Author + information
- Vivek Y. Reddy, MD and
- Jeffrey W. Olin, DO∗ ()
- Zena and Michael A. Wiener Cardiovascular Institute and Marie-Josée and Henry R. Kravis Center for Cardiovascular Health, Icahn School of Medicine at Mount Sinai, New York, New York
- ↵∗Reprint requests and correspondence:
Dr. Jeffrey W. Olin, Zena and Michael A. Wiener Cardiovascular Institute and Marie-Josée and Henry R. Kravis Center for Cardiovascular Health, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, Box 1033, New York, New York 10029.
There are few areas in medicine that have gained as much prominence as quickly, or have been as widely dismissed as quickly, as catheter-based renal sympathetic denervation (RDN) for the treatment of resistant hypertension. Hypertension is a global public health concern affecting 20% to 30% of the world’s population despite advances in medical treatment and patient education (1). Of these cases, it is estimated that 10% have resistant hypertension—defined as a systolic blood pressure (SBP) of at least 140 mm Hg despite adherence to at least 3 antihypertensive medications, including a diuretic, at maximally tolerated doses. Furthermore, it has been shown that more than 50% of patients with resistant hypertension are nonadherent to their medications (2). Given the aging population and the concurrent obesity epidemic, the prevalence of hypertension is only expected to increase worldwide.
The concept of interfering with the sympathetic nervous system to treat essential hypertension has been around for more than 75 years (3–5). In addition to a nonrandomized surgical series in the 1960s (4), RDN emerged as a possible treatment strategy based on: 1) pre-clinical studies demonstrating a pathological role for increased sympathetic tone in animal models of hypertension (6); 2) early clinical studies demonstrating elevated sympathetic tone in patients with hypertension (7); and 3) interventional clinical studies demonstrating a favorable effect of RDN on blood pressure (BP) in patients with resistant hypertension in both a first-in-man nonrandomized clinical series, SYMPLICITY HTN-1 (8), and an open-label randomized controlled clinical trial, SYMPLICITY HTN-2 (9). Although neither of these studies was placebo-controlled, the magnitude of effect on BP in these patients with resistant hypertension was dramatic: a decrease of systolic/diastolic office BP of 32/12 mm Hg with no significant change in the control group (9). Additional follow-up demonstrated sustained reductions in BP of a similar magnitude over 3 years for both SYMPLICITY HTN-1 and -2 (10,11). Hering et al. (12) showed that RDN was not only effective in reducing BP, but that there was sustained reduction in muscle sympathetic nerve activity for 1 year after RDN.
However, enthusiasm surrounding RDN for resistant hypertension virtually evaporated over night with the publication of the single-blind, randomized, sham-controlled trial, SYMPLICITY HTN-3 (13). Unlike prior studies, patients in SYMPLICITY HTN-3 underwent screening renal angiography prior to randomization, so they were blinded to the treatment assignment. Although there were no significant safety concerns with RDN, there was also no significant change in the primary efficacy endpoint of office systolic BP at 6 months: a mean decrease of 14 ± 24 mm Hg in the RDN and 12 ± 26 mm Hg in the placebo group, for a difference of 2.4 mm Hg (p = 0.26 for a superiority margin of 5 mm Hg) (13). Not surprisingly, there has been much controversy over these results and a number of theories on why SYMPLICITY HTN-3 had negative findings. Yet, despite the negative results in SYMPLICITY HTN-3, there continues to be much interest in renal denervation. A PubMed search revealed that approximately 260 papers on RDN have been published in the first 6 months of 2014 for resistant hypertension, diabetes, proteinuria, heart failure, reduction in left ventricular mass, atrial fibrillation, and a host of other indications.
In this editorial, we review some of these theories, particularly in the context of 2 papers published in this issue of Journal, from which we can glean some important insights (14,15).
There have been several clinical trial design concerns raised about SYMPLICITY HTN-3. First is the concern that the BP may not have been stabilized appropriately prior to randomization. Per protocol, medication changes were not allowed in the 2 weeks before randomization, even though many studies have indicated that up to 8 weeks are required to reach a steady state after introducing new drugs or a change in dosage. However, only 31 patients (5.8%) had medication changes during this period, with no significant between-group difference in office BP at screening visits (14). There is also the important effect of regression to the mean, which can occur whenever inclusion into a trial is based on exceeding a threshold of a clinical marker that naturally fluctuates with time, such as BP. That is, a patient has a better chance of meeting the inclusion criteria on the day when their fluctuating BP is above their own average. But during follow-up, the average BP will tend to return to that individual’s true mean pressure, even in absence of an intervention. However, this regression to the mean phenomenon is likely to be minimized when one employs a 24-h ambulatory blood pressure monitor (ABPM), which averages multiple values obtained over the course of 24 h. In this context, the publication by Bakris et al. (14) examined the ABPM changes between groups in SYMPLICITY HTN-3, a powered secondary efficacy endpoint.
