Renal sympathetic denervation for treatment of patients with heart failure: summary of the available evidence

Abstract

Heart failure syndrome results from compensatory mechanisms that operate to restore – back to normal – the systemic perfusion pressure. Sympathetic overactivity plays a pivotal role in heart failure; norepinephrine contributes to maintenance of the systemic blood pressure and increasing preload. Cardiac norepinephrine spillover increases in patients with heart failure; norepinephrine exerts direct toxicity on cardiac myocytes resulting in a decrease of synthetic activity and/or viability. Importantly, cardiac norepinephrine spillover is a powerful predictor of mortality in patients with moderate to severe HF. This provided the rationale for trials that demonstrated survival benefit associated with the use of beta adrenergic blockers in heart failure with reduced ejection fraction. Nevertheless, the MOXCON trial demonstrated that rapid uptitration of moxonidine (inhibitor of central sympathetic outflow) in patients with heart failure was associated with excess mortality and morbidity, despite reduction of plasma norepinephrine. Interestingly, renal norepinephrine spillover was the only independent predictor of adverse outcome in patients with heart failure, in multivariable analysis. Recently, renal sympathetic denervation has emerged as a novel approach for control of blood pressure in patients with treatment-resistant hypertension. This article summarizes the available evidence for the effect of renal sympathetic denervation in the setting of heart failure.

Key messages

• Experimental studies supported a beneficial effect of renal sympathetic denervation in heart failure with reduced ejection fraction.

• Clinical studies demonstrated improvement of symptoms, and left ventricular function.

• In heart failure and preserved ejection fraction, renal sympathetic denervation is associated with improvement of surrogate endpoints.

Keywords:

• Renal sympathetic denervation
• heart failure
• left ventricular function

Introduction

The syndrome of heart failure (HF) results from a constellation of compensatory mechanisms that operate essentially in response to reduced cardiac output in order to restore the systemic perfusion pressure. Sympathetic overactivity plays a pivotal role in HF. Norepinephrine induces systemic vasoconstriction and augments venous tone, contributing to maintenance of the systemic blood pressure and increasing preload. Additionally, it reduces renal blood flow, stimulates renin release, and enhances sodium reabsorption from the renal tubules. Indeed, cardiac norepinephrine spillover increases in patients with congestive HF; this is associated with increased activation of central noradrenergic neurons (1,2). Yet, norepinephrine exerts direct toxic effect on cardiac myocytes through cyclic AMP-mediated calcium overload, with a resultant decrease in synthetic activity and/or viability (3). Moreover, increased cardiac norepinephrine release is associated with left ventricular (LV) hypertrophy (4). Importantly, cardiac norepinephrine spillover was the most powerful predictor of mortality in patients with moderate to severe HF (5). This provided the rationale for the trials that demonstrated survival benefit of beta adrenergic blockers in patients with HF and reduced ejection fraction (EF) (6,7). Nevertheless, the prematurely terminated MOXCON trial demonstrated that rapid uptitration of moxonidine (an inhibitor of central sympathetic outflow) in patients with HF was associated with excess early mortality and morbidity, despite reduction of plasma norepinephrine (8). Interestingly, renal norepinephrine spillover was the only independent predictor of long-term adverse outcome in patients with HF (9). Recently, renal sympathetic denervation (RDN) has emerged as a novel approach for control of blood pressure (BP) in patients with treatment-resistant hypertension. This review summarizes the available evidence for the effect of RDN in the setting of HF.

