兽研背景与目的
心脏骤停治疗后超难治性癫痫持续状态（SUPER-CAT）研究评估了双重抗谷氨酸治疗（联合AMPA和NMDA受体阻断）在心脏骤停后超难治性癫痫持续状态（SRSE）且无不良预后指标患者中的疗效和安全性。
方法
多中心队列研究，纳入符合以下标准的缺氧后SRSE患者：双侧N20皮层反应保留、瞳孔反射完整、无全面性周期性放电。接受双重抗谷氨酸治疗的患者（DUAL组；氯胺酮1.5–7.5 mg/kg/h持续3天，加用吡仑帕奈6–12 mg/d持续5天后逐渐减量）与接受其他抗癫痫药物和麻醉药物组合治疗的患者（OTHERS组）进行比较。主要结局为SE的缓解。次要结局包括恢复遵从指令的觉醒以及不良多模态预后指标少于2项的患者在6个月时的良好神经功能结局（改良Rankin量表0–3分），以及整个队列的6个月死亡率。安全性分析包括院内死亡率和潜在治疗相关不良事件。接受单一抗谷氨酸药物治疗的患者单独描述。
结果
2012年5月至2025年5月期间，1,352例连续的心脏骤停后重症监护病房患者中有95例符合纳入标准（平均年龄66岁，四分位距55–73；女性41%；DUAL组=43例；OTHERS组=41例；单一抗谷氨酸组=11例）。与OTHERS组相比，DUAL组与更高的SE缓解率相关（84% vs 49%；OR 5.4；95% CI 1.9–14.9；p = 0.001），出院前觉醒的可能性更大（56% vs 32%；OR 2.7；95% CI 1.0–7.3；p = 0.053），6个月死亡率更低（52% vs 74%；OR 0.37；95% CI 0.15–0.97；p = 0.043）。DUAL组观察到6个月神经功能结局更优的趋势（39% vs 25%；OR 1.89；95% CI 0.64–5.56；p = 0.245）。与OTHERS组相比，DUAL组可逆性胆汁淤积性肝损伤发生率更高（77% vs 27%；OR 9.00；95% CI 3.34–24.19；p < 0.001）。
讨论
在无不良预后指标的缺氧后SRSE患者中，双重抗谷氨酸治疗在缓解SE方面高度有效，并与短期临床结局改善相关。
试验注册信息
ClinicalTrials.gov NCT05756621。
引言
癫痫持续状态（SE）是心脏骤停后常见且具有挑战性的并发症，影响高达20%–30%的昏迷幸存者，并强烈影响神经功能结局。1 缺氧后SE通常对标准抗癫痫治疗难治，且常进展为超难治性癫痫持续状态（SRSE），定义为在开始联合使用抗癫痫药物（ASMs）和麻醉药物后癫痫活动持续或复发超过24小时。2 缺氧后SRSE的最佳治疗方案、强度和治疗持续时间仍存在争议，特别是考虑到对治疗无效性的担忧。3-7 谷氨酸过度释放驱动的兴奋性毒性在缺氧后神经元损伤中起核心作用，N-甲基-D-天冬氨酸（NMDA）和α-氨基-3-羟基-5-甲基-4-异噁唑丙酸（AMPA）介导的受体过度激活促进持续的癫痫活动。8,9 这为靶向谷氨酸的治疗提供了理论依据。
当前欧洲复苏委员会（ERC）指南10支持使用神经系统检查、脑电图、正中神经体感诱发电位（SSEPs）、神经元特异性烯醇化酶（NSE）和神经影像学进行多模态预后评估，但对二线抗癫痫治疗提供的指导有限，且未专门涉及SRSE。因此，在此背景下积极抗癫痫治疗的作用仍存在争议，通常由个体化临床判断指导。尽管心肺复苏后脑电图癫痫持续状态治疗（TELSTAR）试验显示对全面性周期性放电（GPDs）的积极治疗无效，11 但其研究结果可能不适用于符合Salzburg明确SE标准的患者。12 先前来自独立研究组的观察性研究表明，积极的抗癫痫治疗可能在部分选定患者中产生有利结局。13-16
鉴于氯胺酮和吡仑帕奈具有互补的作用机制，二者提供了针对NMDA介导和AMPA介导兴奋性毒性的靶向策略。17 心脏骤停治疗后超难治性癫痫持续状态（SUPER-CAT）多中心队列研究评估了联合抗谷氨酸治疗在无不良预后指标的缺氧后SRSE患者中的疗效和安全性。
方法
研究设计与设置
SUPER-CAT研究是一项多中心观察性队列研究，在意大利8个重症监护病房（ICUs）和癫痫中心开展，包括2012年5月至2022年9月的回顾性数据和2022年10月至2025年5月的前瞻性数据。患者按标准临床实践接受治疗；未为研究目的分配任何治疗。临床数据和结局从现有记录中收集。数据通过研究电子数据采集云平台上的电子病例报告表（eCRFs）采集。18
标准方案批准、注册和患者知情同意
研究方案涵盖回顾性和前瞻性两个阶段，均经各参与中心的伦理委员会批准。该研究已在ClinicalTrials.gov注册（NCT05756621）。使用了一份统一的知情同意书。对于回顾性病例，在可行的情况下从在世且可联系到的患者处获取知情同意。对于前瞻性患者，尽可能前瞻性地获取知情同意；若患者死亡或始终未恢复决策能力，则豁免知情同意。已从患者或其法定授权代表处获取书面知情同意。在紧急情况下，依据国家法规采用延迟知情同意。对于已死亡或无法联系的患者，依据第101/2018号立法法令第21条和意大利数据保护法第110条批准豁免知情同意。
研究参与者
我们连续纳入2012年5月至2025年5月期间发生心脏骤停并入住ICU的成年患者；各参与中心的具体入组时间段见eTable 1。符合条件的患者为心脏骤停后昏迷并符合以下标准者：在最初24小时内接受了目标温度管理，在事件发生后36小时内启动了连续脑电图监测，符合Salzburg标准12确诊为明确的SRSE，并且双侧瞳孔对光反射保留以及正中神经SSEPs上双侧皮层N20反应保留。若在初始脑电图诊断时或SE首次复发时脑电图上存在GPDs，则排除患者。GPDs根据美国临床神经生理学会（ACNS）重症监护脑电图术语19定义，且为排除目的，特指频率<2.5 Hz且无演变者。其他排除标准为：SE在首轮麻醉药和ASMs治疗后缓解、年龄小于18岁或妊娠。
治疗
DUAL组患者接受双抗谷氨酸治疗，即同日同时启动连续静脉输注氯胺酮和口服吡仑帕奈。该治疗方案称为SUPER-CAT，最初由协调中心基于初步临床经验20开发，随后被其他参与中心采用。氯胺酮经静脉给药3天，剂量根据当地临床实践确定（诱导剂量：1.5–3 mg/kg，随后以1.5–7.5 mg/kg/h持续输注）。氯胺酮剂量根据连续脑电图监测进行滴定，以达到无脑电图癫痫活动的持续α-β背景。根据需要联合给予静脉咪达唑仑（1–15 mg/h）。吡仑帕奈经鼻胃管以压碎片剂给药5天（剂量范围6–12 mg）。随后吡仑帕奈每2天递减2 mg，若未观察到脑电图或临床癫痫复发则停药。
OTHERS组患者根据标准临床实践接受常规ASMs和麻醉治疗，类型和剂量不限，但不包括氯胺酮和吡仑帕奈。接受单药抗谷氨酸治疗的额外队列——单用氯胺酮（SINGLE-K）或单用吡仑帕奈（SINGLE-P）——分别进行分析。
研究结局
主要结局为SE缓解且无复发、无需额外ASMs或麻醉药，评估贯穿整个ICU住院期间。SE缓解后连续脑电图监测的持续时间未预先设定，各中心有所不同。ICU住院期间获得的所有可用脑电图描记均用于判定缓解与复发。
