Study participants with an AI were recruited among 376 consecutive adult patients hospitalized in the Department of Endocrinology and Internal Medicine of the University Clinical Center of the Medical University of Gdańsk between November 2018 through February 2020 due to an adrenal lesion. We included 64 patients with radiological features of an adrenal adenoma/hyperplasia revealed by computed tomography (CT) or magnetic resonance (MR), who agreed to participate in the study, and met none of the following exclusion criteria: 1) age over 75 or under 40; 2) obesity grade III (BMI >40 kg/m²); 3) premenopausal period; 4) adrenal hormone excess other than MACS, i.e. cortisolemia between 50 and 138 nmol/l in the overnight 1-mg dexamethasone suppression test (DST) and no phenotypic features of Cushing’s syndrome (17); 5) kidney disease with eGFR<60 ml/min/1.73m² and/or proteinuria >0.25 g/24h); 6) treatment with a mineralocorticoid receptor antagonist; 7) established and/or overt CVD other than primary HT, including: a) ASCVD (CAD, stroke, transient ischemic attack, peripheral artery disease), b) significant cardiac disease (e.g. pathological arrhythmia, severe valvular heart disease, cardiac tamponade, cardiomyopathy, congenital heart disease, HF), c) vascular diseases (among others venous thromboembolism and vasculitis); 8) active malignancy; 9) decompensated autoimmune disease or immune disease associated with CV and/or renal complications; 10) infectious diseases; 11) current or past addiction to alcohol and/or illicit drugs. Study participants were recruited based on anamnesis, physical examination, additional examinations available for review prior to enrollment and performed in the course of the study. Initially, 73 patients were included, however, three withdrew their consent to participate due to the COVID-19 pandemic, in four patients transthoracic echocardiography (TTE) revealed cardiac post-ischemic lesions, and two were diagnosed with primary aldosteronism.

Based on medical records of our hospital, which included examinations ordered in outpatient clinics and the emergency department, we identified 153 persons with normal adrenal morphology in a CT/MRI scan performed within five years preceding this study. There were 129 who met at least one of the above-listed exclusion criteria, declined participation or were unreachable, therefore, 24 subjects without an AI were enrolled as controls.

The research complied with the Declaration of Helsinki and was approved by the Independent Bioethics Committee for Research of our University. Informed consent for inclusion in the study was obtained in writing from each participant.

Both AI patients and controls underwent the following evaluation: 1) medical interview; 2) physical examination; 3) antecubital venous blood sampling for laboratory analyses; 4) resting 12-lead electrocardiography (ECG); 5) TTE; 6) common carotid artery (CCA) ultrasonography (USG) including CIMT determination, 7) 24-hour ambulatory blood pressure monitoring (ABPM).

Body-mass-index (BMI) was calculated by dividing body weight (W) in kg by the square of height (H) in meters. 2009 International Diabetes Federation criteria were used to diagnose MetS (18). Subjects with HT received hypotensive medications at the time of enrollment or were diagnosed by ABPM based on mean systolic and diastolic BP (SBP and DBP, respectively) of at least 135/85 mmHg for daytime, 120/75 mmHg for nighttime, and/or 130/80 mmHg for the 24-h period (15). Atherogenic dyslipidemia was defined as triglycerides (TGL) ≥150 mg/dL and serum high-density lipoprotein cholesterol (HDL-C) <40 mg/dL for men and <45 mg/dL for women.

In all study participants, 10-year ACSVD risk was estimated using the 2018 calculator provided online by the American Heart Association and the American College of Cardiology based on Framingham Heart Study (FHS-ASCVD Risk) (19, 20). The calculator estimates 10-year risk of developing ASCVD including coronary death, myocardial infarction, coronary insufficiency, angina, ischemic stroke, hemorrhagic stroke, transient ischemic attack, peripheral artery disease, and HF for individuals aged 30 to 74 and without CVD at baseline based on the following predictors: age, type 2 DM (DMt2), smoking, treated and untreated SBP, total cholesterol (TC), HDL-C, and LDL-C.

