Study design and population
This was a retrospective analysis of prospectively collected data from adult cardiac arrest survivors who had undergone TTM at Chonnam National University Hospital, a university-affiliated hospital in Gwangju, Korea, between October 2015 and December 2020. This study was approved by the Institutional Review Board of Chonnam National University Hospital (CNUH-2015-164). Informed written consent was obtained from the patient and/or the patient’s legal representative. This study was conducted in accordance with the principles of the Declaration of Helsinki25.
We included cardiac arrest survivors aged over 18 years who had undergone TTM with an Arctic Sun® system. We excluded patients who did not complete TTM due to transfer or death, patients with a target temperature other than 33 °C, patients who had undergone TTM with devices other than the Arctic Sun® system, and patients with missing core or water temperature data.
We classified patients who had ECMO and/or CRRT during TTM as the extracorporeal group. ECMO was used for extracorporeal CPR or circulatory support after ROSC. OHCA patients who had experienced collapse with bystander CPR, initial shockable rhythm, and 30 min of resuscitation effort (including basic and advanced life support) were indicated for extracorporeal CPR. Circulatory support with ECMO after ROSC was only provided to patients expected to have good neurological outcomes. The attending physician decided whether to apply ECMO. CRRT was used for patients with Kidney Disease Improving Global Outcomes stage 2 or 3. We used ECMO without a thermoregulator and CRRT with a thermoregulator.
TTM and temperature measurement
We provided post-cardiac arrest care, including TTM, to comatose cardiac arrest survivors in accordance with the guidelines1. The target temperature of 33 °C was achieved as soon as possible after ROSC and then maintained for 24 h. After completion of the maintenance phase, patients were rewarmed to 36.5 °C at a rate of 0.25 °C/h. We administered midazolam and remifentanil for analgo-sedation and neuromuscular blockade with continuous infusion during TTM, to enhance TTM efficiency and to reduce brain metabolism. Amplitude-integrated electroencephalography was used to monitor subclinical seizures and propofol was administered to control seizures.
The Arctic Sun® system functions based on the core temperature measured using an esophageal probe and core temperature and water temperature were recorded automatically every minute throughout TTM.
Data collection and definition
We obtained the following demographic data, and cardiac arrest and post-cardiac arrest profiles: age, sex, body mass index, preexisting illness, witness of collapse, bystander CPR, the first monitored rhythm (shockable or non-shockable), etiology of cardiac arrest (cardiac or non-cardiac), epinephrine dose during CPR, time to ROSC, GCS after ROSC, serum lactate after ROSC, blood glucose after ROSC, PaO2 and PaCO2 after ROSC, SOFA within the first day of cardiac arrest26, ECMO, CRRT, time from ROSC to start of TTM, body temperature at the start of TTM, and time to achieve target temperature.
The cooling rate during the induction period was calculated as the difference between the core temperature immediately before initiation of TTM and 33 °C to the time required to reach the target temperature. We calculated various indices to identify the temperature variability during the 24-h maintenance phase. We tried to simultaneously present the extent and time required for the core temperature to deviate from the target temperature of 33 °C during the maintenance phase. Therefore, we introduced the concept of TWCT, which can quantify core temperature and time. We calculated the TWCT based on the core temperature measured every 1 min (Fig. 1). Thereby, we defined the sum of TWCT > 33 °C as positive TWCT, whereas the sum of TWCT < 33 °C was defined as negative TWCT. We defined the sum of TWCT > 33.5 °C as the undercooling value, and the sum of TWCT < 32.5 °C as the overcooling value. The primary outcome was a negative TWCT.
We assessed the Cerebral Performance Category (CPC) scale at 6 months after cardiac arrest via phone interviews and recorded it as follows: CPC 1 (good performance), CPC 2 (moderate disability), CPC 3 (severe disability), CPC 4 (vegetative state), or CPC 5 (brain death or death)27. The secondary outcomes were poor neurological outcomes (CPC 3–5), cooling rate, positive TWCT, overcooling TWCT, and undercooling TWCT.
Categorical variables, reported as frequencies with percentages, were compared between the two groups using χ2 tests with continuity correction in 2 × 2 tables or Fisher’s exact test, as appropriate. Continuous variables, reported as medians with interquartile ranges or mean with standard deviation according to the Shapiro–Wilk test, were compared between the two groups using Mann–Whitney U tests or independent t-tests, as appropriate.
To investigate the association between the use of extracorporeal devices and negative TWCT, we performed multivariate linear regression analyses. We performed multivariate logistic regression analyses using variables with p values < 0.2 (Table 1) and selected covariates for the linear regression analysis. We performed multivariate linear regression analyses in each subgroup of ECMO and CRRT in comparison with the no-extracorporeal group. Patients who received both ECMO and CRRT were included in the ECMO group.
We performed multivariate logistic regression analysis to investigate the association between the use of extracorporeal devices and neurological outcomes. We included all variables with p < 0.2 in the univariate analyses between neurological outcome groups (except cooling rate) as covariates in the logistic regression model. Enter method was used to develop the final adjusting model. The goodness of fit of the final model was evaluated using the Hosmer–Lemeshow test. We report logistic regression analysis results as odds ratios (OR) with 95% confidence intervals (CI). Data were analyzed using IBM SPSS Statistics 26.0 for Windows (IBM Corp., Armonk, NY, USA). A two-sided significance level of 0.05 was used to indicate statistical significance.