Incidence of Atrial Fibrillation and Mineralocorticoid Receptor Activity in Patients With Medically and Surgically Treated Primary Aldosteronism

Key Points

Question  Can the risk of patients with primary aldosteronism developing atrial fibrillation be lowered with mineralocorticoid receptor (MR) antagonists or surgical adrenalectomy?

Findings  In this large cohort study, patients with primary aldosteronism treated with MR antagonists had a 2.5-fold higher risk for incident atrial fibrillation when their renin remained suppressed (a marker of insufficient MR blockade) compared with matched patients with essential hypertension. In contrast, patients with primary aldosteronism whose renin substantially increased with MR antagonists and patients who underwent surgical adrenalectomy had no significant difference in incident atrial fibrillation risk compared with an age-matched cohort of patients with essential hypertension.

Meaning  Activation of the MR by aldosterone may play an important role in the development of atrial fibrillation, and adequate blockade or removal of this aldosterone may prevent incident atrial fibrillation.

Importance  Primary aldosteronism (PA) is an ideal condition to evaluate the role of the mineralocorticoid receptor (MR) in the pathogenesis of atrial fibrillation (AF).

Objective  To investigate whether MR antagonist therapy or surgical adrenalectomy in PA influence the risk for incident AF.

Design  This cohort study included patients aged 18 years and older. Patients with PA and age-matched patients with essential hypertension were identified via electronic health records. Patients with a history of AF, myocardial infarction, congestive heart failure, or stroke were excluded. Data were collected between 1991 and the end of 2016 in an academic medical center, with a mean follow-up duration of approximately 8 years.

Exposures  Patients with PA treated with MR antagonists or surgical adrenalectomy were compared with patients with essential hypertension. Patients with PA who were treated with MR antagonists were categorized by whether their plasma renin activity remained suppressed (< 1 ng/mL/h) or substantially increased (≥ 1 ng/mL/h), as proxies for insufficient or sufficient MR blockade.

Main Outcomes and Measure  Incident AF.

Results  A total of 195 patients with PA who were treated with MR antagonists and 201 patients with PA treated with surgical adrenalectomy were included, as well as 40 092 age-matched patients with essential hypertension. Despite similar blood pressure at study entry and throughout follow-up, patients with PA who were treated with MR antagonists whose renin remained suppressed had a higher risk for incident AF than patients with essential hypertension (adjusted HR, 2.55 [95% CI, 1.75-3.71]). They also had an adjusted 10-year cumulative AF incidence difference of 14.1 (95% CI, 6.7-21.5) excess cases per 100 persons compared with patients with essential hypertension. In contrast, patients with PA who were treated with MR antagonists and whose renin increased and patients with PA who were treated with surgical adrenalectomy had no statistically significant difference in risk for incident AF compared with patients with essential hypertension.

Conclusions and Relevance  When compared with patients with essential hypertension, patients with PA treated with MR antagonists such that renin remained suppressed (as a proxy for insufficient MR blockade) had a significantly higher risk for incident AF; however, treatment of PA with MR antagonists to substantially increase renin (suggesting sufficient MR blockade), or with surgical adrenalectomy (to remove the source of aldosteronism), was associated with no significant difference in risk for developing AF. These findings add to the growing body of evidence suggesting that MR blockade may be a potential therapy to decrease the incidence of AF.

Atrial fibrillation is the most common cardiac arrhythmia; it increases the risk for adverse cardiovascular outcomes such as stroke and reduced cardiac output. Prior studies have demonstrated that long-term aldosterone exposure promotes the development of atrial fibrillation by inducing cardiac fibrosis and conduction disturbances via activation of the mineralocorticoid receptor (MR).^1-7 Recent evidence suggests that blockade of the MR with medications such as spironolactone and eplerenone may provide a new therapeutic approach to prevent or delay the development of atrial fibrillation.^8-10

Primary aldosteronism (PA), a state of autonomous aldosterone secretion,¹¹ offers a classic example of the detrimental effects of chronic and excessive MR activation on the development of atrial fibrillation. Untreated patients with PA have a 3.5-fold higher risk for incident atrial fibrillation compared with patients with similar blood pressures.¹² Although adrenalectomy as a treatment for PA is associated with decreased risk for atrial fibrillation when compared with patients with essential hypertension,⁷ the risk for atrial fibrillation despite lifelong MR antagonist therapy has been reported as high.^7,13 Herein, we conduct a large retrospective cohort study that includes physiologic biomarkers of medical treatment efficacy to examine the risk for incident atrial fibrillation in patients with PA who were treated with MR antagonists or surgical adrenalectomy, compared with patients with essential hypertension.

