Mineralocorticoid Antagonism and Vascular Function in Early Autosomal Dominant Polycystic Kidney Disease: A Randomized Controlled Trial

Abstract

Rationale & Objective:

Vascular dysfunction, characterized by impaired vascular endothelial function and increased large-elastic artery stiffness, is evident early in autosomal dominant polycystic kidney disease (ADPKD) and is an important predictor of cardiovascular events and mortality. Aldosterone excess has been implicated in the development of endothelial dysfunction and arterial stiffness, in part by causing increased oxidative stress and inflammation. We hypothesized that aldosterone antagonism would reduce vascular dysfunction in patients with early-stage ADPKD.

Study Design:

Prospective, randomized, controlled, double-blind clinical trial

Setting & Participants:

61 adults 20–55 years of age with ADPKD, estimated glomerular filtration rate ≥60 mL/min/1.73 m², and receiving a renin angiotensin aldosterone system inhibitor

Intervention:

Spironolactone (maximum dose of 50 mg/d) or placebo for 24 weeks

Outcomes:

Change in brachial artery flow-mediated dilation [FMD_BA] was the primary endpoint and changes in carotid-femoral pulse-wave velocity [CFPWV]) was the secondary endpoint.

Results:

Sixty participants completed the trial. Participants had a mean age of 34±10 (s.d.) years, 54% were female, and 84% non-Hispanic White. Spironolactone did not change brachial artery flow-mediated dilation (8.0% ± 5.5% and 7.8% ± 4.3% at baseline and 24 weeks, respectively, versus corresponding values in placebo group of 8.4% ± 6.2% and 8.0% ± 4.6%; p=0.9 for comparison of change between groups) or carotid-femoral pulse-wave velocity (640 ± 127 cm/s and 603 ± 101 cm/s at baseline and 24 weeks, respectively, versus corresponding values in the placebo group of 659 ± 138 cm/s and 658 ± 131 cm/s; p=0.1). Brachial systolic blood pressure was reduced with spironolactone (median change of −6 [IQR, −15, 1] mm Hg, versus −2 [−7, 10] mmHg in the placebo group; p=0.04). Spironolactone did not change the majority of circulating and/or endothelial cell markers of oxidative stress/inflammation, nor did it change vascular oxidative stress.

Limitations:

Low level of baseline vascular dysfunction; lack of aldosterone measurements

Conclusions:

Twenty-four weeks of aldosterone antagonism reduced systolic blood pressure without changing vascular function in patients with early-stage ADPKD.

Autosomal dominant polycystic kidney disease (ADPKD) is the most common potentially lethal monogenic kidney disorder, occurring worldwide and affecting all races.¹ While ADPKD is characterized by the development and continued growth of multiple kidney cysts that result in end-stage renal disease in the majority of afflicted patients,² the leading causes of death among patients with ADPKD are cardiovascular complications and disorders.³ Hypertension occurs early in the natural history of ADPKD⁴ and is known to be partly responsible for the excess risk of cardiovascular disease in ADPKD.¹

Impaired vascular endothelial function, commonly assessed as reduced endothelium-dependent dilation, as well as increased large-elastic artery stiffness, occur early in the course of ADPKD, even in the presence of preserved glomerular filtration rate (GFR).^{5, 6} Notably, impaired endothelium-dependent dilation and increased large-elastic artery stiffness are both independent predictors of cardiovascular events and mortality.^{7, 8} Both oxidative stress and inflammation are also increased in ADPKD^{9, 10} and may be involved in the pathogenesis of vascular dysfunction and cardiovascular disease in this population. Thus, establishing strategies to delay, minimize and/or prevent arterial dysfunction, and thus mitigate cardiovascular risk, are a high priority in ADPKD.

In ADPKD, renin-angiotensin aldosterone system activation results from expansion of renal cysts and compression of the renal vasculature.^{11, 12} Angiotensin converting enzyme inhibitors (ACEi) are widely used for controlling hypertension in ADPKD, but aldosterone breakthrough may occur,¹³ which can promote vascular oxidative stress, vasoconstriction, and hypertension via vascular mineralocorticoid receptor activation.¹⁴ Aldosterone levels are high in patients with ADPKD and remain elevated after administration of ACEi when compared to patients with essential hypertension.¹² Studies have implicated aldosterone excess in the development of endothelial dysfunction, decreased arterial compliance, and cardiovascular disease in patients with primary hyperaldosteronism, hypertension, heart failure, and even in the general population.^15–18 Additionally, mineralocorticoid receptor antagonists have been shown to reduce vascular dysfunction in numerous chronic disease populations receiving ACEi, including heart failure,^{19, 20} resistant hypertension,^{21, 22} and chronic kidney disease.²³

