Commercial

The American Heart Association (AHA) defines blood pressure (BP) as a force that pushes blood through a network of arteries, veins, and capillaries. The BP reading is the result of two forces: systolic pressure, which occurs as blood pumps out of the heart and into the arteries; and diastolic pressure, which is created as the heart rests between heartbeats. Elevated BP, or hypertension, leads to harm by causing tiny tears in the interior lining (intima) of the arteries and coronary vessels, stimulating a local immune response in the endothelial cells within the arterial walls. In these regions, the arterial intima retains apolipoprotein B, which attracts lipid-rich macrophages (foam cells). 

Recommendations for BP generally target less than 130/80 mm Hg, although BP goal can vary (e.g., comorbidities, life expectancy). Hypertension is estimated to affect approximately 30% of the US population. It accounts for a high burden of morbidity related to stroke, ischemic heart disease, kidney disease, and peripheral arterial disease. An estimated one in four adults with hypertension have their hypertension under control, but the remaining 77% (93 million) remain uncontrolled. Uncontrolled hypertension is diagnosed when an individual's BP remains above targeted levels (typically ≥140/90 mm Hg) when an individual either is not using, or unable to use, treatments to control BP or when hypertension persists despite antihypertensive therapies. The definition of uncontrolled hypertension is inclusive of resistant hypertension, in which BP remains above the targeted range despite the use of three or more antihypertensive medications, including a diuretic, with complementary mechanisms of action. Several factors may contribute to uncontrolled hypertension, including nonadherence to medications, excessive salt intake, inadequate doses of medications, excess alcohol intake, volume overload, drug-induced hypertension, and other forms of secondary hypertension. Also, sometimes it is necessary to address comorbid conditions (i.e., obstructive sleep apnea) to control BP adequately.

Evidence 

Seven randomized controlled trials (RCTs) available for assessment consist of various designs to evaluate renal denervation systems in individuals with hypertension. Three trials—RADIANCE-HTN SOLO, RADIANCE-HTN TRIO, and RADIANCE II—were randomized, double-blind, sham-controlled studies conducted at multiple centers across Europe and the United States. The REQUIRE trial, conducted in Japan and South Korea, was a randomized, single-blind, sham-controlled study. Additionally, two multicenter, sham-controlled, single-blind trials, SPYRAL HTN-OFF and SPYRAL HTN-ON, were conducted across the United States, Canada, Japan, Europe, and Australia. The RADIOSOUND-HTN trial, a three-arm study conducted in Germany, compared uRDN and rfRDN ablation of either the main renal artery or of both the main renal artery, branches, and accessory arteries to each other.

The RADIANCE-HTN SOLO, RADIANCE-HTN TRIO, RADIANCE II, and REQUIRE trials evaluated the safety and efficacy of the uRDN system. These studies were conducted in populations of groups of individuals who were either not using (Off Med) or who were using (On Med) antihypertensive drugs.

The SPYRAL HTN-OFF and HTN-ON trials assessed the safety and efficacy of the rfRDN system, incorporating an adaptive Bayesian design with a pilot study followed by an expansion cohort. Like the RADIANCE trials, these studies were conducted in groups of individuals who were either not using (Off Med) or who were using (On Med) antihypertensive drugs. Finally, the RADIOSOUND-HTN trial employed a prospective, randomized design to directly compare the different renal denervation methods in individuals with resistant hypertension. Participants who were considered medication-adherent were admitted into the study and continued with their antihypertensive medications during the trial, with therapeutic adjustments as needed.

Study quality for RCTs was assessed using the US Preventive Services Task Force (USPSTF) Criteria for Assessing Internal Validity of Individual Studies. Most studies were unblinded during the follow-up period, which could impact the findings. All RADIANCE trials were sponsored by Recor, and all SPYRAL trials were sponsored by Medtronic. REQUIRE was funded by JIMRO Co. and Korea Otsuka Pharmaceutical. RADIOSOUND-HTN was funded by the Leipzig Heart Institute. All studies except RADIOSOUND-HTN received editorial or trial design input or medical writing support from the manufacturers, and all studies had declarations of support (e.g., receiving speaker or consulting fees) from manufacturers for one or more authors.

uRDN

The RADIANCE trials, including SOLO, TRIO, and RADIANCE II, reported adequate randomization and allocation of participants for RADIANCE II, SOLO, and TRIO, adequate blinding of participants (for 6 months), and outcome assessors. Intention-to-treat (ITT) analysis was conducted, and overall attrition was low. Retention in the SOLO trial was 100%, 95%, and 90% at the 2-, 6-, and 12-month follow-up visits, respectively. Retention in the TRIO trial was 100% and 95% at the 2- and 6-month follow-up visits, respectively. RADIANCE II has reported 2-month results with no loss to follow-up at this time.

The authors of the REQUIRE trial reported the possibility of participant sampling error such that a significant number of participants with uncontrolled hypertension and poor drug adherence may have been enrolled in the study and may have improved their drug adherence during the study. Although participants were blinded, clinical staff were not blind to treatment condition, which is a critical study limitation under the USPSTF framework. There was no standardization of antihypertensive medications (AHM). Additionally, authors reported poor drug adherence in nearly half the participants during the study. There were no other major methodological concerns. Randomization and allocation concealment were adequate, as was the blinding of participants and outcome assessors. No ITT analysis was performed, but attrition was low (~95% retention).

rfRDN

The PYRAL HTN-OFF MED and SPYRAL HTN-ON MED trials received ratings of "Good." Randomization and allocation concealment were adequate. There was adequate blinding of participants across studies. In both HTN OFF MED and HTN ON MED, pilot cohort participants remained blinded for 12 months, and expansion cohort participants remained blinded for 6 months. There was adequate blinding of outcome assessors. ITT analyses were performed, and there were no other major methodological concerns.

The RADIOSOUND-HTN was rated “Poor" because only the participants were blinded. Under the USPSTF framework, the lack of blinding of outcome assessors is considered a critical design limitation. Additionally, there was no standardization of AHM, and adherence was not tested. Randomization was adequate, and while no ITT analysis was performed, attrition was low (~95% retention).

