• Yao-Kuang Wu, Chou-Chin Lan, I-Shiang Tzeng, and Chih-Wei Wu.
• Division of Pulmonary Medicine, Taipei Tzu Chi Hospital, Buddhist Tzu Chi Medical Foundation, New Taipei City, Taiwan. Electronic address: drbfci@yahoo.com.tw.
• J Formos Med Assoc. 2021 Mar 1; 120 (3): 1005-1013.

BackgroundAcute exacerbation of chronic obstructive pulmonary disease (COPD) results in deterioration of lung function and mortality. Previous prediction models have been designed for severe exacerbation of COPD, leading to readmission. However, these models lacked newly established predictors such as the eosinophil count. The present study developed a novel CO PD-re admission (CORE) score.MethodsWe retrospectively reviewed medical records of patients visiting Taipei Tzu Chi Hospital between January 1, 2014, and May 31, 2017. We analyzed all covariates by univariate and then multivariate logistic regressions. Numeric or ordinal variables showing statistical significance were transformed into dichotomous variables by cut-off values determined by the Youden Index. The CORE score was designed to predict one-year readmission rates.ResultsA total of 625 patients were recruited. After analysis, the CORE score included five predictors (eosinophil count, lung function, triple inhaler therapy, previous hospitalization, and neuromuscular disease). We observed a highly linear relationship between the CORE score and COPD readmission (R = 0.981; R 2 = 0.963; P < 0.001). The CORE score had a higher predictive accuracy than that for hospitalization in the previous year (area under the curve = 0.703 vs. 0.619; P < 0.001). Patients with higher CORE scores had a shorter time to first COPD readmission (P < 0.001). Using the zero point as a reference, the hazard ratios for each score from 1 to 4 were 1.209, 2.211, 3.359, and 4.510, respectively.ConclusionThe CORE score includes two novel predictors (eosinophil count and triple inhaler therapy). The model has a high predictive power for one-year COPD readmission.Copyright © 2020 Formosan Medical Association. Published by Elsevier B.V. All rights reserved.

12 keywords...

hide…