We performed a cohort study of patients with PA and patients with essential hypertension using the electronic health records at Brigham and Women’s Hospital, Massachusetts General Hospital, and their affiliated partner hospitals (eFigure 1 in the Supplement).¹³ Eligibility for the current analyses required that patients had to be seen between 1991 and 2016 and be 18 years or older. Entry into the study was defined by condition: for patients with PA treated with MR antagonists, by the date of the first plasma renin activity (PRA) measurement at least 1 month after starting MR antagonist therapy; for patients with PA treated with surgical adrenalectomy, by the first follow-up visit 1 to 6 months after undergoing surgical adrenalectomy; and for patients with essential hypertension, by the first follow-up visit 1 to 6 months after the diagnosis of essential hypertension was first entered into the medical record.

This study was approved by the Partners Healthcare System institutional review board. A waiver of informed consent was granted to develop this deidentified cohort.

Patients with PA who were treated with MR antagonists were subdivided by whether PRA remained suppressed at the first measurement at least 1 month after starting MR antagonist therapy (where suppression was defined as < 1 ng/mL/h) or increased substantially to become unsuppressed (≥ 1 ng/mL/h). These test results were considered proxies for insufficient and sufficient blockade, respectively, of the MR in PA.

Patients were excluded if they had a history of atrial fibrillation, myocardial infarction or coronary revascularization, transient ischemic attack or stroke, or congestive heart failure hospitalization prior to study entry. Patients with PA who were treated with MR antagonists who and did not have PRA measurements at least 1 month after initiating treatment were also excluded. Patients with essential hypertension were also excluded if they were treated with MR antagonists at study entry. The population with essential hypertension was frequency-matched by decade of age at study entry to the study population with PA who were treated with MR antagonists.

The analysis investigated the risk for incident atrial fibrillation among patients with PA who were treated with MR antagonists or surgical adrenalectomy compared with patients with essential hypertension. We used multivariate Cox regression models (using the PHREG procedure in SAS version 9.4 [SAS Institute]) to estimate adjusted hazard ratios (HRs) and 95% CIs. All models were adjusted for the following covariates at the time of study entry: age, sex, race, body mass index (calculated as weight in kilograms divided by height in meters squared), diabetes mellitus, estimated glomerular filtration rate, and systolic blood pressure. Patients were censored on the date of incident atrial fibrillation or, if they did not develop atrial fibrillation, on the date of their final follow-up visit. All P values are 2-sided and considered significant if .05 or less. A more detailed description of the study cohort, exposure definition, outcome ascertainment, and statistical analysis is available in the eMethods in the Supplement.

