Renal denervation for hypertension: cross-country cost-effectiveness insights from mainland China, Japan, and Thailand

Abstract

Background

Renal denervation (RDN) has been introduced as a novel non-pharmacological intervention for patients with hypertension that is poorly controlled by pharmacological means. Our study aims to evaluate the cost-effectiveness of the Netrod RDN treatment plus antihypertensives compared with antihypertensives alone for hypertension in Mainland China, Japan and Thailand.

Methods

A Markov decision-analytic model was developed to simulate the long-term clinical events, quality-adjusted life years (QALYs) and related costs among patients who underwent RDN regimen and antihypertensive regimen in line with Netrod-HTN trial, with yearly cycles over a 30-year horizon. This study adopted the perspectives of the healthcare systems. Cost and utility inputs were collected from published literature, price databases, expert consultations, and hospital information systems. Both costs and outcomes were discounted at a rate of 5%. Model validation, univariate and probabilistic sensitivity analyses, and scenario analyses were conducted to verify the robustness of the results.

Results

Compared with the antihypertensive regimen alone, the RDN regimen yielded a 30.61% reduction in cardiovascular, cerebral and renal events. Cost-effectiveness analysis showed the RDN regimen yielded the most favorable incremental cost-effectiveness ratio in Japan at $3,451 per QALY, followed by Thailand at $13,932 per QALY, and Mainland China at $19,049 per QALY. Sensitivity and scenario analyses confirmed the robustness of the findings.

Conclusions

Netrod RDN is a cost-effective intervention from the healthcare system perspective in Mainland China, Japan, and Thailand. However, its cost-effectiveness varies across countries, reflecting differences in socioeconomic contexts. In middle- and low-income countries, appropriate pricing strategies may play a key role in enhancing its affordability and cost-effectiveness.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13561-025-00669-w.

Introduction

Hypertension emerges as the paramount risk factor for cardiovascular disease (CVD), chronic kidney disease (CKD), and all-cause mortality globally [1–3]. The Western-pacific region has a substantial burden of hypertension [4], with a prevalence rate of 28% and an estimated 346 million individuals aged 30–79 afflicted by this condition, representing the highest number among WHO regions [5]. The elevated systolic blood pressure (SBP) contributes to 10–20% of the total burden of disability-adjusted life-years across a majority of countries in the Western Pacific region [6]. Despite the widespread recognition of hypertension’s detrimental effects, hypertension control rate remains alarmingly low in the Western Pacific region, where only 18% of hypertensive patients achieve target levels - a figure significantly lower than the 36% observed in the Americas and 26% in Europe [5]. Beyond its disease burden, hypertension imposes a heavy economic burden on health systems worldwide, with direct health-care spending now estimated at more than US $500 billion each year. In high-income setting like Japan, hypertension and its complications absorb more than 3% of total medical expenditure [7]. By contrast, in China, direct medical cost linked to hypertension and its complications rose to ¥74.06 billion Chinese Yuan (about US$11.0 billion) in 2019, equivalent to 5.1% of national health expenditure [8].

Conventional hypertension pharmacological management strategies, often limited by modest blood pressure reduction and low patient medication adherence [9], have shown poor effectiveness across high-income, middle-income, and low-income nations within the Western Pacific region. For instance, control rates of hypertension stand at 29% in Japan [10], 12% in Mainland China and Thailand [11, 12], according to existing research. This underscores the imperative need for supplementary therapeutic options [13]. Recently, renal denervation (RDN) has been introduced as a novel non-pharmacological intervention in the management of hypertension. Despite existing controversies [14], a substantial body of high-quality, evidence-based medical research has confirmed RDN’s significant efficacy in substantially reducing blood pressure with no major safety events [15–17]. Consequently, both the U.S. Food and Drug Administration approved the Symplicity Spyral and Paradise RDN systems, and the Chinese National Medical Products Administration granted approval for the Netrod RDN system, all for the treatment of hypertension that remains poorly controlled by pharmacological means [18–20].

The Netrod system, a novel basket-like RDN device, has exhibited promising clinical efficacy and safety in a multicenter, randomized, sham-controlled Netrod-HTN trial [21]. In this study, 205 patients with uncontrolled hypertension were randomized into two groups: Netrod RDN plus antihypertensives treatment (n = 139) and antihypertensives alone treatment (n = 66). The Netrod RDN plus antihypertensive group demonstrated a significant difference in office systolic blood pressure (oSBP) reduction (-25.2 vs. -6.2, p < 0.001) compared to the antihypertensive alone group over a 6-month follow-up period. Moreover, there were no significant differences in safety outcomes between the two groups [21].

To comprehensively appraise the integrated value of the technology, it is imperative to evaluate the cost-effectiveness of the Netrod RDN technology for healthcare systems [22]. Previous research has substantiated the cost-effectiveness of the Symplicity Spyral RDN system for treating hypertension within the healthcare systems of developed countries in Europe and the U.S. [23–25], focusing primarily on Caucasian population. The incremental cost-effectiveness ratios (ICER) were 3,071 U.S. dollars ($) per quality-adjusted life year (QALY) in the United States, $2,323 per QALY for women and $2,642 per QALY for men in Germany, and £32,732 per QALY in the United Kingdom [23–25]. The economic viability of RDN technology among Asian populations and in developing countries remains unclear. To address these knowledge gaps, our study aims to provide the first cost-effectiveness evaluation of the Netrod RDN technology in Asian populations across countries with varying levels of economic development. Considering the perspectives of healthcare systems in Mainland China, Japan, and Thailand, this research endeavors to guide policy decisions, optimize resource allocation, and ultimately improve health outcomes for the significant number of individuals affected by hypertension in the Asia-Pacific region.

