Analysis Of Risk Factors And Construction Of Risk Prediction Model For In-stent Restenosis After Stenting For Coronary Artery Disease Based On Disease And Syndrome Combination

Study 1:Systematic review of the risk prediction model for in-stent restenosis(ISR)after coronary stentingObjectiveThrough a comprehensive literature search,we analyzed and systematically evaluated the model construction,performance evaluation,quality assessment and validation of ISR risk prediction models after coronary stent implantation at home and abroad to provide some evidence to support their further rational,safe and effective clinical application.MethodsA comprehensive search of relevant domestic and foreign clinical research literature published in English databases such as PubMed,Web of Science,Embase,Cochrane Library,and Chinese databases such as China National Knowledge Infrastructure(CNKI),Wanfang database,VIP Database for Chinese Technical Periodicals,and China Biomedical Literature Database(sionMed)was conducted.The final eligible literature was screened by inclusion and exclusion criteria,and a systematic and comprehensive analysis of the ISR risk prediction model after coronary stent implantation in terms of model construction,performance evaluation,quality assessment and validation was used.Results1.Ten English clinical research papers published in international journals and seven Chinese clinical research papers published in domestic journals on ISR risk prediction models after stenting for coronary artery disease were finally included,with a total of 57,698 patients.2.Seventeen studies were included to construct 17 risk prediction models for ISR after stenting for coronary artery disease.Fifteen studies were retrospective cohort studies,one study was a randomized controlled study,and one study was a case-control study;only one study reported missing values and treatment methods,and the remaining 16 studies did not report missing values;for predictor screening,10 studies used univariate analysis,3 studies used LASSO regression methods,2 studies used machine learning methods,1 study used stepwise method,and 1 study used single factor analysis combined with LASSO regression method;about the method of model developing,14 studies used Logistic regression,1 study used Cox regression,and 2 studies used machine learning.3.Model predictors ranged from 4 to 12,with the top 5 being disorders of glucose metabolism(12 items),disorders of lipid metabolism(9 items),age,multi-vessel lesions,and the stent diameter(all 5 items).Seventeen predictive models had AUCs ranging from 0.66 to 0.86,with some degree of discrimination.Twelve studies used calibration curves or Hosmer-Lemeshow goodness-of-fit tests for calibration degree assessment.Six studies underwent internal validation and 8 studies underwent external validation.4.The risk of bias for all 17 prediction models was assessed as high risk of bias,and the risk of applicability was assessed as low risk of applicability.ConclusionThe currently constructed models for predicting ISR risk after stenting for coronary artery disease are somewhat discriminatory,but all of them take a high risk of bias and most of them lack valid spatial and domain external validation.Study 2:Analysis of risk factors for ISR after stenting for coronary artery disease based on disease and syndrome combinationObjectiveThis study intends to collect general data,TCM symptom factors,coronary artery lesions,intraoperative management and review coronary angiography results of patients after coronary stent implantation,and conduct statistical analysis to explore the distribution characteristics of TCM symptom factors,ISR risk factors and correlation between different TCM symptom factors and ISR occurrence in patients with ISR,and explore the risk factors for ISR occurrence,aiming at providing discriminative ideas for early prevention and treatment of ISR.MethodsPatients who underwent drug-eluting stent implantation at the Department of Cardiovascular Medicine from January 2018 to December 2021 and had coronary angiograms reviewed 6～24 months after the procedure was selected.TCM symptom factors,ISR occurrence,and covariates such as demographic characteristics,laboratory tests,medication use,coronary artery lesions,and intraoperative management was collected based on the integrated clinical research platform to compare the population characteristics of the ISR group and the non-ISR group.We compared the population characteristics of the ISR group with those of the non-ISR group and analyzed the distribution characteristics of the TCM symptom factors.Logistic regression analysis was used to investigate the relationship between clinically relevant variables and TCM symptom factors and ISR,and multi-model adjustment was performed to identify ISR risk factors.Subsequent stratified analyses were performed to understand the presence or absence of interactions,and finally,sensitivity analyses were performed to determine the robustness of the results.Results1.A total of 602 cases were included,of which 106 cases(17.6%)had ISR.Cases in the two groups with and without ISR showed statistical differences in blood glucose,use of clopidogrel/ticagrelor,multiple vascular lesions,Gensini score,management of multiple vascular lesions,number of stents implanted,length of stents implanted,symptom factors of the blood stasis,phlegm stagnation,internal obstruction of phlegm and blood stasis,the combination of Qi stagnation and blood stasis,the combination of Qi deficiency,blood stasis,and phlegm,the combination of blood stasis,phlegm,and toxic heat.2.Patients with coronary heart disease after stent implantation and patients with ISR mostly belonged to the complex TCM symptom factors of deficiency in origin and excess in superficiality,a combination of the excess syndrome and deficiency syndrome.The origin deficiency is dominated by qi deficiency(47%,53.8%,respectively),and the superficiality excess is dominated by the blood stasis(70.6%,96.2%,respectively)and the phlegm stasis(39.9%,51.9%),while heat toxicity(23.3%,24.5%),qi stagnation(14.1%,21.7%),yin deficiency(11%,3.8%),yang deficiency(9.3%,0%)and other symptom factors may be seen in combination.Combined TCM symptom factors types were most commonly associated with the combination of qi deficiency and the blood stasis,internal obstruction of phlegm,and blood stasis.3.In the Logistic analysis of a fully adjusted model(adjusted for covariates of sex,age,history