危重症患者的炎症与谵妄：综述。
全文（开放获取） · 中文译文
摘要 谵妄在危重症患者中的高发生率及不良影响促使我们更深入地理解其机制。本综述探讨了全身炎症与谵妄之间的关键相互作用，研究了年龄、疼痛和手术等损伤因素如何通过小胶质细胞激活和血脑屏障破坏来协调神经炎症和谵妄。我们的综合分析识别出一系列具有预后价值的炎症和神经元损伤生物标志物，并评估了方兴未艾的抗炎药物治疗领域。我们得出结论，通过生物标志物驱动的试验将这一机制性知识转化为临床实践，对于制定针对谵妄的有效干预措施至关重要。 引言 危重症是指以实际或即将发生的器官功能障碍为特征的急性、危及生命的状况，需要重症监护且常需器官支持治疗（；Marshall et al., 2017）。在临床实践中，危重症患者通常通过入住重症监护病房、需要有创或无创机械通气、血管升压药或正性肌力药物支持、肾脏替代治疗，或存在严重脓毒症、休克、急性呼吸衰竭或多器官功能障碍来识别。病情严重程度通常使用经过验证的评分系统进行评估，如急性生理与慢性健康评估II、序贯器官衰竭评估评分和简化急性生理评分II（；；）。 谵妄是一种急性的、波动的注意力、意识和认知障碍，在短时间内发生，可归因于潜在的医学状况、中毒或戒断、药物暴露或多个相互作用的病因（;）。在危重症患者中，谵妄通常使用经过验证的床旁工具进行评估，如ICU意识模糊评估法和重症监护谵妄筛查检查表（; Bergeron et al., 2001）。谵妄可根据运动表型分为亢进型、低活动型或混合型，并根据临床背景分为ICU获得性谵妄、脓毒症相关性谵妄、术后谵妄、戒断相关性谵妄或叠加于痴呆的谵妄（Peterson et al., 2006; ）。 谵妄在危重症患者中尤为常见，根据疾病严重程度、诊断方法和患者人群的不同，影响约30–50%的普通ICU患者和高达60–80%的机械通气患者（; ; ; ）。ICU谵妄与死亡率增加、机械通气时间延长、ICU和住院时间延长、医疗费用增加以及出院后持续的认知和功能损害相关（; ; ）。更长的谵妄持续时间也与更差的长时认知和执行功能相关（; Wolters et al., 2017）。 危重症中谵妄的高负担反映了重症监护病房（ICU）独特的生理环境。患者经常暴露于全身炎症、低氧血症、休克、脑灌注不足、代谢紊乱、镇静和镇痛药物、睡眠碎片化、疼痛、制动和器官功能障碍中（; ）。这些损害可能通过血脑屏障破坏、内皮功能障碍、神经递质失衡、线粒体损伤、突触信号传导受损、脑连接性改变和神经炎症而汇聚于大脑（; ）。高龄、衰弱、痴呆和多重病症进一步降低大脑储备，并增加对急性脑功能障碍的易感性（; ）。 本综述探讨危重症成人谵妄的炎症相关机制。我们重点关注与ICU级别疾病相关的全身炎症状态，包括脓毒症、需要重症监护的大型手术、疼痛、衰老和衰弱（; ; ）。术后谵妄的证据也被纳入，因为手术应激与危重症重叠，或可为ICU谵妄提供机制性见解（; Berger et al., 2023）。我们并不旨在全面涵盖所有形式的谵妄。相反，我们评估炎症如何与危重症中更广泛的病理生理过程相互作用，并评估炎症生物标志物和抗炎治疗策略是否可为谵妄的预测、预防和治疗提供信息（; ）。 本文是一篇叙述性综述，总结并批判性讨论了危重患者中全身炎症与谵妄之间关系的现有证据。通过检索主要生物医学数据库（包括 PubMed/MEDLINE、Web of Science、Embase 和 Google Scholar），使用关键词组合，如“谵妄”、“ICU 谵妄”、“危重病”、“脓毒症相关性脑病”、“术后谵妄”、“全身炎症”、“神经炎症”、“细胞因子”、“血脑屏障”、“小胶质细胞”、“生物标志物”、“IL-6”、“CRP”、“S100β”、“GFAP”、“tau”、“神经丝轻链”和“抗炎治疗”，确定了相关文献。从所选出版物的参考文献列表中确定了其他相关文章。本综述旨在提供机制、生物标志物和治疗证据的广泛叙述性综合，而非正式的系统综述。 为什么危重患者特别容易发生谵妄 危重症造成一种生物学状态，在此状态下多个稳态系统同时衰竭(; )。脓毒症、创伤或手术引起的全身性炎症可能改变内皮功能并增加血脑屏障通透性，使外周免疫介质能够影响中枢神经系统功能(; )。呼吸衰竭和休克可损害氧输送和脑灌注，而肾和肝功能不全可能改变神经活性药物的代谢和清除(; Hughes et al., 2020)。镇静药、阿片类药物和抗胆碱能药物进一步干扰神经递质系统，包括胆碱能、多巴胺能、GABA能和谷氨酸能信号传导(Hshieh et al., 2018; )。重症监护病房中的环境暴露，如睡眠剥夺、噪音、持续光照和社交隔离，破坏昼夜节律调节和认知定向(; Hshieh et al., 2019)。这些因素与基线易损性相互作用，包括高龄、衰弱、痴呆和多病共存，从而降低谵妄阈值(; )。 炎症与谵妄的病理生理机制 目前，许多假说认为神经炎症是谵妄发生的潜在机制（）。神经炎症引起的神经元线粒体功能障碍、突触损伤和神经元凋亡被认为是导致谵妄的重要机制（）。本文以炎症为中心，从手术创伤、感染、衰老和衰弱等多个角度探讨其与谵妄的关联。 ICU谵妄中感染驱动的神经炎症：从全身炎症到小胶质细胞–C1q介导的突触损伤 既往研究已证实，感染是ICU谵妄的重要危险因素。脓毒症等严重全身性感染可能通过炎症通路引发脑功能障碍()。脓毒症的全身炎症效应以促炎/抗炎平衡紊乱为特征：细菌脂多糖（LPS）等病原相关分子模式激活Toll样受体，触发巨噬细胞和内皮细胞释放IL-1β、IL-6、IL-8和TNF-α等细胞因子。这些炎症介质激活脑内小胶质细胞，启动神经炎症级联反应，最终导致神经元线粒体功能障碍和突触传递受损()。一项研究表明，在脓毒症患者中，海马区小胶质细胞和C1q补体激活导致神经元和突触损伤，从而引起包括谵妄在内的神经认知障碍。靶向小胶质细胞的CSF1-R抑制剂可降低C1q水平和C1q标记突触的数量，防止神经元损伤和突触丢失，并改善神经认知结局()。手术诱发的炎症级联反应 手术创伤引发的系统性炎症反应是术后谵妄（POD）的核心病理机制，尤其是在心脏手术等大手术后。系统性炎症可损害血脑屏障（BBB）并激活小胶质细胞炎症反应（）。手术创伤通过多种途径激活神经炎症，最终导致急性脑功能障碍。血脑屏障（BBB）的破坏在这一过程中起关键作用。研究表明，谵妄患者脑脊液/血浆白蛋白比值（CPAR）和血浆S100B水平显著升高，且升高程度与谵妄严重程度呈正相关（）。这些异常提示血脑屏障通透性增加，促进外周炎症介质侵入中枢神经系统。术中失血和低血压是血脑屏障损伤的重要预测因素，可促进纤溶酶原激活，加剧炎症级联反应和血管内皮损伤（）。此外，围手术期脑脊液（CSF）中纤维蛋白原浓度升高已被证明与血脑屏障破坏、炎症和神经元功能障碍的标志物相关（）。 手术创伤与谵妄。实线箭头表示有临床或实验证据支持的机制；黑色箭头表示直接的病理生理通路（如炎症信号、细胞损伤），而红色箭头表示抑制性或有害过程（如血脑屏障破坏、神经毒性效应）。箭头宽度与相应通路的相对贡献一致。