Telemedicine-delivered myofunctional therapy remodels upper airway anatomy in obstructive sleep apnea: a prospective controlled study

Abstract

Study Objectives

To evaluate whether telemedicine-delivered myofunctional therapy (MT) produces measurable structural changes in the upper airway (UA) of patients with obstructive sleep apnea (OSA).

Methods

This prospective, nonrandomized, blinded study included 60 adults with moderate-to-severe OSA, assigned to three groups: moderate OSA + MT (n = 19), severe OSA + MT + CPAP (n = 20), and OSA + CPAP without MT (control, n = 18). MT consisted of daily oropharyngeal exercises for 3 months, monitored through telemedicine. Pre- and postintervention assessments included sleep studies, submental ultrasound (interarterial distance, tongue thickness, and tongue volume), Iowa Oral Performance Instrument (IOPI) scores, and drug-induced sleep endoscopy (DISE; VOTE classification). Ultrasound measurements were performed by a single blinded examiner.

Results

After 3 months, the moderate OSA + MT group showed significant reductions in tongue volume (−8 cm³; P = .002), tongue thickness (−4 mm; P <.001), and interarterial distance (−5 mm; P < .001). In the severe OSA + MT + CPAP group, tongue volume (−12 cm³; P < .001) and interarterial distance (−7 mm; P < .001) decreased, while tongue thickness remained unchanged. No significant changes occurred in controls. Absence of tongue collapse during DISE increased from 15% to 80% in moderate OSA + MT (P = .039) and from 15% to 55% in severe OSA + MT + CPAP (P = .109).

Conclusions

Telemedicine-based MT promotes measurable UA remodeling. Submental ultrasound provides a practical, noninvasive monitoring tool, supporting MT as a scalable, patient-centered strategy to enhance structural and functional outcomes in OSA management.

Brief Summary

This prospective study explored whether myofunctional therapy (MT), delivered through a telemedicine platform, could lead to measurable changes in the upper airway of patients with obstructive sleep apnea (OSA). Sixty adults were followed for three months with submental ultrasound and drug-induced sleep endoscopy. Patients with moderate OSA who performed MT showed clear reductions in tongue size and airway width, while no significant changes occurred in CPAP-only controls. These results suggest that telemedicine-based MT can expand upper airway structures, particularly in moderate OSA, and may help guide individualized treatment strategies.

Current Knowledge/Study Rationale:

Myofunctional therapy has been reported to improve symptoms of OSA, but its structural effects on the upper airway and the role of telemedicine-based delivery remain poorly documented.

Study Impact:

This study provides evidence that MT can induce measurable airway remodeling, especially in moderate OSA. Submental ultrasound proved useful for monitoring these changes, supporting the use of telemedicine MT as a scalable, patient-centered approach to individualized OSA management.

Supplementary Information

The online version contains supplementary material available at 10.1007/s44470-025-00008-0.

Introduction

Obstructive sleep apnea (OSA) is a highly prevalent disorder associated with significant cardiovascular, metabolic, and neurocognitive morbidity [1]. Anatomical factors, particularly those involving the tongue and oropharyngeal soft tissues, play a central role in upper airway (UA) collapsibility [2]. Recent imaging studies have shown that increased tongue volume, often due to fatty infiltration, is independently associated with OSA severity and may represent a modifiable therapeutic target [3, 4].

Myofunctional therapy (MT), a structured program of oropharyngeal exercises, may improve muscle strength [5, 6] and reduce OSA severity in some patients [7]. However, whether MT is accompanied by measurable structural remodeling of the UA remains controversial [8–10]. Some studies have suggested that the UA dilator muscles, particularly those targeted by MT, may lose neuromotor inputs during sleep, the period when airway collapse occurs, thereby raising questions about the clinical relevance of daytime oropharyngeal muscle training [11, 12]. In addition, although substantial changes are detectable in the upper airway of patients with OSA [13–15], expert consensus reports highlight the current lack of objective evidence linking MT to sustained anatomical modifications [11].

Current treatments, such as continuous positive airway pressure (CPAP), effectively splint the airway open during sleep, but it remains unclear whether CPAP therapy induces long-term anatomical change [16]. Submental ultrasound has emerged as a practical, noninvasive tool for quantifying tongue volume, thickness, and the interarterial distance between the lingual arteries [17–20]. Compared with magnetic resonance imaging [21], ultrasound is more portable, repeatable, and suitable for serial measurements, which makes it an attractive option for tracking potential anatomical changes during MT [21].

Drug-induced sleep endoscopy (DISE) provides a complementary dynamic assessment of airway collapse patterns and allows investigation of the relationships between structural measurements and functional improvements [22–25]. To our knowledge, no prospective study has combined submental ultrasound and DISE to evaluate the anatomical and functional changes in patients undergoing MT, particularly when clinical management is delivered via a telemedicine-based platform.

To address some of these issues, the current study examined whether telemedicine-delivered MT could promote measurable structural changes in the tongue and UA, as assessed by submental ultrasound and DISE, in patients with moderate-to-severe OSA while comparing these changes with a control group of patients receiving CPAP without MT.

