Diagnosis of bone loss in AS

Use of DXA in AS

Currently there is no accurate measure of overall bone strength. BMD measured by DXA is utilized as a surrogate measure of bone strength. However, osteoporosis is characterized by compromised bone strength and high fracture risk. Osteoporosis and low bone density are defined using measurements of BMD. The World Health Organization (WHO) has defined bone loss using DXA based measurements of BMD (Kanis 1994). The World Health Organization defines osteoporosis as BMD 2.5 standard deviations below the mean for matched young adults. Bone loss in men less than 50 years old and premenopausal women are categorized using the ISCD definition (Schousboe 2013). BMD measurements have other advantages as well. For instance, serial BMD readings are useful to demonstrate ongoing bone loss and to monitor the effect of various treatment modalities (Marshall 1996). The three major sites used to generate BMD measurements are lumbar spine (L1to L4), total hip and femoral neck.

Though distal forearm is another important site used to measure BMD, it is recommended only when hip and/or lumbar spine cannot be used for measurements, in obese people and in patients with hyperparathyroidism. There are certain drawbacks of BMD. The main limitation of BMD is that it explains bone strength only to a certain extent (Genant 2008, Bouxsein 2008, Delmas 2004, Watts 2005). In addition, the definitions proposed by the WHO can be applied only to men older than 50 years and postmenopausal women.

Limitations of using DXA in patients with AS

Certain limitations exist regarding the use of DXA imaging in patients with AS. These are related to the sites of DXA used, technical difficulties in positioning of the patients and to the lack of specific data regarding the association between BMD and fracture risk in AS.

Limitations of spine DXA in AS

Conventional DXA machines use antero-posterior view of the lumbar spine to generate values of BMD. Spine is the most affected skeletal organ in AS and patients with AS develop bony changes at the spine as the disease progresses. This leads to false positive readings of BMD at the lumbar spine in patients with long- standing and advanced radiological disease (Ulu 2013, Gilgil 2005). BMD of the lumbar spine is also unreliable in AS to assess longitudinal bone loss since BMD tends to become normal as the disease progresses due to the potential radiological progression of the disease. Patients with advanced radiological disease have fusion and immobility of the spine and this limits their ability to lie down flat on the DXA table while undergoing the study. Improper positioning can produce erroneous results. One solution for better interpretation of lumbar spine BMD using DXA is to perform lateral DXA (Ulu 2013, Gilgil 2005).

Facet joint disease, fusion of interapophyseal joints, calcification of ligaments and presence of syndesmophytes increase the miscalculation (Baek 2005, Devogelaer 1992, Masud 1993). Ulu and colleagues suggested that spine BMD can be falsely high in patients with AS (Ulu 2013). Moreover, the PA spinal BMD tended to be higher in patients who had vertebral fractures (28 % of patients) but the lateral spinal BMD values were significantly lower in the fracture group (p=0.004) and also in patients with syndesmophytes (p = 0.004). In summary, it is clear that BMD measurement at the lateral lumbar spine reflects bone loss and fracture risk better. DXA scan using lateral view of the lumbar spine may be better in diagnosing bone loss in patients with AS. Lateral DXA can isolate the vertebral body from the ankylosed posterior elements, the zygapophyseal joints, endplates and syndesmophytes during the measurement. It can also detect decreased BMD better than postero-anterior views of the spine (MacKay 2000). Besides, proper positioning can be difficult in some patients who are unable to lie in the lateral decubitus position due to spinal deformities. The interpretation of images can be complicated due to the interference of ribs and iliac crests. Newer scanners are capable of measuring lateral DXA from supine positioning (Blake 1994). The main limitation with this approach is that BMD of the lateral spine has lower precision than femur neck or AP spine. Hence, lateral DXA may be unreliable for follow up.

Limitations of hip DXA in AS

The measurement of BMD at the hip in patients with AS is also associated with certain issues. Hip might not be a reliable site because of the pathologic involvement of hip joints in AS (203 Vander Cruyssen 2013, Huang 2013). Changes that happen at the hip include subchondral bone marrow edema, enthesitis, fatty infiltration of the bone marrow, bony erosions, and narrowing
of the joint space. MRI is capturing more information on clinical and subclinical hip involvement than the assessments based on X-rays and clinical symptoms. For example, in a study done by Huang et al in 2013, 74% had abnormalities detected at the hip (86/116) as opposed to 21% (24/116) and 30% (35/116) detected based on radiographs or clinical symptoms (Huang 2013). Also, many patients undergo hip replacement during the course of the disease making it impossible to use hips for DXA scanning. Appropriate positioning and internal rotation of the hip are essential prerequisites for performing a good DXA study. But as stated above, patients with AS sometimes face difficulties in positioning their limbs on the DXA table. The presence of coxitis can limit the ability to internally rotate the limb during DXA scanning (Ghozlani 2009). Even minor changes in the degree of rotation of the hip can affect measurements of the DXA.

