Primary Aldosteronism and Risk of Cardiovascular Outcomes: Genome‐Wide Association and Mendelian Randomization Study

Abstract

Background

Observational studies have reported associations between primary aldosteronism (PA) and cardiovascular outcomes, including coronary artery diseases (CAD), congestive heart failure (CHF), and stroke. However, establishing causality remains a challenge due to the lack of randomized controlled trial data on this topic. We thus aimed to investigate the causal relationship between PA and the risk of developing CAD, CHF, and stroke.

Methods and Results

Cross‐ancestry meta‐analysis of genome‐wide association studies combining East Asian and European ancestry (1560 PA cases and 742 139 controls) was conducted to identify single‐nucleotide variants that are associated with PA. Then, using the identified genetic variants as instrumental variables, we conducted the 2‐sample Mendelian randomization analysis to investigate the causal relationship between PA and incident CAD, CHF, and stroke among both East Asian and European ancestry. Summary association results were extracted from large genome‐wide association studies consortia. Our cross‐ancestry meta‐analysis of East Asian and European populations identified 7 genetic loci significantly associated with the risk of PA, for which the genes nearest to the lead variants were CASZ1, WNT2B, HOTTIP, LSP1, TBX3, RXFP2, and NDP. Among the East Asian population, the pooled odds ratio estimates using these 7 genetic instruments of PA were 1.07 (95% CI, 1.03–1.11) for CAD, 1.10 (95% CI, 1.01–1.20) for CHF, and 1.13 (95% CI, 1.09–1.18) for stroke. The results were consistent among the European population.

Conclusions

Our 2‐sample Mendelian randomization study revealed that PA had increased risks of CAD, CHF, and stroke. These findings highlight that early and active screening of PA is critical to prevent future cardiovascular events.

Nonstandard Abbreviations and Acronyms

APA

aldosterone‐producing adenoma

MR

Mendelian randomization

PA

primary aldosteronism

UKB

UK Biobank

Clinical Perspective.

What Is New?

• We conducted the first 2‐sample Mendelian randomization analysis using 7 genetic loci significantly associated with the risk of primary aldosteronism (PA) identified in our updated cross‐ancestry meta‐analysis of genome‐wide association studies including 1560 (PA) cases and 742 139 controls.

• Our 2‐sample Mendelian randomization revealed that PA had increased risks of coronary artery disease, congestive heart failure, and stroke.

What Are the Clinical Implications?

• Our findings using novel genetic data associated with PA, along with previous results from well‐designed observational studies, provide robust evidence about the cardiovascular burden of PA.

• These findings underscore the necessity of early systematic screening for PA, as a promising avenue for effective cardiovascular prevention.

• Further investigations are needed to examine the heterogeneity of our findings by PA subtypes, somatic mutations, and population.

Primary aldosteronism (PA) is a common cause of secondary hypertension, affecting approximately 5% to 10% of patients with hypertension and often manifesting at a relatively young age. ¹ Characterized by the excessive production of aldosterone, PA can be caused by an aldosterone‐producing adenoma (APA) or bilateral adrenal hyperplasia. ¹ Because APA can be effectively addressed through surgical resection of the adrenal gland and bilateral adrenal hyperplasia can be managed with pharmacological therapy, timely detection and suitable treatment of PA hold significant clinical importance. In PA, the regulatory feedback mechanisms between sodium levels and aldosterone are disrupted, leading to an increased risk of cardiovascular diseases. ² A previous meta‐analysis showed that PA was associated with 2‐ to 3‐fold increased risk of coronary artery disease, stroke, and heart failure. ³ Moreover, such cardiovascular burden could be both directly and indirectly (ie, through elevated blood pressure), ⁴ highlighting the need for renin‐angiotensin‐aldosterone system blockade beyond blood pressure management. However, in observational studies, establishing causality remains a challenge due to the lack of comparability between individuals with PA and those without.