A total of 195 patients with PA who were treated with MR antagonists (including 130 whose PRA remained suppressed and 65 whose PRA substantially increased to unsuppressed levels), 201 patients with PA treated with surgical adrenalectomy, and 40 092 patients with essential hypertension were included in the study. Compared with patients with essential hypertension, patients with PA who were treated with MR antagonists were by design of similar age (patients treated with MR antagonists with suppressed PRA: mean [SD] age, 57.1 [12.3] years; patients treated with MR antagonists with PRA at an unsuppressed level: 56.6 [13.0] years; patients with essential hypertension: 56.8 [11.8] years); in contrast, patients with PA treated surgically were younger (mean [SD] age, 49.3 [10.9] years; Table 1). Patients with PA who were treated surgically had a more severe biochemical PA phenotype compared with patients with PA who were treated with MR antagonists (patients treated with MR antagonists with suppressed PRA: median [interquartile range] serum aldosterone levels, 19.0 [13.0-42.6] ng/dL; patients treated with MR antagonists with PRA at an unsuppressed level: 23.0 [17.0-30.0] ng/dL; surgically treated patients: 30.6 [21.3-46.0] ng/dL). Patients with PA had higher blood pressure prior to study entry (patients with suppressed PRA: mean [SD] systolic blood pressure, 149 [18] mm Hg; patients with PRA at an unsuppressed level: 146[16] mm Hg; surgically treated patients: 145 [17] mm Hg) than patients with essential hypertension (mean [SD] systolic blood pressure, 138 [19] mm Hg); however, at the time of study entry and throughout the follow-up period, blood pressure was similar between patients with PA and those with essential hypertension (patients treated with MR antagonists with suppressed PRA: mean [SD] systolic blood pressure at study entry, 139 [21] mm Hg; patients treated with MR antagonists with PRA at an unsuppressed level: 133 [16] mm Hg; surgically treated patients: 133 [16] mm Hg; patients with essential hypertension: 135 [18] mm Hg; eFigure 2 in the Supplement). Patients with PA whose PRA level became unsuppressed with MR antagonist therapy were prescribed higher-potency equivalents of MR antagonists than patients with PA whose PRA had remained suppressed (by a measure in which mean MRA equivalent potency was calculated by combining spironolactone and eplerenone doses into a single value, with spironolactone considered to be twice the potency of eplerenone; patients with unsuppressed PRA: initial dose: mean [SD], 95 [52] mg/day; maximum dose: 161 [113] mg/day; patients with suppressed PRA: initial dose: mean [SD], 77 [53] mg/day; maximum dose: 149 [135] mg/day; Table 1).

Patients with PA who were treated with MR antagonists and whose PRA remained suppressed had a higher risk for incident atrial fibrillation compared with patients with essential hypertension (adjusted HR, 2.55 [95% CI, 1.75-3.71]) (Figure). In contrast, patients with PA who were treated with MR antagonists and whose PRA rose to become unsuppressed had no statistically significant difference in risk for incident atrial fibrillation compared with patients with essential hypertension (adjusted HR, 1.03 [95% CI, 0.54-2.00]); the same was true for patients with PA who were treated with surgical adrenalectomy (adjusted HR, 0.75 [95% CI, 0.41-1.36]) (Figure; eTable 1 and eTable 2 in the Supplement).

The adjusted 10-year cumulative incidence difference for developing incident atrial fibrillation indicated that patients with PA who were treated with MR antagonists and whose PRA remained suppressed had 14.1 (95% CI, 6.7-21.5) excess cases of atrial fibrillation per 100 persons than patients with essential hypertension (Table 2). In contrast, patients with PA who were treated with MR antagonists and whose PRA became unsuppressed, and patients with PA who were treated with surgical adrenalectomy, had no significant difference in the adjusted 10-year cumulative incidence of atrial fibrillation.

The findings in this study expand the evidence implicating aldosterone and the MR in the pathogenesis of atrial fibrillation. Without targeted treatment, PA is associated with a 3.5-fold higher risk for developing atrial fibrillation than essential hypertension, independent of blood pressure.¹² We previously showed that even when MR antagonists are used in PA, the risk for atrial fibrillation in these patients is 1.9-fold higher than in patients with essential hypertension.¹³ However, the current analyses demonstrate that this risk may be modified by the way medical treatment is implemented: when PA is treated with MR antagonists such that renin remains suppressed, the risk for developing atrial fibrillation is 2.5-fold higher than in age-matched patients with essential hypertension and similar blood pressures. In contrast, treating PA with MR antagonists to induce a substantial increase in renin or with surgery to remove the source of autonomous aldosterone secretion results in no significant difference in the risk for developing atrial fibrillation. Renin suppression in PA is a result of excessive MR activation and sodium and volume expansion,^11,14 and the increase in renin with MR antagonist therapy reflects sufficient antagonism of the MR and decreased sodium reabsorption and volume contraction. Indeed, an increase in renin with MR antagonists in PA has previously been shown to be associated with a lower risk for myocardial infarction, stroke, heart failure, and death.¹³