of diabetes,history of smoking,blood glucose,BMI,diastolic blood pressure,heart rate,bilirubin,total cholesterol,triglycerides,and high-density lipoprotein),multiple vascular lesions,management of multiple vessels,number of stents implanted,and length of stents implanted were positively correlated with the occurrence of ISR(OR>1,P<0.05),and use of clopidogrel/ticagrelor was passively correlated with the occurrence of ISR(OR<1,P<0.05).4.In the Logistic analysis of a fully adjusted model,TCM symptom factors of the blood stasis,phlegm stagnation,internal obstruction of phlegm and blood stasis,the combination of Qi deficiency,blood stasis,and phlegm,the combination of blood stasis,phlegm,and toxic heat were positively correlated with the occurrence of ISR(OR>1,P<0.05).ConclusionThe TCM symptom factors in patients who developed ISR after coronary stent implantation were mostly deficiency in origin and excess in superficiality,a combination of the excess syndrome and deficiency syndrome.Patients in the group with ISR had high admission glucose,did not use clopidogrel/ticagrelor,had multiple vascular lesions,had higher Gensini scores,dealt with multiple vascular lesions,had more stents implanted,had longer stent lengths implanted,and had a higher proportion of symptom factor of the blood stasis,phlegm stagnation,internal obstruction of phlegm and blood stasis,the combination of Qi stagnation and blood Stasis,the combination of Qi deficiency,blood stasis,and phlegm,the combination of blood stasis,phlegm,and toxic heat.The use of clopidogrel/ticagrelor,multiple vessel lesions,management of multiple vessels,number of stents implanted,the total length of stents implanted,TCM symptom factors of the blood stasis,TCM symptom factors of phlegm and turbidity,TCM symptom factors of internal obstruction of phlegm and blood stasis,TCM symptom factors of the combination of Qi deficiency,blood stasis,and phlegm,TCM symptom factors of the combination of blood stasis,phlegm,and toxic heat were all independent risk factors for the development of ISR in patients after stenting for coronary artery disease.Study 3:Construction and validation of a disease and syndrome combination based ISR prediction model after stenting for coronary artery diseaseObjective:In-stent restenosis(ISR)is a post-stenting complication that seriously affects treatment outcomes and patient prognosis.Researching disease and syndrome combination based risk prediction models of chronic disease to provide accurate and targeted health guidance is an important tool for chronic disease prevention and treatment.This study proposes to develop a disease and syndrome combination based postoperative ISR risk prediction model for coronary heart disease stent implantation,which will help clinicians identify high-risk ISR patients and optimize treatment strategies.MethodsIn this part of the study,602 patients from the second part of the study were included,including 106 patients who suffered from ISR and 496 patients who did not suffer from ISR.After missing value treatment,model variables were screened using single/multiple factor analysis,stepwise regression,optimal subset regression,and LASSO regression methods,and a disease and syndrome combination based ISR risk prediction model after stent implantation was constructed using the screened variables and a nomogram was drawn.The subject operating characteristic curve(ROC)and area under the curve(AUC)was used to evaluate the discrimination of the prediction model,and the calibration degree curve,Brier value,and Hosmer-Lemeshow test was used to evaluate the calibration degree of the model.Decision curve(DCA)to evaluate the clinical usefulness of the prediction model.Internal validation of the model was performed using random split validation,cross-validation,Bootstrap bootstrap sampling validation,and the sensitivity analysis using the complete data set.Finally,the methodological quality of the model was evaluated using the PROBAST tool.Results1.The risk prediction formula for ISR after stenting for coronary artery disease was derived from regression coefficient of the regression model assigned to the risk factors as follows:score=-1.299+(1.025×post-stenting)+(1.81×symptom factors of internal obstruction of phlegm and blood stasis)+(0.113×glucose at admission)+(-0.917×HDL-C)+(-1.059×use of clopidogrel/ticagrelor)+(0.222×treatment of 2 vessels)/(1.114×treatment of 3 vessels)/(1.508×treatment of 4 vessels).2.The AUC of this prediction model is 0.752,95%confidence interval is 0.696 to 0.807,which indicates good differentiation of the prediction model.The Brier score of this prediction model is 0.122,which is less than 0.25,suggesting good model calibration and predictive ability.The chi-square value of the Hosmer-Lemeshow goodness-of-fit test was 13.048 with a p-value of 0.11,which was greater than 0.05,suggesting a good model fit.Clinical decision curve analysis showed the best applicability of using this column line plot for predicting ISR when the threshold probability of predicting ISR was in the range of 1%to 83%.3.In randomized split cross-validation,the ROC curve for the validation group showed the AUC is 0.744,95%confidence interval is 0.622 to 0.865.The brier value is 14.8,95%confidence interval is 10.2 to 19.3,the chi-squared value of the Hosmer-Lemeshow Test is 10.664,and the p-value is 0.221.The AUC of the cross-validation group is 0.7253.The AUC of the Bootstrap sampling validation is 0.733.All of the above suggest good validation and good discrimination and calibration of the model.ConclusionIn this study,we developed and validated a disease and syndrome combined predictive model for predicting the risk of ISR after stenting by retrospectively analyzing clinical information and coronary angiography data of patients undergoing stenting,incorporating the predictive variables of blood glucose,high-density lipoprotein,history of previous stenting,non-use of clopidogrel,number of treated diseased vessels,and symptom factors of internal obstruction of phlegm and the blood stasis.This model has good differentiation and calibration,good clinical applicability,and provides reference indicators with TCM characteristics for early screening and intervention in people at high risk of ISR after stenting,which helps clinicians to adopt TCM methods for prevention and intervention at an earlier stage,and has certain clinical significance and reference value for prevention and reduction of ISR.