Materials and methods

Study design and registration

We conducted a prospective, nonrandomized clinical study to evaluate the effects of telemedicine-delivered MT on UA anatomy in patients with OSA. The study protocol was registered prospectively at ISRCTN (ISRCTN92645461) and has been previously published [21]. The ISRCTN registry initially described the primary outcome as “tongue fat” to facilitate public understanding and dissemination, as requested by the registry administrators. However, direct quantification of intramuscular fat by submental ultrasound is technically limited. Therefore, the present study used tongue volume [4], a validated and reproducible ultrasound-derived surrogate that reflects structural remodeling of the tongue and indirectly captures potential changes in fat infiltration. The conceptual aim and hypothesis of the registered protocol, evaluating whether myofunctional therapy induces anatomical changes in the tongue related to upper airway patency, remained unchanged.

Ethical approval was obtained from the Provincial Research Ethics Committee of Málaga, Spain (AWGAP-2023-2). All participants provided written informed consent before enrollment.

Participants

Sixty adults aged 18–75 years with a diagnosis of moderate-to-severe OSA, confirmed by overnight polygraphy according to American Academy of Sleep Medicine criteria, were recruited consecutively between April 2023 and May 2025 from the Sleep Disorders Clinics at Hospital Quirónsalud Marbella (Marbella, Malaga) and Hospital Quirónsalud Campo de Gibraltar. (Palmones, Cadiz) Spain.

Inclusion criteria

• Age 18–75 years.

• Diagnosis of moderate or severe OSA (apnea–hypopnea index (AHI) ≥15 events/h).

• For CPAP users: stable use ≥4 h/night on ≥70% of nights.

• Body mass index (BMI) < 30 kg/m².

• Ability to follow the protocol and provide informed consent.

• All participants in the MT intervention groups exhibited habitual oral breathing, which was considered a clinical indication for therapy. This characteristic was absent in the CPAP-only control group.

Exclusion criteria

• Previous oropharyngeal surgery affecting the tongue or airway anatomy.

• Neurological or muscular disorders affecting swallowing or speech.

• Severe cardiovascular instability.

• Active neoplastic disease.

• Previous or ongoing oropharyngeal muscle rehabilitation or oral appliance therapy.

• Temporomandibular joint (TMJ) dysfunction or severe ankyloglossia.

• Inability to use a smartphone or access telemedicine tools.

• Weight variation > 5 kg during the study period.

• All anatomical evaluations were conducted by an otolaryngologist who was blinded to the subject group assignment to minimize assessment bias and to allow the isolation of MT-related effects.

Interventions

Study groups

Participants were allocated into three groups as follows.

• Moderate OSA + MT group: moderate OSA (AHI 15–30 events/h) treated exclusively with MT for 3 months.

• Severe OSA + MT + CPAP group: severe OSA (AHI > 30 events/h) receiving stable CPAP plus MT.

• Control group (CPAP only): moderate-to-severe OSA with stable and adherent use of CPAP ≥3 months before enrollment but no MT.

Sample size justification

The sample size was calculated to detect a clinically relevant difference in UA anatomy (primary outcome: lingual thickness and area measured by submental ultrasound) between the MT and control groups using an analysis of covariance model in which the posttreatment values were adjusted for the baseline with a two-sided α of 0.05 and 80% power. Based on pilot data (effect size ≈ 0.5, correlation between baseline and postintervention values r ≈.60) 18 participants per group were required. Allowing for a 10–15% dropout rate, the final target was 60 participants (i.e., 20 subjects per group).

Myofunctional therapy

MT comprised a standardized set of isotonic and isometric oropharyngeal exercises targeting the tongue, soft palate, and pharyngeal walls that aimed to strengthen upper airway (UA) muscles and improve nasal breathing. Patients performed MT by interacting with a plug-in connected to a smartphone screen [23]. The AirwayGym^® app provides a series of daily exercises—including both isometric (tension) and isotonic (contraction) movements—designed to be performed conveniently in any setting. It offers the additional advantage of providing immediate feedback, allowing patients to monitor their progress and resolve difficulties with the assistance of a therapist.

Before each exercise, an animated demonstration and video provide step-by-step instructions. After each exercise, patients receive visual, acoustic, and tactile feedback as a score indicating performance quality. The tactile feedback is generated through a patented plug-in device connected to the smartphone: correct performance is recognized only when the oropharyngeal muscle exerts stable and adequate pressure on the screen sensor, triggering a vibrotactile signal confirming the exercise. A chat function is also available, allowing direct contact with the supervising therapist. This app complies with EU data protection regulations (2002/58/EC and 2016/679).

Exercises were performed for 20 min per day, 5–7 days per week, for 3 months, following instructional videos and supervised via the AirwayGym^®telemedicine platform, which has been previously validated for OSA treatment [23, 26–33]. Weekly online reviews ensured correct execution and adherence (Fig. 1, panel a).

Fig. 1.

Multimodal assessment of telemedicine-delivered myofunctional therapy in patients with obstructive sleep apnea (OSA). (A) Tongue proprioceptive training exercise performed with the Airway Gym^® smartphone plug-in. (B) Submental ultrasound assessment of tongue volume, showing probe placement and representative image. (C) Coronal and sagittal ultrasound views with tongue boundaries outlined for volume calculation. (D) Submental ultrasound measurement of the interarterial distance, with calipers placed between the lingual arteries (arrows)

Although all participants completed the same nine standardized exercises, the emphasis on specific tasks was individualized according to the findings from the IOPI and DISE. Patients presenting with low tongue pressure and lingual collapse on DISE were instructed to focus on genioglossus-strengthening exercises, whereas those with velopharyngeal collapse or reduced buccinator pressure were guided to reinforce perioral and soft-palate exercises. The AirwayGym^® app measures the accuracy of each exercise and provides targeted feedback. Exercises can be adjusted to address the specific levels and patterns of upper airway collapse identified during drug-induced sleep endoscopy (DISE). This approach was informed by our previous studies demonstrating correlations between IOPI measurements and DISE collapse patterns [34].

