SCTC = Scleroderma Clinical Trials Consortium; CRA = Canadian Rheumatology Association; ACR = American College of Rheumatology; vs = versus; SSc = systemic sclerosis.
Contribution from the Field
Scleroderma treatment differs between experts and general rheumatologists
Version of Record online: 6 FEB 2006
Copyright © 2006 by the American College of Rheumatology
Arthritis Care & Research
Volume 55, Issue 1, pages 138–145, 15 February 2006
How to Cite
Pope, J. E., Ouimet, J. M. and Krizova, A. (2006), Scleroderma treatment differs between experts and general rheumatologists. Arthritis & Rheumatism, 55: 138–145. doi: 10.1002/art.21714
- Issue online: 6 FEB 2006
- Version of Record online: 6 FEB 2006
- Manuscript Accepted: 11 AUG 2005
- Manuscript Received: 7 DEC 2004
Scleroderma (systemic sclerosis [SSc]) is a rare connective tissue disease with variable involvement of the skin and major organs (1). The hallmarks are fibrosis, vascular instability, and, initially, inflammation. The prevalence of SSc may be as high as 2 of 10,000 (2), but SSc is far less common than rheumatoid arthritis (RA; 0.8%) and systemic lupus erythematosus (0.2%) (3, 4). Therefore, rheumatologists who do not have a specific interest in SSc will rarely treat individuals with this condition. Although there is no strong evidence for effective disease-modifying treatment for SSc (in particular, well-designed randomized controlled trials [RCTs] of early diffuse SSc have been negative &lsqbr;5, 6&rsqbr;), disease management may be achieved through the treatment of associated organ involvement. However, treatments for most organ involvement have shown no significant increase in survival in prospective studies, with the exception of aggressive treatment of SSc renal crisis treated with angiotensin-converting enzyme inhibitors and possibly therapy of pulmonary arterial hypertension (PAH) with prostacyclin and endothelin receptor blocker medications (7, 8). Standard therapies for gastroesophageal reflux disease (GERD; such as proton-pump inhibitors), gastric dysmotility (promotility agents), and Raynaud's phenomenon (calcium-channel blockers) are used.
The rarity of the disease, the clinical heterogeneity, and the lack of effective agents to date render therapy a major challenge. There are currently no widely disseminated therapeutic guidelines for the management of scleroderma. The present qualitative survey study was designed to provide a general synopsis of the current standard of care for patients with SSc treated by general rheumatologists and those with an interest in scleroderma. The objective was to assess whether rheumatologists in North America (defined as all practicing members of the Canadian Rheumatology Association [CRA] and randomly selected members of the American College of Rheumatology [ACR] practicing in the US, exclusive of those also in the CRA) demonstrated similar prescribing/treatment practices for their patients with SSc compared with Scleroderma Clinical Trials Consortium (SCTC) members, and to relate treatment trends to published evidence. Thus, this research was hypothesis generating and was meant to establish a starting point from which future research on standard treatment for SSc could be developed, providing a “snap shot” of scleroderma treatment practices. We hypothesized that SCTC members would treat patients with SSc differently (compared with non-SSc expert rheumatologists) because they would be more aware of published and unpublished results of clinical trials. In other therapeutic areas, there are opinion leaders (perhaps fast adapters) who treat patients more aggressively (such as treatment in RA targeting a low disease activity score) (9).
Subjects and Methods
A 5-page, self-administered survey was designed to assess the number of limited and diffuse patients with SSc followed; frequency and type of therapies for disease manifestations overall and for involvement of internal organs such as gastrointestinal tract, kidneys, lungs, heart, joints, muscles, and pulmonary arteries; and physicians' estimates of comorbidity and mortality frequencies in their patients with SSc. The questionnaire was mailed to 302 members of the CRA, a random subset of 312 members of the ACR, and 139 members of the SCTC. The random subset of ACR members was drawn using the ACR 2002 Directory. The directory was opened to a random page and every twentieth member (excluding those with non-US addresses) was chosen to obtain a sample size similar to that of the CRA. Physicians who followed <3 patients with SSc and those only in research were excluded from subsequent analyses.
