Catheter-based renal denervation emerged as a powerful, innovative intervention, with the promise to treat and even cure drug-resistant hypertension. The early uncontrolled/unblinded studies presented impressive results in office blood pressure reductions of 25–30 mmHg (systolic). The early results generated excitement and optimism among investigators, hypertension experts, and interventionists alike. With the publication of Symplicity HTN-3 , however, the optimism faded, the enthusiasm disappeared, and the whole field fell into a hibernation state. Although many opinions have been expressed as to why Symplicity HTN-3 failed, most authorities believe that incomplete or partial denervation was the main reason . The single-tip catheter used in that study was not suitable to achieve circumferential, 4-quadrant ablation in a predictable way. In the intervening time frame since the publication of Symplicity HTN-3, a great deal of research has been performed to help us all better understand the renal micro-neuroanatomy and how to achieve more complete and more effective renal denervation.

In this issue of the journal, Tim A. Fischell et al, publish their results using a new, innovative technique that can provide complete and permanent renal neurolysis and complete renal denervation.

The authors provide an update of their work—both pre-clinical and clinical—on renal denervation, using alcohol infusion in the perivascular/adventitial space. They used a novel catheter-based, three-needle delivery system (Peregrine™) to achieve chemical neurolysis and denervate the renal arteries of adult swine (n = 17) and humans (n = 18) in a first-in-man feasibility study.

In the pre-clinical testing, ethanol was infused bilaterally into the perivascular space, using two doses of ethanol or saline as sham control. Renal parenchymal norepinephrine (NEPI) concentration and safety were the primary endpoints. Data from the first-in-man study (n = 18) tested primarily the feasibility of the technique and safety but also assessed efficacy. The authors report that in the pre-clinical experiments the device was successful in all 20 animals and 40 renal arteries. There was no study-related morbidity or mortality, and the procedure was not limited by any anatomic variations. There were no dissections, perforations, hematomas, thrombus formation, or device-related complications. There was also no change in renal function at 3 months and in renal artery anatomy. There was, however, significant reduction of renal parenchymal NEPI at 3 months by 68% and 88%, with the two doses of 0.3 and 0.6 ml, as compared to controls. Histological examination revealed marked, deep, and circumferential renal nerve injury at depths of up to 13.4 mm from the intimal surface. Moreover, nerve injury appeared permanent, with damage to the perineural sheath that would prevent nerve regeneration. Furthermore, at 3 months there were no worrying signs: no evidence of thrombi, dissection, aneurism, perforation, hematomas, or neointimal proliferation. There was no evidence of inflammation, injury, or fibrosis and no discernible damage to adjacent tissues, including the kidney, adrenals or bowel.

In the first-in-man study, the authors report 100% device success, no complications, a procedure time of 4.3 ± 3 minutes/artery, and minimal or no patient discomfort during treatment. Angiography at 6 months showed no evidence of renal artery stenosis. There was no change in renal function, and there was evidence of a significant improvement in blood pressure control as compared to baseline. The majority of evaluated patients had a reduction in blood pressure from baseline to 6 months while they reduced the number of antihypertensive medications. A number of patients had substantial blood pressuring lowering in the face of substantial reduction of antihypertensive medication. The authors conclude that the procedure is safe and effective and a promising new approach for renal denervation.

I agree with their conclusions. Indeed, the procedure seems to be safe and effective. In fact, this is the only study that reported substantial blood pressure and medication reduction in all participants following renal denervation. It is also the only study that demonstrated fiber neurolysis at a depth up to 13 mm from the lumen, essentially affecting 100% of the perivascular sympathetic fibers. The reduction of NEPI by > 88% also concurs with this impression. From the totality of the renal denervation literature, it becomes apparent that close-to-compete denervation is necessary to achieve optimal blood pressure reduction in suitable patients. This technique therefore merits further investigation.

It is now believed by many that Symplicity HTN-3 failed to meet its primary efficacy endpoint, because it achieved only partial denervation. Below, therefore, I’ll briefly review the current status of the field of renal denervation and try to put things into perspective.