BBB，血脑屏障；CNS，中枢神经系统；CRP，C反应蛋白；DAMPs，损伤相关分子模式；EV，细胞外囊泡；GFAP，胶质纤维酸性蛋白；HMGB1，高迁移率族蛋白B1；IL，白细胞介素；NfL，神经丝轻链；NLRP3，核苷酸结合寡聚化结构域样受体家族含pyrin结构域蛋白3；ROS，活性氧；S100β，S100钙结合蛋白β；TNF-α，肿瘤坏死因子-α。 术后疼痛控制不足也可能成为谵妄的重要促成因素。从生物学角度看，疼痛与谵妄可能通过共同且重叠的机制相关联（）。对疼痛刺激的反应增强受趋化因子、细胞因子和谷氨酸释放的调控，这些物质可同时促进血脑屏障（BBB）的破坏（）。围手术期神经元损伤的有力证据来自一项分析脑脊液（CSF）样本的大型队列研究（）。该研究将神经炎症与神经毒性代谢产物喹啉酸水平升高相关联，喹啉酸的产生源于犬尿氨酸通路中吲哚胺2,3-双加氧酶的激活。喹啉酸激活NMDA型谷氨酸受体，诱导兴奋性毒性，同时抑制糖酵解。值得注意的是，在外科和内科患者中，脑脊液喹啉酸水平与谵妄、神经丝轻链（NfL）升高及生存结局相关（）。这些发现为谵妄发病机制中炎症与神经元损伤之间的关键联系提供了证据。 各种术后认知功能障碍的实验模型均提示，含NOD、LRR和pyrin结构域的蛋白3（NLRP3）炎症小体通路参与谵妄的发生。抑制该炎症通路及其效应分子caspase-1可降低炎症生物标志物的表达，并减轻围术期认知缺陷 (; )。总之，手术诱导的神经炎症通过涉及血脑屏障（BBB）破坏、神经毒性代谢产物生成以及特定炎症小体通路激活的汇聚机制，促进术后谵妄的发生。 老年期衰弱与谵妄：年龄驱动的神经炎症与脑储备耗竭的两个方面 神经炎症与脑储备耗竭 老年人群中发生谵妄的风险显著升高，65岁及以上住院患者中超过50%会出现该并发症 ()。这种易感性增高从根本上与年龄相关的神经炎症调节改变和脑弹性降低有关。 核心机制：衰老大脑中预激活的神经炎症 老年大脑处于一种慢性、低度炎症状态，其特征是基线小胶质细胞激活和促炎细胞因子水平升高。在暴露于手术或感染等急性生理应激源时，这种预激活状态会引发过度且失调的神经炎症反应。临床前证据证实了这一现象：在脂多糖（LPS）刺激后，与年轻对照组相比，老年小鼠海马IL-1β表达的增加幅度是年轻对照组的三倍，这与显著的认知缺陷相关。这种适应不良反应由小胶质细胞表型转变所驱动；转录组学分析显示，老年小胶质细胞组成性高表达如Clec7a等模式识别受体，并对炎症刺激表现出高反应性()。 神经连接性：复原力受损的一个维度 除了分子炎症之外，老化大脑在功能神经网络上表现出易损性。研究表明，谵妄与左侧颞上回（l-STG）至听觉皮层的反馈连接受损有关。该连接的强度与谵妄严重程度及全身炎症程度均呈负相关()，这表明神经通信中断是大脑应对压力能力下降的另一个关键方面。 衰弱与谵妄：易感性的汇聚通路 高龄、衰弱和促炎状态密切相关。由慢性疾病（如心血管疾病、糖尿病）、久坐不动、营养不良和心理社会压力等因素驱动的全身性炎症，通过包括肌少症、胰岛素抵抗和内皮功能障碍在内的途径促进衰弱()。相比之下，谵妄通常由急性炎症打击（如感染、手术）触发，导致促炎细胞因子激增，进而破坏中枢神经系统稳态()。 统一假说：同一枚硬币的两面 一个汇聚框架认为，衰弱和谵妄都是共同的潜在生物学脆弱性的临床表现。这种脆弱性源于随年龄增长而逐渐耗竭的生理储备和代偿机制。衰弱是这种累积缺陷负担的有力临床替代指标()，代表一种全身复原力降低的状态，在这种状态下，轻微的压力源即可诱发功能衰退。类似地，谵妄可能标志着脑适应性能力的衰竭——一种急性脑功能障碍，当大脑的功能储备不足以满足压力源施加的代谢和炎症需求时出现()。因此，衰弱和谵妄可以被视为同一过程的两个相互关联的面：年龄驱动的全身和大脑领域稳态和储备的侵蚀。 尽管感染、手术、创伤、缺氧和衰弱在临床上有所不同，但它们可能通过重叠的生物学途径促进谵妄。外周炎症激活循环细胞因子、补体蛋白、内皮黏附分子、细胞外囊泡和损伤相关分子模式。这些信号可通过体液途径、神经传入、内皮激活和血脑屏障破坏影响大脑。在中枢神经系统内，炎症信号传导促进小胶质细胞活化、星形胶质细胞功能障碍、氧化应激、线粒体损伤、突触可塑性受损和神经传递改变。衰老和衰弱可能通过小胶质细胞预激活和认知储备减少进一步放大这种反应。因此，危重症中的谵妄不太可能反映单一途径；相反，它源于全身炎症与血管、代谢、胶质细胞和神经元稳态受损的汇聚。 炎症标志物与谵妄的预测 将炎症与谵妄联系起来的临床证据的主要类别总结于 ，危重症患者炎症相关谵妄的候选生物标志物列于 。炎症性和神经退行性生物标志物共同预测谵妄，但它们通过不同的机制层面发挥作用。两项汇总了 >40 项研究的大型荟萃分析确定了九种血清蛋白——IL-6、CRP、IL-8、IL-10、MCP-1、GFAP、IL-1β、神经丝轻链（NfL）和神经营养因子IGF-1——在术后谵妄（POD）中可重复地改变。与此同时，一项多中心前瞻性队列（n = 1,043）证明，IL-6、IL-8、IL-10、TNF-α 和 TNFR1 升高，加上蛋白C降低，可在测量后24小时内独立预测谵妄或昏迷（）。然而，同一ICU队列的1年神经认知随访显示，累积谵妄持续时间预测持续性自我报告认知障碍，且与上述炎症指标无显著交互作用，表明下游神经退行性过程——而非持续的全身性炎症——维持长期缺陷（）。 在新兴的生物标志物中，星形胶质细胞损伤标志物提供了互补的时间分辨率。血浆S100β通过反映血脑屏障破坏独立预测了新发谵妄和谵妄相关死亡率（），而GFAP升高先于活动减退型谵妄的临床表现约12小时，并以82%的敏感度识别出亚综合征状态（）。总tau蛋白是轴突损伤的标志物，在术后第1天升高>15 pg/mL，并以91%的敏感度和79%的特异度预测谵妄；其浓度相关性更强
全文（开放获取） · 英文原文
【Abstract】 The high prevalence and adverse impact of delirium in critically ill patients necessitate a deeper understanding of its mechanisms. This review explores the pivotal interplay between systemic inflammation and delirium, examining how insults like age, pain and surgery orchestrate neuroinflammation and delirium through microglial activation and blood-brain barrier disruption. Our synthesis identifies a spectrum of inflammatory and neuronal injury biomarkers with prognostic value and evaluates the burgeoning field of anti-inflammatory pharmacotherapy. We conclude that bridging this mechanistic knowledge to clinical practice via biomarker-driven trials is essential for developing effective interventions against delirium. 