The SCTC is an incorporated not-for-profit organization of scleroderma researchers particularly interested in research in the conduct and methodology of therapeutic trials in scleroderma. The SCTC is an international group of scleroderma researchers whose mandate is to advance the state of knowledge and therapeutic practices for scleroderma through clinical research trials. Membership occurs via the institution (SCTC members are affiliated with an academic institution) and requires that member institutions have a dedicated scleroderma clinic, have participated in at least 1 SSc clinical trial, and have published at least 1 SSc report in the past 5 years. Provisional membership can be attained for institutions that do not meet all requirements (10). The SCTC is multinational, but the largest membership representation is from the US. We considered the members of the SCTC as scleroderma treatment experts for the purposes of this study. Half of the CRA members are academic based, and we are unaware of the proportion of surveyed ACR members who were affiliated with academic centers (due to a high nonresponse rate).
The questionnaire began with a short demographics section (age, sex, year and location of rheumatology fellowship, location of practice and other subspecialties) followed by a section on patient characteristics (annual number of limited and diffuse patients with SSc seen). Questions 10–18 comprised pages 2–4, and each question was of the format, “How often do you use the following for (e.g., worsening interstitial lung disease [ILD]) in scleroderma?” A list of applicable therapies was provided with checkoff boxes under the options “Never,” “Rarely,” “Occasionally,” and “Frequently” (Appendix A). Survey respondents were instructed to select all options that applied. On the final page respondents were asked to indicate what percentage of their patients with SSc had comorbidities that were presented in a list, with check-box options for proportion ranges (Appendix B). The last formal question (question 20) asked physicians to report the frequencies of various causes of mortality (e.g., “Other causes, with scleroderma a significant comorbidity”), choosing from the same frequency categories used for questions 10–18.
The questionnaire was pretested on the 4 rheumatologists in our institution who were not involved in the development of the study survey. No major revisions were required. Responses to the questionnaire were not validated.
The intial survey was sent via postal mail with a stamped return envelope or international reply coupons for overseas SCTC members. Survey nonrespondents were re-sent the questionnaire by facsimile (with instructions to return by facsimile) one additional time. If subjects did not send their completed questionnaires back after the second request, they were considered nonrespondents.
Data were entered into a JMP statistical software (11) database using a standardized template based on the study questionnaire. Ambiguous responses were coded based on agreement by the 2 authors responsible for data entry, and if no agreement could be reached, the particular question was considered a nonresponse item. Data were verified by reentering a subset of randomly chosen surveys. Between-groups comparisons (SCTC compared with both CRA and ACR) were performed for reported value frequencies using proportion statistical tests. P values for the SCTC versus CRA and SCTC versus ACR were reported together unless they differed in significance.
Interpretation of data coding categories
Due to small numbers of observations in each category of response (“never,” “rarely,” “occasionally,” and “frequently”), data fields were combined in the final analysis to allow for statistical inference testing using contingency tables. For the majority of questions, comparisons were made across physician groups (SCTC, CRA, and ACR) regarding whether they currently use a treatment (i.e., the respondent answered that he or she used that treatment “rarely,” “occasionally,” or “frequently”) versus never use. Comprehensive data on physician-reported frequency of use were also presented for treatments that were used by ≥80% of respondents in all groups (SCTC, CRA, and ACR) and also for methotrexate (MTX) and D-penicillamine for the treatment of skin involvement (despite not being used by ≥80% of respondents), demonstrating that frequent use was relatively uncommon.
