CPAP still outperforms mouth-guard alternatives, but the size of the gap is unclear
Pooling eight trials, oral appliances left nearly 12 more breathing events per hour than CPAP. But the trials disagreed with each other so much that the researchers say that number itself should be read with caution.
| Group | Value (events per hour) |
|---|---|
| Higher than with CPAP | 11.83 (6.61 to 17.04) |
| Lower than with sham or control | 15.78 (8.43 to 23.13) |
CPAP remains the standard first-line treatment for obstructive sleep apnea, but long-term adherence is a persistent, well-documented problem — many patients struggle to use the machine consistently. Mandibular advancement devices (MADs), oral appliances that reposition the jaw to keep the airway open, are increasingly used as an alternative. A meta-analysis published this month in Sleep Medicine pools the randomized evidence comparing the two directly [s1].
The design
Researchers systematically searched PubMed, Scopus, and the Cochrane Library following PRISMA guidelines for randomized controlled trials comparing mandibular advancement devices against either CPAP or sham/control interventions in adults with obstructive sleep apnea [s1]. Meta-analysis used random-effects restricted maximum likelihood models with generic inverse variance methodology; for crossover trial designs, the Elbourne method was used to impute paired standard deviations where needed [s1]. Eight randomized controlled trials involving 642 participants were included [s1].
What it found
Compared with CPAP, mandibular advancement device therapy showed significantly higher residual apnea-hypopnea index (AHI, breathing interruptions per hour) — meaning CPAP controlled breathing events more effectively: mean difference 11.83 events per hour favoring CPAP (95% CI 6.61–17.04) [s1]. But this estimate came with extreme statistical heterogeneity across the pooled trials (I² = 97.75%, an unusually high figure indicating the individual trials' results varied widely from each other, well beyond what chance alone would explain), which the study's authors say should make readers "interpret cautiously" and treat the precision of the estimate skeptically [s1].
Against sham or control devices — rather than CPAP — mandibular advancement devices did show a significant benefit: mean difference −15.78 events per hour (95% CI −23.13 to −8.43), though this comparison also carried substantial heterogeneity (I² = 86.47%) [s1]. Secondary outcomes showed generally favorable but variable responses to MAD therapy [s1].
Why the extreme heterogeneity is the real headline, not just a footnote
An I² value above 97% is unusually high even by the standards of meta-analyses of heterogeneous clinical interventions, and it means the eight pooled trials didn't produce reasonably consistent results that a single averaged number can meaningfully summarize — some trials likely showed MADs performing much closer to CPAP, while others showed a much larger gap, and the pooled 11.83-events-per-hour figure is a statistical average across that wide spread rather than a number any single trial actually demonstrated. The study's own authors flag this directly, and it's the detail that should shape how confidently anyone treats the headline number [s1] — a caution as important as the finding itself.
What likely drives that inconsistency
The meta-analysis doesn't fully resolve why the eight trials diverged so much, but plausible contributors implicit in any pooled analysis of this kind include differences in apnea severity across trial populations (mild versus moderate versus severe apnea likely responds differently to a jaw-repositioning device than to pressurized air), differences in how well individual mandibular devices were fitted and titrated, and differences in trial duration and outcome-measurement methods.
What this study still tells us with more confidence
Despite the caution warranted around the precise magnitude, the overall conclusion — that CPAP remains more effective than MADs at reducing respiratory events, while MADs still meaningfully outperform no treatment — is consistent with the broader clinical consensus and with the general direction (if not the precise size) shown across the individual pooled trials. The practical clinical framing the study's authors offer: CPAP remains the most effective therapy for reducing respiratory events, while MADs represent a reasonable alternative specifically for patients who are intolerant of or poorly adherent to CPAP, rather than a first-line replacement for it [s1].
What this doesn't establish
Eight trials and 642 total participants is a modest evidence base to draw firm conclusions from, particularly given the heterogeneity discussed above. This meta-analysis focuses on the apnea-hypopnea index as its primary outcome — a physiological severity measure — rather than on outcomes patients may care about equally or more, like adherence rates, daytime sleepiness improvement, or cardiovascular outcomes, where the adherence advantage MADs are believed to offer over CPAP might tell a different comparative story than the AHI numbers alone.
What to watch
Larger, better-standardized trials that could help explain the heterogeneity found here, and head-to-head comparisons that weigh adherence and quality-of-life outcomes alongside the raw apnea-hypopnea index. This article is not medical advice; choice between CPAP and oral appliance therapy should be made with a treating clinician.
Sources
Sources
More on
Most snoring is harmless. What quiets it depends on why you snore
The remedies that work are the ones matched to the cause — nose, mouth, tongue or palate. Here is what the trials support, and the sign that means see a doctor first.
Patients with insomnia and apnea saw a bigger blood-pressure drop from CPAP
After four weeks, the comorbid-insomnia group's blood pressure fell 20 points more than the apnea-only group. The finding flips the usual framing of COMISA as simply a worse-outcomes phenotype.
In heart patients with apnea but no sleepiness, CPAP still improved cardiac features
Nearly twice as many CPAP-treated patients saw their HFpEF-like markers improve at one year compared with those who went untreated — in a population sleep specialists have historically deprioritized for treatment.
On night one, fixed CPAP preserved sleep continuity better than the auto version
Both machine types cut breathing interruptions equally well in a 108-patient study. But auto-adjusting PAP's constantly shifting pressure came with more time awake and less total sleep on night one.