Methods for Heart Rate Variability Biofeedback (HRVB): A Systematic Review and Guidelines

Abstract

Heart Rate Variability Biofeedback (HRVB) has been widely used to improve cardiovascular health and well-being. HRVB is based on breathing at an individual's resonance frequency, which stimulates respiratory sinus arrhythmia (RSA) and the baroreflex. There is, however, no methodological consensus on how to apply HRVB, while details about the protocol used are often not well reported. Thus, the objectives of this systematic review are to describe the different HRVB protocols and detect methodological concerns. PsycINFO, CINALH, Medline and Web of Science were searched between 2000 and April 2021. Data extraction and quality assessment were based on PRISMA guidelines. A total of 143 studies were finally included from any scientific field and any type of sample. Three protocols for HRVB were found: (i) "Optimal RF" (n = 37), each participant breathes at their previously detected RF; (ii) "Individual RF" (n = 48), each participant follows a biofeedback device that shows the optimal breathing rate based on cardiovascular data in real time, and (iii) "Preset-pace RF" (n = 51), all participants breathe at the same rate rate, usually 6 breaths/minute. In addition, we found several methodological differences for applying HRVB in terms of number of weeks, duration of breathing or combination of laboratory and home sessions. Remarkably, almost 2/3 of the studies did not report enough information to replicate the HRVB protocol in terms of breathing duration, inhalation/exhalation ratio, breathing control or body position. Methodological guidelines and a checklist are proposed to enhance the methodological quality of future HRVB studies and increase the information reported.

Keywords: Checklist; Deep breathing; Guidelines; Heart Rate Variability Biofeedback; RSA; Resonance frequency.

Fig. 1

Flowchart of the different phases of the searching and selection of studies, following the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). RF resonance frequency. See table for details on the exclusion criteria

Fig. 2

Type of HRVB intervention (n = 145). Optimal RF: interventions that followed the original protocol by Lehrer et al. (2000) and applied a range of breathing rates before each HRVB session or for the total HRVB intervention in order to detect the optimal resonance frequency (RF) for each participant. Individual: those interventions that used a biofeedback system that indicates the actual breathing rhythm and the optimal breathing rhythm in real time for each participant. Thereby, each participant individually adapted the breathing rate to the optimal rhythm or rate indicated by the biofeedback system. Preset-Pace: a fixed breathing ratio is set up for all participants and all HRVB sessions, usually 6 b/m. Referenced: those studies that did not report enough information to determine the type of HRVB intervention, because the authors referred to an HRVB intervention from a previous study

Fig. 3

Characteristics of the HRVB intervention by type of HRVB intervention. Based on the four categories of HRVB interventions (columns for: Optimal RF, Individual, Preset-Pace, and Referenced), we indicated in rows: (1) the total duration of the HRVB intervention in weeks; (2) the number of laboratory sessions per week; (3) minutes of exclusively breathing during laboratory HRVB sessions (in green those studies that only applied home sessions, in purple those studies that only reported the duration of the entire session, but not the breathing time specifically); (4) daily minutes of exclusively breathing during the home HRVB sessions (in green those studies that only applied laboratory sessions). The numbers (n) in the bars indicate the number of studies/interventions per each category

Fig. 4

Design and control group of the included studies and interventions (n = 143). A Study designs: Experimental/BS (between-subjects): studies that applied an experimental (randomized and/or equivalent control group) between-subjects design (each participant was assigned to one experimental group). Experimental/WS (within-subjects): those studies that applied an experimental within-subjects design (each participant was assigned to one order of the experimental groups). Quasi-Experimental: those studies that applied a quasi-experimental design (non-randomized and/or non-equivalent control group). B The type of control group/s or condition/s. Active: participants did something active such as breathing without a biofeedback screen, writing in a diary or running. Passive: participants did not do anything active, they were on a waiting list or sat in the laboratory. Studies that only reported a control group were included in this category. Both: studies that included two or more control groups/conditions and at least one was active, and one was passive. No control group: this category included quasi-experimental studies that did not have a control group. Exp/WS (within-subjects): studies in which all participants received both the control and the experimental conditions. The numbers (n) in the graphs indicate the number of studies per each category

Fig. 5

Characteristics of the samples used in the included studies and experiments (n = 143). A Type of sample divided between those studies that included students and those that did not include students. B Age of the samples based on the reported “mean age” or “age range”. Children & Adolescents: less than 18 years, Early Adulthood: between 18 and 39 years, Middle Adulthood: between 40 and 55 years, Late Adulthood: between 56 and 65 years, Seniors: more than 65 years. C Total males and females considering the final sample of each study. The numbers (n) in the graphs indicate the number of studies per each category. ¹Four studies did not report the biological sex of the participants (n = 139)

Fig. 6

Methodological features and co-variables reported in the selected papers. RF Range: the breathing rates or range applied to detect the optimal resonance frequency (RF). Minutes RF range: the minutes of each breathing rate from the “RF Range”. Check RF each Ses.: the optimal RF was assessed before each HRVB session; “yes” instead of “reported”. Mean RF: the mean of the RF in the total sample. RF/Preset-Pace controlled: any attempt to ensure that participants/patients breathed at the indicated RF or preset breathing rate. Inhalation/Exhalation Ratio: the ratio of inhalation, hold and exhalation during breathing. Body Position: the body position during the HRVB sessions. Eyes: open or closed eyes during the HRVB intervention. In studies that used a screen or a display, “open eyes” are inferred. Room: room conditions including: light, noise, smell and temperature; it was considered “reported” if just one condition was reported. Time of day: the time of day when the HRVB sessions were performed. Num. participants at the same time: how many participants, clients or patients were together during an HRVB session. Instructions: recommendations given to for participants before each HRVB session, like no smoking 3 h before the session. Control of the instructions: it was considered “reported” if just one of the instructions were reported in the study. n = 145, except for: ¹only for “Optimal RF”, n = 37; ²only for “Optimal RF” and “Individual”, n = 85; ³only for “Optimal RF” and “Preset-Pace”, n = 88; ⁴only for those studies that reported any pre-HRVB instruction, n = 43. The numbers (n) in the graphs indicate the number of studies per each category

Fig. 7

Risk of Bias for the experimental and quasi-experimental studies/interventions. A Risk of Bias for Experimental (n = 107) studies; B Risk of Bias for Quasi-Experimental studies (n = 36). Values are expressed in terms of percentage of studies with a certain score. The “Cochrane Risk of Bias Tool for Randomised Trials” tool was used for experimental studies. The 5-Score system of the ROBINS-I was reformulated using the terminology of Cochrane, and reduced to 3 categories, merging Moderate and Serious Risk of Bias into a “Medium Risk” category, in order to simplify its interpretation and assessment. Those items related to “reporting outcomes” were not included, because it was not the objective of this review