Innovations in Obstructive Sleep Apnea Management: From Mechanisms to Precision Therapies

From Mechanisms to Precision Therapies

Sleep Medicine · Seminar week 30 · released September 7, 2026 · includes a discussion video

The 2026 American Academy of Sleep Medicine (AASM) guidelines emphasize innovative approaches to OSA, such as the expansion of pharmacotherapeutics and technological…

Learning Objectives

• Differentiate obstructive sleep apnea phenotypes using symptoms, objective testing, anatomy, physiology, and comorbidity burden.
• Apply the AASM guidance documents that remain current in 2026 to diagnosis, PAP initiation, longitudinal testing, and referral.
• Distinguish the insomnia indication for dual orexin receptor antagonists from treatment of upper-airway obstruction.
• Select appropriate candidates for hypoglossal nerve stimulation and compare currently approved device architectures.
• Integrate dental sleep medicine, positional therapy, weight management, and surgery into individualized treatment pathways.
• Construct a longitudinal OSA plan that verifies efficacy objectively and addresses adherence, residual symptoms, and treatment harms.

Introduction to OSA and Current Challenges

%%FIG0%% Obstructive sleep apnea (OSA) consists of recurrent upper-airway narrowing or collapse during sleep despite continued respiratory effort. Reported prevalence varies with age, sex, obesity, sampling, and hypopnea definition. A frequently cited scoping review estimated that OSA affects 5–10% of the global population (Hirani and Smiley, PMID: 36836743); standardized modeling suggests a much larger burden—approximately 936 million adults aged 30–69 years with mild-to-severe OSA and 425 million with moderate-to-severe disease (global prevalence model, PMID: 31300334). Most disease remains undiagnosed.

The obstructive event begins when sleep-related withdrawal of pharyngeal dilator drive permits inspiratory negative pressure to narrow a vulnerable airway. Reduced airflow produces hypercapnia, intermittent hypoxemia, and increasing respiratory effort until arousal restores muscle tone and ventilation. Repetition generates sympathetic surges, oxidative stress, endothelial dysfunction, intrathoracic pressure swings, and sleep fragmentation. These mechanisms help explain associations with hypertension, atrial fibrillation, stroke, insulin resistance, impaired vigilance, mood symptoms, and motor-vehicle crashes. Association does not guarantee that every patient receives the same cardiovascular benefit from treatment; the SAVE trial, for example, did not demonstrate secondary cardiovascular event prevention with modest-duration CPAP use in a largely nonsleepy population (PMID: 27571048).

Teaching Point: AHI is a frequency count, not a complete biological phenotype. Conventional categories—mild 5.0–14.9, moderate 15.0–29.9, and severe at least 30 events/hour—should be interpreted alongside oxygen nadir, hypoxic burden, event duration, arousal intensity, REM and positional dependence, symptoms, and comorbid disease.

Mechanistically, OSA can be organized into four interacting traits: anatomical collapsibility, inadequate upper-airway muscle responsiveness, unstable ventilatory control or high “loop gain,” and a low respiratory arousal threshold. Obesity, tonsillar enlargement, fluid redistribution, nasal resistance, retrognathia, and reduced lung volume increase anatomical load. A patient with predominantly positional collapse may respond to positional therapy; one with mandibular retrusion may favor an oral appliance or maxillomandibular advancement; another with tongue-base collapse and preserved neuromuscular responsiveness may be suited to hypoglossal stimulation. Endotypes guide probability, not certainty.

Framework: Before labeling symptoms as OSA, separate obstruction from central sleep apnea, opioid-related breathing instability, obesity hypoventilation, COPD–OSA overlap, neuromuscular hypoventilation, insufficient sleep, circadian misalignment, depression, medication effects, restless legs, periodic limb movements, and narcolepsy. A home test reports a respiratory-event index based on recording time and can underestimate disease when wakefulness or insomnia is substantial. It also cannot reliably characterize sleep stages, arousals, hypoventilation, or many competing sleep disorders.

The practical treatment challenge is the gap between efficacy and effectiveness. PAP can eliminate obstruction while worn, but pressure intolerance, mask leak, nasal symptoms, aerophagia, claustrophobia, insomnia, cost, and fragmented care reduce use. The insurer convention of at least four hours on 70% of nights is an administrative threshold, not the therapeutic target; patients remain untreated during every unprotected period of sleep. Early mask fitting, humidification, nasal treatment, pressure review, behavioral support, and data-guided follow-up should precede a declaration of “CPAP failure.”

