Department of Adult Intensive Care Medicine, Lausanne University Hospital, Rue du Bugnon 46, 1011, Lausanne, Switzerland.
Faculty of Biology and Medicine, University of Lausanne, Lausanne, Switzerland.
Neurocrit Care. 2022 Aug;37(1):293-301. doi: 10.1007/s12028-022-01516-0. Epub 2022 May 9.
According to international guidelines, neuroprognostication in comatose patients after cardiac arrest (CA) is performed using a multimodal approach. However, patients undergoing extracorporeal membrane oxygenation (ECMO) may have longer pharmacological sedation and show alteration in biological markers, potentially challenging prognostication. Here, we aimed to assess whether routinely used predictors of poor neurological outcome also exert an acceptable performance in patients undergoing ECMO after CA.
This observational retrospective study of our registry includes consecutive comatose adults after CA. Patients deceased within 36 h and not undergoing prognostic tests were excluded. Veno-arterial ECMO was initiated in patients < 80 years old presenting a refractory CA, with a no flow < 5 min and a low flow ≤ 60 min on admission. Neuroprognostication test performance (including pupillary reflex, electroencephalogram, somatosensory-evoked potentials, neuron-specific enolase) toward mortality and poor functional outcome (Cerebral Performance Categories [CPC] score 3-5) was compared between patients undergoing ECMO and those without ECMO.
We analyzed 397 patients without ECMO and 50 undergoing ECMO. The median age was 65 (interquartile range 54-74), and 69.8% of patients were men. Most had a cardiac etiology (67.6%); 52% of the patients had a shockable rhythm, and the median time to return of an effective circulation was 20 (interquartile range 10-28) minutes. Compared with those without ECMO, patients receiving ECMO had worse functional outcome (74% with CPC scores 3-5 vs. 59%, p = 0.040) and a nonsignificant higher mortality (60% vs. 47%, p = 0.080). Apart from the neuron-specific enolase level (higher in patients with ECMO, p < 0.001), the presence of prognostic items (pupillary reflex, electroencephalogram background and reactivity, somatosensory-evoked potentials, and myoclonus) related to unfavorable outcome (CPC score 3-5) in both groups was similar, as was the prevalence of at least any two such items concomitantly. The specificity of each these variables toward poor outcome was between 92 and 100% in both groups, and of the combination of at least two items, it was 99.3% in patients without ECMO and 100% in those with ECMO. The predictive performance (receiver operating characteristic curve) of their combination toward poor outcome was 0.822 (patients without ECMO) and 0.681 (patients with ECMO) (p = 0.134).
Pending a prospective assessment on a larger cohort, in comatose patients after CA, the performance of prognostic factors seems comparable in patients with ECMO and those without ECMO. In particular, the combination of at least two poor outcome criteria appears valid across these two groups.
根据国际指南，心脏骤停（CA）后昏迷患者的神经预后使用多模态方法进行。然而，接受体外膜肺氧合（ECMO）的患者可能需要更长时间的药物镇静，并表现出生物标志物的改变，这可能对预后产生挑战。在这里，我们旨在评估常规使用的不良神经预后预测因素是否也能在 CA 后接受 ECMO 的患者中表现出可接受的性能。
这是一项对我们登记处的连续昏迷成年 CA 患者进行的观察性回顾性研究。排除 36 小时内死亡且未进行预后测试的患者。<80 岁的难治性 CA 患者，入院时无血流<5 分钟且低血流≤60 分钟时，开始进行静脉-动脉 ECMO。将接受 ECMO 与未接受 ECMO 的患者的神经预后测试表现（包括瞳孔反射、脑电图、体感诱发电位、神经元特异性烯醇化酶）与死亡率和不良功能结局（Cerebral Performance Categories [CPC] 评分 3-5）进行比较。
我们分析了 397 例未接受 ECMO 的患者和 50 例接受 ECMO 的患者。中位年龄为 65（四分位间距 54-74），69.8%的患者为男性。大多数患者有心脏病因（67.6%）；52%的患者有可除颤节律，有效的循环恢复时间中位数为 20（四分位间距 10-28）分钟。与未接受 ECMO 的患者相比，接受 ECMO 的患者的功能结局更差（74%的 CPC 评分为 3-5，59%，p=0.040），死亡率无显著升高（60%对 47%，p=0.080）。除神经元特异性烯醇化酶水平（接受 ECMO 的患者更高，p<0.001）外，两组中与不良结局（CPC 评分 3-5）相关的预后项目（瞳孔反射、脑电图背景和反应性、体感诱发电位和肌阵挛）的存在情况相似，同时存在至少两个此类项目的发生率也相似。在两组中，这些变量的每个特定变量对不良结局的特异性均为 92%至 100%，同时存在至少两个项目的组合特异性为 99.3%（未接受 ECMO 的患者）和 100%（接受 ECMO 的患者）。其对不良结局的组合预测性能（受试者工作特征曲线）在未接受 ECMO 的患者中为 0.822，在接受 ECMO 的患者中为 0.681（p=0.134）。
在对更大队列进行前瞻性评估之前，在 CA 后昏迷的患者中，预后因素的表现似乎在接受 ECMO 和未接受 ECMO 的患者中相似。特别是，至少两个不良结局标准的组合在这两组中似乎都是有效的。