Limitations of DXA in determining fracture risk in AS

The ability of DXA based measurements to predict fracture risk in patients with AS is also unclear. BMD at both lumbar spine and hip have not been consistently linked to the high risk or fractures in patients with AS (Chen 2013, Mitra 2000 and Weiss 2010). For instance, in 1994, results of a study done by Donnelly et al confirmed that low femoral neck BMD was not related to the presence of osteoporotic vertebral fractures (Weiss 2010). In the cross sectional study done by Ghozlani and colleagues, lumbar spine BMD was similar in those with or without VFs (1.059 ± 0.2 vs. 1.052 ± 0.1, p=NS, n=80) (Ghozlani 2009). In the multivariable regression model, the odds ratio of lumbar spine for a Genant grade 2 or 3 vertebral fracture was not significant (OR=12.15; 95% CI=0.41–359.7). The authors also observed that about 30% of patients with osteopenia and 20% with normal BMD were found to have vertebral fracture (grades 2 and 3). However the authors used VFA and not spine X-ray to document
spine fractures. Spine radiograph is the gold standard for diagnosing vertebral fractures.
Finally, there are issues related to the limitations of the DXA technique. The areal DXA measures mineral content of the bone in a two dimensional manner. It cannot account for the size and depth of the bone. The bones in men are larger; therefore corresponding DXA readings are higher in men. Also, DXA cannot distinguish between cortical and trabecular compartments of the bone. This is particularly disadvantageous in inflammatory diseases such as AS because there can be differential involvement of the cortical and trabecular compartments.

Assessment of bone strength in AS

Almost all the current evidence that describes bone loss in patients with AS is based on data obtained from studies done using DXA. As discussed above, DXA has limited resolving power to study the structural and strength or mechanical properties. Given the disadvantages and limitations of DXA in AS, it is crucial to identify better predictors of fracture risk. Bone strength is the main determinant of fracture risk. Information about bone strength would also help in identifying targeted treatment to treat bone loss in AS.
Monitoring bone strength also helps to identify the benefits of anti-osteoporosis or anti-TNF alpha medications on bone that may occur independent of changes in BMD. Moreover, it is likely that bisphosphonates and TNF-alpha inhibitors may exert differential effects on the cortical and trabecular compartments of the bone. Monitoring treatment response using DXA alone might underestimate the benefits of TNFalpha inhibitors or bisphosphonates. For instance, in addition to its ability to increase BMD, alendronate has been shown to improve parameters such as cortical thickness, trabecular BV/TV and estimated failure load (Burghardt 2010). Zoledronic acid has been shown to increase cortical thickness, cortical density and trabecular volume fraction in radius and tibia without much effect on cortical porosity (Hansen 2013). Finite element estimated bone strength was preserved, but not increased with zoledronic acid (Hansen 2013). No data exists on the effect of TNF-alpha inhibitors on bone strength in AS. In summary, there exists an urgent need to study bone strength in AS.

Bone histomorphometric studies on bone structure in AS

Bone structure is studied in detail using histomorphometric of iliac crest specimens obtained from bone biopsies. Histomorphometry or quantitative histology is the gold standard for assessing bone structure (Recker). It is also helpful to study bone formation and resorption (Malluche 2007). However, bone histomorphometric studies in AS are scarce. Szejnfeld et al conducted biopsies of the iliac crest in 16 Caucasian males with AS (mean age: 34+ 3 years and disease duration: 11+2 years) and reported that bone formation was affected more than bone resorption (Szejnfeld 1997). Fourteen patients presented osteopenia, ten had mineralization defects, and three patients presented with osteomalacia. Trabecular bone mass and wall thickness were found to be lower in subjects with AS than the controls. The osteoid volume, osteoid thickness and mineralization lag time were significantly greater than control values (p < 0.05) but both control subjects and patients had similar bone osteoclast interface and the eroded surface. The mineral apposition rate was less than the control group (p < 0.01).
The authors concluded that decreased bone mass seen in AS was due to reduced bone formation rather than an increase in bone resorption since the resorption parameters such as bone osteoclast interface and eroded surface were similar to those of controls. However, despite being the gold standard, bone biopsy has certain limitations. Although the histomorphometric parameters are obtained from the iliac crest, the sites susceptible to fragility fractures are the load-bearing sites and vertebral bodies. Thus, it is likely that iliac crest biopsies may not accurately represent the pathological features that may predispose to
fractures. Moreover, bone biopsy is an invasive procedure and for this reason, it is not practical to do biopsy in a clinical setting on a routine basis.