Objective To determine whether joint synovitis and tendon friction rubs (TFRs) can predict the progression of systemic sclerosis (SSc) over time.
Patients and methods We performed a prospective cohort study that included 1301 patients with SSc from the EUSTAR database with disease duration ≤3 years at inclusion and with a follow-up of at least 2 years. Presence or absence at clinical examination of synovitis and TFRs was extracted at baseline. Outcomes were skin, cardiovascular, renal and lung progression. Overall disease progression was defined according to the occurrence of at least one organ progression.
Results Joint synovitis (HR: 1.26, 95% CI 1.01 to 1.59) and TFRs (HR: 1.32, 95% CI 1.03 to 1.70) were independently predictive of overall disease progression, as were also the diffuse cutaneous subset (HR: 1.30, 95% CI 1.05 to 1.61) and positive antitopoisomerase-I antibodies (HR: 1.25, 95% CI 1.02 to 1.53). Regarding skin progression, joint synovitis (HR: 1.67, 95% CI 1.06 to 2.64) and TFRs (HR: 1.69, 95% CI 1.02 to 2.77) were also independently predictive of worsening of the modified Rodnan skin score. For cardiovascular progression, joint synovitis was predictive of the occurrence of new digital ulcer(s) (HR: 1.45, 95% CI 1.08 to 1.96) and decreased left ventricular ejection fraction (HR: 2.20, 95% CI 1.06 to 4.57); TFRs were confirmed to be an independent predictor of scleroderma renal crisis (HR: 2.33, 95% CI 1.03 to 6.19).
Conclusions Joint synovitis and TFRs are independent predictive factors for disease progression in patients with early SSc. These easily detected clinical markers may be useful for the risk stratification of patients with SSc.
- Systemic Sclerosis
- Outcomes research
Statistics from Altmetric.com
Systemic sclerosis (SSc) is an orphan and incurable connective tissue disease characterised by vascular, immune and fibrotic abnormalities in the skin and some internal organs.1 ,2 It involves progressive organ failure, and it is therefore a severe chronic disease leading to major disability and even premature death.3 A recent meta-analysis found that the mortality rate among patients with SSc has not changed over the past 40 years, underlining that SSc remains a devastating condition.4 There are currently no validated predictors of disease evolution and this significantly limits patient risk-stratification and consequently the use of potentially innovative therapies in the earliest phase of the disease and/or for high-risk patients.5 Thus, improvement in the ability to identify those patients with SSc at risk of major organ involvements is a critical issue for the management of this disease, and needs to be addressed urgently.
A previous cross-sectional analysis of the EUSTAR registry revealed that articular involvement is very frequent in SSc: the prevalence of articular symptoms of any form was 28%.6 In addition, synovitis and tendon friction rubs (TFRs) were more likely to occur in the early stages of the disease, and were associated with disease activity and systemic inflammation.6 ,7 TFRs were also shown to be predictive of the further development of renal, cardiac and gastrointestinal involvement. However, these results were obtained for a limited number of patients and were restricted to patients with early diffuse cutaneous SSc.8 Thus, one hypothesis is that joint and tendon involvement might be a marker of disease progression in patients with early SSc. Therefore, our aim was to determine for the first time the value of synovitis as a predictor of disease progression, and extend in a large cohort the data previously obtained for TFRs in patients with early diffuse SSc. This is now made possible by the implementation of systematic longitudinal follow-up of patients with SSc included in the EUSTAR registry, updated on an annual basis, which has been a major step forward, and now allows the identification of predictors of disease course.
Patients and methods
We conducted a prospective cohort study using the EUSTAR database.
Research study network
All data for patients registered in the EUSTAR database as of January 2013 was exported. We analysed the baseline and follow-up visit data (on a regular basis of 1 year intervals), using a predefined annual data collection protocol created, applied and updated by trained physicians of the EUSTAR network, in order to prevent differential misclassification. The structure of the database, the minimum essential data set, and the inclusion criteria have been described previously.9–12 Each participating centre obtained approval of the local ethics committee and registered patients gave informed consent.
