Abstract
Therapeutic hypothermia, or targeted temperature management (TTM), is recommended for comatose patients after both out-of-hospital and in-hospital cardiac arrest (Callaway et al., 2015). Despite widespread use of temperature management and some supporting animal data, the quality of evidence for therapeutic hypothermia in cardiac arrest is low, and much remains to be learned about the optimal way to apply temperature management. The TTM research group published a systematic review in 2011, which showed that studies performed up until then were at a high risk of bias (Nielsen et al., 2011).
In 2013, the TTM group conducted an international study called the TTM trial (Nielsen et al., 2013). They concluded that, in unconscious survivors of out-of-hospital cardiac arrest of presumed cardiac cause, hypothermia at a targeted temperature of 33°C did not confer a benefit as compared with a targeted temperature of 36°C. A widespread argument of hypothermia for treatment of brain injury has begun since then. In fact, more than five letters were sent to The New England Journal of Medicine (NEJM) at that time commenting on the TTM trial. A notable argument observed that some patients may have prolonged periods of cardiac arrest before the return of spontaneous circulation.
In such cases the neurological structures are already damaged, and using hypothermia to reduce neurological metabolism cannot be expected to have much effect (Perchiazzi et al., 2014). Another letter emphasized the importance of early cooling, noticing that the TTM trial permitted a time of 4 hours before initiating cooling (Nordberg P. Taccone F, 2014). The TTM trial was also questioned regarding whether its results could be widely applied to patients with longer periods to resuscitation, due to the TTM population's short time to Cardiopulmonary Resuscitation (CPR), and its percentage of bystander-initiated CPR (73%), which was much higher than in previous clinical trials (49% to 58%) (Taccone, 2014).
In all accounts of the disputes, target temperature management was still recommended as cardiac arrest guidelines. However, after the TTM trial was published in NEJM in 2013, adherence to recommended TTM decreased significantly. This was explained by a significant decrease in Intravascular Temperature Management use (Garfield et al., 2020). In the United Kingdom, the lowest temperature recorded in OHCA patients within the first 24 hours of admission was higher in the post-TTM trial time than in the pre-TTM trial time (Nolan et al., 2021). The TTM trial continues to bring up many debates in the hypothermia field.
In 2021, NJEM published the TTM2 trial (Dankiewicz et al., 2021), which was a continuation of the collaboration that resulted from the 2013 TTM (hereafter TTM1). As mentioned in the published TTM2 trial protocol, the TTM research group had paid attention to the criticism of the previous trial and designed the TTM2 to respond to that criticism. Using an absolute risk reduction of 7.5% as the anticipated hypothermia intervention effect to calculate the sample size, the research group recruited 1900 adults, twice as many as the TTM1 trial had recruited, to increase confidence limits.
Moreover, the two groups (hypothermia and normothermia) were indeed relatively balanced after randomization. The TTM1 trial had been criticized in that neurocognitive testing follow-up should have been further delayed beyond the 6-month visit used in the trial; TTM2 delayed follow-up until 6 months. In TTM1, the primary outcome was all-cause mortality through the end of the trial; the time for the primary outcome in TTM2 was extended to 6 months. TTM1 was based on the Cerebral Performance Category (CPC) scale and a modified Rankin scale, which were criticized as too simple to assess cognitive prognosis; In TTM2, health-related quality of life, as assessed with the use of the EQ-5D-5L visual analog scale, was added as an important outcome measurement.
In addition, patients in TTM2 were cooled at a similar or faster rate than in most previous trials to counter the criticism of insufficiently early cooling. In general, the TTM2 trial was designed and conducted as an excellent textbook study in cardiac arrest, and the researchers came to the same conclusion as in TTM1, that hypothermia intervention conferred no benefit. Only a few details may need further discussion. For example, the research was carried out in 21 centers in 14 countries, and the number recruited ranged from 152 (Bristol Royal Infirmary) to 1 (Australia). Whether there is a central effect can be further explored.
Yet, some questions still remain to be clarified before abandoning hypothermia in cardiac arrest.