Unlike the ∼10-mm Hg drop in the mean ambulatory systolic BP seen in SYMPLICITY HTN-1 and -2 (Figure 1), there was only a 6.8 ± 15.0-mm Hg drop in the RDN group in SYMPLICITY HTN-3, and this was not statistically different than the 4.8 ± 17.0-mm Hg drop in the sham group (p = 0.98, for a superiority margin of 2 mm Hg). The availability of 24-h BP values also allowed the investigators to examine the difference in BP between daytime and nighttime. Normally, there is a decrease in nocturnal BP, but this normal diurnal BP variation may be absent or even paradoxical (termed a “nondipper” response) in certain hypertensive patients, and this is associated with worsened end-organ damage and clinical outcomes (16). To this point, Bakris et al. (14) now report that in SYMPLICITY HTN-3, there were no significant differences between groups in the daytime systolic ABPM (1.1 mm Hg difference, p = 0.52), nocturnal systolic AMBP (3.3 change, p = 0.06), or the percent of nondippers converting to dippers (21.2% and 15.0% in the denervation and sham groups, respectively, p = 0.30) (14). Although the trend for improved nocturnal ambulatory BP control following RDN is interesting, the overall negative results are directionally consistent with the initially-reported negative conclusion based on the office BP.
The ABPM analysis was also important in the subgroup interpretation of SYMPLICITY HTN-3. Unlike prior RDN studies, which largely included only Caucasian patients, SYMPLICITY HTN-3 was the first to include a significant percentage (∼26%) of black hypertensive patients. The initial subgroup analysis of office BP in SYMPLICITY HTN-3 suggested that RDN was only ineffective in black individuals and would have been effective if these patients had been excluded (13). However, the ABPM data failed to demonstrate any significant interaction with race and BP outcome: there was no significant difference in 24-h systolic BP change between black and nonblack individuals (p value for interaction = 0.643) (14).
Finally, the ABPM analysis is important in elucidating the effect of baseline BP. That is, a high baseline BP has been thought to be a clinical predictor of response to RDN—with a greater response observed in patients with a higher initial BP. Indeed, in the first analysis of the Global Symplicity Registry, which included ∼1,000 patients treated worldwide, the RDN-based 6-month reduction in the office systolic BP was 20 ± 22 mm Hg in those patients with a baseline office systolic BP ≥160 mm Hg, as compared with a 12 ± 25 mm Hg reduction in the full Global Symplicity Registry cohort (17). However, when the SYMPLICITY HTN-3 cohort was divided into tertiles based on the baseline 24-h ABPM, there was no significant difference in the systolic BP between the RDN and sham groups for any of the 3 tertiles (mean reduction of 1.6, 3.4, and 1.2 mm Hg for each tertile; p = NS for each) (18). However, it should be noted that there was a decrease in the SBP of 11.0 mm Hg at 6 months in the sham group. The concept of this study was to enter patients whose BP could not be controlled on a good triple-drug regimen. How, then, was there such a dramatic BP response in the sham group, when these patients were entered because their BP could not be controlled?
Bakris et al. (14) outline several possible reasons why this trial did not meet its primary or secondary endpoints and why BP reduction was not as robust as in previous trials. There did not appear to be a difference in response based on those operators who performed <5 procedures and those who performed ≥5 procedures. There were 111 operators who performed at least 1 procedure (31% performed only 1 procedure). Overall, there were 183 patients treated by physicians who performed ≥5 procedures and 181 patients treated by physicians who performed <5 procedures. There was no difference in the 6-month change in office, ambulatory, or home BPs between these 2 groups. In SYMPLICITY HTN-1 and -2, there were relatively few operators, whereas in SYMPLICITY HTN-3, there were many. Did some centers have a better response to denervation than other centers?
Among the possible explanations for lack of efficacy (lack of validation of medication adherence, Hawthorne effect, poor patient selection, and sympathetic nervous system was not mediating the hypertension), it is the inability to determine that effective renal denervation has occurred that is most disconcerting. To assess the efficacy of renal denervation, Tzafriri et al. (15) evaluated nerve and ganglia distribution as it relates to effective denervation. These investigators conducted a porcine pre-clinical study using a multielectrode radiofrequency ablation catheter. Although this catheter was different than that used in the Symplicity studies, it nonetheless delivered ablation energy in a focal manner. This study demonstrated that the number and density of the nerves and ganglion were the greatest at the proximal portion of the renal artery; however, they are farther from the lumen of the artery. In the more distal portions of the renal artery, the nerves are less abundant but are closer to the lumen, making ablation easier and more effective. These investigators have shown that at the ostium, an ablation up to 5 mm will affect <38% of the nerves.
