Introduction

Pulmonary disease is commonly reported in association with connective tissue disease (CTD). One of the more severe forms of pulmonary involvement in CTD is interstitial lung disease (ILD), which refers to varying degrees of inflammation and fibrosis involving the interstitial compartment of the lung. CTD–ILD can range from an incidental finding on radiographic imaging to a rapidly progressive illness leading to respiratory failure and death. It has been described in up to 90% of patients with a definable CTD depending on the modality of diagnosis. The highest prevalence is observed in patients with systemic sclerosis (SSc), rheumatoid arthritis (RA), and polymyositis/dermatomyositis (PM/DM) . CTD–ILD can have a significant adverse effect on quality of life and is a leading cause of mortality, highlighting the importance of accurate diagnosis and appropriate clinical management . Given the complexities of diagnosis and paucity of treatment trials, management of patients with CTD–ILD is challenging. In this study, we will discuss the distinguishing features of CTD–ILD in SSc, RA, and PM/DM; recent advances in clinical management; and the steps needed to improve our ability to care for patients with CTD–ILD, including the role of longitudinal observational cohorts, well-designed clinical trials, and a structured approach to provide patient care.

Characteristic features of CTD–ILD

CTD–ILD is a broad term including any diffuse parenchymal lung disease in patients with a definable CTD or a constellation of signs, symptoms, and lab abnormalities suggestive of a CTD. Many CTD patients are diagnosed with ILD after overt respiratory symptoms, such as exertional breathlessness or cough, prompt an evaluation with pulmonary function testing (PFT) and/or radiographic imaging. However, a subset of CTD patients have no pulmonary symptoms and are diagnosed with ILD incidentally after chest imaging is obtained for other reasons (e.g., lung cancer screening, coronary calcium screening, evaluation for pulmonary embolism), an entity referred to as “subclinical” CTD–ILD. ILD can also occur in the presence of autoimmune features that do not meet classification criteria for a specific CTD, a condition often referred to as lung-dominant CTD or interstitial pneumonia with autoimmune features (IPAF) .

At present, the exact etiology of CTD–ILD is not known. It has been hypothesized that the lung may be an innocent organ that is targeted by disease-specific autoantibodies generated elsewhere. Another hypothesis is that certain subtypes of CTD begin with a lung-based process. In the latter, it is proposed that a lung injury that triggers local inflammation induces autoantigen expression that can in turn lead to autoantibody generation in the lung. This process could be perpetuated by subsequent binding of the disease-associated autoantibodies and antigen in the lung, leading to further lung inflammation and fibrosis . In line with this hypothesis, in RA patients, generation of disease-specific antibodies has been shown in the lung , and the presence of RA-related antibodies in the lung is associated with more severe lung damage in RA–ILD . Also supporting this hypothesis, ILD can precede joint disease in a subset of patients with RA–ILD .

While there are many similarities among the different subsets of CTD–ILD, some distinguishing features are also important to consider ( Table 1 ). For example, the majority of CTD–ILD patients display a pattern on high-resolution computed tomography (HRCT) or histopathology of nonspecific interstitial pneumonia (NSIP) with or without organizing pneumonia (OP) ( Fig. 1 ). NSIP is a more inflammatory subtype of ILD and the most common pattern observed in SSc (68–77%) , PM/DM (65–82%) , Sjogren’s syndrome (28–61%) , and undifferentiated connective tissue diseases (UCTD) (83%) . By contrast, RA patients with ILD have a higher incidence of a pattern of usual interstitial pneumonia (UIP) on HRCT or histopathology, a more fibrotic subtype of ILD ( Fig. 2 ) . In general, the specific ILD pattern can predict response to treatment, with the more inflammatory ILDs (cellular NSIP and OP) showing a greater response to immunosuppression than the more fibrotic ILDs (fibrotic NSIP and UIP).

Nonspecific interstitial pneumonia (NSIP) in a patient with scleroderma (SSc). A basilar predominate process with reticulation and ground glass and without honeycombing. Courtesy of Joshua Solomon, MD, National Jewish Health.

Usual interstitial pneumonia (UIP) pattern in a patient with rheumatoid arthritis (RA). A basilar and peripheral pattern of reticulation and honeycombing with minimal ground glass opacities. Courtesy of Joshua Solomon, MD, National Jewish Health.

