Osteoporosis is a common disease and a major risk factor for fragility fractures, particularly in older patients.[1] Patients with rheumatoid arthritis (RA) also have an increased risk of vertebral fractures.[2-4] In one study, the number of vertebral deformities in the RA group (female; mean age, 63.0 years) was 2.9 times higher than in the control group.[3] In another, the odds ratio (OR) of vertebral fracture in the RA group was 6.5, greater than that of the control group.[4] In general, among women, the age-adjusted relative risk of mortality following vertebral fracture was 8.64 (95% confidence interval [CI], 4.45-16.74).[5] In community-dwelling volunteers aged ≥65 years, the hazard ratio of mortality in individuals with osteoporosis and vertebral fracture was 1.89 (95% CI, 1.27-2.77).[6]

The Japanese population is aging, and the onset of RA has increased.[7] Therefore, examining the factors associated with vertebral fractures in patients with RA is important. This retrospective study aimed to investigate factors associated with vertebral fractures and treatment goals to prevent new vertebral fractures in patients with RA. Japanese patients with RA were analyzed to determine the prevalence and identify significant predictors of vertebral fractures and treatment goals to prevent vertebral fractures. The hypothesis of this study was that treatment goals for vertebral fracture prevention depend on the presence of risk factors for vertebral fracture. This study should contribute to better clinical management and improved outcomes for patients with RA who are at risk of vertebral fractures.

A total of 304 patients with RA who fulfilled the American College of Rheumatology classification criteria (1987) and/or the ACR/European League Against Rheumatism criteria [8,9] in 2018 were enrolled in this retrospective study. The cohort database from the authors’ hospital was used. Patients with missing data, scoliosis, difficulty walking independently, dementia, hospital transfer, or death were excluded (Fig. 1). The patients underwent spine X-rays and bone mineral density (BMD) measurements annually from 2018 to 2022. In 2017, the American Society for Bone and Mineral Research and the United States National Osteoporosis Foundation formed a working group that described goal-directed treatment for osteoporosis.[10] One of principles of goal-directed treatment for osteoporosis is a fracture-free interval of 3 to 5 years. Therefore, the follow-up period was four years.

Vertebral fractures were evaluated using thoracic and lumbar spine (LS) lateral X-rays of the T8 to L5 vertebrae. The presence of a vertebral fracture indicated grade 1 or higher on a semiquantitative grading scale.[11] A new vertebral fracture indicated an increase of at least one grade on the scale, while a worsening fracture was defined as one that results in a height loss of 20% or more at the vertebra.[11,12]

The BMD of the LS (L2-4), total hip (TH), and femoral neck (FN) were measured using dual energy X-ray absorptiometry (Lunar Prodigy Advance; GE Healthcare, Madison WI, USA).

The clinical data included the following: age, sex, disease duration, body mass index, Clinical Disease Activity Index (CDAI), Health Assessment Questionnaire Disability Index (HAQ-DI), serum Ca level corrected for serum albumin level, and pharmacotherapy status. Pharmacotherapy for RA included the administration of methotrexate (MTX), glucocorticoids (GC), and disease-modifying antirheumatic drugs (DMARDs), including biological DMARDs (bDMARDs) and targeted synthetic DMARDs (tsDMARDs). Pharmacotherapy for osteoporosis includes the use of bisphosphonates, denosumab, teriparatide, romosozumab, selective estrogen receptor modulators, and active vitamin D3. Patient adherence to osteoporosis medications was examined.

The CDAI is a composite disease activity measure of RA and calculated as the sum of swollen joint count, tender joint count, patient’s global assessment, and evaluator’s global assessment.[13] The HAQ-DI is a patient-reported measure of physical disability.[14]

The characteristics and clinical data for the years 2018 to 2022 and the results of the trend tests are summarized in Table 1. The rates of bDMARD or tsDMARD, MTX, and GC use were 53.9%, 69.4%, and 19.4%, respectively. The mean CDAI and HAQ-DI were 4.2±4.5 and 0.3±0.6. This study found that 41.8% of patients received pharmacotherapy for osteoporosis. The adherence rates for denosumab, teriparatide, romosozumab, bisphosphonate, selective estrogen receptor modulator, and active vitamin D were 95.6%, 86.9%, 94.5%, 93.7%, 99.2%, and 97.7%, respectively. Vertebral fractures were noted in 26.3% of patients. The mean T-scores in the LS, TH, and FN 2018 were −0.7±1.4, −1.3±1.0, and −1.7±1.1, respectively. Additionally, the standard error of the mean Tscores in the LS, TH, and FN were 0.08, 0.06, and 0.06, respectively. In 2019, 2020, 2021, and 2022, the mean T-scores in the LS changed by −0.7±1.3, −0.6±1.4, −0.6±1.4, and −0.5±1.4, respectively; those in the TH changed by −1.3±1.0, −1.3±1.0, −1.3±1.0, and −1.3±1.0; and those changed by −1.8±1.0, −1.8±1.0, −1.8±1.0, and −1.7±1.0. The minimal clinically important difference (MCID) in the T-scores was calculated based on 2018. These values for the LS were 0.228 in 2019, 0.241 in 2020, 0.268 in 2021, and 0.317 in 2022; those for the TH were 0.203 in 2019, 0.197 in 2020, 0.252 in 2021, and 0.235 in 2022; and those for the FN were 0.258 in 2019, 0.275 in 2020, 0.305 in 2021, and 0.281 in 2022. Based on the MCID, the T-score did not significantly change for any site.