A total of 1465 participants were enrolled across seven RCTs, with 506 participants in the RADIANCE trials, 366 in the HTN-OFF MED trial, 337 in the HTN-ON MED trial, 136 in the REQUIRE trial, and 120 in the RADIOSOUND trial. Follow-up periods for participants ranged from 2 to 36 months. Across all trials, the mean age of participants ranged from 51.4 years in the HTN-OFF MED trial to 64.6 years in the RADIOSOUND trial. However, the overall proportion of participant population aged 65 and older was not reported.

Overall, these seven trials represented a broad spectrum of hypertensive populations, ranging from mild to resistant forms of hypertension. The RADIANCE-HTN trials targeted two distinct hypertensive groups: those with mild to moderate hypertension not on medications at the time of enrollment (SOLO and RADIANCE II) and those with resistant hypertension (TRIO). In contrast, the SPYRAL HTN-OFF and HTN-ON trials focused on individuals with mild to moderate hypertension, and HTN-OFF was further limited to participants able to discontinue AHMs, while HTN-ON included individuals on stable regimens of one to three AHMs. These two trials were conducted in two phases: an initial pilot cohort to assess feasibility and an expansion cohort utilizing an adaptive Bayesian design. The REQUIRE and RADIOSOUND trials both studied individuals with resistant hypertension.

Across all trials, baseline characteristics between the RDN (uRDN or rfRDN) and sham groups were generally well balanced, with minor variations. Key parameters such as age, sex, body mass index (BMI), and BP showed close alignment in most studies. For instance, age and BMI were comparable across groups in trials like SOLO, Radiance II, and HTN-OFF. BP, both office and 24-hour measurements, also showed similar baseline levels between the RDN and sham groups, ensuring fair comparisons. Minor differences included slightly higher male representation in the sham groups of some trials, such as RADIANCE II and REQUIRE. Racial and ethnic status, however, was not always thoroughly reported, and enrollment of different racial and ethnic groups was limited across studies, with most participants being White. In the studies in which racial backgrounds were properly reported (RADIANCE SOLO, RADIANCE II, and TRIO), the proportion of African-Americans in these studies was similar to their overall proportion in the general US population. However, resistant hypertension is more prevalent in African-Americans (Sarafidis et al., 2013). Hispanic and Asian-American individuals were generally underrepresented in these studies. Despite this, the racial distribution within each trial (when reported) was generally balanced between the groups. Background medical therapy was not standardized in SPYRAL HTN-ON MED, and it was not in the form of single-pill combination therapy in the RADIANCE HTN TRIO trial. This study design left room for adding or withdrawing medications during the studies. 

Ultrasound Renal Denervation (uRDN) 
The RADIANCE studies enrolled individuals aged 18 to 75 years with a documented history of hypertension, suitable renal anatomy for the renal denervation procedure confirmed by recent renal computed tomography angiography (CTA) or magnetic resonance angiography (MRA), and the ability to comply with study procedures. The studies RADIANCE-HTN SOLO (Azizi et al., 2018), RADIANCE-HTN TRIO (Azizi et al., 2021), and RADIANCE II (Azizi et al., 2023) were conducted at multiple centers across Europe and the United States. The REQUIRE trial was conducted in Japan and South Korea (Kario et al., 2022).

The efficacy of uRDN for lowering BP was mixed between the Off-Med (SOLO and RADIANCE II) studies and between the Off-Med and On-Med (TRIO and REQUIRE) studies. Use of uRDN in the Off-M​​ed SOLO and RADIANCE II trials showed consistent reductions in office systolic BP and daytime ambulatory BP (ASBP) compared to the sham intervention at 2 months follow-up in participants with mild-to-moderate hypertension and individuals with stage 2 hypertension, although the evidence is more varied at later time points, and the evidence is limited for durability of effect. In individuals with resistant hypertension, findings differed between the two On-Med trials. In TRIO, uRDN reduced daytime ASBP statistically, but not meaningfully, more than the sham procedure at 2 months but not at 6 months. The smaller REQUIRE trial found no difference in ASBP changes between uRDN and sham at the 3-month follow-up. There were relatively few treatment-related adverse events across studies.

Mild-to-Moderate Hypertension 

In individuals with mild to moderate hypertension (SOLO; Azizi et al., 2018), at 2 months the uRDN group (n=74) achieved a clinically significantly greater reduction in office SBP relative to the sham group (n=72), with a baseline-adjusted difference (AD) between groups of −6.5 mm Hg (95% confidence interval [CI], −11.3 to −1.8; P=0.007) as well as daytime ASBP (−6.3 mm Hg; 95% CI, −9.4 to −3.1; P=0.0001). At 6 months, those in the SOLO trial were still blinded to treatment condition (Azizi et al., 2019) and had received recommended and standardized stepped-care antihypertensive treatment since the 2-month follow-up if home BP control did not achieve the target range (home BP ≥135/85 mm Hg). At the 6-month time point, 65.2% of participants in the uRDN group were treated with antihypertensive medications (AHM) versus 84.5% in the sham group (P=0.008); the number of medications used and the defined daily dose (DDD) were lower in the uRDN group relative to the sham group (AHM, 0.9 ± 0.9 vs 1.2 ± 0.9; P=0.043 and DDD: 1.4 ± 1.5 vs 2.0 ± 1.8; P=0.018; respectively). Reductions in out-of-office BP (OSBP) did not differ between the groups (AD, −3.7 mm Hg; 95% CI, −8.1 to 0.7 mm Hg; P=0.102; adjusted for baseline value and AHM). However, reductions in ASBP were statistically different between the groups (AD, −4.3 mm Hg; 95% CI, −7.9 to −0.6 mm Hg; P=0.024; adjusted for baseline value and AHM), but this difference may not be clinically meaningful. At the 12-month follow-up (Azizi et al., 2020), participants were no longer blind to treatment condition and were treated with AHM according to standard of care (i.e., not standardized). The number of medications (1.0 vs. 1.4; P=0.015), and DDD (1.4 vs. 2.2; P=0.007) were less with uRDN versus sham. There was no difference in daytime ASBP between the groups (AD, −2.3 mm Hg; 95% CI, −5.9 to 1.3; P=0.201; adjusted for baseline value and AHM), but OSBP was clinically meaningfully decreased (AD, −6.3 mm Hg; 95% CI, −11.1 to −1.5 mm Hg; P=0.010).

