2025年3月18日出版的《美国医学会杂志》发表了匹兹堡大学Romain Pirracchio研究团队的最新成果，他们报道了脓毒性休克中血管加压素的最佳启动。

重要性:去甲肾上腺素是脓毒性休克患者的一线血管加压药物。何时以及是否应添加第二种药物，如抗利尿激素，尚不清楚。

目的:推导并验证强化学习模型，以确定接受去甲肾上腺素治疗感染性休克的成人危重患者抗利尿激素的最佳起始规则。

设计、设置和参与者：利用2012年至2023年加州5家医院3608名符合脓毒症-3休克标准的患者的电子健康记录数据，使用强化学习来生成抗利尿激素启动的最佳规则，以改善短期和医院预后。该规则在来自加州数据集和3个外部数据集的628名患者中进行了评估，这些数据集包括10来自227家美国医院的217名患者，使用加权重要抽样和混合逻辑回归与逆概率加权。

暴露 ；临床、实验室和治疗变量在电子健康记录中按小时分组，持续120小时。

主要结局和测量方法 ；主要结局为院内死亡率。

结果:衍生队列（n = 3608）包括2075名男性（57%），中位（IQR）年龄为63（56-70）岁，休克发作时序贯器官衰竭评估（SOFA）评分为5（3-7）。验证队列（n = 10217）为56%的男性（n = 5743），中位（IQR）年龄为67（57-75）岁，SOFA评分为6（4-9）。在验证数据中，该模型显示，与临床医生的作用相比，抗利尿激素启动的患者更多（87%对31%），相对于休克发作更早（中位数[IQR]， 4[1-8]对5[1-14]小时），去甲肾上腺素剂量更低（中位数[IQR]， 0.20[0.08-0.45]对0.37 [0.17-0.69]μg/kg/min）。与临床医生的行为相比，该规则在验证数据中与更大的预期奖励相关（加权重要性抽样差异，31）。如果抗利尿激素启动与不同的规则相似，则调整后的医院死亡率较低（优势比，0.81），这一发现在外部验证集中是一致的。

研究结果表明，在接受去甲肾上腺素治疗的感染性休克成年患者中，抗利尿激素的使用是可变的。在几个观察数据集中开发并验证的强化学习模型建议比平均护理模式更频繁和更早地使用抗利尿激素，并与降低死亡率相关。

附：英文原文

Title: Optimal Vasopressin Initiation in Septic Shock: The OVISS Reinforcement Learning Study

Author: Alexandre Kalimouttou, Jason N. Kennedy, Jean Feng, Harvineet Singh, Suchi Saria, Derek C. Angus, Christopher W. Seymour, Romain Pirracchio

Issue&Volume: 2025-03-18

Abstract: Importance  Norepinephrine is the first-line vasopressor for patients with septic shock. When and whether a second agent, such as vasopressin, should be added is unknown.

Objective  To derive and validate a reinforcement learning model to determine the optimal initiation rule for vasopressin in adult, critically ill patients receiving norepinephrine for septic shock.

Design, Setting, and Participants  Reinforcement learning was used to generate the optimal rule for vasopressin initiation to improve short-term and hospital outcomes, using electronic health record data from 3608 patients who met the Sepsis-3 shock criteria at 5 California hospitals from 2012 to 2023. The rule was evaluated in 628 patients from the California dataset and 3 external datasets comprising 10217 patients from 227 US hospitals, using weighted importance sampling and pooled logistic regression with inverse probability weighting.

Exposures  Clinical, laboratory, and treatment variables grouped hourly for 120 hours in the electronic health record.

Main Outcome and Measure  The primary outcome was in-hospital mortality.

Results  The derivation cohort (n=3608) included 2075 men (57%) and had a median (IQR) age of 63 (56-70) years and Sequential Organ Failure Assessment (SOFA) score at shock onset of 5 (3-7 [range, 0-24, with higher scores associated with greater mortality]). The validation cohorts (n=10217) were 56% male (n=5743) with a median (IQR) age of 67 (57-75) years and a SOFA score of 6 (4-9). In validation data, the model suggested vasopressin initiation in more patients (87% vs 31%), earlier relative to shock onset (median [IQR], 4 [1-8] vs 5 [1-14] hours), and at lower norepinephrine doses (median [IQR], 0.20 [0.08-0.45] vs 0.37 [0.17-0.69] μg/kg/min) compared with clinicians’ actions. The rule was associated with a larger expected reward in validation data compared with clinician actions (weighted importance sampling difference, 31 [95% CI, 15-52]). The adjusted odds of hospital mortality were lower if vasopressin initiation was similar to the rule compared with different (odds ratio, 0.81 [95% CI, 0.73-0.91]), a finding consistent across external validation sets.

Conclusions and Relevance  In adult patients with septic shock receiving norepinephrine, the use of vasopressin was variable. A reinforcement learning model developed and validated in several observational datasets recommended more frequent and earlier use of vasopressin than average care patterns and was associated with reduced mortality.

DOI: 10.1001/jama.2025.3046

Source: https://jamanetwork.com/journals/jama/fullarticle/2831858