Rheumatology Advance Access originally published online on November 17, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Rheumatology 2004; 43: 79-84
© British Society for Rheumatology 2003; all rights reserved
Clinical |
The 5-yr HAQ-disability is related to the first year's changes in the narrowing, rather than erosion score in patients with recent-onset rheumatoid arthritis
Centre Hospitalier Universitaire Dijon, Dijon and INSERM/ERIT-M 0207, University of Burgundy, 1Service d'Immuno-rhumatologie, Centre Hospitalier Universitaire Montpellier, and INSERM U454, Montpellier, 2Centre Hospitalier Universitaire Tours, Tours, 3Centre Hospitalier Universitaire Rangueil, Toulouse and 4Institut de Rhumatologie, Hôpital Cochin, Université René Descartes, Paris, France.
Correspondence to: J. F. Maillefert, Service de Rhumatologie, Hôpital Géneral, 3 rue du Fb Raines, 21000 Dijon, France. E-mail: jean-francis.maillefert{at}chu-dijon.fr
Abstract
Objective. To evaluate the predictive validity of radiological change on 5-yr disability in rheumatoid arthritis (RA).
Methods. The study was designed to be multicentre, prospective, longitudinal, with a 5-yr follow-up. Participants were RA patients (ACR criteria), with a disease duration of <1 yr at entry. Radiographs of the hands and feet in posteroanterior view at baseline and after 12 months of follow-up (van der Heijde's modification of Sharp method) were used for structural evaluation. Disability was evaluated with Health Assessment Questionnaire (HAQ) at yr 5. Analyses consisted of (i) correlation existing between the changes in the radiological scores during the first year and the HAQ value at yr 5 and (ii) determination of the optimal cut-off in the changes in the radiological scoring system, by ROC curve analysis, in which variable to be explained was disability status at yr 5, defined by HAQ value of at least 1.
Results. Due to missing data and/or lost to follow-up, 135 patients (out of the 191 recruited patients) were included in the analyses (mean change in the radiological score = 4.9 ± 8.7 points, mean HAQ at yr 5 = 0.62 ± 0.68). There was a statistically significant correlation between the HAQ-disability status at yr 5 and the changes observed in the radiological total damage and narrowing scores during the first year (r = 0.18, P = 0.046 and r = 0.25, P = 0.006, respectively). Conversely, the short-term changes in the erosion score were not correlated with subsequent HAQ-disability (r = 0.084, P = 0.36). A change of at least 2 points in the total X-ray score was considered as optimal (sensitivity, specificity, positive and negative predictive values of 66.7, 53.9, 32.8 and 82.8%, respectively).
Conclusion. This work shows that early changes in joint damage in patients with recent-onset RA are related to subsequent HAQ-disability. This relationship is due to changes in narrowing, rather than in erosion score, suggesting that the joint narrowing score might be of great importance in the follow-up of RA patients and in the reports of scientific results. The weak performance of the thresholds established using predictive validity for subsequent HAQ-disability compromise their use at the individual level.
KEY WORDS: HAQ-disability, Joint narrowing, Radiological scoring, Recent-onset RA.
Rheumatoid arthritis (RA) is a chronic disease of unclear etiology, which can lead to severe disability. Numerous factors are associated with disability, particularly the inflammation process and joint destruction [1, 2]. The classical disease-modifying anti-rheumatic drugs (DMARDs) have been proved to reduce inflammation in some patients, while at least some of them reduce joint damage [3]. Recently, some new biological agents, which provide improvement of systemic and joint inflammation and might halt the progression of joint damage, have been proposed [4]. Since structural alterations occur rapidly in the disease history, such therapies might be used in patients with early disease. However, RA is a heterogeneous entity, whose clinical course fluctuates and whose prognosis is difficult to predict. In some patients, the disease process is severe and results in progressive disability, while a benign course with mild articular damage and low disability can be observed in others. Consequently, it is necessary to be able to accurately predict RA outcome before using aggressive therapy, particularly new biological agents, in patients with early disease.
