Discussion
In this paper, we found a significant association between longer duration of mild hypothermia and good neurological outcome at discharge. This finding supports the hypothesis that early mild hypothermia exerts a positive effect on neurological outcome in VA-ECMO patients. This observation stands true even after excluding post-cardiotomy patients who did not experience cardiac arrest. Interestingly, mild hypothermia did not influence in-hospital mortality. Although one may hypothesize that bleeding may be exacerbated with ECMO-associated coagulopathy in addition to hypothermia, this was not found to be the case.
We previously showed that early hyperoxia was a strong marker of poor neurological outcome in VA-ECMO patients.14 In this study, hypothermia proved to be yet another significant factor in neurological outcome, after adjusting for important risk factors including hyperoxia, suggesting temperature management may be an early intervention that improves neurological outcome in this population. Nevertheless, despite our findings, there is conflicting evidence regarding the benefit of therapeutic hypothermia in ECMO patients. In agreement with our study, in a pooled analysis of 13 studies, Chen et al. found a significant association between hypothermia (32–34°C) and favorable neurologic outcomes, defined by a cerebral performance category of 1–2.15 In contrast, a recent meta-analysis of 35 studies showed that among extracorporeal cardiopulmonary resuscitation (eCPR) patients, survival and neurological outcomes were not different between patients who underwent therapeutic hypothermia (ranging between 33-36°C) and patients who did not.16 However, these reports, unlike our study, were limited by high heterogeneity of included studies without granular temperature data.
Furthermore, the data on the effect of mild hypothermia in non-eCPR VA-ECMO patients is sparse. While limited data exists regarding the benefit of early, therapeutic hypothermia in ECMO patients, a physiologic explanation for why it might be beneficial can be opined for all 3 cohorts, eCPR, cardiogenic shock, and post-cardiotomy shock. Reperfusion injury secondary to prolonged low flow in eCPR patients portends a significant neurological injury, which may be mitigated by TTM.17 Also, hypothermia may offer hemodynamic benefits following cardiogenic shock including reduced metabolic rate, increased contractility, and increased cardiac output.18,19 A similar advantageous hemodynamic profile could result in improved neurological outcomes in post-cardiotomy shock patients. This study provides supporting evidence that even in the non-eCPR patients, mild hypothermia was associated with good neurological outcome. Therefore, our study is hypothesis-generating, necessitating further research on hypothermia in each VA-ECMO indication.
This study has several limitations. It is a single center observational study. We included ECMO patients with different indications such as cardiac arrest, cardiogenic shock, and post-cardiotomy shock. However, we performed a sensitivity analysis to exclude patients with cardiac arrest and the and the benefit to hypothermia persisted. Given the limited sample size, a multi-center study is necessary to study this question in each VA-ECMO indication. Despite the small sample size, a beneficial effect of hypothermia was large on neurological outcome.