For nondiabetic subjects, 10-year CVD risk was also calculated using Systematic Coronary Risk Estimation 2 (SCORE2) for subjects aged 40-69 and SCORE2-Older People (SCORE2-OP) for those aged 70-75 for high CV risk countries, which include Poland (21). Predictors used in SCORE2/-OP are: age, sex, smoking, SBP, and non-HDL-C. We are aware SCORE2/-OP was developed to estimate risk in treatment-naïve persons, and that a significant portion of our subjects received lipid- and BP-lowering therapy. Nevertheless, we concluded applying this estimation tool along with FHS-ASCVD Risk calculation is of value.

Blood was drawn between 8 and 10 a.m. after a fasting period of at least 8 hours from an antecubital vein, and used for regular examinations in the laboratory of our hospital apart from samples preserved for the determination of plasma Cts in all subjects, serum aldosterone and plasma direct renin concentration (DRC) in controls. These were centrifuged at 2,000 rpms for 20 minutes at 4 degrees C, aliquoted and stored at -80 degrees C until analysis.

Samples were analyzed in Central Clinical Laboratory in Gdańsk using standard laboratory methods (with a Siemens IMMULITE 1000 Immunoassay System for most biochemical tests, and an Abbott Architect analyzer, which applies the spectrophotometric method). Serum Cts was determined by an enzyme-linked immunosorbent assay (ELISA) by using a commercially-available diagnostic kit (SunRedBio, catalogue no: 201-12-8276; sensitivity: 0.268 ng/mL; assay range: 0.3-90 ng/mL). Cts concentrations above 90 ng/ml (n=15) were extrapolated based on ELISA standard curve.

Serum Cts, creatinine, sodium, potassium, aldosterone, renin, lipid profile (TC, HDL-C, LDL-C, and TGL), UA, hs-CRP, 24-h urinary protein and albumin excretion were determined both in AI patients and controls. Morning serum cortisol, dehydroepiandrosterone sulphate (DHEA-S), overnight 1 mg-DST cortisol and 24-h urinary cortisol were determined in AI patients. In most (n=50) AI patients, 24-h urinary meta- and normetanephrine excretion was determined, in others it had been performed prior to hospitalization. Screening for primary hyperaldosteronism based on aldosterone-to-renin ratio (ADRR) was performed without modifying antihypertensive medications in both AI patients and controls; there were no study participants with both HT and an ADRR above 2 ng/dL:μIU/mL.

24h ABPM was conducted using a Spacelabs Ontrak 90227 monitor on the non-dominant arm. During the day BP was recorded every 20 minutes, while during nighttime rest every 30 minutes. ABPM was repeated or not considered in the analysis if more than 30% of measurements were invalid. Normal results were adopted according to the European Society of Cardiology/European Society of Hypertension 2018 guideline: <130/80 mmHg for the 24-h period, <135/85 mmHg for daytime, and <120/70 mmHg for nighttime (22). Patients were classified as ‘non-dippers’ if their mean diurnal SBP and DBP were not at least 10% higher than nocturnal (22).

All measurements were performed in accordance with the recommendations endorsed by the American Society of Echocardiography and the European Association of Cardiovascular Imaging (23). Three on-site cardiology consultants with an expertise in ultrasonography performed TTE using the GE Vivid E9/E95 ultrasound system.

Measurements included left-ventricular (LV) internal dimension in diastole (LVIDd) and systole (LVIDs), LV ejection fraction (LVEF) according to modified Simpson’s rule (24), posterior LV wall thickness (LVPWd), and interventricular septal thickness (IVSd). LV mass (LVM) was calculated with the cube formula: LVM (g)=0.8×1.04 ×[(LVEDd+IVSd+LVPWd)³−LVEDd³]+0.6. LVM was indexed to body surface area (BSA) calculated using the DuBois formula (BSA=0.007184×H^0.725×W^0.425): LVM index (LVMI)=LVM/BSA. Relative wall thickness (RWT) was calculated with the formula: RWT=(2×LVPWd)/LVEDd. Left ventricular hypertrophy (LVH) was defined as LVMI >95 g/m² for females and >115 g/m² for males. RWT was used to further classify LVH as either concentric (RWT >0.42) or eccentric (RWT ≤ 0.42).