The relationship between inflammation, structural damage and disability might depend on the phase of the disease. Disability might be predominantly related to inflammation at early stages, structural damage being of more importance in established RA [1, 2, 5]. Since it has been suggested that joint damage at early stages is related to subsequent joint destruction [6–9], it might also be related to future disability, and be a candidate for prediction of such subsequent disability. However, the assessment of joint damage uses radiological scores, which result in a continuous variable. In the hypothesis that joint damage at early stages, and especially early changes in joint damage, are related to future disability, the use of such damage for individual clinical decision-making necessitates the establishment of a threshold that can accurately discriminate between patients with and without good prognosis. Moreover, a threshold in radiological progression should be useful in therapeutic trials, since the percentage of patients having radiological disease progression is an important endpoint, is easier to understand than mean changes and allows comparison of data from different trials. Several approaches can be used to determine a cut-off above which a change in imaging variables could be considered as relevant [10, 11]. Distribution-based models describe features of the distribution of the measure in the population. In particular, the smallest detectable difference (SDD) is based on measurement error and thus on reproducibility. Authority-based models are based on expert opinion, thus on the intuitive clinician's global assessment, using gained experience and knowledge. Finally, predictive models are experimentally driven and evidence-based approaches.
In the present study, we aimed to determine whether the early changes in joint damage in patients with recent-onset RA are related to subsequent health assessment questionnaire (HAQ) disability, and to determine a cut-off above which a change in the radiological score could be considered as relevant in RA patients, based on prediction of future HAQ-disability.
Patients and methods
Study design
The study was designed to be a 5-yr, multicentre, prospective, longitudinal, follow-up study.
Participants
All consecutive outpatients who were referred from primary care physicians in four French centers (Montpellier, Paris-Cochin, Toulouse, Tours), who fulfilled the ACR criteria for RA, had a disease duration of <1 yr, and had not been previously treated with DMARDs were included between March 1993 and October 1994. Some of the patients primarily participated in a 1-yr randomized trial comparing the effects of methotrexate, sulfasalazine or a combination of both [12]. All patients agreed to be enrolled in a subsequent follow-up study, and provided their informed consent. They were treated subsequently with DMARDs that could be modified during the study according to efficacy and side-effects. The study was approved by the ethical review board in Montpellier.
Clinical assessment
The HAQ was collected by the same investigator at baseline and at 5-yr follow-up. The disease activity score (DAS) 28 was obtained at baseline, and at 1- and 5-yr follow-up.
Structural assessment
Hand, wrist and foot radiographs, in posteroanterior view, were taken at baseline and at 1 yr. They were evaluated blindly and in chronological order by a single independent observer, according to the Sharp method modified by van der Heijde, whose range is 0–448 [13]. For each patient, an erosion score, a narrowing score and a total damage score were noted for hands and feet. Before the definite evaluation, the intra-class intra-observer coefficient of correlation was calculated on 30 chosen pairs of X-rays of hands and feet and was higher than 0.85. No systematic differences were found in any of the scores.
Statistical analysis
Data were analysed using the Statistical Package for the Social Sciences (SPSS 10.1, 2000).
Changes in erosion, narrowing and total damage scores between baseline and 1-yr follow-up evaluations were calculated. Additionally, changes in hand (erosion + narrowing) and feet scores were calculated.
The correlation existing between the changes in radiological scores (total damage, erosion, narrowing, hands and feet scores) between baseline and 1-yr follow-up, and HAQ-disability at yr 5 were evaluated using univariate linear regression. In order to substrate the influence of disease activity, the same analyses were then performed using the changes in the radiological score between baseline and 1-yr follow-up and the HAQ disability at yr 5, in patients with a low activity at yr 5, defined as a DAS < 3.2. Finally, the correlation existing between the changes in the DAS score during the first year and the HAQ-disability at yr 5 were evaluated on the whole population, in order to see if clinical items predicted the outcome as well.