The SCTC, CRA, and ACR response rates to the survey were 56%, 48%, and 19%, respectively. Baseline characteristics are shown in Table 1. The SCTC saw more patients with SSc per year and during training than the ACR or CRA (SCTC was currently seeing 93 patients with scleroderma annually versus 14 in the CRA and 17 in the ACR; P < 0.005). Prescribing practices are presented in Table 2 (drug use versus never use) and Table 3 (frequent use versus all other options).
|Characteristic||SCTC (n = 139)||CRA (n = 302)||ACR (n = 312)||P|
|Respondents, no. (%)||78 (56)||146 (48)||59 (19)||–|
|Age, mean ± SEM years||49 ± 1||48 ± 0.8||49 ± 1.0||0.91|
|Clinical practice (vs. clinical practice and research), no. (%)||14 (19)||89 (64)||41 (77)||0.0001†|
|Practicing other specialty besides rheumatology, no. (%)‡||20 (26)||44 (31)||15 (26)||0.69|
|Time after rheumatology fellowship completed, mean ± SEM years||16 ± 1||17 ± 1||15 ± 1||0.40|
|SSc patients seen per year, mean ± SEM||93 ± 18||14 ± 2||17 ± 2||0.005†|
|SSc patients seen during training, mean ± SEM||69 ± 7||24 ± 6||43 ± 9||0.0001†|
|Estimate of SSc in practice with diffuse disease, %||40 ± 2||36 ± 2||40 ± 3||0.26|
|RCT evidence for treatment in SSc (reference)||SCTC (n = 78)||CRA (n = 146)||ACR (n = 59)||P|
|Methotrexate||N = 2; weak (6, 11)||85||75||62||0.20|
|D-penicillamine||N = 1; neg. (5)||61||76||70||0.12|
|Interstitial lung disease|
|Steroids ( dose)||86||85||82||0.90|
|Calcium-channel blockers||N = 8; pos. moderate effect (18)||100||99||100||0.48|
|ACE inhibitors||N = 1; weak (19)||81||65||75||0.06‡|
|Iloprost||N = 9; pos. (19–21)||84||65||40||0.0001|
|SSRI||N = 1; weak (23)||48||18||36||0.0001‡|
|Steroids ( dose)||30||39||30||0.36|
|Iloprost||N = 3; pos. (19, 21)||81||60||37||0.0001|
|Bosentan||N = 1; pos. (22)||40||12||27||0.0003|
|Bosentan||N = 2; pos. (8, 36, 37)||83||70||79||0.5†|
|Epoprostenol||N = 1; pos. (38)||76||52||61||0.16|
|Treprostinil||N = 1; pos-weak (39)||28||5||18||0.1†|
|Iloprost||N = 1; weak (40)||71||47||29||0.008|
|Manifestation||SCTC (n = 78)||CRA (n = 146)||ACR (n = 59)||P†|
|Interstitial lung disease|
|Steroids ( dose)|
Drug use (including frequent, occasional, and rare use) versus never use
Drug use versus never use was similar between groups for the majority of agents used for skin and joint manifestations. The most commonly used agents (by all rheumatology groups overall) for joint involvement were nonsteroidal antiinflammatory drugs (99.9%), MTX (90%), and prednisone (89%). Seventy-six percent of physicians overall used MTX for skin involvement. However, of those who reported using MTX, only 22% (27% SCTC, 22% CRA, and 14% ACR; P = 0.30) used it frequently. D-penicillamine was used for skin involvement by 70% of respondents, but only 16% reported using it frequently (12% SCTC, 20% CRA, and 12% ACR; P = 0.24).
To treat myositis, all rheumatology groups used (frequently, occasionally, or rarely) prednisone (99.8%) and MTX (91%), with no differences among groups. However, SCTC members were more apt to use azathioprine for myositis (P = 0.03). SCTC members were also more likely to use iloprost for Raynaud's phenomenon (84% versus 65% CRA and 40% ACR; P = 0.0001), iloprost for digital ulcers (81% versus 60% CRA and 37% ACR; P = 0.0001), and bosentan for digital ulcers (40% versus 12% CRA and 27% ACR; P = 0.0003) than either CRA or ACR members. SCTC members were more likely than CRA members (but not ACR members) to use selective serotonin reuptake inhibitors (SSRIs) for Raynaud's phenomenon (48% versus 18% CRA and 36% ACR; P = 0.0001) and antibiotics for digital ulcers (92% versus 76% CRA and 93% ACR; P = 0.006). Results are summarized in Table 2.