Accordingly, the aim of this randomized, placebo-controlled trial was to determine the efficacy of an aldosterone antagonist for improving vascular endothelial function (primary endpoint), as well as reducing large-elastic artery stiffness (secondary endpoint) in patients with early-stage ADPKD receiving the maximal tolerable dose of an ACEi or angiotensin receptor blocker (ARB). Additionally, we sought to obtain insight into the mechanisms by which aldosterone antagonism may reduce vascular dysfunction in ADPKD. We hypothesized that aldosterone antagonism would improve brachial artery flow-mediated dilation (FMD_BA), a measure of endothelium-dependent dilation, and reduce carotid-femoral pulse-wave velocity (CFPWV), an index of large-elastic artery stiffness, and that this would be associated with reductions in circulating and endothelial cell markers of oxidative stress and inflammation, including changes in circulating bioactive lipid mediators.

METHODS

Additional details of the Methods are provided in Item S1.

Study Participants.

Eligible participants were enrolled at the University of Colorado Denver Anschutz Medical Campus between July 2014 and June 2016 (the trial concluded according to enrollment determined by power calculations described below). Participants eligible for inclusion were men and women 20–55 years of age with a diagnosis of ADPKD based on the Ravine criteria (in participants ≥30 years)²⁴ with a PKD1 genotype. Total kidney volume was between 500–2,500 ml, estimated glomerular filtration rate (eGFR) by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation²⁵ was ≥60 ml/min/1.73 m², and all participants had a history of hypertension (SBP > 130 mm Hg and/or diastolic BP > 80 mmHg based on 3 separate measurements within the past year) treated with a stable dose of an ACEi or ARB (standard recommendation at the time the trial was designed) and currently controlled (to <160/90 mmHg).

Participants were excluded if they had an average serum potassium >5.5 mEq/l or any single value >6.0 mEq/l in the past 6 months, received an aldosterone antagonist in the past 6 months, were using a potassium-sparing diuretic or other medication that could contribute to hyperkalemia (e.g. non-steroidal anti-inflammatory agents), body mass index ≥40 kg/m² (for accuracy of vascular measurements), were currently smoking or had a history of smoking in the past 12 months, had a history of severe congestive heart failure (ejection fraction <35%), were hospitalized in the last 3 months, had a history of liver disease, received immunosuppressive therapy within the last year, used warfarin with an INR > 2.5, had an active infection or antibiotic use, had alcohol dependence or abuse, or were pregnant, nursing, or planning to become pregnant. If study participants were using antioxidants and/or omega-3 fatty acids, they were discontinued at least 4 weeks prior to study participants, and if participants used cannabis use was discontinued 2 weeks prior to vascular measurements.

Study Design.

A 24-week (±2 weeks) prospective, randomized, placebo-controlled (1:1 allocation), parallel group, double-blind study with the mineralocorticoid antagonist spironolactone was conducted at the University of Colorado Anschutz Medical Campus Division of Renal Diseases and Hypertension Clinical Vascular Physiology Laboratory. Randomization (without blocking) occurred using a computer-generated procedure run and kept by a statistician. After initial screening, subjects meeting inclusion criteria had baseline vascular measurements performed, as described below. Measurements were made in the supine position following standard recommendations.²⁶ Participants were then randomized by the study coordinator according to a computer-generated random allocation sequence and received either the mineralocorticoid antagonist spironolactone or matching placebo. Participants received a dose of 25 mg/d for 4 weeks, with dosing escalated to 50 mg/d for the remainder of the study if tolerated by an individual participant.

All investigators, coordinators, analysts, and participants were blinded to group assignment, with only the nursing staff not affiliated with the study and the statistician aware of the randomization. Measurements were repeated after 24 weeks of the intervention. The primary outcome was the change in FMD_BA in the spironolactone compared to placebo group.

Procedures.

Vascular Measurements.

FMD_BA was determined using duplex ultrasonography (Xario 200, Toshiba, Tustin, CA) with ECG-gated end-diastolic ultrasound images analyzed by a single blinded analyst using a commercially available software package (Vascular Analysis Tools 5.8.1, Medical Imaging Applications, Coralville, IA), as described in detail previously.^27–29 Doppler flow of the brachial artery was also measured and peak shear rate was calculated as a potential covariate.^4–7 Following a minimum of 15 minutes of rest in the supine position, 30 seconds of B-mode ECG-gated end-diastolic ultrasound images were recorded to determine baseline brachial artery diameter. A forearm blood pressure cuff placed just distal to the olecranon process was inflated to 250 mmHg for 5 minutes. Thirty seconds prior to cuff release, brachial artery dimeter during occlusion was recorded for calculation of shear rate. Peak shear rate was calculated during the 15 seconds following cuff release, and peak diameter was determined as the largest six consecutive R-wave triggered frames during the two minutes following cuff release for calculation of percent FMD_BA. Endothelium-independent dilation (brachial artery dilation to 0.4 mg of sublingual nitroglycerin) was assessed as a standard index of smooth muscle cell sensitivity to exogenous nitric oxide.^{27, 29, 30}

