Lessons from a pragmatic trial of field targeted temperature management

The concept of initiating targeted temperature management (TTM, previously referred to as therapeutic hypothermia) in the field for victims of out-of-hospital cardiac arrest (OHCA) who obtain return of spontaneous circulation (ROSC) seems intuitively attractive. Research has demonstrated that “time is [heart] muscle” and “time is brain” – it seems reasonable to assume that “time is chill”, and that starting cooling as soon as possible after ROSC would maximize its benefits to the ischemic brain. However, the data thus far do not bear this out. A number of trials of various designs, and two meta-analyses of these data, have not demonstrated any significant outcome benefit when TTM is initiated in the field by EMS personnel as compared to when it is initiated in the hospital [[1], [2]].

In this issue of Resuscitation, Scales and colleagues report another randomized controlled trial of TTM applied either in the field or in the hospital – but their primary objective was not to look at patient outcomes, but rather whether initiating TTM in the field can “act as a catalyst to encourage more timely application of TTM by in-hospital clinicians” [3]. Their primary outcome was “successful TTM”, defined as reaching 32–34?°C with 6?h of arrival at the hospital. Among the secondary outcomes was ever receiving TTM in-hospital. Safety parameters of pulmonary edema identified in the ED and re-arrest during transport were included, as these have been identified as potential risks to prehospital TTM in prior work [4]. Survival to 6?h and to discharge, and survival with good neurologic outcome, were also among the secondary outcomes, and the study as designed was adequately powered to detect an improvement from 20% to 28% in the latter. The intervention included chilled intravenous saline and surface ice packs, but also a wristband placed on patients, notifying in-hospital personnel that cooling had been initiated in the field, and encouraging them to continue these efforts.

Significant differences were found in only two of the outcomes: a higher rate of “ever TTM” was seen, and a lower rate of pulmonary edema detected in the ED was seen, among those with TTM initiated in the field. No differences were seen in the primary outcome of successful TTM, or the other secondary outcomes, including any of the three survival measures. The study did not meet its enrollment target, and thus may have been underpowered to detect some differences it might otherwise have found. It is also worth noting that 115 of the 700 eligible patients, or 16%, were not randomized. The authors indicate that this was largely due to paramedics refusing to participate in research. While this loss of eligible subjects is troubling overall, and may have affected the ability of the project to detect some small but potentially important differences between groups, it seems unlikely to have biased the project in one direction or the other.

The lower rate of pulmonary edema detected in the ED among patients with TTM initiated in the field is a bit puzzling. Concern has been expressed that fairly large volumes of chilled saline delivered rapidly to initiate cooling could result in pulmonary edema [5], and this has been seen in both human [4] and animal studies [6]. But the authors’ data show that patients in the field TTM group received, on average, only 170 cc of fluid more than their in-hospital TTM counterparts, and the mean volume of cold saline infused was less than half a liter. By comparison, nearly half of the patients in the earlier study demonstrating increased pulmonary edema received a full two liters in the field [4]. The authors of the current study note that the smaller volumes of cold saline may have decreased the potential for pulmonary edema or recurrent arrest via arrhythmias; at the same time, it could be argued that it may have decreased the potential for positive findings among other outcome measures.

The authors describe their study as a pragmatic trial. Such studies are “designed to evaluate the effectiveness of interventions in real-life routine practice conditions” [7], as compared to traditional clinical efficacy research, which studies discreet interventions in tightly controlled circumstances. Many of us engaged in EMS research have seen the “real-life routine practice conditions” that have adversely affected our enrollment numbers, adherence to protocol, completeness of data collection, etc. Back in 1995, Spaite taught us that essentially all EMS research is “systems research”, noting that “Data collection often involves many individuals…many agencies are involved, and principle investigators frequently do not directly participate in the actual data-collection process. The research environment is of complex and uncontrolled and contains several sources of information. The desired data elements are often difficult to obtain with a high degree of accuracy, and outcome parameters are complex” [8]. While this description seems to support the type of pragmatic real-world study being reported here, Spaite was actually arguing for adopting research methodologies used in engineering, behavioral science, public administration, and social science to develop multidisciplinary systems analysis research methods that might allow us to move even farther forward in our research efforts. While that has happened to some extent, the majority of EMS research that I see still clings to adapting traditional clinical research models to the non-traditional environment, to varying degrees of success.

In the end, the findings of the study by Scales et al. are encouraging in that perhaps EMS can drive meaningful change further down the care continuum. However, despite the improvement noted in “ever received TTM”, no outcome differences were seen ? perhaps due to under-powering, perhaps due to low chilled saline volumes delivered, perhaps for any of a dozen or so other reasons. A recent paper by Ji et al. examining post-admission outcomes in the PARAMEDIC trial similarly found no meaningful differences between those managed with mechanical vs manual CPR among those who survived to hospital admission [9]. Perhaps those who survive to admission, either because of or despite our best efforts in the field, do uniformly well enough to make detection of differences between groups unlikely. This is not to say that we should not look at post-admission factors in our OHCA efforts; but it may also be telling us to work harder to adopt Spaite’s recommendations of over 20 years ago.