Summary
Background
In patients with systemic sclerosis, it is common practice to treat interstitial lung disease (ILD) in patients in whom progression has already occurred. We sought to clarify whether observed progression of systemic sclerosis-associated ILD confers risk for subsequent progression.
Methods
In this multicentre observational cohort study, based on an analysis of prospectively collected data, we included patients with systemic sclerosis-associated ILD aged 18 years or older at diagnosis, who fulfilled the 2013 American College of Rheumatology–European Association of Alliances in Rheumatology systemic sclerosis classification criteria. The main cohort (diagnosed between January 2001 and December 2019) was consecutively followed up annually over 4 years at the Department of Rheumatology at the Oslo University Hospital, Norway, and the Department of Rheumatology at the University Hospital Zurich, Switzerland. We applied four definitions of ILD progression: the primary definition was forced vital capacity (FVC) decline of 5% or more, and secondary definitions included FVC decline of 10% or more, progressive pulmonary fibrosis (PPF), and progressive fibrosing ILD (PF-ILD). We applied these definitions at each annual visit after the first (visit 1). We validated our findings in an enriched cohort that included patients from the main cohort with systemic sclerosis-associated ILD and short disease duration of less than 3 years along with patients diagnosed between January 2003 and September 2019 from the Division of Rheumatology, University of Michigan, Ann Arbor, MI, USA. Multivariable logistic regression analyses were applied to predict ILD progression and its effect on mortality. There was no involvement of people with lived experience in this study.
Findings
Of 231 patients with systemic sclerosis-associated ILD from the main cohort (mean age 48·0 years [SD 14·6], 176 [76%] female and 55 [24%] male), 71 (31%) had ILD progression as defined by an FVC decline of 5% or more between visit 1 and visit 2, 38 (16%) as defined by an FVC decline of 10% or more, 39 (17%) as defined by PPF, and 89 (39%) defined by PF-ILD. In multivariable logistic regression analyses, adjusted for risk factors for progressive systemic sclerosis-associated ILD and immunosuppressive treatment, we found that ILD progression, defined by FVC decline of 5% or more, from visit 1 to visit 2 reduced the risk for further progression from visit 2 to visit 3 (odds ratio [OR] 0·28 [95% CI 0·12–0·63]; p=0·002) and that there was no risk for subsequent progression using the other definitions (FVC decline of ≥10%: 0·57 [0·16–1·99; p=0·38]; PPF: 0·93 [0·39–2·22; p=0·88]; and PF-ILD: 0·69 [0·35–1·36]; p=0·28]). Using the primary definition of progression, we found the same results in the enriched systemic sclerosis-associated ILD cohort, wherein 41 (34%) of 121 patients had progression defined by an FVC decline of 5% or more (OR 0·22 [95% CI 0·06–0·87]; p=0·031). FVC decline of 5% or more was significantly associated with mortality (hazard ratio 1·66 [95% CI 1·05–2·62]; p=0·030) adjusted for other risk factors.
Interpretation
Systemic sclerosis-associated ILD progression does not predict further ILD progression at the next annual follow-up visit, even in an enriched population, but progression was associated with mortality. These results have implications for clinical practice because they support a paradigm shift in treatment strategy, advocating for initiating therapy in patients at risk of progression. Further research is needed to confirm these findings.
Funding
None.
Introduction
Interstitial lung disease (ILD) affects approximately 50% of patients with systemic sclerosis, and a further subset of this group has a progressive disease course.1,2 Although presence of ILD per se is associated with increased mortality, survival is even more impaired in patients with progressive ILD.1,3,4 Several definitions of ILD progression in systemic sclerosis are used in the field. Among these definitions, declining lung function is probably the most widely accepted proxy, with decline in forced vital capacity (FVC) being the primary measure. Additionally, ILD progression is often assessed by worsening of respiratory symptoms and increasing extent of pulmonary fibrosis on high-resolution CT.5 These variables were included in the official 2022 American Thoracic Society (ATS), European Respiratory Society (ERS), Japanese Respiratory Society, and Latin American Thoracic Society Clinical Practice Guideline definition of progressive pulmonary fibrosis (PPF).5 A third definition is derived from the INBUILD trial,6 which evaluated the efficacy of nintedanib in patients with progressive fibrosing ILD (PF-ILD). The PF-ILD definition overlaps with the PPF definition, although worsening can appear within 24 months.
The definitions for ILD progression in systemic sclerosis were developed from those applied for idiopathic pulmonary fibrosis (IPF). However, such adaptation might not be optimal because the natural history of systemic sclerosis-associated ILD differs from that of IPF.7–9 Compared with the prototypical relentless disease progression of IPF, the disease course of systemic sclerosis-associated ILD is more heterogeneous.2 In our recent study2 and in another Canadian study,7 individuals with systemic sclerosis-associated ILD who were followed up over many years had progressive periods followed by stable periods and vice versa.
Current paradigms assume that recent progression of systemic sclerosis-associated ILD identifies patients at risk of further progression; however, whether progression according to any of these criteria reliably predicts subsequent progression in systemic sclerosis-associated ILD remains unclear. Moreover, it is uncertain whether this assumption holds true in patients with systemic sclerosis-associated ILD who have a short disease duration, which is a risk factor for ILD progression.
We aimed to assess whether ILD progression predicts subsequent ILD progression in patients with systemic sclerosis-associated ILD using different definitions of progressive disease in an unselected cohort. Previous studies10 have suggested a higher frequency of progressive ILD in patients with systemic sclerosis of short duration, (eg, <3 years), which has also been found to be a risk factor for progressive ILD in clinical trials. We therefore assessed ILD progression in a cohort enriched with patients with short disease duration.
