Open Access Research article

Irisin and exercise training in humans – Results from a randomized controlled training trial

Anne Hecksteden1*, Melissa Wegmann1, Anke Steffen1, Jochen Kraushaar1, Arne Morsch2, Sandra Ruppenthal3, Lars Kaestner3 and Tim Meyer1

Author Affiliations

1 Institute of Sports and Preventive Medicine, Saarland University, 66123 Saarbrücken, Germany

2 Deutsche Hochschule für Prävention und Gesundheitsmanagement, 66123 Saarbrücken, Germany

3 Institute for Molecular Cell Biology, Saarland University, 66421 Homburg, Germany

For all author emails, please log on.

BMC Medicine 2013, 11:235  doi:10.1186/1741-7015-11-235

Published: 5 November 2013



The recent discovery of a new myokine (irisin) potentially involved in health-related training effects has gained great attention, but evidence for a training-induced increase in irisin remains preliminary. Therefore, the present study aimed to determine whether irisin concentration is increased after regular exercise training in humans.


In a randomized controlled design, two guideline conforming training interventions were studied. Inclusion criteria were age 30 to 60 years, <1 hour/week regular activity, non-smoker, and absence of major diseases. 102 participants could be included in the analysis. Subjects in the training groups exercised 3 times per week for 26 weeks. The minimum compliance was defined at 70%. Aerobic endurance training (AET) consisted of 45 minutes of walking/running at 60% heart rate reserve. Strength endurance training (SET) consisted of 8 machine-based exercises (2 sets of 15 repetitions with 100% of the 20 repetition maximum). Serum irisin concentrations in frozen serum samples were determined in a single blinded measurement immediately after the end of the training study. Physical performance provided positive control for the overall efficacy of training. Differences between groups were tested for significance using analysis of variance. For post hoc comparisons with the control group, Dunnett’s test was used.


Maximum performance increased significantly in the training groups compared with controls (controls: ±0.0 ± 0.7 km/h; AET: 1.1 ± 0.6 km/h, P < 0.01; SET: +0.5 ± 0.7 km/h, P = 0.01). Changes in irisin did not differ between groups (controls: 101 ± 81 ng/ml; AET: 44 ± 93 ng/ml; SET: 60 ± 92 ng/ml; in both cases: P = 0.99 (one-tailed testing), 1−β error probability = 0.7). The general upward trend was mainly accounted for by a negative association of irisin concentration with the storage duration of frozen serum samples (P < 0.01, β = −0.33). After arithmetically eliminating this confounder, the differences between groups remained non-significant.


A training-induced increase in circulating irisin could not be confirmed, calling into question its proposed involvement in health-related training effects. Because frozen samples are prone to irisin degradation over time, positive results from uncontrolled trials might exclusively reflect the longer storage of samples from initial tests.

Trial registration Identifier: NCT01263522.

Myokine; Browning; FNDC5