We have to give a second look at the study population to determine if the results are universal enough for wide application. Data from the TTM2 still showed a very short time to CPR, a high percentage of bystander-initiated CPR (82%, as compared with 73% in France (Lascarrou et al., 2019a) and 70.5% in the United States (Fordyce et al., 2017)), and the remarkable overall 6-month survival rate of ∼50% of OHCA patients, when the historical value is ∼20–40% (Myat et al., 2018). Clinical characteristics of the TTM2 trial were reported as ∼70% of patients presenting with bilateral pupillary reflexes; the median arterial pH was 7.2 and the arterial lactate level was 5.8 mmol/L.
According to the outcome prediction tool NULL-PLEASE designed for cardiac arrest patients (Gue et al., 2020), the predictive score could be very low in the TTM2 trial population, since the population's proportion of shockable rhythm, witnessed arrest, short low-flow period, initial blood pH >7.2, and initial blood lactate <7 mmoL are all considered as favorable manifestations. In particular, 70% of first rhythm monitors in the trial were shockable rhythms, which are thought to have better outcomes than nonshockable rhythms (Cournoyer et al., 2019).
In 2019, NEJM published a similar open-label randomized controlled trial focusing on patients resuscitated from both out-of-hospital and in-hospital cardiac arrest with nonshockable rhythm due to any cause (Lascarrou et al., 2019b); one of its conclusions was an endorsement of therapeutic hypothermia. The TTM2 trial, in contrast, did not recruit all kinds of cardiac arrests, but limited the study population to patients with presumed cardiac conditions or unknown causes. The exclusion of other causes, such as respiratory, made the condition of the study population less severe, since respiratory causes account for a large percentage of cardiac arrests and have a worse outcome than cardiac causes (Orban et al., 2018). The question, then, is whether temperature has same effect on patients with different kinds of cardiac arrest. Moreover, was the condition of the patients in the TTM2 trial too mild to benefit from hypothermia?
We performed a network meta-analysis online (Duan et al., 2022) that included 9005 patients from 42 trials. Compared with other groups, patients in the <20 minutes+TTM group were more likely to have better survival and good neurological outcomes (probability = 46.1 and 52.5%, respectively). In comparing the same time groups with and without TTM, only the survival and neurological outcome of the 20–39 minutes+TTM group was significantly better than that of the 20–39 minutes group [odds ratio, OR = 1.41, 95% confidence interval, CI (1.04–1.91); OR = 1.46, 95% CI (1.07–2.00), respectively].
Applying TTM with <20 minutes or >40 minutes of collapse to Return of Spontaneous Circulation (ROSC) did not improve survival or neurological outcome [<20 minutes vs. <20 minutes+TTM: OR = 1.02, 95% CI (0.61–1.71)/OR = 1.03, 95% CI (0.61–1.75); 40–59 minutes vs. 40–59 minutes+TTM: OR = 1.50, 95% CI (0.97–2.32)/OR = 1.40, 95% CI (0.81–2.44); ≥60 minutes vs. ≥60 minutes+TTM: OR = 2.09, 95% CI (0.70–6.24)/OR = 4.14, 95% CI (0.91–18.74), respectively]. Both survival and good neurological outcome were closely related to the time from collapse to ROSC. Therefore, we believe that for the application of TTM, there is an optimum time window that has a clear correlation with the time between collapse and ROSC. Early or late implementation may not bring benefits.
The results of both TTM trials are at high risk of being misinterpreted. Both trials carefully controlled the target temperature using multiple methods. Whatever the temperature, deliberate care bundles involved in TTM, as guided by the protocol, are needed (Morrison and Thoma, 2021). At this point, there are many knowledge gaps in this field that need to be filled. For example, the target temperature may not be the same for all cardiac arrest patients, and further study may focus on subgroups to determine the effect of therapeutic hypothermia. It may be reasonable to make decisions based on a patient's resuscitation method, etiology, and condition. How to define the criteria for appropriate TTM? We may need more TTM trials to answer those questions.