Introduction There are plausible biological mechanisms for how increased physical activity (PA) may prevent pancreatic cancer, although most studies do not report an inverse association. We investigated whether this may be related to the age at which PA is measured, using a validated questionnaire, and whether the effect of PA is independent of body mass index (BMI).
Methods 23,639 participants, aged 40–74 years were recruited into the EPIC-Norfolk cohort study between 1993 and 1997. These participants completed validated questionnaires on both occupational and leisure time PA. From this, four levels of PA index were derived. The cohort was monitored for up to 17 years to identify those participants who developed pancreatic cancer. The hazard ratios (HRs) of developing cancer were estimated using Cox regression and adjusted for covariates (age, gender, cigarette smoking status and type 2 diabetes). Each analysis was first performed in those recruited of all ages and then in those younger and older than 60 years at recruitment.
Results Within 17 years, 88 participants developed pancreatic cancer (55% female, median age of diagnosis 73 years, range 52–89 years). There was no association between PA and risk of pancreatic cancer in the whole cohort (trend HR=1.03, 95% CI: 0.84–1.27). However, in those recruited at younger than 60 years (n = 29 cases), higher levels of PA were associated with a decreased risk (highest vs. lowest category HR=0.27, 95% CI: 0.07–0.99, trend HR=0.75, 95% CI: 0.53–1.06, p = 0.11). When BMI was included, the associations were similar (highest vs. lowest category HR=0.25, 95% CI: 0.07–0.93, trend HR=0.73, 95% CI: 0.51–1.03, p = 0.08). In participants aged greater than 60 years (n = 59 cases), higher PA was associated with a non significant, increased risk both when BMI was unaccounted for (highest vs. lowest category HR=1.98, 95% CI: 0.94–4.16, p = 0.07, trend HR=1.23, 95% CI: 0.96–1.57, p = 0.10) and when BMI was included (trend HR=1.21, 95% CI: 0.94–1.55, p = 0.13).
Conclusion The association between PA and cancer risk is dependent on the age at which PA is measured. This possibly reflects occupational activity and differences in general medical health with age or residual confounding. The associations were similar when adjusted for BMI, suggesting an independent mechanism of PA. If the inverse association of increased PA in younger participants is causal, one in six cases of pancreatic cancer might be prevented by encouraging more PA. Aetiological studies should measure PA at different ages when investigating pancreatic cancer.
Disclosure of Interest None Declared.