Interventional procedure overview of percutaneous transluminal renal sympathetic denervation for resistant hypertension

Analysis

Follow-up issues: Twelve studies provided information on withdrawals. SYMPLICITY HTN-3 2014 recorded 14 (3.8%) withdrawals from the RDN group and 2 (1.2%) from the control arm. In SYMPLICITY HTN-2 2010, there were 3 withdrawals from both the intervention and control arms. DENER-HTN 2015 reported 5 (10%) withdrawals from the RDN group. In Desch 2015, 6 patients (17%) withdrew from the RDN and 2 (5.55%) from the sham group. Prague-15 recorded 7 (13.7%) and 31 (62%) withdrawals from the RDN and control groups, respectively. Three studies reported no withdrawals. SYMPATHY recorded 8 withdrawals (5.8%) (5 in the RDN and 3 in the usual care group).

Study design issues: This study evaluated the short- and long-term effects of RDN in individuals with resistant hypertension on clinical end points, including fatal and non-fatal cardiovascular events, all-cause mortality, hospital admissions, quality of life, BP control, left ventricular hypertrophy, cardiovascular and metabolic profile and kidney function, as well as the potential adverse events related to the procedure.

Primary outcomes included i) fatal and non-fatal cardiovascular events, including but not limited to myocardial infarction, cerebrovascular accidents, and congestive heart failure; 2) all-cause mortality; 3) any hospitalisation and duration of hospital stay (if long-term data are available); 4) quality of life (assessed using validated scales or any other instrument as reported by authors).

Refractory or resistant hypertension was defined by the presence of a clinic BP above target (higher than 140/90 mmHg, or higher than 130/80 mmHg in individuals with type 2 diabetes mellitus), despite the concomitant use of 3 or more antihypertensive drugs of different classes, including a diuretic.

Two authors independently screened and selected eligible studies, carried out data extraction, assessed the quality of each study using the 'risk of bias' assessment tool.

Study population issues: This review included 15 eligible studies and 25 ongoing trials (115 articles). Of these 15 studies, 12 studies were included in quantitative synthesis (meta-analysis). In 4 studies, RDN was compared with sham procedure; in the remaining studies, RDN was tested against standard or intensified antihypertensive therapy. Most studies had unclear or high risk of bias for allocation concealment and blinding.

Of the 15 studies, RDN was done with the electrode radiofrequency Symplicity catheter system in 11 studies. Ablation was done with an off-the-shelf saline-irrigated radiofrequency catheter in 1 study (RELIEF 2012). In 2 studies (INSPIRED and SYMPATHY), ablation was made with the EnligHTN™ multi-electrode denervation system. In 1 study (Franzen 2012), details of the denervation procedure were not provided.

Other issues: The main limitation was represented by the data obtainable from the included studies. Studies were mainly focussed on small populations and short treatment periods. As a result, most trials were not adequately powered to capture exhaustive information on hard, patient-centred outcomes, such as fatal or non-fatal cardiovascular events. More studies that look at factors important to patients such as quality of life are needed. Studies that last longer and have more participants are needed to find out if denervation can lower BP. Moreover, use of multiple catheter systems could potentially contribute to the heterogeneity observed in the analysis.

Analysis

Study design issues: This study estimated the occurrence of renal artery adverse events following denervation with common radiofrequency systems to determine whether radiofrequency ablation increases the risk of renal artery stenosis in the uncontrolled hypertensive population.

Other issues: Clinical trials and registries did not always mandate renal artery imaging of asymptomatic patients and therefore renal artery abnormalities after the radiofrequency RDN procedure might be missed. Subclinical weakening of the renal artery wall might not become clinically manifest for several months or years, so the current estimated rates could change as trials with longer follow up are reported. However, authors conducted a separate meta-analysis only including trials with ≥12 months of follow-up that resulted in a similar result. Authors also gained consistent results across methodological sensitivity analyses.

Analysis

Follow-up issues: One patient in the RFM-RDN and 2 patients in the RFB-RDN group did not attend follow up. In total, 117 patients were available for analysis.

