Pro: Sympathetic renal denervation in hypertension and in chronic kidney disease

In this Polar View, we will submit that the present more or less accepted indication for renal denervation (RDN), i.e. the so-called resistant hypertension, is incorrect. We will submit the idea to redefine patient groups likely to benefit from RDN based on the present knowledge of the pathophysiology. We will end by concluding that chronic kidney disease (CKD) patients could be one of these groups. In 2009, the first results on the effects of RDN on blood pressure became available. First but also subsequent results suggested a substantial blood pressure-lowering effect in so-called resistant hypertension patients, which is usually defined as sustained high blood pressure despite the use of three or more antihypertensive agents. The effect has been shown to be sustained for 2–3 years of follow-up. The rationale for and the concept of RDN are discussed in detail elsewhere [1–7]. Why was RDN introduced in resistant hypertension patients? At the time of introduction there were no data whatsoever on efficacy and safety. With that in mind, it was perfectly understandable and may have even been very wise to obtain the first results with this therapy in patients who were otherwise untreatable. However, the choice of this group of patients is not supported by any specific knowledge of the pathophysiology suggesting that these patients are especially likely to benefit. On the contrary, in a recent analysis we found that in a group of patients referred to us because of resistant hypertension, more than two-thirds of the patients did not meet the blood pressure criteria, or could be adequately treated with simple adjustments of dosages or had a secondary form of hypertension [8]. The fact that resistant hypertension patients represent a very mixed group of diagnoses was also seen in other studies. It is important to realize that RDN is meant to produce a very localized effect, i.e. disrupting nerve endings located within the renal artery wall. This concept is very different from pharmacological therapy where the intervention is ‘offered’ to the whole body. Therefore, it is especially important to address the question of which patients are likely to benefit. Keeping this in mind it seems logical to hypothesize that RDN is effective in disease conditions characterized by increased activity of afferent and/or efferent renal nerve activity. Consequently, the question arises in which patients are renal afferent and efferent nerves especially active (Figure 1). An important problem in this regard is that this nerve activity cannot be measured directly in humans. Whether increased efferent activity exists in humans as a primary abnormality, i.e. that due to a primary abnormality in the central nervous system (CNS) increased efferent activity exists, is unclear. There is convincing evidence that increased afferent activity results in hypertension in experimental conditions as well as in humans. The primary abnormality seems to be kidney injury and/or failure. There is convincing evidence that this includes CKD over the whole range of kidney function. We have discussed this elsewhere in detail [1–7]. Schematic representation of the kidney involvement of the function on renal afferent and efferent nerves. Minimal kidney damage, not necessarily affecting kidney function, results in area(s) of ischaemia. This results in increased afferent nerve activity to the CNS. Increased central sympathetic outflow affects many organs also including the kidneys and the cardiovascular system. Yes, there is experimental evidence to support this idea. Campese et al. performed a set of important studies. They showed in subtotally nephrectomised rats that blood pressure rapidly increases after surgery, which was abolished by afferent denervation (reviewed in [9]). Importantly, in a later study in rats they showed that a small lesion in one kidney by an intra-renal phenol injection, not affecting kidney function, increases sympathetic activity and leads to a long-term increase in noradrenaline secretion and to hypertension. These effects are also abolished by afferent denervation. Also in humans, there is substantial evidence for this idea. Early clinical observations showed that the presence of diseased kidneys could lead to hypertension characterized by high peripheral vascular resistance. Bilateral nephrectomy resulted in a substantial drop in blood pressure, which was caused by a decrease in peripheral vascular resistance. Muscle sympathetic nerve activity (MSNA) is increased in dialysis patients, whereas it is comparable with normal subjects in bilateral nephrectomised dialysis patients (reviewed in [1–7]). These data provide very convincing evidence that the diseased kidneys are critically involved in the pathogenesis of increased MSNA, which is considered to represent central sympathetic activity outflow towards the resistance vasculature. In subsequent studies, we and others showed that MSNA is not only increased in dialysis patients but already in patients with