Biofeedback-assisted relaxation improves glycemic control across several randomised trials, where it outperformed diabetes education or supportive psychotherapy. Voluntary foot warming with thermal biofeedback also helps ulcers heal. These methods add to diabetes care and never on their own change an insulin prescription.
Updated :
What the research shows
Biofeedback
Assisted relaxation (frontal EMG, digital thermal) for glycemic control — type 1 and type 2 diabetes (adult)
AAPB5Efficacious and specificNeuroLogic5Efficacious and specificsame vs AAPB
Level 5 maintained. Six randomised trials of biofeedback-assisted relaxation added to diabetes care underpin the rating, four of them favourable and from independent settings. Surwit & Feinglos 1983: n = 12 with type 2 diabetes, five frontal EMG sessions against treatment as usual — glucose tolerance and 2-hour postprandial glucose improved only in the biofeedback group. Miley 1989: n = 21 inpatients, three arms — greater within-session blood-glucose reductions than nondirective psychotherapy or treatment as usual in weeks 1 and 2, comparable in week 3. McGrady 1991: n = 18 with type 1 diabetes, ten sessions of EMG and thermal biofeedback with counselling against counselling alone — lower blood glucose and a higher proportion of fasting values on target. McGinnis 2005: n = 30 with type 2 diabetes, ten sessions against three diabetes-education sessions — greater reductions in blood glucose, HbA1c and forehead EMG, maintained at 3 months. Two trials were null on metabolic outcomes: Lane 1993 (n = 38, no group difference in mean glycated haemoglobin) and Jablon 1997 (n = 20, electrodermal, EMG and anxiety measures improved but glucose tolerance, fasting and postprandial glucose and fructosamine did not).
Heart rate variability biofeedback (resonance-frequency breathing) — type 2 diabetes
Ages
Adult (18+)
Techniques
HRV — resonance-frequency breathing
AAPB0Not ratedNeuroLogic2Possibly efficacious
Level 2. AAPB names HRV biofeedback only as a modern home-practice option and reviews no HRV trial, so this row is new. Wu 2025: six weekly 60-minute HRV biofeedback sessions added to usual care in type 2 diabetes, n = 61 (30 biofeedback, 31 control; allocation method not stated) — significant group by time effects on HRV indices, breathing rate, depressive symptoms and diabetes self-care behaviour, but no interaction effect on HbA1c. Herhaus 2025: four weeks of mobile resonance-frequency breathing with HRV biofeedback in 30 healthy adults — no effect on glucose metabolism overall, with lower fasting and 2-hour post-load glucose in the participants who were more insulin resistant beforehand, and higher time- and frequency-domain HRV. Gitler 2022 rates the evidence for HRV biofeedback in diabetes as poor, against high-quality evidence in coronary heart disease, pain and hypertension. One adequately sized controlled study in patients, non-randomised and without a glycaemic effect, supports Level 2 and no more.
Thermal biofeedback for healing of diabetic foot ulcers
Ages
Adult (18+)
Techniques
Thermal biofeedback
AAPB3Probably efficaciousNeuroLogic3Probably efficacioussame vs AAPB
Level 3 maintained. Rice 2001: 32 patients with chronic non-healing lower-extremity ulcers assigned to one session of thermal biofeedback-assisted relaxation with instruction in vascular physiology and five days a week of home practice, or to their own preferred relaxation method — the biofeedback group significantly increased toe temperature, with the training accounting for 64 % of the variability in temperature, and 14 of 16 ulcers (88 %) healed against 7 of 16 (44 %) in the control group. Fiero 2003: 24 patients, six weekly sessions — foot temperature rose by 1.2 °C on average despite mild-to-moderate neuropathy, but lower-extremity sympathetic-autonomic and sensory neuropathy accounted for 41 % of the variance in foot-warming and limited skill acquisition. Shulimson 1986 is a three-patient series in which two ulcers healed or nearly healed and one did not. A single randomised trial from one setting, not replicated; no 2022-2026 study of thermal biofeedback for diabetic ulcers passed screening.
