Moderator's View: Renal denervation: The jury is still out and the verdict will be more complex than initially envisaged

New scientific findings and discoveries that may have an impact upon human health, particularly therapies for conditions that demand a high death toll, have a characteristic trajectory when described in a graph relating the scientific proof of the validity/invalidity of the same research findings to the interest they attract within the medical community and the public at large (Figure 1). Ideally, the two variables should maintain a linear association: the interest should grow in strict proportionality to the strength of the scientific proof underlying the validity or invalidity of the same findings. However, the two variables, the strength of the scientific proof and the professional and public interest, have a polynomial rather than a linear relationship. Initially, discoveries are overlooked and the interest/scientific validity curve lies largely below the ideal, linear, relationship. At some stage, the interest grows exponentially and the relationship crosses the ideal line and the scientific discovery in question becomes overemphasized and fashionable or prematurely abandoned. At this point, scientific scrutiny becomes more stringent and, when the validity of the same discovery is well proven (or negated), the curve plateaus and the interest in the discovery in question eventually becomes independent of the scientific proof underlying its value. This is the stage when the implication of the discovery for human health is either definitively accepted and translated into established treatments or dismissed due to lack of value. This theoretical relationship perfectly serves the scope of describing the trajectory of renal denervation as a therapy for treatment-resistant hypertension. Evidence that selective renal denervation delays the development and attenuates the severity of hypertension was well documented in experimental models in the early nineties [1], and in 1989, Esler et al. [2], the founder of renal denervation as a treatment for hypertension, had already proved that efferent sympathetic nerve activity was raised in human hypertension. Furthermore, in the late nineties, Campese et al. [4] provided experimental evidence that afferent sympathetic neurons transmit signals to the central nervous system that contribute to setting the overall efferent sympathetic tone in renal injury [3] and renal failure. Although research on the sympathetic nervous system was a major fruitful research area during these pioneer years, the potential of renal denervation as a cure for human hypertension was overlooked for nearly a decade. Mainly thanks to well-focused studies in a swine model by Esler's group, renal denervation was eventually developed as a minimally invasive procedure applicable in human studies. The procedure exploits the favorable location of the sympathetic fibers around the renal arteries and their peculiar sensitivity to radiofrequency. After the Simplicity-1 trial, interest in renal denervation literally boomed and health systems in affluent countries including Australia, the country where this technique was founded, the Netherlands, Germany and Italy granted reimbursement for renal denervation in resistant hypertension. Several thousand patients were subjected to the procedure and the potential indications for the application of the technique flourished. From 1 April 2009 to 31 December 2013, as many as 672 papers addressing renal denervation were published in PubMed-indexed journals. At this stage, the curve reached a level well above the underlying evidentiary basis. We are now at a critical point in time: the first properly controlled clinical trial testing sympathetic denervation versus a sham procedure by the Medtronic device (Symplicity HTN-3) in the USA produced negative results, i.e. it failed to show a greater fall in office systolic blood pressure (BP) (the primary endpoint of the study) in hypertensive patients subjected to renal denervation than in sham-treated patients [5]. Among the factors which might have artifactually amplified the apparent BP lowering effect of renal denervation in previous studies not contemplating a proper control group (i.e. treated with a sham procedure), the use of office BP as an outcome measure appears of paramount relevance. Failure to adopt 24-h ambulatory blood pressure measurement (24-h ABPM) as a mandatory criterion for the definition of resistant hypertension seriously undermines the definition of this pragmatically delineated disorder [6]. It is well known that 24-h ABPM, but not office BP, is an independent predictor of adverse cardiovascular outcomes in patients with treatment-resistant hypertension [7]. In many centres, experience has shown that the application of stringent criteria for the definition of resistant hypertension makes the diagnosis of this disorder very rare. Initially, my group was enthusiastic about the potential of renal denervation after the Simplicity HTN-1 study and actively sought treatment-resistant patients from our