PWV was measured as described in detail previously using a transcutaneous custom tonometer.^27–29 Supine blood pressure was also measured using the auscultatory method. Additionally, as secondary indices of arterial stiffness, carotid artery compliance, carotid artery β-stiffness index, carotid intimal medial thickness, carotid augmentation index, and carotid SBP were measured.^27–29

The influence of oxidative stress on FMD_BA was assessed by infusing a supraphysiological dose of ascorbic acid and measuring FMD_BA during the “drip infusion” when peak plasma concentrations occur, as described previously.^{27, 30, 31}

Cellular Markers of Oxidative Stress and Inflammation.

Vascular endothelial cells were obtained from the intima of an antecubital vein and were later stained for positive identification of endothelial cells, assessment of nuclear factor κB (NFκB), interleukin 6 (IL-6), NAD(P)H oxidase (p47^phox), and phosphorylated endothelial nitric oxide synthase (PeNOS) using immunofluorescence, as described previously.^{27, 29, 32}

Circulating Markers of Oxidative Stress, Inflammation, and Bioactive Lipid Mediators.

Serum high-sensitivity C-reactive protein (hsCRP) and IL-6 levels were measured by ELISA. Targeted liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis of markers of oxidative stress (prostaglandins [PG], including 8-isoprostane, as well as PGF_2α, PGD₂, and PGE2), and bioactive lipid mediators (hydroxyoctadecadienoic acids [9-HODE and 13-HODE], hydroxyeicosatetraenoic acids [5-HETE, 8-HETE, 9-HETE, 11-HETE, 12-HETE], epoxyeicosatrienoic acids [8,9-EET, 11,12-EET and 14–15-EET), and hydroxyeicosapentanoic acids [5-HEPE, 12-HEPE] was also performed on serum samples using validated assays, as described in detail previously.^{10, 33, 34}

Outcome Measures.

The primary endpoint was change in FMD_BA in the spironolactone compared to placebo group. Secondary outcomes were change in CFPWV, blood pressure, the change in FMD_BA following acute infusion of ascorbic acid, and circulating and vascular endothelial cell protein expression of markers of oxidative stress, inflammation, and associated pathways. Pre-specified exploratory outcomes included change in carotid artery compliance, carotid artery β-stiffness index, carotid SBP, and carotid intimal medial thickness.

Statistical Analysis.

All statistical analyses were performed by using SAS version 9.4 (SAS Institute, Cary, NC). Differences in baseline variables between groups were assessed using independent samples t-tests, Chi-square tests, or Fishers exact test. The comparison of changes in percent FMD_BA in response to treatment between groups were analyzed using an independent samples t-test. This same statistical approach was performed for other secondary and exploratory endpoints. Non-normally distributed variables were log-transformed prior to analysis. For variables that could not be log-transformed (change in SBP and DBP), a non-paramteric (Wilcoxon rank sums) test was used. In addition to comparison of change in percent FMD_BA with ascorbic acid (compared to isovolumetric saline) between groups (independent samples t-test), a paired t-test was used for within group comparisons of the change in percent FMD_BA with an acute ascorbic acid infusion (compared to saline infusion). Alpha was set at 0.05 (two-sided). All data are reported as means ± S.D (S.E. in figures), medians (interquartile range), or n (%).

A sample size of 24 subjects per group was calculated based on 97% power and a two-side type I error rate of 0.05 in order to detect a difference in mean change in FMD of 3.8 percent (baseline: 5.5%, end-of-study: 9.3%), assuming standard deviations of 2.1 and 3.4, respectively, based on previously published data (open-label) assessing the effect of an aldosterone antagonist of FMD_BA.²⁰ To account for potential dropout and experimental failure of about 20%, approximately 30 participants were randomized to each group.

Study Approval

All procedures were approved by the Institutional Review Board of the University of Colorado Anschutz Medical Campus (13–1440) and adhere to the Declaration of Helsinki. The nature, benefits and risks of the study were explained to the volunteers and their written informed consent was obtained prior to participation. The trial was registered at ClinicalTrials.gov (NCT01853553).

Results

Enrollment and Baseline Clinical Characteristics.

Of the 66 participants who were screened for participation, 61 were randomized to receive either the aldosterone antagonist spironolactone or placebo (Figure 1). One participant in the spironolactone group (group 1) discontinued the intervention prior to the final study visit at 24 weeks due to inability to tolerate the medication (frequent urination interfering with daily living). Participants had a mean age of 34±10 (s.d.) years; 54% were female and 84%, non-Hispanic White. Groups did not differ significantly in terms of baseline characteristics, including gender, race/ethnicity, body mass index, blood pressure, lipids, eGFR, and medications (Table 1).

