Blog entry by Arif Khwaja
There has been much interest in the last couple of years in the potential of renal denervation to manage resistant hypertension particularly after the SYMPLICITY-HTN-2 study was published in the Lancet several years ago.
There have been a number of studies sponsored by medical device companies including SYMPLICITY and EnligHTN that have been active recently. At last we have the data with the publication of SYMPLICITY HTN-3 in the New England Journal of Medicine.
In this study patients with severe resistant hypertension (all patients were on the maximum doses of at least 3 drugs including a diuretic) were randomly assigned in a 2:1 ratio to undergo renal denervation or a sham procedure. The primary end point was change in office blood pressure (probably a poor surrogate for actual blood pressure) and a secondary endpoint including a safety composite of death, end-stage renal disease, embolic events resulting in end-organ damage, renovascular complications, or hypertensive crisis at 1 month or new renal-artery stenosis. A particular strength of the study was that 24 hour ambulatory BP was measured as well as office BP as a secondary endpoint and an impressive 535 patients were included in the study
The bottom line was that there was NO significant difference in the primary endpoint in both groups. Furthermore whilst blood pressure fell significantly in the denervation group it also fell in the sham group ( the mean (±SD) change in systolic blood pressure at 6 months was −14.13±23.93 mm Hg in the denervation group as compared with −11.74±25.94 mm Hg in the sham-procedure group (P<0.001 for both comparisons of the change from baseline) with no significant differences in office BP between the groups. 24 hour ambulatory BP failed to show a positive impact of denervation either: 6.75±15.11 mm Hg fall in BP in the denervation group and −4.79±17.25 mm Hg. Again this difference was not significant.
There were no safety concerns with denervation. Subgroup analysis did not show any positive impact in any subgroup including 10% of patients with 'renal impairment'
The findings are at odds with previous studies such as the unblinded SYMPLICITY HTN-2 trial which showed significant reductions in blood pressure with denervation.
So how come the difference - well what the study highlights is the importance of blinding in a clinical trial. The positive results from earlier trials were from unblinded studies whilst in this study blood pressure fell significantly in the sham-procedure group as well... the reasons of this aren’t clear but we know that there can be a powerful 'placebo' effect in any clinical study. In medical device studies 'sham' interventions are the equivalent of the placebo or control arm - yet such sham interventions are often not performed for 'ethical' reasons of not wanting to subject patients to a sham-intervention that maybe associated with risk to the patient. Yet SYMPLICITY HTN-3 clearly demonstrates the absolute necessity for having such a sham-intervention group when evaluating the impact of medical devices. The reasons underlying the BP fall in the sham group aren’t clear but we know that being enrolled in any clinical trial often brings improvements even in the control/placebo arm. This may simply be as a result of increased contact with medical professionals resulting in improved patient adherence in medications and general good care.
Enthusiasts for denervation may say that one of the problems is that the radiologist has no 'read out' to tell them whether denervation has been successful and this may account in part for these disappointing results. However ablation catheter used in the SYMPLICITY HTN-3 study was no different from that used in the SYMPLICITY HTN-1 and HTN-2 studies.
My colleague Will McKane who has considerable expertise in denervation tells me that he has anecdotal experience of denervation being a huge success in some patients with resistant hypertension with some patients achieving good control on minimal medication after years of intractable, severe hypertension. Indeed the standard deviations indicate that there can be a very broad response to denervation and so it is possible that a positive impact of denervation in selected patients was ‘hidden’ in the mass of data from the trial. However we don’t have good evidence to support this nor is there anyway of identifying those who may respond.
As was pointed out in an accompanying editorial in NEJM there has been enormous hype around denervation with many claiming it to be a potential ‘cure’ for hypertension. That bubble has now burst and at the moment its difficult to see any significant future for denervation in the management of resistant hypertension