Ten-Year Incidence of Self-reported Nonvertebral Fractures in 9,720 Japanese Patients with Rheumatoid Arthritis: A Prospective Observational Cohort Study

Background/Purpose: Although rheumatoid arthritis (RA) is a risk factor for osteoporosis and fracture, few studies have described the correlation between disease activity of RA and incidence of fracture in patients with RA.  We previously reported that disease activity in our Institute of Rheumatology Rheumatoid Arthritis (IORRA) cohort improved significantly from 2001 to 2010; however, the change in the incidence of fractures in this period is unknown.  This study aimed to investigate the change in the incidence of nonvertebral fractures between 2001 and 2010.

Methods: The IORRA is a prospective observational cohort study of Japanese RA patients at the Institute of Rheumatology, Tokyo Women’s Medical University.  A total of 9,720 patients (82% female; mean age, 56 years) with RA were enrolled in the IORRA cohort study from 2000 to 2010.  Clinical parameters, including physician’s assessment, patient’s assessment, and laboratory data, were collected biannually in April and October.  All participants self-reported the occurrence of a nonvertebral fracture within the previous 6 months and the fracture sites.  Statistical analysis was performed using the ƒÔ² test with a confidence interval of 95%.

Results: From 2001 to 2010, the percentage of patients with DAS28 remission increased from 7.8% to 39.7%, the mean score on the Japanese version of the health assessment questionnaire decreased from 0.82 to 0.65, prednisolone intake decreased from 51.4% (mean, 4.7 mg/day) to 41.3% (mean, 4.1 mg/day), and bisphosphonate intake increased from 5.0% to 23.4%.  The incidence of nonvertebral fracture was 1.19% in 2001 and 1.79% in 2010, with no apparent change (mean, 1.78% [0.23%]; Figure 1).  The incidence of fracture increased dramatically after the age of 45 years in women, while the increase was linear in men.  The mean age of fracture (year) and the mean incidence of fracture (%) were 61.8 (12.7) and 0.36 (0.05) in the ribs, 65.9 (11.1) and 0.07 (0.03) in the pelvis, 66.8 (10.4) and 0.11 (0.04) in the shoulder, 67.4 (10.4) and 0.12 (0.04) in the wrist, 68.4 (9.4) and 0.18 (0.06) in the hip, and 62.5 (10.6) and 0.13 (0.04) in the ankle, respectively.  Overall incidence of nonvertebral fracture was significantly higher in the autumn/winter than in the spring/summer (p = 0.02).

Conclusion: Despite the improvement in disease activity and functional disability, the incidence of nonvertebral fracture appears to show no apparent change between 2001 and 2010 in our patients with RA.