基于CAR、NLR的列线图模型在预测老年髋部骨折患者术后1年内死亡中的价值

doi:10.3969/j.issn.1673-8640.2025.09.002

摘要/Abstract

摘要：

目的探讨基于C反应蛋白/白蛋白比值（CAR）、中性粒细胞/淋巴细胞比值（NLR）的列线图模型在预测老年髋部骨折患者术后1年内死亡中的价值。方法选取2019年1月—2021年12月广西壮族自治区人民医院行髋关节置换术的老年髋部骨折患者161例。收集所有患者的临床资料、手术相关资料和术前、术后实验室指标检测结果。采用Cox回归分析评估老年髋部骨折患者术后1年内死亡的独立影响因素。构建预测老年髋部骨折患者术后1年内死亡的Cox死亡风险模型和术后3个月、6个月、1年生存率的列线图模型。采用受试者工作特征（ROC）曲线确定列线图模型总评分区分低危、中危、高危的最佳临界值。采用Kaplan-Meier生存曲线分析低危组、中危组和高危组术后1年的生存情况。结果年龄、查尔森合并症指数（CCI）、美国麻醉医师协会（ASA）分级、术后CAR和术后NLR均是老年髋部骨折患者术后1年内死亡的独立危险因素[风险比（HR）值分别为2.745、2.486、3.568、2.597、2.518，95%可信区间（CI）分别为1.377~5.993、1.274~3.738、1.148~6.993、1.126~3.957、1.343~3.889，P<0.01]。基于年龄、CCI、ASA分级、术后CAR和术后NLR建立预测老年髋部骨折患者术后1年内死亡的Cox死亡风险模型和术后3个月、6个月、1年生存率的列线图模型。老年髋部骨折患者术后CAR、术后NLR与Cox死亡风险得分呈正相关，与术后生存时间呈负相关。列线图模型判断老年髋部骨折患者术后1年内死亡的曲线下面积（AUC）为0.940。低危组、中危组和高危组术后1年总生存率依次降低（P<0.05）。结论基于CAR、NLR的预测老年髋部骨折患者术后1年内死亡的列线图模型具有较高的临床应用价值。

关键词: C反应蛋白/白蛋白比值, 中性粒细胞/淋巴细胞比值, 列线图模型, 髋部骨折, 老年人, 髋关节置换, 死亡

Abstract:

Objective To investigate the role of nomogram models based on C-reactive protein-to-albumin ratio （CAR） and neutrophil-to-lymphocyte ratio（NLR）in predicting 1-year mortality in elderly patients with hip fractures after surgery. Methods A total of 161 elderly patients with hip fractures who underwent hip replacement surgery at Guangxi Zhuang Autonomous Region People's Hospital from January 2019 to December 2021 were enrolled. The clinical data，surgical-related data and preoperative and postoperative laboratory determination results were collected. Cox regression analysis was used to evaluate the independent risk factors for 1-year mortality in elderly patients with hip fractures after surgery. Cox mortality risk models and nomogram models for predicting 1-year mortality and postoperative 3-month，6-month and 1-year survival rates in elderly patients with hip fractures after surgery were established. Receiver operating characteristic（ROC）curve was used to determine the optimal cut-off value for the total score of the nomogram models to distinguish low-risk，medium-risk and high-risk groups. Kaplan-Meier survival curves were used to analyze the 1-year survival of low-risk，medium-risk and high-risk groups after surgery. Results Age，Charlson comorbidity index（CCI），American Society of Anesthesiologists（ASA）classification，postoperative CAR and postoperative NLR were all independent risk factors for 1-year mortality in elderly patients with hip fractures after surgery [hazard ratios（HR）were 2.745，2.486，3.568，2.597 and 2.518，95% confidence intervals（CI）were 1.377-5.993，1.274-3.738，1.148-6.993，1.126-3.957 and 1.343-3.889，respectively，P<0.01]. Cox mortality risk models and nomogram models for predicting 1-year mortality and 3-month，6-month and 1-year survival in elderly patients with hip fractures after surgery were established based on age，CCI，ASA classification，postoperative CAR and postoperative NLR. Postoperative CAR and postoperative NLR in elderly patients with hip fractures were positively correlated with Cox mortality risk scores. It is negatively correlated with postoperative survival time. The area under curve（AUC）for predicting the mortality of elderly hip fracture patients within 1-year after surgery by the combined determination of the 5 factors was 0.940. The 1-year overall survival rates of low-risk，medium-risk and high-risk groups were decreased successively（P<0.05）. Conclusions The nomogram models based on CAR and NLR for predicting the mortality of elderly hip fracture patients within 1-year after surgery has high clinical application value.

