The MetaboHealth score is 40% heritable and is influenced by frailty status, BMI, and smoking in Swedish twins

Abstract The MetaboHealth score, a metabolomics-based biomarker trained on mortality, is associated with ageing-related phenotypes and morbidity, but its heritable component is still uncertain. Understanding genetic versus environmental contributions is essential for its utility as an intervention-responsive biomarker. Hence, the aim of this study was to estimate heritability of the MetaboHealth score in the large Swedish TwinGene cohort and evaluate interactions with modifiable lifestyle factors and health outcomes. We analysed 12,256 twins (55% female; mean age 64.8 years at blood draw) with Nightingale Health-based metabolomics data, followed until February 2024 (median follow-up time of 16.2 years). We subsequently estimated the heritability of the score using additive genetic (A), shared environmental (C), and unique environmental (E) (ACE) modelling via Cholesky decomposition in 1,370 monozygotic and 3,381 dizygotic twin pairs. Cox regression was used to assess mortality associations, with sensitivity analyses stratified by age, smoking, body mass index (BMI), and frailty. The MetaboHealth score was significantly associated with all-cause mortality (Hazard ratio (HR) = 1.91, 95% confidence interval (CI): 1.73–2.11; P = 1.32 × 10⁻ 38 ) in Swedish twins and its heritable component was estimated to be 40% (95% CI: 37–43%), with 57.5% of the variance attributable to unique environment and 2.5% to shared environment. Lifestyle factors and health outcomes, including BMI (P = 8.36 × 10⁻⁶), smoking (P = 0.009), and frailty (P = 6.55 × 10⁻ 4 ), were associated with the score. Notably, mortality associations were strongest in younger adults (HR = 2.76 in individuals ≤ 65 years of age versus HR = 1.77 in individuals > 65 years of age) and smokers (HR = 2.16 in smokers versus HR = 1.59 in non-smokers). Moreover, analyses of monozygotic twins showed reduced score correlations in discordant pairs, where one twin was non-frail and the other pre-frail or frail (r 2 concordant = 0.184 versus r 2 discordant = 0.119; P = 0.042). In conclusion, with only 40% of its variance fixed by genetics, the MetaboHealth score is predominantly environmentally modifiable. Its heightened sensitivity to mortality risk in younger adults and smokers, and association with frailty status and BMI, supports its use as a biomarker for precision interventions in healthy ageing. Graphical Abstract