Rheumatoid arthritis and interstitial lung disease

A dangerous liaison This editorial refers to ‘Interstitial lung disease has a poor prognosis in rheumatoid arthritis: results from an inception cohort’, by Gouri Koduri et al., doi:10.1093/rheumatology/keq035, on page 1483. Although RA is defined as a symmetrical inflammatory arthropathy, a wide variety of extra-articular manifestations are seen in patients with RA. The pleuropulmonary manifestations of RA encompass airway diseases (e.g. bronchiolitis obliterans), pleural diseases (e.g. pleuritis with or without effusion) and parenchymal pulmonary diseases [1, 2]. Interstitial lung disease (ILD) is associated with RA (RA-ILD), but the prevalence, incidence and natural history are not well defined. The prevalence of ILD in RA depends on the population selected for study (mild vs severe disease) and the methods applied for its detection (clinical, physiological, radiological or pathological). The paper by Koduri et al. [3] in this issue of Rheumatology sheds light on the incidence and the natural history of RA-ILD. They report the results from an RA inception cohort, the Early Rheumatoid Arthritis Study (ERAS) [3]. Using a prospective, observational, longitudinal study design, extra-articular features and comorbidity were recorded annually in a well-established inception cohort of RA with 20-year follow-up. Patients with clinically significant pulmonary conditions were further investigated as part of standard clinical practice, which included plain radiology, pulmonary function tests and high-resolution CT (HRCT) scan if RA-ILD was suspected; however, none of the patients had lung biopsy. Fifty-two of 1460 (3%) patients with RA developed RA-ILD: half of them either at baseline or within 3 years of onset, indicating that RA-ILD is an early feature of RA. The incidence of RA-ILD was 4/1000 per year and increased with advancing age and disease severity or activity (as measured by the HAQ). Importantly, 39 of the patients with RA-ILD died (mainly due to respiratory insufficiency, which was the primary cause of death). Median survival following the diagnosis of RA-ILD was only 3 years. This cohort study clearly shows that RA-ILD is an important feature of RA and has a poor prognosis, which is an important message that is clinically relevant. The strengths of this study are the prospective cohort design, large number of consecutive patients with recent-onset RA and long-term follow-up. Previously, only one research group had studied the prevalence of RA-ILD in patients with recent-onset RA [4]. Using several sensitive techniques including HRCT of the chest, Gabbay et al. [4] found abnormalities consistent with ILD in 21/36 (58%) selected RA patients with joint disease of <2 years duration. However, only in a minority of these patients (5/21 patients) were the detected abnormalities categorized as clinically significant. Importantly, the classic chest X-ray is insensitive and imprecise for the detection of the presence of lung involvement in RA [5]. In contrast, HRCT of the chest is definitely a more sensitive technique to detect ILD than chest X-rays [4, 6] (ILD abnormalities were present in 33% of HRCT scans vs 6% of chest X-rays in patients with recent-onset RA [4]). The ERAS study has limitations that should also be taken into account. The physicians did not screen for lung disease; moreover, they did not do a systematic chest X-ray, HRCT or pulmonary function tests. These investigations were only done in patients with clear symptoms or signs of pulmonary disease (i.e. case finding). However, in patients with RA, exertional dyspnoea may be masked by the immobility of the arthritic patient. The case-finding approach, instead of the systematic screening approach in the ERAS study, has two crucial consequences. First, the true prevalence (at baseline) and incidence of RA-ILD is certainly underestimated, because complaints typically occur relatively late in the course of ILD and chest X-ray is insensitive for early detection of RA-ILD. The application of HRCT of the chest and pulmonary function testing [including measurement of total lung capacity (TLC), vital capacity (VC) and carbon monoxide diffusing capacity (DLCO)] would detect more patients with RA-ILD. Secondly, since only RA patients with (severe) symptomatic ILD—relatively late in the course of the disease—were diagnosed with RA-ILD, the prognosis appears to be poor with a median survival of only 3 years. However, in contrast to the start of the ERAS study (in 1986), the diagnosis of RA-ILD should nowadays be made in an earlier phase, thanks to the wide availability of HRCT and the increased awareness of RA-ILD. Even without any beneficial effect of new or existing therapies on the underlying RA-ILD, the prognosis of an RA patient with an ILD detected early in 2010 will be better than the prognosis of an RA-ILD patient diagnosed late in the course of the disease in the 1980s because of lead time bias. New RA inception cohorts are thus needed to update the prevalence, incidence and prognosis of ILD in patients with RA, using HRCT and pulmonary function tests as sensitive screening tools. The diagnostic and therapeutic management of RA patients with ILD is challenging for many reasons. First, the drug-induced lung diseases (e.g. MTX-induced ILD) and respiratory infections (e.g. opportunistic infections such as tuberculosis or CMV pneumonia in immunosuppressed patients with RA) need to be excluded. Secondly, although some patients have progressive RA-ILD, other patients with RA have more stable ILD implicating the need for regular clinical, radiological and physiological monitoring to discriminate between these different courses. A reduced DLCO of <54% of the predicted value appears to be an independent predictor of progression of RA-ILD [7]. Since smoking is an independent predictor of radiographical or physiological abnormalities suggestive of ILD in RA patients [8], smoking cessation should be offered to all RA patients. Thirdly, the current evidence of the efficacy and safety of specific therapies for RA-ILD is very limited. Therefore, additional observational studies addressing the natural history of RA-ILD and randomized controlled trials studying the risk–benefit ratio of drug treatment of RA-ILD are urgently needed. Disclosure statement: The author has declared no conflicts of interest.