Regarding safety outcomes, the SOLO trial reported no major adverse events in either group at 30 days or 6 months (Azizi et al., 2019). At the 12-month follow-up, one individual in the sham group died (suicide), and one experienced a cerebrovascular event. Neither group had any other major adverse events (Azizi et al., 2020). Twelve-month imaging was available in 63 RDN participants. One study enrollee underwent renal artery stent placement at 6 months after mild progression of ostial renal artery plaque. No new renal artery stenosis greater than 70% was detected on CTA or MRA of the renal arteries at 12 months, and the calculated eGFR remained stable from baseline to 12 months in the RDN group (Azizi et al., 2020).

Of note, the SOLO trial was not designed or powered to demonstrate a difference in BP between the uRDN and the sham beyond 2 months, and sham participants with persistent uncontrolled hypertension were eligible to crossover and receive uRDN. A single-arm follow-up study of participants who remained in the uRDN arm from randomization reported findings at 36 months (Rader et al., 2022). In participants with uncontrolled BP at screening (n=36), office systolic BP decreased by 10.8 mm Hg (no estimate of variability; P<0.001) at 36 months. This is a small subset of participants, and these findings regarding durability should be interpreted in that context. No new adverse events were deemed related to the procedure at this follow-up.

Stage 2 Hypertension

In individuals with stage 2 hypertension (RADIANCE II; Azizi et al., 2023) those randomly assigned to uRDN (n=150) versus sham (n=74) demonstrated a clinically meaningful reduction in OSBP with a between-group difference of −5.4 mm Hg (95% CI, -9.0 to −1.8; P=0.004) and in daytime ASBP (AD, −6.3 mm Hg; 95% CI, −9.3 to −3.2; P<0.001) at 2 months follow-up with no major adverse events in either group. This study's follow-up is short, and planned observational data at 6 and 12 months have yet to be reported.

Resistant Hypertension
The findings in individuals with resistant HTN are mixed. Both studies were conducted in individuals receiving concomitant antihypertensive drug therapy. In the TRIO trial, participants with resistant hypertension on fixed-dose, triple combination therapy (Azizi et al., 2021), uRDN (n=69) reduced OSBP (unadjusted median difference: −7.0: 95% CI, −13.0 to 0.0; adjusted P=0.037) and daytime ASBP (unadjusted median difference: −4.5 mm Hg; 95% CI, −8.5 to −0.3; adjusted P=0.022) more than sham (n=67). These findings are statistically different but may not be clinically meaningful. However, at 6 months, the per-protocol analysis indicated no significant difference between treatment conditions in either OSBP (AD, 0.7; 95% CI, −5.3 to 6.6; P=0.93) or daytime ASBP (AD, −0.0 mm Hg; 95% CI, −4.6 to 4.5; P=0.65; Azizi et al., 2022). In REQUIRE, participants with resistant hypertension and standard-of-care medication therapy (Kario et al., 2022), uRDN (n=69) and sham (n=67) treatments did not differ at 3-month follow-up for either OSBP (Least Squares Mean Difference [LSMD], −2.0; Standard Error of the Mean (SEM), 3.0; 95% CI, not reported; P=0.511) or 24-hour ASBP (LSMD, −0.1; SEM, 2.1; 95% CI, −5.5 to 5.3; P=0.971).

Regarding safety outcomes, three adverse events were reported in the TRIO trial after uRDN within 30 days of the procedure (Azizi et al., 2022). Six other cardiovascular or kidney events through 6 months were reported in four participants in each group, and eGFR decreased slightly and similarly from baseline to 6 months in the 2 groups. No new 50% or greater kidney artery stenosis was detected on non-invasive imaging in either group at 6 months (Azizi et al., 2022). REQUIRE did not report any major procedure- or device-related adverse events. However, vasospastic angina was seen in one participant, and a puncture site hemorrhage occurred in another during the RDN procedure (Kario et al., 2022).

It should be noted that the populations differed between TRIO and REQUIRE. Individuals enrolled in TRIO were primarily White and residing in Western countries (e.g., the United States, the United Kingdom, Germany, and France), whereas individuals enrolled in REQUIRE were Asian and resided in Japan or South Korea. Baseline enrollment criteria differed between the two studies, with TRIO enrolling participants with office BP of at least 140/90 mm Hg despite using three or more medications, including a diuretic. Participants in TRIO received a standardized fixed-dose triple-drug regimen. REQUIRE enrolled individuals with a seated office BP of at least 150/90 mm Hg with the same medication treatment requirements. Drug treatment in REQUIRE was defined as standard of care (i.e., not standardized). Although the baseline measures of office BP appear similar between the two trials, the concomitant drug therapies differed between the studies, and there may be differences between the two populations; the data should be interpreted in this context. Some additional limitations were noted in the REQUIRE trial. First, the participants were blinded, but the physicians were not, and there was no objective assessment of medication adherence, as measured by drug concentrations in blood or urine. The authors suggest that there may have been several participants with poor medication adherence who, during the study, were more compliant with treatment, leading to similar improvements in BP between the two treatment conditions, minimizing the impact of uRDN on BP. Additionally, the mean SBPs in this study were similar for AMBP and OBPM, further suggesting possible sampling error or compliance issues.

Of further note, trial blinding was only maintained for 6 months in the TRIO trial, after which medication changes could be made as needed. Beyond the 12-month follow-up, only OSBP measurements were recorded. A single-arm follow-up study of 49 individuals who remained in the uRDN arm from randomization reported findings at 36 months (Bloch et al., 2024). In these participants, OSBP was reduced by an average of −14.5 mm Hg (standard deviation [SD], 26.1) from screening and an average of −8.0 mm Hg (SD, 24.5) from baseline while on an average of 3.7 AHMs. Like the long-term findings for the SOLO trial, these data are from a small subset of participants; thus these results regarding durability should be interpreted in that context. No new adverse events were deemed related to the procedure at this follow-up.

