Preclinical Obesity and Risks of Incident Diabetes, Cardiovascular Disease, and Mortality in Seven International Ageing Cohorts

This preregistered secondary analysis will use harmonised individual participant data from seven international longitudinal ageing cohorts to evaluate the associations of preclinical obesity with incident diabetes, cardiovascular disease, and all-cause mortality among middle-aged and older adults. At each participant’s first eligible study visit, individuals will be classified into four mutually exclusive groups: (1) no obesity and no core clinical abnormality; (2) no obesity but at least one core clinical abnormality; (3) preclinical obesity, defined as obesity without any core clinical abnormality; and (4) clinical obesity, defined as obesity with at least one core clinical abnormality. Obesity will primarily be identified using population-specific BMI thresholds, with additional anthropometric criteria based on waist circumference and waist-to-height ratio used where these measures are available. Core clinical abnormalities will include hypertension or antihypertensive medication use, diabetes or glucose-lowering medication use or a strict metabolic abnormality definition where biomarkers are available, cardiovascular disease, and severe limitations in basic activities of daily living or mobility. The primary comparison will be between participants with preclinical obesity and those with neither obesity nor core clinical abnormalities. Three parallel primary outcomes will be examined: incident diabetes, incident cardiovascular disease, and all-cause mortality. Outcome-specific risk sets will be created so that participants with diabetes at baseline will be excluded only from the incident diabetes analysis, and participants with cardiovascular disease at baseline will be excluded only from the incident cardiovascular disease analysis. A one-stage individual participant data approach will be used. Cox proportional hazards models will use attained age as the time scale, adjust for sex, and allow cohort-specific baseline hazard functions through stratification by cohort while estimating a common cross-cohort hazard ratio. A secondary adjusted model will additionally include smoking, alcohol consumption, physical activity, and education. Death will be treated as a competing event in analyses of incident diabetes and cardiovascular disease. Cause-specific hazard ratios will be the primary association measures, with Aalen–Johansen cumulative incidence estimates and Fine–Gray models used in supplementary analyses. Holm correction will be applied to the three primary comparisons to control the family-wise error rate. Prevalence estimates will be reported descriptively by cohort and by prespecified age groups of younger than 65 years, 65–74 years, and 75 years or older. Sensitivity analyses will examine alternative obesity definitions, exclusion of functional limitation criteria, stricter metabolic definitions, alternative cardiovascular outcome definitions, exclusion of early events, competing-risk methods, missing-data assumptions, and the influence of individual cohorts. The study is expected to clarify whether preclinical obesity consistently identifies individuals at increased risk of diabetes, cardiovascular disease, or death across diverse international ageing populations. The results will also provide epidemiological parameters for subsequent studies of disease burden, treatment targeting, and economic evaluation.