【Introduction】 Critical illness refers to an acute, life-threatening condition characterized by actual or impending organ dysfunction requiring intensive monitoring and, frequently, organ-supportive therapy (; Marshall et al., 2017). In clinical practice, critically ill patients are commonly identified by admission to an intensive care unit, the need for invasive or non-invasive mechanical ventilation, vasopressor or inotropic support, renal replacement therapy, or the presence of severe sepsis, shock, acute respiratory failure or multiple organ dysfunction. Severity of illness is commonly assessed using validated scoring systems such as the Acute Physiology and Chronic Health Evaluation II, the Sequential Organ Failure Assessment score and the Simplified Acute Physiology Score II (; ; ). Delirium is an acute and fluctuating disturbance in attention, awareness and cognition that develops over a short period and is attributable to an underlying medical condition, intoxication or withdrawal, medication exposure or multiple interacting causes (; ). In critically ill patients, delirium is commonly assessed using validated bedside instruments such as the Confusion Assessment Method for the ICU and the Intensive Care Delirium Screening Checklist (; Bergeron et al., 2001). Delirium may be classified according to motor phenotype as hyperactive, hypoactive or mixed, and according to clinical context as ICU-acquired delirium, sepsis-associated delirium, postoperative delirium, withdrawal-related delirium or delirium superimposed on dementia (Peterson et al., 2006; ). Delirium is particularly common in critically ill patients, affecting approximately 30–50% of general ICU patients and up to 60–80% of mechanically ventilated patients, depending on illness severity, diagnostic approach and patient population (; ; ; ). ICU delirium is associated with increased mortality, prolonged mechanical ventilation, longer ICU and hospital stay, greater health-care costs and persistent cognitive and functional impairment after discharge (; ; ). Longer delirium duration is also associated with worse long-term cognitive and executive function (; Wolters et al., 2017). The high burden of delirium in critical illness reflects the distinctive physiological environment of the ICU. Patients are frequently exposed to systemic inflammation, hypoxaemia, shock, cerebral hypoperfusion, metabolic derangements, sedative and analgesic medications, sleep fragmentation, pain, immobilization and organ dysfunction (; ). These insults may converge on the brain through blood–brain barrier disruption, endothelial dysfunction, neurotransmitter imbalance, mitochondrial injury, impaired synaptic signaling, altered cerebral connectivity and neuroinflammation (; ). Advanced age, frailty, dementia and multimorbidity further reduce cerebral reserve and increase vulnerability to