The uRDN procedure effectively reduced BP across all three RADIANCE studies in short-term follow-ups. These reductions exceeded the clinically meaningful threshold of 5 mm Hg (Haberman et al., 2024) and, should the effect be durable, may contribute to a relative risk reduction in cardiovascular events. However, the long-term superiority of the uRDN over sham interventions is not well established. The SOLO, TRIO, and RADIANCE II trials were not designed or powered to assess differences in BP between uRDN and sham interventions beyond 2 months. As a result, individuals on sham intervention with persistent uncontrolled hypertension were eligible to crossover and receive uRDN. Consequently, the long-term treatment durability effects were assessed only in a small subset of participants originally randomly assigned to the uRDN group.

Overall, few serious adverse device- or procedure-related events (<3%) have been seen with uRDN, of which all have been transient and resolved with no long-term sequelae. Injury to the renal artery and/or the kidneys is rare. There have been no reports of new-onset clinically significant renal artery stenosis requiring intervention. The totality of evidence from the RADIANCE trials suggests a positive benefit-risk profile of the uRDN. However, the trials' exclusion criteria were strict, omitting individuals with renal abnormalities, aneurysmal renal arteries, previous renal stenting, or eGFR below 40 mL/min/1.73 m². Renal anatomy suitability was confirmed via renal CTA or MRA before the procedure and renal angiography during the procedure, ensuring safe RDN application. Notably, medication adherence and eGFR analyses were not protocol-mandated beyond the 12-month follow-up, which raises considerations for long-term kidney function monitoring post-procedure. Safety findings should be considered within the context of this highly selected participant population, and applicability may be limited when considering the broader real-world population of participants with HTN.

Catheter-Based Radiofrequency Renal Denervation (rfRDN)

The SPYRAL HTN studies enrolled individuals aged between 20 and 80 with baseline ASBP 140 to 169 mm Hg and OSBP 150 to 179 mm Hg. Key exclusion criteria included recent angina or myocardial infarction; heart failure, stroke, transient ischemic attack, or atrial fibrillation; prior renal denervation; untreated secondary causes of HTN; uncontrolled diabetes; and renal anatomy ineligible for treatment. SPYRAL HTN-OFF MED studies (Townsend et al., 2017; Böhm et al., 2020) were further limited to individuals able to discontinue antihypertensive medications, while SPYRAL HTN-ON MED studies (Kandzari et al., 2018; Mahfoud et al., 2022; Kandzari et al., 2023; Kario et al., 2023) included participants on stable regimens of one to three AHMs. The SPYRAL HTN studies were conducted across the United States, Canada, Japan, Europe, and Australia.

The SPYRAL HTN trials assessed the efficacy of rfRDN in reducing BP in participants with uncontrolled hypertension. The SPYRAL HTN trial began with two international, multicenter, sham-controlled pilot studies exploring the rfRDN catheter system in participants with mild to moderate hypertension: SPYRAL HTN-ON MED, involving individuals on AHMs, and SPYRAL HTN-OFF MED, involving participants off AHM therapy. Each pilot study included 80 participants and, while not powered for efficacy outcomes, demonstrated significant BP-lowering effects following rfRDN with minimal BP changes in the sham groups (Townsend et al., 2017; Kandzari et al., 2018). Following these pilot studies, two larger prospective trials were launched to detect changes in 24-hour ASBP. These trials included the SPYRAL HTN-OFF MED Pivotal trial (for participants without AHMs; Böhm, 2020) and the SPYRAL HTN-ON MED Expansion trial (for participants with AHMs; Kandzari et al., 2023; Mahfoud, 2022). Using an adaptive Bayesian design, these studies incorporated data from the pilot trials as informative priors and conducted interim analyses, allowing for early study termination for efficacy or futility.

Uncontrolled Hypertension

Off Medication

The SPYRAL HTN-OFF Pivotal trial (n=251; Böhm et al., 2020) combined data from this named trial and the HTN-OFF Pilot trial (Townsend et al., 2017) using a Bayesian approach to incorporate data from the pilot (n=80) as an informative prior into the primary analysis (total n=331). Medications were discontinued 3 to 4 weeks before randomization through 3 months postprocedure unless they met safety escape criteria requiring reintroduction of medications (SBP >180 mm Hg or BP-related symptoms or complications). After the primary efficacy follow-up visit at 3 months, participants were treated with a guideline-based AHM escalation protocol between 3 and 6 months postprocedure if needed. Crossover to RDN was permitted after 12 months for the pilot cohort and 6 months for the pivotal cohort.

In the rfRDN group, there was a greater mean reduction in OSBP (MD, −6.6 mm Hg; 95% Bayesian credible interval [BCI], −9.6 to −3.5) and 24-hour ASBP (MD, −4.0 mm Hg; 95% BCI, −6.2 to −1.8) at 3 months relative to sham. For both endpoints, rfRDN met the statistical requirement for superiority (>0.975) with a posterior probability of superiority (>0.999), and the authors indicate that the differences in SBP measurements were clinically meaningful at 3 months. No major procedural or device-related safety events were reported.

On Medication

SPYRAL HTN-ON MED Expansion trial combined data from this trial (n=257; Kandzari et al., 2023) and the HTN-ON proof-of-concept trial (Kandzari et al., 2018) using a Bayesian approach to incorporate data from the pilot (n=80) as an informative prior into the primary analysis (total randomized n=337). Participants were prescribed one, two, or three standard AHMs, and the primary efficacy follow-up was at 6 months. At this time point, there were no between-group differences for the primary endpoint of 24-hour ABSP (MD, −0.03 mm Hg; 95% BCI, −2.82 to 2.77 mm Hg). The secondary effectiveness endpoint was the baseline-adjusted change in OSBP from baseline to 6 months postprocedure. In the rfRDN group, there was a greater reduction in OSBP at 6 months versus the sham group (MD, −4.9 mm Hg; 95% BCI, −7.9 to −1.9), but this may not be clinically meaningful, and the authors note that the study was not powered for this endpoint. The mean number of AHMs at 6 months was lower in the rfRDN group compared with the sham group (1.9 vs. 2.1; P=0.0085); the medication burden (based on number, class, and dosage) at 6 months was 2.9 in the rfRDN group versus 3.5 in the sham group (P=0.043). One individual in the rfRDN group required pseudoaneurysm repair at the right femoral access site.