Assessments

Evaluations were conducted at baseline and at 3 months by the same team of blinded examiners.

Otorhinolaryngological examination

A complete otorhinolaryngological assessment was performed at baseline and at 3 months follow-up, and included anterior rhinoscopy, oropharyngeal inspection, and flexible nasopharyngolaryngoscopy. Tonsil size and tongue position were graded according to the Friedman staging [35], and nasal patency was assessed for septal deviation and turbinate hypertrophy. The presence of oral breathing, pathological lingual frenulum, TMJ dysfunction, or other structural abnormalities that could influence UA collapsibility was recorded. Neck circumference and craniofacial features were also noted. All examinations were conducted by board-certified otolaryngologists blinded to the treatment allocation.

Submental ultrasound measures

The submental ultrasound technique is a noninvasive imaging method used to assess anatomical features of the tongue and adjacent vascular structures. It is particularly useful for evaluating the tongue volume (expressed as cm³; Fig. 1 panel b), tongue thickness (mm; Fig. 1 panel c), and the position of and distance between the lingual arteries (Fig. 1 panel d. The patient was placed in a supine position with the head extended slightly backward to provide optimal access to the submental region. The mouth remained closed, and the tongue was in a natural resting position. Submental ultrasound examinations were performed using the SonoSite 180 Plus system (SonoSite Inc., Bothell, WA, USA) with a linear L38/10–5 MHz transducer. The same machine was used for all measurements throughout the study to ensure consistency. A high-frequency linear array transducer (typically 7–15 MHz) was used in the coronal and sagittal submental approaches [21]. Gel was applied to the submental area to ensure good acoustic coupling. The transducer was positioned transversely under the chin (in the submental region), perpendicular to the sagittal plane, and then rotated or angled as necessary to obtain optimal images of the tongue.

To ensure reliability, all ultrasound examinations were conducted by the same experienced otolaryngologist, who was blinded to group allocation. In a random subset of 10 participants, repeated measurements demonstrated high intra-rater reproducibility, with intraclass correlation coefficients (ICC) > 0.85 for tongue volume, tongue thickness, and interarterial distance. These findings are consistent with prior validation studies confirming the repeatability of submental ultrasound in patients with OSA [17, 19].

Visualization of lingual arteries

Using color Doppler mode, the bilateral lingual arteries were identified in their course through the tongue. The inter-arterial distance (distance between the left and right lingual arteries) was measured at the level of their maximal visibility, typically near the midportion of the tongue.

Tongue thickness measurement

In this study, tongue thickness was measured consistently at the midline, as this approach has been validated as the most reproducible and standardized reference point for submental ultrasound assessment in OSA. Using a single measurement site for all participants minimized variability and ensured comparability across groups.

Tongue volume Estimation

Multiple cross-sectional images were acquired along the anterior–posterior axis of the tongue. The height, width, and thickness of the tongue were taken into account in different measurements to estimate the volume of the tongue using the ultrasound software to estimate the area. Once all the measurements have been saved, the result will appear on the screen: Tongue volume was calculated using the built-in measurement software of the SonoSite 180 Plus ultrasound system (SonoSite Inc., Bothell, WA, USA). The device automatically estimates volume based on three orthogonal two-dimensional measurements—height, length, and thickness—using a geometric ellipsoid model integrated into the manufacturer’s software. All participants were examined using the same ultrasound system and measurement protocol to ensure methodological consistency and reproducibility. As a result, all tongue volume data were derived using a single standardized technique, without inter-method variability.

Tongue strength measurement

The Iowa Oral Performance Instrument (IOPI) was used to measure maximal tongue pressure through a combined elevation–protrusion movement against the anterior palatal rugae, reflecting global tongue strength [36–38]. In individuals with a flat hard palate, the maxillary papilla lies at the same height as the tongue tip, naturally inducing a slight anterior component during elevation. This dual movement engages both genioglossus and intrinsic muscles, offering a comprehensive assessment of tongue strength [37, 38]. The IOPI is a validated device that measures tongue and lip strength by recording the maximum pressure exerted against an air-filled bulb. It provides objective, reproducible data about orofacial muscle performance that are helpful for monitoring functional changes during therapy [36, 37].

Sleep study

Overnigt polygraphy was performed for each participant using a type III sleep study device (SOMNOtouch™, SOMNOmedics GmbH, Randersacker, Germany). During polygraphy, the following parameters were evaluated for this study: mean oxygen saturation and lowest oxygen saturation, time of sleep spent with blood oxygen saturation < 90%, and the AHI. The AHI was defined as the total number of apnea and hypopnea events per hour of sleep recorded in an overnight sleep study. Apnea was defined as ≥90% decrease in airflow for at least 10 s and hypopnea as a reduction in respiratory signals for ≥10 s associated with a minimum of 3% oxygen desaturation [39].

For participants in the CPAP groups, follow-up polygraphy was performed after a washout period of at least 48 h without CPAP use to minimize residual effects of prior therapy, as AHI values have been reported to fluctuate immediately after CPAP withdrawal before stabilizing. All patients had stable CPAP adherence (≥ 4 h/night on ≥ 70% of nights) for at least 3 months prior to enrollment, with an average therapeutic pressure of 10.8 ± 2.1 cm H₂O in the severe OSA + MT + CPAP group and 10.4 ± 2.3 cm H₂O in controls.