It has been shown for a long time in many animal models that surgical renal denervation with stripping of renal nerves in both the arteries and veins and painting the vessels with phenol works. This surgical technique achieves complete renal denervation with reduction of tissue NEPI by up to 95–99%. Evidence suggests that partial denervation may not work. The kidney has a tremendous ability to compensate in many respects and can overcome partial renal denervation. Unilateral renal denervation or even nephrectomy does not have any effect on blood pressure. There is a potent reno-renal reflex that can compensate for deficiencies of the contralateral kidney. The healthy kidney or part of a kidney can compensate by increasing renin production, vasoconstriction, and sodium retention.

Animal studies that employed catheter-based renal denervation used mostly radiofrequency energy sources to thermally injure or destroy sympathetic fibers through the renal artery wall. The algorithm used, however, allowed penetration only to a depth of 3–4 mm from the artery lumen, thus reaching only a portion of sympathetic fibers. Furthermore, thermal energy delivery needs to be applied in a circumferential 4-quadrant fashion. For these reasons, most animal studies that employed radiofrequency energy delivery demonstrated only partial NEPI reduction in the renal tissue, ranging from 40 to 65% reduction . Recent studies, however, suggest that modification of the application technique to access the distal segment of the renal artery and/or branches may achieve much better denervation because in these locations the fibers are closer to the lumen. Revised ablation devices also allow thermal delivery to deeper locations in the adventitia. Adoption of these techniques may improve the degree of renal denervation with radiofrequency energy source.

In a focused analysis of the Symplicity HTN-3 results, Kanzari et al demonstrated convincingly that partial denervation was the main reason for the failure of the study. In that article, blood pressure response was plotted against the number of lesions employed in both renal arteries. Ten or more lesions in both renal arteries were required to see any sham control subtracted blood pressure effect. Some blood pressure reduction was noted for lesions 11–13 and significant blood pressure reduction for > 14 lesions. For those patients who received < 9 lesions in both renal arteries, there was a net increase in blood pressure. This was true for office blood pressure and blood pressure measured by ambulatory monitoring. Furthermore, the greatest blood pressure reduction was noted only in those patients (N = 19) who received 4-quadrant lesions bilaterally. This analysis, although post hoc, leaves little doubt that the number of ablations and the degree of renal nerve injury are the most important determinants of blood pressure response.

In this context, therefore, how can we explain the “impressive” results of the early studies? The early renal denervation studies in humans were single-arm, uncontrolled studies that demonstrated “impressive” results and created a lot of excitement and enthusiasm in the field. These studies, however, had no controls and could not distinguish between true blood pressure reduction due to renal denervation and blood pressure “response” due to sham effect, Hawthorn effect, and/or regression to the mean. There is no question that several factors contributed to the large reduction of office blood pressure seen in most of the early studies . Most of these single-arm, uncontrolled studies followed the study design of Symplicity HTN-1 and -2 and allowed for “unintentional exaggeration” of baseline blood pressure that made the response larger than its true value. In short, it is difficult to ascertain from these studies how much of the blood pressure response was truly due to renal denervation and how much was due to confounders. The effect seen on ambulatory measurements was much smaller and mostly explained by the procedure or sham effect. The discrepancies between office and ambulatory blood pressure response have been pointed out by us and others . The early Symplicity HTN-1 and -2 had only a limited number of patients who had both office and ambulatory blood pressure measurements, but EnligHTN I had monitoring in all participants. Results were pretty similar between these studies. Most of these early studies presented reductions in office blood pressure in the range of 20–30 mmHg, but blood pressure reduction seen on 24-hour ambulatory monitoring was consistently lower.

The clear message from these and other trials that included both office and ambulatory blood pressure monitoring is that the substantial disparity in the magnitude of blood pressure response between the two measures can be attributed to a large extent to unintentional overestimation of office blood pressure at baseline.