capacity for carbon monoxide (DLCO) [10]. Many of the medications used to treat RA have been reported to be associated with the development of drug-induced ILD though evidence of causation is lacking. Methotrexate has long been speculated to contribute to ILD in a minority of RA patients [18] though recent data suggest no increased incidence in patients on long-standing treatment [34]. The anti-TNF agents infiximab [35], etanercept [36], and adalimumab [37] have all been associated with the development of ILD or exacerbation of pre-existing ILD in case reports, but large population-based studies have not supported an increased risk [38, 39].

In addition to clinical risk factors, genetic risk factors are being identi ed. The HLA shared epitope (SE) is known to be a risk factor for RA and, though older studies did notnd an association between the presence of SE alleles and ILD [6, 16, 40], a recent study looking at 450 Japanese RA patients found associations between HLA-DR2 and ILD [41]. RA-ILD patients also appear to share gene variants linked to familial pulmonary brosis such as mutations in the TERT, RTEL1, PARN, and SFTPC genes [42]. Recently, a gain of function variant in the promoter region of a gene encoding mucin 5b (MUC5B), the strongest genetic risk factor for IPF, was found in up to 32% of patients with RA-ILD [43].

Pathogenesis

The pathogenesis of RA in general and in RA-ILD more speci cally remains unknown. Genetics play a role, but they do not explain all of the risks. Environmental exposures resulting in a dysregulated immune response are also involved. A variety of environmental exposures are associated with RA-ILD risk, with tobacco smoking the strongest known. Smoking leads to increased citrullinated protein levels in the lung, as measured by bronchoalveolar lavage (BAL) fuid [44]. Because the RA-related anti-citrullinated protein antibodies target citrullinated proteins, it has been hypothesized that smoking-associated increases in citrullinated proteins may contribute to the pathogenesis of RA [45, 46], but further study is still needed. In general, it is well established that RA-related antibodies (e.g. RF and anti-citrullinated protein antibodies) can be elevated in the blood several years prior to the onset of joint infammation [47]. This nding has led to the currently accepted model of RA development, in which the initial immune dysregulation starts at a mucosal site. Many of the data supporting this hypothesis suggest the lung as one of the main mucosal trigger sites [45, 46]. For example, studies have demonstrated RA-related antibody generation in the lung, using sputum or BAL, in individuals before and after RA onset in the joints [48, 49]. It is less clear what mechanisms lead to the development of RA-ILD, with more research clearly needed.

It has been speculated that RA-ILD may be the coincidental development of IPF in an individual with underlying RA. While there are some shared similarities between IPF and RA-ILD including shared clinical and genetic risk factors mentioned above, the prevalence of clinically signi cant RA-ILD within RA patients is markedly higher than the prevalence of IPF in the general population (5% vs. 0.2%), suggesting pathogenic features speci c to RA are likely involved in the development of RA-ILD.

Clinical Features and Diagnosis

There are no established diagnostic criteria for RA-ILD (Box 13.1). Similar to other progressive ILDs, the onset of RA-ILD is often insidious and is characterized by progressive breathlessness with or without a dry cough. Unlike those with isolated ILD, patients with ILD and an infammatory arthritis can have musculoskeletal limitations to physical activity leading to delayed recognition of respiratory symptoms. RA-ILD is more common in men [9, 10, 12, 18, 32] and the average age of ILD onset is in the seventh decade of life [50, 51] irrespective of the year of RA onset [52]. The duration of RA prior to the onset of ILD depends on the age of onset of RA (with earlier onset RA having a longer duration prior to ILD) and the subtype of ILD (with UIP having a shorter duration) [52]. The majority of patients with RA-ILD will present with RA rst, though ILD can precede RA in up to 17% of cases [51, 52] and be diagnosed before or within a year of the diagnosis of RA in up to a third of cases [51–53]. In diagnosing RA-ILD, clinicians should consider the possibility of drug-induced lung disease and infection as a cause or contributor.

