In this longitudinal observational study, we investigated the formation of new syndesmophytes in the spines of patients with axSpA over a period of 2 years, and identified predictors of new syndesmophyte formation and spinal radiographic progression. About 20% of the patients developed a new syndesmophyte over 2 years. The presence of a syndesmophyte at baseline and low BMD were predictors of the formation of new syndesmophytes and significant mSASSS progression.
Abnormal bone metabolism in axSpA is characterised by pathological new bone formation in the cortical zone of the vertebrae and the loss of trabecular bone from the centres of the vertebral bodies. Osteoproliferation leads to syndesmophytes, while the loss of trabecular bone leads to low BMD [26].
This study is the first to demonstrate that low BMD predicts radiographic progression in axSpA. The main determinants of low BMD in axSpA patients are systemic inflammation and bone-specific inflammation [27]. The inflammatory process is associated with altered systemic bone remodelling, increased bone resorption and impaired bone formation resulting from the effects of inflammatory mediators on the differentiation and activity of osteoclasts and osteoblasts. Proinflammatory cytokines can influence osteoclastogenesis and osteoblastic activity [28]. Thus, the presence of low BMD in axSpA is considered to be a result of altered bone remodelling caused by persistent inflammation. In the present study, serum levels of ESR correlated with BMD values of the lumbar spine, femoral neck and total hip, and Z scores of the femoral neck and total hip. Additionally, CRP levels correlated with the Z score at the lumbar spine and femoral neck (data not shown).
Bone loss resulting from chronic inflammation and the associated changes in bone microarchitecture have been proposed as a potential driving mechanism for the ankylosing process [29]. The inflammatory process induces bone loss, which affects the microarchitecture in the trabecular bone, thereby leading to instability. Reduced bone strength triggers a stabilising anabolic effort that results in bone formation. Trabecular and cortical compartments appear to have different reactions to inflammation; in axSpA, inflammation has a direct effect on the trabecular bone of the vertebrae, but not on the cortical bone [8]. As persistent inflammation in the trabecular bone of the vertebral bodies may prevent the anabolic response from correcting the instability, new bone formation in the cortical bone of the vertebrae may be increased [30]. This would result in the formation of syndesmophytes: compensatory stability for the spine but with a loss of normal mobility [31].
Another explanation is that low BMD may represent the presence of a repairing area that was affected by active inflammation in the past. There is increasing evidence that new bone formation in axSpA is the consequence of previous inflammation in the subchondral bone marrow, with the appearance of granulated repair tissue occurring as a mandatory intermediate step, with this then being followed by new bone formation. It has been proposed that the best current treatment for the prevention of bone formation is the early and effective suppression of bony inflammation [32]. Additionally, if it is possible to detect the presence of repair tissue, the risk of new syndesmophyte formation could also be predicted. The presence of a low BMD means that the repair process may become apparent, to compensate for bone loss resulting from the inflammatory process. Therefore, the presence of low BMD may represent areas with ongoing repair affected by inflammation, such as fatty lesions on MRI.
In the current study, the presence of baseline syndesmophytes in axSpA was found to be the strongest predictor for the formation of new syndesmophytes. This finding is consistent with those of earlier studies. Similar results with respect to spinal radiographic progression have been found in early axSpA patients, as well as in AS patients [5, 33, 34]. However, a substantial proportion of patients with baseline syndesmophytes do not show progression over 2 years. Furthermore, it is arguable whether the presence of baseline syndesmophytes should be considered a true predictor, because the patients already exhibited the features that the model was designed to predict [5].
Smoking is reported to be associated with spinal radiographic progression in early axSpA [6] and the radiographic severity of AS [35], although the exact mechanisms for the influence of smoking on radiographic progression are not known. In the present study, smoking was associated with new syndesmophyte formation only in the univariable analysis, whereas it was independently associated with significant mSASSS progression over 2 years after adjustment for confounding factors. Only the status of smoking at baseline was included in the analysis, and the relatively small number of patients may have influenced the inconsistent results with respect to smoking. The influence of smoking on new syndesmophyte formation should be studied in a larger cohort.
The presence of radiographic sacroiliitis was significantly associated with spinal radiographic progression over 2 years, but this significant association was not present in the multivariable analysis. This finding is consistent with a previous result in early axSpA patients [6]. In the present study, 90% of patients with low BMD had radiographic sacroiliitis. This high proportion of radiographic sacroiliitis in the patients with low BMD could have affected the result. The exact effect of radiographic sacroiliitis on spinal progression should be clarified in a large patient cohort including non-radiographic axSpA patients with low BMD.
Although we did not observe a significant result in this study, systemic inflammatory markers were reported as predictors of radiographic progression in patients with early axSpA in the German Spondyloarthritis Inception Cohort (GESPIC) [6]. This previous study included early axSpA patients with short symptom duration (mean 4.2 years) and low mSASSS (mean 4.25) [6]. In this study, the mean symptom duration and mSASSS were 8.6 years and 7.6 points. These differences suggest that our study patients had more severe and longstanding symptoms and were advanced patients. In a prospective observation of a cohort with longstanding AS, inflammatory markers did not emerge as independent predictors, as per the results of the present study [5]. These discordant results could possibly be explained by the differences in disease duration and structural damage severity. Furthermore, we only analysed the associations between baseline ESR and CRP measurements and new syndesmophyte formation, not time-averaged inflammatory markers. Baseline inflammatory markers may be less reflective of the status of persistent systemic inflammation than time-averaged values. Therefore, our results for the predictive role of systemic inflammation should be interpreted with caution.
Inflammation plays a key role in bone loss in axSpA, and anti-inflammatory drugs are expected to have a beneficial effect on bone through both the increased mobility related to pain relief and the direct effects on bone [28]. The best treatment for the prevention of bone formation/progression is currently the early and effective suppression of bony inflammation [32]. Our results suggest that successful anti-inflammatory treatment reduces inflammation and allows the bone metabolism to normalise, thereby taking away the compensatory anabolic response that leads to new bone formation in the cortical bone of the spine. Although we did not find a beneficial effect of NSAIDs or TNF inhibitors in this 2-year follow-up study, recent long-term follow-up data show that TNF inhibitors suppress radiographic spinal progression [36, 37]. Thus, if patients are treated for a longer time with a TNF inhibitor (preventing new occurrences of the sequence of inflammation, repair and new bone formation), or if they are treated early in the course of their disease, such a treatment seems to be effective in retarding the process of new bone formation [32]. Taken together, active anti-inflammatory treatment is crucial for the prevention of spinal ankylosis in young axSpA patients, especially those with low BMD.
This study has some limitations. First, the number of axSpA patients was relatively small; therefore, the regression analysis could be underpowered. In our patients, those with low BMD received more calcium agents and vitamin D treatments, and these agents could have affected the bone metabolism. Although calcium and vitamin D intake were not significant in the regression analysis, their effects on spinal progression need to be clarified in a large cohort including more patients with low BMD. The BMD in axSpA can be affected by the presence of syndesmophytes or other structural lesions such as an ankylosed posterior arch and periosteal bone formation. Therefore, the BMD of the lumbar spine in patients with syndesmophytes could have been overestimated. Lastly, it is known that there is an association between serum levels of sex hormones and BMD in AS [38], but these were not measured in this study.