次要结局包括：SE缓解的累积概率；早期和长期良好神经功能结局，定义为出院前觉醒至能够遵嘱动作以及6个月时改良Rankin量表（mRS）评分21为0–3分，二者均在根据ERC指南10具有少于2项不良预后指标的患者中评估；以及整个队列的6个月死亡率。意识恢复至能够遵嘱动作通过查阅病历进行评估。mRS通过电话访谈或面访对回顾性和前瞻性患者进行评估。对于回顾性患者，评估指事件后6个月时的神经功能障碍。安全性结局包括ICU内死亡率、γ-谷氨酰转移酶（GGT）升高超过正常上限（UNL）3倍、肌酸激酶（CK）升高超过UNL 10倍、三度房室传导阻滞、心脏骤停复发以及任何可能治疗相关的不良事件。
样本量估算
初步数据提示，接受常规治疗的患者中约20%实现SRSE缓解，而接受双抗谷氨酸治疗者中约60%实现缓解。基于这些估计，计算样本量为80例患者（每组40例），以提供80%把握度检测SRSE缓解率30%的保守绝对差异，双侧α为0.05。样本量估算用于评估总体队列的统计把握度并指导前瞻性入组。
统计分析
采用描述性统计方法总结基线特征，分类变量以频数和百分比表示，连续变量以中位数（四分位距，IQR）表示。组间比较（DUAL 组 vs OTHERS 组）中，分类变量采用 χ2 检验或 Fisher 精确检验，连续变量采用 t 检验或 Wilcoxon-Mann-Whitney 检验。主要疗效、次要及安全性结局在治疗组间采用 χ2 检验进行比较。鉴于样本量有限，治疗-结局关联采用以治疗组为唯一自变量的 logistic 回归模型进行评估。结果以比值比（OR）及95% CI报告。超级难治性癫痫持续状态（SE）缓解和觉醒的累积概率采用累积发生率函数分析，将死亡作为竞争事件，组间比较采用 Gray 检验。6个月生存率采用 Kaplan-Meier 曲线分析，组间比较采用 log-rank 检验。SINGLE-K 组和 SINGLE-P 组的各结局均进行描述性分析。统计学显著性设定为 p < 0.05（双侧）。缺失数据未进行插补，包括因失访导致的缺失数据。分析采用 SAS 软件（版本9.4，SAS Institute，Cary，NC）完成。
数据可用性
合格的研究者可申请获取去标识化的个体患者水平数据及相关试验文件，包括原始数据集和分析就绪数据集、研究方案、注释电子病例报告表（eCRF）、统计分析计划及数据集规范。申请将由 SUPER-CAT 指导委员会评估。经批准后，在签署正式数据共享协议并符合适用伦理和监管要求的前提下授予数据访问权限。
结果
研究人群与神经预后指标
研究期间，心脏骤停后入住 ICU 的1,352例患者中，95例（7.0%）符合纳入标准。其中，84例患者纳入2个主要比较组（DUAL 组和 OTHERS 组），另有11例接受单药抗谷氨酸药物治疗的患者单独分析（eFigure 1和2）。DUAL 组与 OTHERS 组的基线人口学和临床特征总体均衡（表1）。DUAL 组中位年龄为64岁（IQR 53–71），OTHERS 组为69岁（IQR 60–73）。男性分别占48.8%和65.8%。院外心脏骤停发生率分别为62.8% vs 75.6%，可除颤心律分别为51.2% vs 63.4%。中位低流量时间（21 vs 22分钟）、临床肌阵挛（27.9% vs 21.9%）和 NSE 水平（45 vs 31 μg/L）相当，角膜反射保留的发生率（23% vs 23%）及影像学上弥漫性缺氧性损伤的发生率（13% vs 9%）亦相当。
Table 1. Clinical Characteristics and Neuroprognostic Indicators
| DUAL (n = 43) | OTHERS (n = 41) | p Value | |
| Age, y, median (IQR) | 64 (53–71) | 69 (60–73) | 0.154 |
| Sex, male, n (%) | 21 (48.8) | 27 (65.8) | 0.129 |
| OHCA, n (%) | 27 (62.8) | 31 (75.6) | 0.242 |
| Shockable rhythm, n (%) | 22 (51.2) | 26 (63.4) | 0.278 |
| No-flow time, min, median (IQR) | 0 (0) | 0 (0–1) | 1.000 |
| Low-flow time, min, median (IQR) | 21 (10–35) | 22 (12–30) | 0.590 |
| Corneal reflex present, n (%) | 23 (53.4) | 23 (56.1) | 1.000 |
| NSE (highest value 24–72 h), μg/L, median (IQR) | 45 (27–65) | 31 (20–63) | 0.830 |
| Diffuse anoxic injury at CT or MRI, n (%) | 13 (30.2) | 9 (28.1) | 0.804 |
| Clinical myoclonus, n (%) | 12 (27.9) | 9 (21.9) | 0.618 |
治疗暴露
治疗特征总结于表2。DUAL 组所有患者均接受连续静脉注射 ketamine 3天（中位剂量5 mg/kg/h，IQR 4–6），滴定至达到目标 ketamine 脑电图模式（图1B）。30例患者（69.8%；中位剂量2 mg/h，IQR 0–5）联合静脉注射 midazolam。口服 perampanel（中位剂量12 mg，IQR 8–12）同时给药5天，随后逐渐减量。两组患者接受 valproate（88% vs 85%）、levetiracetam（67% vs 70%）和 propofol（97% vs 98%）的比例相当。ICU 住院期间使用的其他抗癫痫发作药物（ASM）见 eTable 2。与 OTHERS 组相比，DUAL 组患者接受额外 ASM（39% vs 53%）、thiopental（14% vs 27%）和 midazolam（63% vs 83%）的可能性更低。DUAL 组的中位 ASM 数量更高（3 [IQR 2–4] vs 2 [2–3]），中位麻醉周期数亦更高（3 [2–3] vs 2 [2–3]）。84例患者中有75例（89.3%）使用 ketamine 以外的麻醉药物（包括在 SE 超级难治期之前给予的药物）滴定至脑电图爆发抑制。DUAL 组 ICU 住院时间更长（33 vs 15天）。
Table 2. ASMs and Anesthetics Received During the ICU Stay
| DUAL (n = 43) | OTHERS (n = 41) | |
