Absence of marked seasonal change in bone turnover: a longitudinal and multicenter cross-sectional study

The effect of season on bone turnover is controversial. No information is available on seasonality of new serum markers of bone resorption. In this study, we have been unable to confirm findings of a marked wintertime increase in bone formation and resorption within the general population. Seasonality was assessed by cosinor analysis.

We investigated the effect of season on seven markers of bone turnover in a longitudinal study (six men and six premenopausal women; age, 24-44 years) and a separate large population-based multicenter European study (n = 2780 women, Osteoporosis and Ultrasound Study [OPUS]).

Measurements included serum Crosslaps, procollagen type I N-terminal propeptide (PINP), osteocalcin (OC), and the N-telopeptide fragment of type I collagen in urine (NTX). Seasonality was assessed by cosinor analysis with Hotelling's T2 test.

Serum 25(OH) vitamin D showed a marked seasonal rhythm. There was no significant seasonal component for any marker of bone turnover in the longitudinal analysis (cosinor analysis, p > 0.05). The percentage of within subject variance accounted for by any seasonal trend was very small for all markers (less than 2.5%). Less than 1% of the between-person variance was accounted for by seasonality in the cross-sectional analysis for all markers (n = 2780). There was a small but statistically significant summertime increase in OC and PINP in the healthy postmenopausal population after exclusions based on disease or medication use (remaining n = 1226, amplitudes 5.6% and 5.4%, respectively, p < 0.001).

We have been unable to confirm findings of a marked wintertime increase in bone formation and resorption within the general population. The absence of marked seasonality was irrespective of age, menopausal status, reported supplemental Vitamin D intake, age or geographical location. The small but statistically significant summertime increase in bone formation in this and other studies is unlikely to confound clinical interpretation of these measurements.