Nomograms to predict outcome for patients undergoing venoarterial extracorporeal membrane oxygenation treatment for septic shock.

Abstract

Venoarterial extracorporeal membrane oxygenation (VA-ECMO) is increasingly being employed to treat patients with refractory septic shock. Despite its growing use, there is a notable absence of prognostic assessment tools specifically designed for septic shock patients who have received VA-ECMO therapy. The aim of this study is to develop and validate a prognostic model for patients with refractory septic shock undergoing VA-ECMO, and to identify those who may derive the greatest benefit from this treatment. This single-center, retrospective cohort study was conducted at a comprehensive intensive care unit in China. Adult patients with refractory septic shock who received VA-ECMO treatment were included. Two hundred consecutive patients were randomly divided into training and validation cohorts in a 7:3 ratio. Least absolute shrinkage and selection operator regression analysis was employed to select relevant variables for the logistic regression model, and its performance was tested in both training and validation cohorts based on discrimination, calibration, and net benefit. Between January 2019 and September 2024, 293 patients were screened, 200 of whom were eligible and were divided into development (n = 140) and validation (n = 60) cohorts. The 28-day survival rate was 23.0%, and median duration of ECMO run was 6.0 days (IQR 2.0-8.0). Age, APACHE II score at ICU admission, immunosuppression status, hypertension, IL-6, and APTT measured within 6 h before ECMO initiation were the six predictors included in the nomograms. The nomogram demonstrated strong discriminative power in the training cohort (area under the curve [AUC]: 0.873, 95% CI 0.812-0.929), as well as in the validation cohort (area under the curve [AUC]: 0.818 (95% CI 0.687-0.920). The model's reliability in predicting outcomes was evident from the high consistency between predicted probabilities and observed proportions during calibration. Decision curve analysis indicated that the model's clinical benefit was advantageous. The novel validated nomogram is designed to predict outcomes after VA-ECMO treatment in individuals with refractory septic shock. It can support physicians in performing precise mortality risk evaluations and making more informed decisions regarding the application of VA-ECMO treatment.