| No. of ASMs, median (IQR) | 3 (2–4) | 2 (2–3) |
| No. of anesthetic cycles, median (IQR) | 3 (2–3) | 2 (2–3) |
| Perampanel dose, mg, median (IQR) | 12 (8–12) | 0 (0) |
| Ketamine (EEG-targeted dose), mg/kg/h, median (IQR) | 5 (4–6) | 0 (0) |
| Valproate, n (%) | 38 (88) | 35 (85) |
| Levetiracetam, n (%) | 29 (67) | 29 (70) |
| Other ASMs, n (%) a | 17 (39) | 22 (53) |
| Propofol, n (%) | 42 (97) | 40 (98) |
| Midazolam, n (%) | 27 (63) | 34 (83) |
| Thiopental, n (%) | 6 (14) | 11 (27) |
| Duration of ICU stay, d, median (IQR) | 33 (26–51) | 15 (9–24) |
[图：一例代表性缺氧后超级难治性癫痫持续状态患者的脑电图病程 脑电图时段显示于（]
主要结局
在ICU住院期间，DUAL组43例患者中有36例（83.7%）SRSE得到缓解，OTHERS组41例中有20例（48.8%）得到缓解（OR 5.40；95% CI 1.95–14.90；p = 0.001；表3）。图1展示了一例接受DUAL治疗的缺氧后患者具有代表性的EEG病程。图2显示了30天内SE缓解的累积概率，Gray检验表明DUAL组的缓解概率显著更高（p = 0.006）。
Table 3. Primary and Secondary Outcomes
| DUAL (n = 43) | OTHERS (n = 41) | OR (95% CI) | p Value | |
| Primary outcome | ||||
| Resolution of status epilepticus during the ICU stay, n/N (%) | 36/43 (83.7) | 20/41 (48.8) | 5.40 (1.95–14.90) | 0.001 |
| Secondary outcomes | ||||
| Early good neurologic outcome a (awakening to follow commands during the ICU stay), n/N (%) | 18/32 (56.2) | 11/34 (32.3) | 2.68 (0.99–7.32) | 0.053 |
| Long-term good neurologic outcome a (mRS 0–3 at 6 mo), n/N (%) | 12/31 (38.7) | 8/32 (25.0) | 1.89 (0.64–5.56) | 0.245 |
| Mortality at 6 mo, n/N (%) | 22/42 (52.4) | 29/39 (74.4) | 0.37 (0.15–0.97) | 0.043 |
[图：癫痫持续状态缓解的累积发生率 累积发生率函数显示缓解概率]
次要结局
在不利预后指标少于2项的患者中，早期良好神经功能结局（定义为ICU住院期间觉醒至遵从指令）在DUAL组32例中有18例（56.2%）发生，在OTHERS组34例中有11例（32.3%）发生（OR 2.68；95% CI 0.99–7.32；p = 0.053）。
3例患者失访（DUAL组n = 1；OTHERS组n = 2）。DUAL组31例中有12例（38.7%）观察到长期良好神经功能结局（6个月时mRS 0–3），OTHERS组32例中有8例（25.0%）观察到（OR 1.89；95% CI 0.64–5.56；p = 0.245）。在整个队列中，DUAL组的6个月死亡率（22/42，52.4%）低于OTHERS组（29/39，74.4%；OR 0.37；95% CI 0.15–0.97；p = 0.043；表3）。觉醒和6个月生存的时间-事件分析见图3，显示DUAL组觉醒的累积发生率更高，且DUAL组6个月生存改善（log-rank p = 0.003）。在住院期间死亡的患者中，19例（51%）在撤除生命维持治疗（WLST）后死亡，17例（46%）死于进行性多器官衰竭，1例死亡归因于脑死亡。
[图：早期神经功能恢复和长期生存 （A）45天内觉醒至遵从指令的累积发生率]
安全性结局
安全性数据总结于表4。ICU住院期间死亡在DUAL组有12例（27.9%），在OTHERS组有25例（60.9%）。接受DUAL治疗的患者中观察到GGT升高超过UNL 3倍者43例中有33例（76.7%），OTHERS组41例中有11例（26.8%）。CK升高（>10× UNL）分别发生于15例（34.8%）和11例（32.3%）患者。DUAL组有1例患者发生三度房室传导阻滞；心脏骤停复发在DUAL组有2例（4.6%），OTHERS组有1例（2.4%）。未观察到其他可能与神经治疗相关的不良事件。
Table 4. Safety Outcomes
| DUAL (n = 43) | OTHERS (n = 41) | OR (95% CI) | p Value | |
| Death during the ICU stay, n (%) | 12 (27.9) | 25 (60.9) | 0.24 (0.09–0.61) | 0.003 |
| GGT elevation (>3 times the UNL), n (%) | 33 (76.7) | 11 (26.8) | 9.00 (3.34–24.19) | <0.001 |
| Highest GGT value during the ICU stay, U/L, median (IQR) | 624 (318–1,084) | 557 (247–910) | NP | NP |
| CPK elevation (>10 times the UNL), n (%) | 15 (34.8) | 11 (32.3) | 1.12 (0.43–2.90) | 0.815 |
| Third-degree atrioventricular block, n (%) | 1 (2.3) | 0 (0) | NP | NP |
| Recurrence of cardiac arrest, n (%) | 2 (4.6) | 1 (2.4) | NP | NP |
单用抗谷氨酸药物