Methods
Study design and participants
This is an analysis of prospectively collected data using a protocolised and standardised follow-up process.11 In this multicentre cohort study, we included all patients with systemic sclerosis-associated ILD from the cohorts from the Department of Rheumatology at the Oslo University Hospital, Norway, and the Department of Rheumatology at the University Hospital Zurich, Switzerland, who were diagnosed between January 2001 and December 2019. We also compiled a combined cohort including patients with systemic sclerosis-associated ILD from the Oslo and Zurich cohorts who had a disease duration of less than 3 years and added patients with systemic sclerosis-associated ILD (diagnosed between July 2011 and September 2019) from the Division of Rheumatology, University of Michigan, Ann Arbor, MI, USA with a disease duration of up to 3 years—an enrichment criterion for progressive ILD in clinical trials. This cohort was called the enriched cohort.12 Patients from all three sites met the following four criteria: presence of ILD on high-resolution CT diagnosed by expert radiologists; age 18 years or older at the time of diagnosis; fulfilment of the 2013 American College of Rheumatology (ACR)–European Association of Alliances in Rheumatology (EULAR) systemic sclerosis classification criteria;13 and availability of consecutive and protocolised FVC measurements performed annually (±3 months), with the first FVC assessment defined as visit 1.
All patients gave written informed consent. The Regional Committee of Health and Medical Research Ethics in South-East Norway (number 2016/119), the Cantonal Ethic Committee of Zurich (BASEC numbers 2016–01515 and 2018–02165), and the University of Michigan Institutional Review Board committee (IRB number HUM00075979) approvals were gained. This study complied with the Declaration of Helsinki. There was no involvement of people with lived experience in this study.
Procedures and assessments
We defined disease duration as time from first non-Raynaud symptom to visit 1. Immunosuppressive treatment was defined as follows: mycophenolate mofetil, cyclophosphamide, tocilizumab, rituximab, prednisone (≥10 mg per day), azathioprine, or methotrexate. None of the included patients used anti-fibrotic treatment and therefore this was not assessed.
Pulmonary function tests with FVC and diffusing capacity for carbon monoxide (DLCO) were performed according to ATS–ERS guidelines with all centres applying the same standardised equations (The Global Lung Function Initiatove [GLI], 2012 for FVC and 2020 for DLCO).14,15 In the Oslo cohort, all high-resolution CT lung images were assessed by expert radiologists for the presence of ILD and analysed as previously described.1,16 In the Zurich cohort, visual assessment of high-resolution CT images was performed by expert radiologists.5,17 Additionally, all images from the Oslo and Zurich cohorts had radiological reports available that stated whether there was radiological progression. If in doubt, images were re-assessed according to radiological evidence of ILD progression (defined by the ATS guidelines for PPF and visual inspection for PF-ILD) by experienced ILD teams at each site.5,6 No assessments were done in the Michigan cohort. Respiratory symptoms were assessed and graded by treating physicians using a modified medical research council definiton, which classified symptoms into four functional classes at every visit:18 no breathlessness, breathlessness at exercise, breathlessness at easy exercise, and breathlessness at rest. The 6-min walk distance with assessment of oxygen saturation at the end of the 6-min walk test was reported as previously described.19
ILD definitions and study outcomes
For the primary analysis, we defined ILD progression as an absolute FVC decline of 5% or more over a 12-month period (appendix 3 p 10). We set the threshold at 5% or more per year because values above 5% are above biological variability and are associated with increased mortality20 For each individual patient, we defined presence or absence of ILD progression for each individual 12-month period with available data (appendix 3 p 10).
To allow for comparisons, we applied three additional and alternative definitions of ILD progression. The first was absolute FVC decline of 10% or more over a 12-month period. The second was PPF, defined as meeting at least two of the following criteria at 12 months: worsening of respiratory symptoms, absolute decline in FVC of 5% predicted or more, absolute decline in DLCO of 10% predicted or more or radiological evidence of ILD progression as specified in the PPF criteria5 The third was PF-ILD, defined as meeting at least one of the following criteria within the past 24 months: relative FVC decline of 10% or more, relative FVC decline of 5% or more but less than 10% and worsening of respiratory symptoms or an increased extent of fibrosis on high-resolution CT, or worsening of respiratory symptoms and an increased extent of fibrosis on high-resolution CT.6 The PPF and PF-ILD criteria were applied in all patients with at least two criteria available.
For our primary analysis we tested whether absolute FVC decline of 5% or more over a 12-month period was predictive for subsequent FVC decline in the main systemic sclerosis-associated ILD and in the enriched systemic sclerosis-associated ILD cohorts. We also assessed whether absolute FVC decline of 10% or more over a 12-month period, PPF, and PF-ILD were predictive for subsequent ILD progression in the main systemic sclerosis-associated ILD cohort. Last, we assessed whether ILD progression, defined as absolute FVC decline of 5% or more over a 12-month period, would translate into a change in mortality.
Statistical analysis
Comparisons of baseline characteristics and clinical features between those with and without ILD progression were evaluated with the Pearson χ2 test, Fisher’s exact test, and the Kruskal–Wallis t test. For correlations, Pearson or Kendall’s tau-β coefficients were calculated. Descriptive statistics of baseline characteristics and clinical features were applied in the enriched cohort.
Univariable and multivariable logistic regression analyses were applied to analyse the predictive ability of absolute FVC decline of 5% or more for subsequent progressive ILD adjusted for other potential risk factors for progressive ILD in the main and the enriched cohort and were reported as odds ratio (OR) and 95% CIs. Based on previous studies and expert opinion, the following potential risk factors for ILD progression in systemic sclerosis were assessed as covariates: age, sex, disease duration, systemic sclerosis cutaneous subtype, systemic sclerosis specific autoantibodies such as anti-topoisomerase I, extent of skin involvement assessed by the modified Rodnan skin score (mRSS), signs of inflammation measured by C-reactive protein (CRP), tendon friction rubs and arthritis;2,10,21,22 ILD measures (FVC% and DLCO%, extent of lung fibrosis, 6-min walk distance, O2 desaturation on 6-min walk test, and functional class); and immunosuppressive treatment.2,10,12,19,21–25
We applied multivariable models using the 1:10 rule-of-thumb, meaning that we included one variable in the multivariable analysis for eight-to-ten events of absolute FVC decline of 5% or more as well as the other progressive definitions.26 All multivariable analyses were preceded by estimation of correlation between variables. In the multivariable analysis, we used age, sex, and treatment as fixed covariates in the main cohort and treatment in the enriched cohort and carried forward variables into the multivariable analysis that were significant in the univariable analysis or had the lowest p values. Models were tested by receiver operator curve (ROC) analysis to discriminate their accuracy, where values of area under the curve (AUC) of more than 0·7 were considered to indicate acceptable discrimination. Next, in sensitivity analyses, we assessed whether initiation of a new treatment at the first follow-up (visit 2) might be associated with less progression at the following visit (visit 3) in the main cohort. We repeated our multivariable logistics regression analysis excluding those who initiated a new treatment at the first follow-up visit (visit 2). We also conducted regression to the mean analysis adjusting for baseline FVC. We standardised the FVC changes for year 1 and year 2 and stratified patients into quartiles on the basis of their baseline FVC. Scatter plots with fitted regression lines were used to visually inspect the associations. Linear regression analyses were conducted within each quartile to quantify the associations and assess their significance.
