- Email: [email protected]
Illusions of truths in the Symplicity HTN-3 trial: generic design strengths but neuroscience failings
Journal of the American Society of Hypertension 8(8) (2014) 593–598
Review Article
Illusions of truths in the Symplicity HTN-3 trial: generic design strengths but neuroscience failings Murray Esler, MBBS, PhD, FRACP* Human Neurotransmitter Laboratory, Baker IDI Heart and Diabetes Institute, Melbourne, Australia Manuscript received and accepted June 4, 2014
Abstract The Achilles heel in catheter-based studies of renal denervation for severe hypertension is the almost universal failure to apply a confirmatory test for renal denervation. When renal denervation efficacy was assessed, using measurements of the spillover of norepinephrine from the renal sympathetic nerves to plasma, the only test validated to this point, denervation was found to be incomplete and nonuniform between patients. It is probable that the degree of denervation has typically been suboptimal in renal denervation trials. This criticism applies with special force to the Symplicity HTN-3 trial, where the proceduralists, although expert interventional cardiologists, had no prior experience with the renal denervation technique. Their learning curve fell during the trial, a shortcoming accentuated by the fact that one-third of operators performed one procedure only. Recently presented results from the Symplicity HTN-3 trialists confirm that renal denervation was not effectively or consistently achieved in the trial. J Am Soc Hypertens 2014;8(8):593–598. Ó 2014 American Society of Hypertension. All rights reserved. Keywords: Resistant hypertension; renal denervation; sympathetic nervous system; antihypertensive device.
Despite the widespread availability and prescribing of angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, diuretics, and calcium channel blockers in a substantial minority of patients with essential hypertension, perhaps 10%,1,2 target blood pressure (BP) cannot be achieved. In these drug-resistant hypertensives a new strategy was needed, and in fact, devised. This was the development of device-based therapies targeting the sympathetic nervous system, the surgically implanted barostimulator
The author is a Senior Director of the Baker IDI Heart and Diabetes Institute Melbourne, and Conjoint Professor of Medicine Monash University, Melbourne. His primary funding is in the form of a Senior Principal Research Fellowship from the Australian National Health and Medical Research Council [ID586601]. He discloses research funding and receipt of consultancy fees, travel expenses, and honoraria from Medtronic, which funded the research from which some of the presented results are drawn, but the author holds no shares in the company or patent rights for renal denervation. In the preparation of this article no financial or other support was provided by Medtronic. *Corresponding author: Murray Esler, MBBS, PhD, FRACP, Baker IDI Heart and Diabetes Institute, PO Box 6492, St Kilda Road Central, Melbourne, Vic 8008, Australia. E-mail: [email protected]
device3 and catheter-based renal denervation,4,5 the latter being the subject of this review article. Central to the development of radiofrequency (RF) renal denervation was knowledge of the physiology of the renal sympathetic nerves and their pathophysiology in experimental and human hypertension. In untreated essential hypertensive patients, the application of regional norepinephrine isotope dilution methodology,6 to measure the outward flux of the transmitter from the sympathetic nerves of the kidneys to plasma (‘‘renal norepinephrine spillover’’), demonstrated that a high level of activation of the renal sympathetic outflow was present.7 This renal sympathetic activation is central to hypertension pathogenesis.7–12 In rodents, the renal nerves have been demonstrated to stimulate secretion of renin from the juxtglomerular apparatus, to promote renal tubular reabsorption of sodium, and to cause renal vasoconstriction, all potentially BP elevating responses.13 The renal tubules receive a dense sympathetic innervation, at all tubular levels. In experimental models of hypertension, surgical sectioning of the renal sympathetic nerves lowers BP or delays or prevents the development of hypertension.13 These facts and knowledge of the anatomy of the postganglionic renal sympathetic nerves in their passage to the kidneys provided the intellectual framework for the
1933-1711/$ - see front matter Ó 2014 American Society of Hypertension. All rights reserved. http://dx.doi.org/10.1016/j.jash.2014.06.001
594
M. Esler / Journal of the American Society of Hypertension 8(8) (2014) 593–598
development of catheter-based renal denervation for the treatment of essential hypertension. In humans, the renal sympathetic nerves pass from the sympathetic chain and ganglia to the kidneys via the outer wall of the renal arteries or just outside in perirenal adipose tissue and connective tissue, within reach of RF energy delivered by a catheter in the artery lumen.14 Seven years after the first patient was treated with the Symplicity RF catheter system, these initial trials, their continuation to later specified endpoints,15,16 accompanying resistant hypertension renal denervation registry files17 and trials with other newly engineered renal denervation devices18,19 have established important therapeutic principles: 1. Efferent sympathetic renal denervation can be achieved with luminal delivery of RF and ultrasonic energy. 2. The treatment can be delivered with very minimal procedural complications, these being typical of invasive cardiological procedures in general, and not consequential on RF energy delivery. 3. The mean BP reduction across the trials shows consistency, office systolic BP falling on average by 20– 30 mm Hg at the primary endpoints. Renal function is preserved. The BP reduction is durable, demonstrably persisting for 3 years and beyond. 4. New renal artery stenoses in the field of RF energy delivery are very uncommon.
Among clinical studies of renal denervation, only in the Simplicity HTN-1 trial4 was renal denervation confirmed, by measurements of norepinephrine spillover (release of the transmitter from the renal nerves to plasma; Figure 1). Measurement of regional norepinephrine spillover is well established as a valid test for denervation, having been applied for two decades in the diagnosis of pure autonomic failure,21 a disorder characterized by spontaneous degeneration of postganglionic sympathetic nerve fibers, and denervation of the heart, kidneys, and other organs. The degree of renal denervation achieved in the Symplicity HTN-1 trial was less than expected (on average 47%)4 but did seem to be sufficient, in that the antihypertensive response was very adequate. This led to complacency, in that renal denervation was said to be procedurally easy, to the point of being ‘‘boring’’, a commonly used descriptor. Subsequent analysis by my group confirms that denervation is often very incomplete and surprisingly nonuniform from patient-to-patient (Figure 2). Achieved denervation can be <25%, no doubt inadequate for a full therapeutic effect. So the denervation catheter technique might look ‘‘easy’’
A disappointing aspect of renal denervation therapy, however, is the failure of some patients to experience a drop in arterial pressure, this being reported in different studies to occur in 15–50% of patients. The most commonly offered explanation for this is that in these patients sympathetic nervous system activation is not the primary pathophysiology. But this might not be true, as surgical renal denervation commonly lowers BP in experimental models of hypertension in which sympathetic nervous activity is actually normal.13 Recent observations, described in the following, suggest that the real reason might more commonly be that in ‘‘nonresponders,’’ the denervation procedure has failed technically. Renal denervation in human hypertension is very incomplete, nonuniform from patient to patient, and achieved with difficulty.20
Testing for Achieved Renal Denervation The belief that renal denervation has been achieved in clinical trials of catheter-based renal denervation is almost invariably based on hope and trust, rather than testing. This contrasts with the studies of surgical renal denervation in experimental hypertension, where good experimental design necessitates that the effectiveness of denervation is always confirmed, typically by documenting 90–95% reduction in the kidney content of norepinephrine.
