The apnea-hypopnea index, or AHI, counts the average number of times per hour your breathing partially or fully stops while you sleep. An AHI below 5 is considered normal, 5 to 14 is mild obstructive sleep apnea, 15 to 30 is moderate, and anything above 30 is severe.1PubMed Central. The STOP-Bang equivalent model and prediction of severity of obstructive sleep apnea: relation to polysomnographic measurements of the apnea/hypopnea index Those cutoffs have guided treatment decisions for decades, but the number is more slippery than it looks, and sleep researchers increasingly argue it misses important parts of what makes sleep apnea dangerous.
What the Number Actually Captures
An “apnea” is a complete pause in airflow lasting at least ten seconds. A “hypopnea” is a partial reduction in airflow, typically by at least 30%, also lasting at least ten seconds and accompanied by either a drop in blood oxygen or a brief arousal from sleep. The AHI adds up every apnea and hypopnea over the course of a night and divides by total sleep hours. If you stop breathing or nearly stop breathing 120 times during an eight-hour study, your AHI is 15.
The gold-standard measurement comes from an in-lab polysomnography study, where sensors track airflow through your nose and mouth, chest and abdomen movement, blood oxygen levels, brain waves, and more. Home sleep tests, which are now common for initial diagnosis, use fewer channels but still produce an AHI (sometimes called a “respiratory event index” instead). Studies comparing portable devices to full lab studies have found strong correlations, with combined correlation coefficients around 0.89 for AHI between peripheral arterial tonometry devices and lab polysomnography.2JAMA Otolaryngology–Head & Neck Surgery. Diagnosis of Obstructive Sleep Apnea by Peripheral Arterial Tonometry: Meta-analysis Home tests can underestimate severity, though, partly because they measure total recording time rather than actual sleep time, diluting the count if you spent part of the night awake.
Why Severity Categories Matter for Health
The mild/moderate/severe labels are not just academic bookkeeping. Higher AHI tracks with worse health outcomes, although the relationship is not always a clean straight line.
Cardiovascular risk climbs with severity. Large prospective studies have linked obstructive sleep apnea to a higher incidence of hypertension, coronary disease, stroke, and heart failure.3PubMed Central. Role of Obstructive Sleep Apnea in Cardiovascular Disease Even within the “mild” category, there are meaningful differences. One study of over 12,000 patients found that people with a mild AHI between 11 and 15 had roughly 80% higher odds of having hypertension compared to nonapneic snorers, and that risk was significantly greater than in those with AHI between 5 and 11.4PubMed Central. Mild obstructive sleep apnea increases hypertension risk, challenging traditional severity classification That finding challenges the common assumption that mild sleep apnea is harmless across the board.
Metabolic consequences show up as well. Sleep apnea is independently tied to insulin resistance even after accounting for obesity, with each additional apnea or hypopnea per hour increasing fasting insulin and insulin resistance by about half a percent.5PubMed. Obstructive sleep apnea is independently associated with insulin resistance That might sound tiny per event, but someone with an AHI of 40 accumulates a meaningful metabolic hit night after night. This association holds in both obese and non-obese individuals, meaning you cannot write off the metabolic risk as purely a weight problem.6PubMed Central. The Correlation between the Severity of Obstructive Sleep Apnea and Insulin Resistance in a Japanese Population
Cognitive effects are also dose-dependent. A pooled analysis found that even mild-to-severe sleep apnea was associated with slightly lower global cognition scores compared to people with an AHI below 5.7JAMA Network Open. Sleep Architecture, Obstructive Sleep Apnea, and Cognitive Function in Adults The steepest cognitive decline over time appears in older adults, men, and carriers of the ApoE4 gene variant associated with Alzheimer’s risk.8European Respiratory Journal. Obstructive sleep apnoea and 5-year cognitive decline in the elderly At the severe end, AHI above 30 combined with poor sleep quality and excessive daytime sleepiness significantly raises the risk of mild cognitive impairment.9PubMed Central. Relationship between sleep apnea severity and mild cognitive impairment in people with obstructive sleep apnea
Mortality risk rises too, though the relationship gets muddied once you adjust for other health factors. In one decade-long cohort study, AHI as a continuous variable was significantly associated with all-cause mortality and cardiovascular events, but the effect weakened after controlling for traditional risk factors like high blood pressure, diabetes, and cholesterol.10PLOS Medicine. Obstructive Sleep Apnea and Risk of Cardiovascular Events and All-Cause Mortality: A Decade-Long Historical Cohort Study A separate analysis found that severe sleep apnea carried a 38% higher adjusted mortality risk compared to no sleep apnea.11PubMed Central. Long-term All-Cause Mortality Risk in Obstructive Sleep Apnea Using Hypopneas Defined by a ≥3 Percent Oxygen Desaturation or Arousal
Your AHI Can Change from Night to Night
One of the more frustrating aspects of AHI is that it is not a fixed number. Sleep position, sleep stage, alcohol consumption, nasal congestion, and even how much fluid has shifted into your neck tissues can all alter the count on any given night.
