Pulmonary Denervation for Heart Failure–Related Pulmonary Hypertension (PADN-HF-PH)

Authors, Journal, Affiliations, Type, DOI

• Authors: Hang Zhang, Qiguang Wang, Ming Liang (the three contributed equally), Xiang Chen, Wei Zhuang, Biao Cheng, Yue Li, Jieyan Shen, Bin Liu, Caojing Zhang, Gangcheng Zhang, Xinxue Liao, Zhenwen Yang, Ruilin He, Yuquan He, Yunshan Cao, Yan Wang, Dajun Chai, Wei Huang, Xiaoping Peng, Wen Gu, Gregg W. Stone, Yaling Han, and Shao-Liang Chen (corresponding), for the PADN-HF-PH Investigators
• Journal: N Engl J Med, published online 30 August 2026. DOI 10.1056/NEJMoa2609056
• Affiliations: 25 sites in China, coordinated from Nanjing First Hospital, Nanjing Medical University, and the General Hospital of Northern Theater Command, Shenyang; co-authors from Icahn School of Medicine at Mount Sinai, New York. One co-author (Wen Gu) is affiliated with Pulnovo Medical, the device manufacturer
• Type: Investigator-designed, multicentre, open-label, randomized, controlled trial; 1:1 allocation, centralized Web-based randomization in blocks of four or six, stratified by PH phenotype (isolated postcapillary vs combined precapillary and postcapillary) and trial site. Independent data and safety monitoring board; clinical events committee adjudicated events in a blinded manner; echocardiographic and haemodynamic measurements analysed at an independent core laboratory by blinded technicians; NT-proBNP analysed centrally and blinded. Registration NCT05824923. Rationale and design published separately (Cardiol Discov 2026;6:99–105)
• DOI: 10.1056/NEJMoa2609056

Overview

PADN-HF-PH is the first adequately-powered randomized trial of transcatheter pulmonary-artery denervation in group 2 pulmonary hypertension — PH caused by left heart disease, which accounts for 65–80% of all PH and has no approved therapy directed at its pulmonary vascular component. In 264 patients with symptomatic heart failure (any LVEF) already on contemporary guideline-directed medical therapy, pulmonary-artery denervation plus GDMT halved the 2-year Kaplan–Meier rate of clinical worsening versus GDMT alone (25.7% vs 51.5%; HR 0.49, 95% CI 0.30–0.82; P=0.006), with the effect persisting when the 6-minute walk distance component was removed from the composite, and with no procedural complications beyond access-site haematomas (2 vs 1). The proposed mechanism is interruption of pulmonary vascular sympathetic signalling, with an acute fall in mean pulmonary-artery pressure in 87.9% of patients at 10 minutes and improved biventricular function and reduced tricuspid regurgitation on serial echocardiography at 6 months. The critical caveat is the design: the trial was open-label with no sham control, which the authors acknowledge leaves placebo effects and ascertainment bias unexcluded, and the individual hard endpoints — death and heart-failure hospitalisation — were not separately significant.

Keywords

Not provided — NEJM original articles do not carry an author keyword list. The document's own indexing terms would be: pulmonary hypertension associated with left heart disease; WHO group 2 pulmonary hypertension; isolated postcapillary (IpcPH); combined precapillary and postcapillary (CpcPH); heart failure; pulmonary-artery denervation; sympathetic nervous system; pulmonary vascular resistance; right ventricular–pulmonary-artery coupling.

Key Takeaways

Background

• Heart failure affects more than 35 million people worldwide and remains a leading cause of hospitalisation and death. Outcomes stay poor in older patients and those with advanced disease despite advances in pharmacotherapy and rhythm management
• Heart failure is characterised by sustained neurohormonal and sympathetic nervous system activation, which contributes to progression and adverse outcomes
• PH secondary to left heart disease accounts for 65–80% of all PH and portends a worse prognosis in heart failure
• Two phenotypes, separated by pulmonary vascular resistance:
  • Isolated postcapillary PH (IpcPH) — mPAP >20 mm Hg, PAWP >15 mm Hg, PVR ≤2 Wood units
  • Combined precapillary and postcapillary PH (CpcPH) — same mPAP and PAWP thresholds but PVR >2 Wood units; associated with especially poor outcomes
• No therapies directly target the pulmonary vascular component of PH associated with left heart failure — including drugs that work in WHO group 1 pulmonary arterial hypertension
• Prior pulmonary-artery denervation studies (PADN-1 first-in-man 2013 in PAH; PADN-5 2019 in CpcPH with left HF, with 3-year outcomes in 2023) suggested haemodynamic and clinical benefit but were underpowered and did not include contemporary medical management

