Device-based treatments in hypertension

The clinical development of device-based treatments for the management of hypertensive patients not sufficiently responding to pharmacotherapy created warm discussions in the hypertension community. Although some proponents may be overly enthusiastic regarding the merits of such treatments, others fervently refuse interventional procedures altogether. Provided that opinions exhibit a Gaussian distribution – which may or may not be true – the majority will be without fear or favor at this stage. We simply do not have the data to inform clinical decisions. Even though some devices are approved for clinical use in Europe, we know surprisingly little about their mode of action, safety, and clinical efficacy. Outcome data do not exist. Furthermore, given their invasiveness and costs, predictors for individual treatment success or lack thereof should be defined. The state of affairs is largely explained by differences in the regulatory approval path between medical devices and medications. Compared with medications, medical devices undergo less rigorous clinical testing before they receive a Conformité Européenne (CE) marking and enter the European market. Market uptake is primarily related to national or regional reimbursement rules. Clinical trials on devices are typically smaller, and true efficacy data are not required. Nevertheless, we should demand such data before applying new device-based treatments to the patient in front of us. Seventy-five years ago, Grimson [1] published a seminal article on surgical sympathectomy in the treatment of hypertension. He pointed out that ‘a proper evaluation of the effect of any form of therapy in hypertension is always difficult because of the variety of symptoms encountered and the difficulty of establishing a definite control’. Indeed, the design of sham-controlled trials testing devices poses practical and ethical difficulties. In this issue of the Journal, Wallbach et al.[2] report data on a series of patients with treatment-resistant arterial hypertension who had been implanted with an electrical carotid sinus stimulator following an unsatisfactory response to catheter-based renal nerve ablation. Patients were recruited in experienced hypertension centers, and secondary causes were ruled out. On average, patients were on more than six antihypertensive drugs. In addition, catheter-based renal nerve ablation had been performed on these patients. Yet, average office blood pressure (BP) at baseline was 182/96 mmHg. Patients were implanted with an electrical carotid sinus stimulator using a unilateral unipolar electrode design. It should be noted that this device, unlike its predecessor with bilateral and tripolar electrodes [3], has only been tested in an uncontrolled clinical trial. Acute studies suggest that the electrode design could affect efficacy and tolerability [4]. Four weeks following implantation, the device was activated and individually adjusted. Office BP was lower at 6 months. Out of 28 patients, 23 underwent ambulatory BP monitoring before and on electrical carotid sinus stimulation. These measurements are a particular strength of the study and may provide a more realistic approximation of clinical efficacy compared with office BP measurements. Remarkably, 24-h SBP was not significantly reduced at month 6. In a subset with another ambulatory BP recording at month 12, 24-h SBP was 14 mmHg lower compared with the baseline measurement. Moreover, 24-h DBP was unchanged at month 6 and showed a numerical albeit nonsignificant decrease by 6 mmHg at month 12. When a clinical response was defined as 5 mmHg decrease in 24-h SBP, 48% of patients were responders, leaving 52% as nonresponders. There was no signal for worsened renal function and a tendency for albuminuria to improve over time. The number of prescribed antihypertensive medications remained unchanged. The study by Wallbach et al.[2] has important limitations, particularly the open design and lack of a control group. Furthermore, it is unknown whether or not catheter-based renal nerve ablation actually denervated the kidneys. Nevertheless, the findings provide useful clinical and scientific information. Using the modest ambulatory BP criteria above, at least 50% of patients with treatment-resistant arterial hypertension do not respond to electrical carotid sinus stimulation using the unilateral unipolar electrode design – with or without previous renal nerve ablation [5]. Provided that part of the response is explained by regression to the mean, a Hawthorne effect or a placebo response, the true proportion of nonresponders may be even higher. Clinicians deciding to implant electrical carotid sinus stimulators outside clinical trials should discuss these facts openly with their patients. It may be prudent restricting use of the treatment to clinical studies until data from controlled trials become available. Future