Significance of Palpable Tendon Friction Rubs in Early Diffuse Cutaneous Systemic Sclerosis


  • Adam Doré,

    1. University of North Carolina, Chapel Hill, and University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania
    Search for more papers by this author
  • Mary Lucas,

    1. University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania
    Search for more papers by this author
  • Dana Ivanco,

    1. University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania
    Search for more papers by this author
  • Thomas A. Medsger Jr.,

    1. University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania
    Search for more papers by this author
    • Dr. Medsger has received speaking fees and/or honoraria (less than $10,000 each) from the American College of Rheumatology and the University of Michigan.

  • Robyn T. Domsic

    Corresponding author
    1. University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania
    • Department of Medicine, Division of Rheumatology and Clinical Immunology, University of Pittsburgh, S724 Biomedical Science Tower, 3500 Terrace Street, Pittsburgh, PA 15261. E-mail:

    Search for more papers by this author



Palpable tendon friction rubs (TFRs) in systemic sclerosis (SSc; scleroderma) have been associated with diffuse skin thickening, increased disability, and poor survival. Our objective was to quantify the prognostic implications of palpable TFRs on the development of disease complications and longer-term mortality in an incident cohort of early diffuse cutaneous SSc (dcSSc) patients.


We identified early dcSSc patients (disease duration <2 years from the first SSc symptom) first evaluated at the University of Pittsburgh Scleroderma Center between 1980 and 2006 and found to have palpable TFRs. These patients were matched 1:1 with the next consecutive early dcSSc patient without TFRs as a control. All had ≥2 clinic visits and 5 years of followup from the first visit.


A total of 287 early dcSSc patients with TFR were identified and matched to 287 controls. The median disease duration was 0.83 years in TFR patients and 1.04 years in controls. The median followup was 10.1 years in TFR patients and 7.9 years in controls. Over the course of their illness, patients with TFRs had a >2-fold risk of developing renal crisis and cardiac and gastrointestinal disease complications, even after adjustment for other known risk factors. Patients with TFRs had poorer 5- and 10-year survival rates.


Patients with early dcSSc having ≥1 TFRs are at an increased risk of developing renal, cardiac, and gastrointestinal involvement before and after their first Scleroderma Center visit and have reduced survival. Patients presenting with TFRs should be carefully monitored for serious internal organ involvement.


Systemic sclerosis (SSc; scleroderma) is a chronic, multisystem autoimmune disease characterized by vascular abnormalities, inflammation, and fibrosis. Its 2 clinical subsets include limited cutaneous SSc (lcSSc; including SSc sine scleroderma) and diffuse cutaneous SSc (dcSSc). Patients with dcSSc generally develop internal organ complications early in the disease ([1]) and have reduced survival ([2-4]).

Tendon friction rubs (TFRs) were first observed by Westphal in 1876 ([5]) and later characterized by Shulman et al ([6]). Rodnan and Medsger ([7]) described a TFR as a “leathery, crepitus feel” on palpation during active or passive motion. In a large prevalent observational cohort of both lcSSc and dcSSc patients, we previously concluded that TFRs were associated with diffuse disease, shorter disease duration, reduced survival, and renal and cardiac involvement ([8, 9]). Khanna et al ([10]) examined a small number of early dcSSc clinical trial participants and found that changes in TFRs correlated with changes in skin thickness score and the Health Assessment Questionnaire disability index over a 2-year followup period.

The objective of this study was to quantify the prognostic implications of palpable TFRs in an incident cohort of early dcSSc patients. It overlaps our previous publication ([8]), but is methodologically superior because of a case–control analysis focused on early dcSSc patients, control for temporal trends, and adjusted risk quantification analysis for organ-specific long-term outcomes.

Patients and methods

Patient selection.

We identified all early dcSSc patients seen for an initial visit at the University of Pittsburgh Scleroderma Clinic between January 1, 1980 and December 31, 2006. Early disease was defined as the presence of skin thickening proximal to the elbows or knees and <2 years from the first symptom attributable to SSc. All patients had ≥2 visits, had 5 years of followup, and resided in the US. TFRs were considered present if palpated in the finger flexors or extensors, wrist flexors or extensors, olecranon bursa, shoulders, knees, or anterior or posterior ankles with active motion.

Study design.

We used case–control methodology. To control for temporal trends, each case (TFR present at the first visit) was matched to the next consecutive patient presenting with early dcSSc who did not have TFRs (control) at any visit.

Data collection.