Figure 1:

Flow diagram of patient enrollment, randomization, and completion.

Table 1.

Baseline Characteristics of Study Participants

Variable	Spironolactone (n=29)	Placebo (n=32)	P-Value
Age, y	34±10	34±9	0.9
Male sex	55%	38%	0.2
Non-Hispanic White	79%	88%	0.4
BMI, kg/m2	27.2±4.7	27.1±5.3	0.9
SBP, mmHg	122±13	120±12	0.5
DBP, mmHg	79±9	76±11	0.2
eGFR, ml/min/1.73m2	96±23	93±19	0.6
LDL, mg/dL	95±30	95±26	0.9
HDL, mg/dL	47±11	51±14	0.2
Total Cholesterol, mg/dL	162±36	167±29	0.5
ACEi/ARB, %	100%	100%	0.9
Diuretic, %	17%	19%	0.9
CCB, %	3%	6%	0.6
Statin, %	31%	16%	0.2

Continuous data expressed as mean±S.D. BMI, body-mass index; SBP, systolic blood pressure; SBP, systolic blood pressure, eGFR; estimated glomerular filtration rate; LDL, low density lipoprotein, HDL, high density lipoprotein; ACEi, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; CCB, calcium channel blocker. P-values are a comparison of the spironolactone and placebo groups using an independent samples t-test, chi-square test, or Fisher’s exact test.

Effect of Aldosterone Antagonism on Vascular Function.

The primary endpoint, change in FMD_BA (expressed as percent change), did not differ after 24 weeks in the spironolactone group as compared to the placebo group (Figure 2A and 2B; Table 2). Results were similar when FMD was expressed as an absolute change (Table 2). Baseline diameter and shear rate did not differ across the study, thus modeling was not performed to include these variables as covariates (Table 2). Endothelium-independent dilation to sublingual nitroglycerin, a measure of smooth muscle cell responsiveness to nitric oxide, was also unaffected by aldosterone antagonism (Table 2).

Figure 2: Changes in brachial artery flow-mediated dilation and aortic pulse-wave velocity with spironolactone and placebo.

Brachial artery flow-mediated dilation (FMD) mean±s.e group values (Panel A) and individual data points (Panel B) at baseline (black bars; closed circles) and following 24 weeks of treatment (white bars; open circles) with spironolactone or placebo. Carotid-femoral pulse-wave velocity (CFPWV mean±s.e group values (Panel C) and individual data points (Panel D) at baseline (black bars; closed circles) and following 24 weeks of treatment (white bars; open circles) with spironolactone or placebo.

Table 2.

Vascular Parameters

	Spironolactone n=28			Placebo n=32
	BL	24 wk	Δ	BL	24 wk	Δ	P-Value
Brachial artery FMD
percent change	8.0±5.5	7.8±4.3	−0.2±3.2	8.4±6.2	8.0±4.6	−0.4±4.7	0.9
absolute change; mm	0.26±0.16	0.26±0.12	−0.00±0.11	0.26±0.14	0.25±0.13	−0.01±0.13	0.8
Baseline brachial artery diameter (mm)	3.6±0.7	3.6±0.6	−0.1±0.2	3.4±0.6	3.3±0.6	−0.0±0.2	0.5
Peak shear rate (s−1)	1009±424	1100±336	91±329	1121±489	1174±426	54±410	0.7
Brachial artery dilation to nitroglycerin
percent change *	27.9±9.1	28.6±6.8	0.7±4.4	28.0±6.1	28.9±7.3	0.9±7.4	0.9
absolute change; mm *	0.94±0.22	0.97±0.19	0.03±0.12	0.95±0.21	0.94±0.17	−0.01±0.19	0.5
Brachial SBP (mmHg)	124±13	117±11	−6 (−15, 1)	124±10	125±13	−2 (−7, 10)	0.04
Brachial DBP (mmHg)	74±10	70±9	−4 (−10, 3)	75±9	75±10	−1 (−7, 9)	0.2
Carotid-femoral PWV (cm/s)	640±125	603±101	−37±78	659±138	658±131	−1±89	0.1
Carotid-radial PWV (cm/s)	924±126	903±147	−21±106	920±185	937±187	18±188	0.3
Carotid intimal medial thickness (mm)	0.48±0.09	0.48±0.09	−0.01±0.05	0.47±0.07	0.47±0.06	−0.00±0.07	0.7
Carotid augmentation index (%)	9.3±15.0	2.5±16.5	−6.8±14.6	8.3±15.4	13.4±14.1	5.2±13.3	0.002
Carotid artery compliance (mm/mm Hg * 10−1)	0.12±0.04	0.12±0.04	0.01±0.02	0.12±0.04	0.12±0.03	0.00±0.03	0.5
Carotid β-stiffness index (A.U.)	6.1±2.0	5.8±1.7	−0.3±1.3	6.0±2.1	5.8±2.1	−0.2±1.9	0.7
Carotid SBP (mmHg)	118±15	110±12	−8±15	120±15	120±16	0±17	0.06