Inclusion and exclusion criteria
Only patients who have been classified according to the 1980 American College of Rheumatology (ACR) criteria were analysed.13 We included patients with early disease (onset of first non-Raynaud's symptom ≤3 before inclusion) and with a follow-up of at least 2 years. Patients who missed the 2-year follow-up visit were excluded. We excluded patients with severe cardiac and vascular involvement at baseline defined as follows: current digital ulcers (DUs), and/or pulmonary hypertension (PH) confirmed by right heart catheterisation, and/or heart failure (left ventricular ejection fraction, LVEF <50% or class III/IV dyspnoea) and/or scleroderma renal crisis (SRC).
We extracted data relevant to the presence or absence at clinical examination of synovitis, defined by one or more tender and swollen joints regardless of its aetiology, and TFRs, defined by a rubbing sensation detected as the tendon was moved. These examination findings were made by the examining physician. Although previous data suggested that synovitis and TFRs tend to occur in the same patients, these two variables were considered as two different entities since they display different pathological and clinical characteristics. This has been highlighted by a recent ultrasound analysis, which showed that tendon involvement relies more on a fibrotic component than inflammatory changes, at the difference of synovitis.14
We also extracted data related to disease progression. Skin progression was primarily defined as a worsening of the modified Rodnan Skin Score (mRSS) of ≥30% and ≥5 points. Peripheral vascular worsening was defined as the occurrence of new DUs, which had to be distal to, or at, proximal interphalangeal joints and not thought to be due to trauma. Lung vascular progression was defined as the new onset of precapillary PH on right heart catheterisation (resting mean pulmonary artery pressure ≥25 mm Hg together with a pulmonary capillary wedge pressure of ≤15 mm Hg).15–17 Cardiac progression was defined as the reduction of the LVEF to below 50% as assessed by echocardiography. Renal progression was defined as the occurrence of SRC. Interstitial lung disease progression was defined as the new onset of pulmonary fibrosis on high resolution CT scan, or the deterioration of lung volume (≥10% of forced vital capacity, FVC). Overall disease progression was defined as the occurrence of at least one of the above prespecified events during the follow-up period.
Data are presented as means±SDs unless stated otherwise. The relationship between baseline nominal variables was measures by the Cramer's V test. Predictors of disease progression were evaluated by univariate and multivariate Cox proportional hazards models and summarised as HRs and 95% CIs. All variables identified by univariate analysis as having a p value <0.05 were included in the multivariate model. Significance was defined as a p value <0.05. All calculations were performed with statistical software (MedCalc V.11.4.4 and XL STAT V.2013.6.02).
Of the 9165 patients with SSc included in the database, 1642 patients had disease duration of less than 3 years and follow-up of at least 2 years; 89 patients were excluded since they missed the 2-year follow-up visit, 120 because of severe cardiovascular involvement at baseline, and 132 because of missing data regarding cardiac evaluation at baseline. Finally, 1301 patients (1079 women) met our inclusion criteria. This sample had sufficient power for detecting impact on studied outcomes (above 0.99 for all outcomes, see online supplementary table S1). For all patients, the first visit was between 2003 and 2010 and the mean±SD follow-up was 5±2 years. Patients with SSc had a mean±SD age of 55±15 years and a mean±SD disease duration of 2±1 years. Joint synovitis and TFRs were observed in 234 (18%) and 166 (13%) patients, respectively, and were significantly more frequent in patients with the diffuse cutaneous subset (126/500, 25% vs 108/801, 13%, p<0.001, for synovitis and 118/500, 24% vs 48/801, 6%, p<0.001, for TFRs). Detailed characteristics of these patients are provided in table 1.
Association between joint synovitis and TFRs at baseline
To study the relationship between joint synovitis and TFRs at baseline, the Cramer's V was applied. The coefficient of association between these two variables was 0.19, suggesting a low association. Thus, in case of a p value <0.05 in univariate analysis, joint synovitis and TFRs were entered as independent covariates in the same multivariate model.