Key words: C-reactive protein-to-albumin ratio, Neutrophil-to-lymphocyte ratio, Nomogram model, Hip fracture, Elder, Hip replacement, Death

中图分类号:

R446.1

楚野, 熊春翔, 黄宇, 曾佳兴, 易波德, 黄能干, 杨屹峰, 杨枫. 基于CAR、NLR的列线图模型在预测老年髋部骨折患者术后1年内死亡中的价值[J]. 检验医学, 2025, 40(9): 834-840.

CHU Ye, XIONG Chunxiang, HUANG Yu, ZENG Jiaxing, YI Bode, HUANG Nenggan, YANG Yifeng, YANG Feng. Role of nomogram models based on CAR and NLR in predicting 1-year mortality in elderly patients with hip fractures after surgery[J]. Laboratory Medicine, 2025, 40(9): 834-840.

图/表 7

表1 老年髋部骨折患者术后1年内死亡的Cox回归分析

项目	单因素Cox回归分析					多因素Cox回归分析
项目	β值	标准误	Wald值	P值	HR值（95%CI）	β值	标准误	Wald值	P值	HR值（95%CI）
年龄	1.173	0.237	12.677	0.000	3.474（1.563~5.663）	1.036	0.209	9.823	0.003	2.745（1.377~5.993）
入院Braden评分	0.615	0.218	7.276	0.010	1.895（1.335~3.869）
CCI	0.843	0.256	11.875	0.000	2.593（1.168~4.774）	0.781	0.211	7.639	0.009	2.486（1.274~3.738）
术前BNP	0.675	0.177	4.234	0.039	2.147（1.378~3.889）
ASA分级	1.362	0.339	13.409	0.000	3.984（1.239~6.675）	1.259	0.345	10.235	0.001	3.568（1.148~6.993）
术后 SCr	0.728	0.187	5.896	0.021	2.245（1.158~3.967）
术后CAR	0.865	0.229	10.324	0.001	2.656（1.143~4.375）	0.847	0.225	8.125	0.006	2.597（1.126~3.957）
术后NLR	0.857	0.178	11.488	0.000	2.627（1.248~4.265）	0.823	0.207	7.838	0.007	2.518（1.343~3.889）

图1 预测老年髋部骨折患者术后1年内死亡的Cox死亡风险模型注：（a）Cox死亡风险值；低危组、高危组；（b）老年髋部骨折患者按Cox死亡风险得分由低到高排序后术后生存时间分析；存活、死亡；（c）将年龄、CCI、ASA分级、术后CAR和术后NLR标准化后与Cox死亡风险值的相关性热图；低危组、高危组。

