Skip Navigation


Rheumatology Advance Access originally published online on December 7, 2006
Rheumatology 2007 46(4):699-702; doi:10.1093/rheumatology/kel401
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
46/4/699    most recent
kel401v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (1)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Nistala, K.
Right arrow Articles by McDonagh, J. E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Nistala, K.
Right arrow Articles by McDonagh, J. E.
Related Collections
Right arrow Juvenile Idiopathic Arthritis
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2006. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Clinical assessment and core outcome variables are poor predictors of hip arthritis diagnosed by MRI in juvenile idiopathic arthritis

K. Nistala, J. Babar, K. Johnson, P. Campbell-Stokes, K. Foster, C. Ryder and J. E. McDonagh1

Birmingham Children's Hospital and 1Institute of Child Health, Diana, Princess of Wales Children's Hospital, Steelhouse Lane, Birmingham B4 6NH, UK.

Correspondence to: K. Nistala, Paediatric Rheumatology Department, Birmingham Children's Hospital, Steelhouse Lane, Birmingham B4 6NH, UK. E-mail: K.Nistala{at}ich.ucl.ac.uk


   Abstract
Top
 Abstract
Introduction
 Methods
 Analysis
 Results
 Discussion
 Acknowledgements
 References
 
Objectives. To compare the diagnostic performance of clinical assessment against magnetic resonance imaging (MRI) diagnosed hip arthritis in a juvenile idiopathic arthritis (JIA) population. To determine the clinical and serological predictors of MRI diagnosed hip arthritis.

Methods. A total of 34 JIA patients with established disease (mean disease duration 6.3 yrs) had their hip MRIs scored for features of active hip arthritis and hip damage. Results were compared with clinical variables (disease subtype, history of hip pain, core outcome variables (COV)) and the clinician's assessment of active hip arthritis.

Results. MRI features of active hip arthritis were found in 45 hips (70%) and hip damage in 36 hips (56%). Clinical assessment had fair agreement with MRI scoring of active arthritis in patients with disease duration <4 yrs ({kappa} score 0.38, P = 0.045). Clinical assessment had a sensitivity of 25.7% and specificity of 91% for detecting MRI diagnosed arthritis. Of the core outcome variables only erythrocyte sedimentation rate predicted inflammation detected on MRI (r = 0.44, P = 0.014).

Conclusions. The association between the clinician's assessment, core outcome variables and MRI findings in this study was limited. This indicates that clinical and laboratory findings are inadequate diagnostic tools for the assessment of hip arthritis when compared with MRI as the gold standard.

KEY WORDS: Coxitis, Magnetic resonance imaging, Juvenile idiopathic arthritis, Hip joint, Clinical examination


   Introduction
Top
 Abstract
 Introduction
Methods
 Analysis
 Results
 Discussion
 Acknowledgements
 References
 
Hip involvement in juvenile idiopathic arthritis (JIA) is common [1] and is a cause of significant functional impairment [2]. Long-term follow-up shows that total hip replacements are carried out in 26–44% of JIA patients, most commonly within the first 10 yrs of disease [3]. The development of hip arthritis (coxitis) is likely to be an indication for treatment with disease modifying agents, such as methotrexate and anti-tumour necrosis factor (TNF) agents. In some cases, injection of intra-articular steroids to the hip can bring about prolonged remission [4]. Unfortunately, the clinical detection of coxitis is difficult as inflamed synovium cannot be directly palpated. As a result, subclinical coxitis can lead to joint damage [5].

For the purposes of classification, coxitis is defined clinically as ‘restricted movements with joint pain’ [6], but these clinical features are not specific for active inflammation and may occur with previous joint damage. Ultrasound is often used to assist the diagnosis by confirming the presence of an effusion, but it remains less sensitive than contrast enhanced MRI in detecting features of active arthritis [7]. It is impractical to perform MRIs in all children and young people with JIA as there are issues of cost, availability and the need for sedation. So it is important to know the clinical context in which MRI is most likely to be useful. We hypothesized that clinical assessment would be poor at detecting coxitis in patients with co-existing hip damage. To test our hypothesis we compared the clinician's clinical assessment with MRI as the gold standard for active hip disease in a cohort of patients with established JIA.