Materials and methods

Study design

Following the Consolidated Health Economic Evaluation Reporting Standards 2022 (Table S1) [26], we conducted a Markov decision-analytic model to simulate the long-term clinical events, life‑years (LYs), QALYs and related cost among patients who underwent RDN regimen and antihypertensive regimen in line with Netrod-HTN trial. LYs were calculated within the Markov framework by summing the proportion of the cohort alive in each annual cycle across the lifetime horizon. For the RDN group, patients were administered Netrod RDN surgical intervention, in conjunction with a daily oral regimen of nifedipine and hydrochlorothiazide. The antihypertensive cohort received a sham surgical procedure, followed by an identical daily pharmacological regimen of nifedipine and hydrochlorothiazide. To evaluate the cost-effectiveness of RDN technology across countries of varying economic statuses in the Western Pacific region, the healthcare systems of Mainland China, Japan, and Thailand were selected as the settings for this research. These three countries were chosen to represent different income levels based on the World Bank’s classification: Japan is a high-income country, China is an upper-middle-income country, and Thailand is a middle-income country [27]. In addition, the three countries have similar ethnic compositions, face a substantial burden of hypertension, and maintain universal health coverage through government-sponsored or public-funded social insurance and publicly run hospitals. Given the relatively low direct non-medical costs and indirect costs associated with RDN, this study is conducted from the perspective of the healthcare system. The willingness-to-pay thresholds were established at three times the gross domestic product (GDP) per capita for 2023 [28], following the World Health Organization guidelines, amounting to $37,624 for Mainland China, $101,849 for Japan, and $21,894 for Thailand [29]. QALYs and costs were discounted at 2% (Japan), 3% (Thailand), and 5% (China) per annum, in line with the respective national pharmacoeconomic guidelines [30–32].

Model structure and assumptions

We developed our decision-analytic model through a two-step process. First, we designed the model’s structure, states, and transitions based on the frameworks used in previous cost-effectiveness studies of hypertension treatments [24, 33]. Second, the face validity of the model was corroborated through expert consultation. A Markov model including six health states: hypertension without events, post-coronary heart disease (post-CHD), post-stroke, post-CHD and Stroke, end-stage renal disease (ESRD) and death was conducted by Excel software (Microsoft, Redmond, WA, USA). In the transitions between health states, the model accounted for the following clinical events: coronary heart disease (inclusive of acute myocardial infarction (AMI), Heart failure (HF) necessitates hospitalization, and sudden cardiac death (SCD)), stroke (comprising ischemic stroke (IS) and hemorrhagic stroke (HS)), and transient ischemic attacks (TIA) (Fig. 1, Figure S1-S3). The model was structured with a one-year cycle, within which individuals could either maintain their current health state, endure clinical events, or die from clinical events or causes unrelated to these events. It was assumed that individuals progressing to an ESRD state would not transition to other states except for mortality, although they remained at risk of experiencing strokes and CHD while in the ESRD state. A half-cycle correction was employed in the model, and the base-case simulation horizon was set at 30 years, aligning with the average life expectancy [34].

Fig. 1.

Structure of the Markov model. AMI, acute myocardial infarction; CHD, coronary heart disease; ESRD, end-stage renal disease; HF, heart failure; HTN, Hypertension; TIA, transient ischemic attacks. Transitions from each state to death encompass two pathways: death due to acute events (e.g., CHD or stroke) and non‑event‑related death

The clinical efficacy of treatment in both the RDN group and the antihypertensive group, as input into the model, aligns with outcomes reported in the Netrod-HTN trial: a reduction in oSBP of 25.2 mmHg in the experimental group and of 6.2 mmHg in the control group. In light of the findings from several existing observational studies [35, 36], we posit that the clinical efficacy of both groups remains unchanged.

Study cohort

The study conducted two simulated patient cohorts, each consisting of 1,000 individuals with an initial age of 50 years, and baseline characteristics that closely replicate those observed in the experimental and control groups of the Netrod-HTN trial. The characteristics evaluated included oSBP, gender, age, total cholesterol, high-density lipoprotein cholesterol (HDL-C), smoking status, and geographical location [21, 37]. Additionally, data on family history and urbanization were obtained from the China PAR cohort (Table S2) [38, 39]. Owing to the absence of waist circumference measurements, estimates were derived using a predictive equation based on body mass index, gender, and age, as documented in the literature [40]. The inclusion criteria for the study were consistent with those of the Netrod-HTN trial, incorporating patients with essential hypertension who had been under treatment with at least two antihypertensive medications for a minimum of eight weeks, including a four-week washout period using nifedipine and hydrochlorothiazide, with blood pressure levels of ≥ 150/90mmHg and < 180/110mmHg, and a 24-hour ambulatory SBP of ≥ 135mmHg.