acute brain dysfunction (; ). This Review examines inflammation-related mechanisms of delirium in critically ill adults. We focus on systemic inflammatory states relevant to ICU-level illness, including sepsis, major surgery requiring intensive monitoring, pain, aging and frailty (; ; ). Evidence from postoperative delirium is included where surgical stress overlaps with critical illness or provides mechanistic insight relevant to ICU delirium (; Berger et al., 2023). We do not aim to cover all forms of delirium comprehensively. Instead, we evaluate how inflammation interacts with broader pathophysiological processes in critical illness and assess whether inflammatory biomarkers and anti-inflammatory therapeutic strategies can inform delirium prediction, prevention and treatment (; ). This article is a narrative review that summarizes and critically discusses current evidence regarding the relationship between systemic inflammation and delirium in critically ill patients. Relevant literature was identified through searches of major biomedical databases, including PubMed/MEDLINE, Web of Science, Embase, and Google Scholar, using combinations of keywords such as “delirium,” “ICU delirium,” “critical illness,” “sepsis-associated encephalopathy,” “postoperative delirium,” “systemic inflammation,” “neuroinflammation,” “cytokines,” “blood-brain barrier,” “microglia,” “biomarkers,” “IL-6,” “CRP,” “S100β,” “GFAP,” “tau,” “neurofilament light chain,” and “anti-inflammatory therapy.” Additional relevant articles were identified from the reference lists of selected publications. This review was designed to provide a broad narrative synthesis of mechanistic, biomarker, and therapeutic evidence rather than a formal systematic review. 【Why critically Ill patients are uniquely vulnerable to delirium】 Critical illness creates a biological state in which multiple homeostatic systems fail simultaneously (; ). Systemic inflammation from sepsis, trauma or surgery may alter endothelial function and increase blood–brain barrier permeability, allowing peripheral immune mediators to influence central nervous system function (; ). Respiratory failure and shock can impair oxygen delivery and cerebral perfusion, whereas renal and hepatic dysfunction may alter the metabolism and clearance of neuroactive drugs (; Hughes et al., 2020). Sedatives, opioids and anticholinergic medications further perturb neurotransmitter systems, including cholinergic, dopaminergic, GABAergic and glutamatergic signaling (Hshieh et al., 2018; ). Environmental exposures in the ICU, such as sleep deprivation, noise, continuous lighting and social isolation, disrupt circadian regulation and cognitive orientation (; Hshieh et al., 2019). These factors interact with baseline vulnerability, including advanced age, frailty, dementia and multimorbidity, thereby lowering the threshold for delirium (; ). 