| BackgroundAcute heart failure (AHF) is a severe acute disease; it is a syndrome which is defined as decreased cardiac myocytes contractility or cardiac output failure secondary to cardiac dysfunction, rapid increased pressure of systemic or pulmonary circulation, and increased resistance of peripheral circulation, resulting in the need for urgent therapy. It has higher readmission rates and mortality.Heart failure (HF) can be divided into two types according to the left ventricular systolic function, left ventricular ejection fraction (LVEF) stands for left ventricular systolic function, according to the number of the LVEF, HF with impaired left ventricular systolic function is called HF with reduced ejection fraction (HFREF), and normal left ventricular systolic function called HF with preserved ejection fraction (HFPEF). Traditional treatment of HF focus on HFREF, and the incidence of HFPEF was equivalent or higher than that of HFREF, although there are similar clinical manifestations in the two type of HF, but the pathogenesis and treatment are often different, so it has important clinical significance to distinguish HFREF and HFPEF. HFREF and HFPEF can be distinguished by LVEF to guide the clinical treatment; the LVEF is detected by cardic uhrasonography. However, obtaining LVEF for clinical emergency is often tardiness. Therefore, the application of simple clinical or laboratory indicators is necessary to help as soon as possible distinguish the two types of HF.At present, There are a lot of clinical indicators for predicting the prognosis of AHF, including age, gender, body mass index, etiology, classification of cardiac function, LVEF, high sensitive C reactive protein, cTn and so on, and those of which are concerned with LVEF. Any of the above indicators have their limitations; different types of AHF may be suitable for different indicators. Therefore, clinical or laboratory parameters which can be quickly and easily obtained are necessary for clinical doctors to take the assessment of the severity of AHF and prognosis, to detect high risk patients, to type acute HF with reduced ejection fraction (AHFREF) and actue failure with preserved ejection fraction (AHFPEF) and to take positive and effective measurement to improve the prognosis of AHF.Many studies have shown that systolic blood pressure (SBP) at admission is an independent predictor of prognosis of AHF. Higher SBP at admission, LVEF relative retention may be higher. The French EFICA (French Epidemiology of Acute Heart Failure) study in 2001 found that patients with SBP>160mmHg have a lower mortality than patients with SBP<160mmHg patients (7% vs.17%) at 4 weeks among AHF patients without cardiogenic shock. OPTIMIZE-HF (Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with HF) enrolled patients with HF from 259 hospital in US from March 2003 to December 2004, and studies of its have shown that either low or relatively preserved SBP levels, SBP at admission was an independent predictor of mortality and morbidity in patients with HF. Therefore, if we can identify those AHF patient with different clinical features, prognosis and pathological physiology according to the patient’s blood pressure levels at admission, classify the risk existed in patients with HF, take a more effective treatment according to the patient’s SBP levels, the prognosis of patients with AHF will be effectively improved.Recent years, people gradually realize that the closed relationship between heart rate (HR) and HF. Abnormally rapid HR is related with the change of cardiac output and myocardial oxygen consumption, which directly or indirectly affects the prognosis of patients with HF. HR increased 1 times per minute, cardiovascular death or hospitalization for HF will increase 3%, and the HR increased 5 times per minute, cardiovascular death or hospitalization for HF risk will increase 16%. Studies indicate: patients with HRs ≥70bpm had significantly increased cardiac mortality (34%, P=0.0041), HF hospitalization (53%, P<0.05), myocardial infarction hospitalization (46%, P=0.0066) and