We conducted the same univariable and multivariable logistic regression analyses using absolute FVC decline of 10% or more, and PPF and PF-ILD definitions for ILD progression. We adjusted again the multivariable models using the 1:10 rule-of-thumb with at least age and sex followed by treatment and the same covariates as for absolute FVC decline of 5% or more as far as the number of events allowed.
For our primary analysis of absolute FVC decline of 5% or more, univariable and multivariable Cox regression analyses were applied to analyse the effect of ILD progression on mortality reported as hazard ratios (HRs) and 95% CI in the main cohort. We also conducted survival analyses to assess 1-year, 3-year, and 5-year survival using Kaplan–Meier estimates.
Missing data were reported but were not imputed. Analyses were performed with IBM SPSS software, version 28, and STATA software, version 18. A two sided p value of <0·5 was considered significant.
Role of the funding source
There was no funder for this study.
Results
We included 231 patients with systemic sclerosis-associated ILD diagnosed between January 2001 and December 2019 from the Oslo (n=159) and Zurich (n=72) cohorts who had annual follow-up visits (main cohort; table 1). Mean age was 48·0 years (SD 14·6), 176 (76%) of 231 patients were female and 55 (24%) were male. The cohort was representative of a typical systemic sclerosis-associated ILD population with 94 (41%) of 229 patients having diffuse cutaneous systemic sclerosis and 83 (37%) of 223 patients being positive for anti-topoisomerase I antibodies. At visit 1, mean disease duration was 10·4 years (SD 11·6), mean FVC % predicted was 89% (SD 19·9) and mean DLCO % predicted was 64% (18·7%). We found no major significant differences between patients from Oslo and Zurich, except for the proportion of patients with anti-topoisomerase I antibodies, mean C-reactive protein concentrations, mean DLCO % predicted, and the proportion of patients prescribed immunosuppressive medication at visit 1 (appendix 3 p 1).
Table 1:
Baseline characteristics in the main cohort of patients with systemic sclerosis-associated ILD and in those with and without ILD progression between visit 1 and visit 2
| Main cohort (n=231) | Patients who had ILD progression between visit 1 and visit 2 (n=71) | Patients without ILD progression between visit 1 and visit 2 (n=160) | p value | |
|---|---|---|---|---|
| Demographics | ||||
| Age, years | 48·0 (14·6) | 48·8 (14·8) | 47·7 (14·5) | 0·56 |
| Sex | ||||
| Female | 176/231 (76%) | 51/71 (72%) | 125/160 (78%) | 0·30 |
| Male | 55/231 (24%) | 20/71 (28%) | 35/160 (22%) | ·· |
| Smoking status* | ||||
| Ever smoker | 92/230 (40%) | 29/71 (41%) | 63/160 (39%) | 0·86 |
| Non-smoker | 138/230 (60%) | 42/71 (59%) | 96/160 (60%) | ·· |
| Systemic sclerosis characteristics | ||||
| Disease duration, years | 10·4 (11·6) | 9·4 (11·2) | 10·7 (11·8) | 0·41 |
| Disease duration ≤3 years | 68/220 (31%) | 28/71 (39%) | 40/149 (27%) | 0·059 |
| Diffuse cutaneous subtype | 94/229 (41%) | 36/71 (51%) | 58/158 (37%) | 0·046 |
| Anti-topoisomerase I antibodies | 83/223 (37%) | 33/71 (46%) | 50/152 (33%) | 0·051 |
| Modified Rodnan skin score | 10·1 (9·4) | 10·8 (10·3) | 9·8 (8·9) | 0·44 |
| Tendon friction rubs | 14/196 (7%) | 7/62 (11%) | 7/134 (5%) | 0·13 |
| Arthritis | 18/105 (17%) | 9/38 (24%) | 9/67 (13%) | 0·18 |
| C-reactive protein, mg/L | 3·7 (7·4) | 1·9 (2·2) | 4·5 (8·6) | 0·048 |
| ILD-related parameters | ||||
| FVC % predicted | 89% (19·9) | 93% (19·6) | 86% (19·8) | 0·014 |
| DLCO % predicted | 65% (18·7) | 66% (16·3) | 65% (19·7) | 0·85 |
| 6-min walk distance, m | 489 (121·9) | 469 (116·7) | 499 (123·4) | 0·10 |
| Oxygen desaturation | 19/142 (13%) | 6/43 (14%) | 13/99 (13%) | 0·90 |
| Dyspnoea class 3–4† | 35/181 (19%) | 16/55 (29%) | 19/126 (15%) | 0·028 |
| Ongoing treatment at visit 1 | ||||
| Mycophenolate mofetil | 16/228 (7%) | 6/71 (8%) | 10/157 (6%) | 0·57 |
| Other immunosuppressive drug‡ | 75/228 (33%) | 20/71 (28%) | 55/157 (35%) | 0·31 |
| Corticosteroids ≥10 mg per day | 65/228 (29%) | 20/71 (28%) | 45/157 (29%) | 0·94 |
Data are mean (SD) or n (%). DLCO=diffusing capacity of the lungs for carbon monoxide. FVC=forced vital capacity. ILD=interstitial lung disease.
Smoking status in one patient was not registered.
Dyspnoea was graded using the modified Medical Research Council definitions.
Included cyclophosphamide, tocilizumab, rituximab, azathioprine, and methotrexate.
71 (31%) of 231 patients in the main cohort had an FVC decline of 5% or more between visit 1 and visit 2 and were defined as having ILD progression. Compared with patients who did not have ILD progression between visit 1 and visit 2, those who had progression had a higher frequency of diffuse cutaneous systemic sclerosis and anti-topoisomerase I antibodies, shorter disease duration, and higher C-reactive protein concentration, and, despite higher FVC % predicted, they were more often in dyspnoea functional class 3–4 (table 1).