Data are mean±S.D or median [IQR]. P-value is a comparison of the change from baseline to 24 weeks (Δ) between the 2 groups using a t-test for normally distributed variables and a non-parametric (Wilcoxon rank sums) test for non-normally distributed variables. Blood pressures were taken in the supine position. BL, baseline; FMD, flow-mediated dilation; NTG, nitroglycerin; SBP, systolic blood pressure; DBP, diastolic blood pressure; PWV, pulse-wave velocity.

^*N=16 for spironolactone and n=15 for placebo for NTG variables.

The secondary endpoint, CFPWV, also did not change after 24 weeks in the spironolactone group as compared to the placebo group Figure 2C and 2D; Table 2). Carotid-radial PWV, a measure of peripheral arterial stiffness, did not change in either group (Table 2). Change in resting brachial SBP was −6 [IQR, −15, 1] mmHg in the spironolactone group, compared with a change of 12 [IQR, −7, 10] mmHg in the placebo group (Table 2; p=0.04). Carotid augmentation index, a measure of wave reflection, was reduced with spironolactone (Table 2; p=0.002), and carotid SBP, an index of central blood pressure, tended to be reduced with spironolactone (Table 2; p=0.06). Additional secondary endpoints (carotid artery intimal medial thickness, carotid artery compliance, and carotid β-stiffness index) were unchanged (Table 2).

Vascular Oxidative Stress, Inflammation, and Bioactive Lipid Mediators.

At baseline, infusion of ascorbic acid significantly raised plasma ascorbic acid levels (pre: 34.0±5.1 μmol/L; post: 1093.8±330.1 μmol/L; p=0.002), as reported previously.²⁹ This infusion significantly improved FMD_BA as compared to isovolumetric saline, as also reported previously.²⁹ At 24 weeks, ascorbic acid infusion also improved FMD_BA in both groups (saline vs ascorbic acid: 7.7±4.4% vs 9.9±5.4% [p=0.001] for spironolactone; 7.1±4.1% vs 9.0±5.7% [p=0.008] for placebo); however, the comparison of change in percent FMD_BA with ascorbic acid between groups revealed no reduction in vascular oxidative stress with aldosterone antagonism (p=0.8).

Serum hsCRP levels were increased at 24 weeks in the spironolactone compared to placebo group (Table 3; p=0.04) without a change in IL-6 levels (Table 3; p =0.2). In contrast to circulating inflammatory markers, vascular endothelial cell protein expression of the pro-inflammatory transcription factor NFκB and IL-6 were unchanged with spironolactone compared to placebo (Figure 3A and 3B; p=0.2 and 0.2, respectively). Protein expression of the oxidant enzyme NADPH oxidase (Figure 3C; p=0.3) and PeNOS (Figure 3D; p=0.3) were also unchanged with spironolactone compared to placebo group. Circulating levels of the oxidative stress marker 8-isoprostane were nominally decreased in the spironolactone compared to placebo group, but this finding was not statistically significant (Table 3, p=0.05). Other bioactive lipid mediators assessed did not differ in change between groups (Table 3). We have previously reported a comparison of these biomarkers at baseline to levels in normal healthy controls.²⁹

Table 3.

