Zhi Yongjin, Bao Shuojing, Mao Jingcheng, Zhu Hui, Zhu Jianfeng
Department of Hematology, Taizhou People's Hospital Affiliated to Nanjing Medical University, Taizhou, People's Republic of China.
Department of General Practice, Zhangdian People's Hospital, Taizhou, People's Republic of China.
Hematology. 2025 Dec;30(1):2536402. doi: 10.1080/16078454.2025.2536402. Epub 2025 Jul 22.
The present study systematically explores the correlation between the Prognostic Nutritional Index (PNI) and anemia, and further analysis of its association with all-cause mortality among populations affected by anemia.
Data were taken from the National Health and Nutrition Examination Survey (NHANES) from 2005 to 2018. We utilized multiple logistic regression models, Cox proportional hazards models, restricted cubic spline (RCS) analysis, and time-dependent receiver operating characteristic (ROC) analysis. As for the verification of the result robustness, subgroup and sensitivity analyses were implemented.
A cohort of 28,511 participants was examined, of whom 2647 (9.28%) had anemia. An inverse relationship of PNI with anemia was detected in the fully adjusted model (OR = 0.82, 95%CI: 0.80-0.84). Cox regression showed that lower PNI levels were linked to higher all-cause mortality in people with anemia (HR = 0.91, 95%CI: 0.88-0.95). A non-linear correlation between PNI and mortality was detected through RCS analysis ( < 0.001). Significant interaction effects of PNI with mortality were observed across diabetes and BMI subgroups ( < 0.05). The predictive power of PNI for all-cause mortality among anemia individuals showed areas under the curve (AUC) values of 0.702, 0.806, and 0.813 for 3-, 5-, and 10-year predictions, respectively.
PNI demonstrates a negative association with anemia and a similar negative relationship with all-cause mortality in individuals with anemia. Future studies are warranted to substantiate these findings.
本研究系统探讨了预后营养指数（PNI）与贫血之间的相关性，并进一步分析其与贫血人群全因死亡率的关联。
数据取自2005年至2018年的美国国家健康与营养检查调查（NHANES）。我们使用了多元逻辑回归模型、Cox比例风险模型、受限立方样条（RCS）分析和时间依赖性受试者工作特征（ROC）分析。为验证结果的稳健性，进行了亚组分析和敏感性分析。
共检查了28511名参与者，其中2647人（9.28%）患有贫血。在完全调整模型中检测到PNI与贫血呈负相关（OR = 0.82，95%CI：0.80 - 0.84）。Cox回归显示，贫血患者中较低的PNI水平与较高的全因死亡率相关（HR = 0.91，95%CI：0.88 - 0.95）。通过RCS分析检测到PNI与死亡率之间存在非线性相关性（< 0.001）。在糖尿病和BMI亚组中观察到PNI与死亡率之间存在显著的交互作用（< 0.05）。PNI对贫血个体全因死亡率的预测能力在3年、5年和10年预测中的曲线下面积（AUC）值分别为0.702、0.806和0.813。
PNI与贫血呈负相关，在贫血个体中与全因死亡率也呈类似的负相关。未来的研究有必要证实这些发现。