Analyses were then performed in order to propose relevant thresholds in 1-yr radiological progression, predicting with the maximal accuracy a significant 5-yr HAQ-disability, defined by HAQ at yr 5 
1, allowing to switch the continuous variable (absolute change in radiological score after 1 yr of follow-up) into a dichotomous variable: progression yes/no. The cut-off value for predicted HAQ was chosen as 1 since this score corresponds to the median value in RA patients entered in a large US sample of RA patients [14], and since a HAQ score of 1 means that, on average, the patient will do every activity of daily living with difficulty [15].
Five analyses were performed. The main analysis used the total damage score, while the secondary analyses used the total erosion, the total narrowing and the hands and feet scores. Each analysis used a ROC curve analysis. The ROC curve is a graph that plots the true positive rate (sensibility) as a function of the false positive rate (100 – specificity) at all possible thresholds. In each analysis, the 5-yr HAQ values were dichotomized into qualitative variable: ‘5-yr evaluation HAQ value 1 yes/no’. Thereafter, different possible thresholds were evaluated: for each change observed after 1 yr of follow-up in the radiological score (1 point per 1 point), the sensitivity (percentage of patients with a change in the radiological score above the threshold, among the patients whose 5-yr HAQ value was 1), and 100 – specificity (percentage of patients with a change in radiological score above the threshold, among the patients whose 5-yr HAQ value was <1) for predicting significant HAQ-disability status at yr 5 were calculated. The choice of the cut-off was based on maximal accuracy by ROC curve analysis [16].
In order to substrate the influence of disease activity, all these analyses were performed again in order to calculate thresholds in the population of patients with a low DAS score at yr 5.
Finally, the threshold in the 1-yr radiological progression, which predicted with the maximal accuracy a significant 5-yr disability, defined by a 5-yr HAQ the median 5-yr HAQ value in the whole population, was calculated.
Results
One hundred and ninety-one patients (140 women, 51 men) were enrolled in this study. The mean age at diagnosis was 50.5 ± 14.7 yr and the mean disease duration at inclusion was 3.6 ± 2.6 months. From these 191 patients, two refused to participate, five died and 16 were lost to follow-up. Thus, 168 patients were followed up for 5 yr. Baseline and 1-yr follow-up hand and feet radiographs were available in 135 out of these 168 patients. These 135 patients, whose baseline characteristics are shown in Table 1, were included in the present study.
|
The mean total radiological score for hands and feet at baseline was 7.3 ± 10.4 (median = 3), while the mean change after 1-yr follow-up, compared to baseline, was 4.9 ± 8.7 points (median = 1). The changes after 1-yr follow-up in the total narrowing and erosion scores and in the hands and feet scores are shown in Table 2. The percentages of patients with no erosion at baseline and after 1 yr of follow-up were 30.7 and 21.3%, respectively, while the percentages of those with a total narrowing score of 0 were 36.7 and 30.7%, respectively. The mean baseline and 5-yr follow-up HAQ values were 1.3 ± 0.7 (median = 1.25) and 0.62 ± 0.68 (median = 0.37), respectively. The 5-yr HAQ-disability was correlated with the changes observed in the total radiological and the total narrowing scores during the first year (r = 0.18, P = 0.046 and r = 0.25, P = 0.006, respectively), while there was no correlation with the changes observed in the erosion or the hands and feet scores (r = 0.084, P = 0.36; r = 0.16, P = 0.07; and r = 0.17 and P = 0.06, respectively).
|
The mean baseline, 1- and 5-yr follow-up DAS 28 scores were 4.1 ± 0.7 (median = 4.1), 2.7 ± 1.3 (median = 2.5) and 2.2 ± 1 (median = 2), respectively. The 5-yr HAQ-disability was correlated with the changes observed in the DAS 28 score during the first year (r = 0.22, P = 0.015).
At yr 5, the DAS 28 score was <3.2 in 84% of patients. In this population of patients, the 5-yr HAQ-disability was still correlated with the changes observed in the total radiological and the total narrowing scores during the first year (r = 0.3, P = 0.005 and r = 0.29, P = 0.006, respectively), and was also correlated with the changes observed in the erosion and feet scores (r = 0.21, P = 0.044 and r = 0.3, P = 0.005, respectively), but there was no correlation with the changes observed in the hands scores (r = 0.2, P = 0.06).