The renal-sympathetic axis

The sympathetic innervation of the kidneys arises from the second sympathetic ganglion; sympathetic fibers travel through the adventitia of the renal arteries to reach the kidneys (10). The renal efferent sympathetic fibers terminate in the glomerular arterioles, renal proximal tubules, and juxtaglomerular apparatus (11). Stimulation of the glomerular arteriolar α-1A adrenergic receptors mediates vasoconstriction and reduced renal blood flow; stimulation of α-1B receptors in the proximal tubules mediates sodium retention; whereas stimulation of β1 adrenergic receptors in the juxtaglomerular apparatus mediates renin release and triggers activation of the renin-angiotensin-aldosterone axis (10,12). On the other hand, stimulation of mechanosensitive receptors in the renal pelvic wall and chemoreceptors in the renal interstitium triggers afferent sympathetic nerve signaling (13). Afferent signaling is activated by stimuli such as renal ischemia, hypoxia, and intrinsic renal disease (11). Acting through the posterior hypothalamus, renal afferent sympathetic signaling regulates central sympathetic outflow to increase BP, and control reflex sympathetic efferent activity (14). Yet, evidence suggests a crucial role of the renal afferent sympathetic signaling; especially from a diseased kidney, in the activation and perpetuation of central sympathetic outflow in various chronic disease states. In a rat model of chronic renal failure, the elevated hypothalamic norepinephrine and BP were both reduced by rhizotomy (15). Moreover, renal denervation normalized the elevated BP in a rat model of polycystic kidney (16). In clinical studies, the rate of muscle sympathetic nerve discharge in patients with end-stage renal disease who undergo hemodialysis (no nephrectomy) was significantly higher than healthy controls; such discharge was comparable to controls in hemodialysis patients who underwent bilateral nephrectomy (17,18). This excess sympathetic nerve activity in hemodialysis patients was not decreased by renal transplantation (retained native diseased kidneys) despite excellent graft function, suggesting that the sympathetic overactivity in such patients is probably mediated by signals from the diseased kidneys, rather than associated with uremia-related toxins (18). Moreover, in patients with resistant hypertension, RDN was associated with substantial reduction of muscle sympathetic nerve activity at 3-month follow-up, along with reduction of systolic and diastolic BP (19); such reduction was maintained at 1-year follow-up (20). Similarly, in patients with resistant hypertension, RDN was associated with reduction of cardiac sympathetic activity at 9 months, although cardiac sympathetic innervation remained unchanged (21). Interestingly, RDN in patients with resistant hypertension did not cause orthostatic dysfunction at 3-month follow-up (22). Cardiac sympathetic activity contributes to diastolic function impairment. In rat model, isoproterenol administration was associated with diastolic dysfunction, along with endocardial injury, cardiomyocyte hypertrophy, and interstitial fibrosis (23). In another rat model of isoproterenol-induced cardiomyocyte hypertrophy, passive cardiomyocyte stiffness was increased due to residual actin-myosin cross-bridge formation (24). Finally, in patients with hypertension, increased muscle sympathetic nerve traffic (a surrogate of sympathetic activity) was associated with impairment of diastolic function indices (25).

Renal denervation in heart failure with reduced ejection fraction

Sympathetic overactivity is a hallmark in patients with clinical systolic HF, as well as in subjects with asymptomatic LV systolic dysfunction (26). Plasma norepinephrine increases in patients with HF and reduced EF; total body (62%), as well as cardiac (227%) norepinephrine spillover increases in such patients, versus control subjects (2). Notably, cardiac norepinephrine spillover was the most potent prognostic marker of poor outcome in patients with moderate to severe congestive HF, in a multivariable model (5). Robust evidence from experimental – and less so from clinical – studies supports the safety, and a probable beneficial effect of RDN in preventing, attenuating, or reversing the deleterious effects of systolic HF, at mid-term follow-up.

Experimental studies

Extensive evidence from experimental studies supported a beneficial effect of RDN in HF with reduced EF (Table 1). In experimental studies of lower animal models of HF induced by myocardial infarction (MI), RDN was associated with better natriuresis, better cardiac remodeling and function, better hemodynamics, better autonomic balance, less neurohormonal activation, and less fibrosis (27–36). In studies of higher animal models of HF induced by rapid pacing, RDN was associated with better cardiac remodeling and function, less neurohormonal activation, better electrical stability, and less fibrosis (37–40,42). Furthermore, in a rat model of HF, increased expression of Na-K-2Ca cotransporter in the thick ascending loop of Henle was abolished by RDN (43). Yet, whether these beneficial effects of RDN are mediated by afferent, in contrast to efferent, renal sympathetic nerve ablation is unclear. Selective RDN of the renal afferent sympathetic fibers by capsaicin applied to the renal medulla might offer a potential for addressing this issue (44). Another point of interest is whether reinnervation occurs after RDN procedure. In a sheep model of RDN, there was almost complete functional and anatomical reinnervation by 5.5 months post-procedural; interestingly, by 11 months there was no difference in the response to electrical stimulation, renal distribution of afferent and efferent nerves, and renal norepinephrine levels, compared with nondenervated controls (45).

Table 1. Experimental studies of renal denervation in heart failure with reduced systolic function.