Recently, 2 genome‐wide association studies (GWAS) identified several genetic variants that contribute to PA susceptibility across East Asian and European populations. ⁵ , ⁶ These discoveries have provided us with a better understanding of the underlying genetic and molecular mechanisms of PA, facilitating discussion on improved diagnostic and therapeutic strategies for this common endocrine disorder. ⁷ Furthermore, the identification of genetic variants related to PA enables us to apply Mendelian randomization (MR) analyses to explore the causal relationship between PA and adverse health outcomes. ⁸ , ⁹ Analogous to randomized controlled trial, the MR approach allows for the evaluation of the cumulative effect of a genetically determined PA on cardiovascular outcomes while minimizing the confounding biases commonly encountered in observational studies under the following 3 assumptions. ⁸ , ⁹ First, the genetic instruments are associated with the exposure of interest (the relevance assumption). Second, genetic instruments should not be associated with any confounders that could affect the exposure‐outcome relationship (the independence assumption). Third, the causal pathway from genetic instruments to the outcome of interest should be only through exposure and not through any other pathway (the exclusion restriction assumption).

In this study, we conducted 2‐sample MR analyses to investigate the causal relationships of PA with coronary artery disease, congestive heart failure, and stroke for the East Asian and European populations, respectively. To identify the genetic variants associated with PA and robustly select genetic instruments, we conducted a cross‐ancestry meta‐analysis of GWAS for PA. Quantifying the cardiovascular burden of PA would provide clinicians with additional insights into hypertension and cardiovascular management.

Methods

The data used in this study are available from within the manuscript or publicly accessible from the resources cited. The GWAS summary statistics of the UK Biobank (UKB) and FinnGen release 7 are available at https://www.ukbiobank.ac.uk/ and https://www.finngen.fi/fi.

Study Design

In this study, we employed a 2‐sample MR design using summary‐level data for East Asian ancestry and European ancestry. We considered PA to be the exposure in our MR study and cardiovascular disease (coronary artery disease, congestive heart failure, and stroke) to be the outcomes. In a 2‐sample MR analysis, ⁸ , ⁹ we examine the genetic instruments associated with the exposure using the first sample. Then we calculate the relationships between these genetic instruments and the outcome in the second sample and use these summary‐level data to estimate the causal relationship between the exposure and the outcome. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology‐MR checklist. ¹⁰ All individuals provided written informed consent in each cohort, and this study was approved by the ethical committee of Osaka University Graduate School of Medicine (no. 734‐16) and Hiroshima University (E2007‐9991‐15) and was conducted following the Declaration of Helsinki.

Data Sets

For the identification of single nucleotide variants (SNVs) associated with primary aldosteronism, we used our previous GWAS results from our Japanese cohort and UKB, ⁵ and publicly available GWAS summary statistics from FinnGen and French cohort. The detailed information on sample inclusion, quality control, and phenotype definitions in our previous GWAS was described elsewhere. ⁵ Briefly, our Japanese cohort included 392 PA cases from Hiroshima University Hospital and 67 363 controls from the BioBank Japan. PA was diagnosed based on the Japan Endocrine Society guidelines. ¹¹ , ¹² The BioBank Japan is a multi‐institutional, hospital‐based database including DNA, serum, and clinical details for individuals of Japanese origin, each diagnosed with at least 1 of 47 specific diseases. ¹³ The UKB includes health‐related information for approximately 500 000 individuals aged 40 to 69, recruited from across the United Kingdom between 2006 and 2010. ¹⁴ We defined PA cases based on the International Classification of Diseases, Tenth Revision (ICD‐10) code E26.0, ⁵ resulting in 84 PA cases and 376 236 controls from the UKB. The FinnGen project is a large‐scale public‐private genomic research initiative that collects genomic and health data from Finnish biobanks and digital health records from Finnish health registries. ¹⁵ We employed GWAS summary statistics of E4_HYPERALDO, which was based on ICD, Eighth Revision (ICD‐8) code 25500, ICD, Ninth Revision (ICD‐9) code 2551, and (ICD‐10) code E26, for PA. Our analysis included 522 PA cases and 297 590 controls from FinnGen. Lastly, we included the French cohort ⁶ in which patients were diagnosed as PA based on the Endocrine Society guidelines. ¹ We incorporated their GWAS summary statistics for PA based on 562 PA cases and 950 controls, which was obtained via GWAS Catalog (GCST90129620). To preserve the variant coverage to test, we did not use their meta‐analyzed summary statistics, in which genotype imputation was not performed. In total, 1560 PA cases and 742 139 controls were included in our analysis.

