重症监护免疫学的范式转变
几十年来，败血性休克的临床管理几乎完全集中在抑制初始的过度炎症反应——所谓的细胞因子风暴。然而，当代研究表明，败血症是一种同时存在免疫激活和严重免疫抑制的状态。这种‘免疫麻痹’通常是晚期死亡的主要驱动因素，表现为 ICU 获得性感染、潜伏病毒再激活和伤口愈合受损。随着我们进入精准医疗时代，识别最有可能从免疫刺激疗法中受益的患者已成为首要挑战。两项最近的里程碑研究突出显示，单核细胞 HLA-DR (mHLA-DR) 表达是最可靠且可操作的生物标志物，用于这种分层。
单核细胞 HLA-DR：免疫麻痹的哨兵
人类白细胞抗原-DR (HLA-DR) 是一种表达在抗原呈递细胞表面（尤其是单核细胞）上的分子。其主要作用是向 T 细胞呈递处理后的肽段，从而启动适应性免疫反应。在危重病的背景下，mHLA-DR 分子密度的下降——以每细胞抗体数 (AB/C) 衡量——作为单核细胞功能和整体免疫无能的替代指标。尽管提出了许多生物标志物，但 mHLA-DR 因其标准化测量协议及其与免疫麻痹临床现象的直接相关性而脱颖而出。
20 年败血性休克队列的见解
Monneret 等人在《重症医学》杂志上发表的研究提供了对 mHLA-DR 预测价值前所未有的深度。通过分析过去二十年中 1,023 名败血性休克患者的真实世界队列，研究人员证明 mHLA-DR 不仅是一个诊断的好奇点；它是一个强大的富集生物标志物。
8,000 AB/C 的阈值
该研究证实，< 8,000 AB/C 的阈值——已在 II 期临床试验中使用——与第 28 天和第 90 天的死亡率显著相关。低于此阈值的患者还表现出更高的 ICU 获得性感染发生率。这些数据强化了使用这一特定截止值来识别不良结果高风险患者的重要性。
A Principal component analysis (PCA) of patients’ main clinical characteristics and mHLA-DR expression at ICU admission. PCA was performed to reduce data dimensionality by retaining quantitative variables commonly used to describe a septic shock cohort and to assess cohort homogeneity over time. The loadings of the eight variables on the first two principal components, together accounting for 41% of the total variance, are shown. B Dots represent individual PCA observations over a 20-year period, with confidence ellipses illustrating group (successive period of inclusion) distributions. Samples were divided into four chronological periods, each comprising approximately 255 patients. C Daily occurrence of mortality and ICU-acquired infections from ICU admission to day 30. Results are shown as the number of events per day using interleaved bar graphs. D Time course of mHLA-DR expression in the whole cohort. Samples were collected at four time points: day 1/2 (n = 899), day 3/4 (n = 886), day 5/7 (n = 650), and at ICU discharge (n = 162). mHLA-DR values are expressed as AB/C (antibodies bound per cell). Results are presented as median with interquartile range (Q1–Q3). The accompanying table shows the numerical values at each time-point. The bottom row indicates the percentage of patients with mHLA-DR ≤ 8000 at each time-point. The dashed line represents the lower limit of reference range of mHLA-DR (i.e., 13,500 AB/C)
Association of mHLA-DR expression with main outcomes. A Time course of mHLA-DR expression according to 28-day mortality (n = 301 deceased patients). B Time course of mHLA-DR expression according to 90-day mortality (n = 378 deceased patients). C Based on cut-off value of 8000 AB/C (antibodies bound per cell) at day 3/4, patients were separated into 2 groups to build Kaplan–Meier survival curves until day 90. The log-rank test was used to test the difference between the curves. D Based on cut-off value of 8000 AB/C (antibodies bound per cell) at day 5/7, patients were separated into 2 groups to build Kaplan–Meier survival curves until day 90. The log-rank test was used to test the difference between the curves. E Time course of mHLA-DR expression according to ICU-acquired infection (n = 190 patients). mHLA-DR values were censored once infection occurred. F Based on cut-off value of 8000 AB/C (antibodies bound per cell) at day 5/7, patients were separated into 2 groups to build Kaplan–Meier infection acquisition curves. The log-rank