Article Text

This article has a correction. Please see:

Download PDFPDF

Concise report
Bone loss in very early inflammatory back pain in undifferentiated spondyloarthropathy: a 1-year observational study
  1. G Haugeberg1,2,
  2. A N Bennett3,
  3. D McGonagle4,
  4. P Emery4,
  5. H Marzo-Ortega4
  1. 1Norwegian University of Science and Technology, Trondheim, Norway
  2. 2Department of Rheumatology, Sørlandet Hospital, Kristiansand, Norway
  3. 3The Academic Unit of Musculoskeletal Disease, Medical Defence Rehabilitation Centre, Headley Court, Chapel Allerton Hospital, Leeds, UK
  4. 4Section of Musculoskeletal Disease, Leeds Institute of Molecular Medicine, Chapel Allerton Hospital, University of Leeds, Leeds, UK
  1. Correspondence to Dr H Marzo-Ortega, Section of Musculoskeletal Disease, Leeds Institute of Molecular Medicine, University of Leeds, Chapel Allerton Hospital, Chapeltown Road, Leeds LS7 4SA, UK; medhmo{at}


Objective Bone loss in patients with inflammatory back pain (IBP) suspicious of early undifferentiated spondyloarthropathy is poorly defined. The aim of this study was to examine changes in bone mineral density (BMD) at the hip, lumbar spine and hand in patients with early IBP and to look for possible biomarkers associated with this change.

Methods In 30 patients with early IBP, clinical data were collected and BMD assessed using dual energy x-ray absorptiometry at baseline, 6 and 12 months. Further imaging performed included MRI of the sacroiliac joints (SIJs) and spine at baseline and x-rays of the SIJs at baseline and after 8 years.

Results After 12 months no significant reduction in hip, spine and hand BMD was seen at the group level. However, hip bone loss was found to be associated with raised baseline C-reactive protein levels, baseline MRI bone marrow oedema of the SIJs and the presence of radiographic sacroiliitis after 8 years. No association was found with change in spine and hand BMD.

Conclusion Systemic bone loss in the hip is an early feature of the inflammatory disease process in patients with IBP in undifferentiated spondyloarthropathy and is related to disease activity. These data highlight the importance of aggressive intervention in the early stages of disease in undifferentiated spondyloarthropathy.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Low bone density at the hip and spine is a feature of established or radiographically evident axial spondyloarthropathy (SpA) or ankylosing spondylitis (AS).1 There is evidence that bone loss in SpA occurs mainly in patients with persistently active inflammation.2

Inflammatory back pain (IBP) is the hallmark of incipient SpA. Following the rheumatoid arthritis (RA) analogy where bone loss is known to occur early, even in the undifferentiated state,3 we would expect bone loss related to inflammation in SpA to occur early in the disease process. To our knowledge, bone loss in patients with IBP suspicious of early axial SpA without radiographic evidence of AS has not been previously studied.

The aim of this study was to examine change in bone mineral density (BMD) assessed at the hip, spine (L2–4) and hand in patients with very early IBP, and to explore any possible associations with biomarkers of disease activity such as C-reactive protein (CRP), MRI and the future development of radiographic sacroiliitis which is a feature still mandatory for the diagnosis of AS.4


At baseline, all patients had a clinical history of low IBP (<2 years duration), normal or equivocal plain x-rays of the sacroiliac joints (SIJs) and a diagnosis that was highly suspicious of SpA. A thorough description of the patient cohort together with MRI and radiographic characteristics has already been reported elsewhere.5 6 In the present study we present data on 30 patients with IBP (50% men) in whom BMD measurements were performed at baseline and at 6- and 12-month follow-up visits, with demographic and clinical data collected at the same time points. At 8-year follow-up, eight patients had evidence of radiographic sacroiliitis (defined as grade 2 bilaterally or grade ≥3 unilaterally), allowing a diagnosis of AS to be made.4 None of the female patients was postmenopausal and no patients received treatment for osteoporosis such as bisphosphonates during the study period.

