Felton, Malika, Hundley, Vanora, Mohan, Vikram, McConnell, Alison and Vargas, Pedro (2026) Acute cardiovascular responses to slow and deep breathing in normotensive men and women.
Slow and deep breathing is recommended as an effective treatment for hypertension using the RESPeRATE device. However, the acute cardiovascular responses to slow and deep breathing, including the potential mechanisms underlying its antihypertensive effect are not fully understood. This study characterised the acute cardiovascular responses to three differing, 10-minute bouts of slow and deep breathing. Twelve participants completed four conditions in a randomised order: 1) RESPeRATE, 2) dynamic slow and deep breathing frequency, 3) fixed breathing frequency of 6 breaths.min-1, 4) spontaneous breathing. Comparing mean values for all variables obscured the cardiovascular perturbations created by slow and deep breathing. However, intra- and inter-breath differences (minimum vs. maximum) in arterial blood pressure were significantly larger during slow and deep breathing compared with spontaneous breathing. The amplitude of systolic blood pressure oscillations increased by up to 10.2% (11.4 mmHg) during inspiration and 8.4% (10.0 mmHg) during expiration (spontaneous breathing; 2.9% (3.4 mmHg) and 3.4% (4.2 mmHg) respectively). Cardiovascular responses were maximised at ~6 breaths.min-1, but further research is needed to identify the optimal breathing frequency to induce maximal cardiovascular perturbations.
| Research / Data Type: | Database |
|---|---|
| Creators: | Creators Email Felton, Malika mfelton@bournemouth.ac.uk Hundley, Vanora vhundley@bournemouth.ac.uk Mohan, Vikram vmohan@bournemouth.ac.uk McConnell, Alison amcconnell@bournemouth.ac.uk Vargas, Pedro vargasp@bournemouth.ac.uk |
| Groups: | Faculty of Media, Science and Technology |
| Collection period: | From To May 2018 June 2018 |
| Temporal Extent: | From To 1 May 2018 30 June 2018 |
| Date: | 8 January 2026 |
| Date type: | Publication |
| Data collection method: | Methods Ethical Approval The experimental protocol was approved by Bournemouth University’s Research Ethics Committee (ID 20679) and all experiments conformed to the Declaration of Helsinki, except for registration in a database. Written informed consent was obtained from all participants prior to participating in the study. Participants Twelve participants took part in the study (6 males & 6 females). All participants were non-smokers with no current diagnosis of cardiovascular or respiratory disease. No participants were pregnant at the time of taking part. Participants refrained from eating for 2 hours and from caffeine, strenuous exercise and alcohol for 12 hours prior to data collection. Slow and Deep Breathing Protocol Participants completed three controlled breathing conditions and one spontaneous breathing condition in a randomised order. Randomisation was conducted using a random number generator (www.randomizer.org). All breathing conditions were 10 minutes in duration with a 10-minute period of normal non-paced breathing prior to each measurement. A 10-minute intervention has been used in previous studies of daily SDB practice using RESPeRATE and was found to be effective at reducing BP (Chaddha et al., 2019). Participants rested at baseline for 5 minutes prior to starting the first breathing condition to ensure cardiovascular variables were in a stable, resting state. During the spontaneous breathing condition (Sfr), participants were instructed to breathe normally and no visual feedback was provided to control breathing. The three SDB conditions were 1) RESPeRATE (Rfr), 2) a dynamic algorithm driven by RSA (Dfr) and 3) a fixed breathing frequency of 6 breaths.min-1 (6Ffr). The RESPeRATE device gradually lowers breathing frequency as users breathe in time with a fluctuating musical tone. Breathing frequency is reduced to ≤10 breaths.min-1 and is measured using a belt worn around either the chest or upper abdomen. A full description of RESPeRATE can be found in Gavish (2010) and Cernes & Zimlichman (2017). Participants completed the dynamic breathing frequency condition (Dfr) using a novel, bespoke biofeedback algorithm that guided breathing dynamically to a personalised frequency. The algorithm created a dynamically-driven target breathing frequency, which strived to maximise cardiovascular perturbation, using the amplitude of RSA as the controlled variable. The algorithm used data measured from a finger sensor (photoplethysmography), which tracked the user’s instantaneous physiological responses to their breathing. The finger sensor was connected via the headphone socket of an iPad, which ran the software algorithm and provided visual biofeedback to guide breathing frequency. The optimal SDB frequency is widely regarded in the extant literature to be ~6 breaths.min-1 (Cullins et al., 2013; Russo et al., 2017); accordingly, the third SDB condition was a fixed frequency of 6 breaths.min-1 (6Ffr). Both the dynamic algorithm and the fixed 6 breaths.min-1 conditions were delivered by an app that provided visual feedback, via an iPad screen, which guided the user’s breathing frequency. The user was instructed to inhale when the dome graphic rose and to exhale when the dome fell. Data Measurement and Acquisition Participants were seated in an upright position, at an approximate angle of 60o for the duration of the data collection. Respiratory airflow was monitored continuously throughout each breathing condition. Participants wore an oronasal mask that covered both mouth and nose (Oro Nasal 7450 