Introduction

Interstitial lung disease (ILD) is a common manifestation of systemic autoimmune diseases (also referred to as connective tissue diseases) including systemic sclerosis (SSc) [1], rheumatoid arthritis (RA) [2], mixed connective tissue disease (MCTD) [3], polymyositis/dermatomyositis [4], and Sjögren’s syndrome [5]. The reported prevalence of ILD in patients with systemic autoimmune disease varies widely depending on the methodology used and the population studied. A proportion of patients with systemic autoimmune diseases develop a progressive fibrosing form of ILD, characterized by increasing fibrosis on high-resolution computed tomography (HRCT), worsening of lung function, and early mortality [6,7,8]. The proportion of patients with autoimmune ILDs who develop progressive fibrosis is not known, but in an online survey, rheumatologists and pulmonologists with experience in managing patients with ILDs estimated it to be 24–31% of patients [9]. Combining information from this survey with a systematic review of published literature, it was estimated that 13–40% of patients with autoimmune ILDs develop a progressive fibrosing phenotype [10].

In this article, we will describe the impact of ILD in patients with systemic autoimmune diseases, the current management of autoimmune ILDs, and the potential role of anti-fibrotic therapy in the treatment of autoimmune ILDs with a progressive fibrosing phenotype.

ILDs in autoimmune diseases

Systemic autoimmune diseases are associated with a range of pulmonary manifestations including pleural disease (pleuritis, pleural effusion, and pleural thickening), airways disease, pulmonary hypertension, vasculitis, diffuse alveolar hemorrhage, and ILD [11]. ILD often appears early in the course of systemic autoimmune disease and may even be the first manifestation of the disorder [11,12,13]. In addition, there exists a group of patients with interstitial pneumonia who show some features of autoimmunity but do not meet diagnostic criteria for any recognized autoimmune disease [14, 15].

ILD is particularly common in patients with SSc; indeed, ILD (defined as pulmonary fibrosis seen on HRCT or chest radiography, most pronounced in the basilar portions of the lungs, or the occurrence of idiopathic “Velcro” crackles on auscultation) is included in the classification criteria for SSc [16]. In the Canadian Scleroderma Research Group registry, ILD was evident on HRCT in 64% of patients who had HRCT scans available, while 26% of patients had ILD according to the presence of basilar “Velcro” crackles on auscultation and 22% had ILD based on chest X-ray [1]. In a recent analysis of data from 21 referral centers in Spain, 43% of 1374 patients with SSc had evidence of pulmonary fibrosis on chest radiograph or HRCT [17]. ILD is a rarer occurrence in patients with RA, but still affects a large number of patients. The prevalence of ILD in patients with RA varies across studies due to differences in the populations studied and the criteria used to define ILD. In an analysis of data from 150 consecutive outpatients with RA (irrespective of signs/symptoms of ILD) at a single UK center, 28 (6.2%) had ILD on HRCT [18], while a multi-center study of 1460 patients with early RA (< 2 years of symptoms) found that 52 patients (3.6%) had ILD at baseline or developed it within 3 years of follow-up [19]. In a retrospective analysis of medical records from 582 patients with RA, the lifetime risk of developing ILD was estimated to be 7.7%, compared with 0.9% in a matched cohort without RA [2]. Patients with more active RA, measured using Disease Activity Score using 28 joints (DAS28) or based on markers of disease activity such as erythrocyte sedimentation rate, appear to be at increased risk of developing ILD [2, 19, 20]. A prospective analysis of 1419 patients with RA at a US center found that patients with high or moderate disease activity defined by DAS28 had a 2-fold increased risk of developing ILD compared with those in remission or with low disease activity over a mean follow-up of 5.6 years. For every unit increase in the DAS28, the risk of ILD increased by 35% [20].

