Introduction
Arterial hypertension is a major global health problem that affects 30–45 % of the general population [
1]. Cardiovascular complications are common in hypertensive patients. Thus, it is mandatory to achieve a fast and adequate blood pressure (BP) control. The first step in the antihypertensive treatment should always be lifestyle modification followed by an appropriate medical therapy [
2], including the use of single-pill combinations and the avoidance of interfering substances [
3]. However, an increasing number of especially elderly patients do not respond to antihypertensive therapy in an adequate manner [
4]. Resistant hypertension is diagnosed when BP levels remain above 140/90 mmHg despite the use of three antihypertensive drugs at optimal dosages, including a diuretic [
1]. As a new invasive treatment for resistant hypertension, the successful use of catheter-based renal denervation (RDN) was first described in the two landmark studies HTN-1 and HTN-2 [
5,
6]. In both studies significant reduction of office-based BP following bilateral RDN was demonstrated with an on-going treatment effect up to three years of follow-up [
7,
8]. In 2014 the first blinded randomised controlled RDN trial with the large number of 535 patients was published. In contrast to the HTN-1 and HTN-2 trials it did not show differences in BP lowering between RDN and a sham-control procedure [
9]. This study has raised important issues, especially by providing evidence for the existence of a placebo effect and raising suspicion for improvement of medication adherence among study patients [
9]. However, despite the major impact of this large randomised sham-controlled trial, it had some important methodological issues. More specifically, some of the operators were inexperienced in the field of RDN (with only three or fewer RDN procedures). Further, the study population comprised 25 % Afro-Americans who are known to respond better to sham procedures than to RDN. Finally, the antihypertensive drug treatment of the patients was not truly stable [
10].
To answer the question whether RDN is effective or not in BP lowering in patients with truly resistant hypertension, more well-conducted studies with longer follow-up are needed. For this purpose, in comparison with office-based BP, the 24‑h ambulatory BP (ABP) measurement is a more objective method to identify resistant hypertensive patients. In addition, the 24‑h ABP is a valuable predictor of adverse outcome in hypertensive patients [
11].
In this paper, we present 1/3/6/12 month follow-up data of an observational, non-controlled study (before – after single group study) by using a standard radiofrequency (RF) ablation catheter for RDN. In order to treat only truly resistant patients, we consequently measured the 24‑h ABP for a selection of eligible patients and also to evaluate the BP changes during follow-up.
Discussion
To our knowledge, this is the study with the largest number of patients in whom RDN was performed with standard RF ablation catheter with to date the longest follow-up. In contrast to most of the currently available RDN studies, we exclusively used 24‑h ABP measurement for initial diagnosis and during follow-up. Using this strict setting, we observed excellent results in BP reduction with a high number of BP responders following the RDN procedure, with a very low rate of renal and cardiovascular complications. No periprocedural complications occurred.
Recently, the “real-world” data of the Global SYMPLICITY Registry, a prospective, open-label, multicentre registry with 998 patients, were published. In this registry, a reduction of −11.6 mmHg in systolic office BP and −6.6 mmHg in systolic 24‑h ABP was demonstrated at 6 months [
14]. This BP reduction, although significant, was smaller than that observed in the HTN-1 and HTN-2 trials [
5,
6]. Except that fewer patients were included in our study, the 24‑h ABP reduction observed was higher than described in the Global SYMPLICITY Registry [
14]. Also the BP responder rate in our study at 6‑ and 12-month follow-up was higher (systolic 24‑h ABP reduction >10 mmHg at 6/12 months: 70 %/64 %; >5 mmHg: 77 %/70 %) in comparison with the Global SYMPLICITY Registry (systolic 24‑h ABP reduction at 6 months >8 mmHg: 48 %; >5 mmHg: 68 %) [
14]. In 2015, two randomised controlled RDN trials (DENERHTN and Prague-15 study) were published [
15,
16]. In the DENERHTN study 101 patients were randomised to receive either standardised stepped-care antihypertensive treatment alone (baseline mean systolic 24‑h ABP 156 mmHg) or in combination with RDN (baseline mean systolic 24‑h ABP 160 mmHg) [
15]. In this study, the observed reduction in systolic 24‑h ABP following RDN with the Symplicity catheter (−15.8 mmHg at 6‑month follow-up) was comparable with our results (−18 mmHg). However, in the standardised hypertensive treatment group the 24‑h ABP reduction was −9.9 mmHg, amounting to an additional systolic BP reduction of −5.9 mmHg in the RDN group [
15]. The responder rate in systolic 24‑h ABP was not described in the DENERHTN study [
15]. In contrast, the Prague 15-study with 106 randomised patients found no significant differences in the 24‑h ABP between intensive pharmacological treatment and the RDN group. However, after 6 months, the number of antihypertensive drugs and the serum creatinine level was higher in the pharmacological treatment group [
16]. Keeping in mind that the baseline patient characteristics were similar in the afore-mentioned studies, some important differences might explain the better BP response in our and in the DENERHTN study. First, the mean systolic 24‑h ABP was 165 mmHg in our and 159 mmHg in the DENERHTN study, whereas in the Global SYMPLICITY Registry the baseline mean systolic 24‑h ABP was 151 mmHg [
14]. As described previously and confirmed in our and in the DENERHTN study, the baseline systolic office BP, or mean baseline systolic 24‑h ABP, are to date the only known independent predictors of BP response following RDN [
15,
17,
18].