MUST ACT: Ask directly about drowsy driving, occupational hazards, sedatives, alcohol, opioids, and upcoming anesthesia. A dangerously sleepy patient should stop driving or performing safety-sensitive work until risk is controlled. During hospitalization or procedural care, communicate the OSA diagnosis and continue established PAP, oral-appliance, or neurostimulation therapy unless contraindicated.

Decision Point: Treatment intensity should reflect symptoms and physiological burden rather than AHI alone. AHI 12 with severe sleepiness, resistant hypertension, or substantial hypoxemia may demand more urgent treatment than AHI 32 composed of brief, minimally desaturating events in an asymptomatic patient—although both require informed longitudinal management.

Audience Poll: Which most often limits success in your practice: delayed diagnosis, PAP intolerance, fragmented referrals, or failure to verify alternative therapies?

Review of 2026 AASM Guidelines for OSA

%%FIG1%% An essential correction is that, as of July 2026, there is no unified “2026 AASM outpatient OSA guideline” introducing Class I and Class IIa recommendations. AASM uses GRADE terminology—strong or conditional—and its active practice-guideline catalog remains a set of question-specific documents. The operative framework combines the 2017 diagnostic guideline, 2019 PAP guideline, 2015 oral-appliance guideline, 2021 surgical-referral guideline, 2021 longitudinal-testing statement, and a new 2025 guideline restricted to medically hospitalized adults.

Diagnosis must be objective. Questionnaires such as STOP-BANG identify risk but do not establish OSA. In an uncomplicated adult at increased risk for moderate-to-severe disease, either polysomnography or a technically adequate home sleep apnea test is acceptable. If one home test is negative, inconclusive, or technically inadequate while suspicion persists, proceed to polysomnography rather than repeating home testing. Use polysomnography initially when there is significant cardiorespiratory disease, possible neuromuscular weakness, awake or suspected sleep-related hypoventilation, chronic opioid therapy, previous stroke, or severe insomnia (AASM diagnostic guideline, PMID: 28162150).

MUST ACT: A negative home test does not safely exclude OSA in a high-probability patient. Review raw signals, total recording time, position, oximetry quality, and whether the patient actually slept; then escalate to polysomnography when indicated.

For treatment, the 2019 guideline strongly recommends PAP for adults with excessive sleepiness and conditionally supports it for impaired sleep-related quality of life or comorbid hypertension. Uncomplicated patients may begin with home APAP or laboratory titration; CPAP or APAP is appropriate for routine ongoing therapy, while BPAP is not routinely superior. Education at initiation is strongly recommended, with early behavioral, troubleshooting, and telemonitoring interventions supported conditionally (AASM PAP guideline, PMID: 30736887). During manual titration, adult CPAP commonly starts at 4 cm H₂O and rises by at least 1 cm H₂O no more often than every five minutes. Consider BPAP when high CPAP is intolerable or obstructive events persist around 15 cm H₂O; uncomplicated comfort complaints alone do not automatically require bilevel therapy.

Decision Point: Home APAP is not a shortcut for complex physiology. Heart failure, chronic opioids, significant lung disease, non-OSA nocturnal hypoxemia, central apnea, and suspected hypoventilation require a more deliberate laboratory or specialist pathway.

Alternative-therapy referral is also explicit. Clinicians should discuss sleep-surgeon referral with PAP-intolerant or PAP-unaccepting adults with BMI below 40 kg/m² and bariatric-surgeon referral when BMI is at least 35 kg/m². Persistent pressure-related nonadherence may justify adjunctive surgical evaluation, although PAP is generally attempted before surgery even when a major anatomical abnormality is present (AASM surgical-referral guideline, PMID: 34351848). This guideline creates referral triggers; it does not endorse every operation or HNS device equally.

The newest OSA-specific document addresses hospitalized medical patients. It advises continuing established sleep-disordered-breathing treatment unless contraindicated and conditionally supports pathway-linked inpatient screening, PAP for newly diagnosed or untreated moderate-to-severe OSA, access to sleep consultation, and a concrete discharge plan (2025 inpatient guideline, PMID: 40838698). It does not cover perioperative patients or replace outpatient guidance.