Study	References	Animal model	Mode of induction of HF	Mode of induction of RDN	Timing of RDN in relation to induced HF	Comparison	Results
Nozawa et al.	(27)	Rat model	Induced MI	Surgical	2 days before MI	HF + prior RDN versus HF without RDN	RDN was associated with better natriuresis, better LV systolic function and lower LV filling pressure, versus control.
Hu et al.	(28)	Rat model	Induced MI	Surgical	3 days before MI	HF + prior RDN with no MT versus 3 arms of HF + MT (BB, ACEI, ARB)	RDN was associated with enhanced natriuresis, better cardiac remodeling and function, and improved autonomic modulation, versus the study arms of MT (BB, ACEI, ARB).
Hu et al.	(29)	Rat model	Induced MI	Surgical	3 days before, 1 day after, and 7 days after MI	HF + RDN versus HF without RDN, comparison of RDN at different times in the course of MI	RDN was associated with improved LV systolic function, hemodynamic indices, lower BNP, and higher urine output and sodium excretion assessed at 4 weeks, versus control. There was no significant difference between RDN procedures performed at different times in the course of illness.
Hu et al.	(30)	Rat model	Induced MI	Surgical	1 day after and 4 weeks after MI	HF + RDN versus HF without RDN, comparison of RDN at different times in the course of MI	RDN was associated with improved LV function and hemodynamics, decreased neuro-hormonal activation (decreased plasma and renal tissue norepinephrine, and decreased plasma renin, angiotensin II, aldosterone, and endothelin), improved cardiac autonomic activities (increased myocardial β-1 receptors and improved heart rate variability), attenuated LV remodeling (decreased LV end-diastolic volume), and increased sodium and water excretion, versus control. The beneficial effect was more when RDN was performed early than late.
Zheng et al.	(31)	Rat model	Induced MI	Surgical	4 weeks after MI	HF + RDN versus HF without RDN	RDN was associated with decreased neuro-hormonal activation (decreased urinary excretion of norepinephrine, and lower BNP levels), reversed down-regulation of myocardial β-1 and β-2 receptors, improved response to isoproterenol, along with blunting of the increased collagen 1 expression in LV, versus control.
Liu et al.	(32)	Rat model	Sustained pressure overload (transverse aortic constriction)	Surgical	5 weeks later	HF + RDN versus HF + sham procedure	RDN was associated with decreased collagen volume fraction in cardiac and renal tissue, reduction of tumor necrosis factor α, aldosterone, and BNP, reduction of the biomarkers for fibrosis, downregulation of TGF-β1 and angiotensin II type I receptor expression in the LV, versus sham procedure.
Li et al.	(33)	Rat model	Sustained pressure overload (transverse aortic constriction	Surgical	6 weeks after MI	HF + RDN versus HF + sham procedure	RDN was associated with attenuated myocardial fibrosis, reduced left atrial enlargement and LV hypertrophy, upregulated myocardial β adrenoceptors and sarco-endoplasmic reticulum Ca ATPase and downregulated myocardial angiotensin II type 1 receptors, and decreased plasma BNP, norepinephrine, angiotensin II, and arginine vasopressin, versus sham procedure.
Liu et al.	(34)	Rat model	isoproterenol-induced cardiomyopathy	Surgical	5 weeks after MI	HF + RDN versus HF + sham procedure	RDN was associated with of fibrosis in cardiac and renal tissue, lower concentration of profibrosis factors (TGF-β1, MMP 2 and Collagen I), inflammatory cytokines (CRP and TNF-α), and collagen synthesis biomarkers, decreased release of norepinephrine, and downregulation of renin-angiotensin-aldosterone axis, versus sham procedure.
Clayton et al.	(35)	Rabbit model	Pacing-induced HF	Surgical	2 weeks before pacing	HF versus non-HF HF + prior RDN versus HF without prior RDN	RDN prevented the development of decreased renal blood flow and increased renal vascular resistance, and the alterations in the expression of angiotensin II receptors, versus control.
Schiller et al.	(36)	Rabbit model	Pacing-induced HF	Surgical (unilateral RDN)	4 weeks before pacing started	Pacing-induced HF versus sham procedure HF + prior RDN versus HF + prior sham procedure	Animals which underwent RDN showed better autonomic balance (preserved baroreflex sensitivity and heart rate variability) due to reduction of sympathetic tone, and lower plasma norepinephrine, compared with those which underwent sham procedure.
Zhao et al.	(37)	Canine model	Pacing-induced HF	Catheter-based RDN	3 days before pacing started	HF + prior RDN versus HF without RDN	RDN was associated with lower circulating angiotensin II, aldosterone, BNP, endothelin-1, and renalase, versus control.
Dai et al.	(38)	Canine model	Pacing-induced HF	Catheter-based RDN	8 weeks before pacing started	HF + prior RDN versus HF without RDN	RDN was associated with higher LVEF, lower LV end-diastolic pressure, less fibrosis, lower expression of BNP, angiotensin II, aldosterone and TGF-β in ventricular tissue, versus control.
Guo et al.	(39)	Canine model	Pacing-induced HF	Catheter-based RDN	3 days before pacing started	HF + prior RDN versus HF without RDN	RDN was associated with reduced ventricular fibrillation episodes, and attenuated substrate and electrical remodeling, versus control.
Zhao et al.	(40)	Canine model	Pacing-induced HF	Catheter-based RDN	8 weeks before pacing started	HF + prior RDN versus HF + prior sham procedure	RDN was associated with lower LV end-diastolic pressure, preserved atrial dimensions, preserved atrial effective refractory period, lower vulnerability to atrial fibrillation induction, less atrial interstitial fibrosis, and lower levels of BNP, angiotensin II, and TNF-α, versus sham procedure.
Booth et al.	(41)	Ovine model	Pacing-induced HF	Catheter-based RDN	8–10 weeks after pacing started	HF + RDN versus HF + sham procedure	Compared with sham procedure, RDN was associated with reduction of the resting diastolic and mean blood pressure, but no change in the resting cardiac sympathetic nerve activity (measured 24 hours after RDN). RDN was associated with blunting of the baroreflex mediated increase of cardiac sympathetic nerve activity in response to fall in blood pressure.
Xie et al.	(42)	Pig model	Pacing-induced HF	Catheter-based RDN	1 week after pacing	HF + RDN versus HF without RDN	RDN was associated with lower LV end-systolic dimension, and improved systolic function; plasma renin and aldosterone were lower; renal function was preserved with no signs of renal damage (assessed 1 week after RDN), versus controls.