Drug-induced sleep endoscopy

DISE was used with a standardized protocol involving target-controlled propofol infusion and video recording. The velum, oropharynx, tongue base, and epiglottis (VOTE) classification was performed by two independent blinded otolaryngologists, and their interrater agreement was assessed by calculating Cohen’s kappa [40, 41].

Adherence monitoring

Adherence was automatically recorded by the AirwayGym^®platform, which tracked both the number and duration of daily exercise sessions. Satisfactory adherence was defined as completing ≥ 80% of the prescribed sessions (20 min/day, 5–7 days/week for 3 months). Participants who failed to meet this criterion were considered nonadherent and excluded from analysis. The mean adherence rates were 92% in the moderate OSA + MT group, 97% in the severe OSA + MT + CPAP group, and 86% CPAP compliance in the control group.

Endpoints

The primary endpoint was the change in tongue volume (cm³) from the baseline to 3 months. Secondary endpoints included the changes in tongue thickness, lingual artery distance, and DISE tongue base collapse classification.

Statistical analysis

Statistical analyses were performed by the Department of Statistics at the University of Málaga, Spain, using IBM SPSS Statistics (v. 31; IBM Corp., Armonk, NY, USA). Continuous variables were expressed as mean ± standard deviation or median (interquartile range), and the distribution was assessed using the Kolmogorov–Smirnov test. Between-group differences were analyzed using analysis of variance or the Kruskal–Wallis test. Within-group changes were analyzed using the paired t-test or the Wilcoxon signed-rank test. Categorical variables were compared using chi-square or Fisher’s exact test. Significance was set at P <.05.

Results

Participants and follow-up

Sixty adults with moderate-to-severe OSA were enrolled and allocated to three groups: moderate OSA treated with MT only (n = 20), severe OSA treated with MT plus CPAP (n = 20), and a control group on stable CPAP not using MT (n = 20). Three participants did not complete the 3-month follow-up: one in the moderate OSA + MT group withdrew due to weight change > 5 kg during the study; BMI values remained stable across all groups, and no significant pre- vs. post-treatment differences were observed. One control participant relocated; and another control participant developed temporomandibular joint dysfunction. No losses occurred in the severe OSA + MT + CPAP group. Thus, the final cohort completing the protocol comprised 19 participants in the moderate OSA + MT group, 20 subjects in the severe OSA + MT + CPAP group, and 18 controls (Fig. 2).

Fig. 2.

CONSORT flow diagram of the study. Sixty adults with moderate-to-severe OSA were allocated to three groups: control (CPAP only, n = 20), moderate OSA treated with myofunctional therapy (MT) alone (n = 20), and severe OSA treated with MT plus continuous positive airway pressure (CPAP; n = 20). Baseline blinded assessments included submental ultrasound, drug-induced sleep endoscopy (DISE), and clinical examination. After 3 months of telemedicine-delivered MT, follow-up assessments were repeated. Losses to follow-up were as follows: n = 2 in the control group, n = 1 in the moderate OSA + MT group, and n = 0 in the severe OSA + MT + CPAP group. Final sample sizes were n = 18 in the control group, n = 19 in the moderate OSA + MT group, and n = 20 in the severe OSA + MT + CPAP group

Baseline characteristics

Baseline characteristics are presented in Table 1. Participants were predominantly middle-aged men. The sex distribution did not differ significantly between groups (P =.235). As expected, the Friedman tonsil and tongue position grades showed significant between-group differences (both P <.001); the severe OSA + MT + CPAP group had a higher percentage of patients with a grade 3 classification.

Table 1.

Baseline data

Variable	Moderate OSA + MT (n = 19)	Severe OSA + MT + CPAP (n = 20)	Control (CPAP only) (n = 18)	P value
Demographics
Age, years	51.35±14.07	51.05±10.78	40.94±10.86	0.15
Male sex, n (%)	13 (68)	18 (90)	15 (83)	0.235
Female sex, n (%)	6 (31)	2 (10)	3 (16)	0.17
Anthropometrics
BMI, kg/m²	27.66 ± 1.96	29.80 ± 3.50	23.09 ± 3.46	0.78
Neck circumference, cm	39.71 ± 2.10	40.06 ± 2.23	36.77 ± 3.19	0.82
Friedman staging
Tonsil grade 1, n (%)	0 (0)	0 (0)	3 (15)	< 0.001
Tonsil grade 2, n (%)	11 (55)	0 (0)	6 (30)	< 0.001
Tonsil grade 3, n (%)	9 (45)	20 (100)	11 (55)	< 0.001
Tongue position grade 1, n (%)	0 (0)	0 (0)	2 (10)	< 0.001
Tongue position grade 2, n (%)	8 (40)	0 (0)	6 (30)	< 0.001
Tongue position grade 3, n (%)	12 (60)	20 (100)	12 (60)	< 0.001
Functional measures
IOPI lips, kPa	20.4 ± 4.50	16.05 ± 4.98	23.6 ± 4.38	< 0.001
IOPI tongue, kPa	44.35 ± 9.0	35.5±10.68	54.75±7.93	< 0.001
Mean therapeutic CPAP level, cm H₂O	-	10.8 ± 2.1	10.4 ± 2.3	0.44
Duration of CPAP use, year	-	3.4 ± 1.2	3.1 ± 1.4	0.53
Submental ultrasound baseline parameters
Tongue volume (cm3)	85 ± 5.3	93 ± 5.5	83 ± 6.5	< 0.001
Tongue thickness (mm)	41 ± 3.5	47 ± 3.6	40.3 ± 5.7	< 0.001
Inter-arterial distance (mm)	31 ± 4.0	36 ± 4.1	32.3 ± 3.8	0.006