Framework: Verify every non-PAP therapy objectively. Follow-up PSG or multichannel HSAT is appropriate after oral-appliance titration, surgery, meaningful weight change, or HNS optimization. Routine retesting is unnecessary in an asymptomatic, effectively treated PAP user, but recurrent symptoms, unexplained device data, new cardiovascular disease, or suspected residual hypoxemia should trigger reassessment (AASM longitudinal guidance, PMID: 33704050).

Nuance: “Precision care” is a useful clinical strategy, not a formal 2026 AASM recommendation grade. Shared decision-making should incorporate physiological phenotype, treatment efficacy, nightly use, adverse effects, access, and patient priorities.

Audience Poll: Which guideline step is most often missed locally: escalation after a negative HSAT, early PAP troubleshooting, surgical referral, or objective testing after non-PAP therapy?

Dual Orexin Receptor Antagonists: Mechanism and Application

%%FIG2%% Orexin-A and orexin-B are produced by lateral hypothalamic neurons that stabilize wakefulness through OX1 and OX2 receptors across the locus coeruleus, tuberomammillary nucleus, raphe, basal forebrain, and other arousal networks. Suvorexant, lemborexant, and daridorexant competitively antagonize both receptors, reducing wake drive without directly potentiating GABA-A signaling. That distinction may account for relatively reassuring respiratory-safety observations, but it does not make DORAs treatments for pharyngeal collapse.

Teaching Point: In comorbid insomnia and sleep apnea—COMISA—the DORA target is insomnia-related hyperarousal. It is not airway anatomy, loop gain, or tongue-muscle recruitment. None of the three agents is FDA-approved to treat OSA, and none should replace PAP, an oral appliance, weight-directed therapy, or surgery.

First determine what “insomnia” means. Repeated awakenings may reflect untreated obstruction, mask leak, excessive pressure, nocturia, restless legs, pain, depression, alcohol, stimulants, circadian delay, or insufficient sleep opportunity. Chronic insomnia requires adequate opportunity for sleep plus persistent initiation or maintenance difficulty and daytime impairment. A two-week sleep diary, Insomnia Severity Index, medication review, and PAP download are often more informative than simply prescribing a hypnotic.

CBT-I remains foundational. In one COMISA randomized trial, four CBT-I sessions before CPAP increased subsequent CPAP use by approximately 61 minutes/night and improved acceptance (PMID: 31403168). MATRICS found better insomnia outcomes with CBT-I plus PAP but no significant PAP-adherence advantage, illustrating that insomnia response and device use are related but distinct outcomes (PMID: 32170307). The actual 2026 AASM insomnia guideline conditionally favors CBT-I plus medication over medication alone but suggests against combination treatment over CBT-I alone; its evidence involved older hypnotics, not DORAs (PMID: 41975142).

Respiratory studies are small and short. In 26 adults with mild-to-moderate OSA, suvorexant 40 mg—twice the U.S. maximum—did not alter mean oxygen saturation, but after four nights increased placebo-corrected AHI by 2.66 events/hour, with wide individual variability (PMID: 26194728). Lemborexant 10 mg produced no significant AHI or oxygenation change after eight nights in mild OSA (PMID: 32187781) or in 33 adults with moderate-to-severe OSA (PMID: 37677076). Daridorexant 50 mg over five nights was similarly neutral at the group level in 28 adults with mild-to-moderate OSA (PMID: 33305817). These are respiratory-safety signals, not evidence of OSA efficacy, cardiovascular benefit, or improved PAP adherence.

All are Schedule IV drugs and contraindicated in narcolepsy. Food may delay onset. Agent-specific labeling should be checked for CYP3A interactions and hepatic impairment.

Decision Point: Consider a DORA only when clinically significant insomnia persists after OSA treatment has been optimized and CBT-I has been offered or is unavailable. Start with the lowest effective dose and reassess within one to two weeks using sleep diary or ISI, morning alertness, falls, mood, parasomnias, and PAP data.

MUST ACT: Avoid combining a DORA with alcohol, another hypnotic, or unnecessary opioids or benzodiazepines. Counsel about next-day impairment, falls, sleep paralysis, hallucinations, cataplexy-like weakness, worsening depression, and complex sleep behaviors; discontinue immediately after sleep-driving or another complex behavior. Seek specialist input before use in untreated severe hypoxemia, hypercapnia, advanced lung disease, or opioid-associated breathing disturbance.