Furthermore, treatment efficacy, as defined by a decrease in renal cortical norepinephrine (NEPI) levels, was only observed in 1 of 8 treated vessels, where ablation involved all 4 circumferential quadrants, reached a depth of 9.1 mm, and affected 50% of nerves. There was a correlation between the kidney NEPI and the percentage of effectively ablated nerves. NEPI levels remained at baseline for treatments that affected ≤20% of nerves in the treated segment (15).
These findings have important implications when planning future clinical trials and highlight the potentially significant influence of procedural technique and technology on RDN outcomes. Indeed, there are 3 pieces of corroborating data that implicate procedural technique as an important factor in BP outcome. First, it is important to recognize that, unlike many interventional procedures, RDN does not have a validated physiological procedure endpoint. Indeed, the only procedural endpoint variable reported in SYMPLICITY HTN-3 was the number of “notches” on the post-ablation angiogram, signifying energy delivery sufficient to cause spasm of the artery: the majority of patients exhibited 0 (41%), 1 (21%), or 2 (16%) notches only (19). This questionable technical success of the procedure is not terribly surprising given that in SYMPLICITY HTN-3: 1) there was no roll-in phase to the trial to permit operators to develop technical expertise; 2) a total of 364 procedures were performed by 111 operators, who on average had only performed ∼3 procedures; and 3) a recent subanalysis of SYMPLICITY HTN-3 demonstrated that only 25% of patients received 4-quadrant ablation in at least 1 renal artery (19). Second, the SYMPLICITY HTN-3 subanalysis also demonstrated a statistically significant positive correlation between the number of ablation lesions/patient and the BP difference between RDN and sham; that is, more ablation lesions correlated to improved BP response (19). Finally, there was a recent case report of a patient who underwent an uncomplicated RDN procedure and died 12 days later from an unrelated dissection of the ascending aorta, thereby allowing for autopsy examination (20). Interestingly, histopathology of the adventitial nerves around the renal arteries revealed that ablation-related damage did not penetrate beyond 2 mm from the luminal surface, with significant tapering toward the adventitia where the nerve bundles are located. At least in this patient, the RDN procedure was minimally effective in interrupting the perirenal sympathetic fibers. Additionally, there is an emerging body of data indicating that there may be a significant operator-to-operator variability in the technical performance of the procedure.
There are critical questions that need to be answered to better delineate the role of RDN for patients with resistant hypertension:
1. Is there a particular population of patients who are responders and others who are less likely to respond? Direct measures of sympathetic overactivity, such as renal NEPI spillover or muscle sympathetic nerve activity, are impractical for even clinical trial use, much less clinical use. But, perhaps there are clinical factors identifying patients with high sympathetic activity such as nonresponse to aldosterone antagonist use at baseline—as suggested in the subanalyses of SYMPLICITY HTN-3 (13,14). Or, perhaps, one may employ pre-procedural testing such as cardiac baroreflex sensitivity as assessed by phase-rectified signal averaging to identify patients with resistant hypertension most likely to respond to RDN (21).
2. How important is technique and experience when performing RDN? It is important for the sponsors of SYMPLICITY HTN-3 to evaluate and make public any differences in response among different centers. Improved technical aspects of the procedure and the ability to identify effective denervation at the time of the procedure are important aspects that will help to improve outcomes.
3. Perhaps, radiofrequency RDN is not the most effective method to deliver a deeper and a more circumferential spread to achieve better sympathetic nerve destruction. There are many companies working on different methods of RDN, such as ultrasound or local pharmacologic injection of neurotoxic agents.
4. Does the information provided by Tzafriri et al. (15) help to design a clinical trial that would provide more effective denervation? Ultimately, it will only be as a result of additional carefully-conducted clinical trials that we can fully appreciate which hypertensive patients, if any, will benefit from renal sympathetic denervation.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. Olin is a minority stockholder in Northwind. Dr. Reddy has reported that he has no relationships relevant to the contents of this paper to disclose.
- American College of Cardiology Foundation
- Tomaszewski M.,
- White C.,
- Patel P.,
- et al.
- Esler M.D.,
- Böhm M.,
- Sievert H.,
- et al.
- Bakris G.L.,
- Townsend R.R.,
- Liu M.,
- et al.,
- for the SYMPLICITY HTN-3 Investigators
- Tzafriri A.R.,
- Mahfoud F.,
- Keating J.H.,
- et al.
- ↵Mahfoud F. Responder and non-responder to catheter-based renal denervation in the real world–insights from the Global Symplicity Registry. Paper presented at: EuroPCR; May 22, 2014; Paris, France.
- ↵Kandzari DE. Symplicity HTN-3: insights from the subgroup analysis. Paper presented at: EuroPCR; May 22, 2014; Paris, France.
- Goldschmeding R.,
- Vink A.,
- Weggemans C.,
- et al.
- Zuern C.S.,
- Eick C.,
- Rizas K.D.,
- et al.