Background Previous short-term follow-up studies and subanalyses of clinical trials in rheumatoid arthritis (RA) have suggested that low body mass index (BMI) is associated with more radiographic joint progression and high BMI with less. The effect of BMI on progression of erosion score (ES) and joint space narrowing (JSN) could be different in underweight and obese patients.

Objectives To investigate the association between BMI in early RA with radiographic damage during 15 years; to examine if known predictors of a worse radiographic outcome could explain differential radiographic outcome in low and high BMI groups.

Methods Four hundred and seventy-three patients from the BARFOT study included from 1992 to 1999 who performed their 15-year assessment were studied. The patients were assessed at inclusion and after 1, 2, 5, 8 and 15 years. The groups were defined by BMI (kg/m²) at inclusion: BMI ≤20, (n=27), BMI 20<25 (n=210), BMI 25<30 (n=179), and BMI≥30 (n=57). X-rays of hands and feet were scored by the Sharp-van der Heijde scoring method (SHS). Linear mixed models with SHS, ESR, CRP, SJC and TJC as outcome, and BMI at inclusion as predictor was used, adjusted for age, sex, initial treatment, ACPA and smoking.

Results At baseline, total score of SHS, ES and JSN did not differ between BMI groups. There were more women and smokers in BMI≤20 group and older patients in BMI≥30 group. The baseline disease characteristics were similar in the BMI groups.

For the patients with BMI≤20 at inclusion, BMI was associated with a higher predicted SHS progression during follow-up, effect size 5.11 (95% CI 1.72 to 15.15) p=0.005, while for the patients with BMI≥30 at inclusion, BMI was associated with lower SHS, effect size 0.92 (0.86 to 0.99) p=0.028. The directions of association between BMI at inclusion and ES and JSN were similar to that for the total SHS. The effect size of the association with erosion progression was however significant only in the BMI≤20 group, 1.15 (2.72 to 6.42) p=0.025 (in the BMI≥30 group 0.95 (0.90–1.00) p=0.074). On the other hand, association between BMI and JSN progression was significant only in the BMI≥30 group, 0.93 (0.87 to 0.99) p=0.033 (in the BMI≤20 group 2.53 (0.83–7.67) p=0.096).

There were no associations between BMI and radiographic damage in BMI 20<25 and BMI 25<30 groups.

We found no significant association between BMI and ESR, CRP, SJC, TJC over time in the BMI≤20 group. In the BMI≥30 group, BMI was associated with a higher predicted CRP during follow-up, effect size 1.06 (1.01 to 1.12) p=0.028, but not with ESR and SJC. Compared to the patients with BMI 20–25, patients with BMI≥30 had higher TJC over 15 years, 3.17 (1.06 to 9.27) p=0.038.

Conclusions Underweight at onset of RA is associated with more radiographic damage up to 15 years and obesity with less joint damage, independent of sex, ACPA and smoking status. The effect of BMI is not explained by measures of disease activity.

Disclosure of Interest None declared