All patients had a comprehensive initial visit with evaluation of symptoms, physical examination, laboratory and serologic results, and objective internal organ tests (pulmonary function tests, chest computed tomography scan, echocardiogram, electrocardiogram, esophagram, etc.). At the followup clinic visits, all patients had an SSc-specific history and examination performed. Living patients not seen in the clinic during 2011 were mailed a questionnaire and release of information to help in obtaining objective test results and to determine the time and severity of internal organ involvement.


Survival as of December 31, 2011 was obtained from the Social Security Death Index. Cause of death was determined by medical record review or discussion with physicians. If no records were available, the National Death Index was used to obtain death certificate information and reviewed in the context of known clinical information. Additional contacts with physicians or family, if available, were made.

Definitions of internal organ involvement.

Organ system involvement was defined as previously published ([11]) for renal, cardiac, gastrointestinal, pulmonary fibrosis or restrictive lung disease; pulmonary hypertension; and muscle disease. Severe peripheral vascular disease was defined as digital tip ulcers or gangrene. Joint contractures of the distal interphalangeal joints, proximal interphalangeal joints, elbows, and knees were assessed. Skin was assessed by the modified Rodnan skin thickness score (MRSS) and also by the skin thickness progression rate (STPR), which is calculated as: skin score at visit/(date of examination − date of skin onset). For example, if skin thickness (fingers swollen and never returned to normal size) began 6 months before the first visit where the MRSS was 18, the STPR is 18/0.5 years = 36/year. The STPR groups were defined as slow (<25/year), intermediate (25–44/year), or rapid (≥45/year) ([11]).

Autoantibody identification methods.

Anti–topoisomerase I was detected by Ouchterlony immunodiffusion and RNA polymerase III was detected by protein immunoprecipitation. All other SSc autoantibodies were identified as previously reported ([12]).

Statistical analysis.

Baseline characteristics were compared between the patients with and without TFRs using t-tests, chi-square tests, and nonparametric tests. The risk of developing internal organ involvement (bivariate analysis) was assessed using conditional logistic regression, and then adjusted analysis was performed with a cutoff value of P less than or equal to 0.05. Cumulative survival was calculated by the Kaplan-Meier method. Age- and sex-adjusted analysis was performed using Cox proportional hazards model.


From 1980 to 2006, 3,010 patients with SSc were initially evaluated. Of these, 1,296 had dcSSc, and 793 presented within 2 years of symptom onset. At the first visit, 403 had ≥1 TFRs; 287 were seen ≥2 times with ≥5 years of followup. There were no differences in the 287 cases and the 116 patients seen only once with respect to age, sex, race, or organ involvement other than a lower skin score (27.8 versus 30.8; P = 0.005) and a greater frequency of gastrointestinal involvement (44% versus 23%; P < 0.001) in the cases at presentation.

Of the combined 574 case and control patients, 293 (51%) were alive as of December 31, 2011. Of the 127 living cases, 71 (56%) had been seen within a year, and 39 (70%) of the other 56 returned questionnaires. Of 166 living controls, 75 (45%) had been seen and 37 (40%) of 91 returned questionnaires. Two hundred thirty-six patients (41%) in this cohort were also included in the Steen and Medsger ([8]) report, although in our study there was a longer mean followup period (10.4 versus 6.3 years) and more updated and accurate organ system information.

Baseline characteristics.

The baseline characteristics of the early dcSSc patients with and without TFRs are shown in Table 1. There was no difference in mean age or sex. There were more whites in the TFR group (94% versus 89%; P = 0.01). Patients with TFRs had a shorter median disease duration at the first visit (0.83 years versus 1.04 years; P < 0.0001) and a higher mean MRSS (27.8 versus 23.0; P < 0.0001). Therefore, TFR patients had increased STPR (P < 0.0001). The mean body mass index (BMI) was lower for TFR patients (P = 0.04), but was normal in both groups. TFRs were noted most frequently at the ankles (59%), followed by the fingers (43%), knees (37%), wrists (36%), olecranon bursae (25%), and shoulders (9%).