Methods

• Design: open-label, randomized, controlled trial of patients with PH secondary to left heart disease with symptomatic heart failure (WHO group 2 PH)
• Population: 25 sites in China; age 18–75; chronic symptomatic heart failure (NYHA class II–IVa, any LVEF) for ≥3 months; on guideline-directed medical therapy at stable maximal dose for ≥1 month; 6-minute walk distance 100–450 m; NT-proBNP >300 pg/mL or BNP >100 pg/mL
• Key exclusions: any cause of PH other than left heart disease, or receipt of a WHO group 1 PAH-specific drug; cardiac index <1.5 L/min/m²; severe valve disease; eGFR <30 mL/min/1.73 m²
• Confirmation by right heart catheterisation in all patients: mPAP >20 mm Hg and PAWP >15 mm Hg
• Intervention: transcatheter pulmonary-artery denervation plus GDMT, versus GDMT alone. Nonselective pulmonary angiography in anteroposterior and 20° left anterior oblique projections, both with 20° cranial angulation. Weight-adjusted unfractionated heparin (bolus 70–100 IU/kg) to maintain ACT >250 s
• Device and protocol: a pigtail-like catheter with 10 distal electrodes, sized 1.1–1.2× the diameter of the main pulmonary artery. Ablation at three sites at the distal main pulmonary artery and the left pulmonary-artery bifurcation. Radiofrequency energy 150 seconds per site under temperature control (50 °C), with effective ablation defined as ≥120 seconds. Pressures remeasured 10 minutes after the procedure. In the medical-therapy group, haemodynamic measurements were taken at baseline and after 10 minutes with no additional procedure
• Post-procedure medication: dual antiplatelet therapy (aspirin plus clopidogrel) recommended unless oral anticoagulation was required
• Follow-up: clinical visits at 1, 6, 12, 24 and 36 months; 6-minute walk distance at 6 and 12 months and again 6 months after the last patient was enrolled
• Primary outcome — clinical worsening: composite of death, heart or lung transplantation, hospitalisation for heart failure, outpatient worsening of heart failure leading to intravenous diuresis, or a decline in 6-minute walk distance of ≥10% or >30 m
• Key secondary outcomes: individual components of clinical worsening; clinical worsening without the 6-minute walk distance component; outpatient worsening leading to intravenous medication or oral diuretic intensification; changes in echocardiographic measures, NT-proBNP and the Kansas City Cardiomyopathy Questionnaire overall summary score (KCCQ-OSS); procedural complications
• Blinding: of necessity, all patients, physicians, research personnel and outcome assessors knew the group assignments. Echocardiographic and haemodynamic measurements and NT-proBNP were analysed centrally in a blinded manner
• Statistics: sample size based on an assumed 24% risk of clinical worsening with medical therapy and 11% with denervation, 18-month accrual and 5% annual dropout; 54 events in 264 patients gave 87% power at two-sided alpha 0.05. Intention-to-treat. Kaplan–Meier estimates with Cox proportional hazards; primary analysis adjusted for heart-failure type (reduced vs mildly reduced vs preserved LVEF), PH phenotype, and geographic region in China. Recurrent events compared with the Wei–Lin–Weissfeld method. Secondary end points were not adjusted for multiplicity

Results — Patients

• 363 patients screened (14 Aug 2023 – 18 Feb 2025); 264 randomized — 134 (50.8%) to denervation plus GDMT, 130 (49.2%) to GDMT alone. Screening losses: 43 clinical exclusions, 51 haemodynamic exclusions (of which 33 had both mPAP ≤20 and PAWP ≤15 mm Hg)
• Median age 66.0 years (IQR 56.0–70.0); 49.2% women; median heart-failure duration 26.2 months (10.8–66.9); 50.8% hospitalised for heart failure in the previous year; 98.5% NYHA class II or III
• LVEF preserved (≥50%) in 53.0%; 40.3% vs 31.5% had HFrEF and 47.8% vs 58.5% had HFpEF across the two arms
• Haemodynamics: median mPAP 30.0 mm Hg (26.0–39.0); median PAWP 19.0 mm Hg (17.0–24.0); median PVR 2.3 Wood units (1.5–3.7). 44.7% IpcPH, 55.3% CpcPH
• Background therapy was contemporary and near-complete: beta-blocker ~84%, ARNI/ACEI/ARB ~98% (ARNI ~52%), mineralocorticoid-receptor antagonist ~81%, SGLT2 inhibitor ~77%, loop diuretic ~75–82%
• Median NT-proBNP 958 vs 883 pg/mL; median 6-minute walk distance 338 vs 345 m; median KCCQ-OSS 68 vs 70