【Pathophysiological mechanisms of inflammation and delirium】 At present, numerous hypotheses suggest that neuroinflammation serves as the underlying mechanism for the onset of delirium (). The dysfunction of neuronal mitochondria, synaptic impairment, and neuronal apoptosis induced by neuroinflammation are considered important mechanisms contributing to delirium (). This paper centers on inflammation and explores its association with delirium from multiple perspectives, including surgical trauma, infection, aging, and frailty. 【Infection-driven neuroinflammation in ICU delirium: from systemic inflammation to microglial–C1q-mediated synaptic injury】 Previous studies have established that infection is a significant risk factor for ICU delirium. Severe systemic infections such as sepsis may trigger brain dysfunction through inflammatory pathway (). The systemic inflammatory effect of sepsis is characterized by the disruption of the pro-inflammatory/anti-inflammatory balance: pathogen-associated molecular patterns such as bacterial lipopolysaccharide (LPS) activate Toll-like receptors, triggering the release of cytokines like IL-1β, IL-6, IL-8, and TNF-α by macrophages and endothelial cells. These inflammatory mediators activate microglia in the brain, initiating a cascade of neuroinflammation, ultimately leading to mitochondrial dysfunction and impaired synaptic transmission in neurons (). A study has shown that in patients with sepsis, microglia in the hippocampus and C1q complement activation cause damage to neurons and synapses, thereby leading to neurocognitive disorders including delirium. CSF1-R inhibitors, targeting microglia, reduce C1q levels and the number of C1q-labeled synapses, prevent neuronal damage and synapse loss, and improve neurocognitive outcomes (). 【Surgery-induced inflammatory cascades】 The systemic inflammatory response triggered by surgical trauma constitutes a central pathological mechanism of postoperative delirium (POD), particularly following major procedures such as cardiac surgery. Systemic inflammation can compromise the blood-brain barrier (BBB) and activate microglial inflammatory responses () Surgical trauma activates neuroinflammation through multiple pathways, ultimately leading to acute brain dysfunction. Disruption of the blood-brain barrier (BBB) plays a critical role in this process. Studies have demonstrated that the cerebrospinal fluid/plasma albumin ratio (CPAR) and plasma levels of S100B are significantly elevated in patients with delirium, with the degree of increase positively correlating with delirium severity (). These abnormalities indicate increased BBB permeability, facilitating the invasion of peripheral inflammatory mediators into the central nervous system. Intraoperative blood loss and hypotension are significant predictors of BBB damage, promoting plasminogen activation and exacerbating the inflammatory cascade and vascular endothelial injury (). Furthermore, perioperative increases in cerebrospinal fluid (CSF) concentrations of fibrinogen have been shown to correlate with markers of BBB disruption, inflammation, and neuronal