Table 1. Characteristics at the first visit for patients with TFRs (cases) and patients without TFRs (controls)*
 TFR cases (n = 287)Controls (n = 287)P
  1. TFR = tendon friction rub; IQR = interquartile range; MRSS = modified Rodnan skin thickness score.
  2. aDefined as the time from the first symptom attributable to scleroderma to the first Pittsburgh visit.
  3. bDigital ulcers or gangrene.
Age at first visit, mean ± SD years49.0 ± 13.748.3 ± 14.50.52
Female sex, %77760.83
White, %94890.01
Disease characteristics   
Disease duration, median (IQR) yearsa0.83 (0.61–1.13)1.04 (0.70–1.48)< 0.0001
Followup from the first visit, median (IQR) years10.1 (4.7–16.8)7.9 (2.5–16.6)0.01
MRSS, mean ± SD27.8 ± 10.923.0 ± 10.9< 0.0001
Skin thickness progression rate, %  < 0.001
Body mass index, mean ± SD kg/m223.5 ± 4.124.5 ± 4.70.04
Preexisting internal organ involvement, %   
Renal crisis1590.06
Pulmonary (restrictive physiology or fibrosis)20320.0006
Skeletal muscle5130.01
Pulmonary hypertension210.13
Severe peripheral vascularb640.24
Joint contractures83690.0002
Autoantibodies, no./total  0.35
Anti–topoisomerase I70/27462/275 
Anti–RNA polymerase III156/259137/254 
Anti–U3 RNP6/24912/237 
Anti–U11/U12 RNP3/2242/246 
Anti–U1 RNP7/2655/265 

At the first Pittsburgh visit, TFR patients already had significantly increased rates of cardiac (15%) and gastrointestinal involvement (51%) compared to 8% and 42%, respectively, in controls (Table 1). For cardiac involvement, there was no difference in the frequency of pericardial effusion or cardiomyopathy. Patients with TFRs more often had joint contractures (83% versus 69%), but had lower rates of pulmonary fibrosis or restrictive lung disease (20% versus 32%; P = 0.0006) and skeletal muscle involvement (5.2% versus 13.3%; P = 0.01). There were no differences in the frequencies of renal crisis or pulmonary hypertension.

There was no overall difference in the autoantibody profile of the patients with and without TFRs (P = 0.35). The majority of all patients were anti–RNA polymerase III positive (57%) and 24% had anti–topoisomerase I antibodies. The remainder had other SSc-associated antibodies, and 9% had no identified SSc-related antibodies. There was no difference in the frequency of prednisone, cyclophosphamide, mycophenolate mofetil, methotrexate, or D-penicillamine use prior to or prescribed at the first visit.

Organ involvement.

Cases were followed for a median of 10.4 years and controls for a median of 7.9 years after the first visit. During followup, cases were 3 times more likely to develop renal crisis (Table 2), and this risk increased to 3.77 times more likely (95% confidence interval [95% CI] 1.70–8.34) after adjusting for anti–RNA polymerase III antibody, which has been previously associated with renal crisis ([13]). After additional adjustment for STPR ([11]), TFR cases continued to demonstrate a 3.67 times increased occurrence of new renal crisis compared with controls without TFR. Patients with TFRs were 3.71 times more likely (95% CI 1.69–8.12) to develop cardiac involvement during followup and 4.6 times more likely (95% CI 2.00–10.36) to develop gastrointestinal involvement (Table 2). There was a marginally increased risk of later pulmonary fibrosis or restrictive lung disease and no significantly increased risk of later pulmonary hypertension, severe peripheral vascular disease, joint contractures, or skeletal muscle involvement.

Table 2. Risk of developing subsequent internal organ involvement in early diffuse cutaneous systemic sclerosis patients presenting with tendon friction rubs at the first visit*
 OR (95% CI)P
  1. OR = odds ratio; 95% CI = 95% confidence interval.
  2. aInterstitial lung disease or restrictive lung disease.
  3. bDefined as the presence of digital ulcers or gangrene.
  4. cAdjusted for race, disease duration, baseline modified Rodnan skin thickness score, and skin thickness progression rate group.
Renal crisis3.05 (1.59–5.87)0.0008
Cardiac3.71 (1.69–8.12)< 0.0001
Pulmonarya1.84 (1.05–3.22)0.03
Gastrointestinal4.57 (2.00–10.36)0.0003
Skeletal muscle4.00 (0.85–18.84)0.08
Pulmonary hypertension1.40 (0.71–2.78)0.33
Severe peripheral vascularb1.10 (0.47–2.59)0.83
Joint contractures2.67 (0.85–18.84)0.15
Adjusted for significant baseline differencesc  
Renal crisis2.66 (1.23–5.75)0.01
Cardiac3.26 (1.35–7.84)0.009
Gastrointestinal5.14 (1.74–15.17)< 0.0001
Pulmonarya2.06 (0.95–4.46)0.07
Pulmonary hypertension1.41 (0.56–3.51)0.47
Severe peripheral vascularb1.70 (0.46–6.31)0.43
Joint contractures0.89 (0.49–1.63)0.58

To further understand the impact of TFRs, we adjusted for baseline differences in the case and control groups, including race, disease duration, and baseline MRSS and STPR (Table 2). After adjustment, patients with TFRs continued to be at a significantly increased risk to develop renal crisis (odds ratio [OR] 2.66, 95% CI 1.23–5.75) and cardiac (OR 3.26, 95% CI 1.35–7.84) and gastrointestinal involvement (OR 5.14, 95% CI 1.74–15.17). There was no increased risk of developing pulmonary involvement, pulmonary hypertension, digital ulcers, or joint contractures in those with TFRs. There were too few new cases of muscle involvement to perform regression analysis.