Circulating Markers

	Spironolactone n=28			Placebo n=31
	BL	24 wk	Δ	BL	24 wk	Δ	P-Value
hsCRP, mg/L	0.97 (0.32, 1.85)	2.39 (0.62, 6.12)	0.28 (−0.21, 2.18)	1.19 (0.52, 4.36)	1.74 (0.62, 4.89)	0.19 (−0.29, 1.1)	0.04
IL-6, pg/mL	0.56 (0.39, 0.80)	0.76 (0.52, 1.22)	0.14 (0.00, 0.44)	0.56 (0.32, 1.24)	0.60 (0.40, 1.01)	0.02 (−6.1, 3.9)	0.2
8-isoprostane, pg/mL	5.0 (3.8, 6.6)	4.7 (4.1, 6.5)	−0.50 (−1.6, 1.3)	4.7 (3.8, 5.8)	5.2 (4.4, 6.7)	0.90 (−0.80, 1.5)	0.05
PGF2a. pg/mL	31.7 (11.6, 80.7)	38.0 (13.7, 74.0)	−0.35 (−32.1, 13.6)	21.3 (13.7, 53.8)	34.0 (27.2, 51.2)	3.8 (−6.8, 19.6)	0.5
PGD2, pg/mL	85.7 (46.5, 163.7)	94.5 (45.8, 141.7)	−7.3 (−51.5, 33.0)	77.9 (55.3, 120.3)	88.0 (59.4, 115.7)	−3.5 (−45.3, 50.1)	0.9
PGE2, pg/mL	195.8 (49.0, 475.4)	203.3 (110.5, 335.1)	7.1 (−175.7, 73.7)	133.0 (63.0, 300.8)	158.9 (103.0, 229.0)	7.2 (−134.4, 118.0)	0.9
9-HODE, ng/mL	13.3 (8.4, 21.0)	10.2 (5.1, 14.6)	−5.1 (−10.0, 0.20)	11.4 (7.1, 20.2)	10.2 (6.9, 17.1)	−1.6 (−6.9, 3.1)	0.3
13-HODE, ng/mL	10.9 (7.6, 19.7)	10.0 (8.1, 13.6)	−1.3 (−9.3, 3.7)	9.6 (6.3, 16.8)	9.9 (7.6, 14.8)	−0.2 (−6.1, 3.9)	0.6
5-HETE, ng/mL	2.1 (0.76, 2.9)	2.2 (1.6, 2.6)	0.33 (−0.82, 1.4)	2.0 (1.1, 3.0)	2.1 (1.1, 3.4)	0.15 (−0.29, 1.3)	0.7
8-HETE, ng/mL	0.79 (0.45, 2.0)	0.87 (0.65, 1.8)	0.00 (−0.32, 0.41)	0.58 (0.39, 0.87)	0.84 (0.60, 1.2)	0.21 (−0.21, 0.50)	0.7
9-HETE, ng/mL	4.1 (1.8, 6.5)	2.9 (1.5, 5.2)	−0.67 (−3.5, 0.57)	1.4 (1.0, 2.6)	1.8 (1.3, 3.1)	0.21 (−0.21, 0.50)	0.2
11-HETE, ng/mL	1.4 (0.58, 3.3)	1.5 (0.80, 2.6)	0.10 (−1.3, 0.67)	1.1 (0.62, 2.2)	1.3 (1.1, 2.0)	0.41 (−0.96, 0.91)	0.9
12-HETE, pg/mL	63.8 (24.9, 138.8)	72.9 (32.3, 139.1)	0.65 (−48.7, 38.7)	39.9 (18.3, 64.3)	49.5 (27.5, 75.5)	13.3 (−26.4, 33.2)	0.5
8,9-EET, ng/mL	4.8 (2.5, 13.6)	5.6 (3.5, 12.6)	−0.40 (−4.4, 3.2)	3.0 (1.8, 5.7)	4.5 (2.7, 6.8)	0.30 (−1.5, 2.1)	0.9
11,12-EET, ng/mL	1.4 (0.45, 4.4)	1.5 (0.79, 2.3)	0.13 (−2.0, 0.71)	1.2 (0.48, 2.7)	1.5 (0.98, 2.1)	0.16 (−1.2, 0.95)	0.7
14,15-EET, ng/mL	1.3 (0.69, 2.2)	1.0 (0.82, 1.6)	−0.08, (−1.1, 0.49)	0.94 (0.62, 1.3)	1.1 (0.87, 1.3)	0.13 (−0.37, 0.53)	0.2
5-HEPE, ng/mL	0.84 (0.56, 1.1)	0.68 (0.49, 1.2)	−0.22 (−0.57, 0.38)	0.63 (0.41, 0.78)	0.71 (0.60, 0.91)	0.24 (−0.04, 0.46)	0.08
12-HEPE, pg/mL	38.3 (13.4, 80.4)	51.0 (21.3, 84.6)	−11.60 (−40.2, 34.2)	19.5 (10.8, 52.9)	30.2 (19.9, 48.4)	4.8 (−16.0, 28.4)	0.5

Data are median [interquartile range]. P-value is a comparison of the change from baseline to 24 weeks (Δ) between the 2 groups using log-transformed values. Non-normally distributed variables were log-transformed prior to analysis.

hsCRP, high-sensitivity C-reactive protein; IL-6, interleukin 6; PGF_2α, prostaglandin F_2α; PGD₂, prostaglandin D₂; PGE2, prostaglandin E2; HODE, hydroxyoctadecadienoic acid; HETE, hydroxyeicosatetraenoic acid; EET epoxyeicosatrienoic acid; HEPE, hydroxyeicosapentanoic acid.

Figure 3. Change in vascular endothelial cell protein expression with spironolactone and placebo.

