The Lung in a Cohort of Rheumatoid Arthritis Patients

An Overview of Different Types of Involvement and Treatment

Ana C. Duarte; Joanna C. Porter; Maria J. Leandro


Rheumatology. 2019;58(11):2031-2038. 

In This Article


A total of 1129 patients with RA were identified. Of these, 87 (7.7%) had documented lung involvement on HRCT. Of patients with radiological lung involvement, 65 (74.7%) were female, 74 (85.1%) Caucasian, with mean ± S.D. age at lung disease diagnosis of 63.2 ± 12.4 years. At last follow-up, patients' mean age was 68.3 ± 12 years and median RA duration was 14 (interquartile range 8–29) years. Current and previous smoking habits were documented in 10 (11.5%; mean 21 ± 7.6 pack years) and 38 (43.7%; mean 18.7 ± 5.2 pack years) patients, respectively. ACPA were positive in 72/82 (87.8%) patients and RF in 74/84 (88.1%). Four patients (4.7%) had neither ACPA nor RF. Of patients tested for ANA, 23/62 (37.1%) were positive, with four patients having positive anti-Ro, two anti-Scl70 and two anti-PL12 antibodies. Patients with anti-PL12 antibodies had overlap seropositive-erosive RA with antisynthetase syndrome, but patients with anti-Scl70 did not have any clinical manifestations of SSc despite one of them having nailfold capillaroscopy with early scleroderma pattern. Secondary Sjögren's syndrome was documented in six (6.9%) patients, cutaneous RN in five (5.7%) and cutaneous vasculitis in one (1.1%). Erosive disease was reported in 47 (54%) patients.

Most of the patients presented initially with articular symptoms, with lung disease diagnosed on HRCT after a median of 9 (interquartile range 1–21.5) years. However, 2/87 (2.3%) patients had lung disease as a prior manifestation, one of them presenting with panbronchiolitis 2 years before and the other with non-specific interstitial pneumonia (NSIP) 1 year before developing joint symptoms.

Regarding types of lung involvement on HRCT, ILD was the most common (45 patients; 51.7%). Among these, NSIP and UIP were documented in 18 (40%) patients each and organizing pneumonia in 4 (8.9%) patients. Five patients (11.1%) had lung fibrosis that was unclassifiable on HRCT. None of the patients underwent diagnostic lung biopsy, including those with unclassifiable disease.

As far as airways disease is concerned, isolated bronchiectasis was reported in 31 (35.6%) patients, with 14 (45.2%) of them being current/previous smokers. Seventeen patients with ILD also had traction bronchiectasis. Panbronchiolitis and obliterative bronchiolitis were documented in two (2.3%) patients each. Another patient, a non-smoker, with positive anti-Ro antibody, was described as having small airways disease secondary to RA, but without any specific pattern.

Lung RN occurred in four (4.6%) patients, with none of them having cutaneous nodules. One of these patients had a concomitant pleural effusion. Pleural effusion was reported in three (3.4%) patients, with two of them needing recurrent thoracentesis. One patient underwent pleural decortication.

Distribution of age, gender, disease duration, smoking habits, RF and ACPA between different types of lung involvement are described in Table 1.

A subsequent analysis compared the distribution of some specific variables among patients with ILD and isolated bronchiectasis. The results are shown in Table 2.

At the time of lung disease diagnosis, 26 (29.9%) patients were taking MTX, 18 RTX (20.7%), 12 (13.8%) anti-TNF agents, 6 (6.9%) sulfasalazine and 1 (1.1%) intramuscular gold. However, only two case of MTX-induced pneumonitis were reported. In these patients, symptoms developed within the first year of MTX treatment and there was clinical improvement after drug withdrawal (one patient needed concomitant i.v. pulses of methylprednisolone). At last follow-up appointment, 22 (25.3%) patients were on treatment with MTX. There were no other cases of presumably DMARDs-induced lung disease.

RTX was used in 26 (57.8%) RA-ILD patients with a mean number of 4 cycles (range 1–12). Two of these patients were also receiving MMF and one AZA for ILD. Of these 26 patients, 17 had subsequent PFTs and/or HRCT at different times during follow-up, as shown in Table 3.

Seven patients stopped RTX, six of them due to lack of response from the arthritis and one patient with advanced metastatic prostate cancer stopped RTX as the arthritis was well controlled on low-dose oral prednisolone. There were no cases of severe hypogammaglobulinemia.

During lung disease follow-up (6.7 ± 4.1 years), 22 (25.3%) patients were admitted to hospital with respiratory infection, with the number of admissions ranging from 1 to 5 per patient (total of 18 admissions). Most (63.6%) of these patients had bronchiectasis (10 isolated and 4 in association with ILD). In 13 patients with positive sputum culture results, Pseudomonas aeurginosa was the most common agent (38.4%), followed by Haemophilus influenza (15.4%). Two patients had Candida albicans identified, and one Mycobacterium avium. There were no reported cases of Pneumocystis jiroveci. In the whole cohort, 13 (14.9%) patients were taking prophylactic antibiotics, mainly azithromycin (n = 7; 53.8%).

In the group of isolated bronchiectasis, five patients treated with RTX developed hypogammaglobulinemia, with two of them having RTX stopped due to recurrent respiratory infections. One of these patients was taking prophylactic cotrimoxazole.

Overall, in the 87 patients there were 18 (20.7%) deaths. Of these, 7 (38.9%) were related to ILD and occurred 5.3 ± 3.5 years after ILD diagnosis. Of the patients who died from ILD, four had UIP and three NSIP. UIP was associated with PFT decline when compared with other ILD patterns [odds ratio (OR) = 9.0, 95% CI 1.39–58.44; P = 0.021]. Two other deaths (11.1%) were due to infection, but neither was considered directly attributable to immunosuppression.

Three patients died due to cancer progression (metastatic prostatic cancer, breast cancer and melanoma) and one due to cardiac failure with restrictive pericardial effusion; the other five patients died of unknown causes.

GAP index was calculated at the time of ILD diagnosis and 31/36 (86.1%) patients were GAP stage I, 4/36 (11.1%) GAP stage II and 1/36 (2.8%) GAP stage III. Of the seven patients that died due to ILD, four were GAP stage I and three were GAP stage II. Of the three patients that were GAP stage II, two died 2 years after lung disease diagnosis and one 3 years after. All the GAP stage I patients died between 6 and 10 years after lung disease diagnosis. The patient that was GAP stage III at diagnosis is still alive but was diagnosed with RA-ILD less than a year ago.