Sleeping on your back reliably increases the number of breathing events. In patients with severe apnea, AHI was higher in the supine position than on the sides, though in very obese patients the count remained clinically high regardless of position.12Sleep. Sleep Apnea and Body Position during Sleep For less obese individuals, position matters more, and a subset of patients have what clinicians call “positional” sleep apnea, where their AHI may be normal or near-normal on their side but elevated on their back.13Annals of the American Thoracic Society. Night-to-Night Repeatability of Supine-related Obstructive Sleep Apnea
REM sleep, the stage associated with dreaming, also produces longer apneas and can push the index higher compared to non-REM sleep. If your sleep study happened to capture more or less REM than a typical night, your AHI will shift accordingly. Overnight changes in leg fluid volume, which can redistribute into the neck when you lie down, have been correlated with night-to-night AHI variability during non-REM supine sleep.14PubMed Central. Night-to-night Variability in Obstructive Sleep Apnea Severity: Relationship to Overnight Rostral Fluid Shift The practical takeaway is that a single night’s AHI is a snapshot, not a permanent measurement. Someone near a severity boundary could easily be classified differently on a repeat study.
How Scoring Rules Shift the Numbers
Your AHI depends not only on your breathing but on which rulebook the technologist uses to score the events. This is a bigger deal than most patients realize. The American Academy of Sleep Medicine has revised its hypopnea definition several times, and each revision substantially changes how many events get counted.
The older “conservative” rule required a 4% or greater drop in blood oxygen to score a hypopnea. A newer “liberal” rule counts any event with at least a 3% oxygen drop or an arousal from sleep. When researchers applied both rules to the same population, the prevalence of moderate-to-severe sleep apnea jumped from 22% under the conservative rule to 45% under the liberal one, and the share of people with a completely normal study fell from 48% to 17%.15Sleep. Sleep Apnea Severity Classification — Revisited A population-based study showed OSA prevalence ranging from about 47% to 75% depending on which of three scoring criteria was applied.16PubMed Central. Effect of Three Hypopnea Scoring Criteria on OSA Prevalence and Associated Comorbidities in the General Population
What this means in practice: if you had a sleep study in 2005 and another in 2020, the raw numbers might not be directly comparable even if your actual breathing pattern stayed the same. The AHI from an older study using the stricter rule could be roughly 30% to 60% of what the same night would produce under newer criteria.17SLEEP. The New Aasm Criteria for Scoring Hypopneas: Impact on the Apnea Hypopnea Index If you are comparing results over time or across different labs, ask which scoring rule was used.
Why Researchers Think AHI Is Not Enough
The AHI treats all breathing events as equal. A ten-second pause that barely dips your oxygen is counted the same as a forty-second obstruction that drops your blood oxygen to dangerously low levels. It also ignores the recovery pattern, the degree of arousal from sleep, and whether the events cluster in certain sleep stages. This has led to a growing consensus that AHI, while useful as a screening and classification tool, is a poor predictor of who actually gets sick.
