Factors associated with prolonged duration of acute exacerbation of COPD (AECOPD): Analysis of ECLIPSE cohort

This abstract will be presented at ERS 2019 in Madrid, Spain.

Leonhard Stadler is presenting this Encore on behalf of all authors with their permissions.

Background: Prolonged duration of AECOPD is assumed to be associated with poorer outcomes. We evaluated factors associated with duration of AECOPD in the ECLIPSE cohort.

Methods: The ECLIPSE observational cohort study included patients with COPD (GOLD grades II – IV). We analysed data from patients who experienced ≥ 1 moderate/severe AECOPD during the 3-year study. Exacerbations were classified as moderate (treated with OCS and/or antibiotics) or severe (requiring hospitalisation). Only the first exacerbation was considered in the analysis. Duration (days) of AECOPD was calculated as the end (resolution) date minus the start date for each exacerbation plus one, with start and end dates obtained directly from the patient/physician report. We used explanatory, logistic regression modelling to identify factors (e.g. AECOPD, patient demographic and risk factors) associated with prolonged exacerbation duration, defined as > 22 days (75th percentile) in this analytical cohort.

Results: Overall, 1554 patients experienced ≥ 1 exacerbations. The first exacerbation event had median duration of 14 days (IQR 8 – 22); 23% of these were prolonged exacerbations. Severity of exacerbation was the only factor statistically significant at the p < 0.05 threshold associated with prolonged duration. Fat free mass index, winter season and severity of index exacerbation were independently associated with increased odds of a prolonged exacerbation events at a threshold of p < 0.1 ([Table 1]).

Table 1 Multivariate analysis (complete case analysis).

		Odds ratio (95% CI)	p-value
* Indicates biological or clinical plausibility for association with prolonged exacerbation, as previously reported in the literature to be predictive of recurrent exacerbations. ** Variables with p < 0.1 in univariate or multivariate analysis. Variables are included based on stepwise selection of variables with clinical/biological plausibility selected by clinician, plus additional variables which had p < 0.1 in univariate analysis.
Fat free mass index (kg/m²)* (N = 1530)		0.932 (0.858, 1.012)	0.0929
Severity of index exacerbation* (N = 1554)	Severe vs. moderate	1.917 (1.269, 2.896)	0.0020
Season of index exacerbation** (N = 1521)	Autumn vs. winter	0.680 (0.449, 1.029)	0.0677
	Spring vs. winter	0.578 (0.368, 0.906)	0.0169
	Summer vs. winter	0.638 (0.382, 1.067)	0.0867

Conclusions: Prolonged duration occurs in nearly one quarter of exacerbations. Severity of exacerbation likely influences the outcome rather than explain it. However, prolonged duration may be influenced by season and fat free mass index. These factors should be taken into consideration and suggest that closer post-discharge monitoring of patients who experience severe AECOPD may be warranted.

Funding: GSK (study ID 210008)