A pre-specified win ratio analysis (to address the potentially confounding effect of medication burden) applied a hierarchical composite of outcomes to compare the rfRDN and control groups using two endpoints. The first endpoint was the difference in 24-hour ASBP change from baseline to 6 months using a threshold of 5 mm Hg. Win-ratio analysis favored rfRDN treatment versus the sham intervention (1.50; 95% CI, 1.13–1.99; P=0.005; Kandzari et al., 2023).

Safety and efficacy follow-up data for SPYRAL HTN-ON MED at 12 and 24 months (Kandzari et al., 2025) were published after our initial analysis of the evidence. Once the primary endpoint data were collected at the 6-month follow-up, participants and clinicians were unblinded, and participants in the sham group were permitted to cross over to the rfRDN group. Half of the control participants elected to receive rfRDN; 54 participants crossed over between the 6- and 12-months visit, and a further 12 participants crossed over between the 12- and 24-months visit. ITT analysis used the last observation carried forward method, and participants who crossed over to the rfRDN group were censored from the dataset at the time of crossover. At 12 months, there were no differences between rfRDN and sham for either 24-hour ASBP (−0.06; P=0.71) or OSBP (−3.1; P=0.15). At 24 months, however, both 24-hour ASBP and OBSP were statistically improved for participants in the rfRDN group (−5.7; P=0.039 and −8.7; P=0.0034, respectively). The authors suggest that, at 12 months, individuals in the sham condition were showing decreases in SBP measures with a statistically significantly higher medication burden than those in the rfRDN group. At 24 months, the ambulatory systolic BP change was significantly greater in RDN participants compared with sham control participants (−12.1±15.3 mm Hg vs −7.0±13.1 mm Hg; treatment difference, −5.7 mm Hg; P=0.039) (Kandzari et al., 2025). At this time point, the sham population was half of the enrolled participants at randomization (50% attrition) due to crossover and other loss to follow-up, and there was only 3.4% attrition in the rfRDN group. Analysis of participants who crossed over to rfRDN from sham showed that there were few differences in baseline characteristics relative to those who did not with the exception of more participants remaining in the sham group being smokers. At 12 months after crossover, participants in this group showed decreases in 24-hour ASBP and OSBP relative to baseline (−14.0±13.3; P<0.0001 and −19.1±16.9; P<0.0001, respectively) and relative to the point of crossover (−4.1±14.1; P=0.027 and −8.5±16.3; P<0.0001, respectively). Adverse events were similar between the two groups, with approximately 2.5% of participants in each group meeting the composite safety endpoint, which consisted of the following events. At 24 months, there were no renal reinterventions or renal stenosis greater than 70% in the rfRDN group. Two deaths occurred, one in each group, two strokes occurred, one in each group, two individuals required treatment for vascular complications in the rfRDN group as did one in the sham group. In the rfRDN group, one individual ​​​experienced a significant embolic event resulting in end-organ damage, one required hospitalization for hypertensive crisis, with no participants in the sham group experiencing these events.

Long-term safety and efficacy at 36 months were assessed in a very small subset of the initial cohort of the SPYRAL HTN-ON MED proof-of-concept trial (Mahfoud et al., 2022). In individuals with uncontrolled hypertension, there were modest BP reductions at 36 months in 24-hour ASBP in the rfRDN group (n=30) compared to sham (n=32), but there were no differences in OSBP (MD, −8.2; 95% CI, −17.1 to 0.8; P=0.073). A second exploratory analysis (Kario et al., 2023) of a subset of these participants taking at least three AHMs at 36 months found similar results as the Mahfoud et al. analysis for 24-hour and OSBP outcomes. Their analysis found that 24-hour SBP was controlled to less than 130 mm Hg in a greater percentage of rfRDN participants both in the morning and at night compared to the sham group (Morning 40% vs. 6%, P=0.021; Night 80% vs. 39%, P=0.019). The authors suggest that rfRDN may be beneficial at the times of day when cardiovascular risk is elevated, but these findings should be interpreted with consideration for the small sample size and the post hoc nature of the analyses. No further safety concerns were noted at 36 months postprocedure.

Townsend et al. (2024) compared AHM and BP changes among different prespecified patient subgroups based on geography (within the United States and outside the United States), age, sex, BMI, baseline 24-hour and OSBP, and race within the United States only (Black vs non-Black). Most participants (n=187; 54%) were enrolled outside the United States, whereas 156 (46%) individuals from the United States were enrolled, including 60 (18%) African-American individuals.

Among participants outside the US subgroup (e.g., Europe, Japan, and Australia), both 24-hour ASBP (AD, −4.8 mm Hg; 95% CI, −7.6 to −2.0; ​P=0.001) and OSBP (AD, −6.7 mm Hg; 95% CI, −10.5 to −2.8; P<0.001) decreased statistically significantly in the rfRDN group compared with the sham group. Within the US subgroup, there were no differences in 24-hour ASBP between rfRDN and sham treatment across all US enrollees (data shown graphically; P=0.21) and when considering African-American individuals (AD, 5.4 mm Hg; 95% CI, −3.4 to 14.1; P=0.22) and non–African-American individuals separately (AD, −0.2 mm Hg; 95% CI, −4.8 to 4.3; P=0.92). The authors noted that the treatment effect in African-American individuals was slightly greater in sham relative to rfRDN, but this difference was not significant. Similarly, there were no differences between treatment groups for OSBP measures (US African-American AD, −3.4 mm Hg; 95% CI, −12.5 to 5.7; P=0.46; Non–African-American US: AD: −2.4 mm Hg; 95% CI, −8.0 to 3.1; P=0.38). In a newly published paper reporting on available data from the complete cohort (Kandzari et al., 2025), the findings are substantively similar to what was reported in this paper. The authors caution that the study was not adequately powered to rigorously determine any subgroup differences.

There were no differences between prespecified participant groups for either 24-hour ASBP or OSBP across individuals younger than 65 or 65 and older (24-hour ASBP, P=0.99; OSBP, P=0.21), sex (24-hour ASBP, P=0.84; OSBP, P=0.18); BMI (24-hour ASBP, P=0.66; OSBP, P=0.49); or baseline SBP (24-hour ASBP, 0.99; OSBP, P=0.37; all data shown graphically; Townsend et al., 2024). Although these findings should be interpreted within the context that subgroup analyses may pose increased risk of type 1 error (false-positives arising from multiple comparisons), reduced statistical power for any observed effects, and potential for misleading interpretations when applied to complex patient characteristics, these data provide some information that may be of relevance to participants eligible for Medicare services.