dysfunction (). Surgical trauma and delirium. Solid arrows indicate mechanisms supported by clinical or experimental evidence. Black arrows represent direct pathophysiological pathways (e.g., inflammatory signaling, cellular injury), whereas red arrows denote inhibitory or deleterious processes (e.g., blood–brain barrier disruption, neurotoxic effects). Arrow width corresponds to relative pathway contribution. BBB, blood–brain barrier; CNS, central nervous system; CRP, C-reactive protein; DAMPs, damage-associated molecular patterns; EV, extracellular vesicle; GFAP, glial fibrillary acidic protein; HMGB1, high-mobility group box 1; IL, interleukin; NfL, neurofilament light chain; NLRP3, nucleotide-binding oligomerization domain-like receptor family pyrin domain-containing 3; ROS, reactive oxygen species; S100β, S100 calcium-binding protein β; TNF-α, tumor necrosis factor-α . Inadequate postoperative pain control may also be an important contributing factor to delirium. From a biological perspective, pain and delirium may be linked via shared and overlapping mechanisms (). Enhanced responsiveness to pain stimuli is regulated by the release of chemokines, cytokines, and glutamate, which can concurrently promote BBB breakdown (). Compelling evidence for perioperative neuronal injury comes from a large cohort study analyzing CSF samples (). This study associated neuroinflammation with elevated levels of the neurotoxic metabolite quinolinic acid, resulting from activation of indoleamine 2,3-dioxygenase in the kynurenine pathway. Quinolinic acid activates NMDA-type glutamate receptors, inducing excitotoxicity, while simultaneously suppressing glycolysis. Notably, CSF levels of quinolinic acid were associated with delirium, increased neurofilament light chain (NfL), and survival outcomes in both surgical and medical patients (). These findings provide a crucial link between inflammation and neuronal injury in the pathogenesis of delirium. Various experimental models of postoperative cognitive impairment implicate the NOD-, LRR- and pyrin domain-containing protein 3 (NLRP3) inflammasome pathway in delirium. Inhibition of this inflammatory pathway and its effector, caspase-1, reduces the expression of inflammatory biomarkers and mitigates perioperative cognitive deficits (; ). In summary, surgery-induced neuroinflammation promotes the development of postoperative delirium through a convergent mechanism involving BBB disruption, generation of neurotoxic metabolites, and activation of specific inflammasome pathways. 【Frailty and delirium in late life: two faces of age-driven neuroinflammation and brain-reserve depletion】 【Neuroinflammation and cerebral reserve depletion】 The risk of developing delirium is markedly elevated in the elderly population, with over 50% of hospitalized patients aged 65 years and older experiencing this complication (). This heightened susceptibility is fundamentally linked to age-associated alterations in neuroinflammatory regulation and diminished cerebral resilience. 