Pulmonary-artery denervation cut heart-failure worsening in an open-label trial

A catheter procedure that burns away nerves lining the pulmonary artery roughly halved the rate of heart-failure worsening in a Chinese trial, according to results that look impressive on the surface — but the trial's design leaves an important door open to doubt: patients knew which treatment they had received [s1]. In a condition whose main endpoints depend partly on how patients feel and how often they are admitted, that lack of blinding is not a footnote.

The problem it targets

Some people with heart failure develop pulmonary hypertension — high blood pressure in the lungs' arteries — because the failing left side of the heart backs pressure up into the lungs [s1]. One theory holds that overactive sympathetic ("fight or flight") nerves make this worse, tightening the pulmonary vessels, straining the right side of the heart, and driving poor outcomes [s1]. Pulmonary- artery denervation is an attempt to interrupt those nerves from inside the vessel with a catheter, much as renal denervation targets nerves near the kidneys for blood pressure. Whether it helps patients with left-heart-related pulmonary hypertension had not been established [s1].

What the trial did

The trial, PADN-HF-PH, was a multicentre randomised study conducted in China [s1]. Patients with pulmonary hypertension associated with left heart disease and heart failure were assigned 1:1 to receive pulmonary-artery denervation plus guideline-directed medical therapy, or guideline-directed medical therapy alone [s1]. The primary outcome was clinical worsening — a composite of death, heart or lung transplantation, hospitalisation for heart failure, outpatient worsening of heart failure, or a decline in the six-minute walk distance — measured through the latest follow-up [s1].

The design detail that matters most: the control group received medical therapy alone, with no sham catheter procedure [s1]. Patients therefore knew whether they had undergone denervation, and several components of the primary endpoint — an outpatient judgement of "worsening," a decision to admit, a walk-test effort — are exactly the kind of soft, expectation-sensitive measures that unblinded trials can nudge.

What it found

A total of 264 patients were randomised: 134 to denervation plus medical therapy and 130 to medical therapy alone [s1]. Over a median follow-up of 338 days, the Kaplan–Meier estimated two-year incidence of clinical worsening was 25.7% in the denervation group versus 51.5% in the medical-therapy group — a hazard ratio of 0.49 (95% confidence interval 0.30 to 0.82; P=0.006) [s1]. On procedural safety, the intervention looked clean in the short term: access-site haematomas occurred in two patients in the denervation group and one in the medical-therapy group, with no other procedural complications, and adverse events during follow-up occurred with similar frequency in the two groups [s1].

How to read it

Taken at face value, a halving of clinical worsening is a large effect for a field with few options, and the procedure appeared safe over the follow-up reported [s1]. But three features argue for caution before treating this as practice-changing. First and most important, the absence of a sham control in an unblinded trial with subjective endpoints is a well-known route to overstated benefit; the honest reading is that some of the 25.7%-versus-51.5% gap could reflect the placebo power of having had a procedure rather than the denervation itself [s1]. Second, the trial was single-country and of modest size — 264 patients at a median of under a year of follow-up — so the durability and generalisability of the effect are unproven [s1]. Third, the endpoint blends hard outcomes such as death and transplantation with softer ones such as outpatient worsening and walk-test decline, and the summary does not tell us how much of the benefit came from the hardest components [s1].

Funding belongs in plain view here: the trial was funded in part by Pulnovo Medical, the company developing the denervation device [s1]. A commercially sponsored, unblinded device trial reporting a large benefit on a partly subjective composite is precisely the combination that calls for independent, sham-controlled confirmation before the result changes care. The protocol was registered before enrolment [s2].

For patients with heart failure and pulmonary hypertension, the practical message is one of cautious interest, not arrival: an early signal worth testing properly, not yet an established treatment. This article describes research and is not medical advice.

What to watch

The decisive test would be a larger, multinational, sham-controlled trial with blinded endpoint assessment and longer follow-up — the same standard that reshaped the renal-denervation field after its own early open-label enthusiasm ran ahead of the evidence [s1]. Until such a trial reports, pulmonary-artery denervation for heart-failure-related pulmonary hypertension is best read as a promising hypothesis under active investigation.

Sources

• Pulmonary Denervation for Heart Failure–Related Pulmonary Hypertension — New England Journal of Medicine, 30 August 2026
• PADN-HF-PH trial registration (NCT05824923) — ClinicalTrials.gov, first posted 24 April 2023