After obtaining approval from the local ethics committee, we identified and contacted 245 patients who were previously enrolled in a local renal denervation (RDN) registry and who underwent radiofrequency RDN with the Symplicity Flex® renal denervation system at the University Hospital Halle (Saale), Germany, between 2011 and 2014 (www.drks.de; identifier: DRKS00004173). All patients underwent baseline evaluation through 24 h ambulatory blood pressure (ABP) measurement and laboratory tests prior to renal denervation. A total of 108 patients were available for long-term follow-up (FU). For 72 of these patients, complete 24 h ABP measurements were available. Patients who declined an on-site visit and ABP measurement (ABPM) were asked to take a telephonic interview and undergo ABPM and laboratory tests through their primary care physician. This study was approved by the local ethics committee.

All patients provided informed consent to this study.

Technical and procedural details of ablation systems have been described elsewhere. (15). The procedure was performed by a single experienced operator (AP), who used the Symplicity Flex® renal denervation system by following the instructions for use and recommendations of the device manufacturer. Accessory renal arteries were treated if the length and diameter were found suitable.

For on-site FU of patients, 24 h ABP readings were done using standardized techniques and validated equipment (Mobil-o-Graph®, AMEDTEC GmbH, Germany) according to guideline recommendations (3, 16). The equipment was applied on-site and patients were instructed to leave the system in place for the purpose of measuring a full day–night cycle. ABP and heart rate were measured in intervals of 20 min from 6 a.m. to 10 p.m. and in intervals of 30 min from 10 p.m. to 6 a.m. Electrocardiogram was recorded prior to the use of the ABP equipment.

Blood samples were drawn to determine the electrolyte, creatinine, urea, and HbA1c levels.

Spot urine tests for determining the albumin and creatinine levels were completed during FU of the original registry up to 12 months of FU.

Patients who were not willing or able to complete an on-site visit at long-term FU were asked to have the relevant laboratory tests, electrocardiogram, and ABPMs performed by their primary care physician. Antihypertensive medication was recorded and divided into nine classes [renin–angiotensin–aldosterone system-inhibitors (ACE inhibitors, angiotensin receptor antagonists, and renin inhibitors), calcium channel blockers, beta blockers, diuretics, mineralocorticoid-receptor antagonists, alpha-adrenergic blockers, centrally acting sympatholytic drugs, direct-acting vasodilators, and other medications].

The primary efficacy endpoint was the change in mean systolic ABP at long-term FU. The safety endpoints were the change in renal function [the estimated glomerular filtration rate (eGFR) was calculated by using the CKD-EPI equation] and new-onset renal artery stenosis.

Early response to RDN treatment was defined as a reduction in mean systolic ABP ≥ 5 mmHg at a 3-month FU.

Continuous symmetrically distributed variables are presented as means ± standard deviation and confidence intervals. Between-group differences were compared using the t-test. The median and the 25% and 75% quartiles were calculated to describe skewed variables. The between-group differences of these variables were compared using the Mann–Whitney test. Normal distribution was tested using the Kolmogorov‒Smirnov test. Between-group differences in categorical variables were compared using the χ² test. Categorical variables between baseline and long-term FU were compared by using McNemar's test. Urine albumin levels between baseline and FU were compared by using the signed-rank test.

To evaluate the changes in systolic and diastolic ABP, the glomerular filtration rate during FU and linear mixed-effects models were used to account for repeated observations in the same patients over time.

To identify independent correlates of the early response to RDN treatment (as defined above), the following baseline characteristics were assessed in a one-step multivariate binary logistic regression analysis: age, male gender, BMI, number of ablations, systolic ABP, diastolic ABP, mean heart rate, coronary artery disease, atrial fibrillation, current smoking, diabetes mellitus, number of prescribed antihypertensive medications, eGFR, BNP > ULN, and HbA1c.

In case of missing observations in any of the recorded variables, all other available follow-up observations were included in the calculations. No imputation of missing data was done. All endpoints were analyzed exploratively.

A score of P ≤ 0.05 was considered statistically significant. Statistical analyses were performed by using SPSS Version 28 (IBM, Armonk, USA), and GraphPad Prism version 9 (GraphPad Software, San Diego, California, USA).