【Core mechanism: primed neuroinflammation in the aging brain】 The aged brain exists in a state of chronic, low-grade inflammation, characterized by baseline microglial activation and elevated pro-inflammatory cytokine levels. Upon exposure to an acute physiological stressor such as surgery or infection, this primed state precipitates an exaggerated and dysregulated neuroinflammatory response. Preclinical evidence underscores this phenomenon: following lipopolysaccharide (LPS) challenge, aged mice exhibit a threefold greater increase in hippocampal IL-1β expression compared to young controls, which correlates with significant cognitive deficits. This maladaptive response is driven by a microglial phenotypic shift; transcriptomic analyses reveal that aged microglia constitutively express high levels of pattern recognition receptors like Clec7a and demonstrate hyper-responsiveness to inflammatory stimuli (). 【Neural connectivity: a dimension of compromised resilience】 Beyond molecular inflammation, the aging brain exhibits vulnerabilities in functional neural networks. Research indicates that delirium is associated with impaired feedback connectivity from the left superior temporal gyrus (l-STG) to the auditory cortex. The strength of this connection is negatively correlated with both delirium severity and the magnitude of systemic inflammation (), suggesting that disrupted neural communication represents another critical facet of the brain’s diminished capacity to manage stress. 【Frailty and delirium: converging pathways of vulnerability】 Advanced age, frailty, and a pro-inflammatory state are deeply interlinked. Systemic inflammation, driven by factors such as chronic diseases (e.g., cardiovascular disease, diabetes), sedentariness, malnutrition, and psychosocial stress, promotes frailty through pathways including sarcopenia, insulin resistance, and endothelial dysfunction (). In contrast, delirium is typically triggered by an acute inflammatory insult (e.g., infection, surgery) that leads to a surge of pro-inflammatory cytokines, which then disrupt central nervous system homeostasis (). 【A unifying hypothesis: two sides of the same coin】 A convergent framework posits that both frailty and delirium are clinical expressions of a common underlying biological vulnerability. This vulnerability stems from the age-progressive depletion of physiological reserves and compensatory mechanisms. Frailty serves as a robust clinical proxy for this accumulated deficit burden (), representing a state of reduced systemic resilience where minor stressors can precipitate functional decline. Analogously, delirium likely signifies a failure of cerebral adaptive capacity—an acute brain dysfunction that emerges when the brain’s functional reserves are insufficient to meet the metabolic and inflammatory demands imposed by a stressor (). Thus, frailty and delirium can be viewed as two interrelated faces of the same process: age-driven erosion of homeostasis and reserve