Treatment of ILD was variable between groups: SCTC members were more likely than CRA members (but not ACR members) to use cyclophosphamide (99% versus 83% CRA and 93% ACR; P < 0.001) and azathioprine (78% versus 55% CRA and 67% ACR; P = 0.01), and were more likely than both CRA and ACR members to use anti–tumor necrosis factor (anti–TNF) drugs (19% versus 4% CRA and 5% ACR; P = 0.005). However, SCTC members were less likely to use D-penicillamine for ILD (17% versus 39% CRA and 32% ACR; P = 0.03). High-dose steroids were used similarly between groups (Table 2).
Frequent use versus all other options (occasional, rare, and never use)
Frequent use of many drugs, including all drugs for skin disease, was reported by <27% of rheumatologists in any group. SCTC members were more likely than CRA or ACR members to use prednisone for myositis (P = 0.01), cyclophosphamide for ILD (P < 0.0001), and antiplatelet agents for digital ulcers (P = 0.0005) (Table 3).
Pulmonary arterial hypertension and gastroesophageal reflux disease
The SCTC members were more likely to treat PAH (58% versus 17% CRA and 45% ACR; P < 0.0001). The most commonly used agents (for all rheumatology groups overall) were anticoagulants (98%), bosentan (79%), and epoprostenol (66%). ACR members were less likely than CRA and SCTC members to frequently use anticoagulants (26% versus 70% and 72%, respectively; P = 0.003). SCTC members were more likely to use (occasionally or rarely) treprostinil (28% versus 5% CRA and 18% ACR; P = 0.01), but none reported frequent use. With respect to GERD, proton-pump inhibitors and promotility agents were prescribed very frequently by all groups, with no differences among groups (P = 0.50 and P = 0.74, respectively). Eighteen percent of CRA members (versus 3% SCTC and 5% ACR; P = 0.006) never used H2 blockers.
Frequency of disease manifestations and causes of mortality in SSc
Respondents were asked to indicate the percentage of their patients with SSc who demonstrated disease manifestations that could be associated with SSc (see Appendix B). The frequency of disease manifestation estimates varied between the groups (but only slightly), with SSc experts reporting a higher frequency of ILD, PAH, small bowel overgrowth, and erectile dysfunction in their patients than CRA and ACR members. Physicians' estimates of comorbidity and mortality frequencies are presented in Tables 4 and 5.
|Small bowel overgrowth|
|Interstitial lung disease|
|Pulmonary arterial hypertension|
|Causes of mortality||SCTC (n = 78)||CRA (n = 146)||ACR (n = 59)||P|
|SSc-related renal involvement|
|SSc-related heart involvement|
|SSc-related lung fibrosis|
|Other causes with SSc a significant comorbidity|
Several observations from this study are notable when considered in the context of the SSc evidence–based literature. One such observation is that 70% of respondents reported ever using D-penicillamine for skin involvement, but only 16% reported using it frequently (12% SCTC, 20% CRA, and 12% ACR; P = 0.24) (Tables 2 and 3). Even this level of use is interesting considering that a negative trial exists in early diffuse scleroderma (5). Another notable observation, given that the results of clinical trials do not show a clear beneficial effect of MTX (6, 12), is that 76% of physicians overall used MTX for skin involvement; however, only 22% (27% SCTC, 22% CRA, and 14% ACR; P = 0.30) used it frequently (Tables 2 and 3).