For the analysis of the association between the identified genetic instruments and cardiovascular outcomes for East Asian subjects, we obtained GWAS summary statistics for congestive heart failure, coronary artery disease, and stroke from previous studies using the first cohort of BioBank Japan between 2003 and 2008 ¹⁶ (which is different from the second cohort of BioBank Japan between 2013 and 2017 which we used for PA GWAS, and thus we avoided bias due to sample overlap ¹⁷ ). For European subjects, we obtained GWAS summary statistics for coronary artery disease from the meta‐analysis between UKB and the CARDIOGRAMPLUSC4D (Coronary Artery Disease Genome Wide Replication and Meta‐Analysis Plus the Coronary Artery Disease Genetics) consortium, ¹⁸ congestive heart failure from the HERMES (Heart Failure Molecular Epidemiology for Therapeutic Targets) consortium, ¹⁹ and stroke from the GIGASTROKE consortium. ²⁰ While the GIGASTROKE included non‐European populations in their meta‐analysis, we obtained the results of meta‐analysis only in European populations via GWAS Catalog (GCST90104541). We meta‐analyzed them with GWAS summary statistics from FinnGen for each outcome (described later in the article).

Selection of Genetic Instruments Associated With Primary Aldosteronism

We identified SNVs associated with PA through an updated cross‐ancestry GWAS meta‐analysis combining the most recent Japanese, UKB, FinnGen, and French GWASs (1560 PA cases and 726 761 controls). The detailed procedure for quality controls of variants was as previously described. ⁵ We employed the inverse variance weighted (IVW) method using METAL software (https://genome.sph.umich.edu/wiki/METAL). A significance threshold of genome‐wide association was set at the level of P=5.0×10⁻⁸. We assessed the inflation of test statistics by the genomic control factor λ_GC. The lead SNVs of individual genome‐wide significant loci were selected as genetic instruments. Indels were excluded due to the unavailability in the GWAS of certain cardiovascular outcomes and the possibility of misidentification between studies. Additionally, SNVs on X chromosome were excluded from genetic instruments in European ancestry because they were unavailable in the GWAS of the cardiovascular outcomes. After identifying genetic instruments for PA, we reconducted meta‐analysis among the European cohorts (UKB, FinnGen, and multicohort in Europe) to obtain the effects of each SNV on PA in Europe ancestry.

Meta‐Analysis for the Association Between PA‐Related SNVs and Cardiovascular Outcomes

We meta‐analyzed the GWAS results from the recent large meta‐analyses and FinnGen for the cardiovascular outcomes in European ancestry to ensure sufficient statistical power. We used FinnGen GWAS summary statistics of I9_CHD and I9_HEARTFAIL as coronary artery disease and congestive heart failure, respectively. This analysis was not applied to stroke because FinnGen cohort had already been included in the meta‐analysis in the GIGASTROKE consortium. ²¹ We applied the IVW method using the METAL software, after performing liftover from GRCh38 to GRCh37 for the FinnGen GWAS summary statistics.

Statistical Analysis

We conducted 2‐sample MR analyses using all PA‐related SNVs to assess the relationship between PA and cardiovascular outcomes (congestive heart failure, coronary artery disease, and stroke) for both East Asian and European subjects, respectively. In our main analysis, the IVW method was employed to estimate odds ratios (ORs) and 95% CIs for cardiovascular outcomes associated with PA using a fixed‐effects model. ²²

We then conducted the following sensitivity analyses. First, instead of the IVW method, we employed (1) the weighted median method, which allows us to obtain robust estimates even when up to 50% of the instruments consist of invalid SNVs; and (2) the weighted mode approach, which is less sensitive to the violations of MR assumptions by individual genetic instruments as it focused on the densest region of the causal effect estimates. Second, to assess the validity of the exclusion restriction assumption, we examined the presence of directional horizontal pleiotropy using the MR‐Egger regression. ²³ We also used the MR pleiotropy residual sum and outlier (MR‐PRESSO) method to detect and correct for horizontal pleiotropic outliers in multi‐instrument summary‐level MR testing. ²⁴ Lastly, we performed leave‐one‐out analyses by successively omitting each SNV to identify potential outlier SNVs.