CKD not yet on dialysis. In a study in polycystic kidney disease patients we showed that MSNA was increased in hypertensive patients with normal kidney function, suggesting that it is kidney injury and not CKD per se that drives the MSNA. Further, unilateral nephrectomy for living kidney donation did not affect MSNA. We also found parallel shifts of renin activity and MSNA along changes in volume status in CKD patients. This clearly underscores the interrelationship between the RAAS and the sympathetic system, which can best be explained by a cause and effect relation or a common origin [9]. Finally, an interaction with the nitric oxide system was found [10]. Taken together, the available data seem to indicate that kidney injury/failure is associated with increased sympathetic activity, quantified by MSNA. In this pathophysiologic model the afferent renal nerves are of crucial importance. Kidney ischemia could be a central mechanism. The above-mentioned studies are discussed in great detail elsewhere [1–7]. Various methods of RDN have been applied in different models of CKD. Recently, we have reviewed this and concluded that most studies indicate various beneficial effects on kidney variables and CKD progression [7]. Also in humans, there is some indication to support this. For instance, Zocali et al. [11] found in dialysis patients that plasma noradrenaline is a predictor for mortality. We found a relation between level of sympathetic activity and left ventricular mass [12]. Furthermore, some studies suggest that the addition of sympatholytic therapy to present day standard therapy might be beneficial in CKD patients [13, 14]. If we accept the considerations mentioned above, it seems attractive to hypothesize that there could be an inverse relation between kidney function and the possible antihypertensive effect of RDN. In the first studies, this was not found. In contrast, we analysed patients shortly before RDN, without their antihypertensive medication, because many agents influence glomerular filtration rate (GFR). Indeed, we found an inverse relation (submitted). A recent analysis of the combined datasets of SYMPLICITY 1 and 2 also suggested a trend towards an association between lower estimated GFR (eGFR) and greater blood pressure-lowering effect (oral presentation during annual meeting of the American Society of Nephrology, Atlanta, November 2013, abstract SA-OR036). If we accept the pathophysiologic model mentioned above, it may also help us to identify patients likely to benefit. It is important to confirm the inverse relation between GFR and blood pressure-lowering potential of RDN. Other variables, indicating kidney injury or kidney ischaemia, could be predictors as well. It seems worth while to address such hypothesis. This knowledge of the pathophysiology is important for yet another reason. The presently available data show a substantial variability of the blood pressure lowering capacity of RDN. Apart from incorrect patient selection, also failure to perform an adequate intervention, i.e. destroy the nerves sufficiently, could be a reason for this. Presently, the intervention is a ‘black box’ procedure, i.e. as there are no methods to directly measure or visualize the effects of the intervention itself. In this respect, a recent report on the use of a stimulation catheter shortly before and after the procedure is of great interest [15]. Cardiac electrophysiologists make use of mapping catheters to localize electrical activity within the cardiac cavity. It would be interesting to find out if localizing (residual) autonomic nerve activity would be possible in the renal arteries with the use of this technique. Finally, it might be possible by using high spatial resolution (intravascular) imaging techniques such optical coherence tomography or magnetic resonance imaging to analyse the effect of an intervention. In light of the considerations above, it also becomes increasingly clear how little we know about the normal anatomy of the renal nerves. Very limited data are available on the location. Little is known on the type of nerves (efferent and afferent) and whether there are differences between the various disease conditions. We basically do not know what an ‘adequate’ denervation procedure really is and whether there are differences in this respect between devices. Whether the differences in design between the various devices are of any clinical relevance, both in terms of efficacy and safety, is at present totally unknown. At the time of writing this Polar View (early February 2014) the press release of Medtronic on the overall results of SYMPLICITY III was available. Details of the study were still unknown, so a detailed interpretation of the results is not possible at this time. But it seems safe to conclude that after the initial optimism on the potential role of RDN, reconsideration of the situation is necessary. We think that the concept of RDN has high potential. However, many issues remain unclear. Of paramount importance are (i) proper patient selection based on the knowledge of