   Methods
Top
 Abstract
 Introduction
 Methods
Analysis
 Results
 Discussion
 Acknowledgements
 References
 
We retrospectively reviewed the last 50 hip MRI scans requested on children and young people with JIA at our institution from 9 June 2003 to 25 January 2005. Eligibility criteria for patients included a diagnosis of JIA as defined by the revised International league of Associations for Rheumatology criteria [6] and disease duration >6 months. Symptoms of hip pain and the clinician's assessment of hip arthritis (active, inactive or unsure), active joint count, demographic data and erythrocyte sedimentation rate (ESR) were collected from case notes. Core outcome variables (COV) are routinely recorded for every patient at every clinic visit and include the marking of the active and restricted joints on a skeletal diagram. COV [8] included physician's global assessment of overall disease activity (VAS-PGA) assessed using a 100 mm visual analogue scale, the childhood health assessment questionnaire (CHAQ), as modified for use in the UK [9] and which also included 100 mm visual analogue scales for overall patient well-being (VAS-global). All patients are examined by either a consultant paediatric rheumatologist or an experienced paediatric rheumatologist in training, during their routine clinic visits at the host institution.

A total of 19 hip MRIs were reported blind to clinical details simultaneously by two experienced paediatric radiologists (K.J., K.F.). Any disagreement was resolved by consensus. A further 15 hip MRIs were independently reported by the same radiologists to assess inter-observer agreement. Hips were scored for signs of active synovitis with 1 point each for synovial effusion, gadolinium-enhancing synovium >2 mm, peri-articular bone oedema. Damage was scored with 1 point each for bony erosions, cartilage loss, bony remodelling and acetabular protrusion. The scoring system was developed for this study, and is at present unvalidated, being based on previous literature [10, 11] and the paediatric radiologists clinical experience. MRI images were considered to be abnormal if the respective activity and damage scores were >0.

All scans were performed on a Siemens (Erlanger, Germany) 1.5 Tesla symphony scanner. The sequences obtained were coronal T1, T2, short tau inversion recovery and axial proton density of the hips and bony pelvis. All children had post gadolinium (0.01mmol/kg) DTPA fat saturated coronal T1 weighted sequences.


   Analysis
Top
 Abstract
 Introduction
 Methods
 Analysis
Results
 Discussion
 Acknowledgements
 References
 
The hips of each individual patient were considered to be independent when comparing lateralizing variable such as clinician's assessment and pain with MRI scores. Non-lateralizing variables such as disease duration and ESR were compared with the patient's combined hip scores. The association between total MRI scores and clinical variables was tested using Pearson's correlation. Individual hip MRI scores were compared with clinician defined active and inactive groups using the Mann–Whitney test. Kappa scores were used to assess inter-observer variability and concordance between clinician's assessment and MRI results [12]: Clinical–radiological correlation was not made if the interval between clinical assessment, and MRI was >3 months as changes in disease activity may have occurred (three patients). Chi-squared test for trend was used to elucidate the effect of damage on the concordance between clinician's assessment and MRI activity scores [13]. For this analysis hips classified by clinicians as ‘unsure’ were excluded. P ≤ 0.05 was defined as being statistically significant in this study. Computer analysis was carried out using SPSS software version 12, from SPSS Inc., Chicago, IL, USA.


   Results
Top
 Abstract
 Introduction
 Methods
 Analysis
 Results
Discussion
 Acknowledgements
 References
 
Sixteen patients had repeat MRIs, which were excluded. Thirty-four MRI scans (68 hips) were analysed. Frequency of JIA subtypes were as follows: twelve enthesitis-related arthritis, nine polyarthritis, seven systemic onset, four extended oligoarthritis and two psoriatic arthritis. Mean patient age 14.4 yrs (range = 4.3–19.7 yrs), disease duration 6.3 yrs (0.8–13.6 yrs), ESR 23 mm/h (2–96) and active joint count of 2 (2–13), CHAQ 0.84 (0–1.88), physician's global (VAS-PGA) 14 mm (0–85), patient well-being (VAS-global) 30 mm (0–81). Eight patients were receiving etanercept, 13 methotrexate and four oral prednisolone.