chronic kidney disease (CKD) and hypertension clinics applying a stringent protocol including 24-h ABPM and verification of compliance with drug prescriptions. Yet, over the past 2 years, we failed to identify even a single patient with truly resistant hypertension. Simple adjustment of drug doses and optimization of drug combinations was associated with a spectacular—28/−11 mmHg—BP fall in a recent series of 10 patients with apparent treatment-resistant hypertension [8]. Truly resistant patients certainly exist, but these patients are rarer than currently estimated by loose protocols based exclusively on office BP and without actual verification of drug compliance. On one side of this controversy, Blankestijn et al. [9] recapitulate the underpinnings and history of renal denervation and specifically focus on the potential of this procedure in CKD patients, a category of patients for whom the relevance of renal nerves in triggering sympathetic overactivity is well established in experimental models and in studies in pre-dialysis CKD and kidney failure as well. Enthusiasm for the technique apart, these authors frankly recognize that several uncertainties remain to date that must be addressed before establishing this technique as a valid instrument to treat resistant hypertension, with or without evidence of renal damage, particularly the need for the development of valid prediction criteria to identify patients that may benefit from renal denervation. The unwarranted use of office BP, the attendant problem of ‘true resistance’, the lack of a sham intervention and several other factors that might have distorted the interpretation of the effect of renal denervation in previous trials, all resound in the counterpoint by Persu et al. [10], a group that has long maintained a cautionary attitude toward this procedure. Of these, lack of documentation of functional effectiveness of renal denervation is of obvious relevance and remains an unmet research need in studies of the kind. In the vein of precognizant scientific wisdom, Persu et al., state that ‘… failure of Symplicyty-3 was not a major surprise …’. These authors also emphasize the concerns for the safety of denervation for renal function and on the suboptimal short- and long-term assessment of the integrity of renal arteries after the procedure in most studies performed so far, all problems of critical relevance that should be addressed in detail in future studies. In the end, neither the PRO nor the CON side sound the death bell for renal denervation. Blankestijn et al. have an ongoing clinical trial (SYMPATHY) which is actively recruiting. This study will assess whether renal denervation added to standard care reduces ABPM when compared with standard care alone in subjects with treatment-resistant hypertension as defined by ABPM (systolic ABPM ≥ 135 mmHg). SYMPATHY will also test the effect of renal denervation in specific glomerular filtration rate strata (20–60 mL/min per 1.73 m2 and >60 mL/min per 1.73 m2). On the other hand, Persu et al., describe in some detail the kind of studies we sorely need and elaborate on a randomized controlled trial (INSPiRED), designed by the same authors, which adequately deals with the several shortcomings of the trials performed so far, from the issue of true resistance to safety and the identification of patients who may benefit from this intervention. This year marks the fifth anniversary of the Symplicity-HTN1 trial publication. Methodological issues apart, 5 years after this landmark pilot trial, several unresolved technical questions remain on the table, from the extent of renal denervation needed to optimize the effect of the procedure on BP to the validity of unilateral denervation or the equivalence of the different instruments now used for nerve ablation. Certainly, after the release of the preliminary analyses in Symplicity-HTN3, Esler's dream [11], that the application of renal denervation can be extended to early, mild hypertension, appears to be a more remote possibility than it was 2 or 3 years ago. As with drugs, the registration and the use of medical devices before they are demonstrated to be efficacious for the reduction of meaningful clinical endpoints, 24-h ABPM and cardiovascular events in our case, may not be in the best interest of patients and may even turn out to be counterproductive to the commercial interest of the industries producing the devices in question. Without proper objective (i.e. well-controlled) assessment, we will never know whether or not these devices are effective for treating the disorder they are presumed to be able to resolve or attenuate. (See related article by Persu et al. Renal denervation for all resistant hypertensive patients: the Emperor's new clothes. Nephrol Dial Transplant 2014; 29: 1116–1119; See related article by Blankestijn et al. Sympathetic renal denervation in hypertension and in chronic kidney disease. Nephrol Dial Transplant 2014; 29: 1120–1123.) You can find more on this issue on the NDT blog: http://www.ndt-educational.org/ckjblog-256-renaldenervationpolarviewsinnephrology.php