BMI body mass index; CPAP continuous positive airway pressure; IOPI tongue maximal tongue pressure (genioglossus strength and intrinsic muscles); IOPI lips maximal lip pressure; MT myofunctional therapy; OSA obstructive sleep apnea. “–” indicates not applicable (patients not treated with CPAP). No significant changes in BMI were observed at 3 months. One participant was excluded due to a weight variation >5 kg P values obtained using independent-sample t test for continuous variables and chi-square test for categorical variables. Baseline submental ultrasound parameters (tongue volume, thickness, and inter-arterial distance) were compared across groups using one-way ANOVA; non-parametric tests were applied when distributional assumptions were not met

Baseline BMI (P =.78) and neck circumference (P =.82) did not differ significantly between groups, but baseline IOPI scores differed significantly between groups (P <.001): lip pressure was highest in the controls (23.6 ± 4.38 kPa), intermediate in the moderate OSA + MT group (20.4 ± 4.5 kPa), and lowest in the severe OSA + MT + CPAP group (16.05 ± 4.98 kPa. Tongue pressure followed a similar pattern (P <.001) and was lowest in the severe OSA + MT + CPAP group (35.5 ± 10.68 kPa) compared with the moderate OSA + MT (44.35 ± 9.5 kPa) and control (54.75 ± 7.93 kPa) groups.

As per study design, AHI was significantly higher in the severe OSA + MT + CPAP group (38.85 ± 5.98 events/h) than in the other two groups — moderate OSA + MT: 22.8 ± 3.61 events/h; CPAP-only controls: 22.6 ± 3.61 events/h — (both P <.001 vs. severe).

Changes in submental ultrasound parameters (Fig. 3)

Fig. 3.

Changes in submental ultrasound parameters after 3 months of intervention. Connected lines represent individual patient trajectories from baseline to follow-up. Black dots with error bars indicate group means ± standard error. (A) Tongue volume, (B) tongue thickness, and (C) interarterial distance. Red asterisks denote statistically significant within-group differences compared to baseline (*P <.05). Significant reductions were observed in the moderate OSA + MT group for all three parameters, and in the severe OSA + MT + CPAP group for tongue volume and interarterial distance. No significant changes occurred in the CPAP-only control group

After 3 months, the moderate OSA + MT group showed significant reductions in tongue volume, tongue thickness, and interarterial distance compared with baseline, while the severe OSA + MT + CPAP group showed reductions in tongue volume and interarterial distance but no significant change in tongue thickness. No significant changes emerged in the control group (Table 2). Between-group analyses confirmed greater improvements in both MT intervention groups than in controls for tongue volume and interarterial distance (all P <.01).

Table 2.

Changes in submental ultrasound parameters after 3 months

Parameter	Group	Pre (mean ± SD)	Post (mean ± SD)	Δ (absolute)	% Change	95% CI for Δ	P value (
Tongue volume, cm3	Moderate OSA + MT (n = 19)	85 ± 5.31	77 ± 8.23	–8	–9%	−11.18 to − 5	0.002
	Severe OSA + MT + CPAP (n = 20)	93 ± 5.45	81 ± 8.45	–12	–13%	−15.10 to − 8.90	< 0.001
	Control (CPAP only) (n = 18)	82.9±6.46	82.43±6.20	NS	—	—	NS
Tongue thickness, mm	Moderate OSA + MT (n = 19)	41 ± 3.48	37 ± 3.18	–4	–10%	−5.27 to − 2.73	< 0.001
	Severe OSA + MT + CPAP (n = 20)	47 ± 3.55	46 ± 3.13	–1	–2%	−2.24 to 0.24	0.08
	Control (CPAP only) (n = 18)	40.27±5.65	40.28±4.93	NS	—	—	NS
Interarterial distance, mm	Moderate OSA + MT (n = 19)	31 ± 4.05	26 ± 4.05	–5	–16%	−6.27 to − 3.73	< 0.001
	Severe OSA + MT + CPAP (n = 20)	36 ± 4.06	29 ± 4.02	–7	–19%	−8.24 to − 5.76	< 0.001
	Control (CPAP only) (n = 18)	32.26±3.84	32.1±3.80	NS	—	—	NS

 Values are presented as mean ± standard deviation (SD) unless otherwise indicated. = the absolute change from the baseline

BMI body mass index; CI confidence interval; CPAP continuous positive airway pressure; MT myofunctional therapy; NS not significant; OSA obstructive sleep apnea

Changes in DISE

DISE video recordings at baseline and at follow-up were reviewed independently by two board-certified otolaryngologists who were unaware of the treatment group allocation and clinical data. Their interrater agreement for VOTE scoring (κ = 0.82) was substantial.

In the moderate OSA + MT group, the percentage of patients without tongue base collapse (T0) increased significantly from 15% at the baseline to 80% after the intervention (P =.039). In the severe OSA + MT + CPAP group, T0 increased from 15% to 55%, a change that approached but did not reach statistical significance (P =.109). No significant differences were observed in the control group (P =.20) (Table 3).

Table 3.