Radiological and histopathological patterns are observed at varying frequencies across different forms of autoimmune disease–associated ILD. In SSc-ILD, the typical pattern on HRCT or histology is non-specific interstitial pneumonia (NSIP) pattern, although a usual interstitial pneumonia (UIP) pattern may also be observed [21, 22]. NSIP is also the predominant pattern on HRCT or biopsy in patients with ILD associated with Sjögren’s syndrome or polymyositis/dermatomyositis [23,24,25]. In contrast, the most common pattern seen on HRCT or lung biopsy in patients with RA-ILD is UIP [2, 13, 26]. In an analysis of 230 patients with RA-ILD in the BRILL network in the UK, 65% of patients had UIP on HRCT while 24% had NSIP [26]. It is important to note that patterns evident on HRCT are not fixed but may evolve over time; for example, patients who have a largely non-fibrotic pattern on HRCT at the time of diagnosis may later develop a more fibrotic form of the disease [27].

The symptoms of cough, dyspnea, and fatigue associated with ILD can contribute to the impairment of quality of life in patients with systemic autoimmune diseases. A survey of 50 patients with RA-ILD at a US center found that the severity of fatigue and dyspnea were the strongest predictors of physical health impairment, while the severity of cough, fatigue, and dyspnea were the strongest predictors of mental health impairment [28]. Dyspnea has been reported by patients with SSc as one of the main factors driving their functional disability [29].

Progression of fibrosing ILDs

Progression of autoimmune ILDs is characterized by deterioration in lung function, which may be assessed through measurement of forced vital capacity (FVC) (lung volume) and diffusion capacity of the lungs for carbon monoxide (DLco) (the lungs’ capacity for gas exchange) [30]. Lung function may decline rapidly after ILD is diagnosed. In a study of 695 patients with SSc in the European Scleroderma Trials and Research (EUSTAR) cohort, approximately one-third of patients had DLco < 50% predicted within 3 years of the onset of Raynaud’s phenomenon [31]. In an analysis of 167 patients with RA-ILD referred to a tertiary care center, the proportion of patients with FVC < 50% predicted increased from 14% at diagnosis to 22% after 5 years, while the proportion of patients with DLco < 40% predicted increased from 29% at diagnosis to 40% after 5 years; patients with a UIP pattern on HRCT were more likely to progress to DLco < 40% predicted than those with NSIP [8]. The course of ILD progression in patients with other autoimmune diseases is not well documented, but it is clear that ILD progresses in a significant proportion of patients with these diseases. Among 107 patients with ILD associated with polymyositis/dermatomyositis treated with immunosuppressants, 16% had a decline in FVC of ≥ 10% predicted and/or a decline in DLco of ≥ 15% predicted over a median follow-up of 34 months [32]. In an analysis of 18 patients with Sjögren’s syndrome and ILD followed for a median of 38 months, 5 (28%) had a decline in FVC of ≥ 10% predicted or a decline in DLco of ≥ 15% predicted despite immunosuppression [33]. In an analysis of 75 patients with anti-synthetase syndrome-ILD receiving anti-inflammatory therapy, 6 patients (8.0%) had a decline in FVC of > 10% predicted and/or a decline in DLco of > 15% predicted 1 year after diagnosis of ILD. Of 36 patients who had an increase in FVC of > 10% predicted and/or increase in DLco of > 15% predicted after 1 year, 12 (33%) had a deterioration over the following 2 years [34].

