• osteocalcin
• ankylosing spondylitis
• disease activity
• bone metabolism

In rheumatic diseases the synovial concentration of osteocalcin, which represents osteoblast activity, is inversely correlated with the extent of joint inflammation.¹ Synovial and serum osteocalcin correlate positively.¹ In ankylosing spondylitis (AS) the serum concentration of osteocalcin has been reported to be low^2,3 or normal.^4–7 Cross sectional studies have shown no significant correlation between osteocalcin serum concentration and erythrocyte sedimentation rate (ESR) or C reactive protein.^5,7

To answer the question whether serum osteocalcin is a useful marker of disease activity in AS, longitudinal studies may be more sensitive and specific. For this reason changes in serum osteocalcin were correlated with changes in ESR, which is probably still the best marker of inflammation in AS.⁸

In 89 patients with ankylosing spondylitis (modified New York criteria; 75 male, 14 female; age 43 (11) years; disease duration 14 (9) years) venous blood was taken at the start and the end of a three week rehabilitation course consisting of physical exercise, physiotherapy, massage therapy, electrotherapy, underwater exercises, and radon treatment as prescribed by the patient's doctor. Patients were advised not to change their drug treatment. The ESR was determined according to Westergren, the result at one hour being used for calculation. Serum was frozen at –18°C until further analysis. Osteocalcin was measured in one batch with a commercially available test kit (IRMA, Biocis, Vienna; normal range according to the manufacturer 7.5–31.5 ng/ml in men, 3.7–31.7 ng/ml in women). Results are given as median (25th, 75th centile). The Mann-Whitney rank sum test and Spearman rank order correlation test were used to test significance.

Values at the first measurement were ESR 18 (8, 28) mm/1st h, serum osteocalcin 25 (20.5, 32.8) ng/ml. The osteocalcin serum concentration was within the normal range in 66 of the 89 patients, and 23 patients had increased serum concentrations. Values at the end of treatment were ESR 16 (8, 26.5) mm/1st h, osteocalcin 26.1 (18.9, 32.7) ng/ml (no significant changes). The ESR and osteocalcin at the first examination did not correlate significantly (r_s=0.07; p=0.5). The changes in ESR (1 (−4, 6) mm/1st h) and changes in osteocalcin (–0.5 (−2.6, 5.7) ng/ml) showed a significant correlation (r_s=0.28; p<0.01).

The results confirm previous findings showing no significant correlation between serum osteocalcin and ESR in cross sectional studies. Changes in osteocalcin after three weeks, however, correlated significantly with changes in ESR, but in view of the weak correlation (r_s=0.28) the clinical relevance of serum osteocalcin determination for assessing disease activity seems limited.