancer diagnosis appears oversimplified. Different cancer types likely have varying latency periods, and the biological relationship between physical activity and cancer development may involve cumulative effects over the life course rather than a fixed temporal window.

Statistical concerns also merit attention. The inability to calculate proper confidence intervals for the potential impact fractions (PIFs) and population attributable fractions (PAFs) because of correlated coefficient estimates limits the ability to assess the precision of these estimates. The occurrence of negative PAF values for certain cancers suggests potential model instability that deserves deeper examination.

The treatment of body mass index (BMI) as either a simple confounder or mediator overlooks the complex bidirectional relationship between physical activity and body composition. A more nuanced analytical approach considering potential effect modification and joint effects may be warranted.

The study does not account for key factors that may influence the relationship between physical activity and cancer risk, including the intensity and type of activity, genetic predispositions, changes in physical activity patterns over time, and the role of sedentary behavior independent of overall activity levels. Socioeconomic and racial/ethnic disparities in access to physical activity resources, differential effects across age groups, interactions with dietary patterns, and the influence of sleep duration and quality are also important considerations.

The selection of ≥30 metabolic equivalent task h/week as the optimal activity threshold appears somewhat arbitrary. The dose–response relationship between physical activity and cancer risk likely varies by cancer type and population subgroup, and a single threshold may not adequately capture these nuances.

The study’s estimates of preventable cancers assume perfect adherence to increased physical activity levels, an unrealistic assumption in real-world settings. Without considering the barriers to physical activity adoption, the practical implications for public health planning remain limited.

Additionally, the reliance on 2005 data may not reflect contemporary physical activity patterns, particularly given significant changes in technology use, sedentary behaviors, and the impact of the coronavirus disease 2019 pandemic on lifestyle habits. These factors may influence the current relevance of the findings.

Finally, the study notably lacks discussion of the biological mechanisms linking physical activity to cancer risk. Although BMI is acknowledged as a mediator, other well-established pathways are overlooked, including insulin resistance, systemic inflammation, sex hormone regulation, oxidative stress, gut microbiota alterations, adipokine signaling, epigenetic modifications, cellular senescence, circadian rhythm regulation, and metabolic and immune modulation.²

Understanding these mechanisms is essential to explaining cancer type–specific effects, optimizing physical activity recommendations, and assessing potential effect modifiers such as age, sex, and genetics. Without addressing these pathways, the study remains largely descriptive, and offers limited insight into how physical activity prevents cancer.

Although this study provides valuable insights, these limitations suggest its estimates should be interpreted with caution. Future research should incorporate more comprehensive measures of physical activity, modern data collection methods, and real-world implementation challenges to refine our understanding of physical activity’s true cancer prevention potential.

Kefah Mokbel is a fractional shareholder of Datar Genetics stock and has received honoraria for offering academic and clinical advice to Merit Medical and Q Medical corporations. Furthermore he owns shares in HCA Healthcare UK. The other authors declare no conflict of interest.