Abstract
Background The incidence of early-onset cancers (EOCs), defined as cancers diagnosed before 50 years of age, has risen globally over the past three decades. The association between modifiable lifestyle factors and EOC risk has been poorly investigated. Methods Participants enrolled before age 50 in EPIC (39,459 men; 110,991 women) and the UK Biobank (52,188 men; 62,528 women) were analysed. We examined the association between smoking, alcohol consumption, body mass index (BMI), physical activity and diet, and risk of lifestyle-related EOCs - including cancer of the breast, lung, colorectum, stomach, liver, cervix, oesophagus, bladder, and others. Each risk factor was scored from 0 (most unfavourable) to 4 (most favourable). Cohort-specific hazard ratios (HR) with 95% confidence intervals (CI) were estimated in Cox regression models and pooled via a meta-analytic approach. Findings We observed 361 lifestyle-related EOCs in men and 2,618 in women, including 1,765 breast cancers. In men, one-point increase in the smoking score, indicating less smoking, was inversely associated with risk of lifestyle-related EOC (HR 0.82, 95% CI: 0.76-0.88). In women, each one-point increase in smoking (less smoking), BMI (lower BMI), and physical activity (more active) scores was inversely associated with risk of lifestyle-related EOC excluding breast cancer (HRs: 0.89 [0.84-0.93], 0.93 [0.89-0.98], 0.95 [0.90-1.00], respectively). For breast cancer, only physical activity showed a statistically significant association (HR 0.95 [0.91-0.995]). Alcohol and diet showed no association. Interpretation These findings support the importance of adopting a healthy lifestyle early in life to reduce cancer risk before the age of 50.</p>