Medical hypnosis versus structured relaxation as adjunct to pulmonary rehabilitation for anxiety in chronic obstructive pulmonary disease (HYPNOBPCO_2): a cluster-randomised, active-comparator trial
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Anxiety affects 22–56% of patients with COPD. It is independently associated with increased exacerbations, readmissions, and mortality. Medical hypnosis transiently alleviates state anxiety in severe COPD, and attenuates experimentally-induced dyspnoea in healthy volunteers. We aimed to assess the efficacy of hypnosis as adjunct therapy for anxiety during Pulmonary Rehabilitation (PR) versus an active comparator controlling for general motivation and relaxation.
Methods
HYPNOBPCO_2 was a single-centre, phase 2, cluster-randomised, active-comparator, parallel-group, superiority trial done at Centre Hospitalier de Bligny, France. Adults ≥30 years with established COPD, mMRC dyspnoea grade ≥2, and ≥10 pack-years were eligible. Consecutive pulmonary rehabilitation cohorts (clusters) were randomly assigned (1:1) to medical hypnosis or structured relaxation, both adjunctive to a 4-week inpatient PR. The primary outcome was the six-item State-Trait Anxiety Inventory (STAI-6) at week 4, analysed in the intention-to-treat population. Secondary outcomes were sensory and affective dyspnoea (Multidimensional Dyspnea Profile sensory and affective sub-scales, COPD Assessment Test) and functional capacity (6-minute walk distance). A moderation analysis tested whether the hypnosis effect varied with baseline sensory and affective dyspnoea burden using Bayesian inference. This trial was registered prospectively ( NCT04868357 ) and the protocol published.
Findings
Between 27/09/2021 and 31/01/2024, 79 participants in 24 clusters were randomised (medical hypnosis n = 36, age 64·9 [8·3], 20 female; relaxation n = 43, age 67·4 [9·1], 23 female). Anxiety improved in both arms (ΔSTAI-6 = −5·72, posterior probability of reduction beyond MCID = 0·93). Medical hypnosis did not show an advantage over relaxation overall (ΔSTAI-6 = 0·95; posterior probability of reduction beyond MCID = 0·12). Medical hypnosis did show an advantage that scaled with baseline dyspnoea profile: in patients with high sensory and low affective burden, the predicted anxiety reduction relative to relaxation was larger and more probable (ΔSTAI-6 = −9·26; posterior probability of reduction beyond the 3-point MCID = 0·81). No clinically important safety issues were associated with either intervention and there were no deaths.
Interpretation
Anxiety improved under both adjunctive mind-body interventions. The usefulness of hypnosis beyond relaxation was predicted by baseline dyspnoea profile, concentrating in patients with predominantly sensory burden. Future trials should focus on evaluating complementary interventions against burden phenotypes, ultimately paving the way for personalized interventions.
Funding
Helebor Foundation; Agence Nationale de la Recherche.
Research in context
Evidence before this study
We searched PubMed, Embase, ClinicalTrials.gov , and ISRCTN.com from database inception to December 1st 2025, using combinations of the terms “COPD”, “anxiety”, “hypnosis”, “relaxation”, “pulmonary rehabilitation”, and “dyspnoea”, with no language restrictions. Pulmonary rehabilitation produces a moderate anxiolytic effect in COPD (SMD −0·53, 95% CI −0·82 to −0·23) but with marked between-patient heterogeneity that is not predicted by demographic or disease-severity variables. Adjunctive mind–body interventions, including hypnosis, cognitive-behavioural therapy, and mindfulness-based interventions, have been studied as augmentations to pulmonary rehabilitation, but prior evidence is limited by waitlist or treatment-as-usual comparators that conflate intervention-specific effects with non-specific benefits of structured psychological contact. A single-session trial of medical hypnosis in inpatient COPD in 2020 reported an immediate anxiolytic effect, and laboratory-induced dyspnea work in healthy volunteers in 2024 demonstrated that hypnosis attenuates both the sensory and affective dimensions of experimentally induced dyspnoea. A single-centre open-label trial (n=106) in 2026 comparing hypnosis as an adjunct to pulmonary rehabilitation against pulmonary rehabilitation alone reported secondary signals on dyspnoea-related anxiety and physical activity at 6 months without meeting its prespecified primary endpoint, but results were muddled as the design could not isolate hypnosis from the non-specific effects of relaxation, motivation or additional psychological contact. At the time of our search, one further registered trial in this area ( NCT07173348 ) had not begun recruitment and has not reported results.
Added value of this study
HYPNOBPCO_2 is, to our knowledge, the first medical hypnosis trial in any clinical indication to satisfy the methodological standards established by the 2024 Task Force for Establishing Efficacy Standards for Clinical Hypnosis: pre-registered protocol, active comparator matched on non-specific intervention effects, blinded outcome assessment, multi-instrument outcome construct, prespecified metrics, and item-level handling of missing data. It is also the first in any setting to prospectively test whether hypnotic anxiolytic response is moderated by baseline dyspnoea sensory and affective burden, providing a viable, targeted alternative for anxiety management in COPD. By using structured relaxation as the active comparator matched on session length, structure, practitioner pool, and self-management framing, the trial isolates the specific cognitive effects of suggestion present in medical hypnosis from non-specific psychological-contact effects in a context close to ecological rehabilitation conditions. At the population level (all participants pooled), adjunctive medical hypnosis did not differ from adjunctive structured relaxation on the prespecified primary outcome or across the confirmatory anxiety construct. However, a continuous moderation analysis based on baseline sensory and affective dyspnea burden revealed a clear sensory gradient on hypnotic advantage across the primary outcome and confirmatory anxiety metrics (posterior probability of the predicted positive interaction coefficient 0·91 on STAI-6 and 0·985 on MDP-A2). At the high-sensory/low-affective end of the burden spectrum, the predicted between-arm contrast on the primary outcome was −9·26 STAI-6 points, i.e. three times the minimal clinically important difference, in so identifying a candidate responder phenotype that warrants further study and options for targeted treatment.
Implications of all the available evidence
Adjunctive medical hypnosis cannot be recommended over structured relaxation for unselected COPD patients in pulmonary rehabilitation. However, it may offer a clinically meaningful anxiolytic advantage for the subgroup with predominantly sensory rather than affective dyspnoea burden, a candidate responder phenotype that requires prospective, stratified confirmation with the potential to guide practice if confirmed by further research.