Transition probabilities

The probabilities of patients transitioning from hypertension without events state to experiencing clinical events were derived from risk equations or estimated through cohort studies. We utilized multivariate risk equations derived from a large-scale Asian population cohort study to predict 10-years atherosclerotic cardiovascular disease (ASCVD) and stroke risk for hypertension patients without events [38, 39]. Furthermore, we applied a fixed ratio of CHD to ASCVD to calculate the incidence of CHD for events-free patients [41]. The input parameters comprised individual patient data from the Netrod-HTN trial. For estimating the risk of CHD and stroke beyond the initial decade, it was assumed that oSBP, smoking status, waist circumference, geographical location, and family history would remain constant, while age, HDL-C, and total cholesterol were extrapolated in accordance with the method reported in published literature [33]. CHD events were categorized into AMI, HF and SCD based on a fixed proportion derived from prior studies [33, 42], while the stroke events can be split into IS and HS using the same method [43]. Due to the absence of risk equations for the progression from hypertension to ESRD in the hypertensive population, we estimated the transition probabilities from events-free to ESRD through a population-based cohort study [44]. Besides, the influence of age on incidence rates was adjusted based on data from another cohort study [45]. The TIA incidence was calculated by a ratio of TIA to stroke [43]. Survivors of stroke, CHD, and ESRD events transition to post-CHD, post-stroke, post-CHD and stroke, and ESRD states, respectively. Within these states, their likelihood of experiencing recurrent events is elevated in comparison to patients without such events. Relative risks (RRs) to adjust the incidence rates of CHD, stroke, TIA, and ESRD were sourced from extensive observational studies (Table S3) [46–50]. In our model, the non-event mortality rates for each state, specific to age groups, along with the acute mortality rates for AMI, HF, HS, and IS, were extracted from prior literature and reports [33, 42, 51–53]. The mortality rate for TIA was assumed to be zero.

Quality of life

In our Markov model, we refined the assessment of health states through quality-of-life adjustments (Table 1). QALYs were determined by the product of the duration spent in a particular health state and the utility value attributed to that state, with utility values sourced from prior investigations that assessed patient health status with the EQ‑5D‑5L instrument, and index scores were generated using the most recent Chinese EQ‑5D‑5L value set [33, 54–56]. Additionally, we incorporated decremental utility values for acute clinical events, adhering to values outlined in the existing literature [33].

Table 1.