Disk summation technique from apical four and two-chamber views was used to determine left atrial volume (LAV), which was indexed to BSA: LAV index (LAVI) (ml/m²) = LAV/BSA (15). Apical four-chamber view was used to record peak blood flow velocity from LV relaxation in early diastole (E) and peak velocity flow in late diastole (A). Since LVEF was normal in all study participants, four criteria were applied to assess diastolic function: (1) LAVI ≥34 ml/m², (2) tricuspid regurgitation velocity (TR) ≥2.8 m/s, (3) ratio of E to average early mitral annular velocity (e’) ≥14, (4) septal e’<7 cm/s or lateral e’<10 cm/s. Indeterminate diastolic function was stated if two criteria were met, and dysfunction if three or four (15).

Maximum carotid intima-media thickness (CIMT) measurements were recorded using echo-tracking technology on the distal wall of the right carotid artery, 1 to 3 cm below the carotid artery bifurcation. The presence of atherosclerotic plaques (ASP) defined as a CIMT ≥1.5 mm, or by a focal increase in thickness of 0.5 mm or 50% of the surrounding CIMT value was also recorded.

Data were analyzed using R-studio. Discrete variables were presented as number (n) or n (percentage). Continuous quantitative data with a normal distribution were presented as mean ± standard deviation (SD), and in the case of a non-normal distribution as median (interquartile range, IQR). We used the Shapiro-Wilk test to determine if a data set was well-modeled by a normal distribution. To compare differences between two independent groups Welch’s t-test was used when variables were normally distributed or the Mann-Whitney U test in the case of non-normal distribution. One-way ANOVA and Tukey’s honestly significant difference (HSD) tests were used to compare three or more independent groups. Simple (bivariate) correlations were computed with the non-parametric Spearman rank-order method (correlation coefficient r is given). Associations between dichotomous categorical variables were examined with Fisher’s exact test, and Benjamini-Hochberg Method was applied to correct for multiple testing.

Multiple regression models were applied to adjust for differences in Cts concentrations depending on potential confounding variables including gender, age, BMI, smoking status, comorbidities (HT, DMt2, MetS), and medications (ACEI/ARB, CCB, BB, diuretics, statins, and PPIs). An exhaustive search method was used to select factors that had the strongest relationship with Cts, i.e.: 1) gender, presence of 2) MetS, 3) HT, therapy with 4) statins, and 5) PPIs. The final multivariate model had a R-squared of 0.1831. Out of five variables included in the model only the presence of MetS (ß=-30, p=0.005) was significantly and negatively associated with Cts. For dichotomous dependent variables (Cts halves, HT, DMt2, and MetS) binary logistic regression was used to adjust for gender, age, and BMI. Significance was set at 0.05.

To assess whether the presence of an AI affected Cts levels, verification of matching between AI patients and controls was undertaken. These groups did not differ in regard to age, sex, BMI, smoking status, and comorbidities (incidence of HT, DM t.2, ASP, and dyslipidemia), see Table 1 . Concerning subjects with HT, the number of patients on mono-, dual- and triple-drug therapy (including betablockers) was also comparable ( Supplementary Table 1 ).

Among AI patients, there were 14 with MACS and 31 classified as NFAI; analyses for AI patients and these two subgroups were performed separately.

Ten-year FHS-ASCVD Risk was comparable between patients with an AI/NFAI/MACS and controls, while in nondiabetic subjects, SCORE2/-OP was significantly higher in patients with MACS than in controls and patients with NFAI: 14% (11-18) vs. 8% (4.5-14), p = 0.021, and 8% (6-12), p = 0.005, respectively ( Table 1 ). Still, this CVD risk index was comparable between controls and all AI patients (p=0.31), which illustrates effective matching between these groups.

Cts distribution was bimodal both in AI patients and controls. Unadjusted Cts was slightly higher in AI patients: 6.45 (4.9-37) vs. 4.5 (3.5-28) ng/ml, p=0.047 ( Figure 1 ). However, after adjusting for potential confounding variables (gender, age, and BMI), solely BMI and male gender were significantly (negatively) associated with Cts (ß=-28.3, p=0.01 and ß=-2.3, p=0.04, respectively) but not the presence of an AI (ß=-8.1, p=0.44).