the treatment rate of coronary artery recanalization (38%, P=0.037). At present, most researches mainly focus on the relationship between chronic HF and HR, few researches focus on relationship between AHF and HR,which is worthy of our further analysis.N-terminal pro-brain natriuretic peptide (NT-pro BNP) belongs to the member of natriuretic peptide family and mainly releases when the ventricular cells feel the increased tension of ventricular wall. It inhibits renin-angiotensin-aldosterone and sympathetic nervous system activity, dilates blood vessels, increases sodium and urine excretion, inhibits cardiac fibroblast proliferation and has many other physiological effects. The BNP precursor protein (pro BNP) encoded by BNP gene is a polypeptide of 108 amino acid fragment. Pro BNP is cleaved into fragments of 76 amino acids NT-pro BNP with biological inactivity and 32 amino acid fragment BNP fragments with biological activity when volume overload, elevated pulmonary capillary wedge pressure, cardiac decompensation etc lead to ventricular wall tension increase. Studies of Januzzi have shown the concentration of NT-pro BNP closely related to HF severity (based on the NYHA classification). The relationship between NT-proBNP levels and left ventricular systolic dysfunction has been very clear, even without the presence of systolic dysfunction, NT-pro BNP levels are also associated with diastolic dysfunction, and plasma NT-pro BNP levels were higher in HFREF than in HFPEF. ESC in the guideline for diagnosis and treatment of HF in 2008 clearly pointed out:patients have HF symptoms but not be untreated; BNP\ NT-pro BNP detection is helpful to the diagnosis of HF. Therefore, NT-pro BNP is an essential indicator for diagnosis and prognosis of HF.More and more researches have shown that age is closely related to the prognosis of HF. The etiology of HF in patients with different ages is often different: rheumatic heart disease is the main cause in most adults in the age of 45-59 years old. Coronary heart disease and hypertension are the main cause in patients at the age of 60-79 years old and more than 80 years old. With age, the proportion of chronic heart failure reduces relatively, while acute left heart failure, acute attack of chronic left heart failure increase significantly, especially in elderly patients. The proportion of diastolic heart failure in the aged group are higher than pre-aged group, suggesting that myocardial contractility decreases only in a part of the elderly patients, and myocardial compliance losses in a considerable part of the elderly patients with HF. So the treatment for senile HF, not only pay attention to enhance myocardial contractility, but also more to improve myocardial compliance. Studies have shown: levels of neurohormones such as noradrenaline, aldosterone (Ald), N-terminal atrial natriuretic peptide (N-ANP) and ANP in the patients with HF have obviously elevated, which are significantly related to age. N-ANP, ANP and NA increase with age, renin and Aid decrease with age. Levels of N-ANP in patients with HF at the age of 80 are 3 times than those in 40 years old. NA may cause ischemia, hypoxia and ventricular remodeling through increased cardiac work, so cardiac function is further deteriorated, mortality is increased. Age can increase the incidence of hypertension, diabetes mellitus and atrial fibrillation; these complications may increase the risk of death in elderly patients with HF. Old age is the prediction of risk factors in patients with HF.As the social aging intensifies, studies on the relationship between age and the prognosis of HF seem to be very important.It has affirmed that blood glucose (GS) has good predictive value for the prognosis of acute coronary syndromes, which is used as a common clinical biochemical indicator and has the advantages of rapid and low cost. Researches have shown glucose may have a similar role in the patients with AHF. Higher GS level in AHF may be a stress response of the body, or induced by autonomic nerve and endocrine system dysfunction, or