Background: Therapeutic hypothermia has been recommended for eligible patients after cardiac arrest (CA) in order to improve outcomes. Up to now, several comparative observational studies have evaluated the combined use of extracorporeal cardiopulmonary resuscitation (ECPR) and therapeutic hypothermia in adult patients with CA. However, the effects of therapeutic hypothermia in adult CA patients receiving ECPR are inconsistent. Methods: Relevant studies in English databases (PubMed, ISI web of science, OVID, and Embase) were systematically searched up to September 2019. Odds ratios (ORs) from eligible studies were extracted and pooled to summarize the associations of therapeutic hypothermia with favorable neurological outcomes and survival in adult CA patients receiving ECPR.Results: 13 articles were included in the present meta-analysis study. There were nine studies with a total of 806 cases reporting the association of therapeutic hypothermia with neurological outcomes in CA patients receiving ECPR. Pooling analysis suggested that therapeutic hypothermia was significantly associated with favorable neurological outcomes in overall (N = 9, OR = 3.507, 95%CI = 2.194-5.607, P < 0.001, fixed-effects model) and in all subgroups according to control type, regions, sample size, CA location, ORs obtained methods, follow-up period, and modified Newcastle Ottawa Scale (mNOS) scores. There were nine studies with a total of 806 cases assessing the association of therapeutic hypothermia with survival in CA patients receiving ECPR. After pooling the ORs, therapeutic hypothermia was found to be significantly associated with survival in overall (N = 9, OR = 2.540, 95%CI = 1.245-5.180, P = 0.010, random-effects model) and in some subgroups. Publication bias was found when evaluating the association of therapeutic hypothermia with neurological outcomes in CA patients receiving ECPR. Additional trim-and-fill analysis estimated four "missing" studies, which adjusted the effect size to 2.800 (95%CI = 1.842-4.526, P < 0.001, fixed-effects model) for neurological outcomes.Conclusions: Therapeutic hypothermia may be associated with favorable neurological outcomes and survival in adult CA patients undergoing ECPR. However, the result should be treated carefully because it is a synthesis of low-level evidence and other limitations exist in present study. It is necessary to perform randomized controlled trials to validate our result before considering the result in clinical practices.