Study design issues: This single-blind, 3-arm randomised trial investigated the effects of ultrasound-based or additional side branch ablation in patients with large renal arteries and compared them with radiofrequency ablation of the main renal artery. The primary end point was change in systolic daytime BP on ABPM at 3 months. Key secondary end points were rate of responders, change in 24-hour systolic ABPM and diastolic BP changes.

Patients were randomly assigned in a 1:1:1 ratio using a time-based nonrestricted computer algorithm. Patients were blinded to the assigned treatment. Authors assumed a change of 12 mmHg in systolic daytime BP on ABPM after 3 months in the USM-RDN and RFB-RDN groups and 6 mmHg in the RFM-RDN group, and a SD of 11 mmHg, as well. To achieve a power of 80% at a 2-sided α-level of 0.05, a sample size of 114 patients was required. To compensate for a potential loss to follow-up, enrollment of 120 patients was planned for the entire cohort. Per protocol analysis was used.

All procedures were done by experienced interventional cardiologists with experience in RDN using all 3 treatment strategies.

Study population issues: Clinical baseline characteristics and medication were well balanced between the groups, except for a numerically different prescription rate of α-blockers and centrally acting sympathicolytics that did not reach statistical significance.

Other issues: The total number of patients was limited, especially considering a 3-arm approach, and the results warrant confirmation in a larger, multicentre trial. Nevertheless, according to power analysis and observed outcomes, the study was adequately powered to assess the primary end point. This trial was carried out in a single-centre and the follow-up period was short (3 months). This trial included patients with larger renal arteries only, based on the assumption that sympathetic fibres are a greater distance from the lumen than in smaller arteries, and so RFB-RDN or higher penetration depth would be more relevant. Therefore, results might have differed in patients with smaller renal artery diameters.

Analysis

Follow-up issues: Of the 143 patients (72 in the RDN group and 71 in the sham control group), all but 1 patient completed the 3-month follow up (1 patient in the sham control group withdrew from the study). Patients were assessed at day 7 after the procedure and then months 1, 2 and 3. Valid ambulatory BP monitoring data at 3 months were available for 69 patients in the renal denervation group and 67 patients in the sham control group.

Study design issues: The sham-controlled REnal denervation on Quality of 24-hr BP control by Ultrasound In Resistant hypertension (REQUIRE) trial assessed the BP lowering efficacy of renal denervation in treated patients with resistant hypertension.

The primary endpoint was the between-group difference in change in 24-hour ambulatory systolic BP from baseline at 3 months. Secondary endpoints were change in daytime and nighttime ambulatory SBP from baseline at 3 months, change in 24-hour, daytime and nighttime ambulatory diastolic BP from baseline at 3 months, and change in seated office SBP and DBP from baseline at 3 months.

Patients were randomised in a 1:1 ratio to have renal denervation using the Paradise^TM Renal Denervation System (ReCor Medical Inc., Palo Alto, CA, USA) or to a sham procedure (renal angiogram only). Randomisation was done using a web-based randomisation tool and was stratified by country (South Korea or Japan), study site, and baseline 24-hour ambulatory SBP (140 to <160 mmHg or ≥160 mmHg). Subjects remained blinded to treatment allocation until 6 months after the procedure. All physicians and study coordinators, including those who interacted with patients, were aware of treatment allocation, but BP assessments were done by study personnel who were unaware of treatment allocation.

Sample size calculation was done based on the assumption that the reduction in 24-hour ambulatory SBP would be 6 mmHg greater in the RDN group than in the sham control group (SD 12 mmHg), it was calculated that the number of patients required to detect a difference between the renal denervation and the sham control groups with 80% power and a 2-sided significance level of 5% was 128 (64 per group). Allowing for a 10% dropout rate over the first 3 months after the procedure, the target sample size was 140 (70 per group).

Study population issues: Demographic and clinical characteristics of the 136 patients at baseline are detailed, below:

Other issues: There were several limitations. There was no standardisation of antihypertensive medications or objective measurement of medication adherence using blood or urine. The nature of the intervention meant that it was not possible to conduct a double-blind study where medical personnel were unaware of treatment group allocation and authors did not prohibit unblinded physicians from participating in follow-up care. There was also no assessment of blinding conducted to determine whether or not the blinding was maintained. There were significant seasonal variation of the temperature and BPs in Japan - morning BP increased in the winter, while the nighttime BP increase in the summer. There was unexpected BP reduction in the sham control group, highlighting study design issues.