across systemic and cerebral domains. Although infection, surgery, trauma, hypoxia, and frailty differ clinically, they may promote delirium through overlapping biological pathways. Peripheral inflammation activates circulating cytokines, complement proteins, endothelial adhesion molecules, extracellular vesicles, and damage-associated molecular patterns. These signals can affect the brain through humoral routes, neural afferents, endothelial activation, and disruption of the blood–brain barrier. Within the central nervous system, inflammatory signaling promotes microglial activation, astrocytic dysfunction, oxidative stress, mitochondrial injury, impaired synaptic plasticity, and altered neurotransmission. Aging and frailty may further amplify this response through microglial priming and reduced cognitive reserve. Therefore, delirium in critical illness is unlikely to reflect a single pathway; rather, it emerges from the convergence of systemic inflammation with impaired vascular, metabolic, glial, and neuronal homeostasis. 【Inflammatory markers and the prediction of delirium】 The major categories of clinical evidence linking inflammation and delirium are summarized in , and candidate biomarkers for inflammation-associated delirium in critically ill patients are listed in . Inflammatory and neurodegenerative biomarkers jointly forecast delirium, yet they operate through distinct mechanistic strata. Two large-scale meta-analyses that aggregated >40 studies identified nine serum proteins—IL-6, CRP, IL-8, IL-10, MCP-1, GFAP, IL-1β, neurofilament light (NfL), and the neurotrophic factor IGF-1—as being reproducibly altered in post-operative delirium (POD). In parallel, a multicenter prospective cohort (n = 1,043) demonstrated that elevated IL-6, IL-8, IL-10, TNF-α, and TNFR1, coupled with reduced protein C, independently predicted delirium or coma within 24 h after measurement (). Nevertheless, 1-year neurocognitive follow-up of the same ICU cohort revealed that cumulative delirium duration predicted persistent self-reported cognitive impairment without significant interaction with the aforementioned inflammatory panel, indicating that downstream neurodegenerative processes—rather than sustained systemic inflammation—maintain long-term deficits (). Summary of major clinical evidence linking inflammation and delirium. Candidate biomarkers for inflammation-associated delirium in critically ill patients. Among emerging biomarkers, astrocytic injury signatures offer complementary temporal resolution. Plasma S100β independently predicted incident delirium and delirium-related mortality by reflecting blood–brain barrier disruption (), whereas GFAP elevations preceded clinical manifestation of hypoactive delirium by approximately 12 h and identified subsyndromal states with 82% sensitivity (). Total tau, a marker of axonal injury, rose >15 pg/mL on post-operative day 1 and predicted delirium with 91% sensitivity and 79% specificity; its concentration correlated more strongly