Sleep apnea and yoga have been tested directly in two small randomized trials, and only one of them measured breathing pauses. In 37 overweight adults with sleep apnea, a yoga-based lifestyle program did not change average severity more than diet and exercise advice alone, but 44% of the yoga group cut their apnea-hypopnea index by more than 40%, against 11% of the other group. Targeted throat and tongue exercises have better evidence: a meta-analysis of nine adult studies found the index fell from 24.5 to 12.3 events per hour, and a randomized trial in 31 people with moderate apnea saw it fall from 22.4 to 13.7. Averages stayed above normal in every study, so exercise is studied as an add-on to treatment and not a replacement.
Anyone told they have sleep apnea eventually asks what they can do with their own body, apart from the machine, the dental device or the surgery. Exercise is the obvious answer. Sleep apnea exercises are easy to find, and the internet is full of routines sold as yoga for sleep apnea. The short version is that throat exercises have a reasonable amount of data, yoga has a little, and no exercise brought the average person back to normal.
Sleep apnea and yoga in two randomized trials
The most direct test is a randomized trial called ELISA. It enrolled 37 patients with obesity and obstructive sleep apnea confirmed by a sleep study (apnea-hypopnea index above 5), 19 in a control group and 18 in a yoga group. Both groups got counselling on diet (a staple Indian diet) and regular exercise. The yoga group additionally followed a yoga module for sleep apnea, with postures, breathing techniques, meditation and principles for daily living. Sleep studies were repeated after a year.
The result is easy to misread in both directions. The change in average severity did not differ significantly between the two groups, and neither did weight loss. Inside the yoga group, though, the average index fell from 51.2 to 36.8 events per hour (P=0.003), while inside the control group the fall from 47.2 to 38.8 was not statistically clear (P=0.08). And 8 of 18 people in the yoga group (44%) reduced their index by more than 40%, against 2 of 19 (11%) in the control group (P=0.02). The authors concluded that a yoga-based lifestyle program may reduce severity in obese patients.
Gupta A, Kaur J, Shukla G, Bhullar KK, Lamo P, Kc B, Agarwal A, Srivastava AK, Sharma G. Effect of yoga-based lifestyle and dietary modification in overweight individuals with sleep apnea: A randomized controlled trial (ELISA). Sleep Med 2023. PMID 37178546
Two cautions apply. The trial was small, and its participants started with severe apnea, an average of around 50 events per hour. After a year of yoga the average was still 36.8, which is severe on the usual scale. The study shows a signal worth following up, not a treatment that stands on its own.
A second trial studied yoga delivered by video call. It randomized 44 adults with sleep apnea (40 men and 4 women) to live group sessions of 60 minutes, three days a week for 12 weeks, or to home thoracic expansion exercises four times a day. The yoga group improved more on respiratory muscle strength, an exercise test, two cognitive tests, sleep duration, sleep efficiency, sleep quality and the Epworth sleepiness score. Its summary reports no apnea-hypopnea index, so it says nothing about breathing pauses. What it does show is that people felt and tested better in several areas that matter in daily life.
Eyuboglu F, Inal-Ince D, Karamancioglu B, Vardar-Yagli N, Kaya-Utlu D, Dalkilic O, Somay G. Effect of tele-yoga on aerobic capacity, respiratory muscle strength, and cognitive performance in patients with obstructive sleep apnea syndrome. Heart Lung 2023. PMID 37536116
Throat and tongue exercises: the bigger evidence base
The exercises with the best data are not yoga. They are specific routines for the muscles around the airway, usually called oropharyngeal or myofunctional exercises. A randomized trial gave 31 patients with moderate sleep apnea either sham therapy or about 30 minutes a day, for three months, of exercises for the tongue, the soft palate and the side walls of the throat. The control group did not change on any measure. In the exercise group, the average apnea-hypopnea index fell from 22.4 to 13.7 events per hour, snoring frequency and intensity dropped, the Epworth score fell from 14 to 8, and neck circumference decreased from 39.6 to 38.5 cm. The authors called it a promising treatment for moderate apnea.
Guimarães KC, Drager LF, Genta PR, Marcondes BF, Lorenzi-Filho G. Effects of oropharyngeal exercises on patients with moderate obstructive sleep apnea syndrome. Am J Respir Crit Care Med 2009. PMID 19234106
A systematic review and meta-analysis pooled this kind of work. It found nine adult studies with 120 patients reporting sleep study results after myofunctional therapy for sleep apnea. The average index fell from 24.5 to 12.3 events per hour, the lowest oxygen level during sleep rose from 83.9% to 86.6%, snoring measured in the sleep lab dropped from 14.05% to 3.87% of sleep time, and the Epworth score went from 14.8 to 8.2. Two small studies in children pointed the same way. The authors concluded that the index decreases by roughly 50% in adults and 62% in children, and that the therapy could serve as an adjunct to other sleep apnea treatments.
Camacho M, Certal V, Abdullatif J, Zaghi S, Ruoff CM, Capasso R, Kushida CA. Myofunctional Therapy to Treat Obstructive Sleep Apnea: A Systematic Review and Meta-analysis. Sleep 2015. PMID 25348130
Look at the "after" numbers in those studies. An average of 12 to 14 events per hour is much better than 22 or 25, and on the usual scale it is still mild apnea (5 to 15), not normal. Some people will have done better than average and some worse, because averages hide that spread.
