Predictive model of hypertension resolution after adrenalectomy in primary aldosteronism: the SPAIN-ALDO score

Resumen:
Purpose: To develop a predictive model of hypertension resolution after adrenalectomy in patients with primary aldosteronism (PA), based on their presurgical characteristics. Methods: A retrospective multicenter study of PA patients in follow-up in 20 Spanish tertiary hospitals between 2018 and 2021 was performed (SPAIN-ALDO Register). Clinical response postadrenalectomy was classified according to the primary aldosteronism surgical outcome (PASO) consensus criteria. The predictive model was developed using a multivariate logistic regression model with the estimation of all possible equations. Results: A total of 146 patients (54.8% females; mean age of 51.5 +/- 10.9 years) with PA who underwent unilateral adrenalectomy were included. After a mean follow-up of 29.1 +/- 30.43 months after surgery, hypertension cure was obtained in 37.7% (n = 55) of the patients. The predictive model with the highest diagnostic accuracy to predict hypertension cure combined the variables female sex, use of two or fewer antihypertensive medications, hypertension grade 1, without type 2 diabetes and nonobesity. The area under the receiver operating characteristic curve of this model was 0.841 [0.769-0.914]. Based on this model, the group of patients with a higher probability of cure (80.4%) were those without type 2 diabetes, BMI <30 kg/m(2), female sex, hypertension grade 1 and who use two or fewer antihypertensives. Our predictive model offered a slightly higher diagnostic accuracy than Wachtel's (area under the curve [AUC]: 0.809), Utsumi's (AUC: 0.804) and Zarnegar's (AUC: 0.796) models and was similar than the Burello's (AUC: 0.833) model. Conclusion: Female sex, use of two or fewer antihypertensive medications, hypertension grade 1, no type 2 diabetes and nonobesity may predict hypertension cure after adrenalectomy in patients with PA. Our score provides a potential tool to guide preoperative patient counseling.