Figure 1. Evidence for catheter-based renal denervation with the Symplicity radiofrequency Arch catheter in patients with resistant hypertension and pigs. Renal sympathetic denervation was quantified in humans with measurements of the fall in renal norepinephrine spillover postprocedure and in pigs using the fall in kidney norepinephrine content, both being validated methods.
M. Esler / Journal of the American Society of Hypertension 8(8) (2014) 593–598
595
arteries. In the Symplicity HTN-3 study,23 we now know that neither of these technical prerequisites was consistently achieved.24
How Much Renal Denervation is Needed? The level of renal sympathetic nerve ablation achieved clinically with the RF denervation procedure, mean 47%,4 contrasts with that achieved surgically in experimental hypertension, 90–95%.13 Which is optimal? If the aim is to ‘‘normalize’’ renal sympathetic activity, 50% ablation of sympathetic nerves in patients who have, on an average, a doubling of renal sympathetic activity7,12 might perhaps be ideal. But as renal denervation lowers BP in some experimental models of hypertension in which sympathetic nervous activity is actually normal,13 normalizing renal sympathetic activity is probably not the desired aim. Greater left-shift in the pressure natriuresis-curve25 and resultant off-loading of more sodium would occur with a higher grade renal denervation than is currently achieved clinically. This could be achieved if delivered energy were to penetrate somewhat beyond 2–3 mm, as some renal sympathetic nerves are more remote from the lumen than this.14,22
Renal Denervation ‘‘Armageddon’’: The Symplicity HTN-3 Trial
Figure 2. The effectiveness of catheter-based radiofrequency renal nerve ablation assessed with renal norepinephrine spillover measurements made before and 30 days after the procedure. The denervation was incomplete, sometimes markedly so, and with pronounced nonuniformity between individual patients. These results contradict the idea that achieving sympathetic denervation is technically ‘‘easy’’. The results suggest that blood pressure nonresponse to renal denervation must commonly be due to very imperfect denervation and not only to the absence of sympathetic nervous activation, which is often invoked as the cause. From unpublished results of the author, Markus Schlaich, Gavin Lambert and Dagmara Hering.
compared with many other interventional cardiological and radiological procedures, but achieving denervation is complex and difficult. Cartoons showing the easy proximity of renal sympathetic nerves to the renal artery lumen were misleading. Older surgical anatomies of the renal nerves14 demonstrated that close to the origin of the renal artery from the aorta, the sympathetic nerves were more distant, but converged on the renal arteries distally, near the renal artery branch point. Contemporary anatomical studies confirm this.22 Clearly ablative energy should be preferentially focused on the distal renal artery, but this is often not done, and on the full circumference of both renal
A challenge to the percutaneous renal denervation treatment of resistant hypertension came with the 9 January 2014 press release concerning the Simplicity HTN-3 trial in drug-resistant hypertension, the pivotal study for US Food and Drug Administration licensure, and in the subsequent New England Journal of Medicine publication on 29 March,23 indicating that the primary efficacy endpoint had not been reached in the trial. A lot was expected of the Symplicity HTN-3 study. Five times larger than the first two Symplicity renal denervation trials and incorporating a blinded sham design, this trial was expected to provide the definitive statement on the value of renal denervation in the treatment of patients with severe hypertension. To many it did—‘‘renal denervation does not work!’’26 The sham design was lauded.26 This trial exemplar had comprehensively exposed the fallacy of imagined renal denervation benefits! How was it possible to argue against the findings of the Symplicity HTN-3 trial (Table 1)?
Symplicity HTN-3 Illusions The perfection of the trial is an illusion (Table 2), only apparent if no discrimination is attempted between its ‘‘form’’ and its ‘‘substance’’. The form of the trial is represented by its generic strengths (large patient enrollment, blinding, sham control), applicable to any BP trial.
596
M. Esler / Journal of the American Society of Hypertension 8(8) (2014) 593–598
Table 1 How is it possible to argue against the findings of the Symplicity HTN-3 Trial? The power of Food and Drug Administration ‘‘branding’’ of a pivotal US study The generic strength of the sham procedure The authority of a New England Journal of Medicine publication .but with very material failings in execution
But this is a neuroscience trial in hypertension, based on a neuroscience theory of causation of severe hypertension (activation of the renal sympathetic outflow), on detailed knowledge of neuroanatomy (of the renal efferent and afferent nerves), and on the engineering of nerve ablative devices. This essence of the trial received inadequate attention and its execution ultimately failed. Among editorialists there was lauding of the sham procedure but neglect of trial neuroscience failings26 Much was amiss with Symplicity HTN-3. At 88, too many centers were recruited for the trial and at 111, too many proceduralists.23 No hands-on experience in renal denervation before the trial was possible in the United States (unlike in the earlier Symplicity trials, where it was mandatory). Experts in their field of interventional cardiology, all participants were novices in the renal denervation procedure. Proctoring was done primarily by company staffers, rather than experienced physicians or renal denervation engineers, in contrast to the earlier Symplicity trials. Energy delivery was not preferentially to the distal renal artery, where it should have been, but inexplicably was more typically to the proximal renal artery; the renal nerves are, in fact, closer to the distal artery.14,22 It is now a matter of record that the denervation procedure fared badly in Symplicity HTN-3.24 Retrospective analysis of stored angiographic and procedural records of all RF energy applications demonstrated that in 74% of patients not even one fully circumferential renal artery application of energy was achieved,24 when it is mandatory that this be achieved bilaterally, making effective nerve ablation impossible.27 Not surprising, the BP fall in these patients was nearly identical to that of the sham group. In contrast, in that minority of patients in whom RF energy was delivered according to the trial protocol in a full circumference in both renal arteries, BP fell very materially, similar in degree to that observes in the earlier Symplicity trials.4,5,24 The obvious rhetorical question is ‘‘what use is a sham design when patients were not effectively or consistently denervated’’? And how could this happen? Presumably Table 2 Form and substance in the Symplicity HTN-3 trial Generic strengths—constituting a plus in any antihypertensive trial (‘‘form’’) Neuroscience failings—in this trial (‘‘substance’’)
because renal denervation was thought and said to be procedurally easy, to the point of being ‘‘boring’’, a commonly used descriptor. Physical aspects of the catheter procedure perhaps looked easy, but achieving denervation was not.4 It should be noted that the field of renal denervation for experimental hypertension is active, in fact energized by the clinical studies. Experimental surgical- and catheterbased denervation for hypertension still works! This was recently well exemplified with catheter-based renal denervation abolishing hypertension in an obese dog model.28 Why should renal denervation, in four mammalian species (rats, dogs, rabbits, and pigs) invariably be antihypertensive,13 but not in the human mammal in Symplicity HTN-3? In the US Pivotal study, a failure to achieve renal denervation came first to mind.20 I believe the Symplicity HTN-3 study failed in its neuroscience execution, and will be judged to be ‘‘on the wrong side of history’’.