还收集了仅接受ketamine治疗（n = 6；未联用perampanel）和仅接受perampanel治疗（n = 5；未联用ketamine）的小队列患者结果，见eTable 3。鉴于样本量有限，结局以描述性方式报告；与DUAL组相比，任一种单用抗谷氨酸药物的SRSE缓解和GGT升高似乎均较少见。
讨论
在这项多中心队列研究中，该研究包括回顾性和前瞻性两个阶段，与常规治疗相比，在选定的心脏骤停后患者中，ketamine和perampanel双联抗谷氨酸治疗与更高的SRSE缓解率相关，提示该联合方案可能代表一种针对病因的潜在治疗策略。主要发现是双联抗谷氨酸治疗的癫痫发作缓解率为84%，而常规联合治疗为49%，这代表了实质性和具有临床意义的差异，支持未来开展前瞻性随机研究以证实这些结果的理由。
双联抗谷氨酸治疗的优越疗效可能反映了针对缺氧后脑损伤基本病理生理的靶向干预。全脑缺血触发大量谷氨酸释放，通过NMDA和AMPA受体过度激活等机制导致兴奋性毒性细胞死亡。8,22 通过同时靶向NMDA和AMPA受体，ketamine和perampanel可能提供互补机制来中断这一病理级联反应，可能优于靶向钠通道或γ-氨基丁酸受体的常规方法。
氯胺酮的神经保护特性不仅限于NMDA受体拮抗作用，还包括抗炎效应和线粒体功能的保护。23 吡仑帕奈作为目前唯一可用于临床的选择性非竞争性AMPA受体拮抗剂，特异性靶向负责大脑中大多数快速兴奋性神经传递的受体系统。24
值得注意的是，两种药物均以非竞争性方式阻断各自对应的受体，从而对谷氨酸能系统产生不可克服的抑制，且该抑制不依赖于缺氧后事件期间释放的谷氨酸水平。尽管氯胺酮和吡仑帕奈单独用于癫痫持续状态时均已显示出疗效，25,26 但联合给药可能在复苏后脆弱期增强癫痫控制的同时提供额外的神经保护。27,28
积极治疗的患者选择时机和标准仍然至关重要。在本研究中，纳入标准仅限于符合最低预后标准组合的患者，即瞳孔反射保留且双侧N20反应存在，而伴有全面性周期性放电的患者被排除。这种严格筛选可能有助于观察到的治疗成功率，并与倡导多模式评估的ERC指南相一致。10 对于以神经功能恢复为重点的次要结局，患者根据ERC推荐进一步分层，排除了具有2项或以上不良预后指标的患者，因为其获得良好结局的可能性极小。
双药治疗组56%的神经功能觉醒率（定义为能够遵嘱执行指令）提示有效癫痫控制与具有临床意义的神经功能恢复之间可能存在关联。这一观察尤为重要，因为心脏骤停后患者入住ICU期间早期或晚期觉醒与良好的长期功能结局相关。29 然而需要指出的是，长期结局不仅受神经功能恢复的影响，还受全身合并症、非神经系统死亡原因以及后期可能发现的认知或行为缺陷的影响，这可能解释了神经功能获益随时间推移而出现的明显减弱。需要更大规模的前瞻性研究来证实双重抗谷氨酸治疗对一线治疗难治性癫痫持续状态的缺氧后患者的临床优势。
这些结果超越了我们先前的研究和初步单中心经验，13,17,20 为双重抗谷氨酸方案的可行性提供了多中心验证。该患者人群中观察到的84%癫痫发作终止率超过了近期一项大型多中心队列中常规治疗的报告率，14 后者终止率为44%，与本研究中对照组观察到的49%密切吻合。我们的发现也与TELSTAR试验的事后分析一致，11 该分析显示抗癫痫治疗对心脏骤停后伴有非周期性癫痫样模式的患者具有潜在获益。
如SUPER-CAT方案所述，双重抗谷氨酸治疗的安全性特征似乎可接受，但值得进一步研究。主要且常见的不良事件（>75%的病例）为胆汁淤积性肝损伤的生化证据，在吡仑帕奈减量后可逆。这一观察与我们先前对8例接受吡仑帕奈口服负荷剂量患者的初步报告一致，其中62.5%的病例出现肝胆指标改变。17 尽管不立即危及生命，但鉴于合并用药负担重以及心脏骤停后综合征中缺氧性肝损伤的作用，该不良事件值得在未來临床研究中进一步探讨，并在个体患者层面进行仔细的风险-获益评估。
我们研究的主要优势在于其多学科协作模式，神经科医生与重症监护医生紧密合作，以及在以病情严重和高死亡率为特征的患者群体中应用了创新的治疗方案。此外，多中心设计以及采用与ACNS术语和ERC指南相一致的标准化定义和预后评估方法，增强了我们研究结果的潜在普适性。总体而言，该方法为改善患者照护提供了初步证据，无论是对于癫痫持续状态的终止还是整体临床结局。若干局限性应予说明。第一，观察性研究设计，包括回顾性和前瞻性阶段，引入了潜在的选择偏倚，尽管我们努力纳入符合入选标准的连续患者并使研究组间基线特征保持均衡。治疗分配并非随机化；然而，缺乏关于难治性癫痫持续状态最佳管理的既定证据，可能最大限度地减少了系统性治疗偏好。第二，尽管样本量对于主要结局是充足的，但检测长期临床结局和罕见不良事件差异的效能可能有限。第三，研究在具有心脏骤停后照护和癫痫持续状态管理专长的专科中心进行，这可能限制普适性。持续脑电图监测和细致多模式预后评估的要求可能并非普遍可及。第四，我们采用保守方法，基于TELSTAR试验的结果决定排除具有全面性周期性放电的患者，该试验显示心脏骤停后全面性周期性放电的治疗无效。因此，无法评估DUAL对全面性周期性放电的效果。第五，不能完全排除维持生命治疗的撤除决策对神经学结局的潜在影响以及生存者偏倚的风险。然而，维持生命治疗的撤除决策系根据机构方案和当代神经预后指南做出，且未在心脏骤停后72小时之前启动，从而将过早撤除治疗的风险降至最低。第六，本研究专门针对缺氧后癫痫持续状态，其在病理生理学和预后方面与其他病因所致的难治性癫痫持续状态存在显著差异；需要进一步研究以评估SUPER-CAT方案在其他患者群体中的疗效和安全性。第七，未系统测量吡仑帕奈血浆浓度。鉴于在无细胞色素P450 3A4诱导剂的情况下其半衰期较长（约105小时），第5天的浓度可能接近6 mg/d方案的浓度，尽管尚未达到稳态。未来研究可考虑更高的负荷剂量，如 suspected inflammatory etiology的癫痫持续状态中所报道的30，以及药代动力学测量。
在经选择的的心脏骤停后难治性癫痫持续状态患者中，与常规治疗相比，氯胺酮和吡仑帕奈双联抗谷氨酸治疗与更高的癫痫发作终止率相关，并显示出改善早期神经功能恢复的强烈趋势，而其对长期结局的益处仍不确定。这些发现支持开展一项前瞻性随机对照试验的理论依据，比较双联抗谷氨酸治疗与当前标准治疗在缺氧后难治性癫痫持续状态中的效果，最好作为多国协作努力进行。
查看英文原文▾
Background and Objectives
The SUPER-refractory status epilepticus (SE) after Cardiac Arrest Treatment (SUPER-CAT) study evaluated the efficacy and safety of dual antiglutamate therapy (combined AMPA and NMDA receptor blockade) in patients with super-refractory SE (SRSE) after cardiac arrest and without unfavorable prognostic indicators.