Treatment of ILD with cyclophosphamide was variable between groups, but was high for all groups (i.e., >83% for all; P < 0.001) (Table 2). This use is relatively consistent with the evidence where positive anecdotal data exist (and 2 ongoing trials) (13–17). It is also interesting that SCTC members were more likely than CRA members (but not ACR members) to use azathioprine (P = 0.01), and were more likely than both CRA and ACR members to use anti–TNF drugs (P = 0.005) (Table 2); no RCTs or published reports exist concerning this finding.
With respect to Raynaud's phenomenon, all groups used calcium-channel blockers, consistent with positive published trials in the literature (18). SCTC members were also more likely than either the CRA or ACR members to use iloprost for Raynaud's phenomenon (P = 0.0001) and for digital ulcers (P = 0.0001) (Table 2), where positive trial evidence has been published for both Raynaud's phenomenon (19–21) and digital ulcers (19, 21). SCTC members also were more likely than CRA or ACR members to use bosentan for digital ulcers (P = 0.0003) (Table 2), where there is one recently published trial of bosentan for the prevention of digital ulcers (it was not published at the time of the survey) (22). SCTC members were more likely than CRA members (but not ACR members) to use SSRIs for Raynaud's phenomenon (P = 0.0001) and antibiotics for digital ulcers (P = 0.006). There is one positive trial for SSRIs in Raynaud's phenomenon that included idiopathic cases as well as those secondary to SSc (23). No RCTs for antibiotics in digital ulcers secondary to SSc have been conducted.
SCTC members were more likely than CRA or ACR members to treat PAH (58%, 17%, and 45%, respectively; P < 0.0001). Anticoagulants were used by 98% of all physicians combined, but ACR members were less likely than CRA and SCTC members to frequently use anticoagulants (26%, 70%, and 72%, respectively; P = 0.003). A positive association with anticoagulant use and survival has been demonstrated using controlled retrospective data only in primary pulmonary hypertension; however, PAH associated with SSc is often treated similarly to idiopathic PAH, where anticoagulation is standard therapy (24). Although SCTC members were more likely to use (occasionally or rarely) treprostinil (P = 0.01), this medication was not approved in Canada at the time of this survey.
The use of self report of treatment practices and estimated patient frequencies in this study allows the potential for recall bias and measurement error. There was no confirmatory review of medical records to determine the accuracy of answers. We are unaware of any rheumatologic study comparing and validating physician response using actual medical records. One could assume that physicians may overreport best possible care. The response rate was low and there is possibly nonresponse bias (i.e., respondents may be different from the nonrespondents). Self-selection bias may have been present, particularly in the ACR sample where the response rate was low (19%). Respondents to the survey may have had opinions that were not representative of all ACR (or other organization) members, and this should be considered when interpreting survey results. Recall bias while reporting the number of patients with SSc seen during fellowship training certainly occurred due to the long time lag; however, we were interested in the paucity of patients with SSc seen by many trainees in the SCTC, CRA, and ACR (mean ± SEM 69 ± 7, 24 ± 6, and 43 ± 9, respectively).
There is likely a referral bias whereby scleroderma experts see more severe disease or may recognize organ involvement more often, or both. The SCTC members were likely better able to answer how many patients they see annually because many have SSc databases, whereas CRA and ACR members may be less likely to have these databases. Members of the CRA were least likely to use TNF inhibitors in SSc. There were some differences in the treatment of SSc between Canada, the US, and the SCTC, perhaps due to the lack of guidelines and proven interventions. The differences in treating SSc between SCTC members and the other study groups (CRA and ACR) may be confounded by the sample selection (or nonresponse error), as 19% of SCTC members were in clinical practice only (with no research) compared with ∼64% of CRA members and 77% of ACR respondents. Therefore, SCTC members with research-based practices may be more likely to have access to even unpublished SSc literature. However, in most cases the usual response for disease modification therapies was “no treatment,” which supports the lack of positive literature in randomized trials. Interestingly, D-penicillamine is at times still used for skin in scleroderma despite a negative trial in early diffuse scleroderma (5).
