Researchers analyzed data from a large cohort of patients with hypertension to examine the association between plasma aldosterone levels and the risk of cardiovascular disease.

The study found that higher aldosterone levels were associated with an increased risk of cardiovascular disease, and the relationship followed a nonlinear dose-response pattern.

These findings suggest that aldosterone may be a useful marker for identifying patients with hypertension at higher cardiometabolic risk.

The study drew on a retrospective cohort of 7,677 hospitalized patients with hypertension who were free of cardiometabolic multimorbidity at baseline, recruited between January 2010 and January 2026 at a specialized hypertension center in Xinjiang, China. Over an average follow-up of 3.52 years, 1,949 participants developed cardiometabolic multimorbidity, defined in this context as the coexistence of at least two major cardiometabolic disorders such as coronary heart disease, stroke, or type 2 diabetes. This large sample size and multi-year observation window allowed the investigators to model the exposure–outcome relationship with considerable statistical precision, including the shape of the dose–response curve rather than assuming a simple linear trend.

The central quantitative finding was that each 1-ng/dL increment in plasma aldosterone concentration corresponded to a 2% higher hazard of incident cardiometabolic multimorbidity after multivariable adjustment. When participants were grouped into quartiles of baseline aldosterone, the risk climbed steadily across categories: compared with the lowest quartile, the adjusted hazard ratios were 1.19 for the second quartile, 1.32 for the third, and 1.65 for the highest, with a highly significant trend across quartiles. Kaplan–Meier curves reinforced this gradient, showing that participants in the top aldosterone quartile accumulated incident multimorbidity at a visibly faster rate than those in the lower three quartiles throughout follow-up.

Restricted cubic spline analysis added an important nuance: the association was not strictly linear. The fitted curve identified an estimated inflection point at approximately 15.34 ng/dL, below which the relationship between aldosterone and multimorbidity risk was relatively flat, and above which the risk curve rose appreciably. This kind of threshold-like behavior is clinically meaningful because it suggests that modest variation in aldosterone within the lower part of the distribution may carry little prognostic consequence, whereas concentrations beyond the inflection point mark a segment of the hypertensive population in which risk escalates more steeply. The nonlinearity was statistically robust, with a P value for nonlinearity below 0.001.

A second layer of complexity emerged from subgroup analyses, which indicated that the strength of the aldosterone–multimorbidity association differed according to serum potassium status, with a statistically significant interaction. To probe this further, the researchers performed restricted cubic spline analyses stratified by potassium level and fitted a bivariate penalized spline model to characterize the joint nonlinear associations of serum potassium and aldosterone with multimorbidity risk. This approach reflects the tight physiological coupling between aldosterone and potassium homeostasis: aldosterone promotes renal potassium excretion, so circulating potassium may modify, or serve as a marker of, the downstream tissue effects of aldosterone excess. The observed interaction implies that a given aldosterone concentration may not carry the same prognostic weight in a hypokalemic patient as in one with normal or high potassium, a consideration that could matter when aldosterone-based stratification is eventually translated into practice.

Beyond association, the study tested whether aldosterone adds genuine predictive value to conventional risk assessment. Using time-dependent receiver operating characteristic analysis with inverse probability of censoring weighting—a technique that accounts for censored observations when evaluating discrimination over time—the investigators found that adding plasma aldosterone to a baseline prediction model raised the area under the curve for 3-year risk from 0.678 to 0.735, a statistically significant improvement. Reclassification metrics told a similar story: the survival-based net reclassification improvement was 0.372, and the integrated discrimination improvement was 0.038, both statistically significant. Calibration assessment and decision curve analysis extended the evaluation, with the latter suggesting that a model incorporating aldosterone would yield greater estimated net benefit than the conventional model across a range of clinically relevant threshold probabilities. In parallel, the Boruta algorithm, a wrapper-based feature selection method built on random forests, independently identified plasma aldosterone as an important predictor among the variables considered, providing a machine-learning perspective that complemented the regression-based findings.

The biological rationale for these observations rests on the expanding view of aldosterone as more than a regulator of sodium and water balance. As the terminal effector of the renin–angiotensin–aldosterone system, aldosterone acts on mineralocorticoid receptors in the kidney, vasculature, heart, and other tissues. Experimental and clinical work has linked aldosterone excess to endothelial dysfunction, oxidative stress, inflammation, and myocardial fibrosis through both mineralocorticoid receptor–dependent and receptor-independent pathways. These mechanisms plausibly connect aldosterone to the individual components of cardiometabolic multimorbidity—atherosclerotic vascular disease, cerebrovascular events, and disturbances of glucose and lipid metabolism—and may explain why elevated aldosterone would promote the co-occurrence of several such conditions in the same patient.

