We included healthy, normal- and slightly overweight subjects of both sexes over 16 years of age. All participants were non-smokers and free of cardiorespiratory and other significant acute or chronic illness and had normal blood pressure. Participants visited the respiratory physiology laboratory on two occasions at least 1 week apart, with one visit acting as control without electrical stimulation and the other during which TES was used during in all postures and gas conditions.

Subjects for the study met all the following criteria: age >16 years, body-mass index (BMI) > 18.5 and <30 kg/m², non-smoker, and clinically stable in the last 28 days.

Subjects were excluded from the study if any of the following conditions were met: history of cardiovascular, respiratory, or neuromuscular disease, cardiac pacemaker, active seizures, current smokers, acute illness, allergy to skin patches, oobesity (BMI>30 kg/m²) or cachexia (BMI<18.5 kg/m²), and vertigo.

Following the baseline visit without electrical stimulation the participants were continuously stimulated using electrical current in the submental area (4 × 4 cm dermal patches; Med-Fit Plus Ltd., Stockport, United Kingdom), at a frequency of 30 Hz and a pulse width of 250 microseconds during the second visit. Intensity of the electrical current was titrated according to individual comfort using a TENS machine (Premier Combo Plus, the TENS + Company Lets, Stockport, United Kingdom, placed in the submental area midway between angle of mandible and the chin (Figure 1) as previously described elsewhere (Steier et al., 2011). Continuous, beat-by-beat arterial blood pressure was measured continuously using digital artery photoplethysmography (Finapres, Ohmeda 2,300, BOC Healthcare, Englewood CO, United States of America). Heart rate was measured from the electrocardiogram (ECG) with electrodes positioned in the lead II configuration (ML132 bioamplifier, ADInstruments, Oxford, United Kingdom).

Respiratory flow was measured via a mouthpiece, with the subject wearing a noseclip using a pneumotachograph (4,800 series, Hans Rudolph Inc., Shawnee Kansas, United States of America) and associated differential pressure transducer (Spirometer, ADInstruments, Oxford, United Kingdom). The distal end of the pneumotachograph was attached to a two way non-rebreathing valve (2,700 series, Hans Rudolph Inc., Shawnee, Kansas, United States of America, deadspace 77 ml) with inspired and expired gases measured continuously using a gas analyser (ML, 206, ADInstruments, Oxford, United Kingdom), connected to a side port on the pneumotachograph via a fine-bore catheter. Blood oxygen saturation (SpO2) was measured using a pulse oximeter (Sat 805 pulse oximeter, Charter Kontron, United Kingdom) attached to the subject’s finger. All data were acquired (PowerLab 16, ADInstruments, Oxford, United Kingdom) with 1 Khz sampling and displayed (LabChart ver 8, ADInstruments, Oxford, United Kingdom). Tidal volume was obtained by digital integration of flow by the acquisition software.

An open circuit (Figure 2) was used to deliver a continuous supply of medical air (wall outlet) to the inspiratory port of the two-way non-rebreathing valve via a low volume (2.5 L) reservoir bag. The inspired gas could be enriched with 100% nitrogen or 100% carbon dioxide from cylinders (BOC, Guildford, United Kingdom) to provide the appropriate inspired gas concentration. Three inspired gas mixtures were used; medical air (21% O², balance N₂), poikilokapnic hypoxia (12% O2, balance N₂) and normoxic hypercapnia (5% CO₂, balance N₂). Symptoms of breathlessness were scored using the modified Borg scale in each posture (seated, supine, and 50°HDT).

An electrically operated tilt table (Plinth2000 Ltd., Stowmarket, United Kingdom) which could be adjusted from 0° (flat) to 50°HDT was used to change posture.

Based on the sample size of 13 subjects, the study detected a treatment difference at a two-sided significance level of 0.025 if the true mean difference in diastolic blood pressure (electrical stimulation on vs. off) was at least 8.049 mmHg (SD 10.6) with 80% power. The variable calculated was the minimal detectable difference in mean diastolic blood pressure, based on previous data (Strollo et al., 2014).

Following testing for normality, data were presented as mean (SD) unless otherwise indicated. Data were analysed using a three-way analysis of variance (ANOVA)) followed by a Tukey’s test using the ‘anovan’ and ‘multcompare’ function of MATLAB (Version 2022B, MathWorks Ltd, Natick/MA, United States of America) to evaluate overall effects of three factors: a) TES (on/off), b) posture (seated, supine, HDT), and c) inspired gas (RA, hypercapnia, and hypoxia); furthermore, baroreflex and chemoreflex interaction was tested with the combination of these three factors. A level of significance was defined as p < 0.05.

A marked reduction in systolic blood pressure during electrical stimulation was observed under hypoxic conditions in the HDT posture; there was also a trend towards reduction in other postures. (Table 1).

A marked reduction in both diastolic and mean arterial blood pressures were also observed during electrical stimulation in supine and HDT postures. There was also a tendency towards a reduction in diastolic blood pressure during electrical stimulation when seated under hypoxic conditions, and during HDT both under hypercapnic and room air conditions. (Tables 2, 3).

There was no significant change to the pulse pressure when applying electrical stimulation, independent of different postures and gas mixtures. (Table 4).