induced by diabetes or abnormal glucose metabolism. HF itself is obvious with insulin resistance (IR), the characteristics of clinical manifestation is hyperinsulinemia after stimulation. A large number of studies have shown that diabetes can cause increased incidence of HF, Higher GS levels in patients without diabetes but complicated with heart disease can increase the rate of hospitalization and mortality. So the disorder of glucose metabolism may play an important role in the occurrence, development and prognosis of HF. GS levels can simply obtain in emergency patients; therefore the evaluation of correlation between GS levels in patients with HF has important significance.Age, blood pressure, HR, GS levels, NT-pro BNP and other indicators are the clinical indicators which can be easily obtained in patients with AHF. So it has important significance to discuss the role of common clinical indicators in the diagnosis, treatment and prognosis of AHF. This study is divided into three parts:part I:The correlation between cardiac function and blood pressure, HR, NT-pro BNP and other indicators in AHF patients at admission in the emergency department, and the difference of these indicators between AHFREF and AHFPEF; part II:The value of combination of multiple clinic indicators (age, blood pressure, HR, GS, NT-proBNP and other indicators) in the differential diagnose of AHFREF and AHFPEF; part III To explore risk factors of short-term prognosis of AHF patients and compare the difference between AHFREF and AHFPEF. Object:1. To investigate the correlation between cardiac function and age, blood pressure, HR, GS, NT-pro BNP and other indicators in AHF patients at admission in the emergency department, and the difference of these indicators between AHFREF and AHFPEF.2. To study the value of combination of several clinic indicators (age, blood pressure, HR, GS, NT-proBNP and other indicators) in the differential diagnose of AHFREF and AHFPEF.3. To explore risk factors of short-term prognosis of AHF patients and compare the difference between AHFREF and AHFPEF. Material:1. The AHF patients admitted to Zhongshan city hospital emergency center between April 2012 and April 2013 (including the incipient AHF and acute onset of chronic heart failure) with routine treatment after discharge. The diagnosis and treatment are accordance with ESC guidelines for the diagnosis and treatment of acute and chronic heart failure 2012.166 patients with AHF were enrolled with age 74.0 (64.0,79.0) years, men 103(62.0%) and women 63(38.0%). All the patients were older than 18 years old, men and women were not limited. The main diseases of 166 cases with AHF include hypertension (n=107,64.5%), coronary heart disease (n=97,58.4%), diabetes mellitus (n=42,25.3%), arrhythmia (n=20,12%), acute exacerbation of chronic obstructive pulmonary disease (n=11,6.6%), pulmonary infection (n=10,6%), dilated cardiomyopathy (n=4,2.4%), infective endocarditis (n=3,1.8%), acute viral myocarditis (n=3,1.8%), etc.2. Exclusion criteria:the patients with heart valve disease (including more than mild ale regurgitation), acute myocardial infarction, renal insufficiency(BUN>9mmol/L, Cr>178μmol/L), undergoing hemodialysis, acute pulmonary embolism, perioperative patients with AHF, malignant tumors and N terminal pro brain natriuretic peptide (NT-pro BNP)< 300 pg/ml were all excluded.3. Grouping:to group the patients accord to the LVEF, LVEF< 0.5 for AHFREF group, LVEF≥ 0.5 for AHFPEF group. The involved patients were followed up via telephone contact and consult of medical recording. Unfavorable prognosis was defined upon the cardiac death and re-hospitalization within 6 months; otherwise they had a favorable prognosis.4. Measurements:to measure the semi-supine blood pressure, HR, bedside detection of NT-pro BNP, glucose at admission and cardiac ultrasound and other indicators within a day after admission of AHF patients in the emergency department. Method1. Data collectionCollect all the general clinical data of patients including:admission number, name, gender, age, detailed