The current prospective observational multicenter study included patients previously treated with uRDN using the Paradise system (ReCor Medical, Palo Alto, California, United States of America) within the context of the international ACHIEVE study [10]. In brief, 96 adult patients with office SBP ≥160 mmHg and 24 h ambulatory SBP ≥130 mmHg despite the use of 3 or more antihypertensive drugs, including a diuretic, were included between January 2013 and December 2014 and followed-up for 12 months. Exclusion criteria were renal artery anatomy ineligible for uRDN (main renal artery length <20 mm, main renal artery diameter <4 mm), renal artery stenosis or an estimated Glomerular Filtration Rate (eGFR) <45 ml/min/1.73m2.

Within the current long-term follow-up study, all patients provided informed consent for the use of their data and the study was conducted in accordance with the Declaration of Helsinki. Local ethical approval was obtained in all countries (Sweden, Germany, The Netherlands) and the study was registered in the German Clinical Trials Register (DRKS-ID: DRKS00029639).

All eight sites involved in the initial study were approached and four sites agreed to participate in the current long-term follow-up study. Participating sites invited all patients still alive at the time of follow-up, to a prospective outpatient clinic study visit, consisting of 24 h ambulatory BP and office BP measurements, renal function testing and data collection on medication regimen and adverse events. Data collected at these visits were used for the primary efficacy analysis.

Retrospective follow-up data were collected in all patients following the end of follow-up of the initial study (12 months post-uRDN). These data involved ambulatory BP, office BP, renal function testing, medication regimen and adverse events. In patients who declined participation in prospective follow-up visits or those initially enrolled in sites that were not willing to participate in active long-term follow-up, efforts were made to collect safety and efficacy data up to the longest available follow-up moment. Patients who declined a prospective visit were asked to provide informed consent only for the use of their retrospective follow-up data (Fig. 1).

Data collection and study outcomes. eGFR, estimated Glomerular Filtration Rate

The primary efficacy outcome was the change in 24 h ambulatory SBP between baseline (pre uRDN) and 8-year follow-up. Secondary efficacy outcomes were the changes in 24 h ambulatory diastolic blood pressure (DBP), daytime ambulatory BP, nighttime ambulatory BP, office BP, the number of defined daily dosages (DDD) and the number of classes of antihypertensive drugs between baseline (pre uRDN) and 8-year follow-up [16].

The primary safety outcome was a composite endpoint, consisting of all-cause mortality, any embolic event causing end-organ damage (including myocardial infarction and stroke), renal failure (defined as eGFR<15 ml/min per 1.73 m2 or need for dialysis), new renal artery stenosis of ≥70% and hospitalization for hypertensive crisis. Secondary safety outcomes were the individual components of the primary safety outcome and the change in renal function (eGFR) between baseline and 8 years.

For all efficacy outcomes, the evolution over time was studied by adding follow-up data from the initial study and retrospective follow-up data to the analyses. A similar approach was used for the safety outcome renal function (Fig. 1).

Categorical variables were expressed as counts (percentages). Continuous variables were described as mean ± standard deviation (SD), or as median [25th–75th percentile], depending on variable distributions. Normality was assessed using quantile-quantile plots and the Shapiro–Wilk test.

The primary efficacy outcome was analyzed in patients that participated in the prospective outpatient clinic study visit, using a linear mixed-effects model to account for between-patient differences in follow-up time. The dependent variable was 24 h ambulatory SBP and fixed effects were used for follow-up time and the number of DDDs of antihypertensive drugs. Random intercepts were used to account for repeated measurements (baseline and prospective follow-up visit) within patients, while random slopes for follow-up time were included only if they significantly improved the model fit. Inverse probability weighting was applied to adjust for loss-to-follow-up and weights were calculated based on age, sex, body mass index, medical history (smoking, diabetes, dyslipidemia, cardiovascular events), renal function (eGFR), 24 h ambulatory SBP and DBP and the number of DDDs of antihypertensive drugs at baseline. To facilitate interpretation of the regression coefficient of the fixed effect of time as the modeled 8-year change in the outcome variable, follow-up time was scaled to an 8-year time horizon in the regression model. This regression coefficient, including the corresponding 95% confidence interval (CI) and P value, was reported.

The primary safety outcome (i.e., the composite outcome) and secondary event outcomes (i.e., the individual components of the composite outcome) were analyzed using Kaplan–Meier analysis. All patients from the ACHIEVE initial study (follow-up 12 months) were included in the survival analyses, irrespective of whether or not they participated in the current long-term follow-up study (prolonged follow-up up until 8 years). The total number of events and the Kaplan–Meier estimate for the cumulative incidence (including the 95% CI) were reported.

The 8-year changes in continuous secondary efficacy outcomes (i.e., 24 h ambulatory DBP, daytime/nighttime ambulatory BP, office BP, antihypertensive drug DDDs and classes) and safety outcomes (i.e., renal function) were analyzed using a similar approach as for the primary efficacy outcome. Fixed effects for the number of DDDs of antihypertensive drugs were only used in models estimating BP outcomes.

In the sensitivity analyses, the evolution of continuous efficacy outcomes (i.e., 24 h/daytime/nighttime ambulatory BP, office BP, antihypertensive drug DDDs and classes) and safety outcomes (i.e., renal function) over time was analyzed using a similar approach as for the primary efficacy outcome. However, follow-up data from the initial study (up until 12 months post-uRDN) and retrospective follow-up data (between 12 months post-uRDN and the prospective follow-up visit) were added to the model. The linear effect of follow-up time was replaced by a non-linear effect (using natural splines with 3 degrees of freedom) in case this significantly improved the model fit. Results were displayed using effect plots and P values for the overall change in the outcome variable over time.

As the maximal sample size in this long-term study was limited to the number of participants in the initial study (n = 96), no formal sample-size calculation was performed. In general, two-sided P values <0.05 were considered statistically significant. Analyses were performed using R version 4.3.0 with packages “ipw”, “nlme”, “splines”, “survival” and “ggplot2” [17].