All analyses were 2 sided and performed using TwoSampleMR and MendelianRandomization packages in R 4.2.3. Because these are 2‐sample MR analyses with binary exposure, the obtained estimates represent ORs for cardiovascular outcomes per 2.72‐fold increase in the odds of PA. ²⁵ The Benjamini–Hochberg method was applied to calculate the adjusted P‐value for multiple comparisons. ²⁶

Results

Cross‐Ancestry Genome‐Wide Meta‐Analysis for Primary Aldosteronism

We conducted an updated cross‐ancestry GWAS meta‐analysis comprising 1560 PA cases and 742 139 controls, targeting 4 777 457 autosomal and 114 446 X‐chromosomal SNVs. There was no evidence of excessive genomic inflation (λ_GC=1.02; Figure S1). We identified genome‐wide significant association signals at 7 loci (1p36, 1p13, 7p15, 11p15, 12q24, 13q12, and Xp11), for which the genes nearest to the lead variants were CASZ1, WNT2B, HOTTIP, LSP1, TBX3, RXFP2, and NDP (Figure 1, Table 1, Figure S2). The SNV with the strongest association was rs2146377 (OR 1.38 [95% CI, 1.27–1.49], P=3.92×10⁻¹⁵), of which the nearest gene is RXFP2. All the loci were reported to be associated with PA in 1 or both of the 2 recent GWAS, ⁵ , ⁶ which mutually reinforces the evidence of their findings. By searching the GWAS catalog database, we found that most of these SNVs were also associated with blood pressure or antihypertensive use. In addition, 2 SNVs (rs880315, rs2023843) were associated with other phenotypes such as cholesterol levels and albuminuria. ²⁷ , ²⁸ rs5906332 was excluded from genetic instruments in European ancestry because it is on the X chromosome. Then, 7 and 6 SNVs were selected as genetic instruments in the following MR analyses for East Asian and European groups, respectively.

Figure 1. Manhattan plot for the cross‐ancestry meta‐analysis of genome‐wide association studies of primary aldosteronism.

This is an updated cross‐ancestry meta‐analysis combining the most recent Japanese, UK Biobank, FinnGen, and French GWAS (1560 PA cases and 726 761 controls). The horizontal axis represents the genome in physical order and the vertical axis shows −log₁₀(P value) for association of individual variants with PA in the cross‐ancestry meta‐analysis. The red horizontal line indicates the genome‐wide significance threshold (P=5.0×10⁻⁸). The lead variants of genome‐wide association loci were marked with their nearest genes. GWAS indicates genome‐wide association studies; and PA, primary aldosteronism.

Table 1.

Lead SNVs for Genome‐Wide Significant Loci Associated With PA Risk in the Cross‐Ancestry Meta‐Analysis