MRI examination
Hip effusions and synovial enhancement were present more often than bone oedema (36, 33 and 15%, respectively). Of all hips scanned, 30% had no MRI features of disease activity, and 15% scored 3 points. Of all hips, 44% had no evidence of damage, and 28% scored 3. None of the patients had MRI features of acetabular protrusion. MRI activity and damage scores did not correlate. There was substantial inter-observer agreement for individual aspects of the activity and damage scoring systems ({kappa} = 0.76–1.0).

Comparison of MRI findings and clinician's assessment
There was a limited relationship between clinician's assessment and MRI activity scores (Fig. 1). MRI scores were higher in patients thought to have clinically active hips vs inactive although this did not achieve statistical significance (1.9 vs 1.2, P = 0.066). Agreement between MRI activity and clinical assessment was fair in cases with arthritis less than 4 years from diagnosis ({kappa} score 0.38, P = 0.045). There was no agreement in longer standing disease ({kappa} score 0.02, P = 0.62).


Figure 1
View larger version (9K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 1. MRI activity scores compared with clinician's assessment. MRI activity for each hip was scored as 1 point each for bone oedema, enhancing synovium, effusion, maximum score = 3. Each triangle represents one hip. Filled triangles have evidence of damage on MRI (MRI damage scores >0). Horizontal lines represent mean MRI activity score.

 
We analysed the agreement between clinical assessment and MRI activity score according the severity of damage. Concordance between clinician and MRI result was highest in undamaged hips with agreement in 11 out of 18 cases. This worsened with increasing damage score ({chi}2Trend = 5.18, 1 df, P = 0.023) and in hips with a damage score of 3, there was agreement in only 4 of 16 cases.

If cases where the clinician was unsure are excluded, clinical examination has a sensitivity of 25.7% and specificity of 91% for detecting MRI-diagnosed arthritis.

Predictors of MRI activity score
COV, history of pain, disease subtype and duration of arthritis were compared with total MRI activity scores using multiple regression to assess their predictive value. In exploratory univariate analyses, only ESR had significant correlation with MRI scores (r = 0.44, P = 0.014, Fig. 2). Physician's global assessment and hip pain were associated with MRI activity score (P= –0.37 and 0.2, respectively) and using a cutoff of P < 0.2 these were selected for inclusion into standard multiple regression analysis. However, the latter two variables did not make a statistically significant contribution to a model based on ESR alone. A receiver operator curve was used to select an optimal cutoff as ESR >7 that predicts hip arthritis on MRI with a sensitivity of 72% and specificity of 80%.


Figure 2
View larger version (12K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 2. Correlation of total MRI activity score with ESR (mmHg). MRI activity was scored as 1 point each for bone oedema, enhancing synovium, effusion per hip, maximum score = 6.

 

   Discussion
Top
 Abstract
 Introduction
 Methods
 Analysis
 Results
 Discussion
Acknowledgements
 References
 
Over the last decade, there has been an increasing move towards earlier and more aggressive treatment of JIA with methotrexate and biological agents in the hope of preventing joint damage [14]. In the context of coxitis, decisions to escalate treatment may be limited because of the difficulties in confirming arthritis by clinical examination. In our study, we have shown that the clinician's assessment of active hip arthritis has only fair agreement with MRI-based diagnosis and this is limited to early disease. Clinicians misjudged coxitis on MRI when hips were more damaged and in particular underestimated MRI features of inflammation. We found that damage worsened with disease duration and this may account for the poor concordance between doctor and MRI results in patients with long-standing disease. It is possible that clinicians are incorrectly attributing their clinical findings of pain and restriction to pre-existing bony damage rather than disease activity. In contrast, clinicians were invariably correct in labelling hips without inflammation on MRI as inactive, as evidenced by the high specificity for clinician's assessment.

