Assisted ventilation in COPD - association between previous hospitalizations and mortality

Abstract

Background: In general, previous studies have shown an association between prior exacerbations and mortality in COPD, but this association has not been demonstrated in the subpopulation of patients in need of assisted ventilation. We examined whether previous hospitalizations were independently associated with mortality among patients with COPD ventilated for the first time.

Patients and methods: In the Danish National Patient Registry, we established a cohort of patients with COPD ventilated for the first time from 2003 to 2011 and previously medicated for obstructive airway diseases. We assessed the number of hospitalizations for COPD in the preceding year, age, sex, comorbidity, mode of ventilation, survival to discharge, and days to death beyond discharge.

Results: The cohort consisted of 6,656 patients of whom 66% had not been hospitalized for COPD in the previous year, 18% once, 8% twice, and 9% thrice or more. In-hospital mortality was 45%, and of the patients alive at discharge, 11% died within a month and 39% within a year. In multivariate models, adjusted for age, sex, mode of ventilation, and comorbidity, odds ratios for in-hospital death were 1.26 (95% confidence interval [CI]: 1.11-1.44), 1.43 (95% CI: 1.19-1.72), and 1.56 (95% CI: 1.30-1.87) with one, two, and three or more hospitalizations, respectively. Hazard ratios for death after discharge from hospital were 1.32 (95% CI: 1.19-1.46), 1.76 (95% CI: 1.52-2.02), and 2.07 (95% CI: 1.80-2.38) with one, two, and three or more hospitalizations, respectively.

Conclusion: Preceding hospitalizations for COPD are associated with in-hospital mortality and after discharge in the subpopulation of patients with COPD with acute exacerbation treated with assisted ventilation for the first time.

Keywords: artificial; chronic obstructive; critical care; hospital mortality; patient readmission; pulmonary disease; respiration.

Figure 1

The inclusion of patients.

Figure 2

Logistic regression model of in-hospital mortality. Note: *Significant at P<0.05. Abbreviations: NIV, noninvasive mechanical ventilation; IMV, invasive mechanical ventilation.

Figure 3

Cox proportional hazards model of mortality after discharge. Note: *indicates significant at P<0.05. Abbreviations: NIV, noninvasive mechanical ventilation; IMV, invasive mechanical ventilation.

Figure 4

Respiratory hospitalizations. Notes: (A) Logistic regression model of in-hospital mortality, (B) Cox proportional hazards model of mortality after discharge. *indicates significant at P<0.05. Abbreviations: NIV, noninvasive mechanical ventilation; IMV, invasive mechanical ventilation.