For the enriched cohort we included 68 patients from the main cohort who had disease duration of 3 years or less and 53 additional patients from the Michigan cohort. There were no major significant differences between patients from Oslo, Zurich, and Michigan, apart from mean disease duration, FVC % predicted, DLCO % predicted, and immunosuppressive medication prescribed at visit 1 (appendix 3 p 2). In the enriched cohort, mean age was 52·0 years (SD 12·2), 83 (69%) of 121 patients were female and 38 (31%) were male; other key characteristics are shown in table 2.
Table 2:
Key characteristics of the enriched cohort
| Enriched cohort (n=121) | |
|---|---|
| Age, years | 52·0 (12·2) |
| Sex | |
| Female | 83 (69%) |
| Male | 38 (31%) |
| Disease duration, years | 1·3 (0.8) |
| Anti-topoisomerase I antibodies | 48 (40%) |
| Modified Rodnan skin score | 11·4 (9·7) |
| FVC % predicted | 81 (30·6) |
| Ongoing treatment at visit 1 | |
| Mycophenolate mofetil | 50 (41%) |
| Other immunosuppressive drug* | 28 (23%) |
Data are n (%) or mean (SD). FVC=forced vital capacity. ILD=interstitial lung disease.
Includes cyclophosphamide, tocilizumab, rituximab, azathioprine, and methotrexate.
In the main cohort, among patients with FVC decline of 5% or more from visit 1 to visit 2, only 14 (20%) of 71 had further ILD progression at visit 3 (figure 1). For comparison, among those without ILD progression between visit 1 and visit 2, 50 (31%) of 160 patients had progression between visit 2 and visit 3 (p=0·048). Based on this observation, we next investigated whether ILD progression was independently associated with further progression at follow-up applying logistic regression. Testing all mentioned risk factors for progression in univariable logistic regression (appendix 3 p 3), FVC % predicted at visit 1 (OR 1·02 [95% CI 1·00–1·03]; p=0·020) was significantly associated with ILD progression (from visit 2 to visit 3), whereas FVC decline of 5% or more from visit 1 to visit 2 was not (0·54 [0·28–1·06]; p=0·070). In multivariable logistic regression, adjusted for age, sex, FVC % predicted at visit 1, mRSS at visit 1, anti-topoisomerase I antibodies, and immunosuppressive treatment at visit 1, ILD progression from visit 1 to visit 2 reduced the risk for further progression from visit 2 to visit 3 (OR 0·28 [95% CI 0·12–0·63] p=0·0020) with an acceptable AUC of 0·71 (figure 2A)
Figure 1:
Disease course in the main cohort of patients with systemic sclerosis-associated ILD
Figure 2: Multivariable logistic regression analysis of predictors for ILD progression defined as FVC decline of ≥5% between visit 2 and visit 3 in the main systemic sclerosis-associated ILD cohort (A) and in the enriched cohort (B).
Continuous variables were age, mRSS at visit 1, and FVC% at visit 1 and categorical variables were sex, anti-topoisomerase I antibody, FVC decline of ≥5% between visit 1 and visit 2, and immunosuppressive treatment between visit 1 and 2. None of the continuous predictors had a presumed non-linear effect with the log odds of the outcome. mRSS=modified Rodnan skin score. FVC=forced vital capacity. ILD=interstitial lung disease.
To further substantiate these findings, we also looked at risk for ILD progression from visit 3 to visit 4 (appendix 3 p 3). We found that 64 (32%) of 203 patients in the main cohort had ILD progression, defined by FVC decline of 5% or more, between visits 3 and 4 (figure 1). In multivariable logistic regression, ILD progression in the year before was associated with a reduced risk of further progression (OR 0·44 [95% CI 0·19–0·99], p=0·046), while no other variables were significantly associated (appendix 3 p 4).
Reduced risk of further ILD progression after an observed progression could possibly be explained by changes of treatment in response to the observed progression. To address this hypothesis, we compared treatment changes at visit 2 in patients with and without ILD progression from visit 1 to visit 2 in a sensitivity analysis (table 3). A higher proportion of patients started a new immunosuppressive drug among those with progression than among those without progression (16 [23%] of 71 vs 20 [13%] of 157), but the difference was not significant (p=0·067). To further assess effects of treatment changes on ILD progression, we excluded patients who initiated a new treatment from the multivariable logistic regression. These analyses showed that ILD progression from visit 1 to visit 2 reduced the risk for further ILD progression irrespective of treatment changes (OR 0·31 [95% CI 0·13–0·74]; p=0·013; appendix 3 p 5). We did not identify other respiratory illnesses or any specific reason as drivers for no change of therapy by chart review of these individuals.
Table 3:
Treatment status and change in use of immunosuppressive drugs at visit 3 in patients with and without ILD progression from visit 1 to visit 2 in the main cohort
| No immunosuppressive drugs ever | Started new immunosuppressive drugs | Same immunosuppressive drugs as on visit 1 | Stopped immunosuppressive drugs | |
|---|---|---|---|---|
| Patients without ILD progression (n=157) | 83 (53%) | 20 (13%) | 49 (31%) | 5 (3%) |
| Patients with ILD progression (n=71) | 36 (51%) | 16 (23%) | 16 (23%) | 3 (4%) |
ILD=interstitial lung disease.
The observed results could have at least partly been explained by regression to the mean. Our analyses did not reveal evidence of regression to the mean in the dataset, as indicated by the scatter plots showing a consistent negative trend in all quartiles. Linear regression analyses confirmed significant negative associations in each quartile adjusted for baseline FVC (p>0·05; appendix 3 p 11). These results indicate a genuine association between FVC % predicted change in year 1 and year 2, not driven by regression to the mean.
In the enriched cohort, 41 (34%) of 121 patients had ILD progression between visit 1 and visit 2 defined by an annual FVC decline of 5% or more and 27 (22%) of 121 had further progression by visit 3 (appendix 3 p 6). In multivariable regression analyses, adjusted for FVC % predicted at visit 1 and presence of anti-topoisomerase I antibodies, both FVC decline of 5% or more from visit 1 to visit 2 (OR 0·22 [95% CI 0·06–0·87]; p=0·031) and immunosuppressive treatment at visit 1 (0·18 [95% CI 0·06–0·56]; p=0·003) reduced the risk for further progression from visit 2 to visit 3, with an acceptable AUC of 0·77 (figure 2B).