BMD was measured at both hips, spine (L2–4, anterior posterior view) and both hands by one technician using the same dual energy x-ray absorptiometry (DXA) equipment (Lunar Expert, Madison, Wisconsin, USA). In vivo short-term precision expressed as a percentage coefficient of variation (CV) was 2.75% at the femoral neck, 1.43% at total hip, 2.33% at spine L2–4 and 1.07% at the hand. Long-term spine phantom CV was 0.80%. For hip and whole hand, we used the mean values from the left and right sides.

Statistical analysis

Paired and independent two-tailed Student t tests were used for group comparisons for variables with normal distribution and the two-tailed Mann–Whitney U test was used for skewed variables. Pearson correlation (Pc) coefficients were applied for normally distributed variables and Spearman correlation (Sc) coefficients for non-normally distributed variables. p Values ≤0.05 were considered statistically significant. All analyses were performed with Statistical Package for Social Sciences (SPSS) program Version 16.0 (SPSS Inc, Chicago, Illinois, USA).


Patient characteristics

The baseline characteristics of the study patients are shown in table 1. Most of the patients had MRI evidence of SIJ bone marrow oedema (BMO) at baseline consistent with active sacroiliitis (n=29, 96.7%). Applying the recently published Assessment of SpondyloArthritis International Society criteria for axial SpA7 retrospectively, 25 patients would have met the imaging arm of the criteria. The other five had a disease duration of <3 months at inclusion and, as such, would have been ineligible. Nevertheless, they all had other associated features which placed them within the SpA spectrum.8

Table 1

Baseline characteristics of 30 patients with very early inflammatory back pain

No differences in baseline characteristics (table 1) were seen between patients defined as having low and high disease activity (CRP cut-off 10 mg/l), apart from disease duration (median 42 weeks (range 2–260) vs 12 weeks (range 2–76), p=0.012).

Imaging characteristics

All the patients were referred for the study on the basis of symptoms of IBP with normal or equivocal x-rays of the SIJs as reported by the referring clinical or local radiologist. However, x-rays were only formally scored at the time of follow-up 8 years later. At this stage, five patients were then reclassified as fulfilling modified New York criteria4 at baseline, the majority with borderline grade 2 changes. This reflects real-life experience as the interscorer reliability for the assessment of SIJ x-rays is known to be poor.9 For the 8 patients with and the 22 patients without radiographic sacroiliitis after 8-year follow-up, no significant difference at baseline was found between the two groups apart from CRP levels (median 22 mg/l (range 5–65) vs 7 mg/l (range 5–57), p=0.032) and BMO score at the SIJs (median 7 (range 3–19) vs 2 (range 0–14), p=0.007). All patients with radiographic changes at 8 years had MRI evidence of sacroiliitis at baseline.

Bone loss and its associations

No change in mean BMD was seen at 6- or 12-month follow-up at the femoral neck (−0.14%, p=0.76 and −0.63%, p=0.34), total hip (0.14%, p=0.74 and −0.11%, p=0.82), lumbar spine L2–4 (−0.47%, p=0.53 and −0.57%, p=0.51) and hand (0.03%, p=0.92 and 0.18%, p=0.58).

When patients were stratified in post hoc analysis according to the mean CRP value (calculated from baseline, 6-and 12-month values; cut-off 10 mg/l), a difference was found between patients with low CRP levels (normal value <10 mg/l, n=19) and high CRP levels (≥10 mg/l, n=11) at the femoral neck (0.78% vs −3.22%, p=0.011) and total hip (1.14% vs −1.76%, p=0.003; figure 1). The same pattern was observed at the lumbar spine (0.52% vs −2.48%, p=0.10), with only minor differences seen between the two groups at the hand (0.37% vs −0.15%, p=0.44). A correlation was also found between baseline CRP levels and change in BMD at the femoral neck (Sc=−0.452, p=0.014) and total hip (Sc=−0.458, p=0.013).

Figure 1

Mean percentage change (with 95% CI) in bone mineral density (BMD) over 12 months in patients with early inflammatory back pain grouped according to mean 12-month C-reactive protein (CRP) level ≥10 mg/l or <10 mg/l. Values shown are baseline BMD with SD and BMD change (Δ% BMD) with SEM for the two groups at the different measurement sites. p Values from paired Student t test.

