A DECAF-based model augmented with age, lactate, and procalcitonin for predicting mortality in acute exacerbations of COPD: development and single-center temporal validation

Background This study aimed to develop a DECAF-based risk prediction model incorporating age, lactate, and log-transformed procalcitonin and to assess its performance for predicting 30-day all-cause mortality in patients with acute exacerbations of chronic obstructive pulmonary disease (AECOPD). Methods This single-center retrospective cohort study included consecutive adult patients presenting to the emergency department with AECOPD. Patients were divided into a development cohort (DC; January 2022–June 2023) and a temporal validation cohort from the same center (TVC; July 2023–December 2024). The augmented DECAF model (A-DECAF) was developed using multivariable logistic regression with four prespecified predictors: DECAF score, age, lactate, and log-transformed procalcitonin. Discrimination, calibration, overall predictive accuracy, reclassification, and potential clinical utility were assessed in the DC. The final model was applied unchanged to the TVC, where discrimination was assessed using receiver operating characteristic curve analysis for 30-day mortality, mechanical ventilation, and hospital admission. Results The DC included 537 patients and the TVC included 519 patients. Thirty-day mortality occurred in 46 patients (8.6%) in the DC and 45 patients (8.7%) in the TVC. In the DC, DECAF score (adjusted odds ratio [aOR] 2.64), log-transformed procalcitonin (aOR 3.32), lactate (aOR 6.17), and age (per 10-year increase; aOR 2.58) were independently associated with 30-day mortality. A-DECAF had a higher area under the curve (AUC) for 30-day mortality than the original DECAF score in the DC (0.983 vs. 0.832; p < 0.001) and TVC (0.958 vs. 0.803; p < 0.001). In the TVC, A-DECAF also had higher AUCs for invasive mechanical ventilation (0.935 vs. 0.830; p < 0.001) and noninvasive mechanical ventilation (0.936 vs. 0.836; p < 0.001). The AUC difference for hospital admission was not statistically significant (0.888 vs. 0.855; p = 0.107). Conclusions A-DECAF had higher observed discrimination for 30-day mortality than DECAF in the development and single-center temporal validation cohorts. These findings should be interpreted cautiously, and further validation in other centers is needed before clinical implementation. Clinical trial number Not applicable.