Box 13.1 Proposed Diagnostic Criteria for RA-ILD

Patient must have all of the following major criteria and at least one minor criterion:

Main Criteria

\1.\ RA by 1987 ACR [54] and/or 2010 ACR/EULAR criteria [55]

\2.\ Radiographic evidence of ILD

\3.\ Exclusion of other known causes of ILD

Minor Criteria

\1.\ Respiratory-related symptoms (i.e. rest or exercise-­ induced breathlessness and/or cough)

\2.\ Physiologic evidence of restriction (i.e. FVC and/or TLC <80% predicted)

\3.\ Gas-exchange abnormalities (exercise-induced desaturation of ≤89% and/or DLCO <80% predicted)

High-resolution computed tomography (HRCT) is the cornerstone of diagnosis and shows patterns similar to those in idiopathic interstitial pneumonias (IIPs). The most common pattern of RA-ILD is one of the usual interstitial pneumonia (UIP, Fig. 13.1), characterized by peripheral and basilar-predominate reticulations and honeycombing/traction bronchiectasis and a paucity or absence of other features such as ground glass opacities and nodules [56, 57]. The second most common pattern is nonspeci c interstitial pneumonia (NSIP, Fig. 13.2), characterized by ground glass opacities and a paucity or absence of honeycombing. Organizing

Fig. 13.1  High-resolution computed tomography (HRCT) ndings in rheumatoid arthritis-associated usual interstitial pneumonia (RA-UIP)

Fig. 13.2  High-resolution computed tomography (HRCT) ndings in rheumatoid arthritis-associated nonspeci c interstitial pneumonia (RA-NSIP)

pneumonia is a rare manifestation of RA and can be secondary to medications or infection [58].

Pulmonary function testing (PFT) shows a progressive restrictive ventilatory defect as the disease progresses, de ned as reduced lung volumes (total lung capacity, FVC, and residual volume). Due to loss of surface area for gas exchange, the DLCO decreases and the reduced compliance of the lung leads to increased elastic recoil manifested by an increase in the forced expiratory volume in 1 s (FEV1)/FVC ratio. Gas-exchange abnormalities are present and correlate with the degree of parenchymal involvement. The correlations between the degree of interstitial involvement and PFTs can be affected by concomitant conditions such as emphysema (leading to a higher FVC and lower DLCO for any degree of interstitial involvement) and pulmonary hypertension (leading to a lower DLCO for any degree of interstitial involvement). Oxygen desaturation initially presents during activity and progresses to be present at rest and during sleep as the disease progresses.

BAL is not necessary for the diagnosis in the majority of cases. Patients with RA-ILD have an increase in the total cell numbers [59] with elevations in neutrophils, lymphocytes, and eosinophils [9, 26, 60]. BAL is also abnormal in subclinical ILD [61] and can help distinguish these patients from those with normal physiology and chest imaging [62]. BAL is particularly useful in excluding infection.

Surgical lung biopsy can help establish the diagnosis of RA-ILD and prognosticate. While histopathologic patterns may have prognostic signi cance [51, 56], chest imaging can often predict the underlying pathologic pattern, with an HRCT pattern of UIP highly speci c for a pathologic pattern of UIP, thus precluding the need for a surgical lung biopsy [51, 63]. Pathologic ndings vary, with UIP the most common pattern identi ed. It is characterized by temporal heterogeneity (i.e. areas of normal lung interspersed with areas containing broblast foci adjacent to areas of establishedbrosis) subpleural accentuation and varying presence of microscopic honeycombing (Fig. 13.3) [51]. In comparison to the UIP pattern seen in IPF, UIP in RA has more lymphoid aggregates and germinal centers [64] and fewer broblast foci [65, 66]. NSIP has either patchy or, more commonly diffuse, brosis and infammation that is temporally uniform (Fig. 13.4) [51].

Treatments

In spite of the prevalence and morbidity of RA-ILD, only one phase 2 randomized treatment trial has been completed (PMID: 36075242) and, there are no approved therapies speci c for RA-ILD. There are limited reports of treatment with methotrexate [67], azathioprine [68], cyclosporine [69], mycophenolate mofetil [70], and TNF-α inhibitors [71, 72]. Mycophenolate mofetil has been evaluated in a retrospective cohort analysis of 125 patients with CTD-ILD and, in the 18

Fig. 13.3  Histopathology of rheumatoid arthritis-associated usual interstitial pneumonia (RA-UIP)

Fig. 13.4  Histopathology of rheumatoid arthritis-associated nonspeci c interstitial pneumonia (RA-NSIP)

patients with RA-ILD, there was acceptable safety and tolerability and a trend toward an improvement in FVC [73]. In a recent open-label registry study of 63 patients with RA-ILD and a mean follow-up of 9.4 months, two-thirds of patients on abatacept, an antagonist of T-lymphocyte co-stimulation, had stable lung function and dyspnea as measured by the Modi ed Medical Research Council Dyspnea Scale [74].

