Comparison Of The Predictive Value Of Three Glasgow Prognostic Scores For The Risk Of All-cause Mortality In Patients With AHF

ObjectTo compare the predictive value of Glasgow Prognostic Scale(GPS),Modified Glasgow Prognostic Scale(mGPS)and Hypersensitive Modified Glasgow Prognostic Scale(Hs-mGPS)for all-cause mortality risk in patients with acute heart failure(AHF).The differences in the predictive value of the three scores in patients with different ejection fractions(LVEF)were compared.MethodA retrospective analysis of 168 patients with acute heart failure admitted to the heart failure specialist ward of Shantou Central Hospital from January 2019 to December 2019.The clinical data,laboratory indicators on admission and other data of patients with acute heart failure were collected.Grouping method: Three different GPS scores were used to classify patients into GPS = 0,1,2,mGPS = 0,1,2,and Hs-mGPS = 0,1,2 points.Univariate and multivariate Logistic regression analysis was used to analyze the independent risk factors for death in AHF patients,and then univariate and multivariate COX regression analysis was performed to analyze the independent risk factors for AHF patients’ mortality within 1 year after admission.The prediction accuracy of the three GPS scores was verified by discriminant analysis,and the predictive value of the three GPS scores on the risk of all-cause mortality within 1 year after admission in AHF patients was compared by receiver operating characteristic curve(ROC)analysis.The patients were divided into three groups:HFp EF,HFmr EF,and HFr EF according to different LVEF.The ROC curve was used to analyze the predictive value of the three GPS scores on the risk of all-cause mortality within 1 year after admission in patients with different LVEF.Result1.The patients were divided into GPS group,mGPS group and Hs-mGPS group according to different scores.In the GPS group,the higher the score,the lower the percentage of lymphocytes,plasma albumin,hemoglobin,and blood sodium concentration(P < 0.05);TNT,CRP,blood urea nitrogen,and EF values were higher(P < 0.05).In the mGPS score,the higher the score,the lower the lymphocyte percentage,plasma albumin,and blood sodium concentration(P < 0.05),and the higher the TNT,CRP,blood urea nitrogen,and EF value(P < 0.05).In the Hs-mGPS score,the higher the score,the lower the lymphocyte percentage,hemoglobin content,and plasma albumin(P < 0.05);the higher the TNT,CRP,serum creatinine,blood urea nitrogen,and EF values(P < 0.05)2.Logistic regression analysis found that in the GPS group,patients with a score of 2 had a 17.78 times higher risk of all-cause mortality than those with a score of 0(OR 17.78,95 % CI 5.16-61.20,P < 0.001).In the mGPS group,patients with a score of 2 had a 13.14 times higher risk of all-cause mortality than those with a score of 0(OR 13.14,95 % CI 4.735-36.466,P < 0.001).In the Hs-mGPS group,patients with a score of 2 had a 10.579 times higher risk of all-cause mortality than those with a score of 0(OR 10.579,95 % CI 2.808-39.846,P < 0.001).Age,lymphocyte percentage,and edema on admission were independent risk factors for all-cause mortality in patients with AHF.3.Cox regression analysis found that in the GPS group,for patients with a score of 2,the one-year all-cause mortality rate was 7.815 times that of patients with a score of 0(HR 7.815,95 % CI 2.928-20.855,P < 0.001).In the mGPS group,all-cause mortality at 1 year was 5.83 times higher for patients with a score of 2 than for those with a score of 0(HR 5.83,95 % CI 2.888-11.768,P < 0.001).In the Hs-mGPS group,all-cause mortality at 1 year was 7.122 times higher for patients with a score of 2 than for those with a score of 0(HR 7.122,95 % CI 2.117-23.957,P = 0.002).Lymphocyte percentage and edema on admission were independent factors affecting all-cause mortality in AHF patients within 1 year after admission.4.Verified by discriminant analysis,the overall prediction accuracy of GPS,mGPS and Hs-mGPS scores for AHF patients within 1 year after admission was75.0 %,76.2 %,and 73.2 %,respectively.5.The area under the ROC curve(AUC)of GPS(0.767,95 % CI 0.693–0.841,P< 0.001)in predicting the risk of all-cause mortality in AHF patients within 1 year was not significantly different from that of mGPS(0.768,95 % CI 0.693–0.845),P <0.001)and Hs-mGPS(0.747,95 % CI 0.672–0.822,P < 0.001).6.In the HFp EF group,the AUC of GPS(0.815,95% CI 0.73-0.899,P = 0.043)was greater than that of mGPS(0.805,95% CI 0.714-0.896,P = 0.046)and Hs-mGPS(0.741,95% CI 0.643-0.84,P = 0.05).In the HFmr EF and HFr EF groups,AUC for Hs-mGPS(0.733,95% CI 0.571-0.895,P = 0.022)、(0.761,95% CI 0.589-0.934,P =0.03)were greater than GPS(0.634,95% CI 0.442-0.827,P = 0.186)、(0.727,95% CI0.53-0.925,P = 0.06)and mGPS(0.673,95% CI 0.485-0.86,P = 0.088)、(0.722,95%CI 0.511-0.932,P = 0.06).Conclusion1.GPS,mGPS and Hs-mGPS have certain predictive value for the risk of all-cause mortality in AHF patients within 1 year after admission,and the predictive abilities of the three scores are basically the same.2.The value of GPS in predicting the risk of all-cause mortality in HFp EF patients within 1 year is higher than that of Hs-mGPS and mGPS;while the value of Hs-mGPS in predicting the risk of all-cause mortality in HFmr EF and HFr EF patients in 1 year is higher than that of GPS and mGPS.