Lung ultrasound as part of a diagnostic algorithm to detect interstitial lung disease in rheumatoid arthritis patients at risk.

Objectives

Rheumatoid arthritis-associated interstitial lung disease (RA-ILD) affects 5-10% of patients with RA. While lung ultrasound (LUS) shows promise for RA-ILD detection, its role remains inconclusive. We aimed to assess its performance as a screening tool in at-risk RA patients.

Methods

We made a cross-sectional study of RA patients, including i) individuals >50 years old with ≥1 risk factor for ILD, and ii) patients with new-onset respiratory symptoms. All patients underwent clinical examination, pulmonary function testing (PFT), LUS, and chest high-resolution computed tomography (HRCT). LUS was considered abnormal in case of ≥5 total B-lines, and diagnostic accuracy was compared against HRCT, using a recent American Thoracic Society definition.

Results

We included 103 RA patients (73% women; mean age 63.4 years, median disease duration 7 years, 62% current or ever smokers, 68% seropositive for RF and/or ACPA). The majority (67 patients, 64.4%) reported no respiratory symptoms and had normal lung auscultation. According to ATS definition, 23 patients (22.3%) had ILD. Use of LUS for RA-ILD showed a sensitivity of 60.9% (95% CI: 40.8-77.8%), specificity of 63.7% (95% CI: 52.8-73.4%), and a negative predictive value (NPV) of 85% (95% CI: 73.9-91.9%). In regression and random forest analyses, LUS did not outperform DLCO for the prediction of RA-ILD. However, combination of negative LUS examination with normal DLCO essentially excluded RA-ILD (NPV 100%, 95%CI: 82-100%).

Conclusions

Negative LUS in combination with normal DLCO reliably exclude ILD in at-risk RA patients, supporting its role in early RA-ILD detection.