Model input values

Parameter	Base case	Ranges	Distribution	Source
Clinical inputs
Character of experimental group
Change in clinic oSBP, mmHg	25.20	(22.68 to 27.72)	Lognormal	(37)
Total cholesterol, mg/dL	183.63	(165.27 to 202.00)	Lognormal	(37)
HDL-C, mg/dL	53.02	(47.72 to 58.32)	Lognormal	(37)
Waistline	92.95	(83.65 to 102.25)	Lognormal	(37, 40)
HTN to ESRD, %	0.028	(0.025 to 0.031)	Beta	(44)
Post-CHD to ESRD, %	0.036	(0.032 to 0.040)	Beta	(44, 50)
Post-stroke to ESRD, %	0.036	(0.032 to 0.040)	Beta	(44, 50)
Character of control group
Change in clinic oSBP, mmHg	6.20	(5.58 to 6.82)	Lognormal	(37)
Total cholesterol, mg/dL	188.76	(169.88 to 207.63)	Lognormal	(37)
HDL-C, mg/dL	54.47	(49.02 to 59.91)	Lognormal	(37)
Waistline	92.77	(83.49 to 102.05)	Lognormal	(37, 40)
HTN to ESRD, %	0.041	(0.037 to 0.045)	Beta	(44)
Post-CHD to ESRD, %	0.053	(0.048 to 0.059)	Beta	(44, 50)
Post-stroke to ESRD, %	0.053	(0.048 to 0.059)	Beta	(44, 50)
Utility inputs
HTN without events	0.94	(0.85 to 1.00) a	Beta	(54)
Post-CHD	0.85	(0.76 to 0.93)	Beta	(33)
Post-stroke	0.79	(0.71 to 0.87)	Beta	(33)
Post-CHD and stroke	0.59	(0.53 to 0.65)	Beta	(33)
ESRD	0.61	(0.55 to 0.67)	Beta	(55)
Death	0		-	Assumption
AMI Disutility	0.01	(0.01 to 0.01)	Beta	(33)
HF Disutility	0.05	(0.05 to 0.06)	Beta	(33)
HS Disutility	0.09	(0.08 to 0.10)	Beta	(33)
IS Disutility	0.09	(0.08 to 0.10)	Beta	(33)
TIA Disutility	0.09	(0.08 to 0.10)	Beta	(33)
Discount rate (%)
Costs	5	(0 to 8)	Beta	(30)
QALYs	5	(0 to 8)	Beta	(30)
Cost input: China ($)
RDN procedure cost	7942	(7147 to 8736)	Gamma	HISb
Annual antihypertensives costs	312	(281 to 344)	Gamma	(65)
Annual hypertension management costs
RDN group	161	(145 to 178)	Gamma	(61), HISb
Antihypertensives group	142	(128 to 156)	Gamma	HISb
Acute cardiovascular events costs
Non-fatal AMI hospitalization	5699	(5129 to 6269)	Gamma	(33)
Non-fatal HF hospitalization	4573	(4116 to 5030)	Gamma	(33)
Non-fatal IS hospitalization	7176	(6458 to 7893)	Gamma	(33)
Non-fatal HS hospitalization	4051	(3645 to 4456)	Gamma	(33)
TIA hospitalization	2022	(1819 to 2224)	Gamma	(33)
Fatal AMI hospitalization	7913	(7122 to 8704)	Gamma	(33)
Fatal HF hospitalization	9364	(8427 to 10300)	Gamma	(33)
Fatal IS hospitalization	6905	(6215 to 7596)	Gamma	(33)
Fatal HS hospitalization	8398	(7558 to 9238)	Gamma	(33)
SCD hospitalization	4225	(3803 to 4648)	Gamma	(33)
Annual post disease management costs
Post CHD (initial year)	2640	(2376 to 2904)	Gamma	(33)
Post CHD (subsequent year)	1430	(1287 to 1573)	Gamma	(33)
Post Stroke (initial year)	4825	(4343 to 5308)	Gamma	(33)
Post Stroke (subsequent year)	757	(681 to 833)	Gamma	(33)
Post CHD and Stroke (initial year)	7465	(6719 to 8212)	Gamma	(33)
Post CHD and Stroke (subsequent year)	2187	(1969 to 2406)	Gamma	(33)
ESRD	15,953	(14357 to 17548)	Gamma	(81)
ESRD end-of-life	4877	(4389 to 5364)	Gamma	(76)
Cost input: Japan ($)
RDN procedure cost	13,370	(12033 to 14707)	Gamma	(64)c
Annual antihypertensives costs	52	(47 to 57)	Gamma	(67)
Annual hypertension management costs
RDN group	225	(202 to 247)	Gamma	(62)
Antihypertensives group	187	(168 to 206)	Gamma	(62)
Acute cardiovascular events costs
Non-fatal AMI hospitalization	13,030	(11727 to 14333)	Gamma	(70)
Non-fatal HF hospitalization	8040	(7236 to 8845)	Gamma	(71)
Non-fatal IS hospitalization	12,312	(11081 to 13544)	Gamma	(70)
Non-fatal HS hospitalization	16,240	(14616 to 17864)	Gamma	(70)
TIA hospitalization	3598	(3238 to 3957)	Gamma	(71)
Fatal AMI hospitalization	18,486	(16638 to 20335)	Gamma	(74)
Fatal HF hospitalization	12,416	(11175 to 13658)	Gamma	(71)
Fatal IS hospitalization	11,121	(10009 to 12233)	Gamma	(74)
Fatal HS hospitalization	10,865	(9778 to 11951)	Gamma	(74)
SCD hospitalization	12,416	(11175 to 13658)	Gamma	(71)
Annual post disease management costs
Post CHD (initial year)	7185	(6467 to 7904)	Gamma	(70)
Post CHD (subsequent year)	1942	(1748 to 2137)	Gamma	(83)
Post Stroke (initial year)	18,129	(16316 to 19942)	Gamma	(70)
Post Stroke (subsequent year)	17,088	(15379 to 18797)	Gamma	(71)
Post CHD and Stroke (initial year)	25,315	(22783 to 27846)	Gamma	Assumption
Post CHD and Stroke (subsequent year)	18,504	(16654 to 20355)	Gamma	Assumption
ESRD	47,209	(42488 to 51930)	Gamma	(77)
ESRD end-of-life	14,305	(12875 to 15736)	Gamma	(78)
Cost input: Thailand ($)
RDN procedure cost	5763	(5187 to 6339)	Gamma	(64)c
Annual antihypertensives costs	141	(127 to 156)	Gamma	(66)
Annual hypertension management costs
RDN group	164	(147 to 180)	Gamma	(63)
Antihypertensives group	49	(44 to 54)	Gamma	(63)
Acute cardiovascular events costs
Non-fatal AMI hospitalization	1117	(1005 to 1229)	Gamma	(75)
Non-fatal HF hospitalization	1782	(1604 to 1960)	Gamma	(73)
Non-fatal IS hospitalization	840	(756 to 924)	Gamma	(75)
Non-fatal HS hospitalization	840	(756 to 924)	Gamma	(75)
TIA hospitalization	540	(486 to 593)	Gamma	(84)
Fatal AMI hospitalization	2980	(2682 to 3278)	Gamma	(72)
Fatal HF hospitalization	1782	(1604 to 1960)	Gamma	(73)
Fatal IS hospitalization	1723	(1551 to 1895)	Gamma	(75)
Fatal HS hospitalization	1723	(1551 to 1895)	Gamma	(75)
SCD hospitalization	2980	(2682 to 3278)	Gamma	Assumptiond
Annual post disease management costs
Post CHD (initial year)	1962	(1765 to 2158)	Gamma	(75)
Post CHD (subsequent year)	560	(504 to 616)	Gamma	(75)
Post Stroke (initial year)	1337	(1204 to 1471)	Gamma	(75)
Post Stroke (subsequent year)	345	(310 to 379)	Gamma	(75)
Post CHD and Stroke (initial year)	3299	(2969 to 3629)	Gamma	Assumption
Post CHD and Stroke (subsequent year)	905	(814 to 995)	Gamma	Assumption
ESRD	11,225	(10103 to 12348)	Gamma	(82)
ESRD end-of-life	4670	(4203 to 5137)	Gamma	(79)