ACEI: angiotensin-converting enzyme inhibitors; ARB: angiotensin receptor blockers; BB: beta blockers; BNP: brain natriuretic peptide; CRP: C reactive protein; HF: heart failure; LV: left ventricular; MI: myocardial infarction; MMP 2: matrix metalloproteinase 2; MT: medical treatment; RDN: renal sympathetic denervation; TGF-β1: transforming growth factor β1; TNF-α: tumor necrosis factor α.

Clinical studies

Evidence from clinical studies on the efficacy and safety of RDN in patients with systolic HF is limited, probably due to concerns about deleterious reduction of BP (Table 2). The Renal Artery Denervation in Chronic Heart Failure (REACH-Pilot) study was a first-in-man small study designed principally to explore the safety of RDN in HF patients; RDN was associated with improved symptoms and exercise capacity at mid-term follow-up (46). However, the small sample size and absence of a control group preclude any meaningful conclusions on a beneficial effect. Two small studies (one observational and one randomized) addressed the efficacy of RDN in patients with systolic HF: both showed improvement of symptoms after RDN; additionally, the randomized study demonstrated improvement of functional capacity, LV EF, and less hormonal activation at mid-term follow-up (47,48). Another unpublished small randomized study reported improvement of LV function in patients with advanced HF who underwent RDN, versus medical treatment (49). Added to the small sample size of the aforementioned studies, the cohort was poorly described, and the device employed was not well specified. Moreover, in a small study of patients with chronic HF and refractory ventricular arrhythmia, RDN was associated with reduced arrhythmia burden at 3-month follow-up (50).

Table 2. Clinical studies of renal denervation in heart failure with reduced systolic function.

