Objective: To establish and validate a nomogram-based predictive model for idiopathic hyperaldosteronism (IHA). Methods: This cross-sectional study was conducted with the collected clinical and biochemical data of patients with primary aldosteronism (PA) including 249 patients with unilateral primary aldosteronism (UPA) and 107 patients with IHA, who were treated at the Department of Endocrinology of the First Affiliated Hospital of Chongqing Medical University from November 2013 to November 2022. Plasma aldosterone concentration (PAC) and plasma renin concentration (PRC) were measured by chemiluminescence. Stepwise regression analysis was applied to select the key predictors of IHA, and a nomogram-based scoring model was developed. The model was validated in another external independent cohort of patients with PA including 62 patients with UPA and 43 patients with IHA, who were diagnosed at the Department of Endocrinology, First Affiliated Hospital of Zhengzhou University. An independent-sample t test, Mann-Whitney U test, and χ2 test were used for statistical analysis. Results: In the training cohort, in comparison with the UPA group, the IHA group showed a higher serum potassium level [M(Q1, Q3), 3.4 (3.1, 3.8) mmol/L vs. 2.7 (2.1, 3.1) mmol/L] and higher PRC [4.0 (2.1, 8.2) mU/L vs. 1.5 (0.6, 3.4) mU/L] and a lower PAC post-saline infusion test (SIT) [305 (222, 416) pmol/L vs. 720 (443, 1 136) pmol/L] and a lower rate of unilateral adrenal nodules [33.6% (36/107) vs. 81.1% (202/249)]; the intergroup differences in these measurements were statistically significant (all P<0.001). Serum potassium level, PRC, PAC post-SIT, and the rate of unilateral adrenal nodules showed similar performance in the IHA group in the validation cohort. After stepwise regression analysis for all significant variables in the training cohort, a scoring model based on a nomogram was constructed, and the predictive parameters included the rate of unilateral adrenal nodules, serum potassium concentration, PAC post-SIT, and PRC in the standing position. When the total score was ≥14, the model showed a sensitivity of 0.65 and specificity of 0.90 in the training cohort and a sensitivity of 0.56 and specificity of 1.00 in the validation cohort. Conclusion: The nomogram was used to successfully develop a model for prediction of IHA that could facilitate selection of patients with IHA who required medication directly.
目的： 建立和验证一个基于列线图的特发性醛固酮增多症（IHA）预测模型。 方法： 横断面研究。收集2013年11月至2022年11月于重庆医科大学附属第一医院确诊的原发性醛固酮增多症（PA）患者（训练队列）的临床和生化资料，包括249例单侧PA（UPA）和107例IHA。均采用化学发光法检测血浆醛固酮浓度（PAC）和血浆肾素浓度（PRC）。应用逐步logistic回归分析筛选预测IHA的关键变量，建立一个以列线图为基础的评分模型。该模型在另一个于郑州大学第一附属医院确诊的PA患者（包括62例UPA和43例IHA）的外部独立队列中进行验证。统计学处理采用独立样本t检验、Mann‐Whitney U检验及χ2检验。 结果： 训练队列中，与UPA患者相比，IHA患者有更高的血钾［M（Q1，Q3），3.4（3.1，3.8）比2.7（2.1，3.1）mmol/L］、立位PRC［4.0（2.1，8.2）比1.5（0.6，3.4）mU/L］；有更低的盐水负荷试验（SIT）后PAC［305（222，416）比720（443，1 136）pmol/L］、单侧肾上腺结节比例［33.6%（36/107）比81.1%（202/249）］，差异均有统计学意义（均P<0.001）。验证队列中，IHA患者的血钾、立位PRC、SIT后PAC、单侧肾上腺结节也有相似表现。对训练队列中差异有统计学意义的所有变量进行逐步回归分析后，建立了一个基于列线图的评分模型，其预测IHA的参数包括：单侧肾上腺结节、血钾、SIT后PAC和立位PRC。当得分≥14分时，该模型在训练队列中的敏感度、特异度分别为0.65、0.90，在验证队列中也得到相似的结果（敏感度、特异度分别为0.56、1.00）。 结论： 基于列线图，成功建立IHA预测模型，可筛选出IHA患者直接进行药物治疗。.