the pathophysiology, (ii) development of a method/variable that can be used to guide the interventionalist during the procedure and (iii) long-term data on safety are necessary. We foresee that gradually this concept of ‘resistant hypertension’ as selection criterion will disappear, simply because it does not make sense from a pathophysiological point of view. Based on the experimental and human data briefly outlined earlier, we submit again the hypothesis that CKD patients are especially likely to benefit from RDN, possibly over the whole range of CKD (Table 1) [1–7]. Present day standard therapy in CKD patients includes RAAS inhibitors, which reduce but on average do not normalize sympathetic activity [16]. Also, blood pressure is often not adequately controlled in CKD [17, 18]. The beneficial effect of renal denervation may not only include a reduction in cardiovascular morbidity and mortality, but also a reduction in progression of CKD. Given the financial burden to society of the ESRD programmes, any reduction of CKD progression could be very cost effective. It would be worth while to perform a clinical trial on the hypothesis that RDN when added to presently accepted standard antihypertensive therapy results in a reduction of important kidney and cardiovascular endpoints without relevant long-term side effects. If such (an) effect(s) would be found, then RDN most likely would be a (highly) cost effective addition to present day treatment options for this patient group. Other potential patients groups could be heart failure and obesity/metabolic syndrome/Type II diabetes. Further, also for patients with a kidney pain syndrome, this therapy could be of great relevance [19–20] (Table 1). Potential patient groups likely to benefit of renal denervation Potential patient groups likely to benefit of renal denervation It is clear to us that we have only reached ‘step one’ in the process of appropriately positioning this new therapy in the field and further development is a truly multidisciplinary task requiring collaboration between the device industry, interventionalists, pathologists, electrophysiologists, vascular internists, radiologists and nephrologists. Given the nature of this intervention, nephrologists should be heavily involved. (See related article by Zoccali and Mallamaci. Renal denervation: The jury is still out and the verdict will be more complex than initially envisaged. Nephrol Dial Transplant 2014; 29: 1124–1126; 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.) We read with interest the contribution of Prof. Blankestijn. In the first part, he states that renal denervation (RDN) is a promising technique but has been applied to the wrong patients. According to him, application of RDN to patients with resistant hypertension is not supported by solid pathophysiological arguments. Furthermore, these patients form a heterogeneous group and most of them either have secondary or spurious resistant hypertension, or are amenable to blood pressure control after skillful treatment adjustment. In this respect, his position is similar to ours. Some readers might even be surprised to find it in the ‘PRO’ side of this controversy. The second and longest part defends the idea that patients with chronic kidney disease (CKD) may be an ideal target population for RDN. The main arguments are (i) CKD is associated with sympathetic overactivity, both in patients and animal models; (ii) in patients with CKD, bilateral nephrectomy was associated with reduced blood pressure; (iii) in animal models of CKD, various methods of RDN and afferent denervation have been associated with beneficial effects on blood pressure, kidney variables and/or CKD progression. As discussed in our review, several elements make us less optimistic about the potential benefits of RDN in patients with CKD/altered renal function: In the SYMPLICITY reports, estimated glomerular filtration (eGFR) rate was not a predictor of blood pressure decrease after RDN; In the cohort of patients followed up within the ENCOReD network, a lower eGFR was associated with a decreased likelihood of BP improvement after RDN. In the context of resistant hypertension, altered renal function may be a hallmark of irreversible vascular damage, predicting a poor blood pressure response to RDN; So far, there is no evidence that RDN might slow down the progression of CKD or have renoprotective effects whatsoever. Notably, when Prof. Blankestijn writes that ‘some studies suggest that addition of sympatholytic therapy to present day standard therapy might be beneficial in CKD patients’, he refers to possible benefits of sympatholytic drugs, not of RDN. The results of Symplicity HTN-1 and EnligHTN I trials do not rule out and may even suggest a deleterious effect of RDN on renal function. Having this in mind, the effect of RDN on blood pressure and renal function in patients with CKD is certainly worth testing. The outcome may depend on age, duration of hypertension, systemic vascular damage and the etiology of renal disease. Alexandre Persu alexandre.persu@uclouvain.be