Prediction of disease progression
Overall disease progression
During the follow-up period, 579 patients (45%) experienced skin and/or lung and/or cardiovascular progression. The mean time to development of overall disease progression after study inclusion was 3±2 years. Univariate analysis identified synovitis as a predictor of overall disease progression (p<0.001); it also indicated that TFRs (p<0.001), the diffuse cutaneous subset (p<0.001), a mRSS >14 (p<0.001), positive antitopoisomerase-I antibodies (p<0.001), muscle weakness (p=0.006), high creatine kinase values (p=0.015) and FVC <75% predicted (p<0.001) were predictors of progression (table 2). Multivariate analysis confirmed synovitis (HR: 1.26, 95% CI 1.01 to 1.59), TFRs (HR: 1.32, 95% CI 1.03 to 1.70), the diffuse cutaneous subset (HR: 1.30, 95% CI 1.05 to 1.61) and positive antitopoisomerase-I antibodies (HR: 1.25, 95% CI 1.02 to 1.53) as independent predictors of overall disease progression (table 2).
During the follow-up period, 99/1253 (8%) patients experienced a skin progression. The mean time to development of skin progression after study inclusion was 3±2 years (figure 1). Information on skin progression was not available for 48 patients. Univariate analysis identified synovitis (p<0.001) as a predictor of skin progression, as well as TFRs (p<0.001). Other predictors identified by univariate analysis were the diffuse cutaneous subset (p=0.021), positive antitopoisomerase-I antibodies (p<0.001) and history of DU (p=0.003) (table 3). Multivariate Cox analysis confirmed that joint synovitis (HR: 1.67, 95% CI 1.06 to 2.64) and TFRs (HR: 1.69, 95% CI 1.02 to 2.77) were strong independent predictors of skin progression, and that positivity for antitopoisomerase-I antibodies (HR: 1.72, 95% CI 1.09 to 2.62) and history of DUs (HR: 1.50, 95% CI 1.01 to 2.23) were also predictive markers (table 3).
Vascular and cardiac progression
During the follow-up period, new DUs occurred in 239/1298 (18%) patients with SSc (time to development after study inclusion: 3±2 years (figure 1); information not available for 3 patients), 36/1210 (3%) experienced a reduction of LVEF <50% (time to development after study inclusion: 4±2 years (figure 1); information not available for 91 patients), 24/1298 (2%) developed SRC (time to development after study inclusion: 3±2 years (figure 1); information not available for 3 patients) and 51/1291 (4%) patients with SSc developed precapillary PH (time to development after study inclusion: 4±3 years (figure 1); information not available for 10 patients).
Joint synovitis and TFRs were predictive of the occurrence of new DUs according to univariate analysis (p<0.001 and p=0.003, respectively); the diffuse cutaneous subset (p<0.001), a mRSS >14 (p<0.001), positivity for antitopoisomerase-I antibodies (p<0.001), history of DU (p<0.001) and FVC <75% of the predicted value (p<0.001) were also found to be predictive of new DUs (table 4). In multivariate analysis, joint synovitis (HR: 1.45, 95% CI 1.08 to 1.96) was confirmed as an independent predictor of the occurrence of new DUs, as well as history of DU (HR: 1.99, 95% CI 1.51 to 2.64) and antitopoisomerase-I antibodies (HR: 1.76, 95% CI 1.30 to 2.40) (table 4).
Reduction of LVEF
Univariate analysis identified the following predictive of the further reduction of LVEF: joint synovitis (p=0.004), the diffuse cutaneous subset (p=0.002), lung fibrosis (p<0.001), muscle weakness (p=0.001), history of DU (p=0.032) and being positive for antitopoisomerase-I antibodies (p=0.002) (table 4). TFRs were not identified as a predictor of reduced LVEF in univariate analysis (p=0.3). Multivariate analysis confirmed joint synovitis (HR: 2.20, 95% CI 1.06 to 4.57), lung fibrosis (HR: 2.21, 95% CI 1.09 to 4.47) and muscle weakness (2.25, 95% CI 1.08 to 4.56) as independent predictors of reduced LVEF (table 4).