history collection and physical examination, measure and record the semi-supine blood pressure and HR in the emergency department. Take cardiac color doppler ultrasound within a day and measure the left ventricular end diastolic capacity and at the left ventricular end contract capacity, etc.2. Measurement of NT-pro BNP and GSExtract venous blood 2ml and Roche CARDIAC(?)roBNP+ was applied to quantitatively detect NT-proBNP levels in heparinized venous whole blood, the test range:60-9,000 pg/ml. Bayer’s BREEZE 2 was applied to quantitatively detect GS levels in 1μL fingertip capillary blood, the test range:0.6-33.3mmol/L.3. EchocardiographyAccording to the measurement method recommended by American Echocardiography Society, with the use of PhillipsIE33 color Doppler ultrasonography, we get every standard section by M-mode ultrasound method. According to the left ventricular long diameter (L) and short diameter measurements (D), the system calculates automatically the left ventricular end-diastolic volume (LVEDV), left ventricular end-systolic volume (LVESV), stroke volume (SV), cardiac output (CO), left ventricular shortening fraction (FS), left ventricular ejection fraction (LVEF) and so on. Related formula:left ventricular volume (VV)=πLD2/6, SV= LVEDV-LVESV, CO= SV*HR, FS= 1-Ds/Dd, LVEF= (LVEDV-LVESV)/ LVEDV, HR for HR, Ds for systolic diameter, Dd for diastolic diameter. The measurement was responsible by someone from the echocardiography department in Zhongshan city people hospital.4. Excel database was established to collect all the information and the data was double-checked after double entry.5. Data analysisData were statistically analyzed using IBM SPSS 19.0 software. Measurement data under normal distribution were expressed as mean±tandard deviation (x±s), otherwise as the median and the interquartile range [M (P25, P75)]. Count data were expressed absolute frequency and constituent ratio. The correlation among each index was analyzed with using bivariate correlation analysis. The data were analyzed with Pearson correlation coefficient when they met the bivariate normal distribution, otherwise with the Spearman correlation coefficient. Measurement data under normal distribution were compared with the t test, otherwise with independent samples nonparametric test.The ratio was compared with Chi-square test. Rank data were compared with the non parameter test.The age, gender, SBP, diastolic blood pressure(DBP), HR, NT-proBNP, GS, LVEF, cardiothoracic ratio, history of hypertension, coronary heart disease and diabetic mellitus as independent variables, and AHF type as a binary dependent variable and to be analyzed by logistic regression analysis. Statistically significant independent variables were screened by univariate analysis, and then independent variables were selected into the equation arguments according to clinical practice. The binary logistic regression analysis regression model was established according to strain samples randomly selected from 70% AHF cases. Calculated each individual (the independent variable joint into the equation) to predict the probability of AHFREF (predicted probability). To make receiver operating characteristic (ROC) curve with the probability predicted and the largest Youden Index based on the ROC curve as the optimal diagnostic cutoff point. The model was tested by test samples from the the remaining 30% AHF cases.The age, gender, SBP, DBP, HR, NT-proBNP, GS, LVEF, cardiothoracic ratio, history of hypertension, coronary heart disease and diabetic mellitus as independent variables, and short-term prognosis of AHF as binary dependent variables and to be analyzed by logistic regression analysis. These independent variables were screened by univariate analysis, and then statistically significant independent variables were analyzesed by multivariate logistic regression analysis. Independent variables filtered by step remove probability of 0.10 into multivariate logistic regression model. P <0.05 was considered as statistically significant.Result1. The relation between blood