急性胰腺炎相关急性呼吸窘迫综合征的形态学亚表型
该回顾性队列纳入2003至2023年ICU收治的151例急性胰腺炎相关ARDS且早期行CT检查的连续成人重症患者,按CT分为局灶型ARDS(81例,54%)与非局灶型ARDS(70例,46%)。两组90天死亡率无显著差异(局灶型41% vs 非局灶型34%,aHR 1.01,95%CI 0.58–1.76,p=0.98)。俯卧位通气与局灶型ARDS更低90天死亡相关(aHR 0.25,95%CI 0.08–0.76,p=0.015),在非局灶型中无此关联;高PEEP与非局灶型ARDS更低90天死亡相关(aHR 0.22,95%CI 0.06–0.85,p=0.029),在局灶型中无此关联。机器学习模型结果与上述差异反应一致。作者认为形态学可能有助于指导该特定ARDS人群的个体化通气策略。
要点
- 急性胰腺炎相关ARDS可按CT形态分为局灶型(81/151,54%)与非局灶型(70/151,46%),两组90天死亡率无显著差异(41% vs 34%,aHR 1.01,95%CI 0.58–1.76,p=0.98)。
- 局灶型ARDS中,俯卧位通气与更低90天死亡相关(aHR 0.25,95%CI 0.08–0.76,p=0.015);非局灶型中未见此关联。
- 非局灶型ARDS中,高PEEP与更低90天死亡相关(aHR 0.22,95%CI 0.06–0.85,p=0.029);局灶型中未见此关联。
- 机器学习模型(XGBoost与SHAP解释)结果与俯卧位通气和高PEEP的差异反应一致。
适用人群与场景
- 2003至2023年ICU收治的急性胰腺炎相关ARDS且早期行CT检查的连续成人重症患者。
诊断标准与关键阈值
- 根据CT扫描将ARDS分为「局灶型ARDS」或「非局灶型ARDS」。
处置流程与药物剂量
- 俯卧位通气与局灶型ARDS更低90天死亡相关(aHR 0.25,95%CI 0.08–0.76,p=0.015),在非局灶型ARDS中无此关联。
- 高PEEP与非局灶型ARDS更低90天死亡相关(aHR 0.22,95%CI 0.06–0.85,p=0.029),在局灶型ARDS中无此关联。
研究设计与关键结果
- 回顾性队列研究,筛查5865例患者,最终纳入151例急性胰腺炎相关ARDS患者进行分析。
- 局灶型ARDS 81例(54%),非局灶型ARDS 70例(46%),两组ARDS管理无显著差异。
- 总体90天死亡率为37%(56/151)。
- 局灶型与非局灶型ARDS的90天死亡率无显著差异(33/81 [41%] vs 33/70 [34%],aHR 1.01,95%CI 0.58–1.76,p=0.98)。
- 俯卧位通气与局灶型ARDS更低90天死亡相关(aHR 0.25,95%CI 0.08–0.76,p=0.015),在非局灶型ARDS中无此关联。
- 高PEEP与非局灶型ARDS更低90天死亡相关(aHR 0.22,95%CI 0.06–0.85,p=0.029),在局灶型ARDS中无此关联。
- 机器学习模型(XGBoost与SHAP解释)结果与俯卧位通气和高PEEP的差异反应一致。
证据类型、等级与局限
- 证据类型:回顾性队列研究。
- 局限:原文未报告局限性。
对床旁实践的改变
- 作者认为,形态学可能有助于指导该特定ARDS人群的个体化通气策略。
提炼说明
- 本条依据论文摘要整理;摘要未涉及「红旗征与不能漏的情况」「指南或共识出处」,故未列出这些栏目;全文请点「阅读原文」。
英文原文(摘要)· 点击展开
BACKGROUND: Acute Respiratory Distress Syndrome (ARDS) encompasses heterogenous subphenotypes. We assessed whether morphological subphenotypes of Acute Pancreatitis-related ARDS have differential responses to prone position and positive end-expiratory pressure (PEEP). METHODS: We analyzed a retrospective cohort of all consecutive critically ill adults with Acute Pancreatitis-related ARDS and early CT-scan admitted to ICU from 2003 to 2023. ARDS was classified according to CT-scan as “Focal ARDS” or “Non-Focal ARDS.” The primary outcome was 90-day mortality. Treatment effects for prone position and high PEEP were assessed. Adjusted Hazard Ratios (aHR) were calculated with a multivariate Cox analysis. Machine-learning models were developed with XGBoost and explained with SHAP. RESULTS: Among the 5,865 patients screened, 151 Acute Pancreatitis-related ARDS were included in the analysis. Eighty-one patients (54%) were classified as Focal ARDS and 70 patients (46%) as Non-Focal ARDS. There was no significant difference in ARDS management. The overall 90-day mortality was of 37% (56/151). There was no significant difference in 90-day mortality between Focal and Non-Focal ARDS (33/81 [41%] vs 33/70 [34%], adjusted Hazard Ratio (aHR): 1.01 [0.58–1.76], p = 0.98). Prone position was associated with lower 90-day mortality in Focal ARDS (aHR: 0.25 [0.08 – 0.76], p = 0.015) but not in Non-Focal ARDS. High PEEP was associated with lower 90-day mortality in Non-Focal ARDS (aHR: 0.22 [0.06–0.85], p = 0.029) but not in Focal ARDS. Machin-learning models concurred with differential responses to prone position and PEEP. CONCLUSION: Two morphological subphenotypes of Acute Pancreatitis–related ARDS were identified and were evenly distributed. Although their clinical characteristics and outcomes were comparable, they exhibited treatment-specific responses: prone positioning was associated with lower mortality in Focal ARDS, whereas high PEEP was associated with lower mortality in Non-Focal ARDS. These findings suggest that morphology may help guide personalized ventilatory strategies in this specific ARDS population.
参考来源
- Pensier J, Touaibia M, Meerun MA, et al. Morphological subphenotypes of acute pancreatitis-related acute respiratory distress syndrome. Crit Care. 2026. PMID: 41546094. doi:10.1186/s13054-026-05837-w
供执业医师抢救与诊疗参考,不替代临床判断;剂量请结合体重、肝肾功能和本院药品规格复核,以指南原文为准。