Age and Sex Shape Obesity’s Impact on Major Diseases, Large Genetic Study Finds

Key Takeaways: 

• A time-resolved genetic analysis of over 360,000 UK Biobank participants shows that obesity’s health risks vary substantially across age and between men and women.
  
• Higher BMI was causally linked to greater risk of type 2 diabetes, coronary artery disease, atrial fibrillation, and osteoarthritis, but the timing and intensity of these effects differed by condition.
  
• The study’s novel genetic approach revealed that preventive interventions such as statin or blood pressure treatment may temporarily dampen obesity-related cardiovascular risk in midlife.
  
  

Understanding obesity’s changing health risks

Nearly one billion adults globally live with obesity, making it a key driver of type 2 diabetes (T2DM), coronary artery disease (CAD), atrial fibrillation (AF), and osteoarthritis (OA). Yet researchers have long struggled to pinpoint when in life excess body weight does the most harm.

Most studies average risk across all adults, masking crucial age-specific patterns. Body mass index (BMI) remains the standard measure of obesity, but its health impact may shift as metabolism, hormones, behaviour, and medical care evolve through life. Moreover, traditional epidemiological studies cannot always distinguish correlation from causation.

Genetic studies using Mendelian randomisation (MR) can infer causal effects, but conventional MR assumes that risks remain constant over time. In a new paper published in Science Advances, researchers introduced a time-resolved MR framework that tracks how obesity’s effects on major diseases change with age and differ between sexes.

Study design and methods

The researchers analysed data from 361,906 unrelated adults of European ancestry within the UK Biobank, a large population-based health resource. Participants had linked genetic and medical record data, and follow-up continued until a median age of around 70 years, capped at 76 to avoid sparse data at older ages.

BMI at study entry was standardised within sex-by-age groups. The primary outcomes were first occurrences of T2DM, CAD, AF, and OA, identified using International Classification of Diseases (ICD-10) codes.

To establish causal relationships, the team employed MR using polygenic scores (PGS) as instruments. They performed genome-wide association studies (GWAS) for BMI in two independent subsamples (each ~180,953 participants) to identify genome-wide significant genetic variants.

To minimise reverse causation, disease-specific BMI PGS were filtered using the Steiger method, which excluded variants that explained more variation in disease outcomes than in BMI itself. The researchers then modelled time-to-event data using Aalen’s additive hazard model, estimating both cumulative (“life-course”) and age-specific (“momentary”) effects.

Sensitivity analyses accounted for potential biases, including lipid-lowering treatment among CAD-free participants, blood pressure (SBP) as an alternative exposure, and cohort selection effects.

Distinct patterns across diseases

Across adulthood, higher BMI was causally associated with increased rates of all four conditions, but with striking differences in timing and trajectory.

• Osteoarthritis (OA): BMI-related risk rose early in life, becoming significant over 20 years before risk for AF increased. This suggests that musculoskeletal strain and inflammatory pathways linked to obesity manifest relatively early.
  
• Atrial Fibrillation (AF): The risk associated with BMI intensified later in adulthood, suggesting that atrial and metabolic factors accumulate over time.
  
• Type 2 Diabetes (T2DM): The effect of BMI increased steadily from midlife but plateaued between ages 60 and 70, indicating that preventive measures or clinical interventions may mitigate risk during this period.
  
• Coronary Artery Disease (CAD): The most distinctive pattern emerged here – a U-shaped curve. Risk decreased markedly around ages 50 to 70 before rising again in older age. This midlife dip was not explained by study participation patterns but appeared more pronounced among individuals on lipid-lowering medication such as statins, suggesting that treatment may blunt BMI-related cardiovascular risk during this window.
  

When the researchers replaced BMI with systolic blood pressure (SBP) as the exposure, AF risk displayed a similar midlife trough, consistent with the effect of antihypertensive therapy. However, no comparable trough appeared for CAD, reinforcing the role of statins rather than blood pressure control in midlife coronary risk reduction.

Sex differences in risk

Sex-stratified analyses revealed generally stronger BMI effects in men for T2DM, CAD, and AF. Osteoarthritis was an exception: both sexes exhibited similar BMI-related risk until about age 60, after which the association appeared to decline slightly in women, although the results carried uncertainty due to diverging confidence intervals.

A particularly notable finding concerned T2DM. Women displayed a distinct, temporary reduction in BMI-related diabetes risk beginning around age 60 and lasting roughly a decade, whereas men’s risk continued to rise. This “female trough” was not accounted for by menopause timing or the use of hormone therapy, suggesting that behavioural or clinical factors – such as greater engagement with weight management or preventive health care – could play a role.

Genetic and methodological insights

Clustering of BMI-associated genetic variants revealed multiple mechanistic pathways underlying obesity’s effects. Different genetic clusters contributed distinct temporal risk patterns for CAD and T2DM. For instance, “high-risk” clusters largely accounted for the CAD trough and the sex differences seen in T2DM.

Importantly, the researchers verified that the strength of genetic effects on BMI declines with age, underscoring the need for age-sensitive models. Simulation studies confirmed that their time-resolved MR method accurately captured dynamic effects even when the genetic influence on BMI varied over time.

Adjustments for potential selection bias slightly reduced the overall magnitude of effects but preserved key age-related patterns, including the midlife risk reductions.

Clinical implications

The findings emphasise that the timing of prevention matters as much as the magnitude of obesity itself. Sustained high BMI elevates the risk for several major diseases, but the most effective period for intervention differs by condition and by sex.

For example:

• Lipid-lowering treatment in midlife may attenuate BMI-related CAD risk.
  
• Blood pressure control could moderate AF risk later in life.
  
• Women may experience a unique window in their 60s when obesity-related diabetes risk temporarily subsides.
  

The authors conclude that prevention strategies should be tailored to life stage and sex, targeting the periods when intervention can avert the greatest number of disease events.

They also note limitations, including the assumption of an immediate biological response to BMI changes and the reduced precision of genetic instruments for early-life BMI. Nonetheless, their time-resolved MR framework offers a powerful new approach for uncovering dynamic, age-specific health risks that static analyses may obscure.