In this study evaluating the database of a randomized clinical trial, we evaluated the differences in temperature control between IC and SFC methods to achieve TTM among survivors of OHCA. The use of IC was associated with a similar time to target temperature (i.e., speed of cooling), a lower temperature variability during the maintenance phase, a slower rewarming rate, but more frequent post-TTM fever than SFC devices. No differences in mortality or poor neurological outcome between the two methods were observed. Also, the rate of adverse events was similar between groups.
The optimal cooling technique to deliver TTM after OHCA remains still unknown. Although international guidelines recommend to use devices with a continuous temperature feedback control (TFC) in this setting, to minimize the risk of overcooling and to provide a more stable control of target temperature [17], invasive methods using IC with TFC are the most precise cooling techniques to provide TTM, even when compared to external SFC devices [13]. However, whether this might influence patients’ outcome, it is still debated. The choice of one or the other method is usually driven by different considerations such as precision, workload for nurses, costs, and the risk/benefit ratio. External SFC methods are applied around the torso and limbs and can adequately induce and maintain TTM in OHCA patients [18]; however, they also require more nursing attention, are frequently associated with shivering, may reduce overall access to the patient, and induce some skin damage [19]. Intravenous cooling with IC devices are effective for TTM in OHCA patients, but they need a trained physician to insert the catheter into the large vein and may be associated with a higher risk of infection and bleeding [19].
Assessment of IC and SFC devices has already been performed in large animal models, showing a faster cooling rate for IC when compared with others [20]. In humans, results are more controversial. Ferreira et al. showed that the time to target temperature was faster and the rewarming rate lower in IC-treated patients when compared with the SFC group [21]. On the opposite, Tømte et al. observed no differences in cooling rates, post-TTM fever, and the occurrence of main adverse events between SFC and IC cooling methods [22]. Similar results on cooling rate and the occurrence of overcooling were observed in another study, although target temperature maintenance was more stable in the IC group than in SFC-cooled patients [23]. Less fluctuation of body temperature during the maintenance phase with the use of IC devices when compared with SFC methods was observed also in other studies [7, 24]. In our study, the cooling rates were similar between groups. However, there was a higher number of patients in the SFC group receiving also cold fluids; administration of cold fluids is actually the most common and rapid method to induce hypothermia, especially when administered in the pre-hospital setting, although some concerns on its effectiveness and potential harms exist [25], and may have contributed to faster-cooling rates in the SFC group. We observed a lower TV in the IC-treated patients than in the SFC group. Despite this finding might again underline a higher precision to maintain the target temperature using IC, TV was somewhat higher in patients with favorable than unfavorable neurologic outcome [16] and might suggest intact thermoregulatory pathways that aim to restore a body temperature close to 37.0 °C rather than a target to optimize TTM. The difference in TV was so limited (< 0.1 °C) between groups that one may argue whether this can translate in clinically relevant benefits on patients’ outcome. Moreover, the proportion of time outside therapeutic ranges was similar between groups, suggesting that despite a more stable core temperature, the effectiveness of the devices to maintain the body temperature in the selected target was similar. Finally, rewarming rate was lower in the IC group, although the difference was numerically small (i.e., 0.06 °C). There are no clinical data suggesting a critical cut-off for harm of excessively rapid rewarming after TTM in OHCA survivors, although high rewarming rate, exceeding 0.5 °C, might be associated with a higher risk of poor neurological outcome (71% vs. 52%) than slower rewarming [26]. As both cooling devices in our study had rewarming rate below this cut-off of 0.5C°/h, it is hard to conclude that the minimal differences in rewarming rate observed in this cohort can result in any clinically relevant result.