The prevalence rates of new vertebral fractures in 2018-2019, 2019-2020, 2020-2021, and 2021-2022 were 2.0%, 1.3%, 2.3%, and 2.0%, respectively. Results of analyses of factors associated with new vertebral fractures are summarized in Table 2. Univariate analysis revealed that the significant variables were the presence of vertebral fractures (P<0.001), and T-scores in the LS (P=0.002), TH (P= 0.003), and FN (P<0.001). In the multivariate analysis, the presence of existing vertebral fractures was associated with new vertebral fractures (P=0.003; OR, 0.241; 95% CI, 0.093-0.624).

The cut-off T-score values for new vertebral fractures in the LS in patients with and without pre-existing vertebral fractures were −0.7 (sensitivity, 40.9%; specificity, 100%; AUC, 0.686) (Fig. 2A) and −1.4 (sensitivity, 69.0%; specificity, 62.5%; AUC, 0.651) (Fig. 2B), respectively.

In this study, the prevalence rates of new vertebral fractures continued at a constant rate over four years. The presence of pre-existing vertebral fractures was identified as a factor associated with the occurrence of new vertebral fractures. The occurrence of vertebral fractures is associated with high C-reactive protein (CRP) levels, high RA disease activity, high HAQ-DI, presence of existing vertebral fractures, and GC use.[15-17] In a previous cross-sectional study, independent factors associated with the presence of vertebral fractures were CRP (OR, 1.76; 95% CI, 1.20-2.94) and DAS28 (OR, 1.66; 95% CI, 1.30-2.12).[16] In another study, HAQ-DI was associated with the presence of vertebral and hip fractures (OR, 1.71; 95% CI, 1.01-2.89). [18] The patients in this study had a mean CDAI of 4.2±4.5 and a mean HAQ-DI of 0.3±0.6. Additionally, 92.1% and 81.3% of the patients had low disease activity and achieved functional remission, respectively. Recently, the rate of clinical remission has increased.[19] We assume that the disease activity of RA may not be associated with new vertebral fractures in patients with controlled disease activity such as those in this study. Although the relationship between GC use and vertebral fractures in patients with RA is controversial,[4,20-22] GC use can inhibit bone formation and affect bone quality.[23-25] GC dosage is related to worse BMD values and increased fracture risk. A high dose (>7.5 mg/d) can cause a higher probability of fractures than low (<2.5 mg/d) and intermediate (2.5-7.5 mg/d) doses.[26] The low percentage of GC use (19.4%) and low doses (3.1±1.7 mg/d) in this study may have affected the results.

Based on these results, the presence of pre-existing vertebral fractures was identified as a factor associated with the occurrence of new vertebral fractures. Pre-existing vertebral fractures were associated with a 4- to 5-fold increased risk of developing new vertebral fractures.[27] Another cohort study found that the prevalence of pre-existing vertebral fractures was associated with the T-score of the LS.[28] In this study, the cut-off values for the T-score of the LS in patients with and without pre-existing vertebral fractures were −0.7 and −1.4, respectively. Patients with pre-existing vertebral fractures require strict clinical management. To prevent new vertebral fractures, it is crucial to achieve and maintain BMD levels exceeding the cut-off values for osteoporosis treatment.

This study has some limitations. First, although there was an inflow and outflow of patients due to the nature of real-world data, the data were obtained from ongoing information in daily practice. As shown in the flowchart of enrolled participants, this study has a selection bias. However, when treating RA and osteoporosis, the decision to continue or discontinue drugs is made for various reasons. Second, 41.8% of the enrolled patients had already received treatment for osteoporosis at baseline. In particular, the effectiveness of bisphosphonates on BMD was stable over the long term.[29,30] Third, the enrolled patients were relatively old, had prolonged clinical courses, and displayed suppressed RA disease activity of RA as determined using CDAI. Therefore, factors associated with new vertebral fractures may have been influenced by the clinical profiles of the patients. Finally, the presence or absence of acute vertebral fractures and their severity were not investigated. Despite these limitations, the results of this study are important for the management of osteoporosis in patients with RA in daily practice.

In conclusion, this study demonstrated that the presence of pre-existing vertebral fractures is an independent factor associated with new vertebral fractures. Based on these results, it is important to tailor treatment goals based on the presence or absence of vertebral fractures to effectively prevent new fractures from forming.