Nuance: Fewer remembered awakenings do not prove that breathing improved. A patient may sleep longer while retaining the same or greater obstructive burden. Continue objective OSA therapy and reassess physiology when symptoms or risk warrant.

Audience Poll: In a patient with COMISA, do you currently treat insomnia before PAP, concurrently with PAP, or only after PAP has been optimized?

Technological Innovations: Hypoglossal Nerve Stimulation

%%FIG3%% Hypoglossal nerve stimulation (HNS) converts selected upper-airway dilator muscles into a programmable therapeutic target. Distal stimulation preferentially recruits medial hypoglossal branches supplying the genioglossus, producing tongue protrusion and airway enlargement. Proximal or bilateral systems recruit a broader pattern of tongue muscles. Benefit depends on electrode placement, airway anatomy, stimulation timing, neuromuscular recruitment, sleep position, and tolerability—not simply whether the implant activates.

The 2026 device landscape is broader than the original seminar suggested, and the correct third device name is aura6000, not “Ora 6000.”

The FDA Inspire expansion permits BMI up to 40, but response is not equivalent across BMI values and many payers retain narrower thresholds. Central plus mixed events exceeding 25% are contraindicating. Inspire requires drug-induced sleep endoscopy (DISE) to exclude complete concentric collapse (CCC) at the soft palate. The Genio approval and aura6000 approval do not use the same CCC labeling, but “no CCC contraindication” does not mean proven efficacy in CCC; pivotal populations and architectures differ.

Teaching Point: Device eligibility begins with a recent diagnostic study documenting severity, the proportion of central or mixed events, PAP experience, weight trajectory, and anatomy. DISE remains essential for Inspire and useful for understanding palatal, lateral-wall, tongue-base, and epiglottic collapse even when another label does not mandate it.

In the STAR pivotal cohort of 126 highly selected patients, median AHI decreased 68%, from 29.3 to 9.0 events/hour, and ODI decreased 70%, from 25.4 to 7.4 at 12 months; 66% met the conventional Sher response definition of at least 50% AHI reduction and final AHI below 20 (PMID: 24401051). ADHERE subsequently demonstrated real-world durability, with median AHI falling from 32.8 to 9.5 and approximately 5.6 hours/night of use; lower BMI predicted greater response (PMID: 31520484).

Newer systems add meaningful but noncomparative evidence. The Genio DREAM trial reported a 63.5% AHI response, 71.3% ODI response, and mean AHI reduction of 18.3 events/hour, with serious adverse events in 8.7% of participants (PMID: 40702817). In the randomized OSPREY trial, aura6000 active therapy produced Sher response in 58.2% versus 13.5% with delayed activation at month seven; no head-to-head trial establishes superiority over Inspire or Genio (PMID: 42008810).

Framework: A typical pathway is implantation, wound healing, activation at about one month, determination of sensation and functional thresholds, gradual home advancement over two to three months, then all-night efficacy testing and parameter optimization. Inspect residual REM and supine disease rather than relying on a single device-generated number. Excess amplitude can cause discomfort, arousals, tongue abrasion, or counterproductive retrusor recruitment.

MUST ACT: HNS is not “set and forget.” Monitor infection, pain, tongue weakness, swallowing or speech symptoms, lead or patch problems, sleep fragmentation, weight gain, and loss of efficacy. Persistent OSA may require reprogramming, positional therapy, weight treatment, oral appliance therapy, adjunctive surgery, or PAP.

Nuance: Objective respiratory improvement has not yet established cardiovascular-event prevention. In the sham-controlled CARDIOSA-12 trial, optimized HNS did not significantly reduce 24-hour systolic pressure or vascular outcomes versus sham (PMID: 38032624).

Audience Poll: Would device architecture, DISE pattern, MRI needs, external-component preference, or local follow-up expertise most influence your referral?

Dental Sleep Medicine Integration

%%FIG4%% Mandibular advancement devices move the mandible anteriorly, tensioning the palatoglossal and palatopharyngeal tissues, increasing retrolingual and lateral airway dimensions, and improving upper-airway stability. They are not generic mouthguards. The joint AASM/AADSM guideline recommends a custom, titratable appliance provided by a qualified dentist for adults who prefer an alternative or cannot tolerate CPAP, with dentist oversight, objective follow-up sleep testing, and periodic sleep-physician review (PMID: 26094920).

