Introduction

Despite continuing advances in pharmacologic therapy, hypertension remains a significant health risk to the population worldwide [1] Blood pressure (BP) control is especially important in patients with diabetes mellitus, among whom hypertension is more common [2] and BP reduction has been shown to lower cardiovascular events [3]. The results from recent randomized, sham-controlled clinical trials suggest that renal denervation (RDN) can safely and effectively lower BP in the absence or presence of antihypertensive medical therapy [4,5,6]. Although these new data are highly encouraging, additional long-term safety and efficacy results of RDN are needed. Specifically, there are limited reports on long-term follow-up after RDN in Asian patients [7, 8].

The ongoing Global SYMPLICITY Registry (GSR) has enrolled over 2700 patients from 45 countries and 196 centers, and most patients have been followed for 3 years. Within the GSR, the effects of RDN on an Asian population have been evaluated in a Korean substudy of GSR (GSR Korea). The aim of the current report is to evaluate long-term outcomes after RDN for patients in GSR Korea and specifically, compare outcomes in patients with diabetes mellitus.

Methods

The study design of the GSR (www.clinicaltrials.gov, NCT01534299) has been previously published [9, 10]. Briefly, GSR is a prospective all-comer registry to evaluate the safety and efficacy of RDN in a real-world population. Patients with uncontrolled hypertension or other conditions characterized by increased sympathetic activity were enrolled. There are currently 196 active centers in Canada, Europe, Latin America, South Africa, the Middle East, Australia, New Zealand and Asia, including South Korea. The trial complied with the Declaration of Helsinki, locally appointed ethics committees approved the clinical protocol at each enrolling center, and informed consent was obtained from all patients.

GSR Korea is a substudy of GSR with additional inclusion and exclusion criteria compared to GSR [10], including the inclusion criteria of office systolic blood pressure (OSBP) ≥160 mmHg, or ≥150 mmHg for type 2 diabetes patients, while receiving 3 or more antihypertensive medications without changes for 2 weeks prior to enrollment and a recommended baseline estimated glomerular filtration rate (eGFR) >45 ml/min/1.73 m2. General exclusion criteria for GSR Korea are possible secondary hypertension, type 1 diabetes mellitus, stenotic valvular heart disease, myocardial infarction, unstable angina, or cerebrovascular accident within 6 months of enrollment. Anatomic exclusion criteria for GSR Korea were prior renal artery intervention, main renal arteries <4 mm in diameter or <20 mm in length, or significant renal artery abnormalities. The patients described in this analysis were treated with the Symplicity Flex catheter (Medtronic, Santa Rosa, CA) for the RDN procedure. The renal denervation procedure was performed according to the instructions in the “Instructions for Use” and according to local hospital requirements. Briefly, radiofrequency ablations lasting up to 2 min each and of 8 watts or less separated both longitudinally and rotationally within each renal artery were performed. Ablations were focused on the main renal arteries, and multiple ablations were allowed if the size and length of the renal artery were within the treatment range.

Patients were followed up for up to 36 months postprocedure. At each follow-up, office BP and changes in prescribed antihypertensive medications were recorded. Office BP was recommended to be measured as the average of three measurements with the patient sitting quietly for at least 5 min before the first measurement and with 1 min between each reading.

Continuous data are summarized as the mean ± standard deviation. Categorical data are summarized as counts and percentages. Antihypertensive medication usage was compared between baseline and 6 months using a paired t-test and McNemar’s test. Comparisons of OSBP and office diastolic BP (ODBP) changes between type II diabetes patients and nondiabetic patients in GSR Korea were compared using analysis of covariance (ANCOVA) models adjusting for baseline blood pressure measurements. Kaplan–Meier time-to-event estimates were used to summarize rates of adverse events based on all available follow-ups. There were no adjustments for multiplicity. All statistical analyses were performed using SAS version 9.2 or higher (SAS Institute, Cary, NC, USA).

Results

The GSR Korea cohort included 102 patients (Supplementary Figure 1), and baseline characteristics are reported in Table 1. The mean age of the cohort was 54.9 ± 14.1 years, and 69.6% was male. Approximately half (46.1%) of the cohort had type 2 diabetes, 7.8% had a previous myocardial infarction, and 19.6% had undergone a previous percutaneous coronary intervention. Baseline OSBP was 169.8 ± 15.5 mmHg, and ODBP was 97.0 ± 14.3 mmHg. The mean follow-up was 1045.5 ± 203.5 days.