From a total of 245 potential participants, data for 108 patients were available for this analysis. A total of 85 patients were lost to FU, 15 declined participation, and 37 were deceased (Figure 1). Of the 108 patients, 70 opted for on-site visits and 38 for remote FU. ABPM recordings were available for 72 patients. The main reason for patients declining participation or on-site visit was related to long travel distances or immobility. In addition, long-term FUs took place during the COVID-19 pandemic, and therefore, contact restrictions or risk of infection were additional reasons mentioned for declining participation.

At the time when the procedure was carried out, the patients were 63.2 ± 8.8 years old, and 60 of them (55.6%) were male.

The patients were mildly obese with a BMI of 30.9 ± 5.1 kg/m², which increased slightly but significantly to 31.6 ± 5.4 during long-term FU. Atrial fibrillation (paroxysmal, persistent, long-standing persistent, or permanent) was noticed in 11.1% of patients at baseline and 35.2% at long-term FU (P < 0.01). The proportion of patients with chronic kidney disease (eGFR < 60 ml/min/1,73 m²) increased from 10.2% to 33.3% (P < 0.01).

Of the patients, 13.9% had known coronary artery disease. Diabetes mellitus was present in 43.5% of patients. The patients were treated with 12.4 ± 3.1 ablations. The baseline characteristics are given in Table 1.

The median time to FU was 9.3 (IQR: 8.5–10.1) years.

Both mean systolic and mean diastolic blood pressure of the overall cohort significantly decreased from baseline to long-term FU (systolic ABP from 150.1 ± 16.9 mmHg to 138.3 ± 16.5 mmHg, P < 0.01, and diastolic ABP from 86.1 ± 12.0 to 77.1 ± 11.1, P < 0.01).

Changes in systolic and diastolic blood pressure over time of patients whose ABPMs were complete at baseline and FU are shown in Figure 2. At 3 months of FU, mean systolic ABP significantly reduced by −5.4 mmHg to 143.2 ± 16.5 mmHg and diastolic ABP significantly declined by −3.6 mmHg to 81.9 ± 12.3 mmHg (P < 0.05 for both). Reduced blood pressure was also evident at all subsequent FU time points (143.3/81.5 ± 15.2/10.2 mmHg at 6 months, 138.7/80.7 ± 13.2/10.4 mmHg at 12 months, and 138.3/77.1 ± 16.5/10.1 mmHg at long-term FU, P < 0.05 for all comparisons vs. baseline, and comparison of diastolic ABP at long-term FU vs. 12 months, except for systolic ABP at 6 months compared with baseline).

A total of 38 of 73 patients (52.1%), whose blood pressure readings were available at baseline and at 3-month FU, responded early to RDN treatment (as defined by a reduction in mean systolic blood pressure ≥5 mmHg). When early responders at the 3-month FU were compared with non-responders (n = 35, 47.9%), with respect to baseline characteristics, no significant differences were found except for baseline systolic ABP, which was significantly higher in the early responder group (154.8 ± 17.3 vs. 143.8 ± 15.0, P < 0.01, Supplementary Table S1), and for the proportion of dippers (defined as a decrease in nighttime systolic blood pressure by ≥10% of daytime systolic blood pressure), which was significantly lower in the group of early responders (23.7% vs. 48.6%, P = 0.03).

In the multivariate logistic regression analysis to identify independent correlates of early treatment response (as defined above), it was found that age (Exp(B) 0.85, 95% CI: 0.73–0.99, P = 0.03), systolic ABP (Exp(B) 1.09, 95% CI: 1.02–1.17, P = 0.01), mean heart rate (Exp(B) 0.92, 95% CI: 0.85–1.00, P = 0.04), and dipping (Exp(B) 0.22, 95% CI: 0.05–0.96, P = 0.04; Supplementary Table S2) were independent predictors of the early response to RDN treatment.

When applying the model to determine whether there was a clinically meaningful reduction in systolic ABP of 10 mmHg during the long-term FU, it was found that only systolic ABP was an independent predictor of the long-term treatment response (Exp(B) 1.11, 95% CI: 1.02–1.21, P = 0.02; Supplementary Table S3).

In the comparison of early responders and non-responders, a significantly pronounced reduction in blood pressure was already observed at the 3-month FU in the early responder group, whereas in the non-responder group, a gradual reduction over the follow-up period was observed after an initial increase in mean systolic and diastolic blood pressure. While systolic ABP was significantly higher in the early responder group at baseline, there was no significant difference between the two groups at long-term FU (Figures 3A,B).

