Multi-modal Ultrasound Evaluation Of Patent Foramen Ovale And Establishment Of A Risk Prediction Model

Background Patent foramen ovale(PFO)is one of the causes of cryptogenic cerebrovascular accident(CVA).However,PFO is a common abnormality of fetal origin,and not all PFOs carry a risk of disease.How to identify PFOs with pathogenic risk is a clinical challenge.Objective To explore the related risk factors of patent foramen ovale(PFO)with pathogenic risk,establish risk prediction model and risk scoring system,and provide reference for the identification and prevention of high-risk individuals.Methods A total of 111 patients with PFO who were evaluated based on multi-modal ultrasound in Yichang Central People’s Hospital from October 2018 to February 2021 were selected,including 46 patients in the CVA group and 65 patients in the asymptomatic control group.General clinical data,transesophageal echocardiography(TEE)morphological evaluation parameters and contrast-transthoracic echocardiography(c-TTE)functional evaluation parameters of PFO patients in two groups were collected.Univariate and multivariate analyses were used to identify the risk factors associated with pathogenic PFO,and the risk prediction model and risk scoring system were established.Receiver operating characteristic curve(ROC)was used to evaluate the predictive efficacy of the prediction model and the scoring system and to establish risk stratification for PFO patients.Results(1)Unifactor analysis: The comparison of general clinical data between the two groups showed that the incidence of hypertension,hyperlipidemia and women in CVA group was higher than that in asymptomatic control group(all P<0.05);Comparison of morphological parameters evaluated by TEE between the two groups showed that the incidence of large PFO,atrial septal tumor(ASA)and high atrial septal mobility was higher in the CVA group than that in the asymptomatic control group(all P<0.05);The comparison of functional parameters evaluated by c-TTE between the two groups showed that the incidence of right-to-left shunt(RLS)in resting state of CVA group was higher than that in asymptomatic control group(P<0.05).(2)Multivariate analysis: Multivariate Logistic regression analysis showed that large PFO,high atrial septal mobility,ASA,moderate to large RLS in resting state and hyperlipidemia were independent risk factors for PFO-related cerebrovascular accidents.(3)Establishment of risk prediction model: According to the risk factors and related risk coefficients established by multivariate Logistic regression analysis,the risk prediction model was Logit(P)=-2.964 + 1.922 × large PFO + 2.635 × ASA + 2.156× high mobility of atrial septum + 2.887 × moderate to large shunt in resting state + 1.505 ×hyperlipidemia.The results of Hosmer-lemeshow test was P= 0.861,and the area under ROC curve was 0.866(P ＜0.001,95%CI: 0.795-0.936).(4)Establishment of risk scoring system:According to the logistic regression analysis,the total score of the risk score system was 7.The ROC curve analysis showed that the area under the ROC curve was 0.845(P ＜ 0.001,95%CI: 0.771-0.920).When the threshold was 2(sensitivity 73.9%,specificity 86.2%),the Youden index reached the maximum value of 0.601,which was the best cutoff point for high-risk PFO patients.Conclusion PFO with large size,ASA,high mobility of atrial septum,moderate to large amount of RLS at rest and hyperlipidemia is at risk.The scoring system established in this study ranged from 0 to 7 points,and PFO patients with scores ≥2 points were identified as high-risk groups.The risk prediction model and scoring system constructed in this study are simple and feasible,which are helpful for the comprehensive evaluation and risk stratification of PFO in clinical practice,and provides a reference for the identification and prevention of high-risk individuals.