PAP generally produces a larger physiological reduction per treated night. A meta-analysis estimated mean AHI reductions of approximately 9.3 events/hour with mandibular advancement devices versus 25.4 with CPAP, although greater appliance use can narrow differences in functional outcomes for selected patients (PMID: 26163056). Favorable features may include mild-to-moderate disease, lower BMI, mandibular retrusion, and positional OSA, but no demographic or cephalometric rule reliably predicts response. Severe OSA is not an absolute prohibition when PAP is refused or intolerable; it requires candid discussion of lower response probability and mandatory verification.

Framework: The sleep clinician establishes the diagnosis and prescription. The dentist evaluates periodontal support, dentition, restorations, occlusion, TMJs, bruxism, nasal patency, manual dexterity, and maximum comfortable protrusion. After impressions or intraoral scans and a protrusive bite registration, therapy commonly begins around 50–60% of comfortable protrusive range with minimal necessary vertical opening. Advance in small, device-specific increments—often 0.25–0.5 mm every several nights or one to two weeks—according to comfort and clinical response.

MUST ACT: Do not titrate by disappearance of snoring alone. Snoring can resolve while silent residual hypopneas, REM obstruction, or hypoxemia persists. Once the position is clinically optimized, confirm efficacy with all-night PSG or multichannel HSAT and continue adjustment or change modality if residual OSA is important.

Early adverse effects include salivation, dry mouth, tooth discomfort, morning bite change, jaw-muscle pain, and TMJ symptoms. Long-term effects can include reduced overjet or overbite, tooth movement, occlusal change, periodontal stress, and restoration failure. Morning jaw exercises or a repositioner, appliance adjustment, and regular dental examination reduce morbidity; maximum protrusion is not automatically the optimal position.

Digital workflows combine intraoral scanning, recorded mandibular position, CAD design, validated milling or additive manufacturing, post-processing, fit verification, and an archived design for rapid replacement. They may improve reproducibility and reduce laboratory steps, but “3D printed” does not itself prove greater efficacy or adherence. Early one-visit studies are encouraging but remain limited (PMID: 41075053).

Nuance: PMID 32618748 supports virtual planning and 3D technology in orthognathic surgery, not 3D-printed oral appliances. Maxillomandibular advancement enlarges both retropalatal and retroglossal airway compartments and is an important option for severe, multilevel, or craniofacial OSA. Observational meta-analysis suggests approximately 80% mean AHI reduction, but neurosensory symptoms, malocclusion, hardware complications, and substantial recovery must be discussed (PMID: 26606321).

Decision Point: Combination therapy is often underused. A mandibular device may lower the PAP pressure needed for a patient with aerophagia, while positional therapy can address residual supine disease. Modern vibratory positional devices have better tolerability than older “tennis-ball” methods, although long-term effectiveness still depends on use (PMID: 32386700).

Audience Poll: Does your referral pathway specify who orders the post-titration sleep study and who acts on residual disease?

Case Studies and Real-World Application

%%FIG5%% A 55-year-old man with BMI 34.6 kg/m², resistant hypertension, gastroesophageal reflux, loud snoring, witnessed apneas, morning headaches, and an Epworth Sleepiness Scale score of 15 presents after nearly falling asleep while driving. He also reports taking 60–90 minutes to fall asleep and awakening four times nightly. A prior home study showed an REI of 31 events/hour and oxygen nadir of 82%, but contained little lateral sleep and could not determine whether awakenings were respiratory or insomnia-related.

MUST ACT: He should stop driving until sleepiness is controlled. Review alcohol, opioids, benzodiazepines, shift schedule, and upcoming procedures immediately. His serum bicarbonate is 25 mEq/L, making awake hypercapnia less likely, but persistent resting hypoxemia, bicarbonate elevation, or restrictive symptoms would require arterial or venous blood-gas assessment and evaluation for obesity hypoventilation.

Because the initial study cannot characterize sleep stages and the treatment decision may include surgery, polysomnography is obtained. It shows AHI 34, supine AHI 61, nonsupine AHI 15, REM AHI 48, central plus mixed events 2%, oxygen nadir 83%, and 6% of sleep time below 90%. Examination shows nasal inflammation, moderate retrognathia, adequate dentition, and no major tonsillar hypertrophy. The phenotype is therefore multicomponent: obesity-related anatomical load, mandibular restriction, strong positional dependence, and probable COMISA.