SNV	Chromosome	Position (GRCh 37)	Band	Gene	Alleles	Risk allele	Cohort	RAF		OR (95% CI)	P value
								PA	Control
rs880315	1	10796866	1p36	CASZ1	T/C	C	Japanese	0.70	0.67	1.14 (0.97–1.35)	1.19×10−1
							UK Biobank	0.39	0.34	1.26 (0.91–1.73)	1.61×10−1
							FinnGen	0.45	0.41	1.15 (1.02–1.31)	2.36×10−2
							French	NA	NA	1.60 (1.36–1.87)	9.10×10−9
							Meta‐analysis			1.26 (1.16–1.37)	2.76×10−8
rs3790604	1	113046879	1p13	WNT2B	C/A	A	Japanese	0.37	0.29	1.41 (1.19–1.67)	8.42×10−5
							UK Biobank	0.10	0.07	1.50 (0.84–2.68)	1.72×10−1
							FinnGen	0.22	0.17	1.38 (1.19–1.60)	1.63×10−5
							French	NA	NA	1.22 (0.91–1.64)	1.93×10−1
							Meta‐analysis			1.38 (1.24–1.52)	1.32×10−9
rs2023843	7	27243221	7p15	HOTTIP	C/T	T	Japanese	0.66	0.59	1.41 (1.20–1.65)	2.88×10−5
							UK Biobank	0.96	0.93	1.61 (0.91–2.87)	1.02×10−1
							FinnGen	0.93	0.90	1.43 (1.14–1.78)	1.89×10−3
							French	NA	NA	1.29 (0.95–1.75)	1.06×10−1
							Meta‐analysis			1.40 (1.25–1.58)	1.63×10−8
rs4980379	11	1888614	11p15	LSP1	C/T	T	Japanese	0.70	0.61	1.43 (1.21–1.69)	2.33×10−5
							UK Biobank	0.42	0.36	1.27 (0.93–1.74)	1.37×10−1
							FinnGen	0.47	0.40	1.28 (1.14–1.45)	3.48×10−5
							French	NA	NA	1.43 (1.23–1.67)	5.24×10−6
							Meta‐analysis			1.35 (1.25–1.47)	6.58×10−14
rs35427	12	115556307	12q24	TBX3	T/G	T	Japanese	0.82	0.75	1.52 (1.26–1.84)	1.30×10−5
							UK Biobank	0.67	0.62	1.24 (0.90–1.71)	1.80×10−1
							FinnGen	0.68	0.64	1.24 (1.09–1.41)	1.00×10−3
							French	NA	NA	1.30 (1.10–1.54)	2.47×10−3
							Meta‐analysis			1.31 (1.20–1.43)	9.28×10−10
rs2146377	13	32179063	13q12	RXFP2	G/A	A	Japanese	0.60	0.57	1.17 (0.99–1.37)	6.64×10−2
							UK Biobank	0.65	0.58	1.34 (0.99–1.83)	5.92×10−2
							FinnGen	0.62	0.53	1.40 (1.24–1.58)	2.68×10−8
							French	NA	NA	1.56 (1.33–1.83)	3.34×10−8
							Meta‐analysis			1.38 (1.27–1.49)	3.92×10−15
rs5906332	X	43833438	Xp11	NDP	A/G	G	Japanese	0.20	0.18	1.08 (0.92–1.27)	3.62×10−1
							UK Biobank	0.50	0.48	1.30 (0.90–1.88)	1.61×10−1
							FinnGen	0.71	0.65	1.19 (1.07–1.32)	1.61×10−3
							French	NA	NA	1.62 (1.41–1.86)	4.59×10−12
							Meta‐analysis			1.28 (1.19–1.38)	5.00×10−11

NA indicates not available; OR, odds ratio; PA, primary aldosteronism; RAF, risk allele frequencies; and SNV, single nucleotide variant.

Association Between Primary Aldosteronism and Risk of Coronary Artery Diseases, Congestive Heart Failure, and Stroke Using Mendelian Randomization Approach

Our 2‐sample MR analysis of East Asian ancestry using all PA‐related SNVs showed that PA was associated with increased odds of coronary artery disease (OR, 1.07 [95% CI, 1.03–1.11]), congestive heart failure (OR, 1.10 [95% CI, 1.01–1.20]), and stroke (OR, 1.13 [95% CI, 1.09–1.18]; Table 2, Figure 2, Figure S3). In the sensitivity analyses using different methods (ie, the weighted median method, the weighted mode method, the MR‐Egger method, and the MR‐PRESSO approach), the point estimates were similar to those in the IVW methods, whereas the 95% CIs of some estimates were larger than our main analysis and included the null. The intercepts in the MR‐Egger analysis were not statistically significant for all cardiovascular outcomes, indicating that direct horizontal pleiotropy was unlikely to be present. In the leave‐one‐out analyses, we did not find any genetic variants that dominated the estimated effects for cardiovascular outcomes (Figure S4).

Table 2.

Association Between Primary Aldosteronism and Risk of Coronary Artery Diseases, Congestive Heart Failure, and Stroke Using Mendelian Randomization Approach

East Asian population	Coronary artery diseases		Congestive heart failure		Stroke
	OR (95% CI)	P value*	OR (95% CI)	P value*	OR (95% CI)	P value*
IVW method	1.07 (1.03–1.11)	<0.001	1.10 (1.01–1.20)	0.021	1.13 (1.09–1.18)	<0.001
Weighted median method	1.08 (1.04–1.13)	<0.001	1.05 (0.98–1.13)	0.126	1.13 (1.08–1.19)	<0.001
Weighted mode method	1.07 (1.03–1.12)	0.023	1.06 (0.99–1.13)	0.082	1.14 (1.07–1.20)	<0.001
MR‐Egger method
Effect estimate	1.14 (1.07–1.21)	<0.001	1.12 (0.91–1.37)	0.287	1.11 (1.01–1.21)	0.045
Intercept	−0.021	0.060	−0.005	0.877	0.008	0.812
MR‐PRESSO approach†	1.07 (1.03–1.11)	<0.001	1.06 (0.98–1.15)	0.144	1.13 (1.09–1.18)	<0.001