图2 预测老年髋部骨折患者术后3个月、6个月、1年生存率的列线图模型

图3 各项指标预测老年髋部骨折患者术后1年内死亡的ROC曲线注： ASA分级； CCI；列线图模型；术后CAR；术后NLR；年龄；参考线。

图4 校准曲线注：实际曲线；校正曲线；理想曲线。

图5 列线图模型预测老年髋部骨折患者术后1年内死亡的决策曲线注：列线图模型；所有患者均死亡；所有患者均存活。

图6 低危组、中危组和高危组术后1年总生存率注：低危组；中危组；高危组。

参考文献 18

[1]	SHARROCK M, HAGAN J, LEE J, et al. Greater length of hospital stay for concurrent hip and upper limb fractures compared to isolated hip fractures:a systematic review of 13 studies including 210 289 patients and meta-analysis[J]. Injury, 2022, 53(7):2617-2624.
[2]	王成刚, 麻彬, 施静, 等. 老年髋部骨折术后死亡的危险因素[J]. 中国矫形外科杂志, 2022, 30(4):289-293.
[3]	王旭, 胡金龙, 张亮, 等. 基于CRP/Hb、TLC、Na+构建老年髋部骨折术后死亡风险的预测模型[J]. 河北医科大学学报, 2022, 43(5):544-549.
[4]	吴天旺, 李永威, 张李宝, 等. 老年髋部骨折手术患者术后1年内死亡风险预测模型建立与验证[J]. 中国老年学杂志, 2022, 42(22):5504-5507.
[5]	TEKIN S B, BOZGEYIK B, MERT A. Relationship between admission neutrophil/lymphocyte,thrombocyte/lymphocyte,and monocyte/lymphocyte ratios and 1-year mortality in geriatric hip fractures:triple comparison[J]. Ulus Travma Acil Cerrahi Derg, 2022, 28(11):1634-1640.
[6]	赵建文, 王晓伟, 孙天胜, 等. 老年髋部骨折术后1年死亡风险及危险因素分析[J]. 实用骨科杂志, 2020, 26(5):399-403.
[7]	王晓伟, 孙天胜, 刘智, 等. 老年髋部骨折患者术后死亡原因分析[J]. 中国骨与关节损伤杂志, 2021, 36(10):1048-1050.
[8]	王金强, 骆洪涛, 彭城, 等. 极高龄髋部骨折患者术后一年病死的危险因素分析[J]. 中国骨与关节杂志, 2019, 8(3):187-190.
[9]	缪旭东, 阮张涛, 邱耿涛, 等. 不同时间点中性粒细胞/淋巴细胞比率与老年髋部骨折患者术后死亡率和并发症的相关性研究[J]. 国际老年医学杂志, 2021, 42(2):91-94.
[10]	LACKNER I, WEBER B, PRESSMAR J, et al. Cardiac alterations following experimental hip fracture-inflammaging as independent risk factor[J]. Front Immunol, 2022, 13:895888.
[11]	LUO T, ZHANG J, ZHOU H, et al. Identification of risk factors for 1-year mortality among critically ill older adults with hip fractures surgery:a single medical center retrospective study[J]. Front Surg, 2022, 9:973059.
[12]	ZHU H, LI Z, ZHOU Y, et al. Neutrophil-lymphocyte ratio as a risk factor for osteoporotic vertebrae fractures and femoral neck fractures[J]. Medicine(Baltimore), 2022, 101(48):e32125.
[13]	SONG B W, KIM A R, MOON D H, et al. Associations of neutrophil-to-lymphocyte ratio,platelet-to-lymphocyte ratio and monocyte-to-lymphocyte ratio with osteoporosis and incident vertebral fracture in postmenopausal women with rheumatoid arthritis:a single-center retrospective cohort study[J]. Medicina(Kaunas), 2022, 58(7):852.
[14]	BALTA O, ALTINAYAK H, GÜRLER BALTA M, et al. Can C-reactive protein-based biomarkers be used as predictive of 30-day mortality in elderly hip fractures ? A retrospective study[J]. Ulus Travma Acil Cerrahi Derg, 2022, 28(6):849-856.
[15]	李雅欣, 冯德宏, 郭宇, 等. CRP/Alb比值对老年髋部骨折患者术后1年内死亡的预测价值分析[J]. 创伤外科杂志, 2021, 23(6):441-445.
[16]	KIM H J, LEE S, KIM S H, et al. Association of C-reactive protein to albumin ratio with postoperative delirium and mortality in elderly patients undergoing hip fracture surgery:a retrospective cohort study in a single large center[J]. Exp Gerontol, 2023, 172:112068.
[17]	ZHU J, CHENG X, LI Y, et al. Low lymphocyte-to-C-reactive protein ratio relates to high 1-year mortality in elderly patients undergoing hemiarthroplasty for displaced femoral neck facture[J]. J Orthop Surg Res, 2022, 17(1):512. DOI PMID
[18]	ANDERSON P M, RUDERT M, HOLZAPFEL B M, et al. Conversion total hip arthroplasty following proximal femur fracture:a retrospective analysis[J]. Technol Health Care, 2023, 31(2):507-516.