Prognosis also differs between subsets of CTD–ILD, although this is mostly reflective of the underlying histopathologic differences. For example, RA–UIP has the worst prognosis of the CTD–ILDs , with a 5-year survival rate of 36%. However, RA–NSIP has a 5-year survival of 94% . Similarly, the 5-year survival of SSc-, PM/DM-, and Sjogren’s syndrome-associated ILD patients, with a predominance of NSIP, ranges from 60% to 85% .

Characteristic features of CTD–ILD

CTD–ILD is a broad term including any diffuse parenchymal lung disease in patients with a definable CTD or a constellation of signs, symptoms, and lab abnormalities suggestive of a CTD. Many CTD patients are diagnosed with ILD after overt respiratory symptoms, such as exertional breathlessness or cough, prompt an evaluation with pulmonary function testing (PFT) and/or radiographic imaging. However, a subset of CTD patients have no pulmonary symptoms and are diagnosed with ILD incidentally after chest imaging is obtained for other reasons (e.g., lung cancer screening, coronary calcium screening, evaluation for pulmonary embolism), an entity referred to as “subclinical” CTD–ILD. ILD can also occur in the presence of autoimmune features that do not meet classification criteria for a specific CTD, a condition often referred to as lung-dominant CTD or interstitial pneumonia with autoimmune features (IPAF) .

At present, the exact etiology of CTD–ILD is not known. It has been hypothesized that the lung may be an innocent organ that is targeted by disease-specific autoantibodies generated elsewhere. Another hypothesis is that certain subtypes of CTD begin with a lung-based process. In the latter, it is proposed that a lung injury that triggers local inflammation induces autoantigen expression that can in turn lead to autoantibody generation in the lung. This process could be perpetuated by subsequent binding of the disease-associated autoantibodies and antigen in the lung, leading to further lung inflammation and fibrosis . In line with this hypothesis, in RA patients, generation of disease-specific antibodies has been shown in the lung , and the presence of RA-related antibodies in the lung is associated with more severe lung damage in RA–ILD . Also supporting this hypothesis, ILD can precede joint disease in a subset of patients with RA–ILD .

While there are many similarities among the different subsets of CTD–ILD, some distinguishing features are also important to consider ( Table 1 ). For example, the majority of CTD–ILD patients display a pattern on high-resolution computed tomography (HRCT) or histopathology of nonspecific interstitial pneumonia (NSIP) with or without organizing pneumonia (OP) ( Fig. 1 ). NSIP is a more inflammatory subtype of ILD and the most common pattern observed in SSc (68–77%) , PM/DM (65–82%) , Sjogren’s syndrome (28–61%) , and undifferentiated connective tissue diseases (UCTD) (83%) . By contrast, RA patients with ILD have a higher incidence of a pattern of usual interstitial pneumonia (UIP) on HRCT or histopathology, a more fibrotic subtype of ILD ( Fig. 2 ) . In general, the specific ILD pattern can predict response to treatment, with the more inflammatory ILDs (cellular NSIP and OP) showing a greater response to immunosuppression than the more fibrotic ILDs (fibrotic NSIP and UIP).

Nonspecific interstitial pneumonia (NSIP) in a patient with scleroderma (SSc). A basilar predominate process with reticulation and ground glass and without honeycombing. Courtesy of Joshua Solomon, MD, National Jewish Health.

Usual interstitial pneumonia (UIP) pattern in a patient with rheumatoid arthritis (RA). A basilar and peripheral pattern of reticulation and honeycombing with minimal ground glass opacities. Courtesy of Joshua Solomon, MD, National Jewish Health.

Prognosis also differs between subsets of CTD–ILD, although this is mostly reflective of the underlying histopathologic differences. For example, RA–UIP has the worst prognosis of the CTD–ILDs , with a 5-year survival rate of 36%. However, RA–NSIP has a 5-year survival of 94% . Similarly, the 5-year survival of SSc-, PM/DM-, and Sjogren’s syndrome-associated ILD patients, with a predominance of NSIP, ranges from 60% to 85% .

