Influencing factors of poor prognosis after extracorporeal membrane oxygenation weaning in patients with cardiopulmonary arrest and the establishment of a nomogram prediction model

Objective To explore the influencing factors of poor prognosis in patients with cardiac arrest (CPA) after extracorporeal membrane oxygenation (ECMO) weaning, and to construct a nomogram prediction model.

Methods A total of 189 CPA patients admitted to our hospital from May 2020 to August 2023 were selected, and were randomly dividedinto validation group (n=57) and modeling group (n=132). According to the prognosis condition, CPA patients in the modeling group with ECMO weaning within 28 days were grouped into good prognosis group (n=49, the survival) and poor prognosis group (n=83, the death). The influencing factors of poor prognosis after ECMO weaning in CPA patients were analyzed using multivariate Logistic regression; the nomogram model for predicting poor prognosis after ECMO weaning in CPA patients was constructed using R4.3.1; receiver operating characteristic(ROC) curve, calibration curve, and clinical decision-making (DCA) curve were applied to evaluate the predictive performance and clinical application value of the nomogram model.

Results The CPA-to-ECMO treatment time, blood lactate levels, and the proportion of patients combining CRRT in the poorprognosis group were longer or higher than those in the good prognosis group, while the hospital stay was shorter than that in the good prognosis group (P < 0.05). Long CPA-to-ECMO treatment time, high blood lactate levels, and combined continuous renal replacement therapy (CRRT) were independent risk factors for poor prognosis after ECMO weaning in CPA patients, while long hospital stay was a protective factor (P < 0.05). The areas under the curve for the modeling and validation groups were 0.913 and 0.896, respectively.In the Hosmer-Lemeshow goodness-of-fit test, chi-square value was 9.511 and P value was 0.301 for the modeling group, and chi-square value was 8.105 and P value was 0.423 for the validation group.When the high-risk threshold predicted by the nomogram model ranged from 0.05 to 0.72, it had high clinical application value.

Conclusion Long CPA-to-ECMO treatment time, high blood lactate levels, combined CRRT, and short hospital stay were factors influencing poor prognosis after ECMO weaning in CPA patients.The nomogram model constructed based on these four factors had high predictive performance.