Framework: Use five steps: protect against immediate harm; validate physiology; rescue the most efficacious tolerable therapy; treat modifiable drivers; then verify the chosen alternative objectively.

He begins APAP 6–15 cm H₂O with a nasal interface, heated humidification, education, and remote review at two weeks. Downloaded data show residual AHI 4 while worn, but use averages 2.2 hours/night, unintentional leak is high, and pressure peaks coincide with aerophagia. Rhinitis is treated with saline and an intranasal corticosteroid; mask fit, expiratory pressure relief, and a narrower pressure range are tested. A supervised BPAP trial is attempted when pressure intolerance persists. Concurrent CBT-I addresses conditioned arousal, irregular time in bed, and mask-related anxiety. Despite these measures, he remains unable to use PAP for most sleep, establishing intolerance rather than merely documenting initial refusal.

Decision Point: At this stage, offer parallel—not mutually exclusive—pathways. A qualified dentist can provide a custom titratable mandibular advancement device because his dentition and retrognathia are favorable. Positional therapy directly targets the supine component. Intensive weight treatment is indicated. Tirzepatide may be considered because he has obesity and moderate-to-severe OSA, after screening for medullary thyroid carcinoma or MEN2 history, pancreatitis risk, gallbladder disease, and relevant drug interactions.

The appliance is titrated gradually. Follow-up HSAT at the optimized setting shows residual AHI 23, with persistent REM and nonpositional obstruction. This is clinically useful but inadequate monotherapy. He then undergoes sleep-surgery consultation. DISE demonstrates anterior-posterior palatal and tongue-base collapse without complete concentric palatal collapse. Central events remain below 25%, and his BMI meets FDA labeling, although payer criteria are verified separately. After shared decision-making, he selects established respiration-synchronized HNS rather than MMA or a newer system with less local follow-up experience.

The implant is activated one month postoperatively and advanced gradually. At three months, discomfort and persistent supine events are addressed through electrode reprogramming and positional therapy rather than indiscriminate amplitude escalation. An efficacy study at six months shows AHI 8, oxygen nadir 90%, and no significant sustained hypoxemia; ESS improves to 6.

His insomnia improves with CBT-I but does not fully remit. A DORA can now be considered specifically for insomnia—for example, lemborexant 5 mg or daridorexant 25 mg at bedtime with at least seven hours available—after counseling about next-day impairment and complex sleep behavior. It is not credited with treating his airway. If used, reassess within one to two weeks and taper when no longer needed.

Nuance: A different result would redirect the plan. Central or mixed events above 25%, hypercapnia, predominant CCC, uncontrolled psychiatric disease, or inability to operate the device could make this HNS pathway inappropriate. Major weight loss would also prompt repeat testing before assuming that every therapy remains necessary.

Teaching Point: Success is not “the patient received an implant.” Success is tolerable all-night use, verified control of obstruction and hypoxemia, improved function, controlled comorbid insomnia, and a plan for weight and cardiovascular risk.

Audience Poll: At the first PAP-intolerance visit, which branch would you activate first: structured PAP rescue, dental referral, weight pharmacotherapy, CBT-I, or surgical evaluation?

Discussion and Future Directions in OSA Management

%%FIG6%% The near-term future of OSA management is combination care organized around mechanism. AHI-only prescribing treats biologically different patients as interchangeable. Emerging approaches incorporate hypoxic burden, event duration, arousal intensity, REM and positional dependence, ventilatory instability, craniofacial anatomy, tongue-fat distribution, cardiometabolic risk, and patient preference. These variables may eventually predict which patient needs airway splinting, weight reduction, mandibular advancement, neuromodulation, pharmacologic endotype modification, or a combination.

The most practice-changing pharmacological development is not a DORA but tirzepatide. The FDA approved it in December 2024, with reduced-calorie diet and increased physical activity, for moderate-to-severe OSA in adults with obesity. In the two SURMOUNT-OSA phase 3 trials, mean AHI fell by 25.3 events/hour versus 5.3 with placebo in participants not using PAP and by 29.3 versus 5.5 among participants assigned to continue PAP (PMID: 38912654). Benefits also included reduced weight, hypoxic burden, systolic pressure, and inflammatory markers.