pressure, HR, NT-pro BNP, glucose etc and heart function1.1 General characteristics of clinical data:166 patients with AHF were enrolled with age 74.0 (64.0, 79.0) years, men 103(62.0%) and women 63(38.0%), SBP M(P25> P7s)143.5 (126.0 ,179.3 ) mmHg, DBP M(P2S, Pis) 89.0 (73.0 ,101.5 ) mmHg, HR M(P25, P75) 102.5(87.0,122.0 )bpm, NT-pro BNP M(P2S, P75) 6312 (2466 ,>9000 )pg/ml, glucose M(P25, P75)9.7 (8.1,12.2 )mmol/L at admission.Echocardiographic indicators of LVEF M(P25 > P75)0.54(0.42,0.62), FS M(P25, P15)0.27 (0.20, 0.32) ,SVM(P25, P7s)59.0 (46.8, 69.3) ml,COM(P25, P7s)4.5 (3.7, 5.7) L/min, Left ventricular end-diastolic diameter M(P2S, P7s)49.0 (47.0, 56.0) mm within 1 day after visit, cardiothoracic ration(CTR) (0.59±0.07) . In 166 patients, 107(64.5%) cases had a history of hypertension, 97(58.4%)had a history of coronary heart disease, 42(25.3%) had a history of diabetes mellitus.Age (Z=-1.220, P=0.222), gender (χ2=0.117, P=0.732), SBP (Z=-0.982, P=0.326) , GS (Z=-0.393, P=0.695) , CTR (t=-0.821, P=0.413) , history of hypertension (χ2=0.233, P=0.629), coronary heart disease (χ2=0.213,P=0.645), diabetes mellitus (χ2=0.065, P=0.799) and prognosis (χ2=1.824, P=0.177) had no significant difference in AHFREF and AHFPEF groups.In AHFREF group than AHFPEF group, DBP [92.0 (77.8,105.3) mmHg vs. 84.0 (71.3,97.0) mmHg, Z = -2.342, P = 0.019]; HR [113.0 (95.8,130.0) bpm vs. 96.0 (83.0,116.5) bpm, Z = -3.186, P = 0.001]; NT-pro BNP elevated [9000 (5568,> 9000) pg/ml vs.4880 (1576,8945) pg/ml, Z = -4.601, P = 0.000]; LVEF [0.39 (0.32,0.45) vs.0.60 (0.55,0.64), Z = -10.898, P = 0.000], FS [0.19 (0.15,0.21) vs.0.31 (0.28,0.34), Z = -10.890, P = 0.000], SV [50.0 (41.0,60.3) ml/beat vs.60.0 (51.3,70.8) ml/beat, Z = -4.738, P = 0.000 ], CO [4.0 (3.2,5.4) L/min vs.4.9 (4.3,6.0) L/min, Z = -3.852, P = 0.000] lowered; interventricular septum [11.0 (10.0,12.0) mm vs.10.0 (10.0,11.0) mm] thickened, Z = -2.044, P = 0.041]; left ventricular dilatation [54.0 (48.0,60.0) mm vs.48.0 (45.3,52.8) mm, Z = -4.296, P = 0.000]enlarged.1.2 Study of correlation:There were correlation between indicators of cardiac function (LVEF, FS, SV, CO) and emergency common clinical indicators (blood pressure, HR, NT-pro BNP, blood sugar, etc.). LVEF were negatively correlated with HR (r = -0.252, P = 0.001), NT-proBNP (r = -0.286, P = 0.000). Left ventricular shortening were positively correlated with age (r = 0.156, P=0.045), and negatively with HR (r = -0.265, P = 0.001), NT-proBNP (r = -0.263, P = 0.001). SV were positively correlated with gender (r = 0.170, P = 0.029), SBP (r = 0.257, P = 0.001), and negatively with HR (r = -0.283, P = 0.000), NT-proBNP (r = - 0.160, P = 0.039). CO were positively correlated with gender (r = 0.193, P = 0.013), SBP {r = 0.185, P = 0.017), cardiothoracic ratio (r = 0.208, P = 0.007), and negatively with glucose (r = -0.202, P = 0.009), history of coronary heart disease (r = -0.197, P = 0.011).Cardiac function and emergency common clinical indicators were also relevant in AHFREF. LVEF and age (r = 0.416, P = 0.001), SBP (r = 0.392, P = 0.001), hypertension (r = 0.345, P = 0.005), history of diabetes (r = 0.273, P = 0.026) were positively related. Left ventricular shortening and age (r= 0.356, P= 0.003), SBP (r = 0.378, P= 0.002), history of hypertension (r= 0.328, P= 0.007), history of diabetes (r= 0.287, P= 0.019) were positively correlated. SV and SBP (r= 0.445, P = 0.000), DBP (r= 0.316, P= 0.010) were positively correlated, glucose (r=-0.366, P=0.002) were negatively correlated. CO and SBP (r= 0.305, P= 0.013), DBP (r= 0.278, P= 0.024) were positively correlated, glucose (r=-0.457, P= 0.000) were negatively correlated.Cardiac function and emergency common clinical indicators were also relevant in AHFPEF. LVEF and history of coronary heart disease (r=-0.250, P= 0.012) were negatively correlated. FS and history of coronary artery disease (r=-0.198, P = 0.048) were negatively correlated. SV and age (r=-0.239, P= 0.017), HR (r=-021 A, P= 0.006) were negatively correlated, gender (r= 0.240, P= 0.016) were positively