During the implementation of the RDN procedure, it was found that the patients were taking an average of 5.4 ± 1.5 antihypertensive agents. During long-term FU, the number of antihypertensive agents being consumed decreased by 0.6 to –4.8 ± 1.6 compared with baseline (P < 0.01). When comparing patients classified as early responders at 3 months with non-responders, it was found that a slightly but not a significantly higher number of antihypertensive agents was being consumed by non-responders during FU. Between baseline and long-term FU, a non-significant decrease in the number of antihypertensive agents being consumed could be observed in both groups (Supplementary Figure S1).

The medication prescribed at baseline and long-term FU is shown in Table 2. While the proportion of RAAS blockers, calcium antagonists, beta blockers, and diuretics remained constant, the proportion of centrally acting sympatholytic drugs (53.7% vs. 34.4%, P < 0.01) and other antihypertensive medication (59.3% vs. 45.8%, P = 0.02) decreased significantly. In contrast, the low proportion of MRA at baseline increased significantly at long-term FU (9.3% vs. 29.6%, P < 0.01). Patients treated with MRA tended to have lower blood pressure than patients without this therapy; however, this difference was not statistically significant [baseline systolic ABP: 144 (±16) mmHg vs. 151 (±17) mmHg, P = 0.29, baseline diastolic ABP: 90 (±10) mmHg vs. 86 (±12) mmHg, P = 0.37; long-term FU systolic ABP: 133 (±16) mmHg vs. 141 (±16) mmHg, P = 0.09, long-term FU diastolic ABP: 75 (±11) mmHg vs. 78 (±11) mmHg, P = 0.36].

The renal function of the patients was determined by measuring their plasma creatinine levels. In addition, urine albumin values were available for the first 12 months of FU. At baseline, most patients (89.8%) had a normal renal function (estimated GFR ≥ 60 ml/min/1.73 m²) as calculated through the CKD-EPI formula using plasma creatinine with a median eGFR of 87.8 (IQR: 81.0–100.0) ml/min/1.73 m², which remained stable over the FU period up to 12 months. Similarly, the estimated GFR in patients with a baseline eGFR < 60 ml/min/1.73 m² and a median eGFR of 56.0 (IQR: 40.9–58.4) ml/min/1.73 m² was stable up to 12 months after RDN. A decrease in the eGFR was observed at the long-term FU in both groups [72.5 (IQR: 55.8–86.6) ml/min/1.73 m², P < 0.01] compared with that at baseline in patients with normal renal function during RDN. There was a non-significant reduction in the eGFR to 39.0 (IQR: 13.5–56.3) ml/min/1.73 m² in the group of patients with impaired renal function during RDN; Figure 4).

Urinary albumin values were available at follow-up time points up to 12 months. In the group of patients with an eGFR ≥ 60 ml/min/1.73 m² during RDN, there was a slight but significant decrease from 8.4 (IQR: 3.2–25.0) mg/L at baseline to 4.8 (IQR: 3.0–13.7) mg/L at 12 months (P < 0.01 for comparison with baseline). Urinary albumin levels of patients with an eGFR < 60 ml/min/1.73 m² at the time of ablation increased during the FU [baseline 5.7 (IQR: 4.4–87.7) mg/L and 12 months 13.5 (IQR: 2.9–26.2) mg/L]. However, the changes observed in patients with impaired renal function were not significant (Supplementary Table S4).

No serious or life-threatening periinterventional complications were observed. A clinically significant bleeding of the arterial access occurred in three patients, which could be treated conservatively.

Of the 245 patients initially enrolled in the registry, 37 died during follow-up. Of these, two died because of heart failure, two because of intracranial hemorrhage, one because of ischemic stroke, six because of malignancies, one because of COVID-19-ARDS, and one because of a traffic accident, according to primary care physicians and family members. The cause of death could not be established in 24 patients. One patient later received additional RDN treatment and one patient was later treated with baroreceptor stimulation. Five patients suffered from non-fatal acute myocardial infarction and 14 patients suffered from non-fatal stroke. Symptomatic heart failure was diagnosed in 39 patients. Postinterventional renal artery stenosis could not be detected in any patient and therefore was not reported.