目的 探讨预后营养指数（PNI）对脓毒症合并心力衰竭患者发生谵妄的预测价值，并评估其在指导护理风险分层中的作用。方法 采用回顾性队列研究，纳入2022年2月—2024年9月广东省第二中医院收治的162例脓毒症合并心力衰竭患者。通过ROC曲线确定PNI预测谵妄的最佳截断值，并以此将患者分为高低PNI组。比较两组基线特征，采用Kaplan-Meier曲线分析谵妄累积发生率和28天院内死亡率，通过Cox比例风险模型评估PNI与谵妄及预后的独立关联，并利用ROC曲线评价其预测性能。最后利用亚组分析技术进一步评估PNI预测价值的普适性。结果 低PNI（≤35.72）是谵妄发生的独立危险因素。竞争风险模型显示，低PNI组10 d谵妄累积发生率达37.8%，高于高PNI组的10.0%。多因素分析表明，PNI每降低一个单位，谵妄风险增加约17.2%。限制性立方样条分析提示PNI与谵妄风险呈线性负相关。ROC曲线分析显示PNI预测谵妄的AUC为0.721，其预测效能优于白蛋白和SOFA评分。亚组分析进一步证实PNI的保护效应在不同临床特征患者中具有一致性。结论 PNI是脓毒症合并心力衰竭患者谵妄风险的独立预测指标，其预测效能优于传统单一指标。基于PNI的风险分层有助于早期识别谵妄高危患者，为实现精准护理干预提供重要依据。

      Objective To explore the predictive value of the prognostic nutritional index（PNI）for delirium in patients with sepsis complicated by heart failure，and to evaluate its role in guiding nursing risk stratification．Methods A retrospective cohort study was conducted，including 162 patients with sepsis and heart failure admitted to Guangdong Provincial Second Hospital of Traditional Chinese Medicine between February 2022 and September 2024．The optimal PNI cut-off value for predicting delirium was determined using ROC curve analysis，and patients were divided into high and low PNI groups accordingly．Baseline characteristics of the two groups were compared．Kaplan-Meier curves were used to analyze the cumulative incidence of delirium and 28-day in-hospital mortality．Cox proportional hazards models were employed to assess the independent association between PNI and delirium as well as prognosis，and ROC curves were used to evaluate its predictive performance．Subgroup analysis was further utilized to assess the generalizability of PNI’s predictive value．Results A low PNI（≤35.72）is an independent risk factor for delirium．The competing risk model showed that the cumulative incidence of delirium in the low PNI group was 37.8% at 10 days，significantly higher than the 10.0% in the high PNI group．Multivariate analysis indicated that for each unit decrease in PNI，the risk of delirium increased by 17.2% approximately．Restricted cubic spline analysis suggested a linear negative correlation between PNI and delirium risk．ROC curve analysis demonstrated that the AUC of PNI for predicting delirium was 0.721，with its predictive performance surpassing that of albumin and SOFA score．Subgroup analysis further confirmed that the protective effect of PNI was consistent across patients with different clinical characteristics．Conclusions PNI is an independent predictor of delirium risk in patients with sepsis and heart failure，with superior predictive performance compared to traditional single indicators．Risk stratification based on PNI aids in the early identification of patients at high risk for delirium，providing an important basis for precise nursing interventions．