| ObjectiveWith the increasing pressure of people's work and life,the population suffer the chronic fatigue stress is increasing.Acute coronary syndrome(ACS)and even cardiac and respiratory arrest are common events caused by long-term stress state.At present,the mortality rate of ACS is as high as 80 people/100,000,it has been accounting for the first cause of death,about 42%.ACS is the most common cause of cardiac arrest,accounting for about 50%of cardiogenic death.It is even as high as 80%in patients over 50 years old,Early CPR is an effective means for ACS to reduce the mortality.The success ratio of rescue can reach 50%within 5 minutes.When increase one minute,the mortality rate increased by 3%.Therefore,early identification of ACS is of great significance to reduce its mortality.A meta-study found that the research on early warning of ACS mainly focused on troponin T,D-dimer and C-reactive protein,but these indicators are involved in more reactions in the body and have poor specificity,searching for new biomarkers for ACS,this is the Concerns of clinical experts and scholars.In recent years,the relationship between stress biomarkers and cardiovascular diseases has draw more and more eyes.It's aim is to explore the predictive value of stress biomarkers for ACS and the predictive value of adverse prognosis.The reason for the high mortality of ACS is not only the high incidence,but also its misdiagnosis rate.It is reported that the rate of myocardial infarction is 3%-5%.The reason for hig misdiagnos is that ACS?aortic dissection and other diseases with chest pain symptoms has the same performance,even the signs,electrocardiogram and laboratory results are very similar,the use of non-invasive examination can not be timely and effective discrimination,which is also a major factor leading to increased mortality.Therefore,it is urgent to build prediction model by big data.The establishment of themodel requires a two-class discriminant method,in which case aortic dissection should be used as the discriminant option of the two-class model.According to reports,the annual incidence of aortic dissection is 25-30/1 million,the mortality rate can reach 50%within 48 hours,and the misdiagnosis mortality rate is more than 90%.The diagnosis of aortic dissection still depends on CT and MRI aortic angiography or transesophageal echocardiography.However,due to its high cost,inconvenient and high risk of detection,the misdiagnosis mortality rate remains high.In view of this situation,it is urgent to establish an early warning and prediction model of ACS,and to provide decision support for reducing the incidence of ACS and the malignant time of cardiopulmonary arrest caused by ACS.Method(1)40 patients with acute coronary syndrome(ACS case group)admitted to emergency department of PLA General Hospital from September 10,2016 to October 10,2016 were collected,and 40 healthy volunteers matched by age and sex(ACS control group)were selected as control group.To retrospectively analyze the general data,clinical test information and stress indicators in order to find out the changes of stress level in acute coronary syndrome.(2)50 patients with sudden cardiac arrest(case group)and 50 healthy persons(control group)matched by sex and age were collected from October 2017 to December 2018.In order to find out the difference of stress level in patients with cardiac arrest and respiratory arrest,the general data,clinical test information and stress index were analyzed retrospectively.(3)279 patients with ACS diagnosed in the Emergency Department of PLA General Hospital from January 1,2012 to October 10,2016 were selected and 250 patients with acute aortic clip were retrospectively analyzed.After data screening,data labeling,missing value supplementation,feature selection,logistic regression,SVM and random forest methods were used to establish the prediction and early warning model of ACS.To improve the early diagnosis rate of ACS and guide the treatment.Result(1)Plasma GDF-15(21.94±14.23vs7.059±5.53,p=0.007),catecholamine(46592.15±30931.27vs5507.14±2083.28,p<0.001)and HSP-70(369.56±300.44 vs107)in the ACS case group.76±54.23,p<0.001)was higher than the control group.GDF-15 plasma level stenosis Gensin score>40 group was significantly higher than<20 group(324.27±198.81 vs 77.43±699.22,p=0.035),and blood vessel stenosis>40 group plasma catecholamine level was significantly higher than<20 grouping(18.71 ±7.32 vs 18.6±46.1,p=0.017),the GDF-15 level in the multivessel disease group was significantly higher than that in the double vessel disease group(618.40±434.42 vs 292.07±219.65,p=0.033).(2)?It is showed that the level of HSP-70(74.44±105.11,p=0.037)?GDF-15(547.24±801.6)in the case group was significantly higher than that in the control group(p<0.05),which was statistically significant Thyrotropin(2.37±2.17 VS 2.056±0.92,p=0.360),neuropeptide-Y(82.92±72.79 VS 72.93±23.27,p=0.605),HSP-70(87.82±136.95 VS 55.97±13.82,p=-1.059),glucocorticoids(12.76±20.35 VS 9.365±13.22,0.669),the remaining indicators were not statistically significant.?Cardiac thyrotropin(2.37±2.17 VS 2.056±0.92,p=0.360),neuropeptide-Y(82.92±72.79 VS 72.93±23.27,p=0.605),HSP-70(87.82±136.95 VS)55.97±13.82,p=-1.059),glucocorticoids(12.76±20.35 VS 9.365±13.22,0.669),GDF-15(564.37±851.76 vs 504.29±726.49)were not statistically different from non-cardiac sudden death.(3)Using logistic regression,SVM and Stochastic Forest methods,ACS prediction and early warning model was established.The accuracy of ACS early warning prediction model is 0.869(6.42e-04),0.867(1.06e-03)and 0.907(1.36e-03),respectively.(3)Using logistic regression,SVM and Stochastic Forest methods,ACS prediction and early warning model was established.The accuracy of ACS early warning prediction model is 0.869(6.42e-04),0.867(1.06e-03)and 0.907(1.36e-03),respectively.Conclusion(1)The stress indicators GDF-15,catecholamine(adrenaline,norepinephrine,dopamine)and HSP-70 in ACS patients were significantly increased,which were related to the severity of coronary artery disease.These three indicators could be used as new predictors of ACS.(2)HSP-70,GDF-15 can be used as a predictor of sudden death in ACS.Glucocorticoid,thyrotropin and neuropeptide Y can not be used as predictors of ACS sudden death.(3)The ACS prediction model can be used as an assistant decision-making tool. |
PDF Full Text Request