Proteomic associations with fluctuation and long-term changes in BMI: A 40-year follow-up study

Introduction: While some studies have explored associations between weight change and blood proteins, most have been intervention-based, offering limited insight into proteomic associations with long-term weight gain. It remains unclear whether plasma proteins are related to BMI fluctuation over time. This study investigates associations of long-term BMI changes and fluctuations with over 1000 plasma proteins involved in cardiometabolic and inflammation functions. Data and Methods: The study included 304 Finnish adult twins (117 men) born before 1958 from the Older Finnish Twin Cohort, with BMI data spanning five time points (1975, 1981, 1990, 2011 and 2012–2014). Proteomic data were derived from blood samples collected at the last BMI measurement. Linear mixed-effects models analysed individual BMI trajectories, producing intercepts (baseline BMI) and slopes (BMI change rates). BMI fluctuation was calculated as the average squared deviation from expected BMI across time points. Associations between BMI changes/fluctuation and (i) 1231 plasma proteins related to cardiometabolic and inflammatory functions and (ii) polygenic risk scores for BMI (PRS_BMI), as well as interaction effects between PRS_BMI and baseline BMI on protein-BMI relationships were studied. Within-pair analyses using monozygotic twins were conducted to account for shared confounding factors. Results: A total of 135 proteins were associated with changes in BMI over 40 years, while 17 proteins were linked to fluctuation in BMI: 12 associations (10 with BMI changes and 2 with fluctuation) remained significant in within-twin pair analyses. PRS_BMI was associated with BMI changes but not with fluctuation. PRSBMI-protein interactions explaining BMI changes or fluctuation were found, though a single interaction between the antigen CD72 protein and baseline BMI was observed. Conclusion: This study highlights significant associations between plasma proteins and long-term BMI changes and fluctuations, with no evidence of PRS_BMI-protein interactions influencing BMI trends. These findings underscore the substantial role of environmental factors in shaping proteome-BMI associations over adulthood.