correlated. CO and age (r=-0.284, P= 0.004), history of coronary heart disease (r=-0.286, P= 0.004) were negatively correlated, gender (r= 0.316, P= 0.001), cardiothoracic ratio (r= 0.323, P= 0.001) were positively correlated.2 The value of combination of several clinic indicators in the differential diagnose of AHFREF and AHFPEF2.1 General characteristics and comparison of clinical data:Set random seed number 20140101, subjects were randomly sampled, about 70% of the total population as training samples (n=118), the remaining subjects as test samples about 30% of the total population (n=48).Total population:166 cases enrolled, age M(P25, P1S) 74.0 (64.0,79.0) years old, male 103 cases (62.0%), female 63 cases (38.0%). SBP M (P25, P79.3) 143.5 (126.0,179.3) mmHg, DBP M(P25, P75) 89.0 (73.0,101.5) mmHg, HRM(P25,P75) 102.5 (87.0,122.0) beats/min, high grade NT-pro BNP in AHFREF group were more and low grade NT-pro BNP in AHFPEF group were more, sugar M(P25, P75) 9.7 (8.1, 12.2) mmol/ L. cardiothoracic ratio 0.59 ± 0.07. In 166 patients, 107 (64.5%) patients had history of hypertension, 97 (58.4%) patients had history of coronary artery disease, and 42 (25.3%) patients had history of diabetes.AHFREF group of 66 cases (39.8%), AHFPEF group of 100 cases (60.2%). In AHFREF and AHFPEF, age (Z = -1.220, P = 0.222), gender (χ2 = 0.117, P = 0.732), SBP (Z = -0.982, P = 0.326) , GS (Z = - 0.393, P = 0.695), cardiothoracic ratio (t =-0.821, P = 0.413), hypertension (χ2 = 0.233, P = 0.629), coronary heart disease (χ2 = 0.213, P - 0.645), diabetes (χ2 = 0.065, P = 0.799) history had no significant difference.AHFREF group than AHFPEF group, DBP [92.0 (77.8, 105.3) mmHg vs. 84.0 (71.3, 97.0) mmHg, Z = -2.342, P = 0.019], HR [113.0 (95.8, 130.0) bpm vs. 96.0 (83.0, 116.5) bpm, Z = -3.186, P = 0.001], and NT-pro BNP (comparison between rank groups, Z=-4.601,P=0.000) elevated.Training population:118 cases enrolled, age M (P25, P75) 74.0 (66.0,78.0) years old, male 76 cases (64.4%), female 42 cases (35.6%). SBP M (P25, P75) 144.5 (127.0, 181.0) mmHg, DBP M(P25, P75) 91.0 (76.5, 103.3) mmHg, HR M (P25, P75) 102.0 (87.0, 122.3) beats/min, high grade NT-pro BNP in AHFREF group were more and low grade NT-pro BNP in AHFPEF group were more, sugar M(P25, P75) 9.8 (8.2, 12.6) mmol/ L. cardiothoracic ratio 0.59 ± 0.07. In 118 patients, 75 (63.6%) patients had a history of hypertension, 70 (59.3%) patients had a history of coronary artery disease, and 33 (28.0%) patients had a history of diabetes.AHFREF group of 49 cases (41.5%), AHFPEF group of 69 cases (58.5%). In AHFREF and AHFPEF, age (Z = -0.894, P = 0.371), gender (χ2 = 0.030, P = 0.863), SBP (Z= -1.491, P = 0.136), DBP (Z=-1.475, i>=0.140) , GS (Z = - 0.885, P = 0.695), cardiothoracic ratio (t=-0.636, P= 0.526), hypertension (x2= 0.110, P= 0.740), coronary heart disease (%2= 0.126, P= 0.723), diabetes (x2= 0.503, P= 0.478) history had no significant difference.AHFREF group than AHFPEF group, HR [112.0 (94.0,129.5) bpm vs.96.0 (78.0,116.0) bpm, Z=-2.729, P=0.007], and NT-pro BNP (comparison between rank groups, Z=-3.678, P=0.000) elevated.DBP, HR, NT-pro BNP between two groups from the total population had statistically significant difference (P<0.05), other indexes between two groups had no significant difference (P>0.05). HR, NT-pro BNP between two groups from the training samples had statistically significant difference (P<0.05), other indexes between two groups had no significant difference (P>0.05).DBP between two groups had no significant difference because of the reduced number of training samples; the other indexes were consistent with total population.2.2 Differential Diagnosis value of combination of several clinic indicators in the differential diagnose of AHFREF and AHFPEF:2.2.1 Screening clinic indicators for the differential diagnose of AHFREF and AHFPEFGrading NT proBNP, DBP and HR in total population as independent variable had statistically significance (P<0.10) by univariate logistic regression with AHFREF group and AHFPEF group as dependent variable. Grading NT proBNP and HR in