Screening for rheumatoid arthritis-associated interstitial lung disease—a Delphi-based consensus statement

For this Delphi consensus statement, a panel of experts was formed, including seven specialists in rheumatology and clinical immunology, four specialists in respiratory medicine/pneumonology, and one specialist in thoracic radiology. The expert panel was selected based on their experience in the topic and their scientific merit. Objectives of this statement were determined a priori. Therefore, the modified Delphi technique in combination with the nominal group technique was used and consisted of five steps [15, 16]:

The aim of this Delphi consensus statement was to develop a stepwise screening algorithm for RA-ILD and to raise awareness among physicians regularly taking care of RA patients. Given the high mortality of RA-ILD and the availability of new therapeutic options such as nintedanib for the progressive pulmonary fibrosis phenotype of RA-ILD, early diagnosis of ILD is essential, and can be achieved by screening [12]. However, the diagnostic methods and the timing of screening are still subject to debate. Monitoring of clinical signs (e.g., by auscultation) of possible lung involvement of RA at the time of diagnosis and during the regular follow-up visits was fully supported by the expert panel, as was the necessity of further investigations in case of suspicious respiratory symptoms (i.e., persistent cough or dyspnea on exertion). Further steps suggested are lung function testing with DLCO testing, body plethysmography, and HR-CT of the lung. DLCO has become an important standard in pulmonary function testing, next to forced vital capacity or total lung capacity [17]. It is a strong and independent predictor of diffusion limitation and often the first sign of lung fibrosis. Particular interest should be given to lung auscultation, since Velcro crackles have been found to be a key finding in early fibrotic lung disease [18, 19]. It was discussed that combined measurements of lung volumes including diffusion capacity and the HR-CT should be used rather than the individual examinations alone. In case of non-corresponding results of lung function testing or HR-CT, a multidisciplinary discussion is mandatory to evaluate the presence of RA-ILD or differential diagnoses. This must be emphasized, since patients with RA may also suffer from coexisting ILD other than RA-ILD, with respective treatment indications. It is important to highlight that a multidisciplinary discussion should at least include specialists in pneumonology, rheumatology, and radiology, each with expertise in ILD [20]. Using HR-CT to detect RA-ILD appears to be the best approach in terms of radiographic imaging. However, RA-ILD patterns vary, and may present as usual interstitial pneumonia (UIP), probable for UIP, non-specific interstitial pneumonia, organizing pneumonia, mixed patterns, unclassifiable patterns, and others [21]. To avoid radiation exposure, the panel gave no general recommendation to screen all newly diagnosed RA patients but rather to focus on a population at higher risk. Therefore, the expert panel recommended evaluating risk factors for RA-ILD in each patient with RA, even in the absence of respiratory symptoms. Well established and known risk factors are current or former smoking (especially more than 25 packyears), age, male sex, high disease activity (e.g., by a Clinical Disease Activity Index [CDAI] score > 10), and a family history of and genetic predisposition for RA-ILD [1, 19, 22‐28]. A particular threshold for age as a risk factor has not yet been approved, but different studies found clinically significant RA-ILD from > 55 years upwards [29‐34].

Nevertheless, it has not yet been resolved how individual risk factors should be graded and translated into a clinically confident situation to further investigate for RA-ILD. Therefore, the panel concluded that the quantity and significance of existing risk factors for initiating continuous RA-ILD screening lies at the discretion of the attending physician on the basis of a case-by-case decision. From the panel’s perspective, there is currently not enough evidence to support any specific scoring or weighting of risk factors. These recommendations will be updated as soon as new and profound evidence is published about the classification of known RA-ILD risk factors.

In this context it must be mentioned that a Spanish recommendation for ILD screening in RA patients was published after the Delphi consensus had been designed [35, 36]. In this Delphi consensus, risk factors were weighted and an arbitrary cut-off of 5 points was chosen for conducting ILD screening in RA patients. However, no confirmation of this score has been published so far, thus making it an expert recommendation based on experience rather than evidence. The important, yet unanswered question is whether or not more risk factors will forecast a higher likelihood of developing RA-ILD.

If a patient qualifies for RA-ILD screening, a baseline pulmonary function test (including DLCO testing) and HR-CT should be performed, as suggested by the panel. For regular follow-up, we suggest looking for clinical sings of RA-ILD and to perform pulmonary function testing including DLCO on an annual basis. Another option is transthoracic ultrasound with evaluation of B lines. This technique has emerged over the past years as an alternative method to evaluate ILD in patients with connective tissue diseases, particularly systemic sclerosis [37, 38], but also RA-ILD [39, 40]. However, transthoracic ultrasound is still rarely applied in the ambulant care of RA patients and might gain increased popularity in the future. The regular implementation of HR-CT was discussed critically in this particular situation, especially in patients without symptoms. Ionizing radiation is a general matter of concern, as it is a recognized cause of cancer [41]; however, there is also currently not enough evidence to suggest that screening using HR-CT could generate a prognostic benefit for the patient. Thus, there is currently no positive benefit–risk profile available to support regular HR-CT in patients with RA and risk factors for RA-ILD but without respiratory symptoms. Therefore, the necessity of regular HR-CT and the interval of diagnostic examinations are at the discretion of the attending physician on the basis of a case-by-case decision. If new evidence is published, this recommendation will be updated.

In recent years, different studies have reported a MUC5B promoter variant rs5705950 as a rare but significant risk factor for developing RA-ILD, even in the absence of respiratory symptoms [30, 33, 42]. Juge et al. also proposed a risk score to evaluate for subclinical RA-ILD, including sex, age at RA onset, RA disease activity using the Disease Activity Score-28 for Rheumatoid Arthritis with erythrocyte sedimentation rate (DAS28-ESR), and the abovementioned MUC5B rs35705950 genetic variant. However, the predictive value of this risk score still needs validation in prospective studies [30]. A future perspective will be the announced European Respiratory Society (ERS)/EULAR connective tissue disease (CTD)-associated ILD management guidelines, which will include screening.

Some limitations and weaknesses of this RA-ILD screening must be mentioned, especially the fact that the current evidence did not allow the expert panel to establish a risk score or a weighting of known risk factors for RA-ILD. Until additional data become available, the task of selecting RA patients based on risk factors for RA-ILD must be left in the hands of the caring physicians, thus making it open to interpretation. However, interdisciplinary awareness for RA-ILD as well as its proper tools for (differential) diagnosis are the major goals of our statement.

This Delphi-based screening strategy was developed for early detection and ongoing surveillance of RA-ILD by a multidisciplinary team of pneumonologists, rheumatologists, and radiologists. The presented algorithm was designed to help carers of patients with RA to identify those at risk for RA-ILD and choose an applicable screening strategy for these patients. Furthermore, disease awareness for RA-ILD among carers and patients is of utmost importance. Since risk factors for RA-ILD are well known but cannot be weighted at the moment, an updated version of this recommendation will be necessary when further evidence on this topic is published.