Study	ref.	Study type	Number of patients	Cohort	Device	Comparison	Follow-up duration	Results
Davies et al.	(46)	Single-arm, observational, pilot safety study	7	Patients with chronic systolic HF (mean age 69 years, BP 112/65 mm Hg, EF 43 ± 15%)	Ardian Symplicity catheter, Medtronic, US	Baseline values were compared versus follow-up	5 days in hospital, then 6 months	No hemodynamic disturbances occurred during in-hospital observation or during follow-up. At 6 months, RDN was associated with improved 6-minute walk distance (221 ± 33 to 249 ± 34 meters, p = 0.03), improved symptoms, and decreased requirement for diuretics in 4 patients. BP (Δ systolic –7.1 ± 6.9 mmHg, p = 0.35; Δ diastolic –0.6 ± 4.0 mmHg, p = 0.88), and renal function (Δ creatinine –5.7 ± 8.4 μmol/l, p = 0.52 and Δ urea –1.0 ± 1.0 mmol/l, p = 0.33) remained stable. Echocardiographic and tissue Doppler indices of systolic and diastolic function remained stable.
Dai et al.	(47)	Observational, pilot safety study	20	Patients with chronic refractory HF (ischemic or nonischemic, NYHA class III-IV)	Not specified	RDN (10 patients) versus standard MT (10 patients)	1 day in hospital, then 6 months	One day after RDN, 24-hour urine volume increased, and neurohormone levels decreased, versus preprocedural values, and versus standard MT group. At 6 months, RDN was associated with improved LV EF, LV end-diastolic dimension, and reduced BNP (p < 0.05 for all). No complications were observed.
Chen et al.	(48)	Pilot randomized controlled trial (1:1)	60	Patients with HF who were stable for ≥6 months	Thermocool catheter, Biosense Webster, US	RDN + MT versus MT alone	6 months	RDN was associated with improved LV EF (p < 0.001), 6-minute walk distance (p = 0.043), NYHA class (p < 0.001), N-terminal-pro-BNP (p < 0.001), and heart rate (p = 0.008), versus control. RDN was associated with improvement in all domains of SF-36 Health Survey scores (p < 0.05 for all), except bodily pain (p = 0.74). No severe adverse events were observed.
Taborsky et al.	(49)	Pilot randomized controlled trial (1:1)	51 (43 analyzed in a nonblinded fashion)	Patients with advanced HF	Symplicity catheter, Medtronic, US	RDN + MT versus MT alone	12 months	RDN was associated with improved LV EF versus baseline (p < 0.01), and versus the control group (p < 0.01); such improvement was not observed in the control group. Similar improvement occurred in the other indices of LV function versus baseline values (p < 0.01 for all), and versus the control group (p < 0.01 for all), such improvement was not observed in the control group. There was a trend toward less frequent HF hospitalization in the RDN versus the control group.

BNP: brain natriuretic peptide; BP: blood pressure; EF: ejection fraction; HF: heart failure; LV: left ventricular; MT: medical treatment; NYHA: New York Heart Association; RDN: renal sympathetic denervation.

Renal denervation in heart failure with preserved ejection fraction

Increasing evidence suggests a prime role of sympathetic overactivity in the phenomena of HF with preserved EF: LV hypertrophy, LV diastolic dysfunction. In an observational study, Brandt et al., reported regression of LV hypertrophy and improvement of LV diastolic function in patients with resistant hypertension who underwent RDN, versus control; regression of LV mass was most evident in patients with LV hypertrophy at baseline: it was due to reduction of LV wall thickness, rather than reduction of LV dimensions (51) (Table 3). Interestingly, regression of LV hypertrophy and improvement of diastolic function were most pronounced in patients with the most marked systolic BP reduction; yet, in six patients assigned as nonresponders for systolic BP reduction, notable reduction of LV mass and E/é were observed (51). In another single-arm observational study, regression of LV mass was comparable in all tertiles of baseline systolic BP, in all tertiles of systolic BP reduction, in all tertiles of baseline ambulatory BP, and in all tertiles of ambulatory BP reduction. Linear regression analysis revealed no correlation between systolic BP change and reduction of LV mass index. Similarly, improvement of diastolic function was comparable in all tertiles of baseline systolic BP, and in all tertiles of systolic BP reduction. Again, no correlation was observed between systolic BP change and improvement of diastolic function. Regression of LV hypertrophy was more pronounced in patients with a higher baseline LV mass index (p < 0.001 for trend) (52). Importantly, the results of the previous two studies suggest that regression of LV hypertrophy and improvement of diastolic function occur independent of BP reduction. Additionally, in a study by cardiac magnetic resonance imaging, RDN in patients with resistant hypertension was associated with reduction of LV mass at 6 months. Interestingly, EF and myocardial contractility improved after RDN in the subgroup with baseline systolic dysfunction (53). In contrast, a small randomized trial showed no benefit of RDN in patients with HF and preserved EF, at 12 months (54). Nevertheless, the trial was terminated prematurely due to recruitment difficulties, and was underpowered for the efficacy endpoints; any lack of difference between the two groups might be due to type II statistical error. Another concern was the potential limitations of the Symplicity TM ablation catheter used in that trial. Finally, in a small retrospective study of patients with resistant hypertension, RDN was associated with regression of LV hypertrophy; EF improved; however, diastolic function did not (55).

