Rheumatology International

, Volume 26, Issue 2, pp 115–120

Ankylosing spondylitis and bone mineral density—what is the ideal tool for measurement?


    • Department of Rheumatology and OsteologyKerckhoff Clinic and Foundation
  • A. Kluge
    • Department of RadiologyKerckhoff Clinic and Foundation
  • J. Strunk
    • Department of Rheumatology and OsteologyKerckhoff Clinic and Foundation
  • J. Teichmann
    • Medical Clinic CCity Hospital
  • G. Bachmann
    • Department of RadiologyKerckhoff Clinic and Foundation
Original Article

DOI: 10.1007/s00296-004-0515-4

Cite this article as:
Lange, U., Kluge, A., Strunk, J. et al. Rheumatol Int (2005) 26: 115. doi:10.1007/s00296-004-0515-4


Ankylosing spondylitis (AS) is characterised by chronic inflammation and partial ossification, yet vertebral fractures due to osteoporosis, although common, are frequently unrecognised. The aim of this study was to (1) show the frequency of changes in the progress of osteopenia/osteoporosis in AS depending on duration and stage of the disease and (2) assess the ranking of two different methods of bone density measurement in this clinical pattern. We measured bone density in 84 male and female patients with both dual X-ray absorptiometry (DXA) and single energy quantitative computed tomography (SE-QCT). In the initial and advanced stages of the disease, a high decrease in axial bone density could be verified (DXA: osteopenia in 5% and osteoporosis in 9.2%; SE-QCT: osteopenia in 11.8% and osteoporosis in 30.3%). Peripheral bone density decrease as in osteopenia could be proven in 17.6% by DXA measurement. With SE-QCT, a decrease in vertebral trabecular bone density could already be observed in the initial stage and continued steadily during the course of the disease; cortical bone displayed the same trend up to stages of ankylosis. With DXA, valid conclusions are more likely to be expected in less marked ankylosing stages of AS. In stages of advanced ankyloses in the vertebral region (substantial syndesmophytes), priority should be given to SE-QCT, due to the selective measurement of trabecular and cortical bone. The DXA method often yields values that are too high, and the replacement of vertebral trabecular bone by fatty bone marrow is not usually recorded as standard. There may already be an increased risk of bone fracture in AS in osteopenia on DXA along with an osteoporosis already established on SE-QCT.


Ankylosing spondylitisBone mineral densityOsteopeniaOsteoporosis


In the past few decades, the pathogenesis of new bone formation and ossification of the ligamentous apparatus was of primary concern in the characterisation of ankylosing spondylitis (AS). Despite the development of syndesmophytes, patients with AS are not protected against fractures, and osteoporosis can be diagnosed in a large percentage of cases [1, 2, 3, 4, 5]. In contrast to rheumatoid arthritis, in AS the anatomical distribution, clinical significance, and point of the occurrence of osteoporosis are for the most part still not clear.

At present, a number of different methods for objectifying a decrease in bone mineral density exist [6, 7, 8, 9, 10, 11]; nevertheless, although the prevalent methods currently employed enable quantitative determination of bone mineral density, analysis of the bone structure is not possible. Nor is it possible to compare the specified methods for measuring bone density, firstly because they are not standardised and secondly because they measure different aspects. Results of the osteodensitometry methods do have predictive value as far as the risk of fracture is concerned, but even the advantages and disadvantages of the methods in question differ considerably (Table 1, Table 2, Table 3) [6, 7, 8, 9, 10, 11]. These studies were conducted with the aim of evaluating bone density decrease in AS by means of two different methods to measure bone density, dual X-ray absorptiometry (DXA) and single energy quantitative computed tomography (SE-QCT), and arriving at a critical assessment of the results.
Table 1

Current problems in bone density measurement

1. Different methods (dual photon absorptiometry, quantitative ultrasonometry, computed tomography)

2. Different topographic measurement (lumbar, radius, hip [forward triangle, trochanter, total hip], heel)

3. Poor comparison of the place of measurement

4. Different norm area (different definition, different standard: mean±2 SD)

5. False interpretation (low bone mineral density equivalent to osteoporosis)

Table 2

Problems of integral bone mineral density measurement methods

1. False high BMD because of degenerative manifestations (i.e. spondylophytes, osteochondrosis, scoliosis, inflammation, ossification, or calcifying/sclerotic manifestations (i.e. aorta, lymph nodes, ligaments)