ILD is a major cause of death in patients with systemic autoimmune diseases. In an analysis of 1072 deaths in patients with SSc in the EUSTAR database, ILD was listed as the cause of death in 16.8% of cases, followed by pulmonary arterial hypertension (14.7%) and cancer (13.1%) (Fig. 1) [35]. A greater extent of fibrosis on HRCT and decline in FVC or DLco are established predictors of mortality in patients with SSc-ILD [7, 36,37,38]. A seminal study found that patients with extensive disease, defined as > 30% extent of fibrosis on HRCT, or 10–30% extent of fibrosis on HRCT plus FVC < 70% predicted, had much greater mortality than patients with more limited disease (HR 3.46) [36]. The presence of ILD also markedly increases mortality in patients with RA [2, 39, 40]. Among 679 patients in a Danish registry, one-year mortality in patients with RA-ILD was 14%, compared with 4% in RA patients without ILD who were matched for age, gender, and time since diagnosis of RA (Fig. 2) [40]. Ten-year mortality in these groups was 60% and 35%, respectively [40]. Predictors of mortality in patients with RA-ILD include the presence of a UIP pattern on HRCT (Fig. 3), a greater extent of fibrosis on HRCT, and lower FVC or DLco [26, 41,42,43,44]. In an analysis of 137 patients with RA-ILD at a single US center, a baseline FVC below the mean of the cohort (68.7% predicted) and a decline from baseline in FVC ≥ 10% predicted at any time over the follow-up period (median 4.8 years) were predictors of mortality (HRs 1.46 and 2.57, respectively) after controlling for age, sex, smoking, and HRCT pattern [41]. A greater extent of fibrosis on HRCT has also been associated with worse survival in patients with MCTD [3].

Fig. 1
figure 1

Causes of death in patients with SSc in the European Scleroderma Trials and Research (EUSTAR) cohort. Adapted from [35]. Republished with permission of Ann Rheum Dis, from Mapping and predicting mortality from systemic sclerosis, Elhai et al., 76(11), 2017; permission conveyed through Copyright Clearance Center, Inc.

Full size image
Fig. 2
figure 2

Kaplan–Meier estimate of mortality in patients with RA-ILD compared with patients with RA without ILD (matched by age, sex, and time since diagnosis of RA). Adapted from [40]. Republished with permission of Ann Rheum Dis, from A population-based cohort study of rheumatoid arthritis-associated interstitial lung disease: comorbidity and mortality, Hyldgaard C et al., 76(10), 2017; permission conveyed through Copyright Clearance Center, Inc.

Full size image
Fig. 3
figure 3

Kaplan–Meier estimate of survival in patients with RA-ILD and UIP or NSIP patterns on HRCT. Adapted from [41]. Reproduced with permission of the © ERS 2019. European Respiratory Journal Feb 2016, 47 (2) 588–596; DOI: https://doi.org/10.1183/13993003.00357-2015. This material has not been reviewed prior to release; therefore, the European Respiratory Society may not be responsible for any errors, omissions or inaccuracies, or for any consequences arising therefrom, in the content

Full size image

Treatment of autoimmune ILDs

Not all patients with autoimmune disease–related ILD require treatment. Pharmacotherapy with immunosuppression is generally reserved for patients with clinically significant, progressive disease [45, 46], while those with non-progressive ILD may remain untreated but should be monitored closely for signs of progression. There is no agreed definition of disease progression in ILD, but in practice, assessment of disease progression is likely to be based on clinicians’ assessment of symptoms, pulmonary function tests, and/or imaging. With this in mind, and considering the other manifestations of the disease that need to be considered when managing patients with autoimmune diseases, cross-disciplinary collaboration between pulmonologists and rheumatologists is key to identifying patients with autoimmune disease-ILD who require treatment and deciding on the therapy that should be given [47].

Immunosuppressants are the standard of care for systemic autoimmune diseases and may be effective in slowing the progression of ILD in some patients. However, other than in SSc-ILD, there is no evidence from randomized double-blind trials to support the efficacy of immunosuppressants in treating ILD. In the absence of a robust evidence base to inform therapeutic decision-making, the choice of therapy for autoimmune ILD is based on clinical experience rather than data.