Analysis

Follow-up issues: Patients completed a masking questionnaire at discharge and at the 2-month follow up.

Study design issues: This adequately powered, sham-controlled, randomised trial reported the primary efficacy and safety results of ultrasound renal denervation in the TRIO cohort of patients with more severe hypertension resistant to 3 or more antihypertensive medications (of patients with combined systolic–diastolic hypertension resistant to a fixed-dose, single-pill, triple combination antihypertensive therapy).

Intention-to-treat analysis was used. The primary efficacy endpoint was the change in daytime ambulatory systolic blood pressure from baseline to 2 months. Secondary efficacy endpoints specified for hierarchical testing at 2 months were change in 24-hour ambulatory systolic and diastolic blood pressures, nighttime ambulatory systolic and diastolic BP, and daytime ambulatory diastolic BP. Prespecified major adverse events were all-cause mortality, renal failure, an embolic event, renal artery or vascular complications needing intervention, or hypertensive crisis within 30 days of the study procedure, and new onset renal artery stenosis greater than 70% within 6 months of the study procedure.

Resistant hypertension was defined as seated office BP of at least 140 mmHg systolic and 90 mm Hg diastolic despite a stable regimen of three or more antihypertensive medications including a diuretic.

Sample size calculation showed that a sample of 128 participants would yield 80% power to detect a 6-mmHg difference in change in daytime ambulatory systolic BP at 2 months between the RDN and sham groups (common standard deviation 12 mmHg, 2-sided type I error rate of 5%). To account for up to 10% missing observations, authors initially planned to randomly assign 146 participants. However, the decision was made to stop enrolment on May 8, 2020, after random assignment of 134 patients with evaluable follow up at 2 months because of the COVID-19 pandemic constraining further recruitment. The decision was consistent with guidance from the US Food and Drug Administration.

Eligible participants were randomly assigned (1:1) to receive ultrasound renal denervation or a sham procedure. The randomisation sequence was generated by computer and stratified by centre using randomised blocks of 4 or 6 and permutation of treatments within each block. Authors randomly assigned patients with resistant hypertension confirmed by ambulatory BP after adjusting their antihypertensive treatment to a single-pill, fixed-dose, triple combination consistent with current guidelines. By reducing pill burden, a high adherence to the standardised treatment was achieved at baseline in both groups. To maintain masking, participants were sedated and wore headphones and eye covers. Patients and clinicians involved in follow-up care were masked to treatment allocation for 6 months after random assignment.

Of the 53 study centres, 35 centres with 40 different interventionalists had patients assigned to the renal denervation group; each interventionalist did a mean of two (range 1 to 6) renal denervation procedures. Circumferential renal denervation treatment was planned based on the pre-procedural imaging.

Study population issues: Baseline characteristics were similar across both groups.

Other issues: Despite all efforts to reduce overall variability, there was still between-patient variation in the response to RDN, some of which might be attributed to variable medication adherence, variable renal nerve ablation (even though the uniform use of circumferential ablations and treatment of accessory renal arteries), or differing contributions of renal nerve signally to hypertension perpetuation.

Analysis

Follow-up issues: Patients were followed up at 6 months, 1 year, 2 years and 3 years. Of the 2,237 enrolled patients who had the procedure, 1,742 patients were eligible for follow up at 3 years. Of the enrolled population, 1,734 patients had office BP measurements available at 6 months, 1,654 at 1 year, 1,258 at 2 years, and 872 at 3 years.

Study design issues: This prospective, open-label, single-arm, observational registry assessed the long-term effectiveness, safety, and effects on renal function in the Global SYMPLICITY Registry up to 3 years after RDN. The primary objective was to assess procedural and long-term safety of RDN in a real-world setting.

Severe treatment-resistant hypertension was defined as office SBP ≥160 mmHg and 24-hour ambulatory BP ≥135 mmHg, despite prescription of ≥3 antihypertensive medications, while less severe hypertension was defined as office SBP and diastolic BP 150 to 180mmHg and ≥90mmHg, respectively, and 24-hour ambulatory SBP 140 to 170 mmHg.