Table 1 Baseline characteristics for the GSR Korea cohort
Full size table

The use of antihypertensive medications at baseline and 36 months is reported in Table 2. The mean number of antihypertensive medication classes used at baseline (3.8 ± 1.0) was similar to the number at 36 months (3.8 ± 1.2). The use of individual antihypertensive medications was also similar between timepoints, except for diuretics, of which fewer were prescribed at 36 months, and direct-acting vasodilators, of which more were prescribed at 36 months.

Table 2 Prescribed antihypertensive medications in the GSR Korea cohort at baseline and 36 months
Full size table

Mean OSBP was significantly reduced between baseline and 6, 12, 24, and 36 months (p < 0.001 at all timepoints; Fig. 1). The majority of patients showed improvement in OSBP from baseline (Fig. 2A). Greater than 50% of patients had OSBP < 140 mmHg at 12, 24, and 36 months (Fig. 2B), while the proportion of patients with persistent OSBP ≥ 180 mmHg was low and steadily decreased at each successive timepoint (5.3% at 12 months, 3.4% at 24 months, and 1.3% at 36 months).

Fig. 1
figure 1

Changes in office systolic blood pressure over 36 months for the GSR Korea cohort at baseline and at 6, 12, 24, and 36 months. Error bars represent standard error. GSR Global SYMPLICTY Registry, SBP office systolic blood pressure

Full size image
Fig. 2
figure 2

A Proportion of patients with ≥10 mmHg reduction, ≥20 mmHg reduction, and no reduction in office systolic blood pressure after renal denervation and B distribution of office systolic blood pressure from baseline to 36 months for the GSR Korea cohort. GSR Global SYMPLICTY Registry

Full size image

Adverse events at 36 months were rare among patients with BP measurements at all follow-up points (Table 3). There was one death of unknown cause.

Table 3 Adverse events in the GSR Korea cohort from baseline to 36 months using Kaplan–Meier time-to-event analysis
Full size table

Analysis of GSR Korea patients with (N = 47) and without (N = 55) diabetes mellitus was conducted, and baseline characteristics are compared in Supplementary Table 1. More diabetic patients had a history of myocardial infarction and previous percutaneous coronary intervention, but other baseline characteristics were similar between the two groups. Diabetic patients had a lower baseline OSBP (166.2 ± 15.1 vs. 172.8 ± 15.3 mmHg, p = 0.002) and ODBP (93.5 ± 11.5 vs. 99.9 ± 15.9 mmHg, p = 0.022) than nondiabetic patients.

Reductions in OSBP and ODBP from baseline to 36 months were similar for diabetic and nondiabetic GSR Korea patients (p > 0.05 at all timepoints), as seen in Fig. 3. Furthermore, there were no significant differences between diabetic patients and nondiabetic patients for adverse events that occurred between baseline and 36 months (Supplementary Table 2). There were no myocardial infarctions, renal artery reinterventions, renal artery stenoses >70%, or vascular complications in either group out to 36 months. Overall, there were two strokes (both in the nondiabetic group) and one death (in the diabetic group, unknown cause).

Fig. 3
figure 3

Reduction in office systolic and diastolic blood pressures over 36 months for diabetic and nondiabetic patients in the GSR Korea cohort. P-values were calculated using the ANCOVA model, adjusting for baseline systolic and diastolic BP. Error bars represent standard error. BP blood pressure, DM diabetes mellitus, GSR Global SYMPLICTY Registry

Full size image

Discussion

In this report, BP responses after RDN were effective and consistent through 36 months in GSR Korea. When reviewing the distribution of OSBP after RDN in GSR Korea, the proportion of patients with OSBP < 140 mmHg was 54.5% at 12 months and remained consistent up to 24 and 36 months. Importantly, the proportion of patients with the highest BP (OSBP ≥ 180 mmHg) progressively decreased from 5.3% at 12 months to 1.3% at 36 months.

In the GSR study, for patients with severe resistant hypertension and a mean baseline OSBP of 179 ± 17 mmHg, the OSBP reduction at 3 years was −26.7 mmHg [11]. Similarly, the RDN trials SYMPLICITY HTN-1 [12] and HTN-2 [13] showed sustained OSBP decreases out to 3 years of −32.0 ± 17.6 and −32.7 ± 24.1 mmHg, respectively, compared to −32.5 ± 18.8 mmHg in GSR Korea at 3 years. However, OSBP reduction at 3 years for GSR overall was reported to be −16.5 ± 28.5 mmHg [14]. HTN-1, HTN-2, and GSR Korea all specified inclusion criteria of baseline OSBP ≥ 160 mmHg despite 3 or more antihypertensive medications. Whether the greater BP reductions in GSR Korea were due to differences in inclusion criteria, medication burden, baseline demographics or ethnicity is unknown. Other possible explanations for the higher OSBP reduction for these patients with high baseline BP could be higher baseline sympathetic activity or regression to the mean [15].

