联合应用老年营养风险指数和改良肌酐指数预测中老年维持性血液透析患者全因死亡的临床评价

Clinical evaluation of combined geriatric nutritional risk index and modified creatinine index predicting all-cause mortality in middle-aged and older patients undergoing maintenance hemodialysis

目的探讨老年营养风险指数(geriatric nutritional risk index, GNRI)和改良肌酐指数(modified creatinine index, mCI)与中老年维持性血液透析(maintenance hemodialysis, MHD)患者全因死亡的相关关系。方法该研究为前瞻性队列研究, 选取2017年4—6月北京11家医院血液透析中心年龄≥50岁的MHD患者作为研究对象。收集患者基线临床资料, 按照GNRI和mCI评分将患者分别分为GNRI升高组(≥98)和GNRI降低组(<98)、mCI升高组(≥20.16 mg·kg^(-1)·d^(-1))和mCI降低组(<20.16 mg·kg^(-1)·d^(-1))。联合GNRI和mCI将患者分为4组:G1组(GNRI升高、mCI升高)、G2组(GNRI升高、mCI降低)、G3组(GNRI降低、mCI升高)及G4组(GNRI降低、mCI降低), 比较4组患者基线临床资料的差异。随访截至2018年6月或患者死亡或失访, 随访终点事件为患者全因死亡。Kaplan-Meier生存曲线法比较各组间患者累积生存率的差异, 多因素Cox回归法分析GNRI和mCI与MHD患者全因死亡的相关性。结果共613例MHD患者被纳入该研究, 年龄(63.65±7.78)岁(范围50~81岁), 男性355例(57.91%), GNRI为(99.35±5.75), mCI为(20.16±2.79)mg·kg^(-1)·d^(-1)。G1组232例(37.85%), G2组177例(28.87%), G3组95例(15.50%), G4组109例(17.78%)。各组患者在年龄、性别分布、糖尿病比例、冠心病比例、体重指数、血清白蛋白及血肌酐等项目上的差异均有统计学意义(均P<0.05)。中位随访52(4, 52)周, 随访期间患者死亡共69例(11.26%)。Kaplan-Meier生存曲线分析结果显示, GNRI降低组患者死亡率高于GNRI升高组(Log-rankχ^(2)=26.956, P<0.001), mCI降低组患者死亡率亦高于mCI升高组(Log-rankχ^(2)=25.842, P<0.001)。联合分组G1组死亡率为3.45%, G2组为10.73%, G3组为9.47%, G4组为30.28%, 4组间差异有统计学意义(Log-rankχ^(2)=57.153, P<0.001)。多因素Cox回归分析结果显示, 当作为连续变量时, GNRI(HR=0.911, 95%CI 0.882~0.941, P<0.001)和mCI(HR=0.873, 95%CI 0.797~0.956, P=0.003)与MHD患者全因死亡相关;当作为分类变量时, 相比GNRI升高组和mCI升高组, GNRI降低组(HR=3.469, 95%CI 2.125~5.665, P<0.001)和mCI降低组(HR=3.255, 95%CI 1.879~5.640, P<0.001)患者的死亡风险均较高;相比G1组, G2组(HR=2.488, 95%CI 1.079~5.738, P=0.033)和G4组(HR=9.449, 95%CI 4.362~20.470, P<0.001)患者的死亡风险均较高。结论 GNRI和mCI是MHD患者全因死亡的独立预测因素, 联合应用GNRI和mCI可更精准地评估MHD患者的全因死亡风险。

Objective To explore the relationship between geriatric nutritional risk index(GNRI)and modified creatinine index(mCI)and all-cause mortality in maintenance hemodialysis(MHD)patients.Methods It was a prospective cohort study.The MHD patients aged≥50 years old at hemodialysis centers of eleven hospitals in Beijing from April to June 2017 were selected as subjects.Baseline clinical data of the patients were collected.The patients were divided into high GNRI group(≥98)and low GNRI group(<98),and high mCI group(≥20.16 mg·kg^(-1)·d^(-1))and low mCI group(<20.16 mg·kg^(-1)·d^(-1)),and further divided into 4 groups:G1 group(high GNRI and high mCI),G2 group(high GNRI and low mCI),G3 group(low GNRI and high mCI)and G4 group(low GNRI and low mCI).The differences of clinical characteristics among the four groups were compared.The patients were followed-up until June 2018 or death or loss,and the endpoint event was all-cause mortality.Kaplan-Meier survival analysis was used to compare the differences of the cumulative survival rates among the four groups.A multivariate Cox regression model was used to analyze the relationship between GNRI and mCI and all-cause mortality.Results A total of 613 patients were included in the study,aged(63.65±7.78)years old(ranged from 50 to 81 years old),with 355 males(57.91%).The GNRI and mCI were(99.35±5.75)and(20.16±2.79)mg·kg^(-1)·d^(-1),respectively.There were 232 patients(37.85%)in the G1 group,177 patients(28.87%)in the G2 group,95 patients(15.50%)in the G3 group,and 109 patients(17.78%)in the G4 group.There were statistically significant differences in age,sex,proportion of diabetes,proportion of coronary heart disease,body mass index,serum albumin and serum creatinine among the four groups(all P<0.05).A total of 69 patients(11.26%)died during a median follow-up time of 52(4,52)weeks.Kaplan-Meier survival curve results showed that the mortality of patients with low GNRI was higher than that of patients with high GNRI(log-rankχ^(2)=26.956,P<0.001),and the mortality of patients with low mCI was higher than that of patients with high mCI(log-rankχ^(2)=25.842,P<0.001).The mortality was 3.45%in group G1,10.73%in group G2,9.47%in group G3,and 30.28%in group G4,and the differences among the four groups were statistically significant(log-rankχ^(2)=57.153,P<0.001).Multivariate Cox regression analysis results showed that as continuous variables,GNRI(HR=0.911,95%CI 0.882-0.941,P<0.001)and mCI(HR=0.873,95%CI 0.797-0.956,P=0.003)were correlated with all-cause death.As categorical variables,compared with high GNRI group and high mCI group,patients with low GNRI(HR=3.469,95%CI 2.125-5.665,P<0.001)and low mCI(HR=3.255,95%CI 1.879-5.640,P<0.001)had higher risks of death.Compared with G1 group,patients in G2 group(HR=2.488,95%CI 1.079-5.738,P=0.033)and G4 group(HR=9.449,95%CI 4.362-20.470,P<0.001)had higher risks of death.Conclusions GNRI and mCI are independent predictive factors of all-cause mortality in MHD patients.The combination of GNRI and MCI can more accurately predict the risk of all-cause death in middle-aged and elderly MHD patients.