Methods
Multicenter cohort study including patients with postanoxic SRSE meeting the following criteria: bilaterally preserved N20 cortical responses, intact pupillary reflexes, and absence of generalized periodic discharges. Patients treated with dual antiglutamate therapy (DUAL; ketamine 1.5–7.5 mg/kg/h for 3 days plus perampanel 6–12 mg/d for 5 days with tapering) were compared with patients receiving other combinations of antiseizure and anesthetic agents (OTHERS). The primary outcome was resolution of SE. Secondary outcomes included awakening to follow commands and good neurologic outcome at 6 months (modified Rankin Scale 0–3) in patients with fewer than 2 unfavorable multimodal prognostic indicators and 6-month mortality for the entire cohort. Safety analysis included in-hospital mortality and potentially treatment-related adverse events. Patients treated with single antiglutamate agents were described separately.
Results
Between May 2012 and May 2025, 95 of 1,352 consecutive post–cardiac arrest intensive care unit patients met inclusion criteria (mean age 66, interquartile range 55–73; 41% female; DUAL = 43; OTHERS = 41; single antiglutamate = 11). Compared with OTHERS, DUAL was associated with a higher rate of SE resolution (84% vs 49%; OR 5.4; 95% CI 1.9–14.9; p = 0.001), a greater likelihood of awakening before discharge (56% vs 32%; OR 2.7; 95% CI 1.0–7.3; p = 0.053) and lower 6-month mortality (52% vs 74%; OR 0.37; 95% CI 0.15–0.97; p = 0.043). A trend toward better 6-month neurologic outcome was observed in the DUAL group (39% vs 25%; OR 1.89; 95% CI 0.64–5.56; p = 0.245). Reversible cholestatic liver injury occurred more frequently in DUAL group, compared with OTHERS (77% vs 27%; OR 9.00; 95% CI 3.34–24.19; p < 0.001).
Discussion
In patients with postanoxic SRSE without unfavorable prognostic indicators, dual antiglutamate therapy was highly effective in resolving SE and was associated with improved short-term clinical outcomes.
Trial Registration Information
ClinicalTrials.gov NCT05756621.
Introduction
Status epilepticus (SE) is a frequent and challenging complication after cardiac arrest, affecting up to 20%–30% of comatose survivors and strongly influencing neurologic outcomes.1 Postanoxic SE is often refractory to standard antiseizure therapy and may frequently progress to super-refractory SE (SRSE), defined as seizure activity persisting or recurring >24 hours after initiation of a combination of antiseizure medications (ASMs) and anesthetics.2 The optimal therapeutic regimen, intensity, and duration for postanoxic SRSE remains controversial, particularly given concerns about treatment futility.3-7 Excitotoxicity driven by excessive glutamate release plays a central role in postanoxic neuronal injury, with overactivation of N-methyl-D-aspartate (NMDA) and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)-mediated receptors promoting ongoing seizure activity.8,9 This provides a rationale for glutamate-targeted therapies.
Current European Resuscitation Council (ERC) guidelines10 support multimodal prognostication using neurologic examination, EEG, median nerve somatosensory evoked potentials (SSEPs), neuron-specific enolase (NSE), and neuroimaging but offer limited guidance on second-line antiseizure treatment and do not specifically address SRSE. Consequently, the role of aggressive antiseizure therapy in this context remains debated and is generally guided by individualized clinical judgment. Although the Treatment of ELectroencephalographic STatus Epilepticus After Cardiopulmonary Resuscitation (TELSTAR) trial showed futility of aggressive treatment of generalized periodic discharges (GPDs),11 its findings may not apply to patients meeting Salzburg criteria for definite SE.12 Previous observational studies from independent groups indicated that active antiseizure treatment may yield favorable outcomes in selected patients.13-16
Given their complementary mechanisms of action, ketamine and perampanel offer a targeted strategy against NMDA-mediated and AMPA-mediated excitotoxicity.17 The SUPER-refractory SE after Cardiac Arrest Treatment (SUPER-CAT) multicenter cohort study evaluated the efficacy and safety of combined antiglutamate therapy for postanoxic SRSE in patients without unfavorable prognostic indicators.
Methods
Study Design and Setting
The SUPER-CAT study was a multicenter, observational cohort study, conducted across 8 intensive care units (ICUs) and epilepsy centers in Italy, including retrospective data from May 2012 to September 2022 and prospective data from October 2022 to May 2025. Patients were treated according to standard clinical practice; no treatments were assigned for research purposes. Clinical data and outcomes were collected from existing records. Data were captured using electronic case report forms (eCRFs) on the Research Electronic Data Capture Cloud platform.18
Standard Protocol Approvals, Registrations, and Patient Consents
The study protocol, covering both the retrospective and prospective phases, was approved by the Ethics Committee at each participating center. The study was registered at ClinicalTrials.gov (NCT05756621). A single informed consent form was used. For retrospective cases, informed consent was obtained when feasible from living, contactable patients. For prospective patients, informed consent was obtained prospectively whenever possible; if patients died or never regained decision-making capacity, informed consent was waived. Written informed consent was obtained from patients or their legally authorized representatives. In emergency cases, deferred consent was applied in accordance with national regulations. For deceased or noncontactable patients, consent waiver was granted in accordance with Article 21 of Legislative Decree 101/2018 and Article 110 of the Italian Data Protection Code.
Study Participants
We included consecutive adult patients who experienced cardiac arrest and were admitted to the ICU between May 2012 and May 2025; detailed enrolment periods for each participating center are presented in eTable 1. Eligible patients were comatose after cardiac arrest and met the following criteria: They underwent targeted temperature management during the first 24 hours, had continuous EEG monitoring initiated within 36 hours of the event, fulfilled Salzburg criteria12 for definite SRSE, and had preserved bilateral pupillary reflexes as well as bilateral cortical N20 responses on median nerve SSEPs. Patients were excluded if GPDs were present on EEG at the time of initial electrographic diagnosis or the first recurrence of SE. GPDs were defined according to the American Clinical Neurophysiology Society (ACNS) critical care EEG terminology19 and, for the purpose of exclusion, were specifically <2.5 Hz and nonevolving. Additional exclusion criteria were resolution of SE after the first cycle of anesthetics and ASMs, age younger than 18 years, or pregnancy.
Treatments
Patients in the DUAL group received dual antiglutamate therapy consisting of the simultaneous initiation, on the same day, of continuous intravenous ketamine and oral perampanel. This treatment protocol, termed SUPER-CAT, was initially developed at the coordinating center based on preliminary clinical experience20 and was subsequently adopted by the other participating centers. Ketamine was administered intravenously for 3 days, with dosing determined according to local clinical practice (induction dose: 1.5–3 mg/kg, followed by continuous infusion at 1.5–7.5 mg/kg/h). Ketamine dosing was titrated according to continuous EEG monitoring to achieve a continuous alpha-beta background without electrographic seizure activity. Intravenous midazolam (1–15 mg/h) was coadministered as needed. Perampanel was administered for 5 days using a nasogastric tube as crushed tablets (dose range, 6–12 mg). Perampanel was subsequently tapered by 2 mg every 2 days and discontinued if no electrographic or clinical seizure recurrence was observed.