Over the course of the illness, including both before and after the first evaluation, patients with TFRs were 2.3 times more likely to develop renal crisis (95% CI 1.52–3.49, P < 0.0001). This risk was unchanged when adjusted for anti–RNA polymerase III antibody and STPR. Patients with TFRs were 2.43 times more likely to develop cardiac disease and 2.25 times more likely to develop gastrointestinal involvement. There was no increased risk to develop pulmonary hypertension, severe peripheral vascular disease, skeletal myopathy, or pulmonary involvement.


Five-year unadjusted cumulative survival was significantly reduced in patients with TFRs (68% versus 81% in controls; P < 0.0001), as shown in Figure 1. Ten-year cumulative survival was also significantly reduced (P < 0.0001) at 55% in TFR cases compared to 64% in controls without TFRs. After age and sex adjustment, TFR cases were 1.8 times more likely to die at 5 years (95% CI 1.21–2.58, P = 0.003). BMI did not change the results. In both cases and controls, 66% of all deaths were judged to be SSc related, with <2% being an undetermined cause of death.

Figure 1.

Five-year cumulative survival in patients with tendon friction rubs compared to controls.


The TFR is a physical examination finding highly specific for SSc and has been linked to diffuse skin involvement, increased disease severity and disability ([10]), and poor survival ([8, 9]). In this large case–control study of an incident cohort of early dcSSc patients, we examined the long-term risks of internal organ involvement in patients presenting with TFRs and demonstrated a substantially increased risk of developing renal crisis and cardiac and gastrointestinal complications compared to those without TFRs. We have quantified these risks and adjusted for other known associations to demonstrate a 3–5-fold greater risk of developing these complications throughout the course of the disease, distinguishing this from prior TFR literature ([8-10]). We confirmed previous reports of decreased survival in patients with TFRs. These findings have important implications for monitoring of early dcSSc patients with TFRs.

Consistent with prior publications ([8, 10]), the most frequent sites of TFRs in this study were the ankles, fingers, and wrists. Although more patients with TFRs present with joint contractures, over time persons with TFRs are not more likely to have joint contractures than those without TFRs. We have also demonstrated an increased risk of developing severe peripheral vascular involvement with digital tip ulcers or gangrene in TFR patients. This difference was not present at baseline, suggesting that TFR patients may later develop more severe peripheral vascular involvement than patients without TFR.

Similar to other studies ([8, 9]), patients with TFRs have a poorer long-term survival that persists after age and sex adjustment. We also showed no increased rate of SSc-associated cause of death compared to controls, which has not been examined in prior studies.

Our earlier report ([9]) suggested an association between TFRs and anti–topoisomerase I antibody, which we did not confirm. The most frequent serum autoantibody in our cohort was anti–RNA polymerase III, which was present in >50% of patients and likely related to the restriction to early dcSSc. However, this antibody positivity did not explain the increased risk of renal crisis, since the 3-fold increased risk of developing renal crisis persisted independent of RNA polymerase III.

Detection of TFRs is admittedly dependent on examiner skill and experience. If TFRs are accepted as important in routine clinical care and in clinical trials, physician training in the proper technique for detecting TFRs should be provided, such as in the case for the MRSS.

There are limitations of this study. First, it is a single-center study from a tertiary care scleroderma center with a predominantly white population, limiting its generalizability. Second, despite our efforts, we were unable to ascertain up-to-date organ assessment testing and information on all patients. The followup questionnaire response rate was higher in patients with TFRs, and these patients had a longer mean followup than the patients without TFR. This could have biased the results by reducing the likelihood of capturing later new organ involvement in controls. Finally, this study covers 3 decades, and therefore differences in the ascertainment of organ system testing may affect the incidence and prevalence of complications and survival. We have attempted to control for such differences by temporally matching cases and controls.

In summary, the TFR is an easily detected physical examination finding that occurs in early dcSSc patients and identifies individuals at high risk for renal crisis and cardiac and gastrointestinal complications. TFRs are associated with decreased survival. Consideration of these factors should be made in monitoring and management of patients presenting with early dcSSc and one or more palpable TFRs.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Domsic had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Doré, Medsger, Domsic.

Acquisition of data. Doré, Lucas, Ivanco, Medsger, Domsic.

Analysis and interpretation of data. Doré, Lucas, Medsger, Domsic.