目的 探讨血浆醛固酮（ALD）在女性原发性高血压左心室肥厚（LVH）形成中的作用，并基于ALD及临床指标构建预测模型，以提高LVH的早期识别与风险评估能力。 方法 纳入原发性高血压女性患者200例，分为左心室肥厚组（LVH组）（n = 104）与非左心室肥厚组（NLVH组）（n = 96）。采集临床资料、实验室指标、ALD、动态血压监测（ABPM）参数。Spearman相关性分析ALD与左心室质量指数（LVMI）的关系。采用多因素logistic回归分析筛选独立预测因子并基于显著变量构建模型，模型1：ALD；模型2：ALD + 糖尿病 + 24 h平均收缩压（24hSBP） + 夜间收缩压（nSBP） + 血压负荷（BPL） + 夜间血压下降率（NBD）；模型3：ALD + 糖尿病 + 血糖（BG） + NT-proBNP + 24hSBP + 24 h平均舒张压（24hDBP）+动脉硬化指数（ASI）。以受试者工作特征（ROC）曲线评价区分度并进行验证。 结果 与NLVH组相比，LVH组患者的糖尿病患病率、BG、NT-proBNP、24hSBP、24hDBP及ALD水平均显著升高（均P < 0.05），夜间血压下降幅度明显减小（P < 0.05）。Spearman 分析显示ALD与LVMI呈显著正相关（r = 0.478，P < 0.001），在多因素logistic回归中仍为LVH独立危险因素（OR = 1.014，P = 0.002），糖尿病史、BG、NT-proBNP、24hSBP、24hDBP及ASI为LVH的预测因子。结合上述变量建立的模型ROC曲线下面积（AUC）为0.852，拟合良好（Hosmer?Lemeshow P > 0.05），预测效能优于单一指标。内部验证Bootstrap后证明模型有良好的预测性能。 结论 ALD升高是EH女性发生LVH的独立危险因素。结合临床指标及动态血压指标建立的多参数预测模型，可显著提高 LVH 的预测准确性，为女性高血压患者的个体化管理与早期干预提供依据。

Objective To investigate the role of plasma aldosterone （ALD） in the development of left ventricular hypertrophy （LVH） in women with essential hypertension （EH）， and to develop a predictive model incorporating ALD and clinical parameters to improve early detection and risk stratification of LVH. Methods A total of 200 female patients with EH were enrolled and classified into the LVH group （n = 104） and the NLVH group （n = 96）. Clinical characteristics， laboratory parameters， ALD levels， and ambulatory blood pressure monitoring （ABPM） data were systematically collected. Spearman′s rank correlation coefficient was used to evaluate the association between ALD levels and left ventricular mass index （LVMI）. Multivariate logistic regression analyses were conducted to identify independent predictors of LVH， with three models constructed： Model 1 included ALD alone； Model 2 adjusted for ALD， diabetes， 24-hour systolic blood pressure （24hSBP）， nighttime SBP （nSBP）， blood pressure load （BPL）， and nighttime blood pressure decline （NBD）； Model 3 further incorporated ALD， diabetes， blood glucose （BG）， NT-proBNP， 24hSBP， 24-hour diastolic blood pressure （24hDBP）， and arterial stiffness index （ASI）. The discriminative performance of each model was assessed using receiver operating characteristic （ROC） curve analysis， with internal validation performed to ensure robustness. Results Compared with the NLVH group， the LVH group exhibited a higher prevalence of diabetes and significantly elevated levels of BG， NT-proBNP，24hSBP， 24hDBP， and ALD， along with a smaller NBD （all P < 0.05）. Spearman′s rank correlation analysis revealed a statistically significant positive correlation between ALD and LVMI （r = 0.478， P < 0.001）. In multivariate logistic regression analysis， ALD remained an independent predictor of LVH （OR = 1.014， P = 0.002）， after adjusting for other significant variables including diabetes， BG， NT-proBNP， 24hSBP， 24hDBP， and ASI. The combined predictive model （Model 3） demonstrated the highest discriminative ability （AUC = 0.852； Hosmer?Lemeshow test P > 0.05）， surpassing models based on single predictors. Internal bootstrap validation confirmed excellent calibration and robustness of the model′s predictive performance. Conclusions Elevated plasma aldosterone concentration is an independent risk factor for LVH in women with essential hypertension. A multimodal predictive model that integrates aldosterone levels， clinical parameters， and dynamic blood pressure measurements can substantially enhance the accuracy of LVH prediction， offering a reliable approach for individualized management and early intervention in female patients with hypertension.