Patients in the OTHERS group received conventional ASMs and anesthetic therapies according to standard clinical practice, of any type and dose, except ketamine and perampanel. Additional cohorts receiving single-agent antiglutamatergic therapy, ketamine alone (SINGLE-K) or perampanel alone (SINGLE-P), were analyzed separately.
Study Outcomes
The primary outcome was resolution of SE without recurrence and without the need for additional ASMs or anesthetics, assessed throughout the entire ICU stay. The duration of continuous EEG monitoring after SE resolution was not predefined and varied across centers. All available EEG tracings obtained during the ICU stay were used to adjudicate resolution vs recurrence.
Secondary outcomes included the cumulative probability of SE resolution; early and long-term good neurologic outcomes, defined as awakening to the ability to follow commands before hospital discharge and a modified Rankin Scale (mRS) score21 of 0–3 at 6 months, both assessed in patients with fewer than 2 unfavorable prognostic indicators according to ERC guidelines10; and 6-month mortality for the entire cohort. Recovery of consciousness up to command following was assessed through review of medical records. The mRS was assessed by either phone interview or in-person visit for both retrospective and prospective patients. For retrospective patients, the assessment referred to neurologic disability at 6 months after the event. Safety outcomes included mortality in the ICU, elevation of gamma-glutamyl transferase (GGT) greater than 3 times the upper normal limit (UNL), creatine kinase (CK) greater than 10 times the UNL, third-degree atrioventricular block, recurrence of cardiac arrest, and any potentially treatment-related adverse events.
Sample Size Estimate
Preliminary data suggested that SRSE resolution occurred in approximately 20% of patients treated with conventional therapies and in approximately 60% of those receiving dual antiglutamate therapy. Based on these estimates, a sample size of 80 patients (40 per group) was calculated to provide 80% power to detect a conservative 30% absolute difference in SRSE resolution, with a 2-tailed α of 0.05. The sample size estimate served to assess statistical power for the overall cohort and guide the prospective enrollment.
Statistical Analysis
Descriptive statistics were used to summarize baseline characteristics, with frequencies and percentages for categorical variables and medians (interquartile range, IQR) for continuous variables. Between-group comparisons (DUAL vs OTHERS) used χ2 or Fisher exact tests for categorical variables and t tests or Wilcoxon-Mann-Whitney tests for continuous variables. The primary efficacy, secondary, and safety outcomes were compared between treatment groups using χ2 tests. Treatment-outcome associations were evaluated using logistic regression models with treatment group as the only independent variable, given the limited sample size. The results are reported as odds ratios (ORs) with 95% CIs. The cumulative probability of SE resolution and awakening was analyzed using cumulative incidence functions accounting for death as competing event, with Gray test for between-group comparisons. Six-month survival was analyzed using Kaplan-Meier curves and compared between groups using the log-rank test. Descriptive analyses for the SINGLE-K and SINGLE-P groups were reported for each outcome. Statistical significance was set at p < 0.05 (2-tailed). Missing data were not imputed, including data missing due to loss to follow-up. Analyses were performed using SAS software (version 9.4, SAS Institute, Cary, NC).
Data Availability
Qualified investigators may request access to anonymized individual patient-level data and relevant trial documentation, including raw and analysis-ready data sets, the study protocol, annotated eCRFs, the statistical analysis plan, and dataset specifications. Requests will be evaluated by the SUPER-CAT Steering Committee. Upon approval, access to the data will be granted after execution of a formal data sharing agreement, in accordance with applicable ethical and regulatory requirements.
Results
Study Population and Neuroprognostic Indicators
Among 1,352 patients admitted to the ICU after cardiac arrest during the study period, 95 (7.0%) met inclusion criteria. Of these, 84 patients were included in the 2 primary comparison groups (DUAL and OTHERS), while 11 patients treated with single antiglutamate agents were analyzed separately (eFigures 1 and 2). Baseline demographic and clinical characteristics were generally balanced between DUAL and OTHERS (Table 1). The median age was 64 years (IQR 53–71) in the DUAL group and 69 years (IQR 60–73) in the OTHERS group. Men accounted for 48.8% and 65.8% of patients, respectively. Out-of-hospital cardiac arrest occurred in 62.8% vs 75.6% and shockable rhythms in 51.2% vs 63.4%. The median low-flow times (21 vs 22 minutes), clinical myoclonus (27.9% vs 21.9%), and NSE level (45 vs 31 μg/L) were comparable, as well as the prevalence of preserved corneal reflexes (23% vs 23%) and diffuse anoxic injury on imaging (13% vs 9%).
Table 1. Clinical Characteristics and Neuroprognostic Indicators
| DUAL (n = 43) | OTHERS (n = 41) | p Value | |
| Age, y, median (IQR) | 64 (53–71) | 69 (60–73) | 0.154 |
| Sex, male, n (%) | 21 (48.8) | 27 (65.8) | 0.129 |
| OHCA, n (%) | 27 (62.8) | 31 (75.6) | 0.242 |
| Shockable rhythm, n (%) | 22 (51.2) | 26 (63.4) | 0.278 |
| No-flow time, min, median (IQR) | 0 (0) | 0 (0–1) | 1.000 |
| Low-flow time, min, median (IQR) | 21 (10–35) | 22 (12–30) | 0.590 |
| Corneal reflex present, n (%) | 23 (53.4) | 23 (56.1) | 1.000 |
| NSE (highest value 24–72 h), μg/L, median (IQR) | 45 (27–65) | 31 (20–63) | 0.830 |
| Diffuse anoxic injury at CT or MRI, n (%) | 13 (30.2) | 9 (28.1) | 0.804 |
| Clinical myoclonus, n (%) | 12 (27.9) | 9 (21.9) | 0.618 |
Therapeutic Exposure
Treatment characteristics are summarized in Table 2. All patients in the DUAL group received continuous intravenous ketamine for 3 days (median dose 5 mg/kg/h, IQR 4–6), titrated to achieve the target ketamine EEG pattern (Figure 1B). Intravenous midazolam was coadministered in 30 patients (69.8%; median dose 2 mg/h, IQR 0–5). Oral perampanel (median dose 12 mg, IQR 8–12) was administered concomitantly for 5 days and subsequently tapered. Patients in both groups received valproate (88% vs 85%), levetiracetam (67% vs 70%), and propofol (97% vs 98%) in comparable proportions. Other ASMs used during the ICU stay are listed in eTable 2. Patients in the DUAL group were less likely to receive additional ASMs (39% vs 53%), thiopental (14% vs 27%), and midazolam (63% vs 83%) than those in the OTHERS group. The median number of ASMs was higher in the DUAL group (3 [IQR 2–4] vs 2 [2–3]), as well as the median number of anesthetic cycles (3 [2–3] vs 2 [2–3]). In 75 of 84 patients (89.3%), anesthetic agents other than ketamine, including those administered before the super-refractory phase of SE, were titrated to achieve EEG burst suppression. Duration of ICU stay was longer in the DUAL group (33 vs 15 days).