Development of a More Generally Accessible Method for Testing Efferent Sympathetic Nerve Ablation? There is a pressing need to incorporate valid testing for renal sympathetic denervation into all clinical trials of the procedure. Growing acceptance has emerged for this important principle, but some of the planned testing for denervation, such as using measurements of heart rate variability, are fanciful. In truth, it is difficult to know how to proceed, as the only validated test for renal denervation, measurement of renal norepinephrine spillover,4,6 has very limited availability and could not be incorporated into an international multicentre trial of the type currently planned by Medtronic, after the technical failure of the Symplicity HTN-3 trial. My own group has commenced exploring a different approach. With ablation of sympathetic nerves, the specific sympathetic nerve proteins, such as tyrosine hydroxylase, will degrade and fragment. Although within the nerves external to the kidneys thermal degradation of nerve proteins might make their identification difficult, sympathetic nerves within the kidneys will escape this process of thermal denaturation. We will now test for the excretion of tyrosine hydroxylase fragments in urine. The undegraded protein is too large to enter urine, but we anticipate tyrosine hydroxylase fragments will. Detection will be with Western blot analysis and enzyme-linked immunosorbent assay testing, presuming the persistence of antibody responsive sites on the tyrosine hydroxylase fragments. It is anticipated that urinary excretion of the marker fragments will be at a peak in the first few days after the denervation procedure, but this will need to be established. Once these principle are established, and at the moment what I provide here is only a promissory note, in a multicentre trial, urine might be collected and sent to a core laboratory for analysis.
M. Esler / Journal of the American Society of Hypertension 8(8) (2014) 593–598
Afferent Renal Nerves and Their Ablation Sensory nerves running from the kidneys to the central nervous system contribute to activation of the sympathetic nervous system in patients with drug-resistant hypertension, through their projection to the hypothalamus.29 Ablation of these nerves during the renal denervation procedure contributes to BP lowering by causing central inhibition of central sympathetic outflow.30 The nociceptive receptors within the kidney, which are responding to a renal injury signal,29 can be targeted with adenosine in a test for afferent nerve ablation which can be conducted in the catheterization laboratory. Adenosine infused into the renal artery increases sympathetic nervous activity and BP. To test for afferent nerve ablation on the catheter day, adenosine is infused before and after the denervation procedure; a reduction in the BP response is indicative of afferent nerve ablation. This test, not fully evaluated as yet, offers prospects of real time evaluation to guide the proceduralist. It is sometimes thought that testing for afferent nerve ablation would also indicate the success or otherwise of efferent, sympathetic ablation. This may not be true as efferent and afferent nerves do not traverse exactly parallel paths. The sympathetic nerves converge on the distal renal artery as they near the kidney but, conversely, the afferent nerves are more distant from the artery as they leave the kidneys, but converge on the renal arteries closer to the aorta.22 The most successful afferent nerve ablation would be expected to result from RF or ultrasound energy delivered through the proximal renal artery, but efferent sympathetic ablation would be less successful at that site.
The Future of Renal Denervation in Treating Severe Hypertension The Symplicity HTN-3 trial has inflicted a ‘‘flesh wound’’ on the renal denervation field, but this is not fatal; the trial is now seen to have failed in its neuroscience execution, not effectively and consistently achieving renal denervation.24 In the pause that this trial created worldwide, it is now time to reflect on what future science is needed (Table 3). Failure to test for renal denervation, as an intrinsic component of all trials but one, represents the Achilles heel of the field. It is probable that in some negative trials, and with nonresponse in some patients, failure to Table 3 The next era of renal denervation neuroscience begins Current levels of achieved denervation are not optimal? Higher energy/multipolar radiofrequency electrodes are needed? Energy should be delivered to both proximal and distal renal artery? Testing for achieved renal denervation is mandatory? Preselection is desirable, but problematic? Renal denervation may not show a ‘‘class effect’’?
597
achieve adequate denervation was responsible. The commonly attributed reason for renal denervation to fail to lower BP, that in such patients sympathetic activity was not increased, has no authenticity in studies where denervation was not tested; failure to denervate is the ever-present confounder. Planned studies, including the second Medtronic US Pivotal Trial in hypertension, will need to incorporate denervation testing. As renal denervation is incomplete and nonuniform between patients, usage of higher RF and ultrasonic energy doses in the future is envisaged (the established safety margin in renal denervation allows this) and multielectrode RF catheters. Energy may need to be delivered to both the aortic and the distal ends of the renal artery, to target both sympathetic and afferent nerves.22 Preselection of patients will remain problematic, until the influence on the responder status of the two determinants, effectiveness of denervation, and hypertension pathophysiology can be discriminated. Will renal denervation show a ‘‘class effect’’, with all energy forms for denervation being equally efficacious? Direct comparison will be needed, as claims and counter claims could be distorted by commercial interest.
