Systemic sclerosis (SSc; scleroderma) is a relatively rare disease whose pathogenesis is characterized by 3 hallmarks: small vessel vasculopathy, production of autoantibodies, and fibroblast dysfunction leading to increased deposition of extracellular matrix. Despite recent advances in certain aspects of disease and further knowledge in some areas of pathophysiology it remains a disabling condition that shortens life expectancy.1

SSc has several clinical variants but essentially presents within two large groups regarding disease-extension: limited cutaneous SSc (lcSSc) where the affected skin is restricted to the distal limbs and face and diffuse cutaneous SSc (dcSSc) where the skin thickening extends proximal to the elbows and may involve the trunk. These two forms may have different degrees of internal organ involvement and be associated with different autoantibodies with different course and prognosis. Therefore, it is important to identify the clinical subset as early as possible.

With the advent of angiontensin converting enzyme inhibitors three decades ago, the principal cause of death which was renal involvement (scleroderma renal crisis) is now controlled more satisfactorily and the principal causes of death today are pulmonary and cardiac.

Pulmonary disease in SSc mainly comprises interstitial lung disease (ILD)-essentially associated with the diffuse form- and pulmonary arterial hypertension (PAH)-usually associated with the limited form of SSc. Over the past 40 years the SSc mortality rate has seen a shift from renal disease to lung involvement (including both ILD and pulmonary hypertension). These have become the primary cause of SSc-related deaths (33% and 28% respectively), replacing SSc renal crisis (with a drop of frequency from 42% to 6% since the awareness of the role of high-dose systemic corticoids in this disease, among other things), whereas the proportion of deaths due to heart disease has not changed significantly over time.2

Cumulative survival of SSc patients from diagnosis is 84.1% at 5 years and 74.9% at 10 years, but when we look particularly into the group of SSc-ILD patients the 10 year survival rate is significantly lower (29–69%) which clearly shows it is important to recognize patients with ILD early and treat them appropriately.3

There are defined risk factors for the development of ILD in patients with SSc: some of them are disease-related like the presence of dcSSc, anti-Scl-70/anti-topoisomerase I antibody and/or absence of anti-centromere antibody 4 and shorter disease duration.6 Others are demographic-related as African–American ethnicity and older age at disease onset. However, none of these risk factors is absolute.

The risk of developing ILD is greatest early in the course of SSc, so much so that particularly in diffuse SSc, the initial approach was to perform a HRCT scan and PFT at diagnosis and annually the first five years. Even though there is some rational knowledge regarding which patient presents an elevated pretest risk of developing SSc-ILD, there is still poor information as to which of these will then progress.5,6

Guler et al. showed different disease behaviour patterns among patients with SSc-ILD, renewing the challenge of finding risk factors and predictors of pulmonary function decline.7 A recent review analyzed the different measures in pulmonary function tests (PFT) used as outcomes for SSc-ILD. It showed that the widespread use of FVC for the Scleroderma Lung Study I and II trials somehow led to a reduction of efforts to identify other possible predictors of SSc-ILD progression, despite the fact that FVC may only weakly reflect the extent of the disease. Ultimately, the best surrogate marker for SSc-ILD onset and progression may be a composite outcome consisting of a combination of two or more measures.8

Over the past decade, several important aspects became apparent regarding ILD as a whole and particularly in SSc, the connective tissue disease more frequently affected and more frequently participating in all drug trials involved when studying ILD in these diseases. The clarification that IPF was not amenable to immunosuppression, the recognition of an ILD in patients without clear-cut connective tissue diseases (IPAF) perhaps responding to immunosuppression, and more recently the demonstration of progression of fibrosis in spite of prior treatment, immunosuppressive or not, have changed our current scenario.9–11 Particularly, the suggestion that a combination of immunosuppression and anti-fibrotic drugs may be useful in some patients has recently been raised.12

Therefore, especially regarding SSc, the early diagnosis of ILD and the identification of possible predictors of progression even under treatment, are very important issues today.

In Latin America and particularly in Argentina, to our knowledge, there is scarce data regarding the epidemiology of SSc, the prevalence of ILD and its course and even less regarding response to treatment.13 A large series from Brazil with almost one thousand patients showed that mortality was greater in males with diffuse SSc and lung and heart disease were the main causes of death.14 A recent report from a University Hospital in Buenos Aires with a cohort of over 200 patients showed an overall prevalence of ILD of 65.3% in the diffuse form and 26.7% in the limited variant. However, no evaluation of progression of ILD was made.15 A study from the centre originating the current data analyzed the response to treatment of 24 patients with SSc and ILD. Eighteen of those had extensive disease as per Goh criteria 16 and 6 had limited lung disease. Twelve of the 18 with extensive disease were treated with cyclophosphamide for a median of 9 months. Although these patients did not improve their lung function, this was stabilized.17

Because of the paucity of overall data in this part of the world, it was our intention to analyze some of the above mentioned aspects in a cohort of SSc patients followed prospectively in a single centre from the Buenos Aires University Hospital (Hospital de Clínicas José de San Martin) evaluated by both rheumatologists and lung specialists.

Our main goal was to describe our SSc population, determine the prevalence of ILD, and see if we could identify differences between patients who progressed and those who did not.