Study population issues: Of the 2,237 patients, 21% had a history of CKD (eGFR <60mL/min/1.73 m2), 38% had Type 2 diabetes mellitus, and nearly half had a history of cardiac disease. At baseline, patients were prescribed 4.5±1.4 antihypertensive medication classes, which in most patients included an angiotensin receptor blocker or ACE inhibitor, a calcium channel blocker, a diuretic, and a beta-blocker.

Other issues: The Global SYMPLICITY Registry is a single-arm registry and as such did not involve control groups to compare outcomes. There was no way to rule out a Hawthorne/placebo effect, which could be caused by participation and care during the study. Comparison of eGFR measurements between patients with and without medication changes was limited since reported medication changes were not verified with medication adherence testing. The device (first-generation, single-electrode SYMPLICITY Flex RDN catheter system) might have made it more difficult to achieve a pattern of 4-quadrant ablations than the current SYMPLICITY Spyral catheter technology, especially within the Global SYMPLICITY Registry study design that did not encourage more treatment ablations or allow for treatment in the renal artery side branches or accessories.

Analysis

Follow-up issues: Follow up was recommended at 2 to 6 weeks, 3 months, 6 months, and on a yearly basis thereafter. Suggested follow-up documentation included office BP, ambulatory BP, renal function, antihypertensive treatment, and long-term safety. To ensure adherence to drug therapy, patients were encouraged to keep a diary.

Study design issues: The Austrian Society of Hypertension created the Austrian Transcatheter RENal Denervation (TREND) Registry in 2011, with an emphasis on ambulatory BP monitoring to monitor safety and efficacy of all RDN procedures performed in Austria. This was the first analysis of the data gathered by the Austrian TREND Registry, reporting efficacy and safety of RDN with respect to office and ambulatory BP in a real-life setting.

Authors did not document quality of ambulatory BP measurements in the registry. Patient management, choice of drug therapy, device selection for RDN as well as vascular access site remained at the discretion of each individual centre.

The Austrian Society of Hypertension suggested RDN for patients on multiple drug treatment, with a mean 24-hour BP >145/90 mmHg, equivalent to an office BP of 160/100 mmHg. Based on this threshold, patients were divided into 2 groups: group A consisted of all patients with a mean baseline 24-hour BP >145/90 mmHg, and all remaining patients were summarised in group B. Responders were defined as follows: office BP responders had a reduction of at least 10 mmHg of office systolic BP after 6 months. Ambulatory BP responders had a 24-h ambulatory BP reduction of at least 5 mmHg after 6 months.

Study population issues: At baseline, patients were on antihypertensive treatment for a median of 10 years (IQR 7 to 15; n=128). Average office BP was 170±16/94±14 mmHg; average 24-hour ambulatory BP was 151±18/89±14 mmHg (n=359). In total, 98% of patients had a systolic office BP >140 mmHg and 91% a systolic 24-h ABP >130 mmHg, respectively. Most prevalent comorbidities were coronary artery disease (37%), diabetes mellitus (36%) and cerebrovascular disease (12%).

Patients in group A were statistically significantly younger, had a higher BMI, and received more antihypertensive medications than patients in group B. Mean 24-hour BP in group A was 159/95 mmHg compared with 132/77 mmHg in group B. This difference was statistically significant (p<0.001). Furthermore, the average office BP in group A was statistically significantly higher, but the difference was smaller (173/96 mmHg compared with 166/90 mmHg). There were no statistically significant differences in comorbidities. Procedural details were available for 279 patients (69%). Procedural details were available for 279 patients (69%). Antihypertensive therapy was paused during the procedure in 44% of cases.

Other issues: 6-month ABPM data was availability in 59% of patients, data needs be interpreted with some caution as selection bias might have occurred. Since regression to the mean phenomenon and the regression of the white coat effect may lower BP readings at subsequent follow-up visits, the data might over-estimate especially office BP reductions. Drug prescriptions and changes of medications were documented in the registry, however, urine analysis or pill count for proving accurate drug intake was not available.

Analysis

Follow-up issues: Clinical follow-up was available in 90% of patients, with a mean duration of office BP follow up of 11 months.

Study design issues: This study reported the UK experience with RDN for treatment-resistant hypertension. It examined the nature of the BP response seen on ambulatory monitoring and the impact of drug changes post denervation on the results. Finally, this study examined the interaction of RDN with the use of aldosterone antagonists.

