Including changes in dyspnea after inpatient rehabilitation improves prediction models of exacerbations in COPD

Reducing the probability of future exacerbations is one of the main goals of pulmonary rehabilitation (PR) in COPD. Recent studies identified predictors of future exacerbations. However, PR might alter both predictors and number of exacerbations.

This secondary analysis examined which predictors assessed at both the beginning and the end of PR predict the risk of moderate (i.e. use of cortisone and/or antibiotics) and severe (hospitalization) exacerbations in the year after PR.

A total of n = 383 COPD patients (34.7% female, mean age = 57.8 years (SD = 7.1), mean FEV1%pred = 51.0 (SD = 14.9)) who attended a 3-week inpatient PR were included. Number of moderate and severe exacerbations were assessed one year after PR (T2) via questionnaires. Potential predictors were assessed at the beginning (T0) and the end (T1) of PR. Negative binomial regression models were used.

The mean numbers of severe (M_s)/moderate (M_m) exacerbations in the year after PR (M_s,t2 = 0.19; M_{m, t2} = 1.07) was reduced compared to the numbers of exacerbations in the year before PR (M_s,t1 = 0.50, p < 0.001; M_m,t1 = 1.21, p = 0.051). Previous exacerbations, retirement, change in dyspnea (for severe exacerbations) and dyspnea at T1 (for moderate exacerbations) were identified as significant predictors.

PR might alter associations between predictors and future exacerbations. Dyspnea at the end of PR or change in dyspnea are better predictors than dyspnea at the beginning of PR.