Primary aldosteronism (PA) is the most common cause of endocrine hypertension. Although adrenal venous sampling (AVS) is recommended as the gold standard procedure for subtype classification in PA, it is a specialized technique with limited availability. The objective of this study was to develop a scoring system that predicted PA subtype using clinical characteristics. Seventy-one patients with PA were studied. The subjects were diagnosed as having either unilateral (n=32) or bilateral disease (n=39) based on AVS, surgery and/or the postoperative clinical course. Variables associated with laterality in the univariate analysis were entered into multivariable logistic regression models and the regression coefficients were used to construct a subtype prediction score. The diagnostic significance of the score was then evaluated using receiver operating characteristic (ROC) curve analysis. The subtype prediction score was calculated as follows: serum potassium ⩽3.4 mEq l(-1), 2 points; plasma aldosterone concentration ⩾165 pg ml(-1), 3 points; and aldosterone to renin ratio ⩾1000 in a post-captopril challenge test (plasma renin activity in ng ml(-1) h(-1)), 3 points. ROC curve analysis for the ability to discriminate between unilateral and bilateral PA showed that a score of 5 points had 75% sensitivity and 95% specificity, and a score of 3 points had a sensitivity of 97% and a specificity of 59%. The area under the ROC curve was 0.920 (95% confidence interval, 0.859-0.979). Our subtype prediction score could discriminate between unilateral and bilateral PA and is useful for selecting patients who should undergo AVS before surgery.