Fig. 1 shows the flow of patients evaluated during the study period. Two hundred and seven fulfilled inclusion criteria. Fifty five were then excluded so as not to confuse results because they had features of overlap with a variety of other connective tissue diseases. One hundred and fifty two then became our study population and their characteristics are shown in Table 1.

Fig. 1.

Flow chart of evaluated SSc patients.

(0,13MB).

Sixty one percent of patients had limited SSc and 32% diffuse, with only 12% with sine scleroderma. There were no significant differences in gender, age at diagnosis, disease duration or smoking history between groups in these patients.

Regarding clinical characteristics, aside from a significantly higher mRSS in the diffuse patients (median 15 vs 4 years), the presence of ILD was more frequent (71% vs 56.6%) but did not reach statistical significance. Other characteristics were similar.

With respect to autoantibodies, as was to be expected, anti-centromere antibodies were significantly more prevalent in limited SSc as anti Scl-70 antibodies were in the diffuse variant.

When considering imaging, diffuse patients had significantly more damage extension in HRCT (68%) than the other groups. We did not have sufficient data regarding HRCT patterns to analyze this information.

Table 2 shows a comparison between patients with and without ILD as per HRCT definition. There were no significant differences in general characteristics such as gender, age, disease duration or smoking history. As expected, ILD was more prevalent in the diffuse variant and patients with ILD had significantly more extensive damage in HRCT.

Patients with ILD had a higher mRSS score but this did not reach statistical significance.

Anti Scl-70 was significantly more prevalent in patients with ILD as opposed to anti-centromere which was significantly more prevalent in those without.

Table 3 shows a comparison between those patients who progressed regarding their ILD versus those who did not. Eighteen patients (19%) progressed. There were no significant differences between the groups regarding gender, subsets of SSc, clinical features or presence of autoantibodies. We did not find any association between degree of HRCT involvement and disease progression.

The only differences we found were a higher risk of progression with a history of smoking, having received immunosuppressive treatment and a shorter disease duration. Using a multivariate logistic regression model, these differences remained, with an OR of 4.97 (1.49–17.96) for a history of smoking and an OR of 15.6 (3.85–106.85) for having received immunosuppressive treatment (Table 4).

Our purpose was to evaluate patients with SSc seen by both lung specialists and rheumatologists at a single University Hospital and study their characteristics particularly regarding interstitial lung disease. These patients were screened routinely at initial visit and did not necessarily have lung symptoms or signs at that time. As opposed to other research analyzing ILD and disease progression 23–25 these were not patients selected because they already had ILD.

Regarding SSc subsets, our patients showed a distribution between limited (61%) and diffuse (32%) in keeping with other reports5,14,15 and particularly the large series arising from EUSTAR with over 3000 patients.24 Clinical and serologic characteristics of both subsets did not differ from previous reports, in particular a higher mRSS and prevalence of anti Scl-70 in the diffuse patients.

The prevalence of ILD in this unselected population was 59%, also similar to other reports.14,15,24,25 Also as previously described, anti Scl-70 was significantly more prevalent in patients with ILD and inversely, anti-centromere significantly more prevalent in those without. In keeping with other series, the prevalence of more damage in HRCT was significantly higher in patients with ILD.

We did not find any significant differences in progression regarding clinical subsets, clinical features or antibody presence. In univariate analysis, disease duration was significantly shorter in those whose ILD progressed. In multivariate analysis the only significant differences were related to smoking status and the use of immunosuppressive treatment.

Regarding disease duration, it is well known that lung involvement preferably occurs within the first five years after onset of SSc so we could speculate that this is more related to this fact than other factors.26 A systematic review of papers evaluating ILD progression have also found a shorter disease duration associated with progression,27 as well as other series.27,28

Regarding immunosuppressive treatment, we may interpret that receiving this was related to the fact that patients with more extensive disease or who had shown progression were those being treated so this may explain this relationship. The OR was high but with a wide confidence interval. However, the minimum value was almost 4.

Other papers have mentioned the limitation of not being able to establish an association with immunosuppression, due to inadequate data because of the retrospective nature of these studies.28

As far as smoking is concerned, it has been analyzed in several reports without finding an association with ILD in SSc.29 We do not have an explanation for this finding.

We did not find an association of progression with autoantibodies,25 pulmonary hypertension, digital ulcers28 or degree of HRCT damage.30

The limitations of our work rely mainly on its retrospective design and the lack of control HRCT. These limitations have been emphasized as inherent to this type of research by several others.6,7,27,28

The strengths lie in the fact that this is a relatively large series for our country, followed by both rheumatologists and lung specialists with interest in this disease with adequate data regarding lung function tests over a relatively prolonged period.

It is clear from several of these reports25–30 that we are lacking a sufficiently useful measure to predict ILD progression in these patients which is essential for the management of this disease. For example, it has been shown that neither CVF decline on its own nor DLCO decline on its own are accurate enough. Periods of stability after decline of these variables and vice versa has been clearly demonstrated.26–28 A lack of association between PFT and imaging has also been reported.30

Finding adequate predictive factors will probably result from prospective trials designed to that effect.

In this local series, with clinical and serologic characteristics similar to those from other parts of the world, we only found that a shorter disease duration, smoking and the use of immunosuppressive therapy were associated with ILD disease progression.

JIE has received fees from Boehringer Ingelheim for lectures and advisory boards.

The other authors have nothing to declare.