Changes in VOTE classification for tongue base collapse on DISE

Group	Baseline T0, n (%)	Baseline T1 + T2, n (%)	Post T0, n (%)	Post T1/T2, n (%)	P value
Moderate OSA + MT (n = 19)	3 (15.8)	16 (84.2)	15 (78.9)	4 (21.1)	0.039
Severe OSA + MT + CPAP (n = 20)	3 (15.0)	17 (85.0)	11 (55.0)	9 (45.0)	0.109
Control (CPAP only) (n = 18)	4 (22.2)	14 (77.8)	5 (27.8)	13 (72.2)	0.200

T0 no tongue base collapse; T1/T2 partial or complete collapse; CPAP continuous positive airway pressure; DISE drug-induced sleep endoscopy; MT myofunctional therapy; OSA obstructive sleep apnea. P values obtained using one-way ANOVA; nonparametric tests applied when distributional assumptions were not met

Changes in IOPI scores

After 3 months, both intervention groups showed significant improvements in tongue and lip strength assessed by the IOPI; no changes were observed in the controls (Table 4). In the moderate OSA + MT group, both tongue and lip pressures increased significantly after 3 months. Similar improvements were also observed in the severe OSA + MT + CPAP group, while no significant changes occurred in the controls (Table 4).

Table 4.

Changes in IOPI measurements after 3 months

Parameter	Group	Pre (mean ± SD)	Post (mean ± SD)	Δ (absolute)	95% CI for Δ	P value(
Tongue pressure, kPa	Moderate OSA + MT (n = 19)	44.35 ± 9.50	54.98 ± 1.57	+ 10.63	5.72 to 15.53	0.002
	Severe OSA + MT + CPAP (n = 20)	35.5 ± 10.68	50.42 ± 1.64	+ 14.92	9.93 to 19.91	< 0.001
	Control (CPAP only) (n = 18)	53.77 ± 7.93	54.16 ± 7.95	NS	–0.84 to 0.066	0.08
Lip pressure, kPa	Moderate OSA + MT (n = 19)	20.4 ± 4.50	33.08 ± 1.34	+ 12.68	10.57 to 14.79	< 0.001
	Severe OSA + MT + CPAP (n = 20)	16.05 ± 4.98	31.29 ± 5.91	+ 15.24	12.58 to 17.89	< 0.001
	Control (CPAP only) (n = 18)	23.36 ± 4.38	23.74 ± 4.72	NS	–1.15 to 0.42	0.30

Values are presented as mean ± standard deviation (SD) unless otherwise indicated. Δ = the absolute change from the baseline

CI confidence interval; CPAP continuous positive airway pressure; IOPI tongue maximal tongue pressure (combined elevation–protrusion effort reflecting genioglossus and intrinsic muscle strength)); IOPI lips maximal lip pressure; MT myofunctional therapy; NS not significant; OSA obstructive sleep apnea. P values obtained using one-way ANOVA; nonparametric tests applied when appropriate

Changes in the apnea–hypopnea index

At 3 months, both MT intervention groups exhibited significant reductions in the AHI compared to baseline, whereas no significant changes occurred in the control group (Table 5).

Table 5.

Changes in apnea–hypopnea index (AHI) after 3 months

Group	Pre (mean ± SD)	Post (mean ± SD)	Δ (absolute)	95% CI for Δ	P value
AHI, events/h
Moderate OSA + MT (n = 19)	22.8 ± 3.61	16.02 ± 4.53	–6.78	–9.50 to − 4.11	0.002
Severe OSA + MT + CPAP (n = 20)	40.9 ± 5.98	29.81 ± 9.51	–11.09	–16.75 to − 5.42	0.0016
Control (CPAP only) (n = 18)	22.6 ± 3.61	20.65 ± 3.32	NS	–4.53 to 0.42	0.071
ODI, events/h
Moderate OSA + MT	21 ± 7	14 ± 6	–7	–10.5 to − 3.5	0.01
Severe OSA + MT + CPAP	35 ± 8	24 ± 9	–11	–16.2 to − 5.8	0.001
Control	20 ± 6	19 ± 6	NS	-	NS
SpO₂ nadir, %
Moderate OSA + MT	82 ± 4	86 ± 3	+ 4	2.0 to 6.0	0.01
Severe OSA + MT + CPAP	79 ± 5	84 ± 4	+ 5	2.0 to 8.0	0.02
Control	83 ± 4	82 ± 5	NS	-	NS

Values are presented as mean ± standard deviation (SD) unless otherwise indicated. Δ = the absolute change from the baseline 

AHI apnea–hypopnea index; CI confidence interval; CPAP continuous positive airway pressure; MT myofunctional therapy; NS not significant; SD standard deviation; OSA obstructive sleep apnea. P values obtained using one-way ANOVA; nonparametric tests applied when appropriate

Both MT intervention groups exhibited significant improvements in ODI and SpO₂ nadir compared to baseline, whereas no changes occurred in controls. Supplementary Table S1 shows that MT led to reductions in both apneas and hypopneas, with a proportionally greater decrease in hypopneas in the moderate OSA group and in apneas in the severe OSA + CPAP group. A responder analysis revealed that > 50% reduction in AHI was achieved in 53% of moderate OSA + MT participants and 60% of severe OSA + MT + CPAP participants, with 37% and 30% respectively also reaching a post-treatment AHI < 10.