Framework: Tirzepatide begins at 2.5 mg subcutaneously weekly for four weeks, then 5 mg; increase by 2.5 mg no more often than every four weeks toward the OSA maintenance dose of 10 or 15 mg weekly as tolerated. Screen for personal or family history of medullary thyroid carcinoma or MEN2 and counsel regarding gastrointestinal effects, gallbladder disease, pancreatitis, dehydration-related kidney injury, hypoglycemia with insulin or sulfonylureas, pregnancy, and delayed gastric emptying around anesthesia.

MUST ACT: Do not stop PAP or another effective treatment when weight medication begins. Improvement develops over months, weight may recur after discontinuation, and non-adiposity mechanisms may persist. Repeat PSG or HSAT after a clinically significant weight change and change airway therapy only after objective reassessment.

Other drug strategies—noradrenergic–antimuscarinic combinations, potassium-channel modulation, carbonic anhydrase inhibition for selected high-loop-gain physiology, and agents targeting upper-airway muscle responsiveness—remain investigational for routine OSA care. Their future depends on matching drugs to endotypes without worsening blood pressure, urinary retention, arrhythmia, sleep quality, or respiratory-event duration.

Technology will increasingly support longitudinal care. PAP telemonitoring can identify leak, residual events, treatment-emergent central apnea, and declining use before the next clinic visit. Wearables and contactless sensors may detect risk or track trends, but consumer-derived oxygen or “sleep apnea” scores should not independently diagnose disease or certify treatment success. Automated scoring must be validated across sex, race, age, body habitus, comorbidity, and device type; algorithms trained on narrow laboratory populations can amplify diagnostic inequity.

Decision Point: Use AI to prioritize review, not to eliminate accountable clinical interpretation. A plausible automated AHI paired with implausible oximetry, poor signal quality, or a symptomatic patient still requires raw-data review and potentially formal testing.

HNS development is moving toward bilateral recruitment, proximal multicontact stimulation, simpler implantation, rechargeable systems, remote programming, and fewer MRI restrictions. Dental workflows will continue to combine digital impressions, reproducible manufacturing, adherence sensors, and remotely guided titration. The important evidence gap is not another uncontrolled pre/post series; it is comparative effectiveness among PAP, oral appliances, obesity treatment, HNS architectures, MMA, and rational combinations.

Future trials should measure all-night use, hypoxic burden, sleepiness, cognition, driving safety, blood pressure, atrial fibrillation burden, cardiovascular events, treatment burden, cost, and durability—not merely whether AHI falls below 20. CARDIOSA-12 reminds clinicians that an impressive respiratory response does not automatically establish vascular benefit. Studies must also include women, older adults, racial and ethnic groups underrepresented in device trials, patients with higher BMI, and people with overlapping insomnia, lung disease, and cardiovascular disease.

Teaching Point: Precision therapy is usually sequencing rather than a single perfect intervention: stabilize immediate risk, choose the most effective acceptable treatment, remove adherence barriers, target obesity and anatomy, add phenotype-specific therapy, and verify the combined result.

Finally, delivery systems matter. Stepped-care and hub-and-spoke models can allow primary care and trained nonspecialists to manage uncomplicated testing and PAP while sleep centers concentrate on diagnostic uncertainty, hypoventilation, refractory symptoms, and multimodal therapy (PMID: 31639334). Precision without access is not precision care.

Audience Poll: Which advance is most likely to change your practice first: obesity pharmacotherapy, expanded HNS options, digital dental workflows, physiological endotyping, or AI-supported longitudinal monitoring?

Tonight on Shift

• Identify immediate danger: ask every high-risk patient about drowsy driving, opioids, alcohol, sedatives, and upcoming anesthesia.
• Match the test to the patient: use PSG for complex disease and escalate a negative or inadequate HSAT when suspicion remains.
• Rescue PAP before declaring failure by reviewing interface, leak, nasal symptoms, pressure, aerophagia, insomnia, education, and device data.
• Treat COMISA deliberately: offer CBT-I first and use a DORA only for persistent insomnia—not as therapy for airway obstruction.
• Route PAP-intolerant patients by phenotype to weight treatment, positional therapy, a qualified dental provider, sleep surgery, or appropriately selected HNS.
• Close the loop with objective PSG or multichannel HSAT after every non-PAP intervention, meaningful weight change, or recurrence of symptoms.

Read this seminar as Markdown · All seminars · Lecture library · Question bank