Vascular endothelial cell protein expression of nuclear factor κ B (NFκB; n=20 spironolactone; n=14 placebo; Panel A), interleukin 6 (IL-6; n=20 spironolactone; n=13 placebo; Panel B), NADPH oxidase (n=20 spironolactone; n=13 placebo; Panel C), and phosphorylated endothelial nitric oxide synthase (PeNOS; n=12 spironolactone; n=10 placebo; Panel D) at baseline (black bars) and following 24 weeks of treatment (white bars) with spironolactone or placebo. Expression is relative to human umbilical vein endothelial cell (HUVEC) control. Representative images shown below. Date are mean±s.e.

Adherence and Adverse Events.

Total compliance, assessed by pill count over the year period, was 96% (95% in the spironolactone group and 96% in the placebo group). Overall, spironolactone was well tolerated. 14 participants experienced an adverse event related or possibly related to the drug (n=7 in the spironolactone group, n=7 in the placebo group, Table 4). There were no serious adverse events in either group. One participant in the spironolactone group dropped out of the study due to inability to tolerate frequent urination, despite dose reduction to 12.5 mg/d. No other participants discontinued the study. The vast majority of participants were able to tolerate the 50 mg/d dose, although two participants were reduced to 25 mg/d (one spironolactone/one placebo) and one participant was reduced to 12.5 mg/d (placebo) due to side-effects. There was no difference in the change in eGFR at 24 weeks between groups (−1.3±11 vs +0.7±17 ml/min/1.73 m² for spironolactone and placebo, respectively; p=0.6).

Table 4.

Adverse Events

	Spironolactone (n=29)	Placebo (n=32)
Dizziness/lightheadedness	3 (10%)	3 (9%)
Muscle cramping/soreness	2 (7%)	3 (9%)
Vision changes	2 (7%)	1 (3%)
Increased urination	2 (7%)	0 (0%)
Hyperkalemia	0 (0%)	1 (3%)
Fatigue	1 (3%)	1 (3%)
Increased thirst	0 (0%)	1 (3%)
Nausea	0 (0%)	1 (3%)
Mildly elevated AST/ALT	0 (0%)	1 (3%)

Spironolactone groups includes n=1 participant who discontinued study participation due to inability to tolerate frequent urination.

ALT, alanine transaminase; aspartate aminotransferase, AST.

Discussion

In a randomized, controlled trial evaluating the efficacy of aldosterone antagonism for reducing vascular dysfunction in individuals with ADPKD, we found no change in FMD_BA or CFPWV following 24 weeks of treatment with spironolactone. This was despite a significant reduction in SBP with spironolactone treatment, as would be clinically expected. Aldosterone antagonism also failed to change the majority of circulating, vascular, and endothelial cell markers of oxidative stress and inflammation, including bioactive lipid mediators.

Vascular dysfunction is present very early in the course of ADPKD^{5, 6, 28} and is an important and independent predictor of cardiovascular events and mortality.^{7, 8} This clinical trial was novel as it evaluated vascular dysfunction (FMD_BA and CFPWV) as endpoints in response to an intervention in ADPKD. We found a modest 6% reduction in CFPWV with spironolactone treatment, although this change failed to meet statistical significance in comparison to change in the placebo group. Additionally, SBP was lowered by 7 mmHg with mineralocorticoid antagonism, which likely contributed to the nominally reduced CFPWV, given the dynamic interconnection between blood pressure and arterial stiffness.³⁵ However, there was no improvement in endothelium-dependent dilation, as measured by FMD_BA (primary outcome).

Mineralocorticoid receptors are expressed extrarenally in the vasculature (endothelial and vascular smooth muscle cells), thus aldosterone has pleiotropic effects influencing vascular endothelial function and blood pressure beyond regulation of sodium and water balance.¹⁴ Aldosterone excess is associated with reduced large-elastic artery compliance^{15, 16} and impaired endothelium-dependent dilation.^{17, 18} Additionally, high levels of aldosterone and mineralocorticoid receptor activation stimulate a pro-inflammatory response via generation of reactive oxygen species.³⁶

Trials of mineralocorticoid receptor antagonism have been efficacious for reducing vascular dysfunction in other chronic diseases with or without uncontrolled hypertension, typically also being treated with an ACEi. Aldosterone antagonism improves both resistance vessel¹⁹ and conduit artery endothelium-dependent dilation²⁰ in individuals with heart failure. Additionally, treatment reduces mortality in heart failure patients.³⁷ Mineralocorticoid receptor antagonism also improves FMD_BA²¹ and reduced CFPWV²² in resistant hypertension. In these populations that have demonstrated improved FMD_BA with aldosterone antagonism, baseline function was notably lower than in the present study, despite treatment with renin-angiotensin aldosterone system blockers. In individuals with hypertension, aldosterone antagonism attenuates a decline in FMD_BA,³⁸ reduces CFPWV,³⁹ reduces arterial wall stiffness and media collagen to elastin ratio,⁴⁰ and reduces left-ventricular mass index,⁴¹ without changing resistance vessel endothelium-dependent dilation.⁴⁰ This is likely mediated by a large reduction in SBP.^38–41 However, aldosterone antagonism does not improve endothelium-dependent dilation in type 2 diabetes mellitus⁴² or coronary artery disease without heart failure.⁴³ Notably, in mild-to-moderate chronic kidney disease, spironolactone²³ but not eplerenone⁴⁴ reduces CFPWV, in addition to promoting a large reduction in SBP and improved left ventricular mass index and aortic distensibility.²³