At yr 5, the HAQ value was 1 in 22% of patients. Using the ROC curve analysis (Fig. 1), a change of at least 2 points in the total damage X-ray score between baseline and 1-yr follow-up was considered as the optimal threshold for prediction of subsequent disability status at yr 5, defined by a HAQ value of at least 1 (sensitivity, specificity, positive and negative predictive values of 66.7, 53.9, 32.8 and 82.8%, respectively). Results of secondary analyses, which are shown in Tables 3, 4 and 5, were comparable to those of main analysis (data not shown).
|
|
|
|
Discussion
This work shows that early changes in joint damage in patients with recent-onset RA are related to subsequent HAQ-disability, that this relationship is due to changes in narrowing, rather than in erosion score, and determines a cut-off above which a change in X-ray scores could be considered as clinically relevant, based upon prediction of subsequent HAQ-disability.
The cohort was DMARD-naive at baseline, then mainly treated with sulfasalazine, methotrexate or a combination of both. Since the 12-month effect of these drugs on radiological outcome was shown to be similar [12], it is unlikely that drug regimen could have influenced the results obtained. The baseline and follow-up HAQ-disability and X-ray scores were comparable to those reported in other cohorts of patients with early-onset RA, but with some differences [5, 8, 16–19]. These differences between study results might be due to several factors, such as difference in disease duration at inclusion, in criteria for inclusion and exclusion or in study design. In the present study, the brief disease duration at inclusion (mean = 3.4 months) might explain, in association with the absence of previous DMARD treatment, that the baseline HAQ scores were higher than observed in some other studies [5, 18]. Moreover, the patients were referred from primary care physicians, so one can suppose that the sample was closer to the community RA population than if directly recruited through rheumatology clinics, which should likely have more severe disease [14, 16]. Finally, the initial inclusion in a therapeutic trial might have introduced a selection bias.
This study suggests for the first time, at least to our knowledge, that a change in the radiological scoring system in a short time span is predictive of subsequent functional impairment. In addition, such subsequent functional impairment was also predicted by change in the short term in DAS. Preventing disability constitutes a major goal of RA treatment. In clinical practice, the choice between less or more aggressive drugs in patients with recent-onset RA can be helped by the collection of different baseline parameters, including baseline structural deterioration, which have been shown to be predictive of subsequent disability. The present results suggest that the changes in X-ray scores in the short term might be used in clinical practice to evaluate treatment effects and to monitor therapy. Moreover, such a relationship highlights the importance of considering radiological changes in therapeutic trials, even of 1-yr duration, or at the 1-yr intermediate analysis.
Interestingly, the correlation between structural short-time-span changes and subsequent HAQ-disability was observed when using the total damage score, and especially when using the narrowing score. Conversely, the changes in the erosion score were not related to subsequent HAQ-disability. Such a discrepancy was not observed when influence of disease activity on disability was minimized, through exclusion of patients with active disease at 5 yr. However, this last analysis still demonstrated a closer relationship between early changes in narrowing score and subsequent disability than between early changes in erosion score and subsequent disability. Such observation is of importance since joint narrowing and erosions are not clearly distinguished in all scoring systems, and since structural results of numerous studies, including therapeutic trials, are given in terms of total damage score, without distinguishing erosions from joint narrowing. Moreover, some studies have suggested that the effects of some drugs, such as corticosteroids, on the progression of these two aspects of joint damage might be different [2]. Thus, although needing to be confirmed by further studies performed on other cohorts of patients, and with a longer follow-up, the present results suggest that the joint narrowing score might be of great importance in the follow-up of RA patients and in the reporting of scientific results.