Table 3. Clinical studies of renal denervation in heart failure with preserved systolic function.

Study	Ref.	Study type	Cohort	Device	Endpoints	Follow-up duration	Results
Brandt et al.	(51)	Observational study	46 patients with treatment-resistant hypertension, versus 18 controls	Symplicity Flex catheter, Medtronic, US	Parameters of LV hypertrophy (IVS thickness, LVMI) and diastolic function (mitral E peak DT, IVRT, mitral annular é, E/é) by echocardiography	1 and 6 months	RDN was associated with decreased IVS thickness from 14.1 ± 1.9 mm to 13.4 ± 2.1 mm, and 12.5 ± 1.4 mm (p = 0.007 for trend, p = 0.032 versus control), and decreased LVMI from 112.4 ± 33.9 g/m^2 to 103.6 ± 30.5 g/m^2 and 94.9 ± 29.8 g/m^2 (p = 0.004 for trend, p = 0.004 versus control). RDN was associated with shortening of the mitral E peak DT (p = 0.003 for trend, p = 0.008 versus control), shortening of the IVRT (p = 0.002 for trend, p < 0.001 versus control), increase of the mitral annular é (p = 0.001 for trend, p = 0.023 versus control), and decrease of the E/é (p < 0.001 for trend, p = 0.001 versus control).
Schirmer et al.	(52)	Single-arm, observational study	66 patients with treatment-resistant hypertension	Symplicity Flex catheter, Medtronic, US	Parameters of LV hypertrophy (LVMI) and diastolic function (E/A ratio, mitral E peak DT, IVRT, mitral annular é) by echocardiography	6 months	Reduction of LVMI was comparable in all tertiles of baseline systolic BP (p < 0.01 in all tertiles, p = 0.52 for trend), in all tertiles of systolic BP reduction (p < 0.01 in all tertiles), in all tertiles of baseline 24-h ambulatory BP (p < 0.01 in all tertiles, p = 0.26 for trend), and in all tertiles of 24-h ambulatory BP reduction (p < 0.01 in all tertiles, p = 0.30 for trend). Improvement of diastolic dysfunction (increase of E/A ratio, decrease of mitral E peak DT and IVRT) was comparable in all tertiles of baseline systolic BP, and in all tertiles of systolic BP reduction. Mitral annular é increased comparably in all tertiles of systolic BP reduction.
Mahfoud et al	(53)	Multicenter Observational study	72 patients with resistant hypertension, versus 17 controls	Symplicity Flex catheter, Medtronic, US	Parameters of LV hypertrophy (LVMI) and systolic function (LV EF, circumferential strain) by cardiac magnetic resonance imaging	6 months	RDN was associated with 7.1% reduction of LVMI versus baseline (46.3 ± 13.6 g/m^1.7 versus 43.0 ± 12.6 g/m^1.7, p < 0.001); no change was observed in the control group (p = 0.653). In the subgroup with baseline systolic dysfunction (LV EF ≤50%) LV EF improved versus baseline (43% versus 50%, p < 0.001); no improvement was observed in the control group (p = 0.537). In the group with reduced myocardial contractility at baseline (circumferential strain of ≥−20%), circumferential strain improved by 21% versus baseline (p = 0.001), no improvement in the control group (p = 0.508).
Patel et al.	(54)	Multicenter small randomized (2:1) controlled trial	25 patients with heart failure and preserved LV EF	Symplicity catheter, Medtronic, US	Minnesota living with heart failure Questionnaire score, peak VO2 on exercise, BNP, E/é, LAVI, LVMI	12 months	There were no differences between the two groups for all the parameters tested. The trial was terminated prematurely due to recruitment difficulties.
de Sousa Almeida et al.	(55)	Retrospective Single-arm, observational study	65 patients with resistant hypertension enrolled (complete 12 month follow-up available in 31 patients)	Symplicity catheter, Medtronic, US (n = 25), EnligHTN, St. Jude Medical, US (n = 4), OneShot, Covidien, US (n = 2)	Parameters of LV hypertrophy (LVMI), LV systolic function (LV EF), and LV diastolic function (E/A ratio, mitral annular é, E/é) by echocardiography	12 months	RDN was associated with reduction of LVMI (152 ± 32 g/m^2 to 136 ± 34 g/m^2, p < 0.001). LV EF improved (p = 0.001). Parameters of diastolic function (mitral inflow E/A ratio, mitral annular é velocity, and mitral E/é) did not improve (p < 0.05 for all).