European population	Coronary artery diseases		Congestive heart failure		Stroke
	OR (95% CI)	P value*	OR (95% CI)	P value*	OR (95% CI)	P value*
IVW method	1.04 (1.02–1.06)	<0.001	1.08 (1.04–1.11)	<0.001	1.08 (1.02–1.15)	0.012
Weighted median method	1.03 (1.01–1.05)	0.014	1.08 (1.04–1.11)	<0.001	1.07 (0.99–1.16)	0.096
Weighted mode method	1.03 (1.00–1.05)	0.138	1.03 (0.99–1.07)	0.216	1.02 (0.93–1.12)	0.663
MR‐Egger method
Effect estimate	0.98 (0.89–1.08)	0.721	0.92 (0.80–1.06)	0.377	0.75 (0.55–1.03)	0.225
Intercept	0.017	0.236	0.046	0.039	0.110	0.063
MR‐PRESSO approach†	1.04 (1.02–1.06)	<0.001	1.08 (1.04–1.11)	<0.001	1.08 (1.02–1.15)	0.012

IVW indicates inverse variance weighted; MR, Mendelian randomization; MR‐PRESSO, MR pleiotropy residual sum and outlier; and OR, odds ratio.

^* P value was adjusted for multiple comparisons using the Benjamini–Hochberg method.

^† rs35427 was excluded as outliers for congestive heart failure among East Asian subjects. No SNVs were excluded for other outcomes.

Figure 2. Plot of Mendelian randomization analyses regarding the association between primary aldosteronism and risk of coronary artery diseases, congestive heart failure, and stroke.

A, The association between PA and coronary artery disease among East Asian population. B, The association between PA and congestive heart failure among East Asian population. C, The association between PA and stroke among East Asian population. D, The association between PA and coronary artery disease among European population. E, The association between PA and congestive heart failure among European population. F, The association between PA and stroke among European population. MR indicates Mendelian randomization; PA, primary aldosteronism; and SNV, single nucleotide variant.

In our 2‐sample MR analysis of European ancestry, we also found the association of PA with coronary artery disease (OR, 1.04 [95% CI, 1.02–1.06]), congestive heart failure (OR, 1.08 [95% CI, 1.04–1.11]), and stroke (OR, 1.08 [95% CI, 1.02–1.15]) (Table 2, Figure 2, Figure S3). We also found consistent results in the weighted median, the weighted model, and the MR‐PRESSO approach but not in the MR‐Egger analysis. For the assessment of the direct horizontal pleiotropy, although the intercepts for congestive heart failure in the MR‐Egger analysis were statistically significant, no SNVs were excluded in the MR‐PRESSO approach. In the leave‐one‐out analyses, we did not find any genetic variants that dominated the estimated effects for cardiovascular outcomes (Figure S4).

Discussion

In this MR study, we uncovered compelling evidence linking PA with an increased risk of coronary artery disease, congestive heart failure, and stroke among individuals of both East Asian and European descent. Through triangulating these results with previous findings from observational studies, our study facilitates a better understanding of the cardiovascular burden of PA and underscores the necessity of early systematic screening for PA, as a promising avenue for effective cardiovascular prevention.

