Clinicians shouldn’t rely on pulmonary function tests (PFTs) alone to screen for interstitial lung disease (ILD). The tests performed poorly in a retrospective study of 212 patients with systemic sclerosis, reinforcing the findings of previous studies.
Any screening algorithm should include high-resolution CT (HRCT), which is good at prognosticating disease, the investigators wrote in Arthritis & Rheumatology. “I think all newly diagnosed systemic sclerosis patients should have a full set of PFTs (spirometry, lung volumes, and diffusion capacity) and an HRCT at baseline to evaluate for ILD,” the study’s lead author, Elana J. Bernstein, MD, said in an interview.
ILD is a leading cause of death in systemic sclerosis (SSc) patients, affecting 40%-60% of those with the disease. HRCT is currently the preferred option for detection of ILD. PFTs are commonly used to screen for ILD but haven’t performed well in previous studies. “Someone can have abnormalities on HRCT that are consistent with ILD but still have PFTs that are in the ‘normal’ range,” explained Dr. Bernstein of Columbia University, New York. One cross-sectional study of 102 SSc patients found that the test’s sensitivity for the detection of ILD on HRCT was just 37.5% when forced vital capacity (FVC) <80% predicted.
Investigators sought to assess performance characteristics of PFTs in patients with early diffuse cutaneous SSc, a cohort at high risk of developing ILD. The study enlisted patients from the Prospective Registry of Early Systemic Sclerosis (PRESS), a multicenter, prospective cohort study of adults with early diffuse cutaneous SSc. Overall, 212 patients at 11 U.S. academic medical centers participated in the study from April 2012 to January 2019.
All patients had spirometry (PFT) and HRCT chest scans. PFTs were conducted per American Thoracic Society/European Respiratory Society guidelines. The investigators calculated test characteristics for single PFT and combinations of PFT parameters for the detection of ILD on HRCT. The HRCTs were ordered at the discretion of treating physicians, and scrutinized for ILD features such as reticular changes, honeycombing, traction bronchiectasis, and ground-glass opacities. The investigators defined the lower limit of normal for FVC, total lung capacity, and diffusion capacity for carbon monoxide (DLCO) as 80% predicted.
Overall, Dr. Bernstein and her colleagues found that PFTs lacked sufficient sensitivity and negative predictive value for the detection of ILD on HRCT in these patients.
An FVC <80% predicted performed at only 63% sensitivity and an false negative rate of 37%. Total lung capacity or DLCO <80% predicted had a sensitivity of 46% and 80%, respectively. The combination of FVC or DLCO <80% predicted raised sensitivity to 85%. However, the addition of total lung capacity to this combination did not improve results.
Dr. Daniel E. Furst
Overall, PFTs had a positive predictive value of 64%-74% and an negative predictive value of 61%-70%. “This means that PFT alone will not accurately predict the presence of ILD in about 35%, and not be correctly negative in about 35%,” observed Daniel E. Furst, MD, professor of medicine (emeritus) at the University of California, Los Angeles, and professor of rheumatology at the University of California, Los Angeles.
While the combination of FVC <80% predicted or DLCO <80% predicted performed better than the other parameters, the sensitivity “is inadequate for an ILD screening test as it results in an false negative rate of 15%, thereby falsely reassuring 15% of patients that they do not have ILD when in fact they do,” the investigators observed.
“This study reinforces the notion that PFTs alone are ineffective screening tools for ILD in the presence of systemic sclerosis, particularly for patients with early systemic sclerosis,” said Elizabeth Volkmann, MD, MS, assistant professor and codirector of the CTD-ILD program in the division of rheumatology at the University of California, Los Angeles.
Dr. Elizabeth Volkmann
The study’s scope was relatively small, yet the results provide further evidence to show that HRCT should be performed in all SSc patients to screen for the presence of ILD, Dr. Volkmann said in an interview.
Other research has demonstrated the value of baseline HRCT as a prognosticator of ILD outcomes. The method provides useful information about the degree of fibrosis and degree of damage in early-stage disease, said Dr. Furst, also an adjunct professor at the University of Washington, Seattle, and a research professor at the University of Florence (Italy). “If there’s honeycombing, that’s a bad prognosis. If it’s ground glass or reticular changes, the prognosis is better.
“Once there’s a lot of damage, it’s much harder to interpret disease with HRCT,” he added.
HRCT and PFT work well together to assess what’s happening in patients, Dr. Furst explained. HRCT provides an idea of anatomic changes, whereas PFT outlines aspects of functional change to diagnose early ILD in early diffuse SSc. The study results should not apply to patients with later disease who have more developed ILD, he noted.
The investigators acknowledged that they weren’t able to categorize and analyze patients according to disease extent because they didn’t quantify the extent of ILD. Another limitation was that the HRCTs and PFTs were ordered at the discretion of individual physicians, which means that not all participants received the tests.
“Although the tests were done in 90% of the population, there is still a probability of a significant selection bias,” Dr. Furst said.
Dr. Bernstein and several other coauthors in the study received grants from the National Institute of Arthritis and Musculoskeletal and Skin Diseases to support their work. Dr. Furst disclosed receiving grant/research support from and/or consulting for AbbVie, Actelion, Amgen, Bristol-Myers Squibb, Corbus, the National Institutes of Health, Novartis, Pfizer, and Roche/Genentech. Dr. Volkmann disclosed consulting for and/or receiving grant support from Boehringer Ingelheim, Corbus, and Forbius.
SOURCE: Bernstein EJ et al. Arthritis Rheumatol. 2020 Jun 25. doi: 10.1002/art.41415.