The didgeridoo study
The strangest trial in this area tested the didgeridoo. Researchers randomized 25 adults with an apnea-hypopnea index between 15 and 30 who complained of snoring. The treatment group took didgeridoo lessons and practiced at home with standardized instruments for four months, on average 25.3 minutes a day on 5.9 days a week. The control group stayed on the waiting list.
Compared with controls, the didgeridoo group had less daytime sleepiness (Epworth difference -3.0, 95% confidence interval -5.7 to -0.3), a lower apnea-hypopnea index (difference of -6.2 events per hour, interval -12.3 to -0.1, P=0.05) and partners who rated their sleep as less disturbed. Sleep quality scores and health-related quality of life did not differ. The idea behind it was that playing reduces the collapsibility of the upper airway, which is close to the logic of the throat exercises. Participants kept at it, practicing on almost six days of the week.
Puhan MA, Suarez A, Lo Cascio C, Zahn A, Heitz M, Braendli O. Didgeridoo playing as alternative treatment for obstructive sleep apnoea syndrome: randomised controlled trial. BMJ 2006. PMID 16377643
What these studies cannot tell you
Some of the limits are built into the designs. Most trials were small, with 25 to 44 people, and several compared an exercise program with waiting or with counselling alone, so participants knew which group they were in. Sleep apnea was measured with sleep studies, but the studies report group averages. If you read the yoga numbers, the same average drop can come from one person improving a lot and another not at all. And none of these studies compared exercise head to head with CPAP or a dental device, so no one can say how an exercise program ranks against them.
What the studies agree on is a direction. Training the muscles of the tongue and throat made the airway somewhat less likely to close, in groups with moderate disease. Weight, which matters for both severity and for how the muscles work, is the other lever (see Weight loss for snoring), and so is sleeping position (Best position to sleep in for sleep apnea). The wider list of do-it-yourself options is in Sleep apnea natural treatments.
Do not stop prescribed treatment to try exercises. If you wake gasping, if a partner sees you stop breathing, or if you fall asleep during the day or while driving, see a doctor. Exercise in these studies was something people did in addition to a diagnosis and a doctor's follow-up.
Telling whether the exercises changed anything
Breathing pauses are hard to judge by feel, which is one reason the studies used sleep labs. At home, a few weeks of recordings before and after a change give a rough picture of snoring. If you want that baseline, record several nights before you start anything new, so that a good or bad single night does not fool you.
For a before and after comparison, SleepTrace listens through the phone's microphone and estimates sleep phases from sound. It recognizes snoring, talking and coughing and keeps short clips for those (3 clips a month can be played back on the free plan, Premium has no limit). Its sleep apnea indicator gives a risk level as a screening, not a diagnosis, and it does not replace the sleep study doctors use to measure apnea. Keep the app open overnight with the phone on a charger; the screen dims by itself.
Frequently asked questions
There is no evidence for a cure. In a randomized trial of 37 overweight adults with sleep apnea, the yoga group's average apnea-hypopnea index fell from 51.2 to 36.8 events per hour over a year, which is still severe by the usual scale, and the difference between the yoga and control groups' averages was not statistically significant. More people in the yoga group (8 of 18) than in the control group (2 of 19) cut their index by more than 40%.
Three kinds have trial data: yoga (postures, breathing techniques and meditation), oropharyngeal or myofunctional exercises for the tongue, soft palate and throat wall, and playing the didgeridoo. The throat exercises have the most studies. A meta-analysis found the apnea-hypopnea index fell by roughly half in adults, from 24.5 to 12.3 events per hour, in nine studies with 120 patients.
The trials only measured at the end. The throat exercise trial ran 3 months with about 30 minutes a day, the didgeridoo trial 4 months with about 25 minutes a day, and the yoga lifestyle trial 12 months. None of them reported how fast the changes appeared along the way.
The trials compared exercises with waiting, sham therapy or counselling, not with CPAP. The meta-analysis authors describe myofunctional therapy as something that could serve as an adjunct to other treatments, and the participants in the trials still had an average index above normal at the end. Decisions about treatment belong with a sleep doctor, especially when sleepiness during the day is already a problem.
In a randomized trial of 25 adults with moderate sleep apnea who also snored, four months of lessons and daily practice reduced daytime sleepiness by 3.0 points on the Epworth scale and the apnea-hypopnea index by 6.2 events per hour compared with waiting-list controls. The trial was small and people knew which group they were in, so it is one interesting result and not a settled treatment.
SleepTrace is a wellness app, not a medical device. This article is general information, not medical advice. If your symptoms are frequent, severe or worrying, please talk to a doctor.
If you suspect apnea, the first step is a doctor and a sleep study, not an experiment with exercises. SleepTrace records sound only, saves snoring clips and marks possible breathing pauses with a time. It is a screening indicator and not a measurement of your apnea index, and the finding belongs with a doctor.
Hear your own night. SleepTrace turns a night of audio into your sleep phases, the sounds you made, and how it all trends. Works without a wearable, better with one. Your phone on the nightstand is enough to start. Get it for iPhone or Android →