Discussion

This study provides, for the first time, objective evidence that telemedicine-delivered MT can induce measurable structural changes in the upper airway of patients with OSA. Submental ultrasound [17, 19] revealed significant reductions in tongue volume, along with a decreased distance between the lingual arteries even after only 3 months of treatment. Interestingly, tongue thickness did not change significantly in the severe OSA group, whereas other ultrasound parameters improved. These anatomical changes were accompanied by measurable gains in tongue strength [28] and improvements in DISE parameters [23–25] as well as in the magnitude of respiratory disturbance during sleep, as illustrated by the AHI. In addition to the reduction in AHI, MT also led to significant improvements in other polygraph-derived parameters. Both intervention groups showed increases in SpO₂ nadir and decreases in ODI compared to baseline, whereas no significant changes occurred in controls (Table 5). Analysis of respiratory events (Supplementary Table S1) revealed reductions in both apneas and hypopneas, with a proportionally greater decrease in hypopneas among moderate OSA patients and in apneas among those with severe OSA on CPAP. A responder analysis further demonstrated that > 50% reduction in AHI was achieved in 53% of moderate OSA + MT participants and 60% of severe OSA + MT + CPAP participants, with 37% and 30% respectively also meeting the combined criterion of > 50% reduction and a post-treatment AHI < 10. Exploratory correlations showed that reductions in tongue volume and increases in IOPI tongue pressure were significantly associated with reductions in AHI, supporting the hypothesis that structural and functional remodeling of the upper airway underlies the observed clinical improvements.

As previously documented in a separate study [28], the AHI decreased significantly after 3 months of MT in both intervention groups and further supported by recent systematic reviews and meta-analyses, myofunctional therapy significantly reduces AHI in patients with OSA [42–44]; These findings suggest that MT can positively influence OSA severity beyond subjective symptom relief [42], likely by improving oropharyngeal muscle tone and reducing UA collapsibility. In contrast, no significant changes in the AHI were observed in the control group, which suggests that the effect of the AHI was attributable to the MT intervention rather than spontaneous night-to-night variability or CPAP alone [45]. Previous studies of MT have primarily focused on functional outcomes, such as reduction in the AHI [41] or improvements in daytime symptoms [46], without objectively demonstrating anatomical remodeling. Current findings fill this important gap and confirm that targeted oropharyngeal exercises can modify both the function and anatomical substrate that contributes to UA collapse.

The clinical implications of these results are potentially far reaching. Our finding that MT can alter UA anatomy supports its role as a viable adjunctive therapy for OSA, particularly in patients with moderate disease or those with poor CPAP adherence [47], or, although not tested here, when surgical intervention is not indicated. By addressing the structural factors that contribute to pharyngeal collapsibility, MT may improve airway stability during sleep and reduce the severity of respiratory disturbance [4].

There has been ongoing debate regarding whether daytime training can counteract the nocturnal reduction in upper airway dilator muscle tone [8]. Previous reports have emphasized the limited structural evidence supporting the use of MT [48]. The present study contributes to this discussion by providing reproducible imaging data that demonstrate structural remodeling following objectively assessed MT.

The telemedicine format of the intervention deserves comment as well. We believe that continued interactions between the patient and the treatment team are an essential factor in the success of this intervention program. Allowing patients to complete their daily exercises at home without an imposed rigid schedule reduced the barriers to adherence compared with traditional face-to-face programs [49]. Remote monitoring ensured that progress could be tracked, and prompt feedback was given if adherence declined [50]. Rather than replacing the therapeutic relationship, this approach appeared to strengthen it, as patients reported feeling more supported and clinicians were able to make timely adjustments to the exercise plan [51].

In addition to ultrasound, we used objective functional measures, including the IOPI [52], to quantify tongue strength. These tools provide reproducible data on neuromuscular performance and can also serve as indirect indicators of MT adherence. Measurable gains were observed only in patients who trained consistently [53]. In our previous work, improvements in these measures correlated with better clinical outcomes, namely reductions in AHI, decreased daytime sleepiness, and improvements in sleep-related quality of life, thereby enabling a valuable complement to imaging for monitoring the effectiveness of MT [54].

The mechanisms underlying the observed anatomical changes are likely multifactorial [55]. Repeated isotonic and isometric activation of the oropharyngeal muscles may increase the baseline muscle tone and promote hypertrophy of the targeted muscle fibers, particularly in the genioglossus and intrinsic tongue musculature. Improved neuromuscular coordination may also enhance the capacity to maintain airway patency during sleep [56]. Prior imaging studies have shown that targeted training and weight loss can reduce intramuscular fat infiltration within the tongue, which is strongly associated with airway collapsibility [4, 57]. Although we did not measure fat content directly, the reductions in tongue volume and thickness observed here are consistent with such possible decrease in fat tissue infiltration in the tongue. Mouth breathing may modulate both the baseline anatomy and the functional response to myofunctional therapy [6]. This pattern is associated with lower tongue posture, diminished oropharyngeal tone, and greater airway collapsibility [36, 37]. The improvement of nasal breathing and tongue position through targeted exercises may therefore have contributed to the structural and functional gains observed after MT [33].

In our cohort, the severe OSA group exhibited a greater reduction in the AHI despite showing smaller changes in submental ultrasound measurements. In contradistinction, the moderate OSA group achieved more pronounced anatomical remodeling but a smaller decrease in the AHI. These differences may reflect distinct mechanisms for the response to MT according to the baseline disease severity. In moderate OSA, improvements could be primarily driven by structural remodeling of the UA and increased muscle tone, as evidenced by significant ultrasound changes and gains in the IOPI score. Conversely, functional adaptations in severe OSA, such as improved neuromuscular coordination, altered pharyngeal dilator recruitment, and reduced collapsibility during sleep as documented on DISE, may play a more important role, even when structural changes are less pronounced. These patterns underscore the multifactorial nature of the effects of MT and suggest that anatomical and functional endpoints should be considered jointly when evaluating MT therapy outcomes.

