Article Text

Download PDFPDF

Concise report
Zoledronic acid does not reduce MRI erosive progression in PsA but may suppress bone oedema: the Zoledronic Acid in Psoriatic Arthritis (ZAPA) Study
  1. Fiona McQueen1,2,
  2. Richard Lloyd1,
  3. Anthony Doyle1,2,
  4. Elizabeth Robinson1,
  5. Maria Lobo2,
  6. Melissa Exeter2,
  7. William J Taylor3,4,
  8. Peter Jones1,
  9. Ian R Reid1,2,
  10. Nicola Dalbeth1,2
  1. 1Department of Molecular Medicine and Pathology, University of Auckland, Auckland, New Zealand
  2. 2Auckland District Health Board, Auckland, New Zealand
  3. 3University of Otago, Wellington, New Zealand
  4. 4Hutt Valley District Health Board, Lower Hutt, New Zealand
  1. Correspondence to Professor Fiona McQueen, Department of Molecular Medicine and Pathology, Faculty of Medical and Health Sciences, University of Auckland, 85 Park Rd, Grafton, Auckland, New Zealand; f.mcqueen{at}


Background The effect of zoledronic acid (ZA) on articular bone in patients with psoriatic arthritis (PsA) was investigated using MRI.

Methods Patients with erosive PsA were randomised to receive 3-monthly infusions of ZA or placebo for 1 year. An additional ‘tests alone’ group received no infusions. Clinical assessments and MRI scans were performed at baseline and 1 year.

Results Paired 1.5T MRI scans were available in 22 patients including 6 who received ZA and 16 who did not (non-ZA = 6 placebo + 10 ‘tests alone’ patients). The Disease Activity Score (28 swollen and tender joints, C reactive protein fell over 12 months to a greater degree in patients on ZA than in non-ZA patients (−1.6 vs −0.3, p=0.023). The MRI bone oedema score decreased in the ZA group (15.5 to 8.5) but increased in the non-ZA group (14.0 to 18.0) (p= 0.0056) with regression of bone oedema at 13.5% of sites in ZA patients vs 1.3% in non-ZA patients (p = 0.0073) and progression in 1.3% of sites in ZA patients vs 6.9% in non-ZA patients (p = 0.072). There was no difference between groups in change in MRI erosion score.

Conclusions In this pilot study ZA reduced the progression of MRI bone oedema, indicating probable suppression of osteitis concordant with reduction in clinical measures of disease activity.

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.


Psoriatic arthritis (PsA) is associated with characteristic joint pathology including bone erosion, proliferation and resorption.1 Bone lysis may be severe and lead to digital shortening and arthritis mutilans.2 MRI allows bone pathology to be studied in life without surgical resection. In rheumatoid arthritis (RA), MRI bone oedema occurs in 40–60% of patients and is strongly associated with the development of bone erosion.3 In RA bone specimens resected at joint replacement surgery, osteitis has been observed where preoperative MRI scans showed bone oedema.4 These regions of bone inflammation contain B and T lymphocytes, plasma cells and macrophages closely apposed to osteoclasts. Comparative MRI/histological studies of bone have not been performed in PsA, but bone oedema has been observed at periarticular and entheseal locations.5

Osteoclasts are known to be intimately involved in the process of bone erosion in RA6 and studies by Ritchlin et al7 have suggested that they are also key effector cells mediating bone resorption and erosion in PsA. Tumour necrosis factor α (TNFα) and Receptor Activator for Nuclear Factor κ B Ligand (RANKL) stimulate osteoclast activation in PsA, mediating bone damage. Zoledronic acid (ZA) suppresses osteoclast recruitment, differentiation and function and also promotes apoptosis. It has been proposed that bisphosphonates act as anti-erosive agents in RA,8 but their effect has not yet been studied in PsA. We investigated the effect of ZA in patients with PsA using MRI to assess bone erosion, oedema and proliferation.


