Asthma is one of the best-documented applications: breathing and heart-rate-variability training reduce symptoms and medication use. This never replaces controller medication — the most direct trial puts the biofeedback effect at about two thirds of a month of inhaled steroids.
AAPB5Efficacious and specificNeuroLogic5Efficacious and specificsame vs AAPB
Level 5 maintained. Lehrer 2004 (n = 64, three months of HRV biofeedback with or without abdominal breathing vs a biofeedback placebo and a waiting list) found greater reductions in respiratory resistance measured by forced oscillation, less asthma medication consumption, fewer exacerbations and a greater fall in an overall asthma severity index; symptoms fell in the placebo arm as well and spirometry did not change (AAPB ch. 5). The earlier three-arm pilot (Lehrer, Carr 1997, n = 17) had already shown respiratory resistance falling only in the HRV arm, and Lehrer 2006 found the effect at least as large above age 40. Lehrer 2018 (n = 68 steroid-naive patients, HRV biofeedback vs a combined stress-management package of EEG alpha biofeedback, paced breathing and music) improved both arms, with exhaled nitric oxide falling only under HRV biofeedback; the biofeedback effect was 65 % of the steroid effect on the methacholine challenge and 81 % on exhaled nitric oxide. Taghizadeh 2019 showed a single session blunting stress-induced lung-function impairment against a behavioural placebo. The acknowledged weakness, stated by the AAPB authors themselves, is that most of this work comes from one laboratory; the 2022-2026 window added no new asthma trial.
The Buteyko method (breathing retraining, without instrumental biofeedback)
AAPB5Efficacious and specificNeuroLogic5Efficacious and specificsame vs AAPB
Level 5 maintained. Randomised superiority over credible active comparators in independent settings, on symptoms, quality of life and medication use. Cowie 2008 (Canada, n = 65 and 64) compared Buteyko with training in relaxed breathing: large improvement in asthma control and quality of life in both arms over six months, with significantly less steroid medication in the Buteyko arm and no spirometry change. Cooper 2003 (United Kingdom, n = 69) compared Buteyko with breathing through a Pink City Lung exerciser as a pranayama analogue. Bowler 1998 (Australia, n = 39, described as blinded) found a significant fall in albuterol use against asthma education, with only borderline changes in steroid use and quality of life and no change in pCO2. Opat 2000 (n = 36) taught Buteyko by video against a placebo video and found better quality of life and less bronchodilator use. Prem 2013 (India) found Buteyko superior to pranayama and to treatment as usual on asthma control and quality of life. Hepworth 2019 improved self-reported control in 169 children. The method does not improve pulmonary function.
Muscle relaxation training and EMG biofeedback (mainly frontalis)
AAPB4EfficaciousNeuroLogic4Efficacioussame vs AAPB
Level 4 maintained. Eight randomised trials, several of them sham-controlled, show frontalis EMG biofeedback improving airflow. Davis 1973: frontalis EMG biofeedback vs progressive muscle relaxation vs a reading-and-instruction control over five sessions — peak expiratory flow rose only under biofeedback and fell in the control. Kotses 1976 (13 children per arm: genuine biofeedback, sham biofeedback, waiting list) and Kotses 1978 (10 children per arm: frontalis, brachioradialis, non-contingent feedback) each found peak expiratory flow rising only with genuine frontalis feedback. Scherr and Crawford 1978 ran three camp trials in consecutive years, 10-12 children per arm, and found greater peak-flow improvement with biofeedback than with false feedback or no treatment every year. Lehrer 1986 (n = 11, relaxation with frontal and trapezius EMG feedback vs a behavioural placebo) improved the methacholine challenge, most in patients with greater upper-airway impairment. Kern-Buell 2000 (n = 16 vs waiting list) improved FEV1/FVC, symptoms and medication use but not daily peak flow. Kotses 1991 (n = 15 per arm) found no meaningful spirometry change. Effect sizes are small but consistent.
AAPB4EfficaciousNeuroLogic4Efficacioussame vs AAPB
Level 4 maintained. Yüce and Taşci 2020, a single-blind randomised trial in 55 adults with asthma, found higher asthma control and asthma-related quality of life after one month of pranayama than in an active control group given relaxation training, with no between-group difference in pulmonary function tests. Saxena and Saxena 2009 trained 25 people over 12 weeks against an active control and found FEV1 and peak flow rate improved. Prem 2013 found pranayama better than treatment as usual on asthma-related quality of life, though weaker than the Buteyko arm. The evidence on lung function is inconsistent: Cooper 2003 gave 30 participants a single pranayama session and found no significant difference from a no-treatment control, and Singh 1990, using a Pink City Lung exerciser to impose a 1:2 inspiration-expiration ratio against placebo, improved FEV1, peak flow, symptoms and inhaler use only non-significantly, although the provocative dose of histamine required for a 20 % FEV1 fall increased. Symptom and quality-of-life benefit is replicated across independent groups; specificity is not established.