Scleroderma renal crisis
Only TFRs were predictive of the further occurrence of SRC (p=0.0007), as were the diffuse cutaneous subset (p=0.01), a mRSS >14 (p=0.0004), being positive for antitopoisomerase-I antibodies (p=0.003), high creatine kinase values (p=0.022), muscle weakness (p=0.006) and FVC <75% of the predicted value (p<0.001). Age, sex, disease duration, lung fibrosis, elevated acute phase reactants and treatment with corticosteroids at baseline did not predict the further occurrence of SRC. TFRs (HR: 2.33, 95% CI 1.03 to 6.19), a mRSS >14 (HR: 3.08, 95% CI 1.24 to 7.61) and being positive for antitopoisomerase-I antibodies (HR: 2.11, 95% CI 1.12 to 6.97) were confirmed as independent predictors of SRC by multivariate analysis.
Precapillary pulmonary hypertension
Joint synovitis and TFRs were not predictive of the development of precapillary PH.
Interstitial lung disease progression
During the follow-up period, 266/1295 (20%) patients experienced interstitial lung disease progression (information not available for 6 patients): 173 (13%) patients developed lung fibrosis as assessed by high resolution computed tomography (HRCT) scanning and 107 (8%) had a deterioration ≥10% of FVC. The mean time to development of lung progression after study inclusion was 3±2 years (figure 1).
Joint synovitis and TFRs were not predictive of lung progression, as well as age, sex, disease duration, lung fibrosis and elevated acute phase reactants. The diffuse cutaneous subset (p<0.001), positive antitopoisomerase antibodies (p<0.001), mRSS >14 (p<0.001), history of DUs (p=0.039), high creatine kinase (CK) levels (p=0.006) and muscle weakness (p=0.031) were each predictive of lung progression. Treatment with immunosuppressive drugs was not found to be associated with lung outcome. Multivariate Cox analysis confirmed a mRSS >14 and positivity for antitopoisomerase-I antibodies as independent predictors of lung progression (HR: 1.67, 95% CI 1.26 to 2.22 and 1.38, 95% CI 1.04 to 1.84, respectively).
We report an analysis using the largest database available worldwide, and describe the first evidence that joint synovitis and TFRs are predictors of the progression of SSc.
Synovitis and TFRs were predictive of overall disease progression. Thus, accurate identification of high-risk patients using joint and tendon involvement may hasten the detection of disease progression; this may facilitate initiation of adequate treatment at the appropriate time for maximum efficacy. Adapted studies are now required to determine whether this is feasible.
We show that synovitis is one of the strongest predictors of skin progression, which directly reflects disease progression. This is consistent with the baseline analysis of the EUSTAR cohort and other published data, which showed that patients with early disease and synovitis were more likely to have the diffuse cutaneous subset.6–8 ,18
Our results also extend to patients with early disease, previous findings indicating that TFRs are predictive of evolution to diffuse cutaneous SSc.7 Moreover, a previous study showed that changes of TFRs over 6 months and 12 months were predictive of changes in mRSS.18 In our analysis, synovitis and TFRs were predictive of skin progression after stratification for cutaneous subset and antibody status: therefore, these clinical signs should be evaluated in all patients with SSc and early disease. These various observations, as a whole, have important clinical implications because the detection of joint and tendon involvement may allow the identification of a subset of patients at risk of skin progression, which in turn, predicts internal organ involvement.19 Joint and tendon involvement may be useful clinical markers in clinical trials in early SSc to identify patients at risk of skin progression, who may also be more prone to respond to selective drugs targeting skin involvement.20
Joint synovitis was identified as a strong predictor of the further occurrence of DU. This is of importance because DUs are frequent in SSc, and have substantial consequences for quality of life. Being positive for antitopoisomerase-I antibodies and the diffuse cutaneous subset were also identified as independent predictors of the occurrence of new DUs. This is consistent with the increased frequency of DUs observed in patients with early onset SSc, more severe skin fibrosis and positive antitopoisomerase-I antibodies.21 ,22
We have also identified joint synovitis as an independent predictor of reduced left heart function. On the other hand, TFRs did not predict cardiac progression, which is not consistent with the higher risk of cardiac complications in patients with TFRs and the diffuse cutaneous subset.7 ,8 This discrepancy may be explained by the small number of cardiac events in this cohort, and by the inclusion of patients with the limited cutaneous subset, whereas previous studies focused only on patients with diffuse SSc.