2. X-ray of the spine is necessary for valid interpretation

3. Strong lumbar lordosis reduces the demarcation of the vertebrae and reduces reproduction

4. No possible statement about morphogenetics

5. No separate measurement of cortical and trabecular bone

Table 3

Problems of computed tomographic measurement

1. Only measurement of the middle layer of the vertebrae (result is not valid for the entire vertebra)

2. High fat error (i.e. SE=10–30%). (Discrepancy between model and reality is the reason for the systematic error)

3. Different selection of the region of interest (ROI)

4. Different norm areas

5. Fractures of the vertebrae must be excluded

6. Higher radiation load (i.e. SE-QCT<100 μS)


Eighty-four Caucasian patients with defined AS (31 women, 53 men) were selected from our clinic. All patients fulfilled the New York criteria [12] and those of the European Spondylarthropathy Study Group. For demographic and disease-specific data, refer to Table 4. The study persons were selected from different groups—group I: 27 patients with only symmetric sacroiliitis; group II: 22 patients with sacroiliitis and with one section of the spine involved; group III: 15 patients with sacroiliitis plus two sections of spine involved; and group IV: 20 patients with sacroiliitis plus three sections of the spine involved. X-ray of the spine (thoracal and lumbar) was performed on all study persons. Bone mineral density was measured at the lumbar spine (L2–4, anteroposterior view) and at the total hip with DXA (Lunar, USA) and lumbar SE-QCT (Somatom Plus 4, Siemens). The coefficients of variation of repeated measurements in vivo were 0.9% (Lunar) and 1% (SE-QCT) for the lumbar spine and 1.6% (Lunar) for the hip. Low bone mineral density was defined by Lunar according to World Health Organisation guidelines as a t- score less than 2.5 standard deviations below the mean of young adults [13]. T- scores from the normal mean obtained from young healthy adults were calculated with SE-QCT; those less than 3.0 SD were defined as osteoporosis, corresponding to a trabecular bone mass of less than 80 mg/ml.
Table 4

Demographic and disease-specific data. NSAIDs nonsteroidal anti-inflammatory drugs, DMARDs disease-modifying antirheumatic drugs, HLA human leukocyte antigen

AS patients

Group I (n=27)

Group II (n=22)

Group III (n=15)

Group IV (n=20)






Age (years)





Age at time of diagnosis (years)





Duration of disease (years)





Peripheral arthritis





Gut inflammation





NSAIDs in the previous 12 months





DMARDsin the previous 12 months





Glucocorticoids in the previous 12 months










Statistical analysis

All data are presented as means±SD and evaluated using the version 10.0.7 statistical software package for PC (SPSS, Chicago, Ill., USA). Data were compared using the t-test. P<0.05 was taken as significant.


In the initial and advanced stages of AS, bone density decrease was verified in a rather high percentage of cases (DXA: osteopenia in 5% and osteoporosis in 9.2%; SE-QCT: osteopenia in 11.8% and osteoporosis in 30.3%). With the DXA measurement, peripheral bone density decrease as in osteopenia was verified in 17.6%. Table 5 shows the osteopenia/osteoporosis frequency in the entire study group as well as in the subgroups according to the X-ray, DXA, and SE-QCT results. Vertebral body fractures as a late sign and a complication of osteoporosis were found in nine patients (one premenopausal woman, eight men) (Table 5). They all had osteoporosis by SE-QCT measurement.
Table 5

Osteopenia/osteoporosis by DXA and SE-QCT


I (n=27)

II (n=22)

III (n=15)

IV (n=20)

Total (n=84)

DXA lumbar






DXA total hip






SE-QCT (trabecular)






Vertebral fractures by thoracolumbar X-ray






aAll patients with vertebral fractures had osteoporosis by SE-QCT measurement (<80 mg/ml trabecular bone mass)

Comparing the same stages of the disease, we found no significant differences between men and women; however, complete ankylosations of the spine were less frequent in women (one woman, 19 men). The axial DXA measurements show the same progression (Fig. 1) as in SE-QCT of the cortical bone (Fig. 2): this was reduced at first in more advanced stages of the disease, only to increase again with the manifestation of ankyloses. The differences were similar on the left femoral neck, with no significance, however.
Fig. 1

Results of dual X-ray absorptiometry, axial

Fig. 2

Results of quantitative computed tomography, cortical bone

Concerning the men, in SE-QCT the bone density of the spongiosa was already reduced in the initial stage, but this is not significant for the entire study group. The mineral salt content then continued to decrease significantly; even in stage IV there was still a significant decrease in bone density (Fig. 3).
Fig. 3

Results of quantitative computed tomography, trabecular bone

It is in stage IV, though, that the variance of the test data was the greatest. Along with highly pathological osteoporosis with almost complete loss of the cortical bone, two male patients had a mixed form with anulus ossifications and hyperostotically modified syndesmophytes showing on normal X-rays, which also verified increased spongiosa. The absolute measurements of spongiosa in later stages are usually influenced by advancing age, but the age/norm measurements also showed a lessening of spongiosa in advanced stages of the disease (Fig. 4).
Fig. 4

Results of quantitative computed tomography, trabecular bone


Secondary osteoporosis in inflammatory rheumatic conditions is by no means rare and represents a problematic and pathogenetic heterogenous complication [14, 15, 16, 17]. With its attendant increased fracture risk, osteoporosis is a common complication in AS [4, 17, 18]. Even early descriptions of the clinical picture arrived at by means of pathological studies contain corresponding findings [19, 20].