Data from randomized controlled trials support the use of cyclophosphamide (CYC) and mycophenolate mofetil (MMF) to treat SSc-ILD. In the Fibrosing Alveolitis in Scleroderma Trial (FAST), subjects were randomized to receive low-dose prednisolone plus CYC for 6 months followed by oral azathioprine for 6 months, or placebo for 1 year. At the end of the treatment period, there was an improvement of 2.4% in FVC % predicted with active treatment versus a decline of 3.0% with placebo [48]. In Scleroderma Lung Study I (SLS I), among 145 patients with SSc who completed 6 months of treatment, the mean decline from baseline in FVC % predicted at 1 year was 1.0% in patients treated with CYC versus 2.6% in those treated with placebo [49]. Based on the results of these trials, guidelines for the treatment of SSc-ILD issued by the European League Against Rheumatism Collaborative Initiative (EULAR) recommend the use of tailored CYC therapy, in particular for patients with progressive disease [50, 51]. However, the toxicity of CYC limits it to short-term use. The latest EULAR guidelines did not consider the findings of SLS II, which showed that treatment with oral MMF for 2 years resulted in the same improvement in FVC as oral CYC for 1 year followed by placebo for 1 year (absolute changes of 2.19% and 2.88% predicted, respectively), with better tolerability [52]. However, MMF is now the most widely used therapy for SSc-ILD and has been endorsed as first-line therapy by consensus panels of experts [46, 53, 54]. Rituximab and azathioprine are also used in the treatment of SSc-ILD, but the evidence that these drugs may preserve or improve lung function in these patients comes solely from retrospective or open-label studies [55,56,57,58].

Beyond SSc-ILD, more robust trial data are needed to determine the role of immunosuppressant therapy for autoimmune-associated ILDs. Although CYC, MMF, rituximab, and azathioprine are also commonly used in the treatment of RA-ILD, the evidence to support their use comes solely from retrospective or observational studies [59,60,61,62]. In a retrospective analysis of 206 patients with RA-ILD in the BRILL network, 21 had received pulsed CYC for progressive ILD; survival time was 72 months in patients treated with CYC compared with 43 months in patients with better baseline lung function who did not receive CYC [61]. In an observational study at a single UK center, lung function changes were assessed in 44 patients with RA-ILD treated with rituximab between 2004 and 2015 (total follow-up 195 patient-years) [62]. The median relative change in FVC was 1.2% predicted in the 6–12 months after initiation of rituximab compared with a decline of 2.4% predicted in the 6–12 months before treatment [62]. At the latest time point with evaluable data, 30 patients (68%) did not meet the criteria for ILD progression (i.e., did not have a decline in FVC of > 10% predicted or DLco of > 15% predicted, worsening of ILD score on HRCT, or death from progressive lung disease) [62]. In an uncontrolled open-label Spanish registry, of 63 patients with RA-ILD treated with abatacept for 12 months, FVC % predicted remained stable in approximately two-thirds of patients and improved by ≥ 10% from baseline in approximately one-fifth of patients [63]. A number of other retrospective/observational studies have shown preservation or improvement in lung function in patients with other autoimmune ILDs treated with immunosuppressants including polymyositis/dermatomyositis [59, 64] and anti-synthetase syndrome [65, 66]; however, these studies were uncontrolled and lung function was measured over limited time periods (1 to 4 years).

There is some evidence linking the use of DMARDs such as methotrexate to the development or worsening of ILD in rare cases [67], but a causal relationship has not been established. In a retrospective analysis of 78 patients with RA-ILD at a Mexican referral center, patients taking methotrexate prescribed to treat ILD had better survival compared with those not receiving methotrexate [68].

Potential role of anti-fibrotic therapy in fibrosing autoimmune ILDs with a progressive phenotype