AMI, acute myocardial infarction; CHD, coronary heart disease; ESRD, end-stage renal disease; HDL-C, high-density lipoprotein cholesterol; HF, heart failure; HIS, Hospital information system; HS, hemorrhagic stroke; HTN, Hypertension; IS, ischemic stroke; oSBP, office systolic blood pressure; RDN, renal denervation; QALYs, quality-adjusted life years; SCD, sudden cardiac death; TIA, transient ischemic attacks. ^a Given that the upper limit for utility values is 1.00, the + 10% upper boundary was not attained in this instance. ^b Hospital information systems from ten nationwide hospitals, which conducted Netrod RDN clinical trials, were utilized to extract data on service fees associated with the RDN procedure and the management of hypertensive patients. ^c Based on the cost of the RDN procedure in China, the base-case expense for Netrod in Japan and Thailand were calculated utilizing the Health Purchasing Power Parity indices published by the World Bank. Apart from the base-case price, we performed a range of pricing scenario analyses, with prices varying from a 100% decrease to a 400% increase. ^d We assumed that the costs associated with SCD are equivalent to those of fatal AMI

Costs

All costs were reported in 2023 $ and were adjusted to 2023 values using the Consumer Price Index [57–59]. Currency conversions were executed employing the average exchange rates in 2023, as reported by the International Monetary Fund, with the Chinese Yuan at 7.085, the Japanese Yen at 140.535, and the Thai Baht at 34.812 [60]. Only direct medical costs were considered since we adopted a perspective of the healthcare system. Direct medical costs included a one-time RDN procedure cost, annual antihypertensive pharmaceutical costs, annual hypertension management costs and disease management costs, which together represent total medical expenditure, including both the portion reimbursed by health insurance and the out‑of‑pocket payments made by patients.

Given the absence of a publicly disclosed price for the Netrod RDN, this study assumed the sale price of the RDN device. The research acquired costs associated with patient examination, procedure, hospitalization, and nursing for undergoing RDN surgery by gathering cost information from the information systems of ten nationwide hospitals that conducted Netrod RDN clinical trials, as well as from price guidelines [61–63]. Upon determining the costs for the Netrod RDN procedure in Mainland China, the base-case costs for the Netrod in Japan and Thailand were calculated using the Health Purchasing Power Parity (PPP) indices published by the World Bank [64]. In addition to the base-case pricing, we conducted a series of price scenario analyses to minimize the uncertainty regarding the generalizability of the study results due to fluctuations in the cost of the RDN procedure. The costs per unit of antihypertensive medication were sourced from price databases, enabling the calculation of the average selling price per milligram for each category of drug [65–67]. The dosages administered to patients were determined based on recommendations outlined in clinical guidelines [68]. The costs of hypertension management for the control group were acquired from hospital information systems, existing literature and price databases [33, 62, 63, 69]. Following expert consultation, the management expenses for the experimental group were augmented to include the cost of an annual ultrasound examination [61–63]. The costs associated with various acute cardiovascular events were segregated into those for fatal and non-fatal incidents. The annual management costs for post-CHD, post-stroke, and post-CHD & stroke states were differentiated between the initial year of management and the costs for subsequent years. For ESRD, the costs were delineated into annual management expenses and end-of-life costs. The specific cost parameters derived from literature and hospital databases were presented in Table 1 [33, 70–84].

Model validation

This study employed multiple strategies to validate the model’s efficacy. Initially, expert consultations were conducted to ascertain the model’s face validity, with nine experienced clinical experts from nine tertiary hospitals and seven pharmacoeconomists evaluating the model’s primary assumptions, structure, input parameters, study duration, intervention plans, and patient cohorts. Subsequently, the study verified internal validity. Two other researchers (S.L., Y.R.) conducted a comprehensive examination of the model. Moreover, the study assessed external validity through two approaches. We compared the incidence rates of AMI, SCD, Stroke, and ESRD in the model’s baseline scenario with those reported in published epidemiological studies. Additionally, we juxtaposed the expected lifespan of the cohort with the life expectancy reported for the population aged 50 in epidemiological studies.

Sensitivity analyses

To evaluate the uncertainty enveloping model parameters, a sequence of univariate sensitivity analyses was executed. This process involved sequentially modifying the value of individual input variables to determine their effect on the ICER. A comprehensive review was conducted across a spectrum of parameters, which included those pertinent to multivariate risk equations, risk adjustments, direct costs, discount rates, and the utilities assigned to various health states and events. Consistent with the guidance of Briggs et al., a standard deviation of ± 10% from their foundational values was applied to most parameters [85], whereas the discount rates were adjusted within a spectrum of 0–8%, adhering to established guidelines [30]. To further evaluate the compounded uncertainty across multiple parameters, Monte Carlo simulations encompassing 5,000 iterations were employed. Within these iterations, a random sampling from the distributions designated for all parameters was conducted concurrently. The efficacy parameters were modelled using lognormal and beta distribution, cost parameters through a gamma distribution, and health utilities via a beta distribution. Table 1 and Table S3 meticulously delineates the predefined values, variable ranges, and the distributions assumed for these model parameters.