To our understanding, this study is among the first to use the MR approach to explore the possible effects of PA on cardiovascular outcomes. This methodology, using genetic variants related to the exposure as instrumental variables, allowed us to infer potential causality while minimizing the risk of confounding bias, often inherent in observational studies. ⁸ , ⁹ Previously, a meta‐analysis involving 31 observational studies suggested a potential increased risk of coronary artery disease, heart failure, and stroke among patients with PA. ³ Although these are well‐designed studies, the estimates could be biased due to uncontrolled confounding given the nature of observational studies. Particularly, because testing for PA is rare even for patients with treatment‐resistant hypertension and those tested are more likely to have higher blood pressure, ²⁹ the exposed group (PA group) in our study could be selected high‐risk population than the unexposed group (non‐PA group) in observational studies such as cohort and case–control studies. Although our estimated ORs are around 1.10, which is smaller than those previously reported in the meta‐analysis mentioned previously (coronary artery disease, 1.77; heart failure, 2.05; stroke, 1.13), it is important to note that these 2 sets of estimates are not directly comparable. Specifically, the estimates from our 2‐sample MR represent ORs for cardiovascular outcomes per a 2.72‐fold increase in the odds of PA, rather than ORs based on having PA or not. ²⁵

Traditionally, PA has been known as a common cause of secondary hypertension, affecting approximately 5% to 10% of all patients with hypertension and between 15% and 20% of those with resistant hypertension. ³⁰ A recent genome‐wide association study among the Japanese cohort reported that 66.7% of the genetic variants associated with blood pressure showed higher ORs for the PA risk than for the hypertension risk. ⁵ Given potential underdiagnosis and underrecognition of this common endocrine disorder, ³¹ our study highlights the necessity of suspecting the presence of PA (and evaluating renin and aldosterone levels) for all clinicians involved in the management of hypertension. ⁷

Several mechanisms might elucidate the deleterious impact of excessive aldosterone, relative to sodium levels, on the cardiovascular system, beyond its role in electrolyte and blood pressure regulation. First, aldosterone excess can promote inflammation and fibrosis within cardiac and vascular tissues. ³² , ³³ It catalyzes the production of inflammatory cytokines and growth factors, thus contributing to vascular inflammation and structural remodeling. Second, aldosterone excess can compromise endothelial function—the single cell layer lining the interior of blood vessels. ³⁴ , ³⁵ Such a dysfunction could suppress the production of the vasodilator, nitric oxide, while simultaneously escalating the generation of vasoconstrictive agents. Third, aldosterone excess can induce reactive oxidative stress, ³⁶ potentially inflicting damage on cardiac and vascular cells. Cumulatively, these pathological derangements have the potential to precipitate left ventricular hypertrophy, coronary artery calcification, and arterial stiffness in patients with PA.

In our updated cross‐ancestry GWAS meta‐analysis, we identified some SNVs that have been implicated in mineralocorticoid receptor activity or adrenal differentiation. For example, WNT2B and NDP play key roles in the Wnt/β‐catenin signaling pathway. This pathway is crucial for adrenal cortex homeostasis, particularly in the zona glomerulosa region, and for aldosterone production. ³⁷ In addition, previous studies have reported the important role of CASZ1 in the pathophysiology of PA through aldosterone biosynthesis and transcriptional activity of mineralocorticoid receptor activity. ³⁸

The present study also sets the stage for several subsequent lines of research that need to be pursued. First, it is plausible that the cardiovascular burden might vary according to the somatic mutation of APA. For instance, some studies have indicated that KCNJ5 mutation ³⁹ —one of the most common somatic mutations of APAs—might be linked with more severe cardiovascular outcomes compared with APAs without KCNJ5 mutation. ⁴⁰ , ⁴¹ Second, as our study used the binary conception of primary aldosteronism, there is also a need for additional MR studies using the information on continuous aldosterone levels under a sodium‐controlled diet to further understand the biological mechanisms involving renin‐independent aldosterone production. ⁴² Such studies could help validate previous observational findings concerning their association (including the spectrum of subclinical PA) with incident hypertension and cardiovascular events. ⁴ , ⁴²