'Responders' to RDN were defined as a reduction in office systolic BP of ≥10 mmHg and reduction in daytime ambulatory systolic BP fall of ≥5 mmHg from baseline to follow up. Absence of normal nocturnal dipping profile on pre-procedural ambulatory BP was defined as a fall in nighttime systolic ambulatory BP of <10%.

Study population issues: Of the 253 patients, 88% were Caucasian and 26.5% had diabetes. Eighty-six percent of patients were seen in a dedicated hypertension clinic with each patient being reviewed by an average of 1.6 hypertension specialists. These included cardiologists, nephrologists, clinical pharmacologists and endocrinologists.

Fifty-eight percent of the cohort had loss of normal nocturnal dipping on ambulatory BP. The median number of antihypertensive drugs prescribed before RDN was 5.0 including 96 % ACEi/ARB; 86 % thiazide or a loop diuretic and 55 % aldosterone antagonist prescription at the time of denervation. The mean number of cases performed per centre within this registry was 15 (SD 6.7).

Other issues: This study was limited by the design (an open-label retrospective registry). However, authors stated that data quality appeared good, as supported by the relatively high frequency of reporting of ambulatory BP results and the close correlation between office BP and ambulatory BP results. Results also appeared consistent across 18 sites. This study did not mandate measures of adherence to prescribed medications and therefore variable levels of compliance pre- and post-procedure could have had a confounding impact on results.

Analysis

Follow-up issues: Of the 311 patients who had RDN, 14 patients (4.5%) were lost to follow up or had missing 3 months BP values and 1 patient (0.3%) died before reaching the 3-month follow up. In total, 296 patients (95.2%) were available for analysis.

Study design issues: This retrospective single-centre registry study investigated the effect of BP reduction after RDN on long-term cardiovascular outcome in patients with resistant hypertension. Clinical events were assessed in patients from previous RDN trials and clinical routine at the study centre, some patients (exact number was unknown) were included in Fengler (2021) and Pisano (2021).

Clinical outcome was assessed using a standardised questionnaire by a single investigator, who was masked to BP outcome. If contacting patients was unsuccessful, or if necessary to complete clinical event assessment, patient's last treating general practitioners were contacted. In addition, hospital database was searched for clinical events for every individual patient. In all patients, antihypertensive drug treatment was kept stable until the 6-month follow up was reached unless indicated otherwise.

BP response was defined as reduction of ≥5 mmHg in 24-hour average systolic BP on ABPM between baseline and 3 months.

Major adverse cardiovascular event was defined as a composite of cardiovascular death, ischemic stroke or intracranial bleeding, acute myocardial infarction, critical limb ischemia as well as acute renal failure. The ischemic events end point was defined as a composite of ischemic stroke, acute myocardial infarction, peripheral artery disease requiring intervention and critical limb ischemia.

To assess the effect of BP reduction, clinical events were compared between BP responders and non-responders. In a second step, a postulated proportional relationship between BP reduction and clinical events was tested.

Study population issues: At baseline, responders had higher systolic and diastolic ABPM values as well as a lower rate of isolated systolic hypertension. Baseline medication and number of antihypertensive drug classes were balanced between responders and non-responders.

Other issues: There were several limitations. First, this retrospective single-centre registry had its limitations such as selection bias, and an underreporting of clinical events during the long-term follow up. Second, drug adherence testing for the patients enrolled was not provided. Therefore, part of the observed effects might also be attributed to alterations in antihypertensive drug intake during follow up, even though this was unlikely. Third, because of the study design and the lack of a control group, it was impossible to separately analyse effects of RDN from effects by BP reduction in general, but the proportional association of BP reduction within the immediate timeframe of RDN on long-term outcomes suggested an at least partial effect. Fourth, the composite end point (major adverse cardiovascular events) herein differed from other, larger-scaled cardiovascular outcome trials as it was a concession to the smaller sample size available. Effects of RDN on hard clinical end points should be tested in larger-scaled analyses in the future. Lastly, the relatively small number of events and patients included gave this study only a hypothesis generating character and all findings warrant confirmation in larger, prospectively designed trials.