While the SPYRAL HTN-OFF MED Pivotal trial demonstrated clinically meaningful reductions in both OSBP and 24-hour ASBP at 3 months, the SPYRAL HTN-ON MED expansion trial did not achieve similar results at 6 or 12 months, although a significant between-group difference was seen at 24 months (Kandzari et al., 2025). Notably, differences in hypertension medication burden between the groups confound interpretation. At this time point, only 50% of the participants randomly assigned to the sham condition remained in that condition, and between the 12- and 24-month follow-up, BP increased in the sham condition while BP continued to decrease in the rfRDN group. The authors note that most participants in the HTN-ON trial had their 6-month follow-ups scheduled during the COVID-19 pandemic, and there were more protocol deviations during the expansion study compared to the pilot study. Further, it was proposed in an analysis of pre-specified subgroups (Townsend et al., 2024) that an unexpectedly large BP decrease in the sham group may have been due to, in part, more common increases in AHMs in the US cohort of African-American individuals. This study had a small sample size, which may limit the interpretation and generalizability of the subgroup analysis.

Across the SPYRAL trials, safety remained favorable, with a 0.04% incidence of major adverse events in the first 253 enrollees, which met the predefined safety endpoint performance goal of 7.1% (P<0.001). The primary safety endpoint was the rate of major adverse events at 30 days postprocedure and renal artery stenosis at 6 months in RDN-treated subjects pooled from HTN-OFF and HTN-ON studies (Böhm et al., 2020). No device-related adverse events were reported up to 24 months of follow-up (Kandzari et al., 2023 and 2025).

Head-to-Head Comparison of uRDN and rfRDN
RADIOSOUND-HTN was a head-to-head comparison of three different techniques of RDN in individuals with resistant hypertension: rfRDN (main artery: RFM-RDN; n=39) as the reference standard, rfRDN (main and branch arteries; RFB-RDN: n=39), and uRDN (USM-RDN; n=42). Key inclusion criteria for the RADIOSOUND-HTN trial were individuals diagnosed with resistant hypertension with daytime SBP greater than 135 mm Hg on ABPM, with a renal artery diameter of 5.5 mm Hg or greater for at least one of the main renal arteries. ​Use of AHM was required to be stable for at least 4 weeks. Participants then underwent ABPM to exclude those with white-coat hypertension. Exclusion criteria were age 75 years and older, pregnancy, life expectancy less than 6 months, evidence for secondary hypertension, participation in any other randomized clinical trial, known renal artery stenosis or anatomy unsuitable for interventional RDN, and any main renal artery diameter less than 4.0 mm. The trial was conducted at a single center in Germany (Fengler et al., 2019).

RDN lowered BP across all groups at all time points. At 3 months, daytime systolic and diastolic BP decreased significantly in the overall cohort by 9.5 mm Hg (SD, 12.3 mm Hg) and 6.3 (SD, 7.8 mm Hg), respectively (P<0.001 for both) and within each treatment group (P<0.001) (Fengler et al., 2019). This effect was maintained at 6- and 12-month follow-up (​P<0.001) (Fengler et al., 2023).

At 3 months, uRDN was found to be more effective than rfRDN (Daytime ASBP MD, −6.7 mm Hg; 98.3% CI, −13.2 to −0.2; adjusted P=0.043). rfRDN did not differ between uRDN (Daytime ASBP MD, −4.9 mm Hg; 98.3% CI, −11.5 to 1.7; P=0.22) or by rfRDN approach (Daytime ASBP MD, −1.8 mm Hg; 98.3% CI, −8.5 to 4.9; P>0.99). Rates of response to RDN were similar across treatment groups. There were several periprocedural adverse events, including transient renal artery spasm, a symptomatic groin hematoma, and a pseudoaneurysm. The authors reported that all events were resolved. During the 3-month follow-up period, two individuals in the rfRDN group experienced symptomatic hypotension. Three individuals (main artery rfRDN: 1; main and branch arteries rfRDN: 2) experienced symptomatic hypertension that required medical treatment. One participant in the rfRDN group experienced acute decompensated heart failure, which required hospitalization. Finally, one person in the rfRDN group died from acute aortic dissection 2 months postprocedure. Examination of this individual's angiogram suggested no evidence of dissection at the time of the procedure. No adverse events were reported in the uRDN group, and there were no renal vascular complications or instances of renal stenosis.

In a research letter by ​Fengler et al. (2023), 6- and 12-month follow-up data were reported for the head-to-head comparisons. At 6 months, ASBP reduction from baseline varied across the treatment arms, with uRDN producing a statistically greater effect than rfRDN (uRDN, −12.1 mm Hg; SD, 11.5; main artery rfRDN, −6.0 mm Hg; SD, 11.0; main and branch arteries rfRDN, −4.8 mm Hg; SD, 12.1; P=0.017 for between-group comparison). Given the wide variability around the point estimates for each treatment, this may not be a clinically meaningful difference. At 12 months, there were no longer any between-group differences in ASBP (Fengler et al., 2023). Harms were not reported.

In addition to the seven RCTs considered previously, a series of observational studies report broadly consistent findings with those from the RCTs, both in terms of enrolled participants and safety and efficacy outcomes. While data on efficacy in these single-arm trials is reported, these findings are limited by the lack of a meaningful comparator, and, while provided for completeness, in this analysis, greater weight is given to data from the RCTs. Observational trials, however, can be useful in identifying safety signals, particularly if longer follow-up times are reported than are available from RCTs. 

