| Objective:The purpose of this study was to investigate the relationship between baseline lymphocyte count and 5-year all-cause mortality in patients with chronic heart failure(CHF).In addition,the predictive value of lymphocyte count combined with NT-proBNP in 5-year all-cause mortality of CHF patients was further analyzed in order to provide reference for clinical practice.Methods:This study was a retrospective cohort study.362 CHF patients hospitalized in the first affiliated Hospital of Guangxi Medical University from July 2012 to July 2013 were selected as subjects.Venous blood samples were taken from all patients within 24 hours after admission.According to the lymphocyte count on admission,all patients were divided into four groups(Quartile 1:0.3 to 1.2 to 10~9/L,Quartile 2:1.2 to 1.6 to 10~9/L).Quartile 3:1.6-2.1-10~9/L;Quartile 4:2.1-3.4-10~9/L).All patients were followed up for 5years,and the endpoint event of this study was 5-year all-cause mortality.The results were compared by odds ratio(OR)and 95%confidence interval(CI).The dimension reduction and feature selection were carried out by lasso(LASSO)regression model,and the confounding factors were adjusted by multivariate logstic regression analysis.In further analysis,curve fitting was used to evaluate whether there was a linear relationship between lymphocyte count and 5-year all-cause mortality in CHF patients.In addition,the correlation between lymphocyte count and NT-proBNP was discussed by pearson correlation analysis,and receiver operating characteristic curve(ROC)to analyze the predictive value of lymphocyte count,NT-proBNP and their combination in 5-year all-cause mortality in CHF patients.Results:During the 5-year clinical follow-up,195 patients died,and the5-year all-cause mortality rate was 53.87%(195/362).The 5-year all-cause mortality rates in the four groups were 66.29%(59/89),61.25%(49/80),43.88%(43/98)and 46.32%(44/95),respectively(P for trend<0.004).Univariate analysis showed that there was a negative correlation between lymphocyte count and 5-year all-cause mortality in CHF patients(Quartile 2 vs Quartile 1:OR=0.80,95%CI:0.43 to 1.51;Quartile 3 vs Quartile 1:OR=0.40,95%CI:0.22 to 0.72;Quartile 4 vs Quartile 1:OR=0.44,95%CI:0.24 to 0.80).Lasso regression analysis showed that there was a non-zero coefficient relationship between eleven variables and 5-year all-cause mortality in CHF patients.After adjusting the selected confounding factors by multivariate logstic regression,lymphocyte count was still negatively correlated with 5-year all-cause mortality in CHF patients(Quartile 2 vs Quartile 1:OR=0.93,95%CI:0.48 to 1.79;Quartile 3 vs Quartile 1:OR=0.41,95%CI:0.22 to 0.77;Quartile 4 vs Quartile1:OR=0.52,95%CI:0.28 to 0.99;P for trend<0.05).The curve fitting showed that there was a nonlinear negative correlation between lymphocyte count and 5-year all-cause mortality in CHF patients.Pearson correlation analysis showed that there was a negative correlation between lymphocyte count and NT-proBNP(log10 transform)(r=-0.486),that is,with the increase of lymphocyte count,NT-proBNP showed a downward trend.The results of ROC curve showed that the AUC of 5-year all-cause mortality rate of CHF patients diagnosed by lymphocyte count was 0.608,95%CI was 0.551 to 0.666,the sensitivity was 0.466,and the specificity was 0.742.At this time,the best lymphocyte threshold was 1.35*10~9/L.The AUC of 5-year all-cause mortality of CHF patients diagnosed by NT-proBNP was 0.631,95%CI was 0.569 to 0.693,the sensitivity was 0.494,the specificity was 0.716,and the best NT-proBNP threshold was 3271(pg/ml).The AUC of 5-year all-cause mortality in CHF patients diagnosed by lymphocyte count combined with NT-proBNP was 0.740,95%CI was 0.667 to 0.813,the sensitivity was 0.659,and the specificity was0.770.Conclusion:Our study showed that there was a nonlinear negative correlation between lymphocyte count and 5-year all-cause mortality in CHF patients,that is,with the decrease of lymphocyte count,the 5-year all-cause mortality rate in CHF patients increased.In addition,the lymphocyte count was also negatively correlated with NT-proBNP,that is,with the increase of lymphocyte count,the NT-proBNP of CHF patients showed a downward trend.On the other hand,the optimal threshold of lymphocyte count for diagnosing5-year all-cause mortality in CHF patients is 1.35*10~9/L,and NT-proBNP for diagnosing 5-year all-cause mortality in CHF patients is 3271(pg/ml).The combination of the two variables can significantly improve the AUC,sensitivity and specificity of predicting 5-year all-cause mortality in CHF patients. |
PDF Full Text Request