References 1. de la Sierra A, Segura J, Banegas JR, Gorostidi M, de la Cruz JJ, Armario P, et al. Clinical features of 8295 patients with resistant hypertension classified on the basis of ambulatory blood pressure monitoring. Hypertension 2011;57:898–902. 2. Persell SD. Prevalence of resistant hypertension in the United States, 2003-2008. Hypertension 2011;57: 1076–80. 3. Bisognano JD, Bakris G, Nadim MK, Sanchez L, Kroon AA, Schafer J, et al. Baroreflex activation therapy lowers blood pressure in patients with resistant hypertension. J Am Coll Cardiol 2011;58:765–73. 4. Krum H, Schlaich MP, Whitbourn R, Sobotka P, Sadowski J, Bartus K, et al. Catheter-based renal sympathetic denervation for resistant hypertension: a multicentre safety and proof-of-principle cohort study. Lancet 2009;373:1275–81. 5. Symplicity HTN-2 Investigators. Renal sympathetic denervation in patients with treatment-resistant hypertension (The Symplicity HTN-2 Trial): a randomized controlled trial. Lancet 2010;376:1903–9. 6. Esler M, Jennings G, Korner P, Blombery P, Sacharias N, Leonard P. Measurement of total and organ-specific norepinephrine kinetics in humans. Am J Physiol 1984;247(Endocrinol Metab 10):E21–8. 7. Esler M, Jennings G, Korner P, Willett I, Dudley F, Hasking G, et al. The assessment of human sympathetic
598
M. Esler / Journal of the American Society of Hypertension 8(8) (2014) 593–598
nervous system activity from measurements of norepinephrine turnover. Hypertension 1988;11:3–20. 8. Grassi G, Colombo M, Seravalle G, Spaziani D, Mancia G. Dissociation between muscle and skin sympathetic nerve activity in essential hypertension, obesity, and congestive heart failure. Hypertension 1998;31:64–7. 9. Lambert E, Straznicky N, Schlaich MP, Dawood T, Hotchkin E, Esler MD, et al. Differing patterns of sympathoexcitation in normal weight and obesity-related hypertension. Hypertension 2007;50:862–8. 10. DiBona GF, Esler M. Translational medicine: the antihypertensive effect of renal denervation. Am J Physiol (Regul Integr Comp Physiol) 2012;298:R245–53. 11. Huggett RJ, Burns J, MackIntosh AF, Mary DASG. Sympathetic neural activation in nondiabetic metabolic syndrome and its further augmentation by hypertension. Hypertension 2004;44:1–6. 12. Rumantir MS, Vaz M, Jennings GL, Collier G, Kaye DM, Seals DR, et al. Neural mechanisms in human obesity-related hypertension. J Hypertens 1999; 17:1125–33. 13. DiBona GF, Kopp UC. Neural control of renal function. Physiol Rev 1997;77:75–197. 14. Oldham JB. Denervation of the kidney. Hunterian Lecture of the Royal College of Surgeons, England. 9th March, 1950. Ann R Coll Surg Engl 1950;7: 222–45. 15. Krum H, Schlaich MP, Bohm M, Mahfoud F, RochaSingh K, Katholi R, et al. Percutaneous renal denervation in patients with treatment-resistant hypertension: final 3-year report of the Symplicity HTN-1 study. Lancet 2014;383:622–9. 16. Esler MD, Krum H, Schlaich M, Schmieder RE, Bohm M, Sobotka PA. Renal denervation for treatment of drug-resistant hypertension. One-year results from the Symplicity HTN-2 randomized, controlled trial. Circulation 2012;126:2976–82. 17. Investigators Symplicity HTN-1. Catheter-based renal sympathetic denervation for resistant hypertension: durability of blood pressure reduction out to 24 months. Hypertension 2011;57:911–7. 18. Worthley SG, Tsioufis CP, Worthley MI, Sinhal A, Chew DP, Meredith IT, et al. Safety and efficacy of a multi-electrode renal sympathetic denervation system in resistant hypertension: the EnligHTN I trial. Eur Heart J 2013;34:2132–40.
19. Ormiston JA, Watson T, van Pelt N, Stewart JT, White JM, Doughty RN, et al. Renal denervation for resistant hypertension using an irrigated radiofrequency balloon: 12-month results from the Renal Hypertension Ablation System (RHAS) trial. EuroIntervention 2013; 9:70–4. 20. Esler M. Louis F Bishop Lecture, treatment of drugresistant hypertension with renal denervation: rationale, results and recent challenges. American College of Cardiology 63rd Annual Scientific Session, 29 March 2014, Washington, DC. 21. Meredith IT, Esler MD, Cox HS, Lambert GW, Jennings GL, Eisenhofer G. Biochemical evidence of sympathetic denervation of the heart in pure autonomic failure. Clin Auton Res 1991;1:187–94. 22. Sakakura K, Ladich E, Cheng Q, Otsuka F, Yahagi K, Kolodgie F, et al. Anatomical distribution of human renal sympathetic nerves: pathological study. J Am Coll Cardiol 2014;63(12_S):A2151. 23. Dhatt D, Kandzari D, O’Neill, D’Agostino R, Flack J, Katzen B, et al. A controlled trial of renal denervation for resistant hypertension. N Engl J Med 2014;370: 1393–401. 24. Kandzari D; lecture on behalf of the Symplicity HTN-3 Investigators. Symplicity HTN-3 Trial: Analysis of potentially confounding factors. Paris: Euro PCR; 2014. 25. Roman RJ, Cowley AW. Characterisation of a new model for the study of pressure natriuresis in the rat. Am J Physiol Renal Physiol 1985;248:F190–8. 26. Shun-Shun M, Howard J, Francis D. Removing the hype from hypertension. BMJ 2014;348:g1937. 27. Esler M. The John Shepherd lecture. Renal denervation for hypertension: observations and predictions of a founder. Eur Heart J 2014. http: //dx.doi.org/10.1093/eurheart/ehu091. 28. Henegar J, Zhang Y, De Rama R, Hata C, Hall M, Hall J. Catheter-based radiofrequency renal denervation lowers blood pressure in obese hypertensive dogs. Am J Hypertens 2014. http://dx.doi.org/10.1093/ajh/hpu048. 29. Campese V, Kogosov E. Renal afferent denervation prevents hypertension in rats with chronic renal failure. Hypertension 1995;25:878–82. 30. Schlaich MP, Sobotka PA, Krum H, Lambert E, Esler MD. Renal sympathetic nerve ablation for treatment of uncontrolled hypertension. N Engl J Med 2009;361:932–4.