Both oxidative stress and inflammation are increased even early in the course of ADPKD,^{9, 10} which may be mediated in part by mineralocorticoid receptor activation.³⁶ In various rodent models, aldosterone antagonism reduces aortic messenger RNA expression of NADPH oxidase p22phox,⁴⁵ NADPH oxidase activity,⁴⁶ plasma thiobarbituric acid-reactive substances,⁴⁶ macrophage oxidation of LDL,⁴⁷ superoxide production,⁴⁷ and hepatic malonyl dialdehyde levels,⁴⁵ in addition to increasing the glutathione-oxidized glutathione ratio.⁴⁵ Less evidence is available in humans; however, 3 months of treatment with spironolactone reduces urinary 8-isoprostane levels in patients with diabetic nephropathy.⁴⁸ We also found a suggestion of reduced circulating 8-isoprostane levels with spironolactone in early-stage ADPKD. In contrast, we found no reduction in endothelial cell markers of oxidative stress in response to aldosterone antagonism, as well as no reduction in vascular oxidative stress, as evidenced by continued improvement in FMD_BA following an acute infusion of ascorbic acid that produces plasma concentrations that have been shown to inhibit superoxide production in vitro.⁴⁹

Aldosterone antagonism also reduces circulating inflammatory markers in patients with hypertension⁴⁰ and diabetic nephropathy,⁴⁸ but not following 60 days of treatment in individuals with type 2 diabetes mellitus and moderately or severely increased albuminuria.⁵⁰ While we found a surprising increase in hsCRP levels with spironolactone treatment, perhaps due to the variability of the test, there was no change in inflammation local to the vascular endothelium, as evidenced by a lack of change in endothelial cell protein expression of NFκB and IL-6. Aldosterone antagonism also increases eNOS protein and mRNA expression in rodents,⁴⁵ and the improvement in endothelium-dependent dilation with spironolactone in heart failure patients treated with an ACEi is mediated by increased NO bioavailability.¹⁹ However, we found no change in endothelial cell protein expression of PeNOS with spironolactone treatment compared to the placebo group.

Notably, there was no difference in adverse events between the active and placebo groups. The major potential risk of spironolactone is hyperkalemia, particularly if receiving an ACEi/ARB. However, this was not a side-effect noted in the present trial. Overall, the study treatment was well-tolerated and drop out from the trial was extremely low.

The major strength of this study is that it was a randomized controlled trial to evaluate vascular endpoints in response to an intervention in individuals with ADPKD, which are both novel and clinically significant endpoints in this population. Additionally, we provided mechanistic insight through the collection of vascular endothelial cells from participants, acute infusion of ascorbic acid to measure vascular oxidative stress, and state-of-the art LC-MS/MS analyses to assess circulating markers. Participant retention in the study was excellent, with only one drop-out in a trial 24 weeks in duration. A potential limitation to the study is a relatively small sample size; however, we were appropriately powered to detect an improvement in the primary endpoint given expected vascular endothelial function at the time the trial was designed. An additional limitation is that it is difficult to separate the effects of blood pressure lowering from potential reductions in CFPWV, as regulation of blood pressure and arterial stiffness are closely related. Furthermore, measurements of plasma or urinary aldosterone were not available; however, accurate aldosterone measurements are challenging due to sensitivity to external factors including sodium intake, medications, position, and time of day, as well as assay limitations with low concentrations. These results may not apply to ADPKD patients who have more advanced chronic kidney disease, who may exhibit greater baseline vascular dysfunction, but would also be at higher risk of hyperkalemia with spironolactone. The overall level of vascular dysfunction in this cohort was lower than previously reported in early ADPKD,²⁹ which may have limited the ability to detect an improvement in vascular function with aldosterone antagonism and altered the assumptions underlying the power calculations for the trial. Finally, there is a potential for inflated type-1 error for all significant differences reported, except for the primary endpoint, given multiple comparisons.

In conclusion, in individuals with early-stage ADPKD, mineralocorticoid antagonism does not improve FMD_BA or reduce CFPWV, despite reducing blood pressure. This is consistent with the general lack of evidence for a reduction in oxidative stress and inflammation and associated pathways. Alternative interventions targeting an improvement in vascular endothelial function and reduced arterial stiffness in early ADPKD should be evaluated.

Supplementary Material

1

Supplementary File 1 (PDF) Item S1. Additional details of methods.