Patients and clinical assessments

With the approval of the New Zealand Multiregion Ethics Committee, 28 patients with PsA9 were recruited from Auckland, Rotorua and Wellington regions from 2005 to 2007. Enrolment required peripheral x-ray erosions confirmed by a radiologist. Of the 28 patients, 15 participated in the randomised placebo-controlled trial in which patients received 4 mg zoledronic acid (ZA) intravenously 3-monthly for 1 year or placebo (infusions given over 2 h with normal saline). Twelve patients completed baseline (T0) and 12-month (T12) assessments and three withdrew. Because of media reports associating ZA with osteonecrosis of the jaw,10 13 patients did not wish to be exposed to ZA and were entered into a ‘tests alone’ group (no infusion, clinical and MRI assessments only). Of these, 10 completed the study. ‘Tests alone’ and placebo groups were combined to form a non-ZA control group. For the study/placebo group, all results were obtained double-blind and, for the ‘tests alone’ group, clinical assessments were unblinded (patients/investigators) but scans were read blinded by the radiologist. Baseline demographics, disease activity and x-ray scores are shown in table 1. Medications included methotrexate (7.5–20 mg/week) in 50% and 56% of ZA and non- ZA patients, respectively, and other disease-modifying antirheumatic drugs (DMARDs) in 33% and 31%, respectively, but none received anti-TNF therapy (unavailable in New Zealand at that time for PsA). Disease activity assessments included swollen/tender joint counts (76/78), visual analogue scores (VAS) for pain and well-being, C reactive protein (CRP) and DAS28-CRP scores. Health Assessment Questionnaire (HAQ), physical function Short Form-36 (PF-SF36) and Psoriasis Area and Severity Index (PASI) scores were also obtained.

Table 1

Demographics, baseline and change in disease activity in ZA and non-ZA patients*

MRI scans

MRI scans of the dominant wrist and fingers were obtained at T0 and T12 in 22 patients using a Siemens Magnetom Avanto 1.5 Tesla (T) scanner with dedicated wrist coil, regardless of whether these regions were involved clinically or radiographically. Sequences and acquisitions have been reported elsewhere.11 Scans were scored in known sequence by one trained reader (AD) for bone erosion, oedema and proliferation at 38 sites using the PsA MRI score (PAMRIS).12 Reliability was established as moderate to high when tested in a two-reader setting.11

Figure 1

Bone oedema sites showing progression (green), regression (red) and no change (yellow) over 1 year. (A) A patient with psoriatic arthritis (PsA) receiving zoledronic acid. Change in bone oedema could only be assessed at 34 of a possible 38 sites due to inadequate imaging of the proximal and distal regions of the fifth proximal interphalangeal joint and the proximal and distal regions of the second distal interphalangeal joint. In this patient there was progression of oedema at two sites (5.9%), regression at five sites (14.7%) and no change at 27 sites (79.4%) (B) Assessment of change in bone oedema in a control patient with PsA over 1 year made at 38 sites. In this patient there was progression at seven sites (18.4%), regression at two sites (5.3%) and no change at 29 sites (76.3%).

Statistical analyses

Median (range) values are reported for disease activity measures and MRI scores. Within-person changes in total scores were calculated for each patient. Percentage regression and progression for oedema, erosion and proliferation scores were calculated for each patient using a site-by-site examination (N=38 bone sites). Wilcoxon paired rank tests and Mann–Whitney U tests were used to compare changes between T0 and T12 scores for the two groups.


Change in disease activity over 12 months: ZA group vs non-ZA group

MRI data were available in 6 patients with ZA and 16 non-ZA patients (6 placebo + 10 ‘tests alone’). Table 1 summarises baseline clinical features, disease activity and change in disease activity for the two groups of patients. Clinical disease activity fell after 12 months in patients with ZA patients compared with non-ZA patients, as revealed by change in swollen joints (−5.0 vs −0.5 in ZA and non-ZA patients, p=0.055) and change in tender joints (−19.0 vs −1.0 in ZA and non-ZA patients, p=0.01). This was matched by the change in DAS28-CRP3v (−1.6 vs −0.3 in ZA and non-ZA patients, p=0.01). However, clinical assessments were performed unblinded in the non-ZA ‘tests alone’ group.

Changes in MRI scores over 12 months

Median MRI erosion scores increased over 12 months but did not differ between groups (35.0–39.5 in ZA group vs 44.5–45.5 in non-ZA group; table 2). However, bone oedema scores decreased in patients with ZA (from 15.5 to 8.5) but increased in non-ZA patients (from 14.0 to 18.0) (p=0.006). A site-by-site analysis (38 sites per patient) revealed regression of bone oedema at 13.5% of sites in patients with ZA compared with 1.3% of sites in non-ZA patients (p=0.007) and progression at 1.3% of sites in ZA patients compared with 6.9% of sites in non-ZA patients (p=0.072). For bone erosion there was progression at 5.5% of sites in ZA patients compared with 5.3% in non-ZA patients (p=0.60) and regression in 2.6% and 0% of sites, respectively (p=0.53). Bone proliferation did not differ between groups.

Table 2

Changes in MRI bone erosion, oedema and proliferation in ZA versus non-ZA groups

Figure 1 shows changes in MRI bone oedema in a ZA patient compared with a non ZA compared with a non-ZA patient.


