Long-Term Leisure-Time Physical Activity and Serum Metabolome

Urho M. Kujala; Ville-Petteri Mäkinen; Ilkka Heinonen; Pasi Soininen; Antti J. Kangas; Tuija H. Leskinen; Paavo Rahkila; Peter Würtz; Vuokko Kovanen; Sulin Cheng; Sarianna Sipilä; Mirja Hirvensalo; Risto Telama; Tuija Tammelin; Markku J. Savolainen; Anneli Pouta; Paul F. 'Reilly9;
Pekka Mäntyselkä; Jorma Viikari; Mika Kähönen; Terho Lehtimäk; Paul Elliott; Mauno J. Vanhala; Olli T. Raitakari; Marjo-Riitta ärvelin ;Jaakko Kaprio; Heikki Kainulainen; Mika Ala-Korpela

CIRCULATIONAHA.112.105551 Published online before print December 20, 2012, doi: 10.1161/​

Abstract

Background—Long-term physical inactivity seems to cause many health problems. We studied whether persistent physical activity compared to inactivity has a global effect on serum metabolome towards reduced cardio-metabolic disease risk.

Methods and Results—Sixteen same-sex twin pairs (mean age 60 yrs) were selected from a cohort of twin pairs on the basis of their over 30-year discordance for physical activity. Persistently (≥5 years) active and inactive groups in three population-based cohorts (mean ages 31 to 52 years) were also studied (1037 age- and sex-matched pairs). Serum metabolome was quantified by nuclear magnetic resonance spectroscopy. We used permutation analysis to estimate significance of the multivariate effect combined across all metabolic measures; univariate effects were estimated by paired testing in twins as well as in matched pairs, in the cohorts, and by meta-analysis over all sub-studies. Persistent physical activity was associated with the multivariate metabolic profile in the twins (P=0.003), and a similar pattern was observed in all three population cohorts with differing mean ages. Isoleucine, α1-acid glycoprotein and glucose were lower in the physically active than in the inactive individuals (P<0.001 in meta-analysis), serum fatty acid composition was shifted towards a less saturated profile, and lipoprotein subclasses towards lower VLDL (P<0.001) and higher large and very large HDL (P<0.001) particle concentrations. The findings persisted after adjustment for body-mass index.

Conclusions—The numerous differences found between persistently physically active and inactive individuals in the circulating metabolome coherently indicate better metabolic health in the physically active than in inactive individuals.