Our results extend previous work on the predictive value of TFRs for the further occurrence of SRC.7 ,8 They also confirm positivity for antitopoisomerase-I antibodies as a marker of poor prognosis.23 ,24
Joint and tendon involvement was not associated with lung outcomes in multivariate analysis. This finding is consistent with previous cross-sectional studies, which have reported no association between synovitis or TFRs and either pulmonary fibrosis or reduced FVC.6 ,7 ,18
There are several strengths to our study. Unlike previous retrospective and cross-sectional clinical studies of articular involvement in SSc, ours was prospective to identify the characteristics and attribution of disease progression using a predefined annual data collection protocol created and applied by trained physicians of the EUSTAR network. In addition, the quality of maintenance of the database has to be underlined, with organisation of regular courses on a 2-year basis, which aim to improve the quality of variables to be collected and teach all investigators contributing to the database. We assessed the value of joint and tendon involvement, very simple and easily available signs that are always collected by physicians. All events used to define disease progression were robust and have been used in previous studies.12 ,16 ,25 ,26 Moreover, the multicentre, international, longitudinal study design allows our findings to be extrapolated to the broader community of patients with SSc.
However, the limitations of this study should be considered when interpreting the findings. This work is uninformative about patients with DU, renal involvement and cardiac failure at inclusion as they were excluded from the study. However, the exclusion of these patients allowed us to focus on a population of patients with early SSc more at risk of disease progression. Information on synovitis and TFRs was only available at baseline, preventing assessment of whether changes in synovitis or TFRs influence changes in disease progression. Synovitis was captured by ‘yes’ or ‘no’ and not by a joint count, which did not allow us to assess the precise distribution of joint involvement. This point has now been improved in the new version of the EUSTAR registry, which will enable to further study this aspect in future studies. Moreover, synovitis that may be related to another process, such as osteoarthritis, may have been included in the analysis. Patients were recruited over several years (2003–2010). However, the prevalence at baseline of joint synovitis and TFRs was stable over time, as well as the incidence of organ progression (see online supplementary table S2). The full spectrum of patients with SSc may have not been included, since we analysed patients fulfilling the 1980 ACR criteria. The development of the new ACR/European League Against Rheumatism (EULAR) classification criteria will allow to increase in the future the number of patients with early disease or with limited cutaneous SSc.27 The assessment of predictors of renal crisis did not include RNA polymerase III antibodies because of the high number of missing data, since the detection of these antibodies is not yet performed in routine care in all the EUSTAR centres. Finally, the duration of follow-up in the current study does not reflect the lifetime experience of all organ-based SSc damage and events. However, focusing on early events is thought to be of high interest in chronic inflammatory diseases, as exemplified by the window of opportunity demonstrated in rheumatoid arthritis.28
In conclusion, this first report of the prospective follow-up of EUSTAR patients provides the first evidence of the value of synovitis and TFRs as predictive indicators of disease progression in patients with early SSc. The results were obtained from the largest database in the world, and argue for the use of these easily evaluated clinical signs for the risk stratification of patients with SSc. These indicators may serve to select high-risk patients, guide therapies and as potential surrogate markers for disease progression.
The authors acknowledge Mrs Carol Desbas and Mrs Fazia Amrouche for their expert secretarial assistance.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
Handling editor Tore K Kvien
Correction notice This article has been corrected since it was published Online First. The study collaborators have been listed.
Contributors Design of the study: JA, YA; Acquisition of data: UAW, EH, GR, GC, PEC, PC, LPA, MM-C, LC, CD, UML; Data interpretation and analysis: JA, YA; Drafting and revisiting the manuscript: JA, UAW, EH, GR, GC, PEC, PC, LPA, MM-C, LC, CD, UML, YA; Final approval of the manuscript: JA, UAW, EH, GR, GC, PEC, PC, LPA, MM-C, LC, CD, UML, YA.