Second, 41 % of the operators in the Global Symplicity registry performed fewer than 15 RDN procedures with the Symplicity catheter. In contrast, in our single-centre study two experienced electrophysiologists performed all the procedures (>30 RDN procedures per operator). Therefore, the operators in our study might be more familiar with the handling of the RF ablation catheter. However, in the DENERHTN study with similar BP responses, the number of RDN procedures per operator was also below 10 cases [
15].
Third, we used a 7‑F 4‑mm tip standard RF ablation catheter that might provide a higher contact pressure, higher tip stability and greater torque control during ablation in comparison with the first-generation Simplicity catheter. Animal studies conducted by our group found that the RF ablation for RDN with a standard RF ablation catheter can be performed without harmful sequelae. Further, we observed an attenuation of neurofilament expression in the renal artery as a surrogate marker of effective RDN [
19]. The performance of these standard RF ablation catheters is non-traumatic as shown in our study (no acute complications in the renal artery), and also during RF ablation of the coronary venous system and the aortic sinus for ablation of ventricular arrhythmias [
20]. In comparison to the results of the DENERHTN study and the Global Symplicity registry, RDN using a 7‑F 4‑mm tip standard RF ablation catheter seems to be at least as effective as RDN with the first-generation Simplicity catheter.
According to our data and to data in the literature, there is a risk of the development of renal artery stenosis following RDN. We observed two new renal artery stenoses (2.7 %) in our patient sample, detected by duplex sonography of the renal artery. However, in one of these patients (with otherwise uncontrollable resistant hypertension) the stenosis was an in-stent-restenosis after renal artery stenting that was performed years ago. In contrast, renal artery stenoses were reported to range between 0.3–1.9 % in the Symplicity HTN-1, -2 and -3 trials [
8,
21]. No renal artery stenosis was reported in the Global SYMPLICITY Registry and the DENERHTN study. However, one could speculate that renal artery stenosis might be overlooked in the Global SYMPLICITY Registry since, as in our study, “there was no mandatory renal artery imaging follow-up” [
14] and, especially in overweight patients, duplex sonography is considered an inferior imaging modality to magnetic resonance angiography or angio-CT. In the EnligHTN 1 study (
n = 46) renal artery stenosis occurred in 4.3 % of the treated patients [
22]. In all these cases this resolved after treatment with renal artery stenting, without further complications.
The effect of RDN between the office BP and the 24‑h ABP measurements differed across the known RDN studies, reviewed by Verloop et al. and Bunte [
23,
24]. One reason for this might be that the 24‑h ABP was not the primary study endpoint (HTN-1 and 2) [
7,
8] or the required level for study inclusion was rather low (systolic 24‑h ABP >135 mmHg in HTN-3 and Global SYMPLICITY Registry) [
9,
14]. However, the 24‑h ABP measurement performed in the patient’s own environment provides more objective results than the office-based BP [
25]. In our opinion, 24‑h ABP measurement with BP values that are beyond 135/90 mmHg as an inclusion criterion should be used for further RDN trials since a systolic 24‑h ABP of 135 mmHg might be too close to the normal range. The weaker criterion of office-based BP should only be used as a sub-criterion.
There are some limitations to our study worth noting. One general concern raised by the negative results of the first randomised sham-controlled study (HTN-3) [
9] is whether observation studies can still be deemed useful for evaluating the efficacy of the percutaneous RDN procedure. As a matter of fact, randomised controlled trials are the gold standard in testing efficacy of medical interventions [
26]. Observational studies can be an alternative, however, when randomised controlled trials are difficult to perform [
27]. As a practical alternative to doing nothing, well-designed prospective observational studies might be used to obtain more information on treatment effect and long-term outcome [
28]. The HTN-3 trial had some methodological deficiencies, mentioned above, and there is evidence from other randomised controlled trials and also from many observational studies that RDN is effective and safe [
6,
8,
14,
17,
22]. The sample size in our study with 70 patients analysed is rather small. As we used 24‑h ABP measurement consequently, it is to date the largest number and longest follow-up of patients in whom RDN was performed by using a standard RF catheter. Because of the observational design of our study, we cannot rule out that the Hawthorne effect (a phenomenon that individuals modify their behaviour if they are aware they are being observed) [
29] might contribute to the BP reduction observed. Another problem of observational studies is the regression to the mean (Extreme outliers on the first measurement tend to be less extreme and closer to the mean on the second measurement. It can make natural variation in repeated data assessments look like a real change.) [
30]. However, all of the patients included in our study had a long history of resistant hypertension with multiple previous clinical consultations and examinations, BP measurements, intensive review, and adaption of medication without significant effects on 24‑h ABP. Furthermore, the patients had to be on a stable antihypertensive treatment for at least 4 weeks before inclusion in our study.
In conclusion, as shown during long-term follow-up our results support the hypothesis that RDN with a standard RF ablation catheter can be used safely to substantially reduce mean 24‑h ABP in resistant hypertension, with an acceptable risk for the development of renal artery stenosis. Thus, our findings support and extend the results of previous studies that RDN is effective in BP reduction. In our study, only the mean systolic 24‑h ABP at baseline was a predictor for responders and should therefore be used as the inclusion criterion for further RDN studies. However, for a final judgement of the efficacy of RDN, randomised trials with long follow-ups are needed.