The current study provides clinically relevant evidence from an extreme state of hyperaldosteronism that using MR antagonists to adequately block the MR, or removal of the unilateral source of aldosterone excess via surgical adrenalectomy, may mitigate the excess risk for atrial fibrillation attributed to pathologic MR activation. These findings are in agreement with recent studies in essential hypertension and heart failure,^8-10,15 which suggest that blockade of the MR may be a potential target for preventing or delaying the incidence of atrial fibrillation and stroke. Our findings may also help contextualize prior studies in cohorts with PA that did not include an evaluation of physiological biomarkers of MR blockade and observed that lifelong MR antagonist therapy did not lower the risk for atrial fibrillation, whereas surgical adrenalectomy did.⁷

Our results must be interpreted within the limitations of this observational study design. Despite blood pressure and other risk factors being similar at study entry and our models adjusting for important confounders, there may still be residual confounding and misclassification. Although the population of patients with PA was relatively large compared with prior cohort studies in PA, the absolute sample sizes of the study populations were still small and could therefore limit the reliability of our conclusions. Although we performed extensive medical record reviews to confirm the atrial fibrillation outcomes, the reliance on medical record documentation for this ascertainment cannot completely exclude some misclassification. Because this study used an open cohort design, we could not capture all atrial fibrillation events that occurred outside of our health care system, which may have resulted in an underestimation of cumulative incidence rates. Patients with PA who were treated with surgery were on average 7 years younger than the other exposure groups, a reflection of the fact that unilateral PA is often more severe and presents earlier in life. However, our models were adjusted for age, and the comparisons using the surgical exposure can be considered to strengthen the differences observed among patients treated with MR antagonists, who were matched by age to patients with essential hypertension. Finally, we defined renin suppression using a binary categorization, rather than using a continuous analysis of renin values, because most patients with PA had renin activity values that were below the lower limit of the assay (0.60 ng/mL/h).

In conclusion, the excess risk for atrial fibrillation that is attributed to aldosterone-MR interactions in PA may be mitigated by MR antagonist therapy that sufficiently blocks aldosterone or by surgical adrenalectomy to remove the source of aldosterone excess. These findings provide additional support for the role of aldosterone and the MR in the pathogenesis of atrial fibrillation and support further investigations to evaluate whether MR antagonists may be an effective option to prevent the onset of atrial fibrillation in the general hypertension population.

Corresponding Author: Anand Vaidya, MD, MMSc, Center for Adrenal Disorders, Division of Endocrinology, Diabetes, and Hypertension, Department of Medicine, Brigham and Women’s Hospital, 221 Longwood Ave, RFB 287, Boston, MA 02115 (anandvaidya@bwh.harvard.edu).

Accepted for Publication: May 29, 2018.

Published Online: July 18, 2018. doi:10.1001/jamacardio.2018.2003

Author Contributions: Drs Hundemer and Vaidya had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Hundemer, Curhan, Vaidya.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Hundemer, Vaidya.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Hundemer, Curhan, Wang, Vaidya.

Obtained funding: Hundemer, Vaidya.

Administrative, technical, or material support: Hundemer, Vaidya.

Study supervision: Curhan, Vaidya.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Vaidya reports receiving grants from National Institutes of Health and the Doris Duke Charitable Foundation and personal fees from The Endocrine Society. Dr Hundemer reports receiving grants from National Institutes of Health/ National Institute of Diabetes and Digestive and Kidney Disease. Dr Curhan reports grants and personal fees from Allena Pharmaceuticals, personal fees from Shire and RenalGuard, and royalties from UpToDate, where he is a section editor and author. No other disclosures were reported.

Funding/Support: This study was supported by the National Institute of Diabetes and Digestive and Kidney Disease of the National Institutes of Health (grant F32 DK114953, Dr Hundemer; grant K24 DK091417, Dr Curhan; and grants R01 DK115392 and R01 DK107407, Dr Vaidya) and the Doris Duke Charitable Foundation (grant 2015085, Dr Vaidya).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.