Time-series metabolomics connects biochemical changes with baseline cardiometabolic risk factors and weight gain in 5,302 individuals across 15-years of follow-up

Ville-Petteri Makinen

South Australian Health and Medical Research Institute, Adelaide, South Australia, Australia

The metabolome tracks cardiometabolic risk, however, population-based time-series data are sparse, which makes it challenging to leverage metabolomics within public health initiatives. We quantified 243 biochemical measures from two visits (1997 and 2012) by the Northern Finland Birth Cohort 1966 (NFBC66; n=3,824; 44% male; age 31±0 years at baseline) and three visits (2001, 2007, 2011) by the Cardiovascular Risk in Young Finns Study (YFS; n=1,478; 45% male; age 32±5 years at baseline). All reported results satisfy P<0.0001. Mean change per decade is expressed in original units or as proportion to baseline concentration. The participants gained +4.4 kg per decade. Glucose (+0.21 mmol/L, +4.3%), triglycerides (+0.24 mmol/L, +20%), LDLC (+0.21 mmol/L, +6.9%), phosphoglycerides (124 umol/L, +5.6%), isoleucine (4.3 umol/L, +7.4%), tyrosine (+4.8 umol/L, +8.7%) and creatinine (+3.1 umol/L, +4.7%) increased. Baseline tertiles were compared for metabolite changes: the top glucose tertile showed a smaller change in glucose (+1.9% vs. +7.0%) and larger changes in VLDL triglycerides (+29% vs. 15%) and isoleucine (+8.5% vs. +5.1%) compared to the bottom tertile. Baseline glucose tertile did not predict insulin change. Baseline insulin did not predict glucose change, but predicted changes in lipids (smaller change in top tertile). The top cholesterol tertile exhibited a smaller change in LDLC compared to the bottom tertile (+1.6% vs. +15%); LDLC in the bottom tertile increased from 3.1 mmol/L to 3.5 mmol/L. The top triglyceride tertile predicted smaller changes in phosphoglycerides (+1.8% vs. +11.3%), sphingomyelins (2.9% vs. 8.0%), alanine (1.9% vs. 5.5%) and phenylalanine (-1.5% vs. +2.6%). The top baseline BMI tertile showed a larger increase in log10 insulin (0.113 vs. 0.017), but a smaller increase in LDLC (+6.5% vs. +12.9%). We observed subtle longitudinal disconnect between insulin, glucose and the metabolome that may indicate progressing insulin resistance and provide new insight into diabetes prevention.


©2018 by Australian Biology of Ageing Conference. Proudly created with Wix.com