Previous studies have also reported more frequent complications with the IC than with SFC devices, although these findings remain discordant and might be due to chance or selection bias among different studies. In one study, infectious and cardiovascular complications were similar between IC and SFC [7]. In a large cohort, shivering, electrolyte disturbances, and arrhythmias were also similar between IC and SFC devices [15]. In a small RCT, pneumonia and ischemic stroke were more frequent in the SFC group, while the occurrence of significant arrhythmias and renal failure was higher in the IC group [11]. In a matched-control analysis, the use of IC was associated with a greater incidence of sepsis [27]. In another study, an increase in bleeding was observed in the IC group (14% versus 2%, p = 0.11) [14]. We also observed relatively similar occurrence of adverse events between IC and SFC devices in our study, except for more frequent arrhythmias and hypokalemia in the IC group, probably due to a larger proportion of patients being cooled for 48 h. Also, post-TTM fever was more frequent in the IC group, which is in contrast with previous publications [22, 23]. These findings might be potentially explained either by a higher number of recorded temperatures in the IC group (i.e., more data available which translates in a higher probability to detect fever) or by the early removal of the endovascular system, because of the risk of infection and bleeding, which would have exposed these patients to a less accurate temperature control in the post-cooling phase. Importantly, as no specific data on shivering control, sedation policies, and/or catheter removal were reported, we could not specifically analyze the determinants of post-TTM fever in this cohort. However, the safety profile of both cooling systems looks comparable, and the selection of one strategy over the other might be influenced by the potential benefits on mortality and neurological recovery.
Together with some advantages in the TTM delivery for IC over SFC devices and a similar safety profile, no significant differences in patients’ outcome were observed between groups. Whether our study is simply underpowered to show any statistical significance between the methods or if this difference is related to an imbalance between IC and SFC (i.e., more patients cooled for 48 h, less arrest occurring at home, and a less significant difference in outcome when adjusted for confounders), it is impossible to conclude from our data. In a recent post hoc analysis of a large RCT including 934 patients [15], mortality and poor neurological outcome was lower in the IC group when compared with SFC devices (46.3% vs. 50.0% and 49.0% vs. 54.3%, respectively). Other studies also showed a non-significant reduction in mortality or poor neurological outcome in the IC group of around 5–10% [7, 14, 22, 23], although this was not consistent in another study [27, 28]. Importantly, this might be explained by chance, selection bias or by different case-mixes (i.e., large referral centers with PCI facilities tend to use endovascular cooling techniques). Moreover, SFC devices are not entirely comparable as they include a wide range of devices, from simple ice bags to sophisticated machines with automatic TFC using blankets containing circulating coolant, and these differences may also impact on outcome in this setting. In one study, Shinada et al. observed that the use of SFC with self-adhesive, hydrogel-coated pads gel-circulating and TFC was associated with a lower mortality and poor neurological outcome (20% vs. 27% and 28% vs. 45%, respectively) than conventional SFC using blankets [18]. In another small RCT, Heard et al. showed that hydrogel-coated pads gel-circulating and TFC had a lower proportion of patients with poor neurological outcome (39% vs. 47%) than the use of cooling blankets and ice [11]. As such, future studies comparing IC and SFC with TFC are needed to better understand the impact of such cooling strategies on patients’ outcome. Also, a systematic review and meta-analysis of the existing literature might help to further quantify the relevance of cooling methods on mortality and neurological outcome of OHCA survivors.
This study has also several limitations. First, selection bias related to the use of different cooling methods in the different participating centers may account for some of the differences between groups. In particular, it is impossible to consider whether a real choice or equipoise existed between IC and SFC at the time point of treatment selection for those centers where both devices were available. Moreover, some centers had only of the two devices available so that differences in outcome, despite similar cooling times, might be influenced by the “site” effect, although this was considered into the adjusted analysis. Second, the study was not powered to detect differences in clinical outcome. Third, the accuracy of intervals, such as time to cooling or target temperature, might be unreliable since this information may be subject to reporting or measurement errors. Forth, different sites of temperature measurement were used, which may have led to a measurement bias. Also, brain temperature typically exceeds body temperature by 0.5 to 2.0 °C after an acute brain injury, so that measuring peripheral temperature is a poor indicator of temperature control within the cerebral tissue [28]. Fifth, this was not a randomized study primarily investigating performance of cooling devices; as such, the risk of bias is high and any association found should be cautiously interpreted. Also, we did not report about the type of devices (i.e., different types of endovascular devices or the proportion of cold-water circulating blankets vs. hydrogel pads in the surface cooling group), and this may contribute to a significant variability in TTM effectiveness and precision within the groups. Moreover, some differences between groups (i.e., arrest location) were not considered in the adjusted analysis of outcomes. Finally, we have analyzed cooling devices by categories rather than comparing individual IC or SFC devices and the heterogeneity among different systems can also account for some observed differences between groups.