To further address the risk conferred by observed ILD progression, we repeated the prediction analyses using different definitions of ILD progression. Of 231 patients in the main cohort, 211 (91%) had high-resolution CT images available and 163 (71%) had respiratory symptoms available and the occurrence of PPF and PF-ILD could be determined in all 231 patients on the basis of at least two criteria. Between visit 1 and visit 2, 39 (17%) of 231 patients had progressed on the basis of PPF and PF-ILD criteria, 89 (39%) had progressed on the basis of PF-ILD, and 38 (16%) had progressed on the basis of an FVC decline of 10% or more. Next, we tested whether ILD progression was an independent predictor for progression at visit 3 applying univariable and multi variable logistic regression (appendix 3 pp 7–9). Irrespective of definition, we did not find that ILD progression was a risk factor for subsequent progression (table 4).
Table 4:
Multivariable logistic regression analysis of risk for further ILD progression at subsequent follow-up (visit 3) in patients with initial ILD progression, defined by PPF, PF-ILD, or FVC decline of ≥10% between visit 1 and visit 2, main cohort
| Progression at visit 3 | Odds ratio (95% CI) | p value | AUC | |
|---|---|---|---|---|
| Initial progression defined by PPF (39 [17%] of 231) | 38/231 (16%) | 0·93 (0·39–2·22)* | 0·88 | 0·59 |
| Initial progression defined by PF-ILD (89 [39%] of 231) | 63/231 (27%) | 0·69 (0·35–1·36)† | 0·28 | 0·63 |
| Initial progression defined by FVC decline of ≥10% (38 [16%] of 231) | 28/231 (12%) | 0·57 (0·16–1·99)‡ | 0·38 | 0·53 |
AUC=area under the curve. FVC=forced vital capacity. ILD=interstitial lung disease PF-ILD=progressive fibrosing ILD. PPF=progressive pulmonary fibrosis.
Adjusted for age, sex, and immunosuppressive treatment used at visit 1.
Adjusted for age, sex, FVC % predicted at visit 1, modified Rodnan skin score at visit 1, and immunosuppressive treatment used at visit.
Adjusted for age and sex.
Over a median of 9·8 years (range 2·2–19·8) of follow-up, 81 (35%) of 231 patients in the main cohort died due to any cause. In univariable regression analysis, absolute FVC decline of 5% or more was significantly associated with death (HR 1·62 [95% CI 1·04–2·54]; p=0·033). In multivariable Cox regression, absolute FVC decline of 5% or more was still significantly associated with death (HR 1·66 [1·05–2·62]; p=0·030) adjusted for age (1·03 [1·01–1·05]; p<0·001), male sex (1·30 [0·80–2·10]; p=0·29), diffuse cutaneous systemic sclerosis (1·10 [0·69–1·73]; p=0·69) and baseline FVC (0·98 [0·97–0·99]; p=0·002). For patients with FVC decline of 5% or more at visit 2, 1-year survival was 93%, 3-year survival was 83%, and 5-year survival was 76% compared to patients without progression, with 98%, 89%, and 82% survival rates, respectively (p=0·030; appendix 3 p 12).
Discussion
In clinical practice, physicians often wait for ILD progression before initiating or escalating therapy in patients with systemic sclerosis-associated ILD. Similarly, patients with systemic sclerosis-associated ILD that have already progressed are recruited into trials to enrich for further progression. These strategies assume that patients with recent ILD progression have a higher risk for further progression; however, this assumption has not been analysed in patients with systemic sclerosis-associated ILD and there is no evidence to support such treatment and recruitment strategies. In this study, ILD progression defined by an FVC decline of 5% or more was not followed by further progression in most patients with systemic sclerosis-associated ILD as verified in an enriched cohort with short disease duration. Furthermore, we showed that ILD progression was protective for further progression in the subsequent period. To verify this initial finding, we used several definitions of ILD progression, including PPF and PF-ILD, which supported our initial finding. We explored several possible confounders including initiation of new immunosuppressive therapies. We also found that an FVC decline of 5% or more over a 12-month period was associated with increased risk of death in adjusted analyses and reduced 1-year, 3-year, and 5-year survival significantly.
These findings have important implications for the clinical care of patients. The disease course of ILD in patients with systemic sclerosis is heterogeneous.2,7 Patients with systemic sclerosis-associated ILD can have periods of longer-term stability with intermittent periods of progression.2 Therefore, in daily practice, when to initiate or escalate treatment for ILD in patients with systemic sclerosis is often unclear. The overall aim is to avoid overtreatment of stable patients, but not to miss patients who are at risk of progression and lung damage. Optimally, clinicians would apply a risk score consisting of multiple parameters clearly separating patients who will progress from those who will be stable. Indeed, several parameters associated with ILD progression have been identified, such as short disease duration, anti-topoisomerase I antibody positivity, progression of skin fibrosis, and increased inflammatory markers.10,23,24 However, their predictive capacity for ILD progression is weak and inconsistent between studies because of differences in clinical ILD phenotype under study.27 Moreover, whether identification of these parameters truly provides aid for treatment decisions or leads to better outcomes in patients has yet to be determined.
Accordingly, because an established risk algorithm for progression is not available, the current expert advice is attentive waiting with careful monitoring for patients who do not have very severe ILD or who are without proven progression.28 Adhering to this advice, clinicians will typically observe a patient over 6–12 months and only initiate or escalate treatment if lung function decline is seen.25,28 The basic premise for this treatment strategy is that progression of systemic sclerosis-associated ILD is a continuous process, where any observed decline in lung function predicts subsequent deterioration. Because this premise does not hold true, as shown in the current study, we are at risk of over-treating some patients displaying recent short-term lung function decline, while under-treating those who appear stable but progress after a longer period of stability.