Lipid profile, hs-CRP, as well as UA were comparable between controls and AI patients, be it with a NFAI or MACS. Proteinuria and albuminuria were normal in all study participants (respectively below 150 and 30 mg/24h). ABPM parameters (SBP, DBP, and pulse rate, PR) were comparable between AI patients and controls ( Table 1 and Supplementary Table 1 ).

Concerning TTE, there were significant differences in IVSd, LVPWd, and LVMI between AI patients and controls (respectively 11 (10-12) vs. 10 (9-10) mm, p=0.003; 10 (9-11) vs. 9 (8-9.5) mm, p=0.007; 86.4 ± 119.2 vs. 84.7 ± 18.5 g/m², p=0.001). Moreover, LVH was more prevalent in MACS patients than controls (42.9% vs. 4.4%, p=0.007) and NFAI patients (42.9% vs. 14%, p=0.028).

Maximum CIMT was higher in patients with an AI, be it with a NFAI or MACS, than in controls: 1 (0.9-1.1) vs. 0.8 (0.8-0.9) mm, p<0.01. However, there were no differences in maximum CIMT between patients with a NFAI and MACS ( Table 1 ). A trend toward a higher prevalence of an ASP in AI, NFAI, and MACS patients (29.7%, 32%, 21.43%, respectively) than controls (9.5%) could be observed (p=0.12) ( Table 1 ).

Upon comparing Cts levels between controls with normal adrenal morphology and AI patients, peptide’s levels were tested in different patient groups. Cts was higher in women than in men: 7 (4.8-100) vs. 4.9 (4-7.4) ng/ml, p=0.015, and the difference between sexes was significant in both AI patients (7.3 (5.5-103) vs. 6 (4.26-7.6) ng/ml, p=0.03) and controls (5.1 (3.8 - 62.6) vs. 2.8 (1.7 - 3.5) ng/ml, p = 0.043), see Figure 2 .

Further, in AI patients and controls analyzed together Cts was lower in hyper- versus normotensive subjects: 5.6 (4-7.1) vs. 15.8 (5.2-103) ng/ml, p=0.003, which was also found for AI patients alone: 5.6 (4.36-6.82) vs. 21.7 (6.85-107), p < 0.001) ( Figure 3 ). Cts was also significantly lower in subjects with MetsS than in those without it: 25.7 (5.8-115) vs. 5.2 (3.9- 6.9) ng/ml, p<0.01 ( Figure 4 ), regardless of potential confounders (gender, age, BMI, presence of an AI and/or HT, statin and PPI use). We confirmed these differences (normo- versus hypertensive subjects as well as those without and with MetS) in women but not men (probably due to their low number). Cts in hypertensive AI females was lower than in normotensive ones: 5.6 (4.7 – 11.6) vs. 45.2 (8.2 – 118) ng/ml, p <0.01, and also lower in those with MetS than without it, both among AI patients: 5.6 (4.8-6.9) vs. 34.3 (7.7 - 121) ng/ml, p= <0.01 and controls: 3.8 (3.7-5.1) vs. 61.2 (8.4-112) ng/ml, p = 0.025 ( Supplementary Figures 1 , 2 ).

There were no differences in Cts between obese and non-obese subjects, smokers and non-smokers, or, among subjects with HT, ‘dippers’ and ‘non-dippers’.

To further investigate associations between Cts and CVD risk, correlations were tested between peptide’s levels and other parameters. In AI patients, weak correlations were found between Cts and: BMI (r=-0.31) ( Figure 5A ), FHS-ASCVD Risk (r=-0.42) ( Figure 5B ), and HDL-C (r=0.32) regardless of statin therapy ( Figure 5C ). Interestingly, among participants without it, there were also positive correlations between Cts and: TC and LDL-C (r=0.36 for both) ( Table 2 and Supplementary Figures 3 , 4 ). In AI patients and controls analyzed as a whole a negative correlation between Cts and UA was also observed (r=-0.27, p=0.01), while for each group analyzed separately, significance has not been reached probably due to sample size.