V2 Mask, Hans Rudolph Inc., Kansas, USA) and respired flow rate was measured continuously using a heated pneumotachograph (Model 3700, Hans Rudolph Inc., Kansas, USA) connected to a flow measurement system (RSS 100-HR, Hans Rudolph Inc., Kansas, USA). Heart rate (fc) was monitored continuously using a 3-lead ECG and non-invasive beat-to-beat arterial BP was estimated using a Finometer (Finapres NOVA, Finapres Medical Systems, The Netherlands). The finger cuff derived BP was calibrated using an arm cuff prior to and halfway through data collection. Stroke volume (SV) was calculated by the Finometer using the Modelflow method (Wesseling, Jansen, Settels, & Schreuder, 1993). Total peripheral resistance (TPR) was calculated as mean arterial pressure divided by cardiac output (Q̇). Peripheral pulse transit time (PTT) was calculated from the time delay between the peak of the R wave of the ECG and the peak of the pressure pulse recorded at the finger. End-tidal CO2 was recorded at the end of each minute using an iWorx CO2/O2 Gas Analyzer (GA-200, New Hampshire, USA). Analogue outputs from the Finapres NOVA (reconstructed brachial pressure waveform, ECG waveform, SV, SBP, DBP) and the respiratory flow meter were sampled continuously at 250Hz via an analogue to digital converter (NI USB-6218 BNC, National Instruments Inc.) and captured using acquisition and analysis software (LabView 2015, National Instruments, Inc.). The LabView software corrected for the 4 second delay between the Finapres NOVA output and the respiratory output. Data were recorded during the baseline period (5 minutes), and during each breathing condition (10 minutes; Sfr, Rfr, 6Ffr, Dfr). Data Analysis Within the LabView software, cardiovascular and respiratory parameters were derived breath-by-breath, and minimum, maximum and mean values were calculated for every inspiration and expiration. Data were averaged in segments of one-minute, as well as mean values for the first 5-minutes, final 5-minutes, and the full 10-minutes for each condition. Data were compared for the three SDB conditions (Rfr, 6Ffr, Dfr) and spontaneous breathing (Sfr). Respiratory sinus arrhythmia (RSA) was calculated using two methods across each breath phase 1) the difference between the average heart rate (fc) during inspiration (fci) and expiration (fce) for every breath (fcΔ); 2) the difference in maximum and minimum beat-to-beat intervals (RR) during inspiration and expiration respectively for every breath (RSA). Because the kinetics of the haemodynamic perturbations created by breathing may lag the breathing phase in which they were generated, the ‘peak-valley’ method was used to analyse all cardiovascular variables; in other words the difference between successive peaks (highest value) and valleys (lowest value) was calculated, independent of the breathing phase in which they occurred. Calculated parameters are as follows; Inter-breath phase indices (Δ) were quantified as the difference between mean inspiration (i) and mean expiration (e) values for all variables. Peak-valley (PV) indices were calculated as maximum minus minimum values during inspiration (Δi) and during expiration (Δe). To represent the largest change in variables within breath phase, inter-breath phase PV indices (ΔPV) were calculated using maximum inspiration minus minimum expiration, or minimum inspiration minus maximum expiration, dependent on which calculation gave the largest difference. The PV indices were also calculated irrespective of the breath phase in which they occurred, and are referred to herein as peak-valley breath phase independent calculations (ΔPV_Ind). Each condition was 10 minutes in duration, but the final 5-minute segment of each SDB condition (Rfr, 6Ffr, Dfr) were used for analysis to ensure relatively steady state values were analysed. For spontaneous breathing (Sfr), the first 5-minute segment was used, as participants were already in a steady state having rested prior to the condition for a minimum of 10 minutes while breathing normally/spontaneously. Dynamic breathing frequencies were also compared across the full 10-minute condition and between the first and final 5 minutes. Values are expressed as mean ± SD unless stated otherwise. Statistical analysis was undertaken using SPSS Statistics 24 (IBM Corp.). After normality was confirmed for cardiovascular variables, repeated measures ANOVA with planned post-hoc pairwise comparisons using Bonferroni corrections were used to compare between the different breathing frequencies. Reported p values are those following adjustment for repeated comparisons. For all analyses, P was set at 0.05. Correlation coefficients were calculated using Pearson Product Moment. |
| Statement on legal, ethical and access issues: | All data is pseudonymised and therefore not identifiable |
| Funder name(s): | Bournemouth University |
| Grant reference number(s): | N/A - Pump priming funding |
| Keywords: | Slow and deep breathing; RESPeRATE; respiratory sinus arrhythmia; blood pressure |
| Status: | Published |
| Publisher: | Bournemouth University |
| Copyright holders: | Malika Felton, Bournemouth University |
| Contact email address: | bordar@bournemouth.ac.uk |
| DOI: | 10.18746/bmth.data.00000513 |
|---|---|
| Depositing User: | Malika Felton |
| Year Deposited: | 08 Jan 2026 10:27 |
| Revision: | 19 |
| Last Modified: | 08 Jan 2026 10:28 |
Available Files
Data
All conditions_LabView_RESPeRATE.xlsx
| ||||||||||||||||||
Individual participa ... thing conditions.zip
|
Related Resources
Acute cardiovas ... men and women.
|
Tools
Tools