Abstract:

Objective　To establish and validate a predictive model for septic shock cardiomyopathy based on the shock index (SI) and cardiac magnetic resonance feature tracking (FT-CMR) parameters, aiming to provide clinical guidance for reducing the incidence of cardiomyopathy in patients with septic shock. Methods　A retrospective analysis was conducted on the clinical data of 137 patients with septic shock admitted to Xi'an Ninth Hospital from May 2020 to October 2023. Patients were divided into a cardiomyopathy group (42 cases) and a non-cardiomyopathy group (95 cases) based on the presence of cardiomyopathy within 72 hours of enrollment. In the cardiomyopathy group, there were 23 males and 19 females, aged (66.52±5.54) years and a body mass index (BMI) of (22.77±1.18) kg/m². In the non-cardiomyopathy group, there were 48 males and 47 females, aged (58.74±4.89) years and a BMI of (22.48±1.16) kg/m². All patients underwent FT-CMR examination and SI assessment at enrollment. Factors influencing the occurrence of cardiomyopathy were analyzed, and a nomogram was constructed to predict the risk of cardiomyopathy. The predictive efficacy of the model for the occurrence of cardiomyopathy was analyzed using receiver operating characteristic (ROC) curves. Statistical analyses were performed using the χ² test and t test. Results　Univariate analysis showed that the age, cardiac troponin (cTnI), lactic acid, N-terminal pro-brain natriuretic peptide (NTpro-BNP) and SI levels in the cardiomyopathy group were higher than those in the non-cardiomyopathy group, while the total left atrial strain (Es), overall longitudinal strain of the left atrium (PLAS), and peak positive strain rate (SRs) were lower in the cardiomyopathy group (all P<0.05). Binary logistic regression analysis revealed that age [odds ratio (OR)=2.769, 95% confidence interval (CI) 1.151-6.663], lactate (OR=4.661, 95% CI 2.645-8.212), cTnI (OR=6.234, 95% CI 3.096-12.550), and SI (OR=4.165, 95% CI 1.526-11.361) were independent risk factors for the development of cardiomyopathy in patients with septic shock, while PLAS (OR=0.186, 95% CI 0.081-0.423) was an independent protective factor. The nomogram prediction model constructed based on these influencing factors showed a C-index of 0.821 (95% CI 0.743-0.925) upon internal validation using the Bootstrap method, with the calibration curve for predicting the occurrence of cardiomyopathy approaching the ideal curve (P>0.05). ROC analysis indicated that the nomogram model had a sensitivity of 85.70% and specificity of 86.60%, with an area under the curve (AUC) of 0.872 (95% CI 0.783-0.960). Conclusion　SI is an independent risk factor for the development of cardiomyopathy in patients with septic shock, and PLAS is an independent protective factor. The nomogram prediction model based on SI and PLAS can better evaluate the risk of cardiomyopathy in patients with septic shock.

Key words:

Septic shock, Cardiomyopathy, Shock index, Cardiac magnetic resonance feature tracking technology, Cardiac function parameters,  , Nomogram, Prediction model

摘要：

目的　基于休克指数（SI）和心脏磁共振特征追踪技术（FT-CMR）的心脏功能评估参数建立脓毒症休克心肌病的预测模型并进行验证，为降低脓毒症休克患者心肌病发生风险提供指导依据。方法　回顾性分析西安市第九医院2020年5月至2023年10月收治的137例脓毒症休克患者临床资料，根据患者入组72 h内是否发生心肌病将其分为心肌病组（42例）和无心肌病组（95例）。心肌病组中男23例，女19例；年龄（66.52±5.54）岁；体重指数（22.77±1.18）kg/m²。无心肌病组中男48例，女47例；年龄（58.74±4.89）岁；体重指数（22.48±1.16）kg/m²。所有患者在入组时均行FT-CMR检查和SI评估。分析影响患者发生心肌病的因素，并以此构建Nomogram列线图模型预测心肌病的发生风险，采用受试者操作特征曲线（ROC）分析预测模型对心肌病发生的预测效能。采用χ²、t检验进行统计分析。结果　单因素分析结果显示，心肌病组的年龄、心肌肌钙蛋白（cTnI）、乳酸、氨基末端B型脑钠肽前体（NT-proBNP）、SI均高于无心肌病组，左心房总应变（Es）、整体纵向应变（PLAS）、正向应变率峰值（SRs）均低于无心肌病组（均P<0.05）。二元logistic回归分析结果显示，年龄［比值比（OR）=2.769，95%置信区间（CI）1.151～6.663］、乳酸（OR=4.661，95%CI 2.645～8.212）、cTnI（OR=6.234，95%CI 3.096～12.550）、SI（OR=4.165，95%CI 1.526～11.361）是脓毒症休克患者发生心肌病的独立危险因素，PLAS（OR=0.186，95%CI 0.081～0.423）是独立保护因素。基于上述影响因素构建的列线图预测模型经Bootstrap法内部验证结果显示，C-index指数为0.821（95%CI 0.743～0.925），预测患者发生心肌病的校正曲线趋近于理想曲线（P>0.05）。ROC结果显示，列线图模型预测患者发生心肌病的灵敏度为85.70%、特异度为86.60%，曲线下面积（AUC）为0.872（95%CI 0.783～0.960）。结论　SI为影响脓毒症休克患者发生心肌病的独立危险因素，PLAS为独立保护因素，基于SI和PLAS建立的列线图预测模型可较好地评估脓毒症休克患者心肌病的发生风险。

关键词:

脓毒症休克, 心肌病, 休克指数, 心脏磁共振特征追踪技术, 心脏功能参数, 列线图, 预测模型