Dynamic phenotypes of preclinical and clinical obesity in relation to new‐onset cancer risk: A longitudinal analysis from the UK biobank

Abstract Aim The definition of clinical obesity was newly announced. Our study aims to investigate the relationship between different states of obesity and dysfunctions due to obesity with cancer incidence and mortality. Methods The prospective cohort study from the UK Biobank included 220 016 participants. Anthropometric parameters, in combination with obesity‐induced dysfunctions, were used to diagnose clinical obesity. Six clusters were categorized according to individual's baseline and follow‐up dysfunction status. Hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) for cancer incidence risk were estimated using the landmark analysis. Results After a mean follow‐up period of 11.0 years, a total of 24 066 cancer incidence was observed. Using Cluster 1 (participants without obesity and dysfunction at baseline and during follow‐up) as the reference group, Cluster 5 (preclinical obesity with follow‐up dysfunctions; HR = 3.17, 95% CI: 3.05–3.29) exhibited the highest multivariable‐adjusted cancer incidence risk, while Cluster 4 (preclinical obesity without follow‐up dysfunctions; HR = 0.88, 95% CI: 0.85–0.92) showed the lowest. Additionally, the fully adjusted HRs for cancer mortality showed the highest in Cluster 6 (clinical obesity; HR = 1.82, 95% CI: 1.65–2.00), compared with Cluster 1. Site‐specific analyses showed consistently higher cancer risks in Cluster 5 and 6 across various types of cancer, notably the incidence of pancreatic cancer and the mortality of prostate or bladder cancer. Conclusion Obesity‐induced dysfunction was significantly associated with cancer risk. For future clinical practice, the early identification and intervention of clinical obesity and obesity‐induced dysfunctions are of critical importance for reducing cancer risks.