Anti-CD20 therapy with rituximab has biological plausibility given the nding of CD20-positive B-cell hyperplasia in RA-ILD [75]. There are a number of case series showing rituximab stabilizing patients with RA-ILD when used as primary or salvage therapy [76–80]. A retrospective evaluation of 10 years of experience at a single center where rituximab was given for joint disease, 52% of patients with ILD were stable and 16% improved [78]. A prospective open-­label study of rituximab for salvage therapy in progressive RA-ILD enrolled ten patients and found stabilization in six of seven patients who completed the study, three of whom had UIP [79].

Response to therapy appears to be dictated by the underlying pattern of ILD, with NSIP having a higher likelihood of response to immunosuppression compared to UIP [50, 51]. In

treating patients with RA-ILD, it is important to remember that adequate control of joint disease does not correlate with adequate control of lung disease and vice versa. Any treatment plan should include input from the patient’s rheumatologist.

Given the radiographic and pathologic similarities between RA-ILD and IPF, there is interest in evaluating the ef cacy of the antibrotics approved for use in IPF, nintedanib, and pirfenidone. RA-ILD was evaluated as a subgroup in the recently released INBUILD trial [81], nding that nintedanib resulted in a signi cantly slower rate of lung function decline compared to placebo for patients with brosing ILD, particularly in those with a UIP pattern of brosis. The only prospective randomized placebo-controlled treatment trial in RA-ILD is the TRAIL1 trial looking at pirfenidone in patients with RA and brotic ILD (PMID: 36075242). Though the primary endpoint wasn’t met, pirfenidone slowed the progression of ILD as measure by decline in FVC in a magnitude similar to that seen in other trials of anti brotics. In addition, the addition of pirfenidone was found to be safe and well-tolerated in the setting of RA-speci c treatments.

In addition to pharmacologic therapy, RA-ILD patients should be vaccinated against common respiratory infections according to published guidelines [82]. Physical therapy should be considered for patients with deconditioning or limitations to physical activity from lung or joint disease. Clinicians should ensure normoxia at rest, with activity, and with sleep through formal oxygen titration and nocturnal oximetry. Given the prevalence of heart disease in subjects with RA [83], providers should have a low threshold to look for obstructing coronary disease in patients whose new or progressive dyspnea cannot be explained by their lung disease alone.Attention to bone health is crucial for any patient on corticosteroids. Smokers are overrepresented in the RA-ILD population and tobacco cessation for active smokers should be regularly stressed. Finally, RA-ILD has an impact on health-related quality of life [84] and attention should be paid to the patients mental health in addition to their physical health.

Lung transplant should be considered in all eligible patients with progressive lung disease. Outcomes in RA-ILD patients are similar to those with IPF and transplant results in signi cant improvement in quality of life [85].

Prognosis

ILD is a leading cause of death in patients with RA and worsens outcome [4, 14, 18]. The presence of ILD in RA increases mortality by twoto tenfold [53] and progression of disease is common—50% of patients with early asymptomatic RA-ILD and 60% of patients with established RA and a UIP pattern of brosis have radiographic progression over 1.5 years [86]. Large population-based medical database studies have found a median survival of 6.6–7.8 years and a 35–40% 5-year mortality for all-comers with RA-ILD (compared to 18% 5-year mortality in non-ILD RA) [14, 53].