research should identify reasons why many patients do not respond to electrical carotid sinus stimulation. Perhaps, baroreflex engagement is not ideal with current electrode designs [4]. It is also possible that activation of peripheral chemoreceptors, that are also located in the carotid sinus, attenuates the response. Finally, optimal placement of stimulator electrodes during surgery is not trivial because anesthesia often normalizes BP even in severely hypertensive patients. Improved device design could address these issues. Similarly, there have been refinements in renal nerve ablation catheters and procedures. For example, histopathological studies revealed that renal nerves are more closely located to the vascular lumen in more distal parts of the renal artery [6]. In pigs treated with catheter-based renal nerve ablation, the majority of nerves were outside the lesioned area [7]. Although increasing the number of ablation points did not result in significant improved efficacy in this model, more distal ablation appeared to be more effective [8]. Meanwhile, new methodologies have been developed such as transcatheter application of ethanol [9]. Both renal nerve ablation and electrical carotid sinus stimulation target the sympathetic nervous system. Yet, sympathetic contributions to BP vary between hypertensive patients [10]. In fact, arterial hypertension in the supine position is common in patients with severe pure autonomic failure independently of efferent sympathetic function [11]. Only one-third of the patients, submitted to surgical sympathectomy, showed satisfactory sustained BP reduction [1]. The response of sympathetic nerve traffic responses to renal denervation is also heterogeneous [12]. Therefore, it is very unlikely that all patients respond to renal nerve ablation, electrical carotid sinus stimulation, or both procedures combined. Animal studies demonstrated that the mechanisms driving arterial hypertension affect the treatment response. Dogs with diet-induced obesity, which is associated with a sympathetically mediated increase in BP, exhibited a robust depressor response to electrical carotid sinus stimulation [13]. In contrast, the response was diminished in dogs with angiotensin II-induced arterial hypertension [14], and aldosterone infusion attenuated the response to both electrical carotid sinus stimulation and surgical renal denervation [15]. In rabbits, angiotensin II infusion chronically lowers renal sympathetic activity assessed using telemetry [16]. Perhaps, more detailed physiological profiling could identify patients who are more likely to respond. How antihypertensive medications interact with interventional treatments should also be systematically assessed. From a scientific point of view, the idea that electrical carotid sinus stimulation may chronically lower BP in a patient with denervated kidneys is thought-provoking. The finding seems to defy physiology textbook knowledge in several ways. It is, therefore, reassuring that surgical renal denervation did not abolish BP reduction with electrical carotid sinus stimulation in dogs [17]. Baroreflexes were thought not to contribute to long-term BP control. Findings with electrical carotid sinus stimulation and elegant physiological experiments [18] challenge this dogma. Furthermore, renal sympathetic innervation was thought to have a critical role in the pathogenesis of arterial hypertension, and the maintained response to electrical carotid sinus stimulation in the setting of denervated kidney is unexpected. This does not mean, however, that renal sympathetic activity does not contribute at all to the BP reduction with electrical carotid sinus stimulation [19]. While waiting for the results of properly controlled clinical trials – which are the responsibility of the manufacturers – we should carefully investigate patients who have been submitted to interventional treatments to learn more about human pathophysiology. The fact that some of our current concepts regarding cardiovascular physiology in general and mechanisms of action of device-based treatments in particular may not be true is a strong incentive. Finally, we should assess whether findings in animal physiology will have a bearing on human hypertension. For example, renal nerves promote sodium excretion in angiotensin II-mediated hypertension [20]. In the Dahl salt-sensitive rat, complete renal denervation lowered BP, whereas selective afferent renal denervation was ineffective [21]. In addition to advancing science, these investigations may help to improve devices and to individualize treatment decisions later on. ACKNOWLEDGEMENTS Disclosures: J.J. served as advisor for Boehringer-Ingelheim, Janssen-Cilag, Novartis, Orexigen, Vivus, and receives research support from Boston Scientific. J.T. receives research support from Boston Scientific. G.G. reports no disclosure. Conflicts of interest There are no conflicts of interest.