AHI Sleep Apnea Scores: Built to Diagnose, Not to Manage

The apnea-hypopnea index (AHI) was built to solve a diagnostic problem: how to turn one night of physiologic data into a number clinicians could compare across patients and centers. It was never designed to track a chronic, decades-long disease. Obstructive sleep apnea changes with weight, aging, medications, and comorbid disease, but AHI only tells you what was true on the night it was measured. For dentists, that distinction matters because oral appliance therapy does not end at delivery. It requires the same kind of ongoing verification that AHI was built to provide once, not the recurring reassessment a chronic disease actually needs. Night-to-night variability is well documented and can affect both diagnosis and treatment decisions.¹ The AHI, while a useful metric, does not account for how an individual patient varies over time.

Ask most dentists in dental sleep medicine what AHI is, and they can define it. Ask what problem it was originally built to solve, and the answer gets fuzzier. That gap matters more than it sounds like it should.

Robyn Woidtke has spent decades inside that history. She was one of the first 500 people to earn the RPSGT credential, and she is careful about how she frames her own authority on the subject. She does not describe herself as a diagnostics expert. She describes herself as someone who has watched sleep medicine’s diagnostic tools evolve from the inside, one development effort at a time.

That vantage point produces a different question than the one most AHI explainers ask. Instead of “what does AHI miss on a given night,” Woidtke’s perspective raises a harder one: AHI was built to answer a one-time diagnostic question, so why are dental and medical practices still using it as if it can also answer an ongoing chronic care management question, year after year, in a disease that keeps changing?

Sleep was not always measurable. Before the 1950s, sleep disorders were understood mostly through patient report and behavioral observation, which meant they were inconsistently diagnosed and often dismissed outright.

That changed in 1953, when Eugene Aserinsky and Nathaniel Kleitman at the University of Chicago published their discovery of rapid eye movement sleep using EEG and EOG recordings.² For the first time, sleep could be shown to be a structured, measurable physiologic process rather than a uniform state of rest. Electromyography was added soon after, and by 1968 the field had its first widely accepted scoring system for sleep stages, published by Rechtschaffen and Kales.³ Sleep could now be classified and compared across laboratories, not just described.

The next inflection point came in the 1970s. Researchers at Stanford, led by William Dement and Christian Guilleminault, identified a recurring pattern of airflow cessation during sleep accompanied by oxygen decline and repeated arousals. Their 1976 paper describing the sleep apnea syndromes gave the field a defined, measurable disease for the first time.⁴ Polysomnography expanded to include airflow sensors, respiratory effort belts, and pulse oximetry to capture it.

As Woidtke has emphasized, the goal of all of this was never simply to collect data. It was to produce decision-grade information: results that could be compared across patients, taught to new clinicians, and defended to payers. Standardization was a requirement and a safety mechanism, not an academic exercise. That is what eventually produced the current AASM scoring manual, revised again as recently as 2017 to keep event definitions and hypopnea criteria consistent across sleep labs.⁵

AHI emerged out of that standardization effort as a summary number. It became the industry standard because it is simple, quantifiable, and easy to sort into apnea type and severity categories that insurers, researchers, and clinicians could all agree on.⁶ But that simplicity came from what the metric was designed to do: classify a single measured night, not track a condition across years.

The longer story of how Sleep became measurable.

What a sleep technologist sees that the report doesn’t show.

This part tends to surprise dentists newer to the field. The limitations of AHI as a standalone number are not a recent critique. Woidtke points out that even early investigators recognized that two patients with an identical AHI could have very different oxygen burdens, arousal profiles, and clinical risk. That distinction was present from the beginning of the field, not discovered decades later.

Woidtke has frequently discussed the limits of using the AHI as the sole indicator of therapeutic success. Event counts are shaped by scoring rules and do not always reflect the physiologic stress or functional impairment a patient actually experiences.⁶ This is where phenotyping and endotyping also come in, characterizing the mechanisms driving an individual patient’s disease rather than only counting its events.⁷ One of the more significant recent developments in sleep medicine focuses on hypoxic burden, a measure of the cumulative depth and duration of oxygen desaturation across the night, which has been shown to predict cardiovascular mortality more strongly than AHI alone.⁸ The number sleep medicine settled on was never claimed to be the whole physiologic picture, even by the people who helped build it.

Here is where Woidtke’s perspective moves past a history lesson and into something dentists can act on.

She uses a comparison that lands quickly with clinicians because it reframes the problem as a systems issue, not a device issue. When a patient is diagnosed with cancer, they are typically placed into a defined surveillance pathway: scheduled imaging, clear escalation triggers, a team that owns follow-up. When a patient is diagnosed with diabetes, many health systems connect them to structured education and a Certified Diabetes Care and Education Specialist to guide long-term self-management.

Obstructive sleep apnea, despite carrying real cardiovascular, metabolic, and neurocognitive consequences, is too often handled as a one-and-done event. A diagnosis is made, a device is delivered or a mask is fitted, and the patient is left to navigate a chronic condition largely on their own, frequently unsure whether the next call belongs to the DME provider, the sleep physician, or their dentist.