There was a trend towards increased tidal volume in the supine posture under hypercapnic conditions with electrical stimulation, although this did not reach significance. (Table 7) Additionally, there was increased minute ventilation with electrical stimulation (p=0.0443; Table 8;Figure 4; details is provided in Supplementary Table S3).

No changes in EtCO₂ during electrical stimulation in seated, supine or HDT posture were observed when breathing room air, under hypoxic or hypercapnic conditions. The SpO₂ did not change when comparing electrical stimulation to baseline in any posture or gas condition studied.

There was no significant change in the breathlessness scores when applying electrical stimulation, independent of posture and gas mixture. (Supplementary Table S2).

In a 3-way ANOVA, BP decreased significantly with TES (systolic: p = 4.3E-06, diastolic: p = 3.01E-09, mean: p = 3.26E-08). Change in gas condition (systolic: p = 0.0386, diastolic: p = 0.0032, mean: 3.2E-03) and different postures (systolic: 6.91E-08, diastolic: p = 6.55E-04, mean: p = 4.91E-05) similarly impacted on BP control. There were no significant interactions between the three different factors: electrical stimulation, gas condition, or posture, except for an effect on minute ventilation (gas condition/posture p = 0.0348; Figure 4 for supplemental information on the analysis please refer to Supplementary Table S3).

A number of pathways are involved in the cardiovascular responses to systemic hypoxia (Marshall, 1994), involving the primary effects of peripheral chemoreceptor stimulation, secondary effects of ventilation, and direct effects of hypoxia on the heart and peripheral vasculature leading to subsequent effects on the autonomic and the central nervous system (Marshall, 1994). The full effects of ventilation, mediated by carbon dioxide and oxygen, on the cardiovascular system remain to be fully elucidated (Heistad et al., 1974). Importantly, there is cardiorespiratory interaction which is mediated via hypoxia and that affects the baroreflex response, as suggested by our observations. It has been described previously that stimulation of the chemoreceptors can lead to an increased heart rate and a change in the blood pressure (cardiovagal baroreflex) in humans (Bristow et al., 1971); (Bristow et al., 1974). This is further supported by recent evidence showing that exposure to hypoxia can alter the arterial baroreflex and change heart rate and sympathetic nerve activity with a higher blood pressure. (Heistad et al., 1974); (Halliwill and Minson, 2002); (Halliwill et al., 2003).

Electrical stimulation targets the upper airway dilator muscles, particularly the genioglossus muscle (GG), and counteracts their diminished neuromuscular state-dependent tone which promotes upper airway collapsibility (Mezzanotte et al., 1996). Both the invasive and transcutaneous approaches to stimulating the upper airway dilator muscles are beneficial for maintaining airway patency during sleep in patients with OSA (Strollo et al., 2014; Pengo et al., 2016) improving the AHI by a mean of 9.1 (95% confidence interval, CI 2.0, 16.2) events/hour and the 4% ODI improved by a mean of 10.0 (95% CI 3.9, 16.0) events/hour (Pengo et al., 2016). Furthermore, the initial feasibility studies used TESLA with a current of 10.1 (3.7) mA (Steier et al., 2011). In the current study, electrical stimulation was well tolerated and had no adverse effects, underlining its safety for the use in the submental area and its efficacy in lowering diastolic blood pressure. In the context of potential long-term treatments for patients with OSA who have a high prevalence of cardiovascular comorbidities it is important to highlight that heart rate did not change significantly. A reduction in the blood pressure, systolic and diastolic, remains a favourable outcome for patients with sleep-disordered breathing, as the cardiovascular risk is typically raised and treatment resistant hypertension is of clinical relevance in this cohort (Antic et al., 2015). There are various interactions between different types of sleep apnoea and cardiovascular variables (e.g., blood pressure). On the one hand, central sleep apnoea is driven by heart failure (Javaheri and Javaheri, 2022). On the other hand, obstructive sleep apnoea leads to an increased sympathetic tone with may impact on the blood pressure contributing to hypertension (Antic et al., 2015); (Pengo et al., 2021); (Pengo et al., 2020).

This prospective physiological study had a relatively small sample size and certain interactions could become more significant with a larger sample size, for example the effect of the three factors on minute ventilation. The effects of TES on the primary outcome variable, diastolic blood pressure, were highly consistently observed in all subjects with a large effect size. Longer steady state periods could have had further advantages over quasi-steady state achieved during the 5 min baseline periods used. The choice of this was pragmatic to allow for completion of what was a lengthy protocol and return of the healthy volunteers for a demanding second session. We were also limited with making causal inferences due to the observational nature of the design of the study. Despite a complex experimental setup, some parameters such as neural respiratory drive, blood gases, and perfusion could have provided helpful additional insights into the interaction between the cardiovascular, the respiratory, the peripheral autonomic and the central nervous system but were not measured on this occasion. In addition, this study focused on healthy subjects with normal blood pressure. Thus, further studies in subjects with hypertension and sleep-disordered breathing need to provide a comprehensive dataset on how electrical stimulation affects the chemo- and baroreceptor response in these clinically relevant cohorts. However, these points do not negate the insightful setup of a highly complex physiological experiment in human beings with a large effect size that enables to derive useful information for future clinical applications.