Reallocating sedentary time to physical activity for chronic disease prevention and healthy longevity : an integrative framework using wearable device data and compositional isotemporal substitution modelling

Abstract

Background: Sedentary behaviour is a major modifiable risk factor for chronic diseases and premature mortality, yet optimal strategies for reallocating sedentary time (ST) to physical activity (PA) remain unclear. The health implications of substituting ST with different PA intensities across genetic susceptibility or biological ageing status are poorly understood. This thesis investigated associations of ST reallocation to different intensities of PA with healthy longevity and chronic diseases using an integrative multi-method approach incorporating wearables, genetics, and compositional data analysis. Methods: This thesis comprises three studies using UK Biobank accelerometry data. Machine learning algorithms classified 24-hour movement patterns into ST, light PA, moderate-to-vigorous PA (MVPA), and sleep. Polygenic risk scores (PRS) were computed from known genetic traits. Compositional isotemporal substitution models employing Cox regression or logistic regression were adjusted for sociodemographic, lifestyle, and clinical confounders. Study One (Chapter 2) (n=55,600) examined ST-to-PA reallocation associations with PhenoAge acceleration and healthspan termination (composite of mortality and seven chronic conditions). Study Two (Chapter 3) (73,272 white-British participants) incorporated PRS for T2D to investigate the genetic modification of ST-to-PA substitution’s association on incident type 2 diabetes (T2D). Study Three (Chapter 4) (15,769 individuals with prediabetes or diabetes) employed multi-ancestry PRS indicative of cardiovascular risk to explore ST-to-PA replacement associations with major adverse cardiovascular events (MACE) across genetic risk strata, whilst using varying levels of baseline ST (5-13 hours/day) as compositional references. Results: Across studies, substituting ST with equivalent PA was linked to more favourable health outcomes, with stronger associations for higher intensity or longer duration of the substitutions. In Study One, reallocating 60 minutes/day of ST to light PA and MVPA was associated with odds ratios (OR) of 0.93 (95% confidence interval [CI]: 0.92-0.94) and 0.79 (95% CI: 0.77-0.81) for accelerated biological ageing, and 2% and 6% lower healthspan termination hazards, respectively. Study Two involving 1,399 incident T2D cases found replacing 30 minutes/day of ST with equivalent MVPA to be associated with 18% lower odds of T2D (OR=0.82, 95% CI: 0.79-0.85); for comparison, ST-to-light PA substitution showed 5% lower T2D odds (OR=0.95, 95% CI: 0.94-0.97). These associations were consistent across genetic risk strata, with relatively greater benefits among individuals at high genetic risk. Among participants with prediabetes or diabetes in Study Three (1,374 MACE cases), reallocating 30 minutes/day of ST to light PA and MVPA was associated with 3% (hazard ratio [HR]=0.97, 95% CI: 0.96-0.99) and 20% (HR=0.80, 95% CI: 0.75-0.85) lower MACE risk, respectively. Analyses using varying levels of baseline ST as compositional references revealed progressively stronger MVPA associations, from HR=0.91 (95% CI: 0.88-0.93) at 5 hours/day of ST to HR=0.66 (95% CI: 0.58-0.75) at 13 hours/day of ST as the compositional reference. Conclusions: Substituting ST with light PA or MVPA is associated with decelerated ageing, extended healthspan, and reduced chronic disease risk. Higher intensity or longer duration of ST-to-PA reallocations yields more pronounced benefits. The findings support precision public health strategies prioritising movement behaviour modifications, with MVPA offering greater health benefits whilst light PA provides an accessible alternative for health promotion.