Review Article
Illusions of truths in the Symplicity HTN-3 trial: generic design strengths but neuroscience failings Murray Esler, MBBS, PhD, FRACP* Human Neurotransmitter Laboratory, Baker IDI Heart and Diabetes Institute, Melbourne, Australia Manuscript received and accepted June 4, 2014
Abstract The Achilles heel in catheter-based studies of renal denervation for severe hypertension is the almost universal failure to apply a confirmatory test for renal denervation. When renal denervation efficacy was assessed, using measurements of the spillover of norepinephrine from the renal sympathetic nerves to plasma, the only test validated to this point, denervation was found to be incomplete and nonuniform between patients. It is probable that the degree of denervation has typically been suboptimal in renal denervation trials. This criticism applies with special force to the Symplicity HTN-3 trial, where the proceduralists, although expert interventional cardiologists, had no prior experience with the renal denervation technique. Their learning curve fell during the trial, a shortcoming accentuated by the fact that one-third of operators performed one procedure only. Recently presented results from the Symplicity HTN-3 trialists confirm that renal denervation was not effectively or consistently achieved in the trial. J Am Soc Hypertens 2014;8(8):593–598. Ó 2014 American Society of Hypertension. All rights reserved. Keywords: Resistant hypertension; renal denervation; sympathetic nervous system; antihypertensive device.
Despite the widespread availability and prescribing of angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, diuretics, and calcium channel blockers in a substantial minority of patients with essential hypertension, perhaps 10%,1,2 target blood pressure (BP) cannot be achieved. In these drug-resistant hypertensives a new strategy was needed, and in fact, devised. This was the development of device-based therapies targeting the sympathetic nervous system, the surgically implanted barostimulator
The author is a Senior Director of the Baker IDI Heart and Diabetes Institute Melbourne, and Conjoint Professor of Medicine Monash University, Melbourne. His primary funding is in the form of a Senior Principal Research Fellowship from the Australian National Health and Medical Research Council [ID586601]. He discloses research funding and receipt of consultancy fees, travel expenses, and honoraria from Medtronic, which funded the research from which some of the presented results are drawn, but the author holds no shares in the company or patent rights for renal denervation. In the preparation of this article no financial or other support was provided by Medtronic. *Corresponding author: Murray Esler, MBBS, PhD, FRACP, Baker IDI Heart and Diabetes Institute, PO Box 6492, St Kilda Road Central, Melbourne, Vic 8008, Australia. E-mail: [email protected]
device3 and catheter-based renal denervation,4,5 the latter being the subject of this review article. Central to the development of radiofrequency (RF) renal denervation was knowledge of the physiology of the renal sympathetic nerves and their pathophysiology in experimental and human hypertension. In untreated essential hypertensive patients, the application of regional norepinephrine isotope dilution methodology,6 to measure the outward flux of the transmitter from the sympathetic nerves of the kidneys to plasma (‘‘renal norepinephrine spillover’’), demonstrated that a high level of activation of the renal sympathetic outflow was present.7 This renal sympathetic activation is central to hypertension pathogenesis.7–12 In rodents, the renal nerves have been demonstrated to stimulate secretion of renin from the juxtglomerular apparatus, to promote renal tubular reabsorption of sodium, and to cause renal vasoconstriction, all potentially BP elevating responses.13 The renal tubules receive a dense sympathetic innervation, at all tubular levels. In experimental models of hypertension, surgical sectioning of the renal sympathetic nerves lowers BP or delays or prevents the development of hypertension.13 These facts and knowledge of the anatomy of the postganglionic renal sympathetic nerves in their passage to the kidneys provided the intellectual framework for the
1933-1711/$ - see front matter Ó 2014 American Society of Hypertension. All rights reserved. http://dx.doi.org/10.1016/j.jash.2014.06.001
594
M. Esler / Journal of the American Society of Hypertension 8(8) (2014) 593–598
development of catheter-based renal denervation for the treatment of essential hypertension. In humans, the renal sympathetic nerves pass from the sympathetic chain and ganglia to the kidneys via the outer wall of the renal arteries or just outside in perirenal adipose tissue and connective tissue, within reach of RF energy delivered by a catheter in the artery lumen.14 Seven years after the first patient was treated with the Symplicity RF catheter system, these initial trials, their continuation to later specified endpoints,15,16 accompanying resistant hypertension renal denervation registry files17 and trials with other newly engineered renal denervation devices18,19 have established important therapeutic principles: 1. Efferent sympathetic renal denervation can be achieved with luminal delivery of RF and ultrasonic energy. 2. The treatment can be delivered with very minimal procedural complications, these being typical of invasive cardiological procedures in general, and not consequential on RF energy delivery. 3. The mean BP reduction across the trials shows consistency, office systolic BP falling on average by 20– 30 mm Hg at the primary endpoints. Renal function is preserved. The BP reduction is durable, demonstrably persisting for 3 years and beyond. 4. New renal artery stenoses in the field of RF energy delivery are very uncommon.
Among clinical studies of renal denervation, only in the Simplicity HTN-1 trial4 was renal denervation confirmed, by measurements of norepinephrine spillover (release of the transmitter from the renal nerves to plasma; Figure 1). Measurement of regional norepinephrine spillover is well established as a valid test for denervation, having been applied for two decades in the diagnosis of pure autonomic failure,21 a disorder characterized by spontaneous degeneration of postganglionic sympathetic nerve fibers, and denervation of the heart, kidneys, and other organs. The degree of renal denervation achieved in the Symplicity HTN-1 trial was less than expected (on average 47%)4 but did seem to be sufficient, in that the antihypertensive response was very adequate. This led to complacency, in that renal denervation was said to be procedurally easy, to the point of being ‘‘boring’’, a commonly used descriptor. Subsequent analysis by my group confirms that denervation is often very incomplete and surprisingly nonuniform from patient-to-patient (Figure 2). Achieved denervation can be <25%, no doubt inadequate for a full therapeutic effect. So the denervation catheter technique might look ‘‘easy’’
A disappointing aspect of renal denervation therapy, however, is the failure of some patients to experience a drop in arterial pressure, this being reported in different studies to occur in 15–50% of patients. The most commonly offered explanation for this is that in these patients sympathetic nervous system activation is not the primary pathophysiology. But this might not be true, as surgical renal denervation commonly lowers BP in experimental models of hypertension in which sympathetic nervous activity is actually normal.13 Recent observations, described in the following, suggest that the real reason might more commonly be that in ‘‘nonresponders,’’ the denervation procedure has failed technically. Renal denervation in human hypertension is very incomplete, nonuniform from patient to patient, and achieved with difficulty.20
Testing for Achieved Renal Denervation The belief that renal denervation has been achieved in clinical trials of catheter-based renal denervation is almost invariably based on hope and trust, rather than testing. This contrasts with the studies of surgical renal denervation in experimental hypertension, where good experimental design necessitates that the effectiveness of denervation is always confirmed, typically by documenting 90–95% reduction in the kidney content of norepinephrine.