For MRI of the SIJs, an association was found between the severity of baseline BMO and the 12-month change in BMD at the femoral neck (Sc=−0.417, p=0.022) and hand (Sc=−0.385, p=0.036) but not at the hip (Sc=−0.208, p=0.27) or lumbar spine (Sc=−0.035, p=0.85). For spine MRI, no association was found with change in BMD.

Interestingly, a difference in the change in BMD was seen at the total hip (−1.61% vs 0.75%, p=0.03) and femoral neck (−2.56% vs 0.11%, p=0.05) but not at the lumbar spine (−0.71% vs −0.52%, p=0.91) or the hand (−0.65% vs 0.51%, p=0.10) between patients who developed (n=8) and those who did not develop (n=22) radiographic sacroiliitis after 8 years. In addition, the mean CRP level after 12 months of follow-up was higher in patients who developed radiographic signs of sacroiliitis after 8 years (15.8 mg/l vs 9.5 mg/l, p=0.04).

For the other demographic and disease variables listed in table 1, no association was found with change in BMD during follow-up.


The current data support the view that generalised bone loss in patients with AS/SpA starts early in the disease process and is dependent on active inflammation. Indeed, both CRP—which is a fluctuating marker of inflammation—and radiographic sacroiliitis—which can be considered a cumulative marker of inflammatory disease activity—were found to be associated with bone loss in our cohort of patients with early IBP. Furthermore, an association was also found between the severity of MRI-determined BMO at the SIJs at baseline and the change in generalised BMD at the femoral neck at 12 months.

In established AS, an increased rate of bone loss at the spine and/or hip has previously been reported in patients with persistent active disease, as defined by raised erythrocyte sedimentation rate and/or CRP, but not in patients with inactive disease.2 10 Bone loss in patients with AS is also associated with high levels of biochemical markers of bone resorption.11 Likewise, raised levels of interleukin 6 and tumour necrosis factor α (TNFα) have been found in patients with SpA with active disease.2 12 These are both proinflammatory cytokines involved in bone resorptive processes acting via the osteoprotegerin/receptor activator for nuclear factor κB ligand pathway. TNF-blocking agents have been shown not only to reduce signs and symptoms of disease activity in SpA,13 but also to arrest hip and spine bone loss.14 15

An interesting observation from our study is that bone loss was seen mainly at the hip and spine but not at the hands in patients with IBP with high CRP levels, which confirms previous observations.16 This is in accordance with the clinical examination as none of our patients had evidence of hand synovitis. Furthermore, it supports the hypothesis that bone loss at the spine and hip in patients with SpA is the result of a more localised effect of inflammatory activity and/or immobility, similar to early RA where a high rate of hand bone loss occurs in patients with hand involvement.3

In this study we only found an association between baseline BMO on MRI at SIJs and the 12-month change in generalised BMD at the femoral neck. We expected to find a stronger association between BMO on MRI both at the spine and hip, as MRI-determined BMO is a recognised diagnostic marker of disease activity in early SpA and is responsive to potent anti-inflammatory treatment with TNF blockers.17 One explanation may be that the presence of BMO and the intensity of the MRI signal do not correlate with the strength of the underlying pathophysiological mechanisms of inflammation in SpA.18 The most likely explanation, however, may be related to the small number of patients included in our study.

In SpA, an increased risk of vertebral fracture but not of non-vertebral fracture has been reported.19 This risk is associated with reduced hip BMD,20 but may also be explained by impaired bone material properties and increased stiffness of the spine caused by syndesmophytes and ankylosis which make the spine in these patients more susceptible to fracture, independent of the BMD level.21

In conclusion, this study provides evidence that bone loss seen in patients with undifferentiated axial SpA is a result of systemic inflammation and starts early in the disease process. This emphasises the need for early intervention in AS, although further studies are warranted to clarify whether this bone loss is associated with a risk of fracture.


The authors thank Mrs Sheena Stewart for her help and expertise with DXA scanning.



  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the Leeds Teaching Hospitals NHS Trust.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • Miscellaneous
    BMJ Publishing Group Ltd and European League Against Rheumatism