Background Rheumatoid arthritis (RA) is characterised by progressive joint destruction and loss of periarticular bone mass. Hand bone loss (HBL) is measured by Digital X-ray Radiogrammetry (DXR) which has been proposed as an outcome measure for treatment effect in RA. A definition of increased HBL adjusted for age- and gender-related bone loss is lacking. Furthermore, it is unknown to which extent HBL is normalised in RA patients during treatment with tumour necrosis factor alpha inhibitors (TNF-I).
Objectives To establish a reference material for HBL and to investigate whether HBL normalises in RA patients during TNF-I treatment in clinical practice.
Methods Hand bone mass (DXR-BMD) was measured with DXR, a computerised method of estimating cortical bone mineral density in the diaphysis of the 2nd – 4th metacarpal bone in a reference population and a patient cohort. The reference population consisted of 1,533 men and 2618 women randomly selected from the urban county of Østerbro in Denmark who had hand x-rays performed in the cross-sectional Copenhagen Osteoarthritis Study. Linear regression analyses were used to calculate normal HBL (defined as the 95% Confidence Interval (95%CI) for the age-related mean changes in DXR-BMD between subsequent age-groups).
The patient cohort was 135 patients from the DANBIO registry with hand x-rays obtained ∼2 years before start of TNF-I (pre-baseline, all patients treated with conventional synthethic Disease-Modifying Anti-Rheumatic Drugs (csDMARD)), at start of TNF-I (baseline) and ∼ 2 years after start of TNF-I (follow-up). Annual HBL during csDMARD and TNF-I treatment were calculated in individual patients and compared with the lower 95%CI of mean DXR-BMD change in the gender- and age-matched reference group to assess if increased HBL was present.
Results Table 1 presents the HBL reference material. The 135 RA patients (85% women, 71% IgM-RF positive, median age 55 (range 23-84) years; median disease duration 5 (range 1-53) years) had a pre-baseline median DAS28 of 4.3 (range 1.6-6.9) and a baseline DAS28 of 5.3 (1.4-8.2). TNF-I treatment was infliximab (74%), etanercept (13%) or adalimumab (13%). At follow-up (DAS28 3.1 (1.4-7.7) 59% received the initial TNF-I, 27% had switched to another biological drug and 14% had withdrawn.
Compared to the reference population, 84 (62%) patients had increased HBL during csDMARD treatment and 60 (44%) had during TNF-I treatment (p=0.10,Chi-Sq). In 42 patients who had elevated HBL during csDMARD treatment HBL was normalised during TNF-I. Eighteen patients had normal HBL during csDMARD treatment but increased HBL during treatment with TNF-I.
Conclusions We have established a reference material for HBL in the general population and found significant age-related decreases in DXR-BMD in both men and women. Increased HBL was present in the majority of RA patients initiating TNF-I treatment in clinical practice and was normalised in only a minority of patients during treatment.
Disclosure of Interest None declared