Results — Procedure

• Denervation was completed successfully in all 134 patients. Most had no or minimal discomfort; only 1.5% received additional analgesia
• Median procedural time 101 minutes (86–120); median ablation time 7.1 minutes (7.0–7.7)
• Safety: access-site haematomas in 2 patients (denervation) vs 1 (medical therapy); no other procedural complications of any kind — no perforation, arrhythmia, acute pulmonary oedema or thrombus reported. Adverse events during follow-up occurred with similar frequency in both groups
• Acute haemodynamic response at 10 minutes: mPAP fell by ≥1% in 87.9% of patients and by ≥10% in 53.4%; pulmonary vascular resistance and PAWP also fell acutely and cardiac output rose. Intracardiac pressures were unchanged at 10 minutes in the medical-therapy group

Results — Primary and Secondary Outcomes

Median follow-up 338 days (IQR 218–467); 6-month clinical follow-up completed in all patients.

• Restricted mean time free from clinical worsening: 19.6 vs 16.7 months — a difference of 87.6 days (95% CI 26–150)
• Multiple imputation, leave-one-component-out, causal sensitivity, and missing-not-at-random analyses are reported as supporting the primary result; the treatment effect was consistent regardless of site enrolment size
• Secondary outcomes were not adjusted for multiplicity and the article states their confidence intervals "should not be used to infer treatment effect"
• The 6-minute walk distance component was measured in 227 patients at a median of 8.7 months; follow-up walk distance was not measured in 13 denervation patients and 24 medical-therapy patients
• Notably, loop diuretic use during follow-up was higher in the medical-therapy group, so the benefit was observed despite less diuretic exposure in the intervention arm

Results — Subgroups (all unadjusted for multiplicity; exploratory)

• The trial's strongest signals sit in subgroups that are not its stated target: IpcPH rather than CpcPH, and LVEF >40% rather than ≤40%. The authors explicitly describe the possibly greater effect in women as uncertain and warranting additional study

Discussion

• The authors state the relative frequency of clinical worsening was reduced by 52% even after removal of the 6-minute walk distance from the composite, and the annualized rate of recurrent events by 35.5%, with "treatment of only 3.9 patients required to prevent one heart-failure–related event"
• Mechanism proposed: sympathetic nervous system activation is a hallmark of both heart failure and PH; increased sympathetic activity contributes to pulmonary-artery vasoconstriction and remodelling, RV–PA uncoupling, and progressive RV dysfunction. Experimental work shows bidirectional interaction between pulmonary vascular sympathetic signalling and systemic neurohumoral activation
• Acute and sustained reductions in pulmonary-artery pressure are proposed to reduce RV afterload while improving LV filling and ventricular interaction, lowering left-sided filling pressures
• Supporting observational data within the trial: serial echocardiography showed improved biventricular function, lower RV pressures and reduced tricuspid regurgitation at 6 months; natriuretic peptide levels were lower than with medical therapy alone
• Prior smaller sham-controlled randomized experience in CpcPH is cited as concordant
• A large-scale confirmatory international sham-controlled trial in CpcPH secondary to left heart failure has recently begun (NCT07214376)

Limitations of the document

Authors' own stated limitations

• No sham control and no follow-up right heart catheterisation. Although the clinical-events committee and core laboratories were blinded, "the treatments were known to the patients, caregivers, and outcome assessors; as such, placebo effects and ascertainment bias cannot be ruled out." The authors argue that changes in cardiac structure and function and in natriuretic peptide levels are less prone to bias — but those were measured as secondary outcomes, not the primary one
• Not powered for the individual components of clinical worsening, such as death or hospitalisation for heart failure. Secondary outcomes and subgroup findings were "neither powered nor adjusted for multiple comparisons, so those findings should be considered exploratory." The authors single out the possibly greater treatment effect in women as uncertain
• No serial invasive haemodynamic assessment during follow-up — the sustained-reduction-in-PVR claim rests on prior studies, not on this trial's own data
• Learning curve and site/operator experience unknown, though effects were similar between the highest- and lowest-enrolling sites
• All patients enrolled in China — confirmation in multinational populations is required to establish external validity
• Additional long-term follow-up is needed to determine whether midterm improvements translate into sustained reductions in mortality and heart-failure hospitalisation