The second part of the study established thresholds in short term changes in X-ray scores in patients with recent-onset RA, based upon predictive validity for subsequent HAQ-disability. Changes on radiographs in RA constitute an important endpoint in therapeutic trials, but do not offer information on the individual patient basis [11]. Moreover, they are commonly used by physicians for individual therapeutic decision-making, but one can suppose that the absence of a validated cut-off value induces some disparities in decisions. Several approaches can be used in order to determine a cut-off above which a change in radiological changes could be considered as relevant in RA. Lasserre et al. [20] proposed to use the SDD. Since it is not known whether the SDD represents a clinically relevant progression or not, the same group determined the minimal clinically relevant difference, using an expert opinion approach [21]. However, the predictive models are considered as the ultimate goal in such a definition [10, 11]. In the present study, a model based on prediction of subsequent HAQ-disability was used. This model has several advantages. It is probably clinically relevant since (i) disability is a central outcome in RA, (ii) maintenance of functional capacity is a major goal of therapy and (iii) the HAQ is the most common assessment tool [22]. However, two problems in the results might compromise the usefulness of the obtained cut-off. Firstly, the thresholds established using the total damage score were lower than the one obtained on the same cohort using the SDD (3.4 vs 2 points) [6]. Although differences smaller than the SDD can be considered as relevant, such differences cannot be distinguished from measurement errors. Secondly, the obtained thresholds were of somewhat moderate accuracy with, in particular, good negative, but poor positive predictive values. Thus, although the threshold might be useful to assess the absence (but not the presence) of future disability, such performances might compromise the use of the thresholds at the individual level. However, since the negative predictive values of the thresholds were good, particularly in the analysis which diminished the influence of disease activity, the thresholds might be used to determine which patients are of good prognosis and should not receive aggressive treatment.
Several hypotheses can be put forward to explain this result. Firstly, individual patients show great variations in HAQ-disability over time [22]. Secondly, structural factors other than erosions and narrowing may contribute to HAQ-disability [1, 2]. Thirdly, factors other than disease activity and structural damage contribute to HAQ-disability [18, 22, 23]. Finally, the follow-up may have been too brief. The respective effects of disease activity and joint destruction on disability change over the course of the disease. Disability might be predominantly related to inflammation at early stages, structural damage being of more importance in established RA. A minimum of 10-yr follow-ups might be necessary to obtain more accurate thresholds in early changes in X-ray scores [11].
Conflicts of interest
The authors have declared no conflicts of interest.
References
- Van der Heijde D. Radiographic progression in rheumatoid arthritis: does it reflect outcome? Does it reflect treatment? Ann Rheum Dis 2001;60 (Suppl III):iii47–50.
- Kirwan JR. Links between radiological change, disability, and pathology in rheumatoid arthritis. J Rheumatol 2001;28:881–6.
[Abstract/Free Full Text] - Van Riel PLCM, van der Heijde DMFM, Nuver-Zwart IH, van de Putte LBA. Radiographic progression in rheumatoid arthritis: results of 3 comparative trials. J Rheumatol 1995;22:1797–9.[Web of Science][Medline]
- Lipsky PE, van der Heijde DMFM, St Clair EW et al. Infliximab and methotrexate in the treatment of rheumatoid arthritis. N Engl J Med 2000;343:1594–602.
[Abstract/Free Full Text] - Welsing PMJ, van Gestel AM, Swinkels HL, Kiemeney LALM, van Riel PLCM. The relationship between disease activity, joint destruction, and functional capacity over the course of rheumatoid arthritis. Arthritis Rheum 2001;44:2009–17.[CrossRef][Web of Science][Medline]
- Combe B, Dougados M, Goupille P J et al. Prognosis factors for radiographic damage in early rheumatoid arthritis. A multiparameter prospective study. Arthritis Rheum 2001;44:1736–44.[CrossRef][Medline]
- Boers M, Kostense PJ, Verhoeven AC, van der Linden S. Inflammation and damage in an individual joint predict further damage in that joint in patients with rheumatoid arthritis. Arthritis Rheum 2001;44:2242–6.[CrossRef][Web of Science][Medline]
- Van der Heide A, Remme CA, Hofman DM, Jacobs JWG, Bijlsma JWJ. Prediction of progression of radiologic damage in newly diagnosed rheumatoid arthritis. Arthritis Rheum 1995;38:1466–74.[Medline]
- Van Zeben D, Breedveld FC. Prognostic factors in rheumatoid arthritis. J Rheumatology 1996;23 (Suppl 44):31–3.