Recognizing the limitations of clinical examination, we questioned if other clinical or core outcome variables would better predict MRI changes. Our data from JIA patients with established disease suggests that from COV only ESR is predictive of active hip arthritis on MRI. Taken from a clinical perspective a highly elevated ESR is specific for hip inflammation, but a normal ESR does not exclude coxitis on MRI. MRI scores did not differ significantly between disease subtypes, but patients with systemic JIA had the highest mean scores, consistent with reported results by Argyropoulou et al. [11].

The MRI features of hip disease in JIA have been reported [15, 10, 11] but not standardized and do not currently play a role in the diagnostic criteria for JIA [6]. Our study piloted a scoring system for disease activity and damage on hip MRI that showed very good levels of inter-observer agreement. Damage and activity scores were independent suggesting that separate disease characteristics were being assessed. Our study did not include a healthy control group. Although normative data on paediatric hip MRI is available [16], formal validation of a scoring system will require control subjects to ensure that MR abnormalities are not being over-called.

Our study has the weakness of being a retrospective review although standardized clinical data recording systems mitigate against some of these flaws. The numbers scanned are small and are biased by being a cohort selected for MRI on clinical grounds and are, therefore, at increased risk of coxitis. The results may not be applicable to all JIA patients particularly those with less aggressive disease. Our findings of inflammation on hip MRI in the absence of clinical signs, are likely to reflect the high sensitivity of MRI, and are consistent with results from MRI of knee arthritis [17]. However, it is difficult to make treatment recommendations on MR findings alone as the long-term significance of these abnormalities is still unclear. Similar difficulties are faced in the MRI of sacroilitis in ankylosing spondylitis (AS) [18]. In AS, there is evidence that juxta-articular bony inflammation can be suppressed with anti-TNF agents [19] and prospective studies are underway to see if this favourably influences long-term outcome. Finally, as the use of MRI remains constrained by cost and the need for sedation in the paediatric age group, comparisons between MRI and more sensitive ultrasound techniques such as power Doppler would be important in the future.

In conclusion we have reported pilot data that show that clinical assessment of active hip arthritis has a limited relationship with MRI features of inflammation, particularly when there is co-existent damage. Of core outcome variables only ESR is of value in predicting MRI results. Our results highlight the contribution of MRI when there is clinical uncertainty between active and damaged hips. A prospective cohort study is now planned that will facilitate the formal validation of the scoring system for hip MRI abnormalities in JIA detailed here.

Formula


   Acknowledgements
Top
 Abstract
 Introduction
 Methods
 Analysis
 Results
 Discussion
 Acknowledgements
References
 
J.E.McD. is an ARC funded Clinical Senior Lecturer in Paediatric and Adolescent Rheumatology (www.arc.org.uk). P.C.S. is funded by the New Zealand Arthritis Foundation. The authors are indebted to the clinical staff in the paediatric rheumatology department for data collection and Carole Cummins (Birmingham Children's Hospital) and Tim Cole (Institute of Child Health, London) for statistical advice.

The authors have declared no conflicts of interest.


   References
Top
 Abstract
 Introduction
 Methods
 Analysis
 Results
 Discussion
 Acknowledgements
 References
 

