Lower serum osteocalcin concentrations are associated with brain microstructural changes and worse cognitive performance

ObjectiveRodent models have found that osteocalcin crosses the blood-brain barrier and regulates behaviour. No data are available on osteocalcin's effects on brain microstructure and cognitive performance in humans. We evaluated the association between serum osteocalcin concentrations and (i) brain microstructural changes on magnetic resonance imaging (MRI) and (ii) neuropsychological performance. Design, Patients and MeasurementsWe studied 24 consecutive obese subjects (13 women; age, 498 81 years; body mass index [BMI], 439 454 kg/m(2)) and 20 healthy volunteers (10 women; age, 488 +/- 95 years; BMI, 243 +/- 354 kg/m(2)) in a cross-sectional study within the multicentre FLORINASH Project. FLAIR signal intensity and DTI-metrics (primary ((1)), secondary ((2)) and tertiary ((3)) eigenvalues; fractional anisotropy (FA); and mean diffusivity) in the caudate, hypothalamus, thalamus and putamen, and in subcortical white matter were assessed. Cognitive performance evaluated by neuropsychological test battery. ResultsLower osteocalcin concentrations were associated with BMI, higher (1,) (2) and (3) values at the caudate and lower FLAIR signal intensity at the caudate and putamen. Obese patients with lower osteocalcin concentrations had higher FA at putamen and thalamus. Lower osteocalcin concentrations were associated with higher Iowa Gambling Task (IGT) scores. FLAIR signal intensity at the caudate <601832 yielded 857% sensitivity, 643% specificity, 706% negative predictive value and 818% positive predictive value for IGT score. Lower osteocalcin was an independent predictor of worse cognitive performance on multivariate analysis (F = 3551, P = 001343; R-2 = 0103). Bayesian information criterion demonstrated that osteocalcin had the predominant role in predicting IGT score. ConclusionsLower serum osteocalcin concentrations are associated with brain microstructural changes and worse cognitive performance.