Table 2. ASMs and Anesthetics Received During the ICU Stay
| DUAL (n = 43) | OTHERS (n = 41) | |
| No. of ASMs, median (IQR) | 3 (2–4) | 2 (2–3) |
| No. of anesthetic cycles, median (IQR) | 3 (2–3) | 2 (2–3) |
| Perampanel dose, mg, median (IQR) | 12 (8–12) | 0 (0) |
| Ketamine (EEG-targeted dose), mg/kg/h, median (IQR) | 5 (4–6) | 0 (0) |
| Valproate, n (%) | 38 (88) | 35 (85) |
| Levetiracetam, n (%) | 29 (67) | 29 (70) |
| Other ASMs, n (%) a | 17 (39) | 22 (53) |
| Propofol, n (%) | 42 (97) | 40 (98) |
| Midazolam, n (%) | 27 (63) | 34 (83) |
| Thiopental, n (%) | 6 (14) | 11 (27) |
| Duration of ICU stay, d, median (IQR) | 33 (26–51) | 15 (9–24) |
[图：EEG Course in a Representative Postanoxic Patient With Super-Refractory Status Epilepticus EEG epochs are shown before (]
Primary Outcome
Resolution of SRSE during the ICU stay occurred in 36 of 43 patients (83.7%) in the DUAL group and 20 of 41 patients (48.8%) in the OTHERS group (OR 5.40; 95% CI 1.95–14.90; p = 0.001; Table 3). Representative EEG course in a postanoxic patient treated with DUAL therapy is illustrated in Figure 1. The cumulative probability of SE resolution over 30 days, with Gray test indicating a significantly higher probability in the DUAL group (p = 0.006), is shown in Figure 2.
Table 3. Primary and Secondary Outcomes
| DUAL (n = 43) | OTHERS (n = 41) | OR (95% CI) | p Value | |
| Primary outcome | ||||
| Resolution of status epilepticus during the ICU stay, n/N (%) | 36/43 (83.7) | 20/41 (48.8) | 5.40 (1.95–14.90) | 0.001 |
| Secondary outcomes | ||||
| Early good neurologic outcome a (awakening to follow commands during the ICU stay), n/N (%) | 18/32 (56.2) | 11/34 (32.3) | 2.68 (0.99–7.32) | 0.053 |
| Long-term good neurologic outcome a (mRS 0–3 at 6 mo), n/N (%) | 12/31 (38.7) | 8/32 (25.0) | 1.89 (0.64–5.56) | 0.245 |
| Mortality at 6 mo, n/N (%) | 22/42 (52.4) | 29/39 (74.4) | 0.37 (0.15–0.97) | 0.043 |
[图：Cumulative Incidence of Status Epilepticus Resolution Cumulative incidence functions show the probability of resolution ]
Secondary Outcomes
Among patients with fewer than 2 unfavorable prognostic indicators, early good neurologic outcome, defined as awakening to follow commands during the ICU stay, occurred in 18 of 32 patients (56.2%) in the DUAL group and 11 of 34 patients (32.3%) in the OTHERS group (OR 2.68; 95% CI 0.99–7.32; p = 0.053).
Three patients were lost to follow-up (DUAL, n = 1; OTHERS, n = 2). Long-term favorable neurologic outcome (mRS 0–3 at 6 months) was observed in 12 of 31 patients (38.7%) in the DUAL group and 8 of 32 patients (25.0%) in the OTHERS group (OR 1.89; 95% CI 0.64–5.56; p = 0.245). In the entire cohort, 6-month mortality was lower in the DUAL group (22/42, 52.4%) compared with the OTHERS group (29/39, 74.4%; OR 0.37; 95% CI 0.15–0.97; p = 0.043; Table 3). Time-to-event analyses for awakening and 6-month survival are presented in Figure 3, showing a higher cumulative incidence of awakening in favor of DUAL, as well as improved 6-month survival in the DUAL group (log-rank p = 0.003). Among patients who died during hospitalization, 19 (51%) died after withdrawal of life-sustaining therapy (WLST), 17 (46%) from progressive multiorgan failure, and 1 death was attributed to brain death.
[图：Early Neurologic Recovery and Long-Term Survival (A) Cumulative incidence of awakening to follow commands within 45 days]
Safety Outcomes
Safety data are summarized in Table 4. Death during the ICU stay occurred in 12 patients (27.9%) in the DUAL group and 25 patients (60.9%) in the OTHERS group. GGT elevation greater than 3 times UNL was observed in 33 of 43 patients (76.7%) receiving DUAL therapy vs 11 of 41 (26.8%) in the OTHERS group. CK elevation (>10× UNL) occurred in 15 (34.8%) and 11 (32.3%) patients, respectively. One patient in the DUAL group experienced third-degree atrioventricular block; recurrence of cardiac arrest occurred in 2 patients (4.6%) in the DUAL group and 1 patient (2.4%) in the OTHERS group. No other adverse events potentially related to the neurologic treatment were observed.
Table 4. Safety Outcomes
| DUAL (n = 43) | OTHERS (n = 41) | OR (95% CI) | p Value | |
| Death during the ICU stay, n (%) | 12 (27.9) | 25 (60.9) | 0.24 (0.09–0.61) | 0.003 |
| GGT elevation (>3 times the UNL), n (%) | 33 (76.7) | 11 (26.8) | 9.00 (3.34–24.19) | <0.001 |
| Highest GGT value during the ICU stay, U/L, median (IQR) | 624 (318–1,084) | 557 (247–910) | NP | NP |
| CPK elevation (>10 times the UNL), n (%) | 15 (34.8) | 11 (32.3) | 1.12 (0.43–2.90) | 0.815 |
| Third-degree atrioventricular block, n (%) | 1 (2.3) | 0 (0) | NP | NP |
| Recurrence of cardiac arrest, n (%) | 2 (4.6) | 1 (2.4) | NP | NP |
Single Antiglutamate Agents
The results from small cohorts of patients treated with ketamine alone (n = 6; without concomitant perampanel) and perampanel alone (n = 5; without concomitant ketamine) were also collected and are presented in eTable 3. Given the limited sample size, outcomes are reported descriptively; resolution of SRSE and GGT elevation appeared less frequent with either single antiglutamate drug than in the DUAL group.