Analyses in subjects of each sex revealed Cts correlated with HDL (r=0.31, p=0.014), BMI (r=-0.29, p=0.019), and UA (r=-0.27, p=0.031) in women, but not in men (respectively r=0.13, p=0.53; r=-0.06, p=0.78; r=-0.12, p=0.56). A negative correlation between Cts and FHS-ASCVD Risk was also recorded in women with an AI (r=-0.3, p=0.049) but not in female controls (r =0.025, p=0.92), nor men with an AI (n=19, r=- 0.37, p=0.12). Correlations for male controls were not tested due to a low number of these subjects (n=5).

Concerning hormonal tests, only a weak correlation between Cts and DRC was observed, but not with aldosterone, nor ADRR ( Supplementary Table 2 ). Cts did not correlate with ABPM, CCA USG, nor TTE parameters (LVPWd, IVSd, LVMI, LAVI) ( Table 2 and Supplementary Table 2 ).

Further analyses were performed among AI patients based on Cts categories. First, adjustment for gender, age and BMI in binary logistic regression analysis revealed AI patients in the lower half of Cts concentrations (median 6.5, IQR 4.9-37 ng/ml) compared to those in the upper had a higher prevalence of HT (OR 0.17, CI 0.05-5.37, p=0.003), and MetS (OR 0.21, CI 0.06-7.51, p=0.018). Moreover, BMI, 24-SBP, and FHS-ASCVD Risk were also higher in the former (respectively 30.1 ± 4 vs. 27.2 ± 3.6 kg/m², p=0.004; 123 ± 7.4 vs. 117 ± 9.5 mmHg, p=0.022; 13.2%(8.9-19.2) vs. 6.3%(4.2-10.8), p=0.002), as summarized in Table 3 .

Second, based on Cts distribution, we divided AI participants into four subgroups, i.e. with: ‘very low’ (Cts <4.9 ng/ml, n=17), ‘low’ (≥4.9 and <6.5 ng/ml, n=15), ‘intermediate’ (≥6.5 and ≤45.2 ng/ml, n=17), and ‘high’ (≥100 ng/ml, n=15) Cts levels ( Table 3 ). The first two comprised subjects from two lower quarters, while the ‘high Cts’ subgroup corresponded to almost all patients in the fourth quarter (15 instead of 16 patients were included since there were none in the 45.2 - 100 ng/ml range, see Figure 1 ).

These four Cts subgroups differed significantly in male-to-female ratio, prevalence of HT and MetS, mean/median BMI, HDL-C, 24h SBP, and FHS-ASCVD Risk ( Table 3 ). Post hoc analysis revealed male gender was more prevalent in the ‘very low’ versus ‘high’ Cts subgroup (53% vs. 6.7%), while HT and MetS in the ‘very low’ versus ‘intermediate’ (82.4% vs. 35.3%, p=0.04 for both) and ‘high’ Cts subgroups (82.4% vs. 33.3%, p=0.04 for both). HDL-C was lower in the ‘very low’ than in the three remaining Cts subgroups (42.9 ± 42.9 vs. 61.6 ± 17.8, 56.7 ± 13.7, and 55.8 ± 9.5 mg/dL, adjusted p=0.001, 0.01, and 0.03, respectively). What is more, 24h SBP in the ‘low’ Cts subgroup was higher than in the ‘intermediate’ (124.7 ± 6.7 vs. 114.5 ± 8.5 mmHg, adj. p=0.008).

Ten-year FHS-ASCVD Risk in the ‘very low’ Cts subgroup was higher than in the ‘intermediate’ and ‘high’: 14.3% (19.2-24.5) vs. 7.1% (4.7-14.2) and 5.6% (3.8-9.8), respective adjusted p=0.014 and 0.005, in line with differences in gender proportions, prevalence of metabolic disorders and HT between the subgroups.

Fisher’s exact test revealed differences in LVH prevalence between four Cts subgroups, yet, without significance in pairwise comparisons with correction for multiple testing (Holm–Bonferroni method). No other significant differences were recorded between Cts halves and subgroups in TTE and CCA USG parameters ( Table 3 ). Finally, clinical, laboratory and TTE parameters were also analyzed in the same Cts subgroups for females only ( Supplementary Table 3 ). Significant differences were recorded for: HT and MetS prevalence, HDL-C and hs-CRP concentrations, mean 24h DBP, and FHS-ASCVD Risk. Calculations for male AI subjects were not performed due to their low number.