We conducted a series of scenario analyses to validate the robustness of outcomes under various circumstances. In the first scenario analysis, we evaluated changes in the ICER by adjusting the price of the RDN procedure by increments of 5,000 Chinese yuan (equivalent to $705.66) in China, while also conducting a range of price scenario analyses in Thailand and Japan, with price variations ranging from a 100% reduction to a 400% increase. In the second scenario analysis, the impact on the ICER of setting the study period to 20 years and 50 years was investigated. The third scenario analysis took into consideration that patients with superior blood pressure control (SBP < 140mmHg) might have a higher quality of life than those with suboptimal control (SBP > 140mmHg) [86, 87]. Consequently, we assigned different utility values to the two groups of patients with hypertension without events (0.97 for RDN group and 0.95 for control group) [87], to assess their impact on the ICER. Finally, in light of the absence of direct clinical data for the Japanese population from the Netrod RDN trial, we performed a scenario analysis incorporating the blood pressure reduction effect from the SYMPLICITY HTN-Japan study (-16.6 mmHg for RDN group and − 7.9 mmHg for control group) to mitigate uncertainty [88]. To maintain comparability, all sensitivity and scenario analyses were conducted using the base‑case scenario with a 5% discount rate.

Patient and public involvement

Due to privacy considerations, funding constraints, and the limited timeframe of the study, patients and the public were not directly involved in shaping the research question, study design, outcome measures, or its implementation. Since the data sources were anonymized, it is not possible to directly share the results with study participants. However, the findings will be communicated to the public and healthcare professionals through a press release written in clear and accessible language.

Results

Model validation

Clinical experts and health economists have validated the face validity of the model, confirming the appropriateness of its assumptions, structure, and parameter inputs. Two authors conducted a back-to-back examination of the model’s internal validity, ensuring the absence of errors. Comparisons with multiple epidemiological studies of Asian population reports have shown that the model predicts similar incidence rates and aligns with other studies in terms of life expectancy (Table S4), thereby affirming its favorable external validity.

Base-case results

In the base-case scenario, the RDN group exhibited a 30% reduction in cardiovascular, cerebral, and renal events in comparison to the control group (Table 2). In the cost-effectiveness analysis, the RDN approach yielded the most favorable ICER in Japan at $3,451 per QALY (discounted at 2%), followed by Thailand at $13,932 per QALY (discounted at 3%), and Mainland China at $19,049 per QALY (discounted at 5%). These ICERs correspond to 0.10 times the GDP per capita in Japan, 1.90 times the GDP per capita in Thailand, and 1.52 times the GDP per capita in Mainland China (Table 2), respectively, demonstrating the cost-effectiveness of the Netrod RDN across these three nations.

Table 2.

Base-case results

	RDN groups	Control groups	Differences	Relative risk
Number of events
CHD	60	83	-23	0.72
AMI	47	64	-17	0.73
HF	21	30	-9	0.70
SCD	13	18	-5	0.72
Stroke	255	370	-115	0.69
IS	132	191	-60	0.69
HS	123	178	-56	0.69
TIA	105	152	-47	0.69
ESRD	13	19	-6	0.68
Cardiovascular, stroke, and renal event mortality	116	164	-48	0.71
Discount rate at 5%
Mean LYs	14.27	13.95	0.32	-
Mean QALYs	13.27	12.90	0.37	-
Mean costs per person ($)
Mainland China	17,260	10,170	7091	-
Japan	34,592	28,703	5889	-
Thailand	10,994	4099	6896	-
ICER ($/QALY)
Mainland China	19,049
Japan	15,821
Thailand	18,525
ICER ($/LY)
Mainland China	22,046
Japan	18,310
Thailand	21,410
Discount rate at 3%
Mean LYs	17.47	17.02	0.45
Mean QALYs	16.23	15.72	0.51
Mean costs per person ($)
Mainland China	19,586	12,722	6864
Japan	41,542	37,953	3589
Thailand	12,287	5190	7097
ICER ($/QALY)
Mainland China	15,254
Japan	7046
Thailand	13,932
ICER ($/LY)
Mainland China	14,927
Japan	8013
Thailand	15,845
Discount rate at 2%
Mean LYs	19.53	19.00	0.53
Mean QALYs	18.13	17.53	0.60
Mean costs per person ($)
Mainland China	21,105	14,388	6717
Japan	46,183	44,105	2079
Thailand	13,117	5887	7230
ICER ($/QALY)
Mainland China	11,153
Japan	3451
Thailand	12,005
ICER ($/LY)
Mainland China	12,674
Japan	3894
Thailand	13,544

AMI, acute myocardial infarction; CHD, coronary heart disease; ESRD, end-stage renal disease; HF, heart failure; HS, hemorrhagic stroke; ICER, incremental cost-effectiveness ratio; IS, ischemic stroke; LY: life year; SCD, sudden cardiac death; QALY, quality-adjusted life year; RDN, renal denervation; TIA, transient ischemic attacks

Sensitivity analyses

In the univariate sensitivity analysis, the factors that most significantly influenced the ICERs across the countries were the discount rates for QALYs and Costs, the utility values for hypertension without events and post-stroke states, the procedural costs of RDN, and the blood pressure reduction, baseline parameters such as HDL-C and total cholesterol in the experimental and control groups. Except for Thailand, where the ICER exceeds the threshold with variations in the QALY discount rate (Fig. 2C), the ICERs in other countries remain below their respective cost-effectiveness thresholds, demonstrating the cost-effectiveness of the intervention (Fig. 2A, B).