Importantly, if clinicians only treat patients with observed progression, they lose the opportunity of very early treatment intervention, potentially leading to lung damage and worse outcomes. Indeed, severe ILD and progressive ILD have been found to be associated with worse outcomes,1,4,29 and in this study we show that even a single FVC decline of 5% or more is associated with increased mortality and reduced survival. Therefore, preventing progression of ILD is of high importance. In IPF, the international guideline recommends to start treatment in patients with preserved lung function before lung function decline, with an aim to prevent progression and to improve quality of life and long-term outcomes.8,30 We suggest that this strategy should also be considered in patients with systemic sclerosis-associated ILD, for whom even more treatment options are available.28,31 However, more research is necessary to identify patients with systemic sclerosis-associated ILD who have a high likelihood of stability to avoid overtreatment. Importantly, this study did not address treatment strategies after progression has occurred. Because a single episode of progression over a 12-month period is associated with long-term mortality, we advise to follow treatment guidelines to consider escalation or switching of therapy after progression has occurred.32,33
These data might also be informative for the design of clinical trials. Enrichment for patients at high risk of ILD progression by recruiting patients who recently progressed has become a popular design of ILD studies; however, these data do not support such a strategy for systemic sclerosis-associated ILD. Our findings suggest that using progressive ILD as an inclusion criterion in studies might be recruiting patients who are subsequently stable and less at risk of ILD progression. Importantly, our findings must be confirmed in clinical trial cohorts. Patients in clinical trials might have a different disease course than patients in clinical practice, although objective criteria that are used for the definitions of progression, such as FVC and imaging, should not be profoundly different between these settings. Additionally, results from the INBUILD trial6 indicate that our findings might be specific for patients with systemic sclerosis-associated ILD or specific forms of ILD. In the INBUILD trial,6 patients with PF-ILD with aetiologies other than IPF were included, and over the 52-week study period, 49% and 67% of patients in the placebo group had an FVC decline of 10% or more and of 5% or more, respectively. How many of these patients had an FVC decline in the year before study inclusion is not reported, and only a minority of these patients had systemic sclerosis-associated ILD. Nevertheless, this high number of patients with ILD progression indicates that either other forms of ILD are in general more progressive than systemic sclerosis-associated ILD, or that continuous progression is much more common in them, and enrichment of cohorts via inclusion of patients with progression might be a successful enrichment strategy in those forms of ILD. Thus, separate analysis of different ILD aetiologies is required to identify risk factors for ILD progression and to characterise specific disease courses.
This study has some limitations. First, by only including patients with consecutive lung function assessments (51% of the total Oslo cohort and 54% of the total Zurich cohort), we could have potentially enriched for patients with severe or progressive ILD. However, the number of patients with progression of ILD per year in this study is consistent with anticipated numbers.2 Second, at the time of study inclusion, patients had varying disease duration, but even by assessing patients with short disease duration in our enriched cohort, which is a known risk factor for ILD progression in systemic sclerosis-associated ILD from the Michigan cohort,10 we obtained the same results as in our main analysis. Third, our results cannot be extrapolated to other forms of ILD that might have different disease courses. Fourth, most patients in the study were White (data not shown) and so our findings need to be validated in other ethnic groups. Fifth, we did not adjust our analyses for socioeconomic status or unobserved confounding, which could possibly have an effect. Finally, we cannot exclude that progression might have happened between the annual visits. We chose to use 12-month periods of analysis because most of the validated progressive definitions that we applied (FVC decline of ≥5% and ≥10% and PPF) are based on 12-month changes. Additionally, in clinical practice, follow up is often conducted annually, and outcome assessments of randomised controlled trials are often conducted after 1 year. Furthermore, high-resolution CTs are rarely conducted in shorter periods than 12 months.
The strength of our study is that we included patients from three well defined systemic sclerosis-associated ILD cohorts with geographical differences but no major significant clinical differences. Serial clinical, pulmonary, treatment, and high-resolution CT imaging data were very comprehensive and derived from multidisciplinary collaboration including specialties involved in the care of patients with systemic sclerosis-associated ILD. Therefore, we could apply all ILD progression definitions and adjust for treatment and risk factors at several follow-up visits. To our knowledge, this study is the first to apply different definitions of ILD progression in a systemic sclerosis-associated ILD cohort. Furthermore, a possible explanation for our findings could have been regression to the mean, but our comprehensive analyses revealed no evidence to support this.
In conclusion, ILD progression in patients with systemic sclerosis-associated ILD does not predict further progression at the following annual visit when applying ILD progression definitions including FVC decline, PF-ILD, and the PPF criteria, and even a single FVC decline of 5% or more is associated with worse outcome. These results challenge current treatment practices because, although initiating or escalating treatment should still be done in patients with ILD progression according to guidelines, our data suggest that treatment should be initiated or escalated in patients at risk of progression to prevent poor outcomes, although confirmation is needed to show a benefit of this strategy on improved outcome. These findings might also have important implications for clinical trial design because selection of patients with systemic sclerosis-associated ILD with ILD progression does not seem to result in a population with a higher likelihood of subsequent ILD progression.
Supplementary Material
Research in context.
Evidence before this study
We previously showed that individual patients with systemic sclerosis-associated interstitial lung disease (ILD) with follow-up over many years had periods when ILD progressed followed by stable periods and vice versa. Current paradigms assume that recent progression of systemic sclerosis-associated ILD identifies patients who are at risk of further progression; however, whether ILD progression according to any progressive criteria reliably predicts subsequent progression in patients with systemic sclerosis-associated ILD is unclear. Moreover, it is common practice to treat patients in whom progression has already occurred rather than treating those at risk of progression, but there is limited evidence supporting this strategy.
Added value of this study
These findings have implications for clinical practice because they do not lend support to the current practice of waiting for progression before initiating or escalating treatment in patients with systemic sclerosis-associated ILD. Additionally, we provide crucial insights for clinical trial design, cautioning against the inclusion of only patients with systemic sclerosis-associated ILD who have recently progressed, because such selection might inadvertently select for individuals who are subsequently stable. We also show that progression is associated with mortality and reduces survival.
Implications of all the available evidence
These findings suggest that a proactive treatment approach should be considered in patients with systemic sclerosis-associated ILD who are at risk of ILD progression, and that clinical trials should refine patient selection criteria to avoid inadvertently including stable patients. These changes could improve long-term outcomes and enhance the accuracy of clinical trial results.
