In this longitudinal cohort of patients with early active RA we conducted a cross-sectional analysis and found a prevalence of vertebral fractures of 15% after 5 years of DAS-steered treatment. The majority of these patients (73%) had only one vertebral fracture and most deformities were grade 1 (69%).
The prevalence is lower than previously reported in non-DAS steered cohorts (19-36%)
[1–3, 7, 10, 19]. This suggests that adequate control of disease activity in earlier RA, with a treat-to-target approach, may protect against the occurrence of vertebral fractures. However, the prevalence of vertebral fractures in the general population is about 5%
[2, 10], which suggests that vertebral fractures may be a symptom of RA even if the disease activity appears relatively well controlled.
Most previous studies have focused on the prevalence of vertebral fractures in female RA patients, and postmenopausal women appeared to be especially at risk
. In a population-based cohort study of men and women older than 55 years, the incidence of vertebral fractures was higher in women than in men with a decrease in BMD as an independent predictor
. We found that the prevalence of vertebral fractures in male patients in our cohort was not statistically different from the prevalence in female patients (17% versus 14%, respectively). The prevalence of vertebral fractures in males was similar to that in postmenopausal women (16%) and higher than in premenopausal females (9%). This is probably related to the fact that males were of similar age as the postmenopausal females (57 versus 58 years, p = 0.281). We found higher age to be independently associated with a high prevalence of vertebral fractures. Also smoking was a risk factor for vertebral fractures.
In contrast to other studies, we found no association with the use of prednisone
[3, 6, 10]. No difference was found in vertebral fractures between the four treatment groups, where in group 3 patients were initially treated with a high dose of prednisone (60 mg daily, tapered in 7 weeks to 7.5 mg daily). Ever use of prednisone was not associated with more vertebral fractures compared to never use, nor was use of prednisone longer than a year compared with use of prednisone less than a year. This suggests that the beneficial effects of prednisone on suppression of disease activity outweigh the potentially deleterious effects on bone quality and vertebral fractures
In this cohort, use of bisphosphonates appeared not to be associated with a lower prevalence of vertebral fractures. This is at odds with previous reports that the use of bisphosphonates reduces the risk of vertebral and non-vertebral fractures
[21–23]. In daily clinical practice, bisphosphonates are only prescribed if risk factors for bone mineral density loss are present, implying confounding by indication. In addition, bisphosphonates could have been started after a fracture has occurred, which also might explain why we found no protective effect. Most important independent risk factor of vertebral fractures after 5 years was a higher age.
It has been demonstrated that lower BMD is associated with the presence of vertebral fractures.
[3, 10] We previously reported that BMD loss in the BeSt study was not depending on prednisone use but on suppression of disease activity
. Since also rheumatic disease activity over time was slightly higher in patients with vertebral fractures compared to patients without vertebral fractures (mean difference 0.19), these findings support the hypothesis that there is a relation between disease activity and poor vertebral bone quality. We did not find a statistically significant association between BMD and vertebral fractures, although mean BMDs over time of the spine and hip of patients with vertebral fractures were somewhat lower than patients without vertebral fractures. Probably, this is related to the small changes in BMD over time, as a result of adequate disease control. In addition, after adjustment for age, itself significantly associated with a higher prevalence of vertebral fractures, the association between BMD and vertebral fractures becomes smaller. It thus appears that age related factors play an important role in the development of vertebral fractures.
More of clinical importance, functional disability over time was higher in patients with vertebral fractures compared to those without. This association was earlier described in a cross-sectional study, which sheds no light on the possible mechanism behind it
. We found that the higher functional disability in patients with vertebral fractures was independent of disease activity and age, suggesting that patients with vertebral fractures suffer functional impairment due to the vertebral fractures or the underlying bone condition, independent of rheumatic symptoms. It may also be possible that HAQ deterioration occurred before the vertebral fractures, but as HAQ deterioration is dependent on disease activity in the early phases of RA
 and as we found that patients with vertebral fractures have higher HAQ over time irrespective of DAS over time, it seems unlikely that HAQ had preceded vertebral fractures, as earlier proposed
. Therefore, prevention of vertebral fractures is important in terms of reducing functional disability.
A limitation of this study is the absence of baseline radiographs of the spine. We are therefore not sure when the vertebral fracture occurred in relation to the onset of RA and the start of antirheumatic and anti-resorptive treatment. It is possible that fractures occurred before RA was diagnosed, in which case the prevalence of vertebral fractures in well suppressed RA is even less than we observed, or that bisphosphonates were started after a fracture has occurred. Although the prevalence of vertebral fractures was determined cross-sectionally, the differences in disease activity over time between patients with and without vertebral fractures do suggest that inflammation plays a role in the occurrence of vertebral fractures.
In conclusion, in a cohort of patients with recent-onset active RA, after 5 years of DAS-steered treatment the prevalence of vertebral fractures was with 15% higher than in the healthy population and similar for males and females. Independent predictors for vertebral fractures were age and smoking. Mean BMD over time of the spine and hip were not associated with the presence of vertebral fractures and vertebral fractures did not occur more in patients who used prednisone than in patients who did not. Patients with vertebral fractures suffer greater functional disability over time than patients without, independent of a slightly higher rheumatic disease activity. This indicates that vertebral fractures can be seen as bony damage of rheumatoid inflammation and may be prevented by optimal disease activity suppression.