BNP: brain natriuretic peptide; BP: blood pressure; DT: deceleration time; IVRT: isovolumetric relaxation time; IVS: interventricular septum; EF: ejection fraction; LAVI: left atrial volume index; LV: left ventricular; LVMI: left ventricular mass index; RDN: renal sympathetic denervation.

Ongoing studies of renal denervation in heart failure

Renal Artery Denervation in Chronic Heart Failure Study (REACH) is a prospective randomized trial (RDN versus sham procedure) with the primary objective to explore improvement in symptomatology following RDN in patients with chronic HF due to systolic dysfunction who are maintained on the maximal tolerated medical therapy (Table 4). Other endpoints include peak VO2 on exercise, 6-minute walk distance, change of chemoreceptor sensitivity, NYHA class, and incidence of major adverse events, at 12-month follow-up (ClinicalTrials. gov Identifier: NCT01639378). The Symplicity HF is an ongoing phase II trial with the primary objective to explore safety of RDN in patients with symptomatic HF due to systolic dysfunction, moderate renal impairment, who are maintained on optimal medical treatment. Secondary outcome measures include LV function and renal function (ClinicalTrials. gov Identifier: NCT01392196). Another ongoing randomized study is designed to enroll patients with symptomatic systolic HF, whose systolic BP is ≥90 mm Hg. The primary outcome measure will be safety of RDN at 6 months; secondary outcome measures include: LV function, glomerular filtration rate, and symptomatology/quality of life (ClinicalTrials. gov Identifier: NCT02085668). The ongoing DIASTOLE randomized trial investigates the effect of RDN in patients with HF and normal EF. The primary objective is improvement of E/é at 12 months; secondary objectives include safety of RDN, magnetic resonance imaging parameters (LV mass, LV volume, EF, and left atrial volume), mIBG uptake and washout, brain natriuretic peptide, BP, heart rate variability, exercise capacity, and quality of life (ClinicalTrials. gov Identifier: NCT01583881) (56). The RESPECT-HF (Renal Denervation in Heart Failure Patients With Preserved Ejection Fraction) is another ongoing randomized study (RDN plus optimal medical treatment or optimal medical treatment alone) of patients with symptomatic HF and preserved EF, who experience episodes of acute decompensation (with or without background hypertension, controlled or uncontrolled), and whose systolic BP is ≥105 mm Hg. The primary outcome measure is left atrial volume index and LV mass index, at 6 months; secondary outcome measures include: VO₂ max and 6-minute walk distance, E/é, biomarkers of cardiac load and interstitial fibrosis, measures of ventricular-vascular function, quality of life, and a composite of death or hospitalization for HF (ClinicalTrials. gov Identifier: NCT02041130).

Table 4. Ongoing trials of RSD in patients with heart failure.

Official title	Design	Number of patients	Inclusion criteria	Intervention	Primary outcome	Status	Estimated completion date	Clinical trial identifier
Impact of renal artery denervation in patients with chronic heart failure compared with Sham procedure	Prospective randomized double-blind	100	HF, EF <40%, NYHA class ≥ II, OMT	RDN + medical treatment versus sham procedure + medical treatment	Improvement of symptoms at 12 months (Kansas City Cardiomyopathy Questionnaire)	Recruiting	August 2017	NCT01639378
Renal denervation in patients with chronic heart failure & renal impairment	Prospective single-arm open-label	39	HF, EF <40%, NYHA class II-III, GFR 30-75 mL/min/1.73 m^2, OMT	RDN	Safety (adverse events) at 6 months	Ongoing but not recruiting	July 2016	NCT01392196
A prospective, multicenter, randomized, open-label, feasibility, safety and efficacy study of renal denervation in patients with chronic heart failure (CHF)	Prospective randomized open-label	100	HF, EF 10-40%, NYHA class II-III, GFR >30 mL/min/1.73 m^2, OMT, BNP >100 pg/ml or NT-Pro-BNP >400 pg/ml	RDN + medical treatment versus medical treatment	Safety (periprocedural adverse events) at 6 months	Not yet open for recruitment	May 2019	NCT02085668
Sympathetic renal denervation in heart failure with normal lv ejection fraction: denervation of the renal sympathetic nerves in heart failure with normal lv ejection fraction	Prospective randomized open-label	60	Symptomatic HFPEF: EF ≥50%, LV diastolic dysfunction, controlled hypertension, GFR >30 mL/min/1.73 m^2	RDN + medical treatment versus medical treatment	Change of E/é from baseline to 12 months	Ongoing but not recruiting	July 2016	NCT01583881
Renal sympathectomy in heart failure (the RESPECT-HF Study) - a study of renal denervation for heart failure with preserved ejection fraction	Prospective randomized open-label	144	HFPEF: EF ≥50%, NYHA class ≥ II, LV diastolic dysfunction (E/é>15) and/or BNP >220 pg/ml, GFR >30 mL/min/1.73 m^2	RDN + medical treatment versus medical treatment	LAVI and LVMI assessed by cardiac magnetic resonance imaging at 6 months	Recruiting	December 2016	NCT02041130