Figure 1. Evidence for catheter-based renal denervation with the Symplicity radiofrequency Arch catheter in patients with resistant hypertension and pigs. Renal sympathetic denervation was quantified in humans with measurements of the fall in renal norepinephrine spillover postprocedure and in pigs using the fall in kidney norepinephrine content, both being validated methods.
M. Esler / Journal of the American Society of Hypertension 8(8) (2014) 593–598
595
arteries. In the Symplicity HTN-3 study,23 we now know that neither of these technical prerequisites was consistently achieved.24
How Much Renal Denervation is Needed? The level of renal sympathetic nerve ablation achieved clinically with the RF denervation procedure, mean 47%,4 contrasts with that achieved surgically in experimental hypertension, 90–95%.13 Which is optimal? If the aim is to ‘‘normalize’’ renal sympathetic activity, 50% ablation of sympathetic nerves in patients who have, on an average, a doubling of renal sympathetic activity7,12 might perhaps be ideal. But as renal denervation lowers BP in some experimental models of hypertension in which sympathetic nervous activity is actually normal,13 normalizing renal sympathetic activity is probably not the desired aim. Greater left-shift in the pressure natriuresis-curve25 and resultant off-loading of more sodium would occur with a higher grade renal denervation than is currently achieved clinically. This could be achieved if delivered energy were to penetrate somewhat beyond 2–3 mm, as some renal sympathetic nerves are more remote from the lumen than this.14,22
Renal Denervation ‘‘Armageddon’’: The Symplicity HTN-3 Trial
Figure 2. The effectiveness of catheter-based radiofrequency renal nerve ablation assessed with renal norepinephrine spillover measurements made before and 30 days after the procedure. The denervation was incomplete, sometimes markedly so, and with pronounced nonuniformity between individual patients. These results contradict the idea that achieving sympathetic denervation is technically ‘‘easy’’. The results suggest that blood pressure nonresponse to renal denervation must commonly be due to very imperfect denervation and not only to the absence of sympathetic nervous activation, which is often invoked as the cause. From unpublished results of the author, Markus Schlaich, Gavin Lambert and Dagmara Hering.
compared with many other interventional cardiological and radiological procedures, but achieving denervation is complex and difficult. Cartoons showing the easy proximity of renal sympathetic nerves to the renal artery lumen were misleading. Older surgical anatomies of the renal nerves14 demonstrated that close to the origin of the renal artery from the aorta, the sympathetic nerves were more distant, but converged on the renal arteries distally, near the renal artery branch point. Contemporary anatomical studies confirm this.22 Clearly ablative energy should be preferentially focused on the distal renal artery, but this is often not done, and on the full circumference of both renal
A challenge to the percutaneous renal denervation treatment of resistant hypertension came with the 9 January 2014 press release concerning the Simplicity HTN-3 trial in drug-resistant hypertension, the pivotal study for US Food and Drug Administration licensure, and in the subsequent New England Journal of Medicine publication on 29 March,23 indicating that the primary efficacy endpoint had not been reached in the trial. A lot was expected of the Symplicity HTN-3 study. Five times larger than the first two Symplicity renal denervation trials and incorporating a blinded sham design, this trial was expected to provide the definitive statement on the value of renal denervation in the treatment of patients with severe hypertension. To many it did—‘‘renal denervation does not work!’’26 The sham design was lauded.26 This trial exemplar had comprehensively exposed the fallacy of imagined renal denervation benefits! How was it possible to argue against the findings of the Symplicity HTN-3 trial (Table 1)?
Symplicity HTN-3 Illusions The perfection of the trial is an illusion (Table 2), only apparent if no discrimination is attempted between its ‘‘form’’ and its ‘‘substance’’. The form of the trial is represented by its generic strengths (large patient enrollment, blinding, sham control), applicable to any BP trial.
596
M. Esler / Journal of the American Society of Hypertension 8(8) (2014) 593–598
Table 1 How is it possible to argue against the findings of the Symplicity HTN-3 Trial? The power of Food and Drug Administration ‘‘branding’’ of a pivotal US study The generic strength of the sham procedure The authority of a New England Journal of Medicine publication .but with very material failings in execution
But this is a neuroscience trial in hypertension, based on a neuroscience theory of causation of severe hypertension (activation of the renal sympathetic outflow), on detailed knowledge of neuroanatomy (of the renal efferent and afferent nerves), and on the engineering of nerve ablative devices. This essence of the trial received inadequate attention and its execution ultimately failed. Among editorialists there was lauding of the sham procedure but neglect of trial neuroscience failings26 Much was amiss with Symplicity HTN-3. At 88, too many centers were recruited for the trial and at 111, too many proceduralists.23 No hands-on experience in renal denervation before the trial was possible in the United States (unlike in the earlier Symplicity trials, where it was mandatory). Experts in their field of interventional cardiology, all participants were novices in the renal denervation procedure. Proctoring was done primarily by company staffers, rather than experienced physicians or renal denervation engineers, in contrast to the earlier Symplicity trials. Energy delivery was not preferentially to the distal renal artery, where it should have been, but inexplicably was more typically to the proximal renal artery; the renal nerves are, in fact, closer to the distal artery.14,22 It is now a matter of record that the denervation procedure fared badly in Symplicity HTN-3.24 Retrospective analysis of stored angiographic and procedural records of all RF energy applications demonstrated that in 74% of patients not even one fully circumferential renal artery application of energy was achieved,24 when it is mandatory that this be achieved bilaterally, making effective nerve ablation impossible.27 Not surprising, the BP fall in these patients was nearly identical to that of the sham group. In contrast, in that minority of patients in whom RF energy was delivered according to the trial protocol in a full circumference in both renal arteries, BP fell very materially, similar in degree to that observes in the earlier Symplicity trials.4,5,24 The obvious rhetorical question is ‘‘what use is a sham design when patients were not effectively or consistently denervated’’? And how could this happen? Presumably Table 2 Form and substance in the Symplicity HTN-3 trial Generic strengths—constituting a plus in any antihypertensive trial (‘‘form’’) Neuroscience failings—in this trial (‘‘substance’’)
because renal denervation was thought and said to be procedurally easy, to the point of being ‘‘boring’’, a commonly used descriptor. Physical aspects of the catheter procedure perhaps looked easy, but achieving denervation was not.4 It should be noted that the field of renal denervation for experimental hypertension is active, in fact energized by the clinical studies. Experimental surgical- and catheterbased denervation for hypertension still works! This was recently well exemplified with catheter-based renal denervation abolishing hypertension in an obese dog model.28 Why should renal denervation, in four mammalian species (rats, dogs, rabbits, and pigs) invariably be antihypertensive,13 but not in the human mammal in Symplicity HTN-3? In the US Pivotal study, a failure to achieve renal denervation came first to mind.20 I believe the Symplicity HTN-3 study failed in its neuroscience execution, and will be judged to be ‘‘on the wrong side of history’’.
