The complexity after simplicity: How to proceed with renal denervation in hypertension?

Recently the results of the Symplicity HTN-3 trial were presented and simultaneously published.1 In this study, patients with so called resistant hypertension were randomized to either catheter based renal denervation or a sham procedure. The main results showed no statistically significant difference in the effect of renal denervation on office blood pressure at 6 months follow up (primary efficacy endpoint) between the two arms. Clearly, based on the results of Symplicity HTN-1 and HTN-2 a different outcome of Symplicity HTN-3 was expected.2,3 So, it seems appropriate to discuss possible explanations. The main observations of Symplicity HTN-3 are that: 1) the effect on blood pressure is smaller than reported in Symplicity HTN-1 and HTN-2 and in various other renal denervation studies and 2) the effect of the sham procedure was much larger than the effect in the control arm of Symplicity HTN-2.3 Office systolic blood pressure decreased in Symplicity HTN-2 with 32/12 and 1/0 mmHg in the renal denervation versus the control group and in Symplicity HTN-3 with 14/6 and 12/5 mmHg in the renal denervation versus the sham procedure group. There are several considerations that need to be taken into account when interpreting these results. First, the previous findings may have been biased through placebo effects or by unblinded blood pressure measurements. After all, Symplicity HTN-2 was an open study, and a specific strength of Symplicity HTN-3 is the blinding of both patients and research personnel. The control group underwent a sham procedure, which included cannulation of the femoral artery and an angiogram. Although a decrease in blood pressure due a placebo effect in trials that study an effect of an antihypertensive intervention is well established, a mechanism of a sustained (6 months) blood pressure lowering effect specific to the angiographic procedure is difficult to define. Second, some other particular features of Symplicity HTN-3 regarding the choice of patients and the nature of the procedure may explain a smaller or absent effect of the intervention. Symplicity HTN-3 was done exclusively in the USA. Ethnicity differed in that 25–30% of the study population was of African American origin, whereas in previous studies almost exclusively Caucasians are included. Pre-specified subgroup analysis, as briefly mentioned in the rationale and design paper,4 showed that in non-African Americans the blood pressure decrease after renal denervation was significantly greater than in the sham group (change in office systolic blood pressure in the renal denervation group versus sham group: 15.2 vs. 8.6 mmHg, p = 0.01). In contrast, in the African American subgroup the effect on office systolic blood pressure was comparable in the two arms (resp. 15.5 versus 17.8 mmHg, ns). It is conceivable, but presently unknown, that renal denervation may have different effects in various ethnic groups. However, the most striking observation is that the decrease in blood pressure in the African Americans in the sham group is much greater than that in the non-African Americans. Symplicity HTN-3 does not provide an explanation for this finding. We can only speculate on possible explanations of this specific finding. Finally, in Symplicity HTN-3 a total of 112 interventionalists performed the 364 procedures, resulting in an average of 3.3 procedures per interventionalist. In fact, approximately 50% of interventionalists only performed one or two procedures, and only five interventionalists (4,5%) performed 10 or more procedures. While the authors argued that this is not relevant and cannot explain the results of the study, it would suggest that most interventionalists were well into the learning curve and denervations may have been suboptimal or incomplete in a significant number of patients. This also very much points to a major limitation of renal denervation procedures in general. Presently, the intervention is a “black box”, i.e. there are no methods to directly measure or visualize the effects of the intervention itself. Further, it becomes increasingly clear that we basically do not know what an “adequate” denervation procedure really means, whether there are differences in that respect between devices and whether these differences are of any clinical relevance, both in terms of efficacy and safety. The concept of renal denervation is supported by both experimental and early human evidence from the previous century when non-selective surgical sympathectomy procedures were done and found to have huge effects on blood pressure in some but not all patients (reviewed in 5–9). Also, the panellists who commented on the presentation during the Late Breaking trial session during the 2014 Meeting of the American College of Cardiology and the authors of the Editorial that accompanied the publication of the Symplicity HTN-3, concluded that the research on the role of renal denervation should continue.10 Next steps in the research in the renal denervation field should include a focus on patient selection. It is clear that this so called “resistant hypertension” cannot at all be seen as a distinct disease condition, which from a pathophysiologic point of view is likely to benefit from renal denervation. On the contrary, it represents a very mixed group of patients.11,12 It seems attractive to hypothesize that renal denervation is especially effective in patients likely to have active renal sympathetic nerves, such as in chronic kidney disease, heart failure or obesity/metabolic syndrome/type II diabetes.9,13 Secondly, it is important to realize that the interventions in the Symplicity HTN-3 study are all done with one specific device, i.e. the single polar Medtronic Symplicity device. There are data suggesting that with this device a highly variable degree of renal nerve destruction is obtained.14 This was also evidenced by the finding that the reduction in renal noradrenaline spillover showed an average of 47% (95% confidence interval 28–65%).2 Various devices of different design and with different techniques for renal denervation are now being developed. It is certainly possible, but presently unknown, that these differences are of clinical relevance. Further, a major difficulty in research in this type of patient is the question to what extent the patients use their prescribed medication. In Symplicity HTN-3, an average of five antihypertensives was prescribed. Compliance to antihypertensive medication is notoriously poor. Indeed, in a recent analysis in patients with “resistant hypertension”, it was found that low adherence to medication was the most common cause for poor blood pressure control.15 This underscores the fact that medication use is a major and until now poorly controlled confounder of any effect on blood pressure. In unblended studies differences in subsequent compliance could have led to major differences in attained blood pressure levels. The above considerations clearly show that much experimental and human research in this field is still necessary. Some next questions to be addressed are summarized in Table 1. Some open questions to be addressed in the renal denervation field. Some open questions to be addressed in the renal denervation field. The results from Symplicity HTN-3, the most rigorously controlled trial on renal denervation in hypertension to date, have been sobering. There is a need to rethink how to position renal denervation. History has learned that it is not uncommon that treatments initially fail but revive following subsequent insights. Shortly after the introduction of captopril, it was found that this compound may cause kidney failure. Currently ACE inhibition is one of the cornerstones of preventive treatment in a range of conditions. Also in device technology there are several examples of early disappointing results, for instance with drug eluting stents. The prospects of renal denervation are sufficiently promising to call for concerted research actions by all disciplines involved in close cooperation with the producers of the devices. But the true value and positioning of renal denervation is probably more complex than the simple and relatively straightforward procedure tested in the patients recruited for Symplicity HTN-3.