A small step forward for better patient selection in renal denervation

More than 15 years ago, renal denervation (RDN) emerged as a potential treatment option for patients with resistant hypertension [1]. However, despite technical and technological advancements, the nonresponder rate remains high. A significant challenge for those patients undergoing the procedure is the lack of a robust prediction model that would allow for better patient selection and reduce the rate of nonresponders. The variations between different renal denervation approaches make creating a standardized RDN method challenging. In the paper titled “Long-term prediction of blood pressure reduction after renal denervation for arterial hypertension”, the authors included 80 patients with resistant hypertension who underwent an ultrasound-based renal denervation and prospectively investigated the predictive value on of a model combining preprocedural 24 h BP monitoring and the ascending aortic distensibility measured via MRI (mAAD), at 6 and 12 months of follow up. The follow-up was performed with 24 h ambulatory BP monitoring. The results on BP reduction at six and twelve months were compatible with other published data [2]. The bivariate regression model combining baseline 24-h systolic BP and mAAD showed a high predictive accuracy for a reduction in BP at 6 months, with an area under the curve (AUC) of 0.819. This accuracy remained high at 12 months, although it slightly decreased to an AUC of 0.787. To date, this is the first prospective study to find predictive value in a prospective manner, which in the field of renal denervation is a step forward in choosing the optimal patients for the procedure. However, there are a few drawbacks. The first is the relatively small number of patients and the high rate of follow-up dropouts. At 12 months, 22 patients were lost to follow-up, which means that the data for more than a quarter of the initial population is missing, and in a study of that type, this can impede the generalizability of the results. Secondly, the authors report a high variability in patient therapy. At 6 months, 60.3% maintained stable medication regimens; at 12 months, no change in medication was reported in only 35.1%. And the main reason for change was not downtitration, but it was adding a new drug or uptitrating an existing drug. That sort of variability is the “Achilles’ heel” in most of the RDN trials. Since the effect of RDN is compatible with adding an extra drug class, this weakness makes the interpretation of the results even harder. However, the authors state that the subgroup analysis showed a similar trend independent of the medication stability, but this creates at least four subgroups- one with stable therapy, one with downtitration, one with uptitration and one with a new class of antihypertensive drug included. To prove a prediction model prospectively, probably the best approach is to design a trial similar to SPYRAL HTN-OFF MED [3], which will allow us to isolate the result from that huge real-world variability. Finally, the inclusion of the MRI-measured ascending aortic distensibility (mAAD) has the benefit of being noninvasive, but in terms of applicability and wide adoption as a predictive tool, it has a few drawbacks. The first is the burden of additional testing for the patient, albeit noninvasive. In the current study, a significant number of screen failures (16 from 22 screen failures) were associated with the MRI examination. Secondly, mAAD is not a readily available measure, and it will lead to a prolonged hospital stay or additional patient visits. Sagmeister et al. acknowledge the limitations mentioned above in the discussion and limitations section of the publication, and despite the constraints highlighted in this commentary, the study has contributed with additional knowledge regarding RDN predictors of success. Furthermore, the search for possible predictors for RDN response is additionally made more difficult by the variability in techniques and technology available for renal denervation. The current ESC guidelines [4] upgraded the RDN, and now it may be considered for patients with uncontrolled BP after shared decision-making, and the data presented in that publication will be an additional tool to use when choosing “the ideal” patient, although with limited applicability. ACKNOWLEDGEMENTS Conflicts of interest There are no conflicts of interest.