Radiographic damage and its progression in early rheumatoid arthritis (RA) can be predicted by markers and regulators of bone metabolism. Studies of bone formation in RA patients measuring osteocalcin or the N-terminal telopeptide of type I procollagen (P1NP) have produced varying results, whereas measurements of bone resorption using serum or urine C-terminal crosslink of type I collagen (β-CTX) mostly show increased values.1,–,3 RA patients often have elevated serum levels of the osteoclast-activating cytokine receptor activator of nuclear factor κB ligand (RANKL), as well as of osteoprotegerin, which prevents osteoclast activation.4

We previously reported increased levels of autoantibodies and acute phase reactants in blood donors years before the onset of RA symptoms.5 6 We now investigated whether this subtle increase of inflammation has an influence on preclinical bone metabolism and on later occurring radiographic damage.

From 79 patients who had been blood donors before the onset of RA,7 three serum samples (18 patients two samples, 14 patients one sample) were selected at 1, 2 and 5 years before the onset of symptoms, together with a control sample matched for gender, age and time of blood donation to ensure identical storage conditions. Differences in osteocalcin, P1NP, β-CTX, osteoprotegerin and RANKL over time in preclinical RA patients were compared with controls using random coefficient analysis.8 The models were corrected for age, gender, time of donation, autoantibodies (IgM rheumatoid factor and antibodies against cyclic citrullinated peptides) and markers of inflammation (C-reactive protein and secretory phospholipase A₂).5,–,7 Variables differing from controls were associated with the radiographic Sharp/van der Heijde score (SHS) at baseline (median SHS 1, IQR 0–7) as well as with progression per year after a mean follow-up of 6 years (median progression 3, IQR 1–10).

There were statistically significantly increased mean levels of only P1NP and osteoprotegerin in the group of preclinical RA patients compared with the control group (an increase of 5.0 ng/ml and 4.1 pmol/l for P1NP and osteoprotegerin, respectively; table 1). After correction for age, gender and autoantibodies, preclinical levels of P1NP and osteoprotegerin were negatively associated with SHS, but these differences were not statistically significant.

The elevated P1NP suggests that significant differences in biochemical bone formation activity can be detected above a certain level of ‘disease activity’ in the presymptomatic phase of RA, whereas the elevated levels of osteoprotegerin indicate an increased osteoclast activity in the preclinical phase. The trend for a negative association between preclinical levels of P1NP and osteoprotegerin and later radiographic damage is not readily explainable. Possibly the elevation of osteoprotegerin is an, albeit failing, compensation mechanism for a slight and not measurable increase in bone degradation.

In conclusion, there appears to be an alteration of bone metabolism parallel to inflammation and autoimmunity in the asymptomatic preclinical phase of RA, which may reflect the beginning of joint destruction.