An additional limitation of this study was that because this was a cross-sectional study meant to represent a current overview of prescribing trends in the treatment of SSc, it was beyond the scope of our research to assess how physicians' treatment styles change over time in response to new information in the medical literature. We do not know why someone does or does not prescribe a certain drug. Validation of the study responses was also beyond the scope of this study. The poor response rates make generalization to all members of the organizations problematic.
We conclude that scleroderma experts see more patients with SSc and use more immunosuppressives, including unproven treatments. Scleroderma experts may treat more severe cases, but it also may be that ILD and PAH are underrecognized by nonexperts. Erectile dysfunction (25–30) and depression (31–35) seem to be frequent problems in scleroderma, each with a substantial impact on patients, and this has been reported in the literature. Guidelines for the evidence-based treatment of scleroderma are needed to help unify treatment and establish screening for organ involvement, particularly for nonexperts, where effective treatment is known. The systematic dissemination of clinical trial results will be ever more important as several randomized trials will soon be reported, including 3 in ILD (2 with cyclophosphamide, 1 with bosentan) and 1 in healing of digital ulcers with bosentan. Other randomized trials, including high-dose cyclophosphamide versus stem cell transplant in diffuse scleroderma, are recruiting subjects. The data from these trials will undoubtedly help standardize scleroderma treatment, especially if results are positive.
- 1Systemic sclerosis (scleroderma), eosinophilic fasciitis, and calcinosis. In: McCartyDJ, editor. Arthritis and allied conditions. Philadelphia: Lea and Febiger; 1985.
- 3Rheumatoid arthritis and other synovial disorders. In: KlippelJ, DieppeP, editors. Rheumatology. Philadelphia: Mosby; 1998. Section 5: 2.1–6., .
- 10Scleroderma Clinical Trials Consortium. Membership information. URL: http://www.sctc-online.org/membersh.htm.
- 11JMP statistical software, version 4. Cary (NC): SAS Institute; 2000.
- 172005.. Personal communication to Dr. Janet Pope, February
- 20Iloprost and cisaprost for Raynaud's phenomenon in progressive systemic sclerosis. Cochrane Database Syst Rev 2000:CD000953., , , , , , et al.
- 30Erectile dysfunction in men with scleroderma. Scleroderma Care Res 2004; 2: 14–9., .
- 37351 and BREATHE-1 Study Groups. Bosentan for scleroderma associated pulmonary arterial hypertension: a subgroup analysis of two placebo controlled trials [abstract]. Arthritis Rheum 2003; 48 Suppl 9: S456., , for the
|13. How often do you use the following for worsening interstitial lung disease in scleroderma?|
|19. What percentage of your scleroderma patients have the following clinically relevant conditions:|
|a) Small bowel overgrowth?||□ <5%||□ 5–10%||□ 11–20%||□ 21–30%||□ >30%|
|b) Interstitial lung disease?||□ <5%||□ 5–10%||□ 11–20%||□ 21–30%||□ >30%|
|c) Renal crisis?||□ <5%||□ 5–10%||□ 11–20%||□ 21–30%||□ >30%|
|d) Pulmonary hypertension?||□ <5%||□ 5–10%||□ 11–20%||□ 21–30%||□ >30%|
|e) Cardiomyopathy?||□ <5%||□ 5–10%||□ 11–20%||□ 21–30%||□ >30%|
|f) Inflammatory arthritis?||□ <5%||□ 5–10%||□ 11–20%||□ 21–30%||□ >30%|
|g) Depression?||□ <5%||□ 5–10%||□ 11–20%||□ 21–30%||□ >30%|
|h) Erectile dysfunction?||□ <5%||□ 5–10%||□ 11–20%||□ 21–30%||□ >30%|
|i) Other:||□ <5%||□ 5–10%||□ 11–20%||□ 21–30%||□ >30%|