OBJECTIVES

Bone fragility fracture (BFF) is a serious incident in treating rheumatoid arthritis (RA). We hypothesized that pain degree during treatment RA correlated with incident BFF and validated how pain affects incident BFF (inc-BFF).

METHODS

Postmenopausal RA patients treated for at least 3 years were recruited. The primary endpoint was the development of inc-BFF. Follow-up began with the first bone mineral density measurement (baseline) and continued until the development of the first BFF or termination of the study. Clinical indicators at baseline, including pain score using a visual analog scale (PS-VAS), were analyzed statistically using Cox regression analysis, receiver operation characteristics (ROC), Kaplan-Meier survival curve analysis (K-M), and chi-square test.

RESULTS

A total of 239 patients were recruited. Using a multivariate Cox regression analysis, the baseline's PS-VAS and prevalent BFF (pr-BFF) demonstrated significantly higher risk ratios. For ROC, pr-BFF and PS-VAS had significant cutoff index (COI) (positive, 21.0) and an area under-curve of 0.692 (P < 0.001) and 0.616 (P < 0.01), respectively. PS-VAS > COI had a 2.24-fold higher hazard ratio than PS-VAS ≤ COI using K-M. When these 2 conditions were combined, patients with pr-BFF-positive and PS-VAS-positive had a sensitivity of 42.3% and a specificity of 88.8% for the inc-BFF. PS-VAS > COI had no statistical significance in the subgroup without pr-BFF, whereas the existence of pr-BFF had a significantly higher risk ratio in the PS-VAS ≤ COI.

CONCLUSIONS

The PS-VAS during RA treatment is a good indicator for predicting the inc-BFF in postmenopausal RA patients with pr-BFF.