Purpose: To evaluate whether the clinical, biochemical and radiological features of patients with primary aldosteronism (PA) can predict both main subtypes of PA.Methods: A retrospective multicenter study of PA patients followed in 27 Spanish tertiary hospitals (SPAIN-ALDO Register). Only patients with con rmed unilateral or bilateral PA based on adrenal venous sampling (AVS) and/or postsurgical biochemical cure after adrenalectomy were included. Supervised regression techniques were used for model development.Results: 328 patients [270 unilateral PA (UPA), 58 bilateral PA (BPA)] were included. The area under the curve (AUC) for aldosterone/potassium ratio and aldosterone responses following saline infusion test were 0.602 [95%CI 0.520 to 0.684] and 0.574 [95% CI 0.446-0.701], respectively, to differentiate UPA from BPA. The AUC was 0.825 [95% 0.764-0.886] when the prediction model with seven parameters -comorbidities (dyslipidemia, cerebrovascular disease, SAS), SBP, plasma aldosterone levels (PAC), hypokalemia and unilateral adrenal nodule >1 cm and normal contralateral adrenal gland on CT/MRI -was used. In patients without comorbidities, hypokalemia, SBP >160 mmHg, PAC >40 ng/dL, and unilateral adrenal lesions were associated with a likelihood of having a UPA of 98.5%. The chance of BPA was higher in individuals with comorbidities, SBP <140 mmHg, normokalemia, low PAC levels, and no adrenal tumors on the CT/MRI (91.5%).Conclusion: A combination of high PAC, SBP >160 mmHg, low serum potassium, a unilateral adrenal nodule and no comorbidities could predict a UPA with a 98.5% accuracy.