Are interventions to keep people sleeping on their side the best way to treat obstructive sleep apnoea?

What is obstructive sleep apnoea?

Obstructive sleep apnoea (OSA) is sleep disorder where the walls of the throat relax and narrow during sleep. This causes pauses in the breathing. The pauses can last for a few seconds to a few minutes and can happen many times in the night. This disrupts the person's sleep. Bed partners can be disturbed by associated loud snoring, chocking and snorting sounds. People living with OSA can be very tired in the day or even fall to sleep. This can be dangerous. In children, sleep apnoea can cause problems at school or hyperactivity.

What is positional OSA?

OSA that improves on changing position of the person while sleeping is known as positional sleep apnoea (POSA). People tend to have apneas when lying on their backs (supine) and the apneas may be reduced or go away when they lay on their side.

What is the standard in the treatment of OSA?

The standard treatment is a device called continuous positive airway pressure (CPAP) that provides a continuous jet of air to the airway as the person breathes, which helps to keep the throat from narrowing during sleep.

What is positional therapy?

Positional therapy is an intervention that helps to keep the person on their side during sleep. Examples include something on the person's back to stop them from rolling over (like a tennis ball), special pillows, or alarms that vibrate when the person rolls onto his or her back.

How is the severity of OSA estimated?

The severity of OSA is measured using a scale called Apnoea-Hypopnea Index (AHI). AHI refers to number of times the breathing stops or becomes shallower per hour of sleep. AHI is measured by using a study done in sleep known as polysomnography.

The severity of OSA can be measured indirectly using a questionnaire called Epworth Sleepiness Scale (ESS). This assesses how sleepy a person is during the day.

What was the aim of this review?

We wanted to compare positional therapy with the CPAP therapy as well as with inactive control (no positional therapy or sham therapy).


We found eight studies with 323 participants. The studies compared positional therapy with CPAP (72 participants) and positional therapy with inactive control (251 participants).

When studies compared positional therapy and CPAP, they found no difference in ESS between groups. CPAP therapy showed a greater improvement in AHI (6.4 fewer events per hour with CPAP) compared with positional therapy. In one small study, people adhered to positional therapy for 2.5 hours longer than they did for CPAP. No difference in quality of life or quality of sleep between the two groups was found.

In the comparison between positional therapy and inactive control, the studies found that positional therapy appeared to be better than control for ESS and AHI (ESS was 1.58 lower in positional therapy and AHI was 7.38 fewer events per hour with positional therapy). Another study noted adverse effects in 10% of participants. Common adverse effects were sleep disturbance and pain in the back and the chest. One study reported that there was no difference in quality of life and quality of sleep between positional therapy and inactive control.

All these studies lasted for a short time and included small number of participants.


1. Positional therapy was less effective than CPAP for reducing Apnoea-Hypopnoea Index (AHI). People may use positional therapy for longer than CPAP in the night. In terms of other outcomes no differences were seen.

2. Positional therapy was shown to be better than inactive control for AHI and Epworth Sleepiness Scale (ESS).

Authors' conclusions: 

The review found that CPAP has a greater effect on improving AHI compared with positional therapy in positional OSA, while positional therapy was better than inactive control for improving ESS and AHI. Positional therapy may have better adherence than CPAP. There were no significant differences for other clinically relevant outcomes such as quality of life or cognitive function. All the studies were of short duration. We are unable to comment on the long-term effects of the therapies. This is important, as most of the quality-of-life outcomes will be evident only when the therapies are given over a longer period of time. The certainty of evidence was low to moderate.

Read the full abstract...

The modalities of therapy for obstructive sleep apnoea (OSA) include behavioural and lifestyle modifications, positional therapy, oral appliances, surgery and continuous positive airway pressure therapy (CPAP). Though CPAP has proven efficacy in treating OSA, adherence with CPAP therapy is suboptimal. Positional therapy (to keep people sleeping on their side) is less invasive and therefore expected to have better adherence. This review considered the efficacy of positional therapy compared to CPAP as well as positional therapy against no positional therapy. Devices designed for positional therapy include lumbar or abdominal binders, semi-rigid backpacks, full-length pillows, a tennis ball attached to the back of nightwear, and electrical sensors with alarms that indicate change in position.


To compare the efficacy of positional therapy versus CPAP and positional therapy versus inactive control (sham intervention or no positional therapy intervention) in people with OSA.

Search strategy: 

We identified studies from the Cochrane Airways' Specialised Register (including CENTRAL, MEDLINE, Embase, CINAHL, AHMED and PsycINFO),, and the World Health Organization trials portal (ICTRP). It also contains results derived from handsearching of respiratory journals and abstract books of major annual meetings. We searched all databases from their inception to September 2018, with no restrictions on language of publication or publication type.

Selection criteria: 

We included randomised controlled trials comparing positional therapy with CPAP and positional therapy with inactive control.

Data collection and analysis: 

Two review authors independently selected studies and extracted the data. We used a random-effects model in the meta-analysis to estimate mean differences and confidence intervals. We assessed certainty of evidence using the GRADE approach.

Main results: 

We included eight studies. The studies randomised 323 participants into two types of interventions. The comparison between positional therapy and CPAP included 72 participants, while the comparison between positional therapy and inactive control included 251 participants. Three studies used supine vibration alarm devices, while five studies used physical positioning like specially designed pillows or semirigid backpacks.

Positional therapy versus CPAP

The three studies included for this comparison were randomised cross-over trials. Two studies found that there was no difference in Epworth Sleepiness Scale (ESS) scores between CPAP and positional therapy. Two studies showed that CPAP produced a greater reduction in Apnoea-Hypopnoea Index (AHI) with a mean difference (MD) of 6.4 events per hour (95% CI 3.00 to 9.79; low-certainty evidence) compared to positional therapy. Subjective adherence, evaluated in one study, was found to be significantly greater with positional therapy (MD 2.5 hours per night, 95% CI 1.41 to 3.59; moderate-certainty evidence).

In terms of secondary outcomes, one study each reported quality-of-life indices and quality-of-sleep indices with no significant difference between the two groups. One study reported cognitive outcomes using multiple parameters and found no difference between the groups. There were insufficient data to comment on other secondary outcomes like respiratory disturbance index (RDI), and frequency and duration of nocturnal desaturation. None of the studies clearly reported adverse effects.

Positional therapy versus inactive control

Three studies of positional therapy versus no intervention were randomised cross-over trials, while two studies were parallel-arm studies. Data from two studies showed that positional therapy significantly improved ESS scores (MD −1.58, 95% CI −2.89 to −0.29; moderate-certainty evidence). Positional therapy showed a reduction in AHI compared with control (MD −7.38 events per hour, 95% CI −10.06 to −4.7; low-certainty evidence). One study reported adherence. The number of participants who continued to use the device at two months was no different between the two groups (odds ratio (OR) 0.80, 95% CI 0.33 to 1.94; low-certainty evidence). The same study reported adverse effects, the most common being pain in the back and chest, and sleep disturbance but there was no significant difference between the two groups in terms of device discontinuation (OR 1.25, 95% CI 0.5 to 3.03; low-certainty evidence). One study each reported quality-of-life indices and quality-of-sleep indices, with no significant difference between the two groups. One study reported cognitive outcome, and found no difference between the groups. There was insufficient evidence to comment on other secondary outcomes (RDI, frequency and duration of nocturnal desaturation).