Risk Assesment Of Death In Patients With Chronic Heart Failure Within 1 Year

Objective To explore the risk factors that affect one-year all-cause mortality in patients with chronic heart failure,and establish a scale for predicting mortality in patients with chronic heart failure.Methods The clinical data of 820 patients diagnosed with CHF in First People's Hospital of Changde City from January 1,2017 to December 30,2018 were analyzed.All the patients were divided into death group(131 cases)and non-death group(689 cases)according to whether patients died within one year.1.Baseline demographic data,basic diseases,laboratory indexes(including blood routine,renal function,N terminal brain natriuretic peptide precursor,albumin,etc.),conventional electrocardiogram(heart rate,QRS duration and correction qt interval,etc.),echocardiography(left atrial,left ventricular end-diastolic diameter and left ventricular ejection fraction,etc.)were compared in two groups.multivariate logistic regression analysis were performed for risk factor affecting all-cause death in CHF patients within 1 year.2.The evaluation value of selected continuous independent risk factors were analyzed by ROC diagnostic test.The adjacent integers of the cut-off value of the optimal ROC diagnostic test parameters were used as the boundary points to convert the continuity variables into dichotomies.The risk factors of dichotomies were analyzed by multivariate logistic regression analysis again,then building a prediction model and carrying out the internal validation.3.The ?-factor assignment of binary risk factors is converted into an integer as the score of the predictive factor,and then a score table is produced.4.The score table were used to evaluate the consistency of the calibration curve and the degree of differentiation of the C index respectively.When the C index analysis is performed,the best score cut-off value for the death prediction of CHF patients within 1 year will be obtained.The cut-off value can be used as the risk score for predicting death within 1 year for CHF patients.P<0.05 was considered statistic.Statistical analysis was performed using SPSS 19.0 statistical software and R statistical software,and the ROC curve was plotted using R statistical software.Results 1.The comparison between the dead group(131 cases)and the non-dead group(689 cases)showed that,(1)Two groups in gender,heart rate,systolic blood pressure,diabetes,chronic obstructive pulmonary disease,chronic renal insufficiency,percutaneous coronary intervention or coronary artery bypass surgery,white blood cell count,hematocrit,platelet count,and average platelet volume,Total cholesterol,high-density lipoprotein,low-density lipoprotein,calcium ion,sodium ion,QT interval adjusted for heart rate,and bundle branch block were not statistically different(P>0.05).(2)Age,NYHA cardiac function grade,atherosclerotic vascular events,anemia,atrial fibrillation,Logarithmic value of N-terminal Pro-brain Natriuretic Peptide(Ln NT-pro BNP),Neutrophil to lymphocyte ratio,erythrocyte distribution width-standard deviation,coefficient of variation of erythrocyte distribution width(RDW-CV),blood urea nitrogen(BUN),creatinine,uric acid,lactate dehydrogenase,myoglobin,potassium,QRS duration is greater than or equal to 110ms(QRSd?110ms),Lown's classification is greater than or equal to ?(Lown's classification??),left atrium,left ventricular end-diastolic diameter,aortic valve moderate-to-severe reflux,mitral valve moderate-to-severe reflux,and tricuspid valve moderate-to-severe reflux in the death group were higher than those in the non-death group,and the differences were statistically significant.(P<0.05).Hypertension,hyperlipidemia,lymphocyte count,lymphocyte percentage,hemoglobin,albumin,triglycerides,and left ventricular ejection fraction were lower than those in the non-dead group.(P<0.05).(3)In terms of treatment,the use of spironolactone,furthamil in the death group were higher than those in the non-death group(P<0.05).There was no significant difference between the two groups inthe use of aspirin,statins,beta-blockers,Angiotensin converting enzyme in hibitor or angiotensin receptor inhibitor and digoxin(P>0.05).2.Multivariate logistic regression analysis and step-by-step forward analysis affecting all-cause death in chronic heart failure showed that age,NYHA cardiac function class ?-?,Ln NT-pro BNP,lymphocyte percentage,RDW-CV,BUN,Lown's classification?? and QRSd?110ms are affecting CHF independent risk factors for all-cause death within 1 year(P<0.05).3.The above-mentioned continuous risk factors were transformed into two-category variables which were performed multi-factor logistic regression analysis again.It shows that age is greater than or equal to 75 years(?=0.77,OR=2.16,95%CI:1.43-3.25),NYHA cardiac function class ?-?(?=0.81,OR=2.25,95%CI:1.21-4.17),Lown's classification??(?=0.58,OR=1.79,95%CI: 1.09-2.91),Ln NT-pro BNP?9(?=0.85,OR=2.34,95%CI:1.53-3.58),lymphocyte percentage?18%(?=0.47,OR=1.60,95%CI:1.06-2.41),RDW-CV?15%(?=0.62,OR=1.85,95%CI:1.21-2.82),BUN?10mmol/L(?=0.51,OR=1.67,95%CI:1.06-2.65),QRSd?110ms(?=0.45,OR=1.56,95%CI:1.00-2.43)were independent risk factor affecting all-cause death in patients with CHF within 1 year(P <0.05).4.The internal verification results of this prediction model show thatthe area under the ROC curve of the modeled population is 0.774,In the verification group,the area is 0.763,indicating that this prediction model has a certain extrapolation.5.Based on the beta coefficients of the two classification variabl es,a risk score table was made.NYHA cardiac function class ?-? a nd Ln NT-pro BNP?9 as 2 points respectively,age is greater than or equal to 75 years,Lown's classificatio??,RDW-CV?15%,lymphocyt e percentage18%,BUN?10mmol/L,QRSd?110ms take 1 point respect ively,constitute a score sheet with a total score of 10 points.6.The evaluation of calibration of the risk score table shows that the calibration curve of the risk score table is close to the standard line,which indicates that the risk score table has a good calibration ability.This score sheet has a discrimination evaluation showing that an area under the ROC curve of 0.771 and a total score of 5 is the best cutoff value of the score sheet.At this time,the score sheet has a sensitivity of 67%,a specificity of76%,There is a certain degree of risk prediction efficiency.Conclusion 1.The findings of this study,Age,NYHA ?-? grade,Lown's grade? ?,NT-pro BNP,lymphocyte percentage,RDW-CV,BUN,QRSd ? 110 ms are independent risk factors affecting all-cause death in patients with chronic heart failure within 1 year.2.This risk score table is based on NYHA cardiac function class?-?and Ln NT-pro BNP?9 to calculate 2 points respectively,Lown's classification??,age is greater than or equal to 75 years,lympho cyte percentage?18%,RDW-CV?15%,BUN?10mmol/L,and QRSd?110ms to calculate 1 points respectively that constitute a score sheet with a total score of 10 points.It can be used to predict the risk of all-cause death in CHF patients within 1 year.3.When the score more than 5 points,the score table has a higher diagnostic sensitivity and specificity and can be used as a high-risk value to evaluate the mortality risk of patients with chronic heart failure.