- Lassere MND, van der Heijde D, Johnson KR. Foundations of the Minimal Clinically Important Difference for imaging. J Rheumatol 2001;28:890–1.
[Abstract/Free Full Text] - Van der Heijde D, Lassere M, Edmonds J, Kirwan J, Strand V, Boers M. Minimal clinically important difference in plain films in RA: group discussions, conclusions, and recommendations. J Rheumatol 2001;28:914–7.
[Abstract/Free Full Text] - Dougados M, Combe B, Cantagrel A et al. Combination therapy in early rheumatoid arthritis: a randomized, controlled, double-blind 52 week clinical trial of sulfasalazine and methotrexate versus the single components. Ann Rheum Dis 1999;58:220–5.
[Abstract/Free Full Text] - van der Heijde DM, van Riel PL, Nuver-Zwart HH, Gribnau FW, van de Putte LBA. Effects of hydroxychloroquine and sulfasalazine on progression of joint damage in rheumatoid arthritis. Lancet 1989;I:307–10.
- Wolfe F, O'Dell JR, Kavanaugh A, Wilske K, Pincus T. Evaluating severity and status in rheumatoid arthritis. J Rheumatol 2001;28:1453–62.
[Abstract/Free Full Text] - Wolfe F, Cathey MA. The assessment and prediction of functional disability in rheumatoid arthritis. J Rheumatol 1991;18:1298–306.[Web of Science][Medline]
- Van der Heijde DMFM, van Leeuwen MA, van Riel PLCM, van de Putte LBA. Radiographic progression on radiographs of hands and feet during the first 3 years of rheumatoid arthritis measured according to Sharp's method [van der Heijde modification]. J Rheumatol 1995;22:1792–6.[Medline]
- Eberhardt KB, Fex E. Functional impairment and disability in early rheumatoid arthritis – Development over 5 years. J Rheumatol 1995;22:1037–42.[Web of Science][Medline]
- Drossaers-Bakker KW, de Buck M, van Zeben D, Zwinderman AH, Breedveld FC, Hazes JMW. Long-term course and outcome of functional capacity in rheumatoid arthritis. The effects of disease activity and radiographic damage over time. Arthritis Rheum 1999;42:1854–60.[CrossRef][Web of Science][Medline]
- Hulsmans HMJ, Jacobs JWG, van der Heijde DMFM, van Albada-Kuipers GA, Schenk Y, Bijlsma JWJ. The course of radiologic damage during the first six years of rheumatoid arthritis. Arthritis Rheum 2000;43:1927–40.[CrossRef][Web of Science][Medline]
- Lassere M, Boers M, van der Heijde D et al. Smallest detectable difference in radiological progression. J Rheumatol 1999;26:731–9.[Web of Science][Medline]
- Bruynesteyn K, van der Heijde D, Boers M et al. Minimum clinically important difference in radiological progression of joint damage over 1 year in rheumatoid arthritis: preliminary results of a validation study with clinical experts. J Rheumatol 2001;28:904–10.
[Abstract/Free Full Text] - Wolfe F. A reappraisal of HAQ disability in rheumatoid arthritis. Arthritis Rheum 2000;43:2751–61.[CrossRef][Web of Science][Medline]
- Escalante A, del Rincon I. How much disability in rheumatoid arthritis is explained by rheumatoid arthritis? Arthritis Rheum 1999;42:1712–21.[CrossRef][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. J. Plant, M. M. O'Sullivan, P. A. Lewis, J. P. Camilleri, E. C. Coles, and J. D. Jessop What factors influence functional ability in patients with rheumatoid arthritis. Do they alter over time? Rheumatology, September 1, 2005; 44(9): 1181 - 1185. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Doan and E. Massarotti Rheumatoid Arthritis: An Overview of New and Emerging Therapies J. Clin. Pharmacol., July 1, 2005; 45(7): 751 - 762. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||