Collaborators *The EUSTAR collaborators: Serena Guiducci (Florence), Alan Tyndall (Basel), Giovanni Lapadula, Florenzo Iannone (Bari), Oliver Distler (Zurich), Radim Becvar (Prague), Stanislaw Sierakowsky, Otylia Kowal Bielecka (Bialystok), Maurizio Cutolo, Alberto Sulli (Genova), Gabriele Valentini (Napoli) Simona Rednic, Ileana Nicoara (Cluj), Panayiotis G. Vlachoyiannopoulos (Athens), Carlomaurizio Montecucco, Roberto Caporali (Pavia), Srdan Novak (Rijeka), Carlo Chizzolini (Geneva), Eugene J. Kucharz, Anna Kotulska (Katowice), Franco Cozzi (Padova), Blaz Rozman (Ljublijana), Carmel Mallia, Bernard Coleiro (Balsan), Armando Gabrielli (Ancona), Dominique Farge-Bancel, Sondess Hadj-Khelifa (Paris), Paolo Airò (Brescia), Roger Hesselstrand, Dr. Agneta Scheja (Lund), Duska Martinovic (Split), Alexandra Balbir Gurman, Yolanda Braun-Moscovici (Haifa), Nicolas Hunzelmann (Köln). Raffaele Pellerito (Torino), Lisa Maria Bambara (Verona), Jadranka Morovic-Vergles (Zagreb), Carol Black, Christopher Denton (London), Nemanja Damjanov (Belgrade), Ina Kötter (Tübingen), Vera Ortiz Santamaria, (Barcelona), Stefan Heitmann (Stuttgart), Dorota Krasowska (Lublin), Matthias Seidel (Bonn), Paul Hasler (Aarau), Harald Burkhardt, Andrea Himsel (Frankfurt), José Antonio Pereira Da Silva, Maria João Salvador (Coimbra), Bojana Stamenkovi, Aleksandra Stankovic (Niska Banja), Bianca Marasini, Laura Belloli (Rozzano), Mohammed Tikly (Johannesburg), Lev N. Denisov (Moscow), Raffaella Scorza (Milano), Merete Engelhart, Gitte Strauss (Gentofte), Gabriella Szücs, Szilvia Szamosi (Debrecen), Carlos de la Puente, Paloma García de la Pena Lefebvre (Madrid), Øyvind Midtvedt, Torhild Garen (Oslo), David Launay, (Lille), Guido Valesini, Valeria Riccieri (Roma), Ruxandra Maria Ionescu, Daniela Opris (Bucharest), Fredrick M. Wigley (Baltimore), Viktor Stoica, Carmen Marina Mihai (Bucharest), Jörg Distler (Erlangen), Pierluigi Meroni, Silvana Zeni (Milan), Luc Mouthon (Paris), Vanessa Smith (Ghent), Dr. Francesco Paolo Cantatore, Ada Corrado (Foggia), Susanne Ullman, Line Iversen (Copenhagen), Maria Rosa Pozzi (Monza), Kilian Eyerich, Rüdiger Hein (Munich), Jacek Szechinski, Piotr Wiland (Wroclaw), Brigitte Krummel-Lorenz (Frankfurt), Martin Aringer (Dresden), Rene Westhovens, Ellen De Langhe (Leuven), Branimir Anic, Marko Baresi (Zagreb), Maria Üprus, Kati Otsa (Tallin), Sule Yavuz. (Istanbul), Sebastião Cezar Radominski, Carolina de Souza Müller (Curitiba), Sergio Jimenez, Joanna Busquets (Philadelphia), Sergei Popa, Svetlana Agachi (Chisinau), Thierry Zenone (Valence), Alessandro Mathieu, Alessandra Vacca (Monserrato), Lisa Stamp, Peter Chapman (Christchurch), Cristina-Mihaela Tanaseanu, Monica Popescu (Bucharest), Rosario Foti (Catania), Paloma García de la Peña Lefebvre, Silvia Rodriguez Rubio (Madrid).
Competing interests None.
Ethics approval Each EUSTAR participating centre has obtained approval of the local ethics committee and registered patients gave informed consent.
Provenance and peer review Not commissioned; externally peer reviewed.
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.