Uncontrolled Observational Studies

Five uncontrolled observational studies met the criteria for inclusion in this review. Two of these studies were registry derived (Rosch et al., 2023) and pooled longer term data from two of the other included observational studies (Fengler et al., 2017, 2022); one was a cohort review, and two were prospective single-arm studies (one a feasibility trial). Two single-arm uRDN studies reported short-term efficacy follow-up findings for 24-hour systolic ABPM at 3 months (Fengler et al., 2017, 2022). Both studies demonstrated a statistical reduction from baseline for 24-hour systolic and diastolic ABPM at 3 months. Fengler et al. (2017) reported very few adverse events, and the same authors (2022) did not report safety outcomes. Of note, 50% of the participants enrolled in the Fengler et al. 2017 study had previously undergone treatment with rfRDN and did not display a sufficient treatment response. The treatment populations differed on important baseline characteristics, and as this study was small (total n=50), interpreting efficacy in these populations would be challenging. Further, 19 of 31 participants did not show an adequate treatment response (35% of first-treated and 40% of re-treated). Regarding safety, during the procedure, 10% of participants experienced transient vascular spasm, 4% required transient noninvasive ventilation, and one individual required single-sided ablation to facilitate balloon catheter placement. No deaths and no new renal artery stenosis were reported. Daemen et al. (2019) reported findings for 24-hour and office BP outcomes at 12 months from a prospective, nonrandomized study of 96 participants treated with an uRDN device. The results indicated statistical reductions at 12 months in 24-hour systolic and diastolic ABPM and OBP. One participant experienced a hypertensive crisis requiring hospitalization, and five participants experienced minor groin complications. One individual died (presumed myocardial infarction), but the authors indicated that this participant had a history of coronary heart disease, thus the event was not thought to be caused by the procedure. No renal artery stenosis was observed.

Whitbourn et al. (2015) conducted a prospective, observational, open-label feasibility study of 50 participants treated with rfRDN (Symplicity Spyral). The authors noted statistical reductions in office-based systolic and diastolic BP from baseline at 3, 6, and 12 months and 24-hour systolic and diastolic ABPM reductions at 6 and 12 months. At 6 months, six adverse events were reported: one myocardial infarction, two elevated serum creatinine (>50%), and three vascular complications. At 12 months, three additional adverse events were reported: one myocardial infarction and two elevated serum creatinine values (>50%). There were no deaths or reports of renal artery stenosis.

Evidentiary Limitations 

The majority of the studies are industry sponsored and funded by each device's manufacturer. Only one poor-quality study directly compared uRDN and rfRDN, and the sample sizes in this study were fewer than 50 participants per arm. Highly selective participant selection across all trials was observed, with a range of 35% (REQUIRE) to 6% (RADIOSOUND-HTM) of individuals who were screened ultimately being enrolled. Overall, sample sizes in the RDN studies were relatively small when considering how common hypertension is as a condition. Strict inclusion/exclusion criteria may limit generalizability (particularly in older individuals), those from different racial and ethnic groups, and those with multiple comorbidities. 

Primary endpoints in each trial were assessed at a relatively short follow-up (2 to 3 months), and most trials did not maintain blinding long-term. Additionally, variations in AHM standardization, differences in medication treatment load, and measures of medication treatment adherence could contribute to the observed variability of the outcome measures across studies.

These studies were not powered to assess long-term health outcomes. Because hypertension treatment is often lifelong and BP is a surrogate outcome, long-term follow-up and demonstration of improved health outcomes are very important. These studies only measured BP and were not designed to capture long-term health sequalae of hypertension, including preventing hypertension-associated end-organ damage and survival. Because hypertension treatment is usually required for life, long-term studies are needed to demonstrate the durability of treatment and improved health outcomes. 

In addition to the above-identified limitations of the reviewed studies, other evidentiary gaps raised in the literature include participant selection and facility/operator experience, which strongly influence the benefit and utility of RDN. A volume-outcomes association has been demonstrated for most invasive procedures, especially during the earlier stages of technology adoption. Operators should have expertise in renal vascular anatomy (for instance, the presence of accessory or aberrant renal arteries), prompt recognition and management of potential complications, including vascular access complications and renal arterial injuries such as dissection, embolization, or perforation. Careful selection of individuals who may benefit from the procedure through a multidisciplinary team approach that includes a hypertension expert in a specialized center is also an important consideration. Such centers should be able to thoroughly evaluate individuals for secondary causes of hypertension. For instance, up to 10% of participants with stage 1 hypertension and between 20% and 25% with stage 2 or resistant hypertension may have evidence of primary aldosteronism (O'Malley et al., 2023). Some individuals may be misdiagnosed (white coat hypertension, high-salt diet, incorrect BP recording technique). Such centers should also be proficient in the judicious pursuit and interpretation of 24-hour ambulatory BP recordings. Assessment for medication adherence through direct questioning and biochemical screening of serum or urinary drug levels (now available in clinical practice) may be necessary. The assessment and treatment of obstructive sleep apnea is another important factor in addressing apparent treatment-resistant hypertension.

SUMMARY

Collectively, these trials demonstrate that second-generation RDN devices are effective for lowering BP in some, but not all, persons with hypertension, achieving modest reductions comparable to those seen with a single antihypertensive medication. The findings also suggest that RDN is safe, with minimal impact on renal function and consistent efficacy across the studied populations and device types. The benefit of these devices is mainly that they function as an “always on" treatment, which is useful for individuals who may have difficulty adhering to or have contraindications to medical treatment; they do not appear to produce side effects that challenge some medication options.

However, the current evidence is inadequate to fully assess which individuals, practitioner, or facility characteristics predict the most successful health outcomes from RDN. Studies of RDN with an active comparator or larger studies of head-to-head comparisons of RDN devices are needed. While one study found a greater BP reduction for uRDN at 3 months compared to rfRDN (with and without branch artery treatment), the differences were no longer evident at 12 months (Fengler et al., 2023, 2019). Additionally, studies enrolling individuals that better reflect real-world individuals are needed to understand the risks and benefits of this technology when combined with other AHTs. One recent analysis suggests that BP reductions after RDN may be dependent on baseline BP, with higher baseline BPs associated with larger BP reductions. Even so, many individuals with severe, resistant hypertension will continue to require multiple AHTs after RDN. For context, adding one AHT can similarly achieve 4 to 6 mm Hg in ambulatory SBP (equivalent to 7–10 mm Hg reduction in OSBP) with RDN. In particular, spironolactone reduced 12-week averaged home SBP by 8.7 mm Hg compared with placebo among individuals with treatment-resistant HTN. 