Acknowledgments
During the preparation of this work, the authors used ChatGPT in order to verify spelling and grammar. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.
Declaration of interests
A-MH-V: speakers bureau fees from Boehringer Ingelheim, Janssen, Medscape, Merck Sharp & Dohme, Roche; consultant fees from Calluna Pharma, Boehringer Ingelheim, Genentech, Janssen, Medscape, Merck Sharp & Dohme, Pliant, Roche; and grant or research support from Boehringer Ingelheim and Janssen. HF: speakers bureau fees and consulting fees from Bayer and Boehringer Ingelheim; and travel support from Actelion and GSK. MOB: speakers bureau fees and consulting fees from GSK, Amgen, Novartis, and Vifor. HJB: grant and research support from Jannsen. CBruni: grants from FOREUM, European Scleroderma Trials and Research Gruppo Italiano Lotta alla Sclerodermia, Scleroderma Clinical Trials Consortium, AbbVie Foundation, and Wellcome Trust; speakers bureau fees and consulting fees from Boehringer Ingelheim and Eli Lily; and travel support from Boehringer Ingelheim. CC: speakers bureau fees and consulting fees from Boehringer Ingelheim, GSK, AstraZeneca, Sanofi, Vifor, Grifols, OM Pharma, Daiichi Synkyo, and CSL Behring. PPD: speakers bureau fees and consulting fees from Boehringer Ingelheim. RD: speakers bureau fees and consulting fees from Actelion and Boehringer Ingelheim; and grants from Actelion and Pfizer. MTD: speakers bureau fees and consulting fees from Boehringer Ingelheim. ME: travel support from Janssen and AstraZeneca. TF: speakers bureau fees and consulting fees from Bayer and Boehringer Ingelheim; and travel support from AstraZeneca. CM: speakers bureau fees and consulting fees from MEDtalks Switzerland, Mepha, PlayToKnow AG, Medbase AG, Romanian Society of Rheumatology, Boehringer Ingelheim, and Mepha; and travel support from Roche and Boehringer Ingelheim. MS: grants from AbbVie. JO: consulting fees from Boehringer Ingelheim, Lupin Pharmaceuticals, AmMax Bio, Roche, and Veracyte; patent on TOLLIP TT genotype for NAC use in idiopathic pulmonary fibrosis; participation in data monitoring and safety or advisory boards for Endeavor Biomedicines, Novartis, and Genentech; and stock options on Gatehouse Bio. DK: consulting and speaker fees for Amgen, AbbVie, Argynx, AstraZeneca, Boehringer Ingelheim, Cabaletta, Merck, Mirador, Mitsubishi Tanabe, Novartis, Renovare Therapeutics, Philikos, and Zura Bio; and grant or research support from AbbVie, Amgen, Boehringer Ingelheim, and the Scleroderma Research Foundation. OD: speakers bureau fees and consulting fees from 4P-Pharma, AbbVie, Acceleron, Alcimed, Altavant, Amgen, AnaMar, Arxx, AstraZeneca, Baecon, Blade, Bayer, Boehringer Ingelheim, Corbus, CSL Behring, Galderma, Galapagos, Glenmark, Gossamer, iQvia, Horizon, Inventiva, Janssen, Kymera, Lupin, Medscape, Merck, Miltenyi Biotec, Mitsubishi Tanabe, Novartis, Prometheus, Redxpharma, Roivant, Sanofi, and Topadur; a patent issued “mir-29 for the treatment of systemic sclerosis” (US8247389, EP2331143); grant or research support from Boehringer Ingelheim, Kymera, and Mitsubishi Tanabe; and cofounder of Citus AG. All other authors declare no competing interests.
Footnotes
For the German translation of the abstract see Online for appendix 1
For the Norwegian translation of the abstract see Online for appendix 2
For the modified medical research council definitions see https://www.pcrs-uk.org/sites/default/files/resources/MRC-Score.pdf
See Online for appendix 3
Data sharing
An anonymised, minimal dataset can be made available upon reasonable request to the corresponding author (a.m.hoffmann-vold@medisin.uio.no) after publication of this Article and after approval of the respective local ethic committees.
References
- 1.Hoffmann-Vold A-M, Fretheim H, Halse A-K, et al. Tracking impact of interstitial lung disease in systemic sclerosis in a complete nationwide cohort. Am J Respir Crit Care Med 2019; 200: 1258–66. [DOI] [PubMed] [Google Scholar]
- 2.Hoffmann-Vold A-M, Allanore Y, Alves M, et al. Progressive interstitial lung disease in patients with systemic sclerosis-associated interstitial lung disease in the EUSTAR database. Ann Rheum Dis 2021; 80: 219–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Volkmann ER, Tashkin DP, Roth MD, Goldin J, Kim GHJ. Early radiographic progression of scleroderma: lung disease predicts long-term mortality. Chest 2022; 161: 1310–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Goh NS, Hoyles RK, Denton CP, et al. Short-term pulmonary function trends are predictive of mortality in interstitial lung disease associated with systemic sclerosis. Arthritis Rheumatol 2017; 69: 1670–78. [DOI] [PubMed] [Google Scholar]
- 5.Raghu G, Remy-Jardin M, Richeldi L, et al. Idiopathic pulmonary fibrosis (an update) and progressive pulmonary fibrosis in adults: an official ATS/ERS/JRS/ALAT clinical practice guideline. Am J Respir Crit Care Med 2022; 205: e18–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Flaherty KR, Wells AU, Cottin V, et al. Nintedanib in progressive fibrosing interstitial lung diseases. N Engl J Med 2019; 381: 1718–27. [DOI] [PubMed] [Google Scholar]
- 7.Guler SA, Winstone TA, Murphy D, et al. Does systemic sclerosis-associated interstitial lung disease burn out? specific phenotypes of disease progression. Ann Am Thorac Soc 2018; 15: 1427–33. [DOI] [PubMed] [Google Scholar]
- 8.Lederer DJ, Martinez FJ. Idiopathic pulmonary fibrosis. N Engl J Med 2018; 378: 1811–23. [DOI] [PubMed] [Google Scholar]
- 9.Wijsenbeek M, Cottin V. Spectrum of fibrotic lung diseases. N Engl J Med 2020; 383: 958–68. [DOI] [PubMed] [Google Scholar]