Importantly, the adverse associations observed in this cohort were not confined to patients with overt aldosterone excess. Primary aldosteronism, the classic syndrome of autonomous aldosterone production, is well established as a driver of disproportionately high rates of stroke, coronary heart disease, and heart failure. But prior studies in populations in which primary aldosteronism had been excluded have shown that higher aldosterone concentrations remain continuously associated with adverse cardiovascular outcomes even within the so-called normal range. The present findings extend that literature by focusing specifically on the transition from single-disease hypertension to cardiometabolic multimorbidity, a stage at which mortality risk rises above that of any single cardiometabolic condition alone.

Several methodological features strengthen the study’s internal validity. The multivariable Cox proportional hazards models adjusted for a panel of potential confounders, and the findings remained consistent across subgroup and sensitivity analyses, suggesting that the aldosterone association was not an artifact of a particular analytic choice or a vulnerable patient subgroup. The combination of classical epidemiological tools—Cox regression, Kaplan–Meier estimation, restricted cubic splines—with modern prediction metrics such as net reclassification improvement, integrated discrimination improvement, and decision curve analysis represents a fairly comprehensive appraisal of whether a biomarker is worth adding to existing risk models, addressing not only statistical association but also potential clinical utility.

Nevertheless, the retrospective design imposes limits on interpretation. Plasma aldosterone was measured at a single baseline time point in hospitalized patients, and single measurements may not capture long-term hormonal status, particularly given that aldosterone secretion varies with sodium intake, posture, circadian rhythm, and concurrent medications, including the antihypertensive agents themselves. Residual confounding by factors such as dietary patterns, adherence to therapy, and socioeconomic characteristics cannot be excluded in an observational cohort. The cohort was drawn from a single hypertension center in northwestern China, and the generalizability of the 15.34 ng/dL inflection point to other populations, laboratories, and assay platforms remains to be established, since aldosterone reference intervals can vary with measurement method. Reverse causation is another consideration: subclinical cardiometabolic disease may itself influence aldosterone regulation, although the exclusion of patients with prevalent multimorbidity at baseline and the consistent findings across sensitivity analyses partially mitigate this concern.

From a clinical standpoint, the study positions plasma aldosterone as a candidate for risk stratification among the large population of hypertensive patients who carry residual cardiovascular risk despite adequate blood pressure control. The improvement in the area under the curve from 0.678 to 0.735, while meaningful, indicates that aldosterone complements rather than replaces conventional predictors, and the decision curve results suggest that the added discrimination could translate into better-targeted preventive decisions at clinically relevant thresholds. The potassium-dependent modification of the association adds a practical dimension: any future use of aldosterone for stratification may need to be interpreted alongside serum potassium rather than in isolation.

Prospective validation in independent cohorts, ideally with repeated aldosterone measurements and diverse populations, would be the natural next step. Interventional questions also follow: if aldosterone marks a mechanistically relevant pathway rather than a passive correlate of risk, then mineralocorticoid receptor antagonism or other strategies that suppress aldosterone signaling might be tested for their ability to reduce the incidence of cardiometabolic multimorbidity in hypertensive patients with concentrations above the identified inflection point. Until such evidence accumulates, the study’s principal contribution is the demonstration that a routinely measurable hormone, long appreciated for its role in volume regulation, carries independent, nonlinear, and clinically quantifiable prognostic information about the burden of coexisting cardiometabolic disease in hypertension.

Subject of Research: Aldosterone levels may predict cardiometabolic risk in hypertension

Article Title: Dose–response association and predictive value of plasma aldosterone concentration for cardiometabolic multimorbidity in patients with hypertension: a retrospective cohort study

Article References: Zhang, Z., Li, N., Zhu, Q., Hu, J., Hong, J., Yang, W., Zhou, P., Ma, H., Ma, R., & Wu, T. (2026). Dose–response association and predictive value of plasma aldosterone concentration for cardiometabolic multimorbidity in patients with hypertension: a retrospective cohort study. Frontiers in Endocrinology, 17, Article 1965927. https://doi.org/10.3389/fendo.2026.1965927

Image Credits: AI Generated

DOI: 10.3389/fendo.2026.1965927

Keywords: aldosterone, hypertension, cardiovascular disease, cardiometabolic risk, cohort study, Dose, response, association, predictive, value, plasma, concentration