Prediction of disease behavior over time in individual patients with RA-ILD is challenging and hindered by a limited understanding of the natural history of disease and no validated surrogate markers of progression [87]. It has long been known that HRCT and histopathologic pattern can predict outcome in ILDs. In RA-ILD, a UIP pattern on HRCT or histopathology carries a poor prognosis [50, 51, 80, 88–95], though radiologist-determined honeycombing (i.e. the distinction between “de nite” and “probable” UIP based on the current international guidelines for the diagnosis and management of IPF [96]) does not seem to infuence outcome [88]. Older studies found that outcomes in patients with RA and a UIP pattern of brosis were similar to that of IPF [50, 89–91] with median survival of 3.2–5.5 years [50, 89, 91] and 5-year survival rates of 36.6% [89] though more recent studies have found median survivals of 7.9–10.4 years [80, 88, 94]. A radiographic pattern of NSIP carries a more favorable diagnosis with 3- to 5-year survival rates ranging from 70% to 100% [51, 88, 89]. Semi-quantitative and quantitative HRCT assessments have been looked at in RA-ILD and both a radiologist-determined brosis score [97] and a CALIPER score vessel-related structure threshold of 4.4% [98] independently predict outcome. A risk prediction tool developed in IPF (the GAP (gender, age, physiology) model) has been shown to predict mortality in RA-ILD [99].

In addition to the chest imaging or histologic pattern of ILD, older age [16, 91, 92, 100], male gender [50], lower DLCO [50, 86, 92], and the presence of brosis on pathology [91] also predict mortality. Physiology at baseline and changes in physiology over time have been shown to be strong predictors of outcome [92], with a 10% decline in FVC at any point in follow-up more than doubling a patient’s mortality [94].

Acute exacerbations (AE, de ned as subacute worsening over 30 days without a precipitant) have an impact on progression and are a signi cant contributor to death in patients with RA-ILD [89]. Of all the CTDs, RA-ILD has the highest reported AE rate, estimated at 3–11% yearly and similar to that seen in the IIPs [101–106]. Exacerbations do not correlate with RA disease activity; quiescent extra-thoracic features do not preclude anAE in patients with RA-ILD. Whether from an AE or infection, patients with RA-ILD that are hospitalized for an acute respiratory worsening have a poor prognosis [106, 107], with one study reporting a median survival of 3.5 years after discharge [108].

Rheumatoid Arthritis-Associated Airways

Disease

Airways disease in RA (RA-AD) can take on many forms including small airway disease (i.e. asthma, chronic obstructive pulmonary disease (COPD), constrictive bronchiolitis,

and follicular bronchiolitis) and medium/large airway disease (bronchiectasis and arthritis of the cricoarytenoid joint). Similar to RA-ILD, the incidence of RA-AD depends on the population studied, the de nition of airways disease, and the prevalence of smoking in the population studied. A signi - cant number of patients with RA-AD will have physiologic or radiographic abnormalities without accompanying symptoms.

Epidemiology

The overall prevalence of RA-AD ranges from 15% to 44% [13, 109–112] in all-comers and 14–30% in non-smokers [109, 111]. The lifetime risk of developing airways disease (de ned as a reduced FEV1/FVC ratio on spirometry and a physician-diagnosed airway or parenchymal lung disease) is 9.6% in all-comers with RA in comparison to 6.2% in the general population [113]. In a study of 100 consecutive RA patients who had normal CXRs, screening PFTs found 32% of patients had airfow obstruction as measured by decreased FEV1/FVC and/or forced expiratory fow 25–75 [111], a prevalence higher than in matched controls. Radiographic abnormalities suggestive of RA-AD are more common, with one study of early RA nding HRCT evidence of air trapping in 69% of patients [114]. A recent population-based study from the UK also found that airways disease, including COPD and asthma, is more prevalent in the years preceding the diagnosis of RA compared to controls (7% vs. 4% for COPD and 7% vs. 14% for asthma) [115], suggesting a potential link with RA pathogenesis. In this study, COPD preceded the diagnosis of RA by a median of 4.5 years and asthma preceded RA by a median of 12.5 years.

Risk Factors

Smoking is a well-documented risk factor for RA-AD with a hazard ratio for symptomatic AD of 4.38 in RA ever-smokers (95% CI 2.14–8.99) [113]. Other identi ed risk factors for AD in this population include advanced age [116], male gender [113], high titer RF and anti-citrullinated protein antibodies [116], more severe RA [113, 116], and a longer duration of RA [112, 116, 117].

Clinical Features, Diagnosis, and Outcome

RA-airways disease can be diagnosed with pulmonary physiology (i.e. reduced FEV1/FVC ratio [airfow obstruction] and elevated residual volume [gas trapping]) and HRCT (infammation/structural changes in large and small airways