- 10.Distler O, Assassi S, Cottin V, et al. Predictors of progression in systemic sclerosis patients with interstitial lung disease. Eur Respir J 2020; 55: 1902026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hoffmann-Vold A, Distler O, Murray B, et al. Setting the international standard for longitudinal follow-up of patients with systemic sclerosis: a Delphi-based expert consensus on core clinical features. RMD Open 2019; 5: e000826. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ramahi A, Lescoat A, Roofeh D, et al. Risk factors for lung function decline in systemic sclerosis-associated interstitial lung disease in a large single-centre cohort. Rheumatology (Oxford) 2023; 62: 2501–09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.van den Hoogen F, Khanna D, Fransen J, et al. 2013 classification criteria for systemic sclerosis: an American College of Rheumatology/European League Against Rheumatism Collaborative Initiative. Ann Rheum Dis 2013; 72: 1747–55. [DOI] [PubMed] [Google Scholar]
- 14.Quanjer PH, Stanojevic S, Cole TJ, et al. Multi-ethnic reference values for spirometry for the 3–95-yr age range: the global lung function 2012 equations. Eur Respir J 2012; 40: 1324–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hall GL, Filipow N, Ruppel G, et al. Official ERS technical standard: Global Lung Function Initiative reference values for static lung volumes in individuals of European ancestry. Eur Respir J 2021; 57: 2000289. [DOI] [PubMed] [Google Scholar]
- 16.Hoffmann-Vold AM, Aaløkken TM, Lund MB, et al. Predictive value of serial high-resolution computed tomography analyses and concurrent lung function tests in systemic sclerosis. Arthritis Rheumatol 2015; 67: 2205–12. [DOI] [PubMed] [Google Scholar]
- 17.Goh NS, Desai SR, Veeraraghavan S, et al. Interstitial lung disease in systemic sclerosis: a simple staging system. Am J Respir Crit Care Med 2008; 177: 1248–54. [DOI] [PubMed] [Google Scholar]
- 18.Hoffmann-Vold A-M, Allanore Y, Bendstrup E, et al. The need for a holistic approach for SSc-ILD—achievements and ambiguity in a devastating disease. Respir Res 2020; 21: 197. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wu W, Jordan S, Becker MO, et al. Prediction of progression of interstitial lung disease in patients with systemic sclerosis: the SPAR model. Ann Rheum Dis 2018; 77: 1326–32. [DOI] [PubMed] [Google Scholar]
- 20.Zappala CJ, Latsi PI, Nicholson AG, et al. Marginal decline in forced vital capacity is associated with a poor outcome in idiopathic pulmonary fibrosis. Eur Respir J 2010; 35: 830–36. [DOI] [PubMed] [Google Scholar]
- 21.Meier FM, Frommer KW, Dinser R, et al. Update on the profile of the EUSTAR cohort: an analysis of the EULAR Scleroderma Trials and Research Group database. Ann Rheum Dis 2012; 71: 1355–60. [DOI] [PubMed] [Google Scholar]
- 22.Hoffmann-Vold AM, Brunborg C, Airò P, et al. Cohort enrichment strategies for progressive interstitial lung disease in systemic sclerosis from European Scleroderma Trials and Research. Chest 2023; 163: 586–98. [DOI] [PubMed] [Google Scholar]
- 23.Nihtyanova SI, Schreiber BE, Ong VH, et al. Prediction of pulmonary complications and long-term survival in systemic sclerosis. Arthritis Rheumatol 2014; 66: 1625–35. [DOI] [PubMed] [Google Scholar]
- 24.Khanna D, Tashkin DP, Denton CP, Renzoni EA, Desai SR, Varga J. Etiology, risk factors, and biomarkers in systemic sclerosis with interstitial lung disease. Am J Respir Crit Care Med 2020; 201: 650–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Roofeh D, Distler O, Allanore Y, Denton CP, Khanna D. Treatment of systemic sclerosis-associated interstitial lung disease: lessons from clinical trials. J Scleroderma Relat Disord 2020; 5 (suppl): 61–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 1996; 49: 1373–79. [DOI] [PubMed] [Google Scholar]
- 27.Kuwana M, Avouac J, Hoffmann-Vold A, et al. Development of a multivariable prediction model for progression of systemic sclerosis-associated interstitial lung disease. RMD Open 2024; 10: e004240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hoffmann-Vold A-M, Maher TM, Philpot EE, et al. The identification and management of interstitial lung disease in systemic sclerosis: evidence-based European consensus statements. Lancet Rheumatol 2020; 2: e71–83. [DOI] [PubMed] [Google Scholar]
- 29.Moore OA, Goh N, Corte T, et al. Extent of disease on high-resolution computed tomography lung is a predictor of decline and mortality in systemic sclerosis-related interstitial lung disease. Rheumatology 2013; 52: 155–60. [DOI] [PubMed] [Google Scholar]
- 30.Wells AU, Brown KK, Flaherty KR, Kolb M, Thannickal VJ. What’s in a name? That which we call IPF, by any other name would act the same. Eur Respir J 2018; 51: 1800692. [DOI] [PubMed] [Google Scholar]
- 31.Hoffmann-Vold AM, Distler O, Crestani B, Antoniou KM. Recent advances in the management of systemic sclerosis-associated interstitial lung disease. Curr Opin Pulm Med 2022; 28: 441–47. [DOI] [PubMed] [Google Scholar]
- 32.Del Galdo F, Lescoat A, Conaghan PG, et al. EULAR recommendations for the treatment of systemic sclerosis: 2023 update. Ann Rheum Dis 2024; 84: 29–40. [DOI] [PubMed] [Google Scholar]
- 33.Raghu G, Montesi SB, Silver RM, et al. Treatment of Systemic Sclerosis-associated Interstitial Lung Disease: Evidence-based Recommendations. An Official American Thoracic Society Clinical Practice Guideline. Am J Respir Crit Care Med 2024; 209: 137–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
An anonymised, minimal dataset can be made available upon reasonable request to the corresponding author (a.m.hoffmann-vold@medisin.uio.no) after publication of this Article and after approval of the respective local ethic committees.