Development of a More Generally Accessible Method for Testing Efferent Sympathetic Nerve Ablation? There is a pressing need to incorporate valid testing for renal sympathetic denervation into all clinical trials of the procedure. Growing acceptance has emerged for this important principle, but some of the planned testing for denervation, such as using measurements of heart rate variability, are fanciful. In truth, it is difficult to know how to proceed, as the only validated test for renal denervation, measurement of renal norepinephrine spillover,4,6 has very limited availability and could not be incorporated into an international multicentre trial of the type currently planned by Medtronic, after the technical failure of the Symplicity HTN-3 trial. My own group has commenced exploring a different approach. With ablation of sympathetic nerves, the specific sympathetic nerve proteins, such as tyrosine hydroxylase, will degrade and fragment. Although within the nerves external to the kidneys thermal degradation of nerve proteins might make their identification difficult, sympathetic nerves within the kidneys will escape this process of thermal denaturation. We will now test for the excretion of tyrosine hydroxylase fragments in urine. The undegraded protein is too large to enter urine, but we anticipate tyrosine hydroxylase fragments will. Detection will be with Western blot analysis and enzyme-linked immunosorbent assay testing, presuming the persistence of antibody responsive sites on the tyrosine hydroxylase fragments. It is anticipated that urinary excretion of the marker fragments will be at a peak in the first few days after the denervation procedure, but this will need to be established. Once these principle are established, and at the moment what I provide here is only a promissory note, in a multicentre trial, urine might be collected and sent to a core laboratory for analysis.
M. Esler / Journal of the American Society of Hypertension 8(8) (2014) 593–598
Afferent Renal Nerves and Their Ablation Sensory nerves running from the kidneys to the central nervous system contribute to activation of the sympathetic nervous system in patients with drug-resistant hypertension, through their projection to the hypothalamus.29 Ablation of these nerves during the renal denervation procedure contributes to BP lowering by causing central inhibition of central sympathetic outflow.30 The nociceptive receptors within the kidney, which are responding to a renal injury signal,29 can be targeted with adenosine in a test for afferent nerve ablation which can be conducted in the catheterization laboratory. Adenosine infused into the renal artery increases sympathetic nervous activity and BP. To test for afferent nerve ablation on the catheter day, adenosine is infused before and after the denervation procedure; a reduction in the BP response is indicative of afferent nerve ablation. This test, not fully evaluated as yet, offers prospects of real time evaluation to guide the proceduralist. It is sometimes thought that testing for afferent nerve ablation would also indicate the success or otherwise of efferent, sympathetic ablation. This may not be true as efferent and afferent nerves do not traverse exactly parallel paths. The sympathetic nerves converge on the distal renal artery as they near the kidney but, conversely, the afferent nerves are more distant from the artery as they leave the kidneys, but converge on the renal arteries closer to the aorta.22 The most successful afferent nerve ablation would be expected to result from RF or ultrasound energy delivered through the proximal renal artery, but efferent sympathetic ablation would be less successful at that site.
The Future of Renal Denervation in Treating Severe Hypertension The Symplicity HTN-3 trial has inflicted a ‘‘flesh wound’’ on the renal denervation field, but this is not fatal; the trial is now seen to have failed in its neuroscience execution, not effectively and consistently achieving renal denervation.24 In the pause that this trial created worldwide, it is now time to reflect on what future science is needed (Table 3). Failure to test for renal denervation, as an intrinsic component of all trials but one, represents the Achilles heel of the field. It is probable that in some negative trials, and with nonresponse in some patients, failure to Table 3 The next era of renal denervation neuroscience begins Current levels of achieved denervation are not optimal? Higher energy/multipolar radiofrequency electrodes are needed? Energy should be delivered to both proximal and distal renal artery? Testing for achieved renal denervation is mandatory? Preselection is desirable, but problematic? Renal denervation may not show a ‘‘class effect’’?
597
achieve adequate denervation was responsible. The commonly attributed reason for renal denervation to fail to lower BP, that in such patients sympathetic activity was not increased, has no authenticity in studies where denervation was not tested; failure to denervate is the ever-present confounder. Planned studies, including the second Medtronic US Pivotal Trial in hypertension, will need to incorporate denervation testing. As renal denervation is incomplete and nonuniform between patients, usage of higher RF and ultrasonic energy doses in the future is envisaged (the established safety margin in renal denervation allows this) and multielectrode RF catheters. Energy may need to be delivered to both the aortic and the distal ends of the renal artery, to target both sympathetic and afferent nerves.22 Preselection of patients will remain problematic, until the influence on the responder status of the two determinants, effectiveness of denervation, and hypertension pathophysiology can be discriminated. Will renal denervation show a ‘‘class effect’’, with all energy forms for denervation being equally efficacious? Direct comparison will be needed, as claims and counter claims could be distorted by commercial interest.