Challenges of CTD–ILD diagnosis

Diagnosis of ILD among CTD patients is challenging and requires interdisciplinary discussions and knowledge of unique nuances related to specific CTD subtypes. Particular challenges exist around symptom and physiologic evaluation. Dyspnea may be underreported because of functional limitations from joint or muscle disease or from generalized fatigue from systemic inflammation; cough may be underreported in myositis patients with vocal cord weakness or be a symptom of gastrointestinal reflux (in SSc) or xerotrachea (in RA with secondary Sjogren’s syndrome). PFTs in PM/DM may be misleading as restrictive physiology and reduction in the uncorrected diffusing capacity for carbon monoxide (DLCO), features suggestive of ILD, may be a sign of diaphragmatic weakness only. In SSc, PFTs may be normal in up to 62% of patients with significant radiographic ILD or reveal restrictive physiology in the absence of ILD due to severe chest wall cutaneous fibrosis. In RA, reductions in forced vital capacity (FVC) may not be apparent until there is an extensive parenchymal involvement and the presence of airway abnormalities, commonly observed in RA , may affect the interpretation of the presence or severity of restriction. Moreover, an isolated reduction in DLCO in PM/DM and SSc patients may be a sign of pulmonary hypertension, given its high incidence in these subtypes of CTD .

Given the risks associated with HRCT screening and the aforementioned challenges with physiologic testing, the solution for routine/periodic evaluation to detect early ILD may be to develop an interdisciplinary evaluative strategy combining rheumatologic and pulmonary-specific variables predictive of ILD. Studies examining the correlation of clinical and physiologic variables with HRCT-defined ILD are a first step to develop strategic diagnostic algorithms, but the addition of disease specific biomarkers may also be required to accurately predict CTD patients who will develop ILD. In RA, it has been shown that the combination of symptoms (phlegm and dyspnea), clinical features (prednisone use and smoking status), and CTD-specific autoantibodies can be used to predict the presence of abnormal PFTs in a cohort of patients who are not suspected of ILD . In a multicenter study of RA patients, male gender, rheumatoid factor, older age at onset of RA, and smoking were linked with the presence of ILD on HRCT . In SSc, anti-topoisomerase I antibodies are predictive of ILD on HRCT among patients with normal FVC on PFT .

Subtypes of CTD–ILD

SSc-associated ILD SSc is characterized by multiorgan involvement, endothelial dysfunction, and excessive collagen deposition . Lung involvement is highly prevalent and clinically relevant in many, but not all of these patients. When unselected SSc patients are screened with HRCT, up to 90% have some degrees of ILD . With improvement in the management of renal disease over the past decades, lung disease, including ILD and pulmonary hypertension, has become the most common cause of death in patients with SSc .

ILD occurs early in the course of SSc , and is more common in patients with anti-topoisomerase antibodies and diffuse skin involvement . HRCT most commonly reveals the pattern of NSIP with end-stage fibrosis less commonly observed . Histologically, patients with SSc display bland paucicellular fibrosis occurring uniformly throughout the interstitium with preservation of the architecture in a fibrotic NSIP pattern .

Although CT abnormalities are common, only 13% of patients with SSc will have severe ILD (defined as an FVC ≤ 50% predicted) . Outcome in SSc–ILD does not depend on the underlying histopathology; those with either a UIP or NSIP pattern on biopsy have 5- and 10-year survival rates of 82–90% and 29–69%, respectively . Patients who develop severe SSc–ILD tend to have progression in their lung disease in the first 2 years after its diagnosis and the progression of SSc–ILD is independent of the duration of the underlying SSc . A normal HRCT at baseline predicts a low likelihood of developing SSc–ILD (85% will have a normal CT at 5 years of follow-up) .

RA-associated ILD RA is characterized by a symmetric inflammatory arthritis that can lead to joint damage and disability. Patients with RA can also develop a variety of extra-articular manifestations, including a high prevalence of lung disease. Although all compartments of the respiratory system can be involved in RA, ILD is of the highest concern. ILD is seen in 19–58% of patients depending on the screening method used , is clinically significant in 7–14% , and accounts for 10–20% of all deaths .

RA–ILD can occur throughout the course of disease. It has been identified in patients with recent-onset RA , and ILD has even been reported to precede joint disease in some patients who develop classifiable RA . Risk factors for ILD include smoking, male gender, and high-titer anti-cyclic citrullinated peptides (CCP) . Unlike the other CTD–ILDs, the fibrotic pattern of UIP is most common in these patients, with NSIP observed less frequently. Lung biopsy is often not needed as radiographic pattern has a good correlation with underlying histopathology .