Fig. 2.

Univariate sensitivity analyses of RDN strategy in Mainland China (A), Japan (B), and Thailand (C). HDL-C, high-density lipoprotein cholesterol; HTN, Hypertension; IS, ischemic stroke; QALYs, quality-adjusted life years; RDN, renal denervation

The results of the probabilistic sensitivity analysis indicate that the RDN holds a substantial probability of being cost-effective in Mainland China, Japan, and Thailand at thresholds of three times the GDP per capita, specifically at 98.04% for Japan, 90.78% for Mainland China, and 65.98% for Thailand (Figure S4-S6). The cost-effectiveness acceptability curves for each country are presented in Figure S7-S9.

Scenario analysis results confirm the reliability of base-case findings, showing that the RDN regimen is still cost-effective below the threshold of three nations. Assumptions of blood pressure reduction, discount rate and utility values demonstrate the largest impact of ICER (Table 3). Exploratory analyses of the pricing of the RDN procedure in Japan and Thailand indicate that when the cost of the RDN procedure in Japan is reduced by 43%, it dominates antihypertensive therapy, being associated with lower costs and improved efficacy. Conversely, when the cost of the RDN procedure in Japan increases by 338%, it is not cost-effective. Similarly, an increase of 22% in the cost of the RDN procedure in Thailand also results in a lack of cost-effectiveness (Figure S10-S11).

Table 3.

ICERs in various scenario analyses

Scenario	Mainland China	Japan	Thailand
Reduction in RDN procedure cost by CNY 5, 000 a	17,194	12,629	17,150
Increase in RDN procedure cost by CNY 5,000	20,996	19,012	19,901
Time horizon set to 20 years	30,501	31,951	28,566
Time horizon set to 50 years	15,656	12,160	15,374
Allocation of different utility values	10,962	9104	10,660
Substitution of Clinical Efficacy Parameters	-	57,878	-

CNY, Chinese Yuan; ICER, incremental cost-effectiveness ratio; RDN, renal denervation. ^a 5000 Chinese yuan equivalent to $705.66

Discussion

This study confirmed that the BP reduction benefits conferred by the Netrod RDN procedure could significantly reduce the number of long-term cardiovascular and renal events, demonstrating favorable cost-effectiveness across the healthcare systems of the respective countries. Both sensitivity analyses and scenario analyses emphasize the robustness of the outcomes. Particularly, this study adopts the internationally recognized threshold of three times the GDP per capita, also recommended by WHO cost-effectiveness analysis guidelines, to ensure comparability between countries. This threshold may differ from those actually used by payers in various countries [30, 89, 90]. Nevertheless, decision-makers can still assess the cost-effectiveness of RDN in their respective countries by comparing the ICER results obtained from this study with their national reference thresholds.

The disparities in the ICERs across different countries reflect the variations in costs and treatment strategies within their healthcare systems. Moreover, our study found that the pricing of RDN devices may offset the impact of these differences on the ICER. Using Health PPP, we determined that the price of the RDN procedure in Thailand was $5,763, while in China it was $7,942. This substantial difference resulted in Thailand having a slightly lower ICER compared to China, despite the overall lower healthcare costs in the latter country. Our study highlights the value of innovative interventions like RDN in developed countries (Japan) with high management costs for CHD, stroke, and ESRD. Despite the substantial upfront costs associated with RDN, its ability to prevent these costly cardiovascular and renal complications results in favorable cost-effectiveness in healthcare systems with higher disease management costs. Differences in the costs of managing cardiovascular, cerebrovascular, and renal events across China, Japan, and Thailand result partly from variation in macro‑economic development and partly from structural distinctions in their health‑care systems, including divergence in healthcare labor costs and provider payment methods. China and Thailand employ Diagnosis‑Related Groups based case payment, whereas Japan reimburses inpatient care through a national per‑diem fee schedule under the Diagnosis Procedure Combination framework [91, 92]. Consequently, health insurance providers in developed countries may be more willing to cover these expensive innovative therapies, enabling patients in these nations to benefit more readily from innovative products like RDN. The elevated ICERs in China and Thailand may present challenges for reimbursement and insurance coverage of RDN. Policymakers in these countries may need to carefully weigh the long-term benefits of RDN against the upfront costs and budget impact when making coverage decisions [93, 94]. To improve equity in global hypertension management, governments, health insurance providers, nongovernmental organizations and manufacturers in low- and middle-income countries should collaborate to enhance the accessibility and affordability of innovative products in these regions [95]. This may involve implementing targeted reimbursement policies, negotiating price reductions, or exploring innovative financing mechanisms to ensure that patients in resource-limited settings can also benefit from advanced hypertension management options. By working together to bridge the gap in access to innovative therapies, stakeholders can promote more equitable health outcomes and reduce the global burden of hypertension-related complications [96].