References 1. de la Sierra A, Segura J, Banegas JR, Gorostidi M, de la Cruz JJ, Armario P, et al. Clinical features of 8295 patients with resistant hypertension classified on the basis of ambulatory blood pressure monitoring. Hypertension 2011;57:898–902. 2. Persell SD. Prevalence of resistant hypertension in the United States, 2003-2008. Hypertension 2011;57: 1076–80. 3. Bisognano JD, Bakris G, Nadim MK, Sanchez L, Kroon AA, Schafer J, et al. Baroreflex activation therapy lowers blood pressure in patients with resistant hypertension. J Am Coll Cardiol 2011;58:765–73. 4. Krum H, Schlaich MP, Whitbourn R, Sobotka P, Sadowski J, Bartus K, et al. Catheter-based renal sympathetic denervation for resistant hypertension: a multicentre safety and proof-of-principle cohort study. Lancet 2009;373:1275–81. 5. Symplicity HTN-2 Investigators. Renal sympathetic denervation in patients with treatment-resistant hypertension (The Symplicity HTN-2 Trial): a randomized controlled trial. Lancet 2010;376:1903–9. 6. Esler M, Jennings G, Korner P, Blombery P, Sacharias N, Leonard P. Measurement of total and organ-specific norepinephrine kinetics in humans. Am J Physiol 1984;247(Endocrinol Metab 10):E21–8. 7. Esler M, Jennings G, Korner P, Willett I, Dudley F, Hasking G, et al. The assessment of human sympathetic
598
M. Esler / Journal of the American Society of Hypertension 8(8) (2014) 593–598
nervous system activity from measurements of norepinephrine turnover. Hypertension 1988;11:3–20. 8. Grassi G, Colombo M, Seravalle G, Spaziani D, Mancia G. Dissociation between muscle and skin sympathetic nerve activity in essential hypertension, obesity, and congestive heart failure. Hypertension 1998;31:64–7. 9. Lambert E, Straznicky N, Schlaich MP, Dawood T, Hotchkin E, Esler MD, et al. Differing patterns of sympathoexcitation in normal weight and obesity-related hypertension. Hypertension 2007;50:862–8. 10. DiBona GF, Esler M. Translational medicine: the antihypertensive effect of renal denervation. Am J Physiol (Regul Integr Comp Physiol) 2012;298:R245–53. 11. Huggett RJ, Burns J, MackIntosh AF, Mary DASG. Sympathetic neural activation in nondiabetic metabolic syndrome and its further augmentation by hypertension. Hypertension 2004;44:1–6. 12. Rumantir MS, Vaz M, Jennings GL, Collier G, Kaye DM, Seals DR, et al. Neural mechanisms in human obesity-related hypertension. J Hypertens 1999; 17:1125–33. 13. DiBona GF, Kopp UC. Neural control of renal function. Physiol Rev 1997;77:75–197. 14. Oldham JB. Denervation of the kidney. Hunterian Lecture of the Royal College of Surgeons, England. 9th March, 1950. Ann R Coll Surg Engl 1950;7: 222–45. 15. Krum H, Schlaich MP, Bohm M, Mahfoud F, RochaSingh K, Katholi R, et al. Percutaneous renal denervation in patients with treatment-resistant hypertension: final 3-year report of the Symplicity HTN-1 study. Lancet 2014;383:622–9. 16. Esler MD, Krum H, Schlaich M, Schmieder RE, Bohm M, Sobotka PA. Renal denervation for treatment of drug-resistant hypertension. One-year results from the Symplicity HTN-2 randomized, controlled trial. Circulation 2012;126:2976–82. 17. Investigators Symplicity HTN-1. Catheter-based renal sympathetic denervation for resistant hypertension: durability of blood pressure reduction out to 24 months. Hypertension 2011;57:911–7. 18. Worthley SG, Tsioufis CP, Worthley MI, Sinhal A, Chew DP, Meredith IT, et al. Safety and efficacy of a multi-electrode renal sympathetic denervation system in resistant hypertension: the EnligHTN I trial. Eur Heart J 2013;34:2132–40.
19. Ormiston JA, Watson T, van Pelt N, Stewart JT, White JM, Doughty RN, et al. Renal denervation for resistant hypertension using an irrigated radiofrequency balloon: 12-month results from the Renal Hypertension Ablation System (RHAS) trial. EuroIntervention 2013; 9:70–4. 20. Esler M. Louis F Bishop Lecture, treatment of drugresistant hypertension with renal denervation: rationale, results and recent challenges. American College of Cardiology 63rd Annual Scientific Session, 29 March 2014, Washington, DC. 21. Meredith IT, Esler MD, Cox HS, Lambert GW, Jennings GL, Eisenhofer G. Biochemical evidence of sympathetic denervation of the heart in pure autonomic failure. Clin Auton Res 1991;1:187–94. 22. Sakakura K, Ladich E, Cheng Q, Otsuka F, Yahagi K, Kolodgie F, et al. Anatomical distribution of human renal sympathetic nerves: pathological study. J Am Coll Cardiol 2014;63(12_S):A2151. 23. Dhatt D, Kandzari D, O’Neill, D’Agostino R, Flack J, Katzen B, et al. A controlled trial of renal denervation for resistant hypertension. N Engl J Med 2014;370: 1393–401. 24. Kandzari D; lecture on behalf of the Symplicity HTN-3 Investigators. Symplicity HTN-3 Trial: Analysis of potentially confounding factors. Paris: Euro PCR; 2014. 25. Roman RJ, Cowley AW. Characterisation of a new model for the study of pressure natriuresis in the rat. Am J Physiol Renal Physiol 1985;248:F190–8. 26. Shun-Shun M, Howard J, Francis D. Removing the hype from hypertension. BMJ 2014;348:g1937. 27. Esler M. The John Shepherd lecture. Renal denervation for hypertension: observations and predictions of a founder. Eur Heart J 2014. http: //dx.doi.org/10.1093/eurheart/ehu091. 28. Henegar J, Zhang Y, De Rama R, Hata C, Hall M, Hall J. Catheter-based radiofrequency renal denervation lowers blood pressure in obese hypertensive dogs. Am J Hypertens 2014. http://dx.doi.org/10.1093/ajh/hpu048. 29. Campese V, Kogosov E. Renal afferent denervation prevents hypertension in rats with chronic renal failure. Hypertension 1995;25:878–82. 30. Schlaich MP, Sobotka PA, Krum H, Lambert E, Esler MD. Renal sympathetic nerve ablation for treatment of uncontrolled hypertension. N Engl J Med 2009;361:932–4.