Associations Between 40-Year Trajectories of BMI and Proteomic and Epigenetic Aging Clocks: Deciphering Nonlinearity and Interactions.

BMI at age 18, BMI at approximately age 60, and changes in BMI over 40 years were associated with increased biological aging across most aging clock estimates.

• Analyses were conducted in 401 Finnish twins with up to nine self-reported or measured BMI values collected over 40 years.
• Olink proteomic and Illumina DNA methylation data were generated from blood drawn at the last BMI measurement.
• Four proteomic and five epigenetic age estimates were derived from these data.
• BMI trajectories were modeled using mixed-effects models, and generalized additive models were applied to examine associations with biological aging, adjusting for chronological age.

Statistical evidence of nonlinearity was found in approximately one-third of the significant associations between BMI trajectories and biological aging.

• Nonlinear associations were identified in about one-third of the significant associations.
• Nonlinearity was mostly observed for proteomic clocks rather than epigenetic clocks.
• The authors conclude that 'assuming linearity in associations between BMI trajectories and biological aging is a critical oversight.'
• Generalized additive models were used specifically to allow detection of nonlinear relationships.

Suggestive evidence was found for interactions between BMI at age 18 and BMI at approximately age 60 in explaining variability in two proteomic clocks.

• Interactions between baseline BMI (age 18) and BMI at ~60 years were identified for two proteomic clocks.
• The p-values for these interactions were p = 0.07 and p = 0.09, described as 'suggestive evidence.'
• The analysis also examined interactions of baseline BMI with BMI change over time.
• These interaction analyses were conducted using generalized additive models.

Proteomic aging clocks demonstrated particular utility in capturing nonlinear associations with BMI trajectories compared to epigenetic clocks.

• Nonlinear associations were 'mostly observed for proteomic clocks.'
• The study used four proteomic age estimates derived from Olink proteomic data.
• Five epigenetic age estimates were derived from Illumina DNA methylation data for comparison.
• The authors highlight 'the potential of proteomic clocks in obesity research' as a key conclusion.

The study design incorporated longitudinal BMI data spanning 40 years in a Finnish twin cohort to model BMI change trajectories.

• The sample consisted of 401 Finnish twins.
• Up to nine self-reported or measured BMI values per participant were collected over 40 years.
• BMI was assessed at ages 18 and approximately 60 years, as well as at multiple intermediate time points.
• Mixed-effects models were used to model BMI change over time before applying generalized additive models for association analyses.