  1. McDonagh JE. (2003) The hip joint in juvenile idiopathic arthritis. In Banta JV and Scrutton D (Eds.). Hip disorders in childhood.(MacKeith Press, London) pp. 130–45.
  2. Bekkering WP, ten Cate R, Suijlekom-Smit LW, et al. (2001) The relationship between impairments in joint function and disabilities in independent function in children with systemic juvenile idiopathic arthritis. J Rheumatol 28:1099–105.[Abstract/Free Full Text]
  3. Packham JC and Hall MA. (2002) Long-term follow-up of 246 adults with juvenile idiopathic arthritis: functional outcome. Rheumatology 41:1428–35.[Abstract/Free Full Text]
  4. Neidel J, Boehnke M, Kuster RM. (2002) The efficacy and safety of intraarticular corticosteroid therapy for coxitis in juvenile rheumatoid arthritis. Arthritis Rheum 46:1620–8.[CrossRef][Web of Science][Medline]
  5. Fedrizzi MS, Ronchezel MV, Hilario MO, et al. (1997) Ultrasonography in the early diagnosis of hip joint involvement in juvenile rheumatoid arthritis. J Rheumatol 24:1820–5.[Web of Science][Medline]
  6. Petty RE, Southwood TR, Manners P, et al. (2004) International League of Associations for Rheumatology classification of juvenile idiopathic arthritis: second revision, Edmonton, 2001. J Rheumatol 31:390–2.[Free Full Text]
  7. Soini I, Kotaniemi A, Kautiainen H, Kauppi M. (2003) US assessment of hip joint synovitis in rheumatic diseases. A comparison with MR imaging. Acta Radiol 44:72–8.[CrossRef][Web of Science][Medline]
  8. Giannini EH, Ruperto N, Ravelli A, Lovell DJ, Felson DT, Martini A. (1997) Preliminary definition of improvement in juvenile arthritis. Arthritis Rheum 40:1202–9.[Web of Science][Medline]
  9. Nugent J, Ruperto N, Grainger J, et al. (2001) The British version of the childhood Health Assessment Questionnaire (CHAQ) and the Child Health Questionnaire (CHQ). Clin Exp Rheumatol 19:S163–7.[Web of Science][Medline]
  10. Murray JG, Ridley NT, Mitchell N, Rooney M. (1996) Juvenile chronic arthritis of the hip: value of contrast-enhanced MR imaging. Clin Radiol 51:99–102.[CrossRef][Web of Science][Medline]
  11. Argyropoulou MI, Fanis SL, Xenakis T, Efremidis SC, Siamopoulou A. (2002) The role of MRI in the evaluation of hip joint disease in clinical subtypes of juvenile idiopathic arthritis. Br J Radiol 75:229–33.[Abstract/Free Full Text]
  12. Landis JR and Koch GG. (1977) An application of hierarchical kappa-type statistics in the assessment of majority agreement among multiple observers. Biometrics 33:363–74.[CrossRef][Web of Science][Medline]
  13. Fleiss JL. (1981) Statistical methods for rates and proportions.(Wiley, New York) pp. 144.
  14. Ramanan AV, Whitworth P, Baildam EM. (2003) Use of methotrexate in juvenile idiopathic arthritis. Arch Dis Child 88:197–200.[Abstract/Free Full Text]
  15. Kramer J, Laub G, Czerny C, Recht MP. (2006) MR and MR arthrography. In Davies AM, Johnson K, Whitehouse RW (Eds.). Imaging hip and bony pelvis.(Springer-Verlag, Berlin) pp. 31–48.
  16. Herwig I. (2006) Arthritis: hip. In Davies AM, Johnson K, Whitehouse RW (Eds.). Imaging hip and bony pelvis.(Springer-Verlag, Berlin) pp. 283–98.
  17. Gardner-Medwin J, Ryder CAJ, Bradshaw K, Johnson K. (2004) Magnetic resonance imaging identifies subclinical features predicting extension of arthritis in children with monoarthritis. Clin Exp Rheumatol 22:525.
  18. Rudwaleit M, Khan MA, Sieper J. (2005) The challenge of diagnosis and classification in early ankylosing spondylitis: do we need new criteria? Arthritis Rheum 52:1000–8.[CrossRef][Web of Science][Medline]
  19. Sieper J, Baraliakos X, Listing J, et al. (2005) Persistent reduction of spinal inflammation as assessed by magnetic resonance imaging in patients with ankylosing spondylitis after 2 yrs of treatment with the anti-tumour necrosis factor agent infliximab. Rheumatology 44:1525–30.[Abstract/Free Full Text]
Submitted 30 March 2006; revised version accepted 27 October 2006.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
46/4/699    most recent
kel401v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (1)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Nistala, K.
Right arrow Articles by McDonagh, J. E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Nistala, K.
Right arrow Articles by McDonagh, J. E.
Related Collections
Right arrow Juvenile Idiopathic Arthritis
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?