Only in the past few years have there been publications of international studies dealing with axial osteoporosis in AS [2, 3, 4, 21, 22, 23, 24]. The aetiopathogenesis of secondary osteoporosis in AS is still unclear. According to Schilling [21, 22, 25], it is possible to differentiate between two forms of osteoporosis: a so-called early type, associated with the spondylarthritis type in young people (onset before age 20), and a late type (syndesmophytic development) in adults. The categories presented by Schilling have been corroborated by recent studies which were able to verify decreased bone density already in early forms of AS [1, 2, 3, 4, 23, 24, 26]. Whereas segment inactivity resulting from ensuing ankylosis is assumed as a pathogenetic factor in the late type of osteoporosis, inactivity atrophy in the early type of osteoporosis is quite unlikely a cause [21, 22, 25].

Longitudinal studies in early AS have shown that spine and hip bone mineral density decreased predominantly in patients with active AS [24, 27, 28] who maintained normal spinal mobility and were physically fit and active. Recently, a prospective study showed a clear relationship between loss of bone mass and inflammatory activity in early AS measured according to biological parameters, with no association between reduction in bone mass, vertebral mobility, and daily physical activity [29]. Our study group linked high disease activity in AS with altered vitamin D metabolism and increased bone resorption [30, 31] and reported a correlation between plasma tumor necrosis factor alpha, insulin-like growth factor 1, biochemical markers of bone metabolism with markers of inflammation and disease activity and clinical manifestations in AS [32].

To summarise: various factors such as treatment, hormone disorders, and decreased mobility or physical activity may contribute to the development of osteopenia/osteoporosis in AS. Lately it was shown that the vitamin D receptor gene may be involved in bone mineral density differences, bone metabolism, and inflammatory processes in AS [33].

Hitherto, examinations of bone density in AS were carried out mainly with projective methods (single and dual photon absorptiometry). Consequently, a decrease in bone density was observed up to the occurrence of ankylosations, with increased bone density in pronounced ankylosing stages [3, 23, 26]. One of the disadvantages of these methods is that there is no separate measurement of spongiosa and cortical bone. Since osteoporosis is accompanied initially by spongiosa and only later involves cortical bone loss, selective measurement of spongiosa and cortical bone was taken with SE-QCT along with the DXA determination of bone density in this particular study.

A considerable decrease in bone density could be verified in the initial and advanced stages of AS. Correspondingly, in SE-QCT during the initial stage of sacroiliitis, there was evidence of decreased mineral salt content of spongiosa with a further constant decrease in advanced stages of the disease. With projective methods, erroneously high values are often obtained following degenerative changes, scoliosis, anulus ossification, and calcification/sclerosis outside the bone [9, 34]. Thus, at present the most reliable means of verifying beginning and manifest spongiosa loss is the SE-QCT method.

In the studies conducted here, reduction of cortical bone was evident parallel to spongiosa loss and then, in advanced stages, there was a tendency for it to increase. Since cortical bone and spongiosa decrease in parallel up to stages of ankylosation, osteoporosis can not be seen solely as a result of static relief due to the thickening of the cortical bone. Possible immobilisation, especially in the initial phase, is also not an explanation for osteoporosis. This is a possible indication of the involvement of humoral factors in the occurrence of osteoporosis.

Kessler-Leonhardt and Droste [3] were also able to verify with X-rays and DXA measurements an axial osteopenia/osteoporosis in over 50% of young AS patients. In addition, they found vertebral fractures in 18–19% of the osteopenia patients, or 15.3% of the entire study group of 72 patients. Surprisingly, there were no neurological symptoms accompanying the complications of vertebral body fractures, and in as many as 50% of the patients there were no clinical symptoms. For nine patients in the present study, there were radiological indications of vertebral body fractures with no clinical signs of neurological deficit. A clearly pathological sclerotisation of the spongiosa was noticeable in some patients in the final stage of the disease, with predominantly mixed changes on the spine, whereas in the purely syndesmophytic type of development there were usually obvious signs of degenerative spongiosa. These findings should be further differentiated and narrowed down in future studies in order to provide answers to the following questions:
  1. 1.

    Are vertebral-body fractures more common in the syndesmophytic type of AS? (According to Kessler-Leonhardt and Droste [3], over 80% of the patients with vertebral body deformations had syndesmophytes, and osteoporosis was observed in 70–90%; but vertebral fractures were also present in cases of mild osteopenia.)

  2. 2.

    Does the so-called mixed form possibly protect against osteoporotic fractures?

  3. 3.

    Is there the likelihood of a so-called fracture threshold?

  4. 4.

    Are neurological complications more likely to be found in advanced stages of AS?


In summary, it is interesting that the study presented here verifies a bone density decrease already in the initial stage of AS and continuing in advanced stages as well, not only in the region of the spine but also on the femoral neck. Histological examinations point to an osteoclast/osteoblast imbalance [3, 35], which would indicate the prescription of antiresorptive therapy with a biphosphonate. An initial study has shown its positive effects on bone density [36]. New data suggest that the benefit of anti-tumour-necrosis-factor alpha therapy on bone mineral density in patients with spondylarthropathy may be due to an uncoupling effect on bone cells [37].

Copyright information

© Springer-Verlag 2004