Future studies are needed to better define the most appropriate population(s) for RDN (e.g., resistant HTN, isolated systolic HTN, early HTN, high lifetime cardiovascular risk) and whether this BP reduction translates into improvements in surrogate markers (such as left ventricular hypertrophy) or hard clinical endpoints (e.g., major adverse cardiovascular events, stroke) as has been noted with studies of antihypertensive medications. Individuals with features of sympathetic overactivity, including combined systolic-diastolic hypertension, orthostatic hypertension, and elevated renin levels, may benefit more from RDN. In contrast, the procedure may have less effectiveness for the treatment of isolated systolic hypertension because the mechanism of hypertension is mainly driven by aortic stiffening rather than sympathetic overactivity. Additionally, given that a significant proportion of hypertensive individuals do not respond to RDN, statistical models are needed to define the predictors of treatment response. The observed limitations within the evidence base does not allow firm conclusions to be drawn on the net overall impact of renal denervation​ on health outcomes for individuals with uncontrolled hypertension.

PROFESSIONAL SOCIETY GUIDELINES

In 2023, the European Society of Hypertension (ESH) published guidelines for the management of arterial hypertension. In the guidelines, the EHS describes the mixed results of trials and meta-analyses as described above, the lacking predictive factors of success, and the lacking understanding of durability (including the potential for nerve regrowth). These limitations lead to the following class 2 recommendations:

• “Renal denervation can be considered as a treatment option in patients with an eGFR of > 40 ml/min/1.73m² who have uncontrolled blood pressure​ despite the use of anti-hypertensive drug combination therapy or if drug treatment elicits serious side effects. (Class of Recommendation: II, Level of Evidence: B)
• Renal denervation can be considered as an additional treatment option in patients with resistant hypertension if eGFR is > 40 ml/min/1.73m². (Class of Recommendation: II, Level of Evidence: B)"

In 2024, the European Society of Cardiology (ESC) published guidelines for the management of elevated BP and hypertension. As the EHS did,  the ESC notes the mixed results of available clinical studies. The Society goes on to note six different major limitations, including but not limited to (1) RDN is an invasive procedure providing a modest reduction in BP, (2) the reduction (or lack thereof) in cardiovascular events is yet to be established, (3) the lacking characterization of long-term safety data, and (4) no direct evidence to understand if denervation was successful or if reinnervation may occur (or has occurred). Several class 2b and class 3 recommendations were provided, demonstrating a lower level of recommendation compared to other hypertension treatment options.

The AHA (2024) published a scientific statement on renal denervation for the treatment of hypertension. The statement provides a review of clinical trials, while also providing specific selection criteria for RDN use. However, the AHA also notes that there are no clinical features aside from BP that is associated with a greater response to RDN. Overall, the AHA writes, “Given the novelty of a device-based approach to the treatment of hypertension and the heterogeneity in existing study outcomes, thoughtful and informed discussions between experienced clinicians and patients who are candidates for this therapy are paramount." Additional conclusions include the need for further research into the long-term efficacy, reinnervation of renal nerves, increased understanding of predictive factors of success, and cost-effectiveness.

In 2025, the AHA along with 12 other organizations provided an update to their 2017 guidelines on the management of high BP. In this update, the following Class 2b recommendation is made, representing a weak recommendation:

• “In carefully selected patients with systolic and diastolic hypertension (office SBP 140-180 mm Hg and DBP > mm Hg) and eGFR > 40 mL/min/1.73 m² who have resistant hypertension despite optimal treatment, or intolerable side effects to additional antihypertensive drug therapy, renal denervation (RDN) may be reasonable as an adjunct treatment to BP medications and lifestyle medication to reduce BP."

TRANSURETHRAL RENAL DENERVATION

Although not FDA approved at this time, the Verve Medical Phoenix system (Verve Medical, Paradise Valley AZ) is the subject of ongoing research. The system uses a radiofrequency generator and an ablation device with four electrodes. A guidewire is advanced under fluoroscopy via a rigid cystoscope, followed by a sheath that allows for placement of the ablation device beyond the ureteropelvic junction. Once activated, the temperature rises to 60 degrees Celsius for two minutes, which is repeated in the other kidney as well.

The Verve Medical Phoenix system has only been researched in a small feasibility trial. Hering et al. (2022) published the primary results of the feasibility trial, which occurred in Tbilisi, Georgia. Eighteen individuals were included in the single-arm, open-label trial. Inclusion criteria consisted of those with a mean BP measured by 24-hour ambulatory BP monitoring of 135-170/<105 mm Hg, up to three antihypertensive medications (not quite meeting the definition of resistant hypertension), and ages between 18 and 70. The primary effectiveness endpoint was the mean change in daytime systolic BP from baseline to 2 months. After 2 months, when necessary, medication intensification was permitted. The cohort consisted of seven women and 11 men, who took a mean of 2.7 antihypertensives, 10 of whom taking a diuretic. While mean baseline daytime systolic BP was 148.4 mm Hg, 2 months postprocedure, mean daytime systolic BP was 129 mm Hg. Reductions in 24-hr systolic BP, nighttime systolic BP, as well as all corresponding diastolic BP measurements were observed at 2 months compared to baseline.

In 2024, Weber et al. published the 12-month results of the feasibility study. Seventeen individuals completed the follow-up, while one individual was lost to follow-up. After the 2-month primary endpoint, individuals were allowed to have medication intensification; however, at 12-months, there was no reporting of medication intensification, and medication reduction only occurred in two individuals. Mean ambulatory systolic and diastolic BP measurements at 12 months were similar to those taken at 2 months. Overall, 13 of 17 individuals had a reduction in 24-hour systolic BP greater than 10 mm Hg.

While the results of the feasibility trial are encouraging, the results of the Renal Pelvic Denervation Pilot Trial (NCT07005050) will provide additional information regarding the safety and effectiveness of transurethral renal denervation. Currently, there is insufficient reliable evidence to permit conclusions regarding the safety and effectiveness of this technique. Larger, sham-controlled, randomized controlled trials, that demonstrate a long-lasting impact on clinical endpoints are needed, as has been demonstrated with antihypertensive medications. Additionally, if traditional radiofrequency and/or ultrasound renal denervation for uncontrolled hypertension is established as routine clinical care, head-to-head comparison trials may also be warranted.