Direct biofeedback of lung activity; informational feedback on medication compliance
AAPB3Probably efficaciousNeuroLogic3Probably efficacioussame vs AAPB
Level 3 maintained for both halves of this row. Direct feedback of lung activity: Khan 1974 randomised 20 children to respiratory-resistance biofeedback or no treatment and reported fewer attacks, fewer emergency-room visits and less medication at one year; Mussell and Hartley 1988 gave three sessions of trachea-noise biofeedback to eight patients against a placebo drug (n = 8) and found a borderline-significant FEV1 improvement after exercise challenge. The case series are inconsistent — Steptoe 1981 (two of eight patients improved, two worsened), Erskine-Milliss and Cleary 1987 (nine patients, eight sessions, no significant effect), Mass 1993 (resistance fell in 15 patients) — and AAPB warns that patients may learn only to adjust the glottis. Informational feedback on adherence: Sulaiman 2018, a randomised trial in severe uncontrolled asthma, found greater adherence and greater clinical improvement with device feedback plus education than with education alone; O'Dwyer 2020 found the same for digital inhaler feedback in community pharmacies. Walsh 2022 (n = 88 from that trial) found similarly high self-efficacy in both arms and no relation between self-efficacy and objectively measured adherence, which does not lower the level.
End-tidal CO2 biofeedback; relaxation training with laryngeal biofeedback
Ages
Adult (18+)
Techniques
Respiratory / capnometry, Other biofeedback
AAPB2Possibly efficaciousNeuroLogic3Probably efficacioushigher vs AAPB
Level 3. The AAPB Level 2 rested on the absence of randomised trials, all work coming from one laboratory. Randomised data now exist: Meuret 2023 reports a trial in which 120 adults with asthma were randomised to capnometry-assisted respiratory training to raise PCO2 or to feedback to slow respiratory rate; among the patients with clinically relevant baseline anxiety, the capnometry arm produced greater reductions in anxiety sensitivity and negative affect at post-intervention (Cohen d values at least 0.58) and lower anxiety sensitivity, negative affect and depression scores at six months (d at least 0.54), while patients with low baseline anxiety showed no differential outcome. Kadam 2023, a meta-analysis of eight randomised trials of breathing retraining including capnometry biofeedback, found a significant large effect on end-tidal CO2 after the intervention and at three and six months, rated moderate-quality evidence. These sit on the AAPB's case series: Meuret 2007 (12 patients, two months, pCO2 normalised with fewer symptoms and less peak-flow variability), Ritz 2009 and Jeter 2012. Not higher, because the randomised asthma work remains one laboratory and the reported randomised outcomes are anxiety rather than asthma endpoints, and because the laryngeal-biofeedback half of this row still rests on two uncontrolled case series (Earles 2003; LeBlanc 2021, n = 25).
In short
Clinical reading
NeuroLogic levels: HRV 5 and Buteyko 5 (same as AAPB) for symptoms, quality of life and medication dose; muscle relaxation with EMG biofeedback 4 and pranayama 4 (same); direct biofeedback of lung activity and adherence feedback 3 (same); end-tidal CO2 and laryngeal biofeedback raised from 2 to 3 on the strength of the randomised capnometry trial in 120 adults (Meuret 2023) and the Kadam 2023 meta-analysis. Only HRV has improved pulmonary function across several studies, though mostly from a single laboratory.
Protocols
Resonance-frequency HRV (around 6 breaths/min), 10-15 supervised sessions with daily home practice; four weeks of capnometry with a portable capnometer; frontalis EMG relaxation; structured breathing protocols for Buteyko and pranayama; adherence feedback from an inhaler sensor.
Limits
Independent replication for pulmonary function is insufficient, most HRV and capnometry trials coming from two laboratories; muscle relaxation can have an opposite immediate effect through parasympathetic rebound; direct feedback of respiratory resistance remains the one strand AAPB judges insufficiently supported; paediatric data concern mainly EMG relaxation and Buteyko, not pranayama or capnometry.
Study base
Several RCTs per modality; robust symptom effects, more modest spirometry effects. 2022-2026 base: 1 publication indexed in the archive (Kadam 2023 meta-analysis), with the sweep additionally returning the Meuret 2023 secondary analysis.
Brendan's perspective
Five of the six rows sit exactly where the AAPB put them, and that is the honest result. The single move is capnometry and laryngeal work, 2 to 3, because Meuret 2023 finally gives that strand a randomised comparator — and look at the comparator: another breathing training. That is the right design. It is also why I do not read this file as more settled than it is. Almost all the HRV work is Lehrer's, and a level 5 built inside one laboratory is a level 5 with a question mark over independence: reproduced by whom else? The other level 5, Buteyko, feeds nothing back at all, which is worth remembering before we credit the instrument. Asthma is not an indication I see often, but resonance-frequency HRV is a skill I teach anyway, and I would run it beside the respiratory physician, never instead. The honest version for the client: symptoms, quality of life and medication use move; spirometry mostly does not. Controller medication stays.
Lehrer et al. (2018) Heart rate variability biofeedback does not substitute for asthma steroid controller medication doi:10.1007/s10484-017-9382-0
Cowie et al. (2008) A randomised controlled trial of the Buteyko technique as an adjunct to conventional management of asthma doi:10.1016/j.rmed.2007.12.012
Yüce & Taşci (2020) Effect of pranayama breathing technique on asthma control, pulmonary function, and quality of life: a single-blind, randomized, controlled trial doi:10.1016/j.ctcp.2019.101081
Sulaiman et al. (2018) A randomised clinical trial of feedback on inhaler adherence and technique in patients with severe uncontrolled asthma doi:10.1183/13993003.01126-2017
Meuret et al. (2023) Biofeedback Training to Increase P co2 in Asthma With Elevated Anxiety: A One-Stop Treatment of Both Conditions? doi:10.1097/PSY.0000000000001188
Kadam et al. (2023) Effect of breathing re-training on end-tidal carbon-di-oxide in patients with asthma: a systematic review and meta-analysis doi:10.1080/02770903.2022.2162413
