Longitudinal Evaluation and Subgroup Evaluation of Cough Severity in Fibrotic Interstitial Lung Disease

To the Editor:

We thank Channick and Swigris (1) for their thoughtful editorial regarding our publication on the epidemiology and significance of cough severity in fibrotic interstitial lung disease (ILD) (2). Channick and Swigris raised several interesting questions regarding longitudinal evaluation of the Cough Severity Visual Analog Scale (VAS) that have prompted us to conduct additional analyses.

We agree with the authors that thresholds of clinical relevance should not be conflated with the minimum important difference (or minimal within-patient change threshold), which in fibrotic ILD we suspect to be below the 30-mm change threshold that has been reported in chronic cough. As suggested, we evaluated the proportion of patients with 5-mm, 10-mm, 20-mm, and 30-mm annualized change in Cough Severity VAS, with these thresholds representing 5%, 10%, 20%, and 30% of the full spectrum of the 100-mm scale. Among 616 patients with idiopathic pulmonary fibrosis (IPF), 41% had a 5-mm increase in Cough Severity VAS, 30% had a 10-mm increase, 19% had a 20-mm increase, and 13% had a 30-mm increase. Similar findings were seen in 1,667 patients with non-IPF fibrotic ILD, with 31% having a 5-mm increase in Cough Severity VAS, 24% having a 10-mm increase, 15% having a 20-mm increase, and 11% having a 30-mm increase. In both IPF and non-IPF fibrotic ILD, patients meeting guideline criteria for disease progression (3) had a concurrent increase in Cough Severity VAS irrespective of the threshold used to define worsening cough (Figure 1). Multivariable analyses adjusted for age, sex, baseline FVC, and Dl_CO percent predicted, ILD-targeted therapy use, and study site showed annualized increases in the Cough Severity VAS at each threshold were generally independently associated with more rapid annualized FVC percent predicted decline in both cohorts, but not Dl_CO (Figure 1). Various thresholds of increase in the Cough Severity VAS were consistently associated with transplant-free survival on multivariable analyses in patients with non-IPF fibrotic ILD but not in those with IPF (Figure 1).

Figure 1.

Proportions of patients with disease progression between those with and without annualized increment in cough severity of different thresholds, and multivariable analyses of associations between annualized increment in cough severity of different thresholds with annualized change in FVC percent predicted, Dl_CO percent predicted, and transplant-free survival in idiopathic pulmonary fibrosis (IPF) and non-IPF fibrotic ILD cohorts. Between-group comparisons of patients with and without annualized increment in cough severity of different thresholds were performed using the Fisher’s exact test. Multivariable analyses for annualized lung function percent-predicted changes and transplant-free survival were adjusted for age, sex, FVC% predicted, Dl_CO% predicted, ILD-targeted therapy use, and study site. CI = confidence interval; ILD = interstitial lung disease; VAS = visual analog scale.

Potential treatment effects of supplemental oxygen on cough in patients with fibrotic ILD have been reported (4, 5). In our IPF cohort, annualized changes in cough severity were comparable between those receiving and not receiving ambulatory and/or continuous oxygen therapy (oxygen users: 2.5 mm; 95% confidence interval [CI], 1.7–3.3 mm; non–oxygen users: 2.2 mm; 95% CI, 1.5–3.0 mm; P = 0.66). A greater increase in cough severity was observed in patients with non-IPF fibrotic ILD who received either ambulatory and/or continuous oxygen therapy than in those who did not receive either (oxygen users: 2.6 mm; 95% CI, 1.9–3.3 mm; non–oxygen users: 0.6 mm; 95% CI, 0.2–1.0 mm; P < 0.001), which may be related to underlying lung disease severity and progression (i.e., confounding by indication). Prospective clinical trials are needed to better understand the effects of supplemental oxygen on cough in patients with fibrotic ILD. Annualized changes in cough severity were also comparable between patients with and without suspected pulmonary hypertension (defined as right ventricular systolic pressure ⩾ 35 mm Hg [6]) in both 224 patients with IPF (suspected pulmonary hypertension: 1.9 mm; 95% CI, 0.6–3.2 mm; non–suspected PH: 2.6 mm; 95% CI, 1.5–3.8 mm; P = 0.41) and 793 with non-IPF fibrotic ILD (suspected pulmonary hypertension: 0.9 mm; 95% CI, 0.3–1.6 mm; non–suspected pulmonary hypertension: 0.8 mm; 95% CI, 0.3–1.4 mm; P = 0.86).

These additional analyses provide further insights into cough in patients with fibrotic ILD; however, more work is clearly needed to better appreciate the pathophysiology, epidemiology, measurement, clinical impact, and management of cough in fibrotic ILD. We again thank Channick and Swigris for their additional suggestions that help move us another step toward this understanding.