Previous cost-effectiveness analyses of RDN predominantly focused on Medtronic’s Spyral system, with earlier publications using the blood pressure reduction data from the Simplicity-HTN1 (20 mmHg) or Simplicity-HTN2 (32 mmHg) trials, which do not reflect the outcomes of the latest clinical evidence [23, 24]. Recent research has demonstrated the cost-effectiveness of Spyral alongside Paradise’s Endovascular Ultrasound RDN within the perspective of UK healthcare system, with base-case Incremental ICERs estimated at £5,600 ($7,164.64) and £13,482($17,248.87), respectively, affirming RDN as a cost-effective strategy [25, 97]. Differences in ICERs between studies are primarily attributable to variations in blood pressure reduction effects, target populations, healthcare costs and discount rates.

This study features several highlights. To our best knowledge, it represents the first exploration into the cost-effectiveness of RDN technology for treating hypertension in Asian populations. Second, diverging from previous cost-effectiveness analyses of RDN that predominantly focused on the Symplicity Spyral RDN system, this research is the first to investigate an innovative RDN device, the Netrod RDN system. Third, the study examines the economic viability of RDN technology within the healthcare systems of both high-income and middle-income countries, filling the gap that previous research primarily concentrated on RDN applications in high-income countries. Our findings offer valuable insights to inform reimbursement policies and clinical decision-making for hypertension management in different resource settings. Based on our study, decision-making elsewhere could adapt our cost-effectiveness analysis and its method to fitting their needs, and reinput cost parameters to their own economic and service‑delivery contexts to ensure an accurate cost‑effectiveness appraisal of RDN.

This study revealed several limitations. First, although we employed a relatively complex model to simulate the clinical progression of hypertensive patients, it inevitably simplifies the potential clinical realities, as all models do. For instance, the study did not consider peripheral vascular diseases caused by hypertension, nor did it account for heart failure cases that do not require hospitalization, potentially leading to an underestimation of the value provided by Netrod RDN’s blood pressure reduction treatment. Second, clinical studies of Netrod RDN were only conducted in the Chinese population, and we extrapolated the blood pressure reduction benefits of Netrod treatment to the Japanese and Thai populations, which are similar to the Chinese ethnicity. Limited by the scarcity of risk equations developed from large cohort population samples, we utilized Chinese risk equations to predict cardiovascular event risk, necessitating the extrapolation of our predicted clinical effects to other populations. Although this approach may limit the generalizability of our findings, existing meta-analyses have confirmed similar cardiovascular disease incidence rates among East Asian and Southeast Asian populations [98]. Third, given that existing follow-up studies on RDN have confirmed a gradually increasing trend in the blood pressure-lowering effect of RDN technology over a 10-year follow-up period [35], the cost-effectiveness derived from the six-month follow-up blood pressure effect of RDN in this study may underestimate the overall benefits to the population. Fourth, as the current pricing for the Netrod RDN procedure has not yet announced in Japan and Thailand, our study employed PPP to estimate these costs. Given the detailed reporting of the formulas for RDN cost input parameters and ICER outcomes in this study, researchers can replace the RDN procedure cost parameters with the actual prices once they are released, allowing for the recalculation of the ICER to support decision-making. Lastly, our analysis, based on the original experimental reports of Netrod RDN’s safety outcomes, did not consider the costs associated with surgery-related adverse reactions. Nevertheless, real-world evidence from the large-scale application of RDN may supplement safety event data for subsequent research.

The Netrod RDN system plus antihypertensives represents a cost-effective intervention compared to the treatment with antihypertensives alone for hypertension from the perspective of the healthcare systems in Mainland China, Japan, and Thailand. Our study found that the cost-effectiveness of RDN varies across countries, reflecting differences in healthcare costs and economic contexts. However, by strategically reducing product prices, RDN can also exhibit favorable cost-effectiveness in resource-limited settings. These findings can inform the development of country-specific reimbursement policies, clinical guidelines, and patient management strategies that optimize the value of RDN in diverse healthcare settings, ultimately contributing to improved hypertension control and cardiovascular outcomes on a global scale.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AMI

Acute myocardial infarction

ASCVD

Atherosclerotic cardiovascular disease

ESRD

End-stage renal disease

GDP

Gross domestic product

HDL-C

High-density lipoprotein cholesterol

HF

Heart failure (necessitates hospitalization)

HS

Hemorrhagic stroke

ICER

Incremental cost-effectiveness ratio

IS

Ischemic stroke

oSBP

Office systolic blood pressure

PPP

Purchasing Power Parity

QALY

Quality-adjusted life year

RDN

Renal denervation

RR

Relative risks

SCD

Sudden cardiac death

TIA

Transient ischemic attacks

Author contributions

All authors have read and approved the final version of the manuscript. SL and YC conceptualized the study. DL, YR and JZ curated the data. DL conducted the formal analysis. Validation involved YS, RG, DX, NL, MZ, SL and YC. DL wrote the original draft, while all authors contributed to reviewing and editing. All authors read and approved the final version of the manuscript. SL and YC is responsible for the overall content as guarantor.

Funding

This work was supported by the Shanghai Golden Leaf MedTech (Grant No. 20231H03164). The funders had no role in considering the study design or in the collection, analysis, interpretation of data, writing of the report, or decision to submit the article for publication.

Data availability

Individual participant data will not be available. Other requests will be considered upon receipt by the corresponding author.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

Funding from Shanghai Golden Leaf MedTech Company Limited for the submitted work; no financial relationships with any organizations that might have an interest in the submitted work in the previous three years. no other relationships or activities that could appear to have influenced the submitted work.