It has been suggested that Asian populations may respond better to RDN due to unique mechanisms of sympathetically mediated hypertension that differ from other ethnicities [16]. In a cohort of GSR patients in the substudy GSR-Taiwan, patients treated with the Symplicity Flex catheter (n = 8) experienced a mean OSBP reduction at 3 years of −29.7 ± 25.9 mmHg [8]. Enrollment in SYMPLICITY HTN-Japan ended prematurely after the publication of SYMPLICITY HTN-3 results, but of the 22 patients followed for 3 years, OSBP reduction was −32.8 ± 20.1 mmHg [7]. We previously reported differences in the degree and timing of BP reduction between the GSR Korea cohort and a matched GSR cohort of Caucasian ethnicity [10]. Variations in time courses may be explained by the existence of multiple mechanisms of action in RDN, which could occur at different rates for people of different ethnicities. For example, there is no clear consensus as to whether RDN primarily ablates the afferent or efferent renal nerves, but the possibility exists that both types of renal nerves are targeted with varying rates of response and results for different individuals, including ethnic differences [17]. Efferent-dependent mechanisms may lead to greater short-term BP reductions due to diuresis and natriuresis, whereas afferent effects may take longer to manifest as peripheral vascular remodeling and lower total peripheral resistance occurs over a longer period of time [18, 19].

For antihypertensive medication burden, patients in GSR Korea were prescribed a similar mean number of antihypertensive medication classes at baseline and 36 months, even while long-term BP reduction was maintained up to 3 years after RDN. The roles and mechanisms of the specific classes of antihypertensive medications used and their impact on BP reduction after RDN are not well defined. In addition, the strategy of down- or up-titration of antihypertensive medications is not well established and remains an area of future investigation. Other areas of future investigation suggested by consensus statements from Kario [20] and Wang [21] on the state of RDN in Asian patients include the importance of identifying predictors of response to RDN as well as optimization of the RDN procedure.

Several studies have reported the BP-lowering effects of RDN in diabetic patients [19, 22,23,24,25]. Some studies also found improved glucose metabolism or beneficial effects on insulin sensitivity after RDN [19, 22, 24]. However, the Denervation of the Renal Arteries in Metabolic Syndrome (DREAMS) study, which evaluated the effects of RDN on patients with metabolic syndrome, did not show improvements in insulin sensitivity 12 months after treatment, although a significant reduction in mean 24-h BP was measured [25]. Our results demonstrate similar reductions in OSBP and ODBP for diabetic and nondiabetic patients from 6 to 36 months after RDN. Similarly, diabetic and nondiabetic patients had similar reductions in 24-h and office SBP in the overall GSR cohort, as reported by Mahfoud et al. [14]. Laboratory values were only measured in a small number of patients at follow-up in GSR Korea, preventing evaluation of glucose metabolism and insulin sensitivity in diabetic patients. None of the aforementioned clinical studies enrolled more than 50 diabetic patients, highlighting the need for larger randomized controlled trials to evaluate the effects of renal denervation on this population.

Diabetic patients are known to be at higher risk for acute myocardial infarction and stroke and increased mortality risk [26, 27]. However, in this GSR Korea post hoc analysis, there were no significant differences in the occurrence of adverse events between diabetic patients and nondiabetic patients at 36 months. Whether RDN can reduce mortality and cardiovascular events has not been evaluated in a large randomized clinical trial. There were no vascular complications, stenoses, or reinterventions in either group, suggesting that renal denervation is equally safe for diabetic and nondiabetic patients.

GSR is a registry without a sham control group, so the impact of a possible placebo effect could not be assessed. Patients in this analysis were treated with the Symplicity Flex catheter rather than the newer Symplicity Spyral catheter. Medication adherence was not objectively tested, so the effect of changes in antihypertensive medications could not be fully evaluated. Because ambulatory BP measurements were not mandated in GSR Korea, this report is limited to office BP measurements. A further limitation is that office BP measurements were only obtained from three-quarters of the patients at 3 years; however, the mean follow-up for all patients was 2.86 years. Finally, the comparison of diabetic vs. nondiabetic patients in GSR Korea was a post hoc analysis without a control group, and larger randomized controlled trials are warranted to further analyze the safety and efficacy of RDN in diabetic patients.

Conclusions

A sustained reduction in OSBP was observed up to 36 months post-RDN without safety concerns in a cohort of Korean patients from the Global SYMPLICITY Registry. Future clinical studies are needed to evaluate how differences in ethnicity and/or baseline characteristics, including medications used, contribute to the BP-reducing effect of RDN. In addition, the efficacy of RDN in diabetic and nondiabetic GSR Korea patients was similar up to 36 months.