Zhonghua Wei Zhong Bing Ji Jiu Yi Xue. 2026 Jan;38(1):146-151. doi: 10.3760/cma.j.cn121430-20250319-00163.
ABSTRACT
OBJECTIVE: To investigate the factors influencing the apnea test (AT) and its clinical effects in brain death determination under updated criteria, and to provide evidence for optimizing and reducing the risk of false-negative results and complications.
METHODS: Based on the data from the Anhui Provincial Brain Injury Evaluation Quality Control Center, the data of brain-dead patients who completed AT with an off ventilator duration of 5-11 minutes were analyzed retrospectively. Data from January 2018 to March 2025 were used as the model development cohort, and the data from June to December 2025 were used as the external validation cohort. Demographic characteristics, clinical data, evaluation and examination indicators, AT operation details, etc. were extracted using standardized case report form. Temporal trends of AT positive rate and the incidences of severe hypercapnia, acidosis, hypoxemia and other complications when offline for 5-11 minutes were evaluated using the Cochran-Armitage trend test, and the key factors affecting the change of arterial partial pressure of carbon dioxide (PaCO2) and pH were analyzed by multiple linear regression model.
RESULTS: The model development cohort included 384 patients with brain death, and the external validation cohort included 47 patients with brain death. There was no significant difference in baseline characteristics between the two cohorts (all P>0.05). With the extension of offline time, the positive rate of AT was gradually increased (Cochran-Armitage trend test: Z=3.52, P<0.001), rising from 76.5% (13/17) at 5 minutes to 91.7% (11/12) at 7 minutes, and plateaued after 7 minutes. The trend analysis of complications in the same period showed that the incidence of severe hypercapnia (PaCO2>80 mmHg, 1 mmHg=0.133 kPa) showed a significant increasing trend (Z=4.09, P<0.001), and was higher at 10 minutes than at 9 minutes [44.7% (59/132) vs. 21.6% (8/37), P<0.05]. Severe acidosis (pH<7.20) became more frequent over time (Z=-4.69, P<0.001), and was higher at 10 minutes than at 7 minutes [73.5% (97/132) vs. 58.3% (7/12), P<0.05]. The incidence of hypoxemia [arterial partial pressure of oxygen (PaO2) <60 mmHg] showed a decreasing trend (Z=-5.21, P<0.001), with no statistically significant difference in incidence between 7-11 minutes (F=0.859, P=0.525). The prediction model was established by multiple regression, indicated that offline time, pre-AT PaCO2, pre-AT pH, heart rate, and body weight collectively influenced post-AT PaCO2 (R2=0.284, P<0.001). Offline time, pre-AT pH, heart rate, and hemoglobin were associated with post-AT pH (R2=0.455, P<0.001). External validation indicated good performance for the pH model (mean absolute error was 0.038, R2=0.69) and acceptable performance for the PaCO2 model (mean absolute error was 6.21 mmHg, R2=0.62).
CONCLUSIONS: When implementing the dual-criteria standard (PaCO2 and pH), an offline time window of 7 to 9 minutes can balance diagnostic efficacy for brain death with patient safety. Pre-intervention strategies, such as lowering pH or raising PaCO2 before disconnection, may shorten the time needed to reach AT targets. However, should be guided by a comprehensive assessment of individualized patient factors, including baseline pH, PaCO2, heart rate, hemoglobin, and body weight.
PMID:41876240 | DOI:10.3760/cma.j.cn121430-20250319-00163