Research on daytime sleepiness, one of the most common symptoms of sleep apnea, illustrates the problem. A study comparing AHI to measures of oxygen desaturation severity found that the depth and frequency of oxygen drops were significantly better predictors of objective daytime sleepiness than AHI alone. A 10% increase in desaturation severity roughly doubled the risk of severe daytime sleepiness, compared to a more modest increase from the same jump in AHI.18PubMed Central. Severity of Desaturations Reflects OSA-Related Daytime Sleepiness Better Than AHI
For cardiovascular outcomes, a metric called “hypoxic burden” has emerged as a more informative alternative. Instead of just counting events, hypoxic burden captures both the depth and duration of oxygen drops associated with each breathing event, essentially measuring the total area under the desaturation curve. In two large community-based cohorts, the top fifth of hypoxic burden carried roughly twice the risk of cardiovascular death compared to the bottom fifth, after adjusting for traditional risk factors. AHI, by contrast, was not significantly predictive of cardiovascular death in those same cohorts.19SLEEP. Metrics of sleep apnea severity: beyond the apnea-hypopnea index Other commonly used severity metrics like the oxygen desaturation index, which also just counts events without weighing their severity, similarly failed to predict major adverse cardiovascular events in clinical data.20American Journal of Respiratory and Critical Care Medicine. Sleep Apnea–Specific Hypoxic Burden and Not the Sleepy Phenotype as a Novel Measure of Cardiovascular and Mortality Risk in a Clinical Cohort
The AHI’s prognostic weakness is now openly acknowledged in the sleep medicine literature, where it is described as poorly correlated with adverse outcomes and having limited ability to predict who will develop complications.21PubMed Central. Hypoxic burden – definitions, pathophysiological concepts, methods of evaluation, and clinical relevance Despite that, AHI remains the standard because it is easy to calculate, universally understood, and embedded in insurance reimbursement criteria. Hypoxic burden and other proposed replacements have not yet been adopted into routine clinical scoring, though that may change as evidence accumulates.
What AHI Looks Like During Treatment
Once you start using CPAP or another treatment, the goal is to bring your AHI down, typically below 5. Most CPAP machines report an estimated AHI each night based on airflow patterns, which is useful for tracking trends but not perfectly precise. When researchers compared the automated AHI from CPAP machines to manually scored data from the same nights, the machines consistently underestimated the count. About a third of patients had a residual AHI above 5 by the machine’s count, but when a human scored the same flow data, over 60% did.22PubMed Central. Residual Events during Use of CPAP: Prevalence, Predictors, and Detection Accuracy
Not everyone responds fully to CPAP. About a quarter of patients in one study still had an AHI above 10 after three months of use, and those patients had worse outcomes and lower compliance with the machine.23PubMed. Residual sleep apnea on polysomnography after 3 months of CPAP therapy: clinical implications, predictors and patterns One predictor of persistent residual events on CPAP is the emergence of central apneas during treatment, where the brain temporarily stops sending breathing signals. A high treatment-phase central apnea index predicted ongoing residual disease and was also linked to patients dropping out of therapy entirely.24PubMed. Predictors and consequences of residual apnea during positive airway pressure therapy If your CPAP-reported AHI looks good but you still feel terrible, or if it is not coming down as expected, a follow-up study or machine data review is worth requesting.
