Tracking dyspnea up to supplemental oxygen prescription among patients with pulmonary fibrosis

Amy L. Olson, Bridget Graney, Susan Baird, Tara Churney, Kaitlin Fier, Marjorie Korn, Mark McCormick, David Sprunger, Thomas Vierzba, Frederick S. Wamboldt, Jeffrey J. Swigris

Dyspnea is the hallmark symptom of pulmonary fibrosis. Supplemental oxygen (O2) is prescribed to many patients with pulmonary fibrosis in hopes of alleviating dyspnea and improving physical functioning. We used response data from the University of California San Diego Shortness of Breath Questionnaire (UCSD) which was administered monthly in the context of a longitudinal, observational study to plot a rich trajectory for dyspnea over time in patients with pulmonary fibrosis. We used other data from that study to identify clinical predictors of being prescribed O2 and to provide additional information for how UCSD scores could be used for clinical purposes. We used linear mixed-effects models and multivariate Cox proportional hazards to model change in dyspnea scores over time and to identify significant predictors of time-to-O2-prescription among a pool of clinically-meaningful candidate variables. In the longitudinal study, all decisions, including whether or not to prescribe O2, were made by subjects' treating physicians, not members of the research team. One-hundred ninety-four subjects with pulmonary fibrosis completed more than one UCSD or were prescribed O2 at some point during the follow-up period (N = 43). Twenty-eight of the 43 had analyzable, longitudinal data and contribute data to the longitudinal UCSD analyses. All 43 were included in the time-to-O2-prescription analyses. Subjects prescribed O2 had more severe dyspnea at enrollment (38.4 ± 19.6 vs. 22.6 ± 18.7, p < 0.0001) and a steeper increase in UCSD scores over time (slope = 1.18 ± 0.53 vs. 0.24 ± 0.09 points per month, p = 0.02) than subjects not prescribed O2. Controlling for baseline UCSD score and FVC%, subjects with a clinical summary diagnosis of idiopathic pulmonary fibrosis (IPF) were far more likely to be prescribed O2 than subjects with other forms of pulmonary fibrosis (hazard ratio = 4.85, (2.19, 10.74), p < 0.0001). Baseline dyspnea and rise in dyspnea over time predict timing of O2 prescription. Accounting for disease severity, patients with IPF are more likely than patients with other forms of pulmonary fibrosis to be prescribed O2. UCSD scores provide clinically useful information; frequent administration could yield timely data on changes in disease status in patients with pulmonary fibrosis. The longitudinal study is registered on ClinicalTrials.gov ( NCT01961362 ). Registered October 9, 2013.