RA–ILD is progressive in a significant number of patients and outcome is related to the underlying histopathology. A recent study found 5-year survival rates of 36% in patients with UIP and 94% in those with NSIP . Baseline physiology and changes in physiology over time are predictive of outcome .

Myositis-associated ILD PM and DM are inflammatory myopathies characterized by varying degrees of muscle inflammation and non-musculoskeletal manifestations. Although there are multiple pulmonary complications in these patients, ILD is the most common non-musculoskeletal manifestation, which can be observed in up to 75% of patients . ILD can predate the myositis in up to 20% of patients and has been shown to be an independent risk factor for death .

Clinically, patients often present insidiously with cough and exertional breathlessness, although a small subset of them have rapidly progressive ILD with respiratory failure , particularly those associated with melanoma differentiation-associated protein 5 (MDA-5) antibodies and clinically amyopathic DM . The most common HRCT pattern in these patients is NSIP with overlapping areas of organizing pneumonia, with UIP less commonly observed . Muscle weakness must be considered when evaluating the physiologic impairments in these patients. The 5-year survival rate of patients with PM/DM and ILD ranges from 60% to 80% .

A subset of patients with PM/DM will have the antisynthetase syndrome, characterized by ILD, anti-aminoacyl-tRNA synthetase antibodies (anti-ARS antibodies) and some combination of fever, arthralgia, Raynaud’s phenomenon, and exanthema on the hands (known as mechanic’s hands) . Although the presence of any of the anti-ARS antibodies predicts ILD, its strongest association is with anti-Jo-1 and anti-PL-12 .

CTD–ILD versus IPF

Idiopathic pulmonary fibrosis (IPF) is a specific subtype of a chronic progressive fibrotic ILD that is seen in older adults. It is limited to the lung and has no identifiable cause. Exclusion of an underlying CTD is critical in the diagnosis of IPF. In fact, recent guidelines for the diagnosis and management of IPF have recommended that all patients with IPF undergo serologic evaluation to determine the presence of rheumatoid factor, anti-CCP and antinuclear antibody testing even in the absence of signs or symptoms of CTD . The guidelines further state that patients with IPF who have a mildly positive antinuclear antibody and/or rheumatoid factor without any other clinical features of CTD are still considered to have IPF and not CTD–ILD. Complicating this distinction between CTD–ILD and IPF is the recent recognition that ILD may be the initial manifestation of CTD (with the extra-pulmonary clinical manifestations of the CTD to follow months or years later) and that the signs and symptoms of CTD may be subtle. Given these subtleties, interdisciplinary assessments involving pulmonologists and rheumatologists are crucial in the evaluation of patients with IPF and may change diagnosis and pharmacologic management .

Fortunately, some notable differences exist between CTD–LD and IPF that can aid in making the correct diagnosis ( Table 2 ). Patients with CTD–ILD are often younger , more commonly female , diagnosed earlier in the course of their disease , and have better lung function at diagnosis than those with IPF. Smokers may also be overrepresented in the IPF group . As mentioned earlier, CTD–ILD patients will more commonly have an NSIP pattern on HRCT . CTD–ILD patients can also have other features on HRCT suggestive of a CTD, such as pleural effusions and pericardial abnormalities , esophageal dilatation and pneumomediastinum . If a surgical lung biopsy is obtained, CTD–ILD often displays features of NSIP with or without OP . Even in patients with CTD–ILD and a UIP pattern on biopsy, there is typically a higher prevalence of germinal centers and lymphoid follicles and fewer fibroblastic foci than those with IPF.

The distinction between CTD–ILD and IPF is clinically relevant in terms of prognosis and treatment. Patients with CTD–ILD have a more favorable outcome than IPF patients , and this can help physicians prognosticate for their patients. In the largest study to date, Park and colleagues have studied 93 CTD patients with biopsy-proven ILD and found survival of the group as a whole was better than an idiopathic cohort with similar lung function (mean 131 months compared with 81 months, p = 0.001) . This finding was primarily driven by a higher prevalence of NSIP in the CTD group and improved survival in CTD–UIP patients compared with idiopathic UIP patients. RA–ILD is one notable exception, where survival is comparable to IPF . An accurate ILD diagnosis is crucial because there are now two Food and Drug Administration-approved therapies for IPF (pirfenidone and nintedanib ) as well as treatments that are found to be harmful in these patients (the combination of prednisone, azathioprine (AZA), and n-acetylcysteine; and warfarin ).