Woidtke is careful not to frame this as a limitation of sleep medicine itself. She points instead to the practical forces driving the gap: reimbursement constraints, limited access to trained sleep specialists, and a healthcare system that remains more fragmented than cohesive.⁹ The most persistent breakdown she has observed across decades in the field is not missed diagnosis. It is the failure to sustain care after diagnosis.

That breakdown is not really an AHI problem. AHI did exactly what it was built to do at the moment of diagnosis. The problem is asking a one-time snapshot metric to carry the weight of an ongoing management plan it was never designed to support.

This is the part of the history that Woidtke returns to most often when she talks with dentists specifically, and it is the reason this gap is not purely a cautionary tale.

Dental practices already operate on a recall model. Patients come back routinely for dental care, many every six months, which creates an ideal opportunity for follow-on care. That structure was largely built for restorative and periodontal maintenance, but the same standard applies equally well to dental sleep practices, and it provides a needed opportunity for ongoing communication with the sleep medicine practitioners managing the patient.

What has historically been missing is the consistent integration of objective verification into that existing structure, not the structure itself. A dentist who schedules recall visits already has more built-in contact with an OSA patient over the following years than most physician follow-up pathways provide. Woidtke’s argument is that this contact should not be limited to checking whether the appliance still fits. It can include investigating how the patient is doing, how they are feeling, and how they are managing their care.

For dentists in dental sleep medicine, closing that continuity gap looks like a few concrete habits rather than a philosophical shift:

• Post-titration testing that confirms the appliance produced a measurable change in respiratory events, not just a comfortable night’s sleep.¹⁰
• Periodic reassessment built into existing recall intervals, since weight change, aging, medication changes, and progression of comorbid disease can all shift a patient’s disease severity years after the original diagnosis.
• Validated patient-reported outcome tools, such as PROMIS measures, collected at recall visits and shared back with the prescribing physician, so both sides of the care team are working from the same picture rather than two disconnected snapshots.
• Explicit handoffs with physician partners so it is clear, in writing, who owns the next step if a patient’s symptoms return or their appliance stops working as well as it once did.

None of that requires stepping outside a dentist’s scope. It requires recognizing that the recall structure already in place is, functionally, the longitudinal surveillance pathway that sleep medicine has lacked since AHI became its standard metric.

AHI is not a flawed metric. It did what it was built to do: turn a single night of physiologic data into something comparable, teachable, and defensible to payers. That achievement is why obstructive sleep apnea became a treatable, insurable, standardized diagnosis in the first place.

The mistake is asking that same number to also function as a management tool for a disease that keeps changing years after the appliance is delivered. Those are two different jobs. Sleep medicine built AHI to do the first one. Nobody built a comparable tool to do the second, which is exactly the gap Woidtke points to when she talks about dentistry’s opportunity.

For dentists, the practical takeaway is not to distrust the AHI on a diagnostic report. It is to stop treating that number as the end of the conversation with a patient, and start treating the recall visit that follows it as part of the same diagnostic tradition AHI came from: measurement over assumption, repeated over time rather than checked once.

1. Montero A, Farook TH, Lechat B, et al. Mechanisms of night-to-night variability in obstructive sleep apnoea. Sleep Med Rev. Published online August 26, 2026. doi:10.1016/j.smrv.2026.102361
2. Aserinsky E, Kleitman N. Regularly occurring periods of eye motility, and concomitant phenomena, during sleep. Science. 1953;118(3062):273-274. doi:10.1126/science.118.3062.273
3. Rechtschaffen A, Kales A, eds. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. US Department of Health, Education, and Welfare; 1968.
4. Guilleminault C, Tilkian A, Dement WC. The sleep apnea syndromes. Annu Rev Med. 1976;27:465-484. doi:10.1146/annurev.me.27.020176.002341
5. Berry RB, Brooks R, Gamaldo CE, et al. AASM scoring manual updates for 2017 (version 2.4). J Clin Sleep Med. 2017;13(5):665-666. doi:10.5664/jcsm.6576
6. Kapur VK, Auckley DH, Chowdhuri S, et al. Clinical practice guideline for diagnostic testing for adult obstructive sleep apnea. J Clin Sleep Med. 2017;13(3):479-504. doi:10.5664/jcsm.6506
7. Malhotra A, Mesarwi O, Pepin JL, Owens RL. Endotypes and phenotypes in obstructive sleep apnea. Curr Opin Pulm Med. 2020;26(6):609-614. doi:10.1097/MCP.0000000000000724
8. Azarbarzin A, Sands SA, Stone KL, et al. The hypoxic burden of sleep apnoea predicts cardiovascular disease-related mortality: the Osteoporotic Fractures in Men Study and the Sleep Heart Health Study. Eur Heart J. 2019;40(14):1149-1157a. doi:10.1093/eurheartj/ehy624
9. Weaver TE, Grunstein RR. Adherence to continuous positive airway pressure therapy: the challenge to effective treatment. Proc Am Thorac Soc. 2008;5(2):173-178. doi:10.1513/pats.200708-119MG
10. Ramar K, Dort LC, Katz SG, et al. Clinical practice guideline for the treatment of obstructive sleep apnea and snoring with oral appliance therapy: an update for 2015. J Clin Sleep Med. 2015;11(7):773-827. doi:10.5664/jcsm.4858