BNP: brain natriuretic peptide; EF: ejection fraction; GFR: glomerular filtration rate; HF: heart failure; HFPEF: heart failure with preserved ejection fraction; LAVI: left atrial volume index; LV: left ventricular; LVMI: left ventricular mass index; NT-Pro-BNP: N-Terminal Pro-brain natriuretic peptide; NYHA: New York Heart Association; OMT: optimal medical treatment; RDN: renal sympathetic denervation..

Renal denervation for treatment-resistant hypertension

So far, seven randomized controlled trials assessed the efficacy and safety of RDN in the treatment of patients with resistant hypertension, at various levels of BP control (57–63). Despite the initial encouraging results of the Simplicity HTN-2 (n = 45, open-label) trial that demonstrated better reduction of 6-month office systolic BP in patients who underwent RDN, versus those who were maintained on optimal antihypertensive treatment alone, the results of the larger and more well-designed Simplicity HTN-3 (n = 491, blinded) trial were disappointing; reduction of 6-month office systolic BP in patients who underwent RDN was comparable to those who underwent sham procedure (57,58). Similarly, post hoc analysis of the Simplicity HTN-3 trial did not show a benefit of RDN, versus sham procedure, for reduction of the 24-hour ambulatory systolic BP, at the same time point (64). Yet, the recently published DENERHTN (n = 101, open-label) trial demonstrated significant reduction of the daytime ambulatory systolic BP, versus standardized stepped-care antihypertensive treatment, at 6-month follow-up (60). A recent meta-analysis of all seven available randomized controlled trials (Simplicity HTN-3 represents >50% of the pooled cohort) again demonstrated no significant benefit of RDN in terms of reduction of office or ambulatory BP, versus control, at 6-month follow-up (65). Advocates of RDN argue that the negative results of the Simplicity HTN-3 trial were due to inefficient denervation procedure, and poor selection of hypertensive patients.

Summary of the available evidence

Evidence supports the safety – and probably efficacy – of RDN in HF (both with reduced and preserved EF) at mid-term follow-up. Extensive evidence from experimental studies – and modest evidence from clinical ones – supported a benefit of RDN in HF with reduced EF. In experimental studies, RDN was associated with better natriuresis, better cardiac remodeling, function, and hemodynamics, better autonomic balance and electrical stability, less neurohormonal activation, and less fibrosis (27–40,42). Consistent evidence from a few “underpowered” – mostly flawed – clinical studies in patients with HF and reduced EF demonstrated that RDN is associated with improvement of symptoms, exercise capacity, LV systolic function, less LV remodeling, and less neurohormonal activation, at mid-term follow-up (level of evidence C). Inadequate sample size, lack of control in most studies, and being based on surrogate endpoint evaluation limit the interpretation and the conclusiveness of the results (46–49). Remarkably, data showing a potential benefit of RDN in animal models of HF may not be readily translated to patients with HF; possible explanations include different renal sympathetic nerve anatomy, poor patient selection, less efficacious RDN technique, and probable sympathetic nerve regrowth in humans. Yet, inconsistent evidence from small observational studies in patients with HF and preserved EF suggested that RDN is associated with regression of LV hypertrophy, and improvement of LV diastolic function, at mid-term (level of evidence C); again being underpowered and based on surrogate endpoints limits the conclusiveness of the findings (51–55). Evidence is awaited from adequately powered well-conducted randomized controlled trials that address “hard” clinical endpoints at long-term follow-up.

Disclosure statement

The authors report no conflicts of interest.

Funding

The current research article did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.