Cerebral Oxygen Saturation Monitoring In Stanford Type A Aortic Dissection Surgery And Prediction Of Neurological Dysfunction

Objective:To evaluate the clinical effect of cerebral oxygen saturation monitoring during the operation of Stanford type A aortic dissection,and to analyze the factors influencing the prognosis of neurological function,so as to provide the basis for clinicians to better prevent and deal with neurological complicationsMethod:According to the inclusion and exclusion criteria,who was admitted to the cardiovascular surgery department of the second hospital of Jilin University from May 2013 to June 2019.The same surgeon performed moderate hypothermia circulatory arrest(MHCA)and bilateral selective antegrade perfusion(bSACP).The patients with Stanford type A aortic dissection were screened and all the selected patients were enrolled in a retrospective cohort study.Thepropensity score matching methodwas used to reduce the confounding bias of baseline data,and the difference between the prognosis data of matched cerebral oxygen saturation monitoring group and non-cerebral oxygen saturation monitoring group was compared to evaluate the monitoring effect of near-infrared spectroscopy(NIRS).The patients monitored by NIRS were analyzed.Compare the left and right regional cerebral oxygen saturation.Therelationship between intra-operative cerebral oxygen saturation and postoperative neurological dysfunction was analyzed by multivariableLogistic regression.LASSO regression modelwas used for data dimension reduction,feature selection.Multivariable Logistic regression analysis was used to develop the predicting model,the related factors of postoperative neurological dysfunctionwas presented with a Nomogram.The performance of the Nomogram was assessed with respect to its calibration,discrimination,and clinical usefulnessResult:1.A total of 245 patients were included in the retrospective cohort study.Among them,13 patients(5.3%)died,16 patients(6.5%)had permanent neurological dysfunction,33 patients(13.5%)had temporary neurological dysfunction,31 patients(12.7%)had renal insufficiency treated with CRRT,and ICU stay was 136.25 hours(86.50-207.84),with an average hospitalization time of 20.00 days(16.00-26.00).2.According to whether the patients used near-infrared spectrometer during the operation,they were divided into 51 patients in the cerebral oxygen saturation monitoring group and 194 patients in the non-cerebral oxygen saturation monitoring group.The preoperative and intra-operative clinical data were mixed confounding bias between the two groups by propensity score matching method,and further use the weighted matching method to reduce the bias.A total of 47.56 cases in cerebral oxygen saturation monitoring group and 47.32 cases in non-cerebral oxygen saturation monitoring group were matched.There was no statistical difference between the two groups in preoperative and intra-operative data.In the weighted matching sequence,the incidence of postoperative permanent neurological dysfunction(PND)in the cerebral oxygen saturation monitoring group was significantly lower than that in the non-cerebral oxygen saturation monitoring group(2.0%vs 10.1%,P=0.014).In addition,the postoperative mortality(2.4%vs 12.1%,P=0.022),Continuous renal replacement therapy(CRRT)for renal dysfunction(5.6%vs 15.7%,P=0.052),ventilator-assisted ventilation time(38.64h vs 66.99h,P=0.028)were significantly lower than the non-cerebral oxygen saturation monitoring group.3.There was no significant difference in cerebral oxygen saturation between the left and right sides of 51 patients in the brain oxygen monitoring group(P>0.05).4.According topostoperative neurological complications,the patients were divided into two groups:transient neurological dysfunction(TND)group and non-neurological dysfunction(NND)group Multivariable logistic regression analysis showed that the minimum value of rSO2 was an independent factor of TND(OR:0.893,95%Cl:0.788?0.984,P=0.040).The area under curve(AUC)of the receiver operating curve(ROC)was 0.713,95%confidence interval was 0.506?0.921,the cutoff value was 49.50%,sensitivity was 55.6%,specificity was 86.8%.5.The results of Logistic regression analysis showed that CPB time was an independent predictor of NND(OR:0.985,95%Cl:0.978-0.996,P=0.003).Five potential predictors of PND were selected by lasso optimal error method.Multivariate Logistic regression analysis showed that age(OR:1.086,95%Cl:1.029-1.154,P=0.004),total leukocytes(OR:1.164,95%Cl:1.014-1.284,P=0.025),cardiopulmonary bypass time(OR:1.019,95%Cl:1.004-1.034,P=0.012),D-Dimer(OR:1.027,95%Cl:1.009-1.046,P=0.002)were independent predictors of postoperative PND.The PND diagnostic Nomogram was composed of age,total leukocyte count,cardiopulmonary bypass time and D-Dimer.The PND prediction model was validated by the original cohort(C-index=0.781)Conclusion?It is significant and effective to applying near-infrared spectrometer during the operation,which can reduce thepostoperative permanent neurological complications,ventilator-assisted ventilation time,renal insufficiency and mortality,and thus improve the prognosis of patients.The lowest intra-operative rSO2 was an independent factor of postoperative transient neurological dysfunction.In addition,this study proposes a Nomogram combined with preoperative and intra-operative risk factors,which can be used for the individualized prediction of postoperative permanent neurological dysfunction.