Body mass index across adulthood, weight gain and cancer risk: a population-based cohort study

Abstract

Background: Although the association between excess weight and cancer risk is well established, it is not known how this association evolves across the lifespan. We aimed to investigate the strength of the association of excess weight at different ages in adulthood and adult weight gain with cancer risk.

Methods: We used data from a German population-based cohort study of 9,218 participants aged 50-75 (mean 62) years recruited between 2000 and 2002. Participants provided socio-demographic, medical, and lifestyle data, including self-reported current height and weight (at ages 20, 30, 40, 50 and baseline). Main exposures were body mass index (BMI, kg/m²) at different ages and weight change (kg) since age 20. The outcome was obesity-related cancer (13 types). Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using multivariable Cox models.

Results: During a median follow-up of 17.1 years, 852 diagnoses of obesity-related cancers were recorded. Overweight and obesity in early and middle adulthood showed no significant associations with obesity-related cancer risk, whereas significant positive associations were observed for overweight and obesity at age 50 years and older. For weight change since age 20, strong associations were found, with HRs (95% CI) of 1.42 (1.11-1.81), 1.57 (1.24-1.99) and 1.96 (1.56-2.47) for the 2nd, 3rd, and 4th quartile compared to the lowest quartile, respectively. After mutual adjustment for adult weight gain and BMI at baseline, the estimates for weight gain persisted, while those for BMI at baseline disappeared. The main limitation of the study is that the weights were self-reported.

Conclusions: Our findings suggest that excess weight may have a varying effect on cancer risk through life with its impact potentially being more pronounced in later adulthood, and that adulthood weight gain might be a better indicator of obesity-related cancer risk than BMI measured at a single point in time.

Keywords: Breast cancer; Cancer; Colorectal cancer; Excess weight; Obesity; Overweight; Weight gain.

Fig. 1

Flow chart showing the selection of the study population

Fig. 2

Association of BMI at different ages with obesity-related cancer risk. Restricted cubic spline regression model– knots placed at 5th, 25th, 50th, 75th, and 95th percentile, and median used as the reference. Adjusted for age, sex, education, previous lower endoscopy, physical activity, alcohol consumption, smoking (pack-years), 1st-degree family history of cancer, red meat intake, fruit intake, vegetable intake, current nonsteroidal anti-inflammatory drugs use, history of mammography (women) and ever use of hormone replacement therapy (women), parity (women) and age at menarche (women)