We present the first study to investigate the efficacy of ZA in modifying bone pathology in PsA. This powerful nitrogen-containing bisphosphonate is preferentially deposited in bone beneath osteoclasts.13 Its mechanism of action depends on interference with the mevalonate pathway by inhibiting farnesyl pyrophosphate, preventing the attachment of a lipid anchor to cell membranes. This inhibits osteoclast recruitment, differentiation, formation of the ruffled border and acid production and induces osteoclast apoptosis.13 Ritchlin et al documented RANKL and TNFα-activated osteoclasts within subchondral bone in the joints of patients with PsA as well as circulating mononuclear cells with features of osteoclast precursors (OCPs), especially in those with radiographic erosions.7 We have restudied this recently in our own cohort and found that the Sharp van der Heijde score for radiographic joint damage correlated with the percentage of peripheral blood OCPs.14 Targeting osteoclasts with bisphosphonates has therefore been investigated as an anti-erosive strategy. Using a TNF-transgenic mouse model, Redlich et al15 showed that pamidronate plus osteoprotegerin (OPG) were effective in preventing TNFα-mediated bone destruction by reducing osteoclast numbers in inflamed joints. Bone destruction was also reduced in mice with collagen-induced arthritis following administration of ZA.16 Jarrett et al studied ZA in patients with RA and showed that fewer bones developed new MRI erosions in ZA-treated patients while MRI bone oedema also progressed less frequently.8 Maksymowych et al showed that pulse intravenous pamidronate was associated with regression of bone oedema (indicated by reduced contrast enhancement of the bone marrow) in patients with spondyloarthropathy with refractory peripheral arthritis.17 Interestingly, there was an associated reduction in tender joint count and CRP, as reported here.

In the current study we found no evidence that ZA reduced the progression of MRI erosions over 12 months, but it did appear to reduce bone oedema both as a total score and when progression or regression at individual regions were examined. There was a concomitant improvement in clinical disease activity in ZA-treated patients, but assessments were unblinded for some non-ZA patients so these results are less reliable. The discrepancy observed between effects on bone oedema and erosion is surprising as bone oedema is likely to lead to erosion in PsA as in RA,18 and concordance would be expected. This could reflect a type II error as many of our patients had severe erosive disease with a significant proportion having arthritis mutilans (reflecting the highly selected group attending tertiary referral centres from which they were recruited). Minor progression on this background could be difficult to detect with MRI, especially where bone sclerosis or arthrodesis have occurred. Changes in bone oedema are also likely to occur earlier and be captured over the course of 1 year. Marzo-Ortega et al noted a marked reduction in bone oedema after 20 weeks of infliximab treatment in patients with PsA,19 and Andarajah et al noted a change in bone oedema over a 6-month period when examining the effect of etanercept.20 However, a 1-year time frame might be too short to examine erosive progression and future studies need to be designed accordingly.

Our results raise questions about the mechanism of action of ZA on bone in this disease. A reduction in bone oedema suggests a reduction in osteitis.4 As the primary effect of ZA is on osteoclasts, a reduction in these cells could lead directly to a reduction in bone oedema. However, suppression of other inflammatory cells might also be required before there was a sufficient decrease in the T2-weighted signal (proportional to water-containing cell bodies) to reduce MRI bone oedema scores. Thus, downregulation of other inflammatory cells might also be implicated. Bugatti et al observed a close relationship between inflammatory cell and osteoclast numbers in regions of RA osteitis, suggesting that mutual regulation may occur.21 There might also be bisphosphonate-mediated toxic effects on inflammatory cells, as shown in vitro where pamidronate suppressed production of proinflammatory cytokines in cultured macrophage cell lines.22 Studies of a model simulating osteonecrosis of the jaw have demonstrated substantial toxicity from prolonged contact of non-bone cells with ZA-labelled bone (Professor I Reid, personal communication).

In summary, we present the findings of a pilot study investigating the effect of ZA on articular bone in PsA using MRI scanning. We did not find evidence for an influence on erosive progression, but bone oedema was reduced in ZA-treated patients concordant with a reduction in disease activity. Thus, ZA could represent an alternative therapy for patients with PsA where the disease is not adequately controlled with traditional or biological therapies or where anti-TNF agents are contraindicated. Larger studies are indicated to determine whether ZA could be a useful addition to the therapeutic armamentarium in this disease.


The authors gratefully acknowledge the assistance of technical staff at the Centre for Advanced MRI at the Auckland School of Medicine, University of Auckland in performing the MRI scans for this study. The authors also thank the Auckland rheumatologists who referred patients for this study.



  • Funding This investigator-initiated study was funded by the Auckland Medical Research Foundation. Novartis supplied the study drug and an unrestricted educational grant. Richard Lloyd (biomedical science student) was sponsored by this grant for a summer studentship.

  • Ethics approval This study was conducted with the approval of the New Zealand Multiregion Ethics Committee.

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