training population as independent variable had statistically significance (P <0.10) by univariate logistic regression with AHFREF group and AHFPEF group as dependent variable.Co-linear test of continuous variables in total population, including age, SBP, DBP, HR, blood sugar, and cardiothoracic ratio, showed significant collinearity (characteristic root= 0.006, condition index= 34.663), and co-linear (Characteristic root = 0.011, condition index = 23.008) were under control after cardiothoracic ratio excluded according to clinic practice.Correlation analysis of qualitative variables in total population showed that gender, hypertension, coronary heart disease and diabetes had not a strong correlation(r<0.3), so not to remove other variables.2.2.2 Establishing regression, ROC curve and the optimal cut-off point of diagnosisCombined with results of the first part and this part of this study and clinic practice , age, SBP, DBP, HR, NT-pro BNP as covariates were included into binary logistic regression analysis, the regression equation P=(I (1+ exp (- (-1.218-0.020 * Age + 0.519 * NTproBNP-0.032 * SBP + 0.046 * DBP + 0.011 * HR))).The regression equation was statistically significant (x2 = 32.177, P <0.001), COX and snell coefficient of determination was 0.239, nagelkerke decision coefficient 0.321. Hosmer-Lemeshow test model (x2 = 8.654, P = 0.372) showed predicated and observed value of the model had not statistically significant differences. Therefore, the model had statistically significance and good fitting effect.Predicted probability of the best diagnostic point according to max Youden index from the ROC curve was 0.375, sensitivityO.796 and specificity 0.739, right rate 76.3%. the area under ROC curve was 0.804 (P<0.001) , and its 95%CI 0.723-0.886.2.2.3 Model testThe remaining subjects as test samples were from about 30% of the total population. The cutpoint sensitivity and specificity in test samples were calculated by the cutpoint from training samples. The sensitivity was 0.765, specificity 0.774, and right rate 77.1% in test samples when cutpoint was 0.375. the area under ROC curve was 0.829 (P<0.001) , and its 95%CI 0.713~0.946。 It consistented with the conclusion form training samples showed that the model established was reliable, and had good adaptivity and practice.3. To study related factors of short-term prognosis for AHF and its differences between AHFREF and AHFPEF3.1 General characteristics of clinical data:166 AHF patients enrolled, age M (P25, P75) 74.0 (64.0, 79.0) years old, male 103 cases (62.0%), female 63 cases (38.0%). In 66 AHFREF group, 19 cases had poor recent prognosis (28.8%, including 4 deaths); in 100 AHFPEF group, 39 cases had poor recent prognosis (39.0%, including 5 deaths), prognosis and mortality in AHFREF group AHFPEF group had no statistically significant difference, (respectively χ2 = 1.824, P = 0.177; χ2 = 0.087, P = 0.768). UFPG group compared with the FPG group, SBP [135.0 (116.8, 169.0) mmHg vs. 151.0 (129.3,180.8) mmHg, Z = -2.574, P = 0.010] significantly reduced, NT-pro BNP [8835 (3437,> 9000) pg/ml vs. 2187 (5236,> 9000) pg/ml, Z = -2.486, P = 0.013] and glucose GS (mmol/L) [10.6 (8.7,13.7) mmol/L vs. 7.9 (9.2 ,11.2) mmol/L, Z = -2.520,.P = 0.012] were significantly increased. AHFREF UFPG group compared with patients with FPG group, SBP [117.0 (104.0, 131.0 mmHg vs.) 157.0 (131.0, 188.0) mmHg, Z = -4.590, P = 0.000], DBP [82.0 (72.0, 92.0) mmHg vs. 99.0 (89.0,110.0) mmHg, Z = -3.422, P = 0.001] and LVEF [0.30 (0.23, 0.45) vs. 0.40 (0.35, 0.45), Z = -2.505, P = 0.012] significantly reduced. AHFPEF UFPG group compared with patients with FPG group, NT-pro BNP [6312 (2000,> 9000) pg/ml vs.4390 (1334, 6691) pg/ml, Z = -2.405, P = 0.016] were significantly higher.3.2 Logistic regression analysis of prognosis factors:According to P <0.10 for the univariate screening criteria into the multivariate logistic regression analysis, SBP, DBP, NT-pro BNP, GS for AHF covariates;Gender, SBP, DBP, NT-pro BNP, LVEF, history of hypertension for AHFREF; NT-pro BNP, blood sugar for AHFPEF.By