Frailty and in-hospital mortality in older patients with acute exacerbation of COPD: A real-world prospective cohort study

Abstract

Background: Few evidence exists for the effect of frailty on the patients admitted with an acute exacerbation of chronic obstructive pulmonary disease (AECOPD).

Objective: We explored the link between frailty and in-hospital death in AECOPD, and whether laboratory indicators mediate this association.

Methods: This was a real-world prospective cohort study including older patients with AECOPD, consisting of two cohorts: a training set (n = 1356) and a validation set (n = 478). The independent prognostic factors, including frail status, were determined by multivariate logistic regression analysis. The relationship between frailty and in-hospital mortality was estimated by multivariable Cox regression. A nomogram was developed to provide clinicians with a quantitative tool to predict the risk of in-hospital death. Mediation analyses for frailty and in-hospital death were conducted.

Results: The training set included 1356 patients (aged 86.7 ± 6.6 years), and 25.0 % of them were frail. A nomogram model was created, including ten independent variables: age, sex, frailty, COPD grades, severity of exacerbation, mean arterial pressure (MAP), Charlson Comorbidity Index (CCI), Interleukin-6 (IL-6), albumin, and troponin T (TPN-T). The area under the receiver operating characteristic curve (ROCs) was 0.862 and 0.845 for the training set and validation set, respectively. Patients with frailty had a higher risk of in-hospital death than those without frailty (HR,1.83, 95%CI: 1.14, 2.94; p = 0.013). Furthermore, CRP and albumin mediated the associations between frailty and in-hospital death.

Conclusions: Frailty may be an adverse prognostic factor for older patients admitted with AECOPD. CRP and albumin may be part of the immunoinflammatory mechanism between frailty and in-hospital death.

Keywords: Acute exacerbation of chronic obstructive Pulmonary disease; Albumin; CRP; Frailty; In-hospital mortality; Mediation analysis.