Establish A Fisher Discriminant Model To Predict The Outcome Of Postoperative Blood Pressure In Primary Aldosteronism

Objective: To explore the predictive factors for the outcome of postoperative blood pressure in primary aldosteronism(PA) and establish a model to predict the outcome before surgery by Fisher discriminant analysis.Methods: Collect data of patients with PA who accepted unilateral adrenal operation from January, 2010 to September, 2015 in the First Affiliated Hospital of Chongqing Medical University. A total of 83 cases were enrolled and divided into two groups: cure group and not cure group according to the postoperative blood pressure. The demographic characteristics, hypertensive information, biochemical tests and imaging characteristics were compared between two groups. Those factors which showed significantly difference were selected into Fisher stepwise discriminant analysis to establish a discriminant model. The model was compared with Aldosteronoma Resolution Score(ARS) and nomogram model by receiver operating characteristic curve(ROC).Results: 1) Hypertension was cured in 52 cases(62.65%), and 31(37.35%) remain hypertension. 2) Uncured group was older, had larger body mass index(BMI), longer duration of hypertension, higher serum triglyceride(TG), more antihypertensive drugs classes, higher incidence of diabetes, smoking and alcohol intake, also less typical nodule on computed tomography(CT) imaging, lower estimated glomerular filtration rate(eGFR) and high density lipoprotein cholesterol(HDL-C)(P<0.05). 3) BMI, number of the antihypertensive medications, typical nodule on CT, eGFR and TG were incorporated into the Fisher discriminant model. The cutoff value was 0.1595, with a sensitivity of 86.5% and a specificity of 83.9%. Area under the curve(AUC) was 0.857, higher than that of ARS(AUC 0.733) and the nomogram model(AUC 0.735).Conclusion: The Fisher discriminant model had a high value to help predict the outcome of postoperative blood pressure in PA.