Discussion
In this multicenter cohort study, which included both retrospective and prospective phases, dual antiglutamate therapy with ketamine and perampanel was associated with a higher rate of resolution of SRSE compared with conventional treatments in selected postcardiac arrest patients, suggesting that this combination may represent a potential etiology-specific therapeutic strategy. The primary finding, 84% seizure resolution with dual antiglutamate therapy vs 49% with conventional combinations, represents a substantial and clinically relevant difference, supporting the rationale for future prospective randomized studies to confirm these results.
The superior efficacy of dual antiglutamate therapy likely reflects targeted intervention against the fundamental pathophysiology of postanoxic brain injury. Global cerebral ischemia triggers massive glutamate release, leading to excitotoxic cell death through both NMDA and AMPA receptor overactivation among other mechanisms.8,22 By simultaneously targeting NMDA and AMPA receptors, ketamine and perampanel may provide a complementary mechanism for interrupting this pathologic cascade, potentially offering advantages over conventional approaches targeting sodium channels or gamma-aminobutyric acid receptors.
Ketamine's neuroprotective properties extend beyond NMDA receptor antagonism, including anti-inflammatory effects and preservation of mitochondrial function.23 Perampanel, as the only selective noncompetitive AMPA receptor antagonist available for clinical use, specifically targets the receptor system responsible for the majority of fast excitatory neurotransmission in the brain.24
Notably, both drugs block their respective receptors in a noncompetitive manner, resulting in insurmountable inhibition of the glutamatergic system that is independent of glutamate levels released during the postanoxic event. Although both ketamine and perampanel have demonstrated efficacy when used individually in SE,25,26 their combined administration may enhance seizure control while providing additional neuroprotection during the vulnerable postresuscitation period.27,28
The timing and selection of patients for aggressive therapy remain critical. In our study, inclusion was restricted to patients meeting a minimum set of prognostic criteria, specifically preserved pupillary reflexes and bilateral N20 responses, while patients with GPDs were excluded. This careful selection likely contributed to the observed treatment success and aligns with the ERC guidelines advocating for multimodal assessment.10 For secondary outcomes focused on neurologic recovery, patients were further stratified according to ERC recommendations, excluding those with 2 or more unfavorable prognostic indicators, given their minimal likelihood of achieving a favorable outcome.
The 56% rate of neurologic awakening, defined as the ability to follow commands, in the dual-therapy group suggests a potential association between effective seizure control and clinically meaningful neurologic recovery. This observation is particularly relevant because early or late awakening during ICU stay has been associated with favorable long-term functional outcomes in postcardiac arrest patients.29 It should be noted, however, that long-term outcomes are influenced not only by neurologic recovery but also by systemic comorbidities, non-neurological causes of death, and cognitive or behavioral deficits that may be identified later, which may account for the apparent attenuation of neurologic benefit over time. Larger prospective studies are warranted to confirm the clinical advantages of dual antiglutamate therapy in postanoxic patients with SE refractory to first-line treatment.
These results extend beyond our previous study and preliminary single-center experience,13,17,20 providing multicenter confirmation of the feasibility of a dual antiglutamate approach. The 84% seizure resolution rate observed in this patient population exceeds that reported with conventional therapies in a recent large multicenter cohort,14 in which the resolution rate was 44%, closely corresponding to the 49% observed in the control arm of our study. Our findings are also consistent with a post hoc analysis of the TELSTAR trial,11 which showed a potential benefit of antiseizure treatment in patients with nonperiodic epileptiform patterns after cardiac arrest.
The safety profile of dual antiglutamate therapy, as outlined in the SUPER-CAT protocol, seems acceptable but deserves further investigation. The primary and frequent adverse event (>75% of cases) was biochemical evidence of cholestatic liver injury, which was reversible following perampanel tapering. This observation aligns with our prior preliminary report in 8 patients receiving perampanel oral loading, where hepatobiliary alterations were noted in 62.5% of cases.17 Although not immediately life threatening, this adverse event warrants further investigation in future clinical studies, with a careful risk-benefit assessment at the individual patient level, given the high burden of concomitant medications and the contribution of hypoxic liver injury in postcardiac arrest syndrome.
The major strength of our study lies in its multidisciplinary approach, with neurologists and intensivists working closely together, and in the application of an innovative treatment protocol in a patient population characterized by severe illness and high mortality. In addition, the multicenter design and the use of standardized definitions and prognostic approach aligned with the ACNS terminology and ERC guidelines enhance the potential generalizability of our findings. Collectively, this approach provides preliminary evidence of improved patient care, for both SE resolution and overall clinical outcomes. Several limitations should be acknowledged. First, the observational study design, including both retrospective and prospective phases, introduces potential selection bias, despite efforts to include consecutive patients meeting eligibility criteria and balanced baseline characteristics between study arms. Treatment assignment was not randomized; however, the absence of established evidence for optimal SRSE management likely minimized systematic treatment preferences. Second, although the sample size was adequate for the primary outcome, it may have limited power to detect differences in long-term clinical outcomes and rare adverse events. Third, the study was conducted at specialized centers with expertise in postcardiac arrest care and SE management, which may limit generalizability. The requirement for continuous EEG monitoring and meticulous multimodal prognostic assessment may not be universally available. Fourth, we adopted a conservative approach and decided to exclude patients with GPDs based on the results of the TELSTAR trial, which showed futility of treatment for GPDs after cardiac arrest. Therefore, the effect of DUAL on GPDs could not be assessed. Fifth, the potential influence of WLST decisions on neurologic outcomes and the risk of survivorship bias cannot be completely excluded. However, WLST decisions were made according to institutional protocols and contemporary neuroprognostication guidelines and were not initiated before 72 hours after cardiac arrest, minimizing the risk of premature treatment withdrawal. Sixth, this study specifically addresses postanoxic SE, which differs substantially from other etiologies of SRSE in terms of pathophysiology and prognosis; further studies are warranted to evaluate the efficacy and safety of the SUPER-CAT protocol in other patient populations. Seventh, perampanel plasma levels were not systematically measured. Given its long half-life (∼105 hours) in the absence of cytochrome P450 3A4 inducers, concentrations by day 5 may approximate those of a 6 mg/d regimen despite not reaching steady state. Future studies could consider higher loading doses, as reported in SE with suspected inflammatory etiology,30 and pharmacokinetics measurements.
In selected patients with SRSE after cardiac arrest, dual antiglutamate therapy with ketamine and perampanel was associated with higher rates of seizure resolution and showed a strong trend toward improved early neurologic recovery compared with conventional therapies, while its benefit on long-term outcomes remains uncertain. These findings support the rationale for a prospective, randomized controlled trial comparing dual antiglutamate therapy with the current standard of care in postanoxic SRSE, ideally conducted as a multinational effort.