A Predictive Model Based on Blood Indicators for Admission to the ICU with AECOPD

AECOPD is the most common cause of death among infectious diseases in developing countries, and also an important cause of mortality and morbidity in developed countries. In recent years, related scoring systems such as the mMRC score and CAT questionnaire have been widely used to assess the severity of AECOPD. However, they both have some shortcomings in predicting the admission of AECOPD patients to the ICU. This study aimed to develop a new prediction model to predict the admission of AECOPD patients to the ICU based on objective blood indicators.

This was a retrospective study. Enrolled patients with AECOPD underwent blood gas analysis as well as biomarker testing for serum inflammatory markers, including white blood cell count (WBC), neutrophils, D-dimer, procalcitonin (PCT), high-sensitivity C-reactive protein (hs-CRP), and erythrocyte sedimentation rate (ESR). General characteristics such as age and gender were also recorded. The main observation was admission to the ICU. Univariate analysis and binary logistic regression analysis were used to explore independent risk factors for admission to the ICU in patients with AECOPD, which could be used as components of a new predictive model. Subject receiver operating characteristic curves (ROC) were used to assess the sensitivity and specificity of the new model, which consisted of all independent risk factors predicting the primary outcome.

Initially, 369 patients with AECOPD were admitted to the general ward, of which 119 were subsequently transferred to the ICU (119/369). PaCO2, WBC, D-dimer, PCT, and hs-CRP were independent risk factors for admission to the ICU in patients with AECOPD. The AUC of the new prediction model (combined model consisting of PaCO2, WBC, D-dimer, PCT, and hs-CRP) was 0.94 (95% CI 0.92 - 0.97). The sensitivity was 80.7% and the specificity was 94.8%.

The model for predicting the admission of AECOPD patients to the ICU based on blood indicators has a high specificity and sensitivity.