Additional limitations evident in the design

• The primary composite is heavily weighted toward soft, unblinded endpoints. Of its five components, one is an open-label 6-minute walk distance and another is outpatient worsening leading to diuresis — both susceptible to ascertainment bias in an unblinded trial where patients and assessors know their allocation. The 6MWD-free analysis holding (HR 0.48) is reassuring, but it still contains the outpatient-diuresis component
• The individual hard endpoints were null. Death from any cause: HR 0.83 (0.19–1.74). Heart-failure hospitalisation: HR 0.56 but with a confidence interval crossing 1 (0.30–1.04). Cardiovascular death: 3 vs 3 events
• Follow-up walk-distance data were missing in 37 of 264 patients (13 denervation, 24 medical therapy) — asymmetrically, and in the direction that would favour the intervention
• A pre-randomization 10-minute haemodynamic assessment was performed only in the intervention arm after the procedure; the medical-therapy group's 10-minute measurement involved no procedure. The acute haemodynamic "response" is therefore not a blinded comparison
• Industry involvement is substantial. Funded by Pulnovo Medical (grant Pulnovo-CO-2022-01), the manufacturer of the ablation catheter, which also supplied the catheters and employs a co-author. The company states it had no role in design, data collection or analysis, manuscript preparation or submission. Additional funding from the Jiangsu Provincial Special Program of Medical Science and the National Natural Science Foundation of China
• The trial's target population and its strongest subgroup signals diverge — see the subgroup table above. The stated rationale rests on CpcPH being the phenotype with the worst outcomes and the one where pulmonary vascular remodelling matters, yet the larger effect appears in IpcPH, where PVR is ≤2 Wood units by definition and a pulmonary-vascular mechanism is least plausible. This tension is not addressed in the discussion
• Publication status: published online 30 August 2026; the version ingested is the online-first article (11 pages)

Key Concepts Mentioned

• concepts/Pulmonary-Artery-Denervation — the therapy tested; this is its first adequately-powered RCT
• concepts/Pulmonary-Hypertension — WHO group 2 (PH due to left heart disease), IpcPH vs CpcPH, and the absence of any approved pulmonary-vascular therapy in group 2
• concepts/RV-PA-Coupling — the proposed mechanism runs through sympathetic-driven RV–PA uncoupling, though no follow-up catheterisation was performed to demonstrate it
• concepts/Renal-Denervation — the same procedural-denervation space, and the reason a sham-controlled result is the standard the field is now held to
• entities/Heart-Failure — guideline-directed medical therapy as the comparator arm; device therapy in symptomatic HF despite optimal medical therapy
• Sympathetic overactivation — the mechanistic premise: neurohormonal activation in HF driving pulmonary-artery vasoconstriction and remodelling
• KCCQ-OSS — health-status measure used as a secondary outcome
• 6-minute walk distance — both a trial entry criterion (100–450 m) and a component of the primary composite

Key Entities Mentioned

• entities/Heart-Failure — primary entity
• entities/HFpEF — 53% of the trial population had preserved LVEF, and the LVEF >40% subgroup showed the larger effect
• entities/HFrEF — 40.3% vs 31.5% of the two arms had reduced LVEF
• concepts/Pulmonary-Hypertension — WHO group 2 as the disease category
• Nanjing First Hospital, Nanjing Medical University; General Hospital of Northern Theater Command, Shenyang — coordinating centres
• Pulnovo Medical — device manufacturer, funder, catheter supplier, and employer of a co-author

Wiki Pages Updated

• Created: wiki/sources/padn-hf-ph-nejm-2026.md
• Created: wiki/concepts/Pulmonary-Artery-Denervation.md
• Updated: wiki/concepts/Pulmonary-Hypertension.md
• Updated: wiki/entities/Heart-Failure.md
• Updated: wiki/entities/HFpEF.md
• Updated: wiki/concepts/RV-PA-Coupling.md
• Updated: wiki/concepts/Renal-Denervation.md
• Updated: wiki/sourceindex.md
• Updated: wiki/wikiindex.md