Positional therapy for positional sleep apnea

Some people stop breathing mainly on their backs. Rosalind Cartwright’s 1984 paper in Sleep is still the usual definition of that pattern: the apnea-hypopnea index at least twice as high supine as it is off the back. Positional therapy tries to keep you off your back. A tennis ball sewn into a T-shirt is folklore with a short half-life. It does not replace CPAP when apnea is just as bad on your side.

What is positional obstructive sleep apnea?

Cartwright studied 30 men in a row who came for clinical polysomnography. In the 24 who spent time both on the back and on the side, the apnea index was twice as high supine. The difference tracked inversely with obesity. Five men who met diagnostic criteria for the whole night fell into the normal range when they were on their side. That is the phenotype in one paragraph: gravity and a throat that still has a chance in the lateral position.

Later authors tried stricter rules. Marklund wanted a supine AHI of at least 10 and a lateral AHI under 10. Mador and others wanted the non-supine AHI under 5. Bignold required enough minutes in each posture and a non-supine AHI under 15. Ravesloot’s 2017 review in the Journal of Clinical Sleep Medicine walks those variants and notes that Cartwright’s 50 percent cutoff remains the one most papers still use, arbitrary as she called it.

Ravesloot also summarized the epidemiology that matters at the visit. In about 56 to 75 percent of people with OSA, position influences how often and how long the events last. Most positional patients have mild or moderate disease (she cites 70 to 80 percent). They tend to be younger and leaner than people whose apnea is bad in every posture. As severity and BMI rise, the positional slice shrinks. Someone with a high non-supine AHI does not have a tennis-ball problem. They have an airway that collapses on the side too.

How do you know you have the phenotype?

You do not know from a spouse saying you snore more on your back. You know from a sleep study that recorded body position against the breathing tracing, with enough time in both postures to make the ratio mean something. A home test can show a positional pattern if it has a position channel and you actually slept in both positions. A lab night that was entirely supine can make everyone look positional. A night with twelve minutes on the back can hide it.

Ask for the supine AHI and the non-supine AHI, not only the headline number. If the overall AHI is 18, the back is 40, and the side is 4, positional therapy is in the conversation. If the overall is 18, the back is 22, and the side is 16, it is not. Cartwright’s doubling rule is the start of that conversation, not a treatment order.

The study also has to be current enough to reflect today’s weight and today’s alcohol pattern. People who lose a large amount of weight sometimes become positional. People who gain can lose the positional pattern. Ravesloot notes both directions. Repeat testing is how you find out, not a memory of a 2014 report.

Does the tennis-ball trick work?

Mechanically, a bulky object on the back makes supine sleep unpleasant, so you roll. Oksenberg’s small efficacy series, cited inside Bignold’s later compliance paper, saw large drops in supine time and in overall AHI while people used it. The problem is not the physics. The problem is that almost nobody keeps doing it.

Bignold and colleagues mailed a questionnaire to 108 patients prescribed the tennis ball technique at an Australian sleep lab. Sixty-seven answered. Mean follow-up was 2.5 years. Four people (6 percent) were still using it. Nine said they had learned to stay off their backs without the ball. Fifty-four (about 81 percent of respondents) were neither using the device nor avoiding supine sleep. Among those who quit, the leading reason was that it was too uncomfortable (34 of 54). Others said the ball migrated, sleep and alertness did not improve, it did not keep them off their backs, or it caused backache. Bignold’s conclusion is the one to keep: long-term compliance appears very poor, with fewer than 10 percent reporting continued use at about 30 months.

Ravesloot’s review put the same folklore in a range: short-term compliance in the literature around 40 to 70 percent, long-term around 10 percent, discomfort as the usual reason. Sewing a pocket into a nightshirt is cheap. Untreated positional apnea after you throw the shirt in a drawer is not a treatment plan.

What are the newer positional devices?

The devices Ravesloot meta-analyzed are small packs worn on the chest or neck that vibrate when you roll supine, a nudge rather than a bruise. Combined data in that 2017 review showed a mean AHI drop of 11.3 events per hour (a 54 percent reduction) and a 33.6 percentage-point drop in the share of sleep spent on the back (an 84 percent reduction in that metric). Under short-term study conditions, people used the devices. The authors were explicit that long-term compliance could not be judged because the reliable data were not there. Additional high-quality long-term studies were needed. That sentence is still the honest one to tell a patient in 2026: the short-term physiology looks real; the decade-long adherence story is thinner than CPAP’s download record, and thinner than anyone selling a belt would like.

A 2023 Thorax systematic review of vibrotactile devices reached a similar practical point. Traditional bulky methods are efficacious and poorly tolerated, with long-term compliance cited at 10 percent. Newer haptic devices can be monitored, which at least tells you whether the belt was on. Monitoring is not the same as a cure. If the download shows you still spend a third of the night on your back, or you stopped wearing it in week three, the AHI on the original study is the AHI you have again.

This clinic does not need to name a consumer brand to make the clinical point. If positional therapy is offered, it should be after a study that showed the Cartwright pattern, with a plan to retest or to review device data, and with a backup if you cannot wear it. A belt from a late-night ad, without a diagnosis, is snoring theater.

Does positional therapy replace CPAP?

Not for non-positional disease. If the side-sleeping AHI is already in the moderate or severe range, keeping you off your back leaves a disease that still needs treatment. Ravesloot’s prevalence figures run the other way from the marketing: positional therapy is most plausible in milder, leaner, younger patients whose events really do cluster supine. It is a poor substitute for CPAP when obstruction is posture-independent.

Even in true POSA, CPAP still splints the airway in every position. People who can use the machine will usually get a lower residual AHI than people who only avoid the back. Positional therapy’s pitch is adherence and simplicity in a subset, not superiority on the tracing. If you already tolerate PAP, there is no guideline that tells you to take the mask off and sleep with a vibrator on your sternum instead.

Combination use exists in the literature (a mandibular device plus position, or PAP plus position). That is individual, and it still depends on the non-supine AHI. Do not add gadgets to hide an undertreated number.

What should you expect from adherence?

Expect the same human problem CPAP has, with worse documentation unless the device logs wear. Tennis-ball users quit because it hurts. Vibrotactile users quit because the buzz wakes them, the pack is annoying, they travel without it, or they feel fine and stop. Short-term trials select people who will show up for a second sleep study. Your third month on a weeknight in Westerville is a different experiment.

If you try positional therapy, decide in advance what failure looks like: still sleepy, still snoring on the side, device unworn, or a follow-up study that shows the non-supine AHI was never safe. Then use one of the actual alternatives: PAP, a custom oral appliance when that is the right tool, weight management, or, in a tightly defined group, hypoglossal nerve stimulation. The CPAP alternatives page is the map. This article is only the positional corner of it.

Alcohol and sedatives make the back worse for many people and can create events on the side that were not there on the diagnostic night. A positional plan that assumes you never drink is a plan for someone else.

How do you find out if you are even a candidate?

You get a sleep study that records position, then a physician reads the ratio. Testing is the prerequisite. Self-training with a backpack full of tennis balls before the diagnosis is how people delay care for obstructive sleep apnea that is not positional at all.

Bring the old report if you have one. If it lacks position-specific indices, it cannot support this treatment. Call (614) 898-9340 and ask whether the study you already had is enough, or whether it needs to be repeated with the channels that make Cartwright’s ratio visible.