For people who cannot tolerate CPAP, hypoglossal nerve stimulation has become an increasingly studied alternative. This surgically implanted device stimulates the nerve controlling tongue movement, keeping the airway open during sleep. Studies have shown significant reductions in AHI and daytime sleepiness with hypoglossal nerve stimulation.25PubMed Central. Hypoglossal Nerve Stimulation Therapy for the Treatment of Obstructive Sleep Apnea A meta-analysis comparing it to other airway surgeries found that hypoglossal nerve stimulation achieved a lower post-treatment AHI, with significantly higher rates of getting the AHI below 10 and below 15.26PubMed. Comparative effectiveness of hypoglossal nerve stimulation and alternative treatments for obstructive sleep apnea: a systematic review and meta-analysis It is currently approved for selected patients with moderate-to-severe obstructive sleep apnea who have failed CPAP.
Obstructive Versus Central Events in the AHI
The AHI is sometimes reported as a single number, but it actually bundles together two distinct types of breathing disruption. In obstructive events, your airway physically collapses or narrows. In central events, the brain simply fails to send the signal to breathe. The treatment implications are very different: CPAP works by splinting the airway open, which directly addresses obstruction but does nothing for a brain-signaling problem.
How common central events are depends heavily on how you define them. When central apneas are counted strictly (ignoring whether hypopneas are central or obstructive), central sleep apnea shows up in about 5% of patients. But when scoring distinguishes between central and obstructive hypopneas, the prevalence of central sleep apnea jumps to roughly 20%.27PubMed. Multidimensional phenotyping to distinguish among central (CSA), obstructive (OSA) and co-existing central and obstructive sleep apnea (CSA-OSA) phenotypes in real-world data A meaningful share of patients, around 11 to 19% depending on the definition, have both types simultaneously. Newer wearable devices are getting better at distinguishing the two. One validated system using peripheral arterial tonometry and respiratory movement analysis achieved 95% agreement with lab polysomnography in differentiating central from obstructive events.28PubMed Central. Detecting central sleep apnea in adult patients using WatchPAT-a multicenter validation study
Sex Differences and Underdiagnosis in Women
Sleep apnea is often framed as a condition affecting overweight middle-aged men, and while that group is heavily represented, the framing has real consequences for women. Women with sleep apnea tend to present differently. They report more insomnia, fatigue, and mood symptoms, and their breathing events are more often partial obstructions (hypopneas) rather than full apneas. Their AHI tends to be lower than men’s at similar levels of symptoms and impairment.29European Respiratory Review. Sex differences in obstructive sleep apnoea
Because the diagnostic system is anchored to the AHI, and the severity thresholds were largely established using male-predominant study populations, women can have a “mild” AHI on paper while experiencing substantial disruption to their health and daily functioning. Multiple studies have concluded that women with obstructive sleep apnea are consistently underdiagnosed and undertreated, partly because of these clinical differences and partly because sociocultural factors make women less likely to be referred for a sleep study in the first place.30PubMed. Gender differences in clinical and polysomnographic features of obstructive sleep apnea: a clinical study of 2827 patients31PubMed. Gender differences in obstructive sleep apnea syndrome: a clinical study of 1166 patients
How AHI Thresholds Apply in Children and Adolescents
The adult AHI thresholds do not transfer to pediatric patients. In children, an AHI above 1 is generally considered abnormal, and anything above 5 can be clinically significant. The breathing events themselves are scored differently too: pediatric rules count shorter apneas (lasting at least two missed breaths rather than a strict ten seconds) and use different oxygen desaturation thresholds.
Adolescents fall into a gray zone. The 2007 American Academy of Sleep Medicine scoring manual did not specify whether to use pediatric or adult rules for teenagers. When researchers compared the two approaches in adolescents, they found that using the stricter adult scoring rule (requiring a 4% oxygen desaturation for hypopneas) led to significantly fewer diagnoses compared to pediatric criteria, while the more liberal adult rule (3% or arousal) produced results similar to the pediatric rules with few patients changing diagnostic category.32SLEEP. Differences in Overnight Polysomnography Scores Using the Adult and Pediatric Criteria for Respiratory Events in Adolescents For parents of teenagers undergoing sleep studies, it is worth asking the interpreting physician which scoring criteria were applied, because the answer can determine whether your child gets diagnosed at all.