Factors predicting mortality in the cardiac ICU during the early phase of targeted temperature management in the treatment of post-cardiac arrest syndrome - The RAPID score

Bettina Nagy; Ádám Pál-Jakab; Gábor Orbán; Boldizsár Kiss; Alexa Fekete-Győr; Gábor Koós; Béla Merkely; István Hizoh; Enikő Kovács; Endre Zima

doi:10.1016/j.resplu.2024.100732

Factors predicting mortality in the cardiac ICU during the early phase of targeted temperature management in the treatment of post-cardiac arrest syndrome - The RAPID score

Resusc Plus. 2024 Aug 16:19:100732. doi: 10.1016/j.resplu.2024.100732. eCollection 2024 Sep.

Authors

Bettina Nagy¹, Ádám Pál-Jakab¹, Gábor Orbán¹, Boldizsár Kiss¹, Alexa Fekete-Győr², Gábor Koós¹, Béla Merkely¹, István Hizoh¹, Enikő Kovács^{1

3

4}, Endre Zima^{1

4

5}

Affiliations

¹ Semmelweis University Heart and Vascular Center, Budapest, Hungary.
² Chelsea and Westminster Hospital NHS Foundation Trust, London, United Kingdom.
³ Semmelweis University, University Department of Anaesthesiology and Intensive Therapy, Hungary.
⁴ Hungarian Resuscitation Council, Hungary.
⁵ Institute of Anesthesiology and Perioperative Care, Semmelweis University, Budapest, Hungary.

Abstract

Introduction: Survival rates after out-of-hospital cardiac arrest (OHCA) remain low, and early prognostication is challenging. While numerous intensive care unit scoring systems exist, their utility in the early hours following hospital admission, specifically in the targeted temperature management (TTM) population, is questionable. Our aim was to create a score system that may accurately estimate outcome within the first 12 h after admission in patients receiving TTM.

Methods: We analyzed data from 103 OHCA patients who subsequently underwent TTM between 2016 and 2022. Patient demographic data, prehospital characteristics, clinical and laboratory parameters were already available in the first 12 h after admission were collected. Following a bootstrap-based predictor selection, we constructed a nonlinear logistic regression model. Internal validation was performed using bootstrap resampling. Discrimination was described using the c-statistic, whereas calibration was characterized by the intercept and slope.

Results: According to the Akaike Information Criterion (AIC) heart rate (AIC = 9.24, p = 0.0013), age (AIC = 4.39, p = 0.0115), pH (AIC = 3.68, p = 0.0171), initial rhythm (AIC = 4.76, p = 0.0093) and right ventricular end-diastolic diameter (AIC = 2.49, p = 0.0342) were associated with 30-day mortality and were used to build our predictive model and nomogram. The area under the receiver-operating characteristics curve for the model was 0.84. The model achieved a C-statistic of 0.7974, with internally validated acceptable calibration (intercept: -0.0190, slope: 0.7772) and low error rates (mean absolute error: 0.040).

Conclusion: The model we have developed may be suitable for early risk assessment of patients receiving TTM as part of primary post-resuscitation care. The calculator needed for scoring can be accessed at the following link: https://www.rapidscore.eu/.

Keywords: Cardiopulmonary Resuscitation; Out-of-Hospital Cardiac Arrest; Prediction of Mortality; Sudden Cardiac Arrest; Targeted Temperature Management.