test was used to test the difference between the curves. G Additional mHLA-DR expression (n = 226 patients for which a fourth sample was available, mean time of sampling = 16 days [Q1–Q3 = 11–22] following ICU admission) according to 90-day mortality. H Based on cut-off value of 8000 AB/C (antibodies bound per cell) for the additional sample, patients were separated into 2 groups to build Kaplan–Meier survival curves until day 90. The log-rank test was used to test the difference between the curves. Results are presented with median and Q1–Q3. Mann–Whitney test was used for comparison between groups. ns: non-significant, *p value < 0.05, **p value < 0.01, ***p value < 0.001, ****p value < 0.0001
Multivariate analysis. A Association with mortality over time. mHLA-DR expression was modeled as a time-dependent covariate using a Cox proportional hazards model, incorporating repeated measurements as a longitudinal exposure. The threshold of 8000 AB/C (antibodies bound per cell) was applied. The model was adjusted for age, SAPS II, and SOFA scores at admission. B Association with mortality at discrete time points. Three separate landmark Cox regression models were performed at days 1/2, 3/4, and 5/7. The threshold of 8000 AB/C (antibodies bound per cell) was applied. Models were adjusted for age, SAPS II, and SOFA scores at admission. C Association with ICU-acquired infections. Cumulative incidence of ICU-acquired infections was analyzed at days 1/2, 3/4, and 5/7 using Fine and Gray competing risk models, treating death and ICU discharge as competing events. Subdistribution hazard ratios were adjusted for age, SAPS II, and SOFA scores
轨迹的重要性
这项 20 年分析中最显著的发现之一是早期下调与持续免疫抑制之间的区别。研究人员指出，早期免疫下调（前 48 小时内）可能代表对初始系统性炎症反应的生理适应。然而，第一周内未能恢复 mHLA-DR 表达水平的患者风险最高。通过 K 均值轨迹聚类，研究显示在 ICU 初期后跟踪 mHLA-DR 是至关重要的。过早进行单一测量可能无法区分那些自然恢复和那些进入慢性免疫功能障碍状态的患者。
肝移植中的预后预测
mHLA-DR 的应用范围不仅限于一般败血症人群，还扩展到实体器官移植等专业领域。Delignette 等人的一项前瞻性观察研究重点关注接受肝移植 (LT) 的患者。这些患者由于手术前已存在的肝硬化相关免疫功能障碍 (CAID) 而特别脆弱。
感染风险与 mHLA-DR 恢复
在接受 LT 的 99 名患者中，超过 35% 在术后早期出现感染。研究发现，淋巴细胞计数或功能性 T 细胞检测均不能预测这些结果。相反，mHLA-DR 动力学是决定性因素。具体而言，mHLA-DR 恢复延迟（定义为第 7 天 < 11,000 AB/C）与感染风险增加 12 倍（比值比 12.1）相关。
生存率和临床相关性
免疫恢复延迟的临床影响非常明显。未能在术后第 7 天达到 mHLA-DR 阈值的患者一年生存率显著较低（77.8% 对比正常恢复者的 98.3%）。当结合术前 MELD 评分 > 30 时，mHLA-DR 监测提供了一个强大的预测模型，用于识别需要加强术后监测和潜在修改免疫抑制方案的患者。
专家评论和临床整合
这两项研究的汇合表明，应将 mHLA-DR 整合到标准的重症监护和移植监测协议中。然而，仍有一些临床考虑事项。
方法标准化
这些发现的优势在于使用了标准化流式细胞术。为了将 mHLA-DR 用作通用富集生物标志物，实验室必须严格校准协议，以确保不同机构的 AB/C 测量结果具有可比性。
治疗意义
识别‘低表达者’为靶向免疫刺激疗法（如粒细胞-巨噬细胞集落刺激因子 (GM-CSF) 或干扰素-γ (IFN-γ)）打开了大门。通过使用 mHLA-DR 轨迹，临床医生可以避免过度治疗已经自然恢复路径的患者，而是专注于那些持续免疫麻痹的患者。
研究局限性
虽然 20 年队列提供了大规模数据，但其数据收集本质上是回顾性的，可能会因过去二十年中标准败血症护理的变化而引入偏差。在肝移植研究中，样本量虽为前瞻性，但较小且基于单一中心。移植中 11,000 AB/C 阈值的进一步多中心验证是必要的。
结论：免疫导向医学的未来
证据明确：mHLA-DR 是应对 ICU 复杂情况的关键工具。无论是管理败血性休克的后果还是肝移植的恢复阶段，临床医生都必须超越免疫系统的静态快照。通过监测 mHLA-DR 轨迹，我们可以识别干预的‘机会窗口’，可能减少继发感染的负担并提高最脆弱患者的长期生存率。
资金和临床试验信息
肝移植研究在 ClinicalTrials.gov 注册，注册号为 NCT03995537（2019 年 6 月 20 日）。败血性休克队列研究由机构资助，并代表同类研究中最大的真实世界数据集之一。
参考文献
1. Monneret G, Lafon T, Gossez M, et al. 单核细胞 HLA-DR 表达在败血性休克患者中的表现：来自 1023 例 20 年真实世界队列的见解。Intensive Care Med. 2025;51(10):1820-1832。
2. Delignette MC, Riff A, Antonini T, et al. ICU 中个体 mHLA-DR 轨迹作为肝移植后早期感染的预测指标：一项前瞻性观察研究。Crit Care. 2025;29(1):79。