Thermal biofeedback for intermittent claudication and lower-extremity perfusion
Ages
Adult (18+)
Techniques
Thermal biofeedback
AAPB2Possibly efficaciousNeuroLogic2Possibly efficacioussame vs AAPB
Level 2 maintained. Rice & Schindler 1992 is the only group study: 40 adults with diabetes (21 type 1, 19 type 2) in a two-phase within-subject design — four weeks of relaxation by the patient's own preferred method, then four weeks with temperature monitoring of the great toe and recorded relaxation instructions. Toe temperature and blood volume pulse rose more in the second phase, but the design is confounded by order, since every participant received the phases in the same sequence, and the second phase monitored temperature without ever feeding it back, so biofeedback as such was not tested. The remaining evidence is two case studies: Aikens 1999 (five sessions of toe temperature feedback in a woman with type 2 diabetes — toe temperature up 0.9 to 1.7 °C, with improved walking distance, speed, stair climbing and lower-extremity blood pressure) and Saunders 1994 (21 sessions; no claudication episodes from session 12 onward, 4.5 miles walked daily at 12- and 48-month follow-up). No randomised trial, and no 2022-2026 publication passed screening.
In short
Clinical reading
AAPB levels maintained by NeuroLogic: 5 for glycemic control (Surwit & Feinglos 1983, Miley 1989, McGrady 1991, McGinnis 2005 against bona fide comparators; two null trials, Lane 1993 and Jablon 1997), 3 for ulcer healing (Rice 2001, 88 % vs 44 % healed), 2 for intermittent claudication. New row: HRV biofeedback Level 2 (Wu 2025, n = 61, self-care and HRV improved but HbA1c unchanged).
Protocols
EMG-assisted relaxation and digital thermal biofeedback, five to ten weekly sessions with twice-daily home practice; toe temperature feedback for ulcers and claudication; resonance-frequency breathing with HRV biofeedback. Modern portable devices (EMG, HRV, temperature) make home training more reliable than in the days of alcohol thermometers.
Limits
The trials are old (the most recent is from 2005) and small (12 to 38 patients), and several never verified that patients learned to lower EMG or warm their fingers — the two null trials among them. Neuropathy limits the acquisition of foot warming. Claudication rests on one order-confounded study and two case reports. Nothing published in children or adolescents.
Study base
Six RCTs on glycemic control, one RCT on ulcers, one quasi-experiment on claudication and several case studies. 2022-2026 base: one publication indexed in the archive (HRV biofeedback).
Brendan's perspective
Level 5 for glycaemic control, unchanged, and I think the number is right: four favourable trials from independent settings against comparators that were doing real work — diabetes education (McGinnis 2005), nondirective psychotherapy (Miley 1989), counselling (McGrady 1991). What holds me back is that the two null trials, Lane 1993 and Jablon 1997, are exactly the ones that never verified that patients had learned to lower frontal EMG or warm their fingers. Without a learning check you do not know what was actually delivered. The HRV row is new and stays at level 2: Wu 2025 improved heart rate variability and self-care behaviour, not HbA1c, and I see no reason to dress that up. In clinic, diabetes is not a presenting complaint I see; where it sits alongside something else I train resonance-frequency breathing, and digital temperature if the feet are involved, always beside the diabetes team. None of this changes an insulin prescription.
Surwit & Feinglos (1983) The effects of relaxation on glucose tolerance in non-insulin-dependent diabetes doi:10.2337/diacare.6.2.176
McGrady, Bailey & Good (1991) Controlled study of biofeedback-assisted relaxation in type I diabetes doi:10.2337/diacare.14.5.360
Rice et al. (2001) Effect of biofeedback-assisted relaxation training on foot ulcer healing doi:10.7547/87507315-91-3-132
Rice & Schindler (1992) Effect of thermal biofeedback-assisted relaxation training on blood circulation in the lower extremities of a population with diabetes doi:10.2337/diacare.15.7.853
Wu et al. (2025) Effect of Heart Rate Variability Biofeedback on Cardiac Autonomic Activation and Diabetes Self-Care in Patients with Type II Diabetes Mellitus doi:10.1007/s10484-024-09666-x