Research is ongoing into whether the addition of anti-fibrotic therapies that have been shown to slow the progression of lung fibrosis in patients with IPF might be added to immunosuppressant therapy to slow the progression of fibrosing ILD in patients with systemic autoimmune diseases. Non-clinical studies suggest that there are commonalities in the mechanisms that drive progressive fibrosis between ILDs associated with immunological diseases and ILDs with other causes [69,70,71,72]. Progressive fibrosing ILDs are believed to be triggered by repetitive injuries at alveolar epithelial or microvascular endothelial sites, which lead to cell destruction and unregulated repair. Fibroblasts orchestrated to the sites of injury are activated to become myofibroblasts, which secrete excessive amounts of extracellular matrix (ECM), resulting in increased tissue stiffness, which in turn further activates and stimulates fibroblasts [69, 70]. Macrophages and lymphocytes recruited to the site of injury release pro-fibrotic mediators, and vascular damage leads to the activation and degranulation of platelets, which release pro-fibrotic mediators. These mediators further promote fibroblast activation, driving a self-sustaining process of progressive fibrosis [70, 73, 74]. Coagulation pathways that are activated by tissue damage may also drive fibrosis [75, 76]. As fibrosis progresses, the accumulation of ECM increases the diffusion distances between blood vessels and cells, reducing the oxygen supply to tissues [77]. Vascular alterations and reduced capillary density may also reduce oxygen supply [77]. In a self-sustaining loop, tissue hypoxia may stimulate further production of ECM proteins [78, 79]. Several epigenetic modifications have also been implicated in the activation of fibroblasts in fibrosing lung diseases [80, 81].

The commonalities in the pathogenic mechanisms that drive fibrosis with different triggers suggest that drugs that slow disease progression in patients with IPF might also inhibit the progression of other ILDs. Nintedanib and pirfenidone reduce the rate of decline in lung function in patients with IPF [82, 83]. The mechanism of action of pirfenidone remains unclear, but it has been shown to reduce the proliferation of fibroblasts and myofibroblasts and to inhibit ECM synthesis and deposition [84, 85]. Nintedanib has demonstrated anti-fibrotic, anti-inflammatory, and vascular remodelling effects in several animal models resembling aspects of fibrosing ILDs [72, 86,87,88,89,90,91,92,93]. In Fra2 mice, a model of the fibrotic and vascular manifestations of SSc, nintedanib reduced hydroxyproline levels and myofibroblast counts in the lung, as well as reducing the thickening of the walls of pulmonary arteries [90]. In transgenic SKG mice, a model resembling aspects of RA-ILD, nintedanib reduced hydroxyproline levels and α-smooth muscle actin staining in the lungs [93].

The Phase III SENSCIS® trial assessed the efficacy and safety of nintedanib in 576 patients with SSc-ILD [94]. Nintedanib reduced the annual rate of decline in FVC versus placebo in a broad range of patients, including those with diffuse cutaneous SSc and limited cutaneous SSc and patients who were and were not taking mycophenolate at baseline. The safety profile of nintedanib was similar to that observed in patients with IPF, characterized mainly by mild or moderate gastrointestinal events, particularly diarrhea [94], and the gastrointestinal adverse event profile was similar in patients who were and were not taking mycophenolate [95]. Other large trials of anti-fibrotic therapies in patients with ILDs related to systemic autoimmune diseases are ongoing (Table 1). The INBUILD® trial (NCT02999178) is investigating the effects of nintedanib versus placebo in patients with progressive fibrosing ILDs other than IPF, including those associated with autoimmune diseases [96]. SLS III (NCT03221257) is studying the efficacy and safety of pirfenidone added to MMF versus MMF alone in patients with SSc-ILD, while the effects of pirfenidone versus placebo in patients with RA-ILD are being investigated in TRAIL I (NCT02808871). Based on their metabolism [98, 99], drug-drug interactions between anti-fibrotic therapies and immunosuppressant therapies are not expected, but further data are needed.

Table 1 Ongoing and recently completed trials in patients with ILDs related to systemic autoimmune diseases
Full size table

Conclusions

ILD is a major cause of morbidity and mortality in patients with systemic autoimmune diseases. Immunosuppressant therapy is the mainstay of therapy for these diseases, but there is very limited evidence to support its efficacy in treating ILD apart from SSc-ILD. Clinical and mechanistic similarities among progressive fibrosing ILDs suggest that drugs known to slow disease progression in patients with IPF may also slow the progression of ILD associated with systemic autoimmune diseases. In the SENSCIS trial, nintedanib reduced the annual rate of decline in FVC in a broad range of patients with SSc-ILD, including those taking mycophenolate, suggesting that in future, the paradigm for treating autoimmune ILDs may shift to a combination of immunosuppressant and anti-fibrotic therapy.