Limitations of the Study

Our study has several limitations that warrant consideration. First, our analysis used SNVs determined by the updated cross‐ancestry meta‐analysis in the MR study for each ancestry under the assumption that causal variants are consistent across ancestries. Second, discrepancies in methods of association analysis between cohorts, including different software and covariates, could potentially influence results. In the present study, the allele frequencies of some PA‐related variants were different across the cohorts. Because our GWAS primarily aimed to identify risk‐associated variants used for subsequent MR analysis, we did not perform fine‐mapping or incorporating heterogeneity modeling in our analytical models. This might partly explain the different allele frequencies of some of the PA‐associated variants across cohorts. Third, variations exist in case definitions between cohorts (ICD‐10 in UK Biobank, cohort‐specific definitions in FinnGen, and clinical diagnoses according to guidelines in French and Japanese cohorts) that may lead to exposure misclassification across cohorts. A previous retrospective study showed that even patients with resistant hypertension are not frequently screened for PA, raising a concern about underdiagnosis of PA. ²⁹ However, such misclassification (ie, some patients with PA are misclassified in the non‐PA group) generally underestimates the PA–outcome relationship. Additionally, this bias could be mitigated by the large size of correctly classified control groups in our study, known as “positive‐unlabeled problem.” ⁴³ Fourth, our primary IVW method and the MR‐Egger method yielded inconsistent findings among the European subjects. Considering several differences between East Asian and European populations such as sodium intake levels and clinical guidelines for PA (including screening procedures and pharmacological therapy ¹ , ¹¹ , ¹² ), the potential heterogeneity in the cardiovascular burden of PA across populations should be the subject of future research. Fifth, as rs880315 and rs2023843 were associated with not only PA and blood pressure but also other phenotypes (ie, cholesterol levels and albuminuria), our findings might suffer from bias due to horizontal pleiotropy. Indeed, although the results of MR‐PRESSO approach were consistent with our findings of IVW methods, the intercept of MR‐Egger method was marginally significant for some outcomes. However, we found consistent results in our leave‐one‐out analysis, suggesting that there is no single SNV that had a large horizontal pleiotropic effect influencing the MR estimates. We did not conduct manual pruning of SNVs that may have horizontal pleiotropic effects because such pruning could lead to an instrument that is no longer biologically meaningful, and thus, is not generally recommended. ⁴⁴ Sixth, our study did not separately assess the causal relationship of the 2 main forms of PA (ie, APA and bilateral adrenal hyperplasia) with cardiovascular outcomes. Accurate and comprehensive diagnoses of APA and bilateral adrenal hyperplasia, based on imaging and adrenal venous sampling, are of paramount importance in future research. Seventh, we did not conduct multivariable MR due to the limited number of SNVs. Given the potential direct effect of PA on cardiovascular health not through hypertension, ⁴ further detailed investigations (eg, direct comparisons with essential hypertension, multivariable MR, and mediation analysis) are needed with large cohorts. Lastly, our study does not provide quantitative evidence about the clinical usefulness of treating PA. There is a strong need for randomized controlled trials to assess the clinical usefulness of surgical treatment via adrenalectomy or medical therapy to cure PA and prevent future cardiovascular events in the general population.

Conclusions

In conclusion, our MR analysis provides robust evidence supporting the association between PA and elevated risks of cardiovascular and cerebrovascular events, such as coronary artery disease, heart failure, and stroke. These findings significantly emphasize the necessity for early identification and targeted treatment initiation for PA. Further investigations are needed to examine the heterogeneity of our findings by PA subtypes, somatic mutations, and populations, and explore whether early intervention and treatment can indeed lead to a reduction in the occurrence of cardiovascular events.

Sources of Funding

Kosuke Inoue was supported by grant 22K17392 and 23KK0240 from the Japan Society for the Promotion of Science, the Japan Agency for Medical Research and Development (AMED; JP22rea522107), and the Program for the Development of Next‐generation Leading Scientists with Global Insight (L‐INSIGHT) sponsored by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. Study sponsors were not involved in study design, data interpretation, writing, or the decision to submit the article for publication. Yukinori Okada was supported by JSPS KAKENHI (22H00476), and AMED (JP22ek0410075, JP23km0405211, JP23km0405217, JP23ek0109594, JP23ek0410113, JP223fa627002, JP223fa627010, JP233fa627011, JP23zf0127008), JST Moonshot R&D (JPMJMS2021, JPMJMS2024), Takeda Science Foundation, Bioinformatics Initiative of Osaka University Graduate School of Medicine, Institute for Open and Transdisciplinary Research Initiatives, Center for Infectious Disease Education and Research (CiDER), and Center for Advanced Modality and DDS (CAMaD), Osaka University. Study sponsors were not involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the article; or decision to submit the article for publication.

Disclosures

None.

Supporting information

Figures S1–S4