multivariate logistic regression analysis of prognostic factors, related factors of short-term prognosis were SBP (β = -0.013, P = 0.006, 95% CI 0.977 ~ 0.996), NT-pro BNP (β = 0.299, P = 0.008, 95% CI 1.080 ~ 1.685 ), GS(β = 0.091, P = 0.025, 95% CI 1.012 ~ 1.186) in AHF, SBP (β = -0.099, P = 0.020, 95% CI 0.832 ~0.985) AHFREF, NT-pro BNP (β = 0.327, P = 0.018, 95% CI 1.057 ~ 1.818) in AHFPEF.ConclusionThe first part of the conclusions(1) SBP and left ventricular function have a positive correlation. Blood pressure and SV (r = 0.257, P = 0.001), CO (r = 0.185, P = 0.017) were positively correlated in AHF patients at admission. And this relationship after typing more obvious in AHFREF, namely systolic and LVEF (r = 0.392, P = 0.001), FS (r = 0.378, P = 0.002), SV (r = 0.445, P = 0.000), CO (r = 0.305, P - 0.013); DBP and SV (r = 0.3,16, P = 0.010), CO (r = 0.278, P = 0.024); high blood pressure history and LVEF (r = 0.345, P = 0.005), FS (r = 0.328, P = 0.007) were positively correlated. Meanwhile in AHFPEF this relationship was not statistically different, suggesting the role of blood pressure in both types of HF and the development process may be different.(2) HR and left ventricular function were negatively correlated. At admission HRs were significantly higher in AHFREF group than AHFPEF group [113.0 (95.8, 130.0) bpm vs. 96.0 (83.0, 116.5), Z = -3.186, P = 0.001], HR and LVEF (r = -0.252, P = 0.001), FS (r = -0.265, P = 0.001), SV (r = -0.283, P = 000) were negatively correlated, this is similar with relatively stable chronic heart failure, suggesting rapid HR may indicate poor heart function.(3) NT-pro BNP and left ventricular function were negatively correlated. At admission NT-proBNP and LVEF (r = -0.286, P = 0.000), FS (r = -0.263, P = 0.001), SV (r = -0.160, P = 0.039) had negative correlation; AHFREF group than AHFPEF group, NT-pro BNP elevated [9000 (5568,> 9000) pg / ml vs.4880 (1576,8945) pg / ml, Z = -4.601, P = 0.000]. According to NT-pro BNP levels in patients with AHF at admission cardiac function can be determined.(4) Sugar negatively correlated with left ventricular function. GS and CO (r =-0.202, P = 0.009) were negatively correlated, especially in AHFREF glucose and SV (r = -0.366, P = 0.002), CO (r = -0.457, P = 0.000) were negatively correlated, so the evaluation of the correlation of GS levels in patients with AHF has important significance.(5) Relationships between each index in AHFREF and AHFPEF and cardiac function were different SBP and LVEF (r = 0.392, P = 0.001), FS (r = 0.378, P = 0.002), SV (r = 0.445, P = 0.000), CO (r = 0.305, P = 0.013) were positively correlated; glucose and SV (r = -0.366, P = 0.002), CO (r = -0.457, P = 0.000) negative correlation; hypertension and LVEF (r = 0.345, P = 0.005), FS (r = 0.328, P = 0.007) were positively correlated; a history of diabetes and LVEF (r = 0.273, P = 0.026), FS (r = 0.287, P = 0.019 ) a positive correlation in AHFREF. history of coronary artery disease in patients with LVEF (r = -0.250, P = 0.012), FS (r =-0.198, P = 0.048), CO (r = -0.286, P = 0.004), left atrium ( r = -0.23, P = 0.020) negative correlation; HR and SV (r = -0.27, P = 006) were negatively correlated; age and SV (r = -0.239, P = 0.017), CO (r = -0.284, P = 0.004) negative correlation; gender and SV (r = -0.240, P = 0.016), CO (r = -0.316, P = 0.001) negative correlation in AHFPEF.In short, at admission reduced SBP and rapid HR, higer NT-pro BNP, hyperglycemia indicated reduce left ventricular systolic function, relationship between each index in AHFREF and AHFPEF and cardiac function were different. With a history of clinical to select well treatment measures.The second part of the conclusion(1) LVEF negatively correlated with NT-pro BNP (r=-0.286, P= 0.000) in the first part of the study; in the end regression equation from the second part of the study, NT-proBNP (β= 0.519, P=0.001,95% CI 1.225～2.252), showed that NT-pro BNP for the diagnosis of HF is not only important, but also the identification of HFREF and HFPEF also meaningful, the gre... |
PDF Full Text Request