One participant discontinued MT because of the onset of TMJ discomfort, despite having undergone a negative TMJ evaluation before therapy initiation. Although preintervention screening can reduce this risk, this isolated incident highlights the potential for TMJ symptoms to emerge during therapy, possibly related to changes in oral posture and increased muscular activity. For this reason, ongoing monitoring of TMJ status is advisable throughout MT protocols, and collaboration with maxillofacial specialists should be considered when symptoms arise.

A methodological strength of this study is the complementary use of two independent assessment modalities. Although submental ultrasound provides a noninvasive and repeatable method to quantify tongue volume, thickness, and interarterial distance, it remains partially operator dependent and subject to measurement variability [16]. To mitigate this limitation, we incorporated DISE for the objective dynamic evaluation of UA collapse. All DISE procedures were recorded, anonymized, and reviewed blindly by two independent expert otolaryngologists, who achieved optimal interrater agreement [43, 58]. This dual approach strengthens the validity of our anatomical findings and also ensures that the changes observed in airway structure are corroborated by real-time functional assessment during induced sleep, thereby reducing the potential bias associated with any single measurement technique. Of course, overnight polygraphy at the beginning and upon completion of the trial further reinforce the validity of the findings.

Limitations

One of the limitations of this study is the reliance on submental ultrasound for anatomical assessment because operator-dependent variability can affect reproducibility. We minimized this potential issue by using the same experienced examiner for all measurements and blinding the scorer to group allocation. The inclusion of a control group of moderate to severe OSA patients on stable CPAP therapy who were assessed under identical conditions also helped mitigate the potential for measurement bias.

A second limitation is the nonrandomized study design, which leaves open the possibility of selection bias despite similar baseline characteristics among the three groups. Another limitation concerns the relatively short follow-up period of 3 months, which does not allow us to draw conclusions about the long-term stability of the observed changes or to establish whether further improvements are possible. Thus, longer interventions and future studies including discontinuation of MT with longitudinal monitoring of functional and structural changes are warranted.

The cohort included predominantly patients with a low average BMI (26.97 kg/m²), and no morbidly obese patients were evaluated. Consequently, our findings may not be fully generalizable to patients with OSA and obesity, whose tongue fat content and airway collapsibility may substantially differ and therefore respond differently to MT. We also did not quantify lingual fat content directly, which could have helped clarify whether fat reduction contributed to the decrease in tongue volume.

Moreover, we did not systematically re-titrate CPAP pressure requirements after the intervention. Therefore, we cannot determine whether the observed improvements in airway anatomy and function translated into lower CPAP pressure needs. Future studies including CPAP retitration after MT would help clarify whether MT reduces therapeutic pressure requirements by decreasing upper airway collapsibility.

Lastly, we did not apply formal corrections for multiple comparisons. Given the exploratory and hypothesis-generating nature of this study and the relatively small sample size, we prioritized sensitivity to potential treatment effects. However, this increases the risk of type I error, and therefore the findings should be interpreted with caution and confirmed in larger randomized trials.

Conclusions

Telemedicine-delivered MT over a period of 3 months resulted in measurable UA remodeling in patients with OSA, as shown by the significant reductions in tongue volume and thickness, improved tongue strength, and favorable changes in tongue base collapse on DISE, particularly in those with moderate disease, as further illustrated by improvements in nocturnal respiratory disturbance. These findings provide objective structural and functional evidence that support MT as a noninvasive adjunctive treatment of non-obese or overweight OSA patients. Integrating ultrasound and DISE with functional measures enables comprehensive monitoring, and the telemedicine format enhances patient accessibility and adherence. Larger, long-term randomized trials are needed to confirm the durability and clinical impact of telemedicine-delivered MT.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AHI

Apnea–hypopnea index

BMI

Body mass index

CI

Confidence interval

CPAP

Continuous positive airway pressure

DISE

Drug-induced sleep endoscopy

IOPI

Iowa Oral Performance Instrument

MT

Myofunctional therapy

OSA

Obstructive sleep apnea

TMJ

Temporomandibular joint

T0

Percentage of patients without tongue base collapse

UA

Upper airway

VOTE

Velum, oropharynx, tongue base, and epiglottis

Author contributions

C.R.A. and C.O.R.: data analysis, manuscript drafting, critical revision. L.R.A.: screening, data collection and analysis, manuscript drafting. J.M.I.: critical revision. G.P. and P.B.: study conception, screening, bias assessment, drafting, critical revision. D.G.: screening, critical revision. All authors reviewed and approved the final manuscript.

Funding

This study did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Consent for publication

Written informed consent for publication of anonymized data and images was obtained from all participants included in this study.

Ethical approval and consent for participation

The study was approved by the Provincial Research Ethics Committee of Málaga, Spain (approval code: AWGAP-2023-2). All procedures were conducted in accordance with the ethical standards of the institutional and national research committees and with the 1964 Declaration of Helsinki and its later amendments. Written informed consent was obtained from all participants prior to their inclusion in the study.

Competing interests

C.O.R. is the creator of the Airway Gym^® mobile application. The other authors declare no conflicts of interest.

Clinical trial registration

How muscle exercise affects the amount of fat in the tongue of patients with obstructive sleep apnea, using ultrasound to measure the changes. ISRCTN (ISRCTN92645461); 10.1186/ISRCTN92645461.