The apnea-hypopnea index (AHI) is the average number of apneas and hypopneas you have per hour of sleep. It is an important part of a sleep study, but it is not a complete diagnosis, a stand-alone measure of health risk, or a grade of how unwell you should feel.
When you read a sleep report, first check whether the number is actually an AHI. A home sleep apnea test may report a respiratory event index (REI), and a positive airway pressure (PAP) machine may display its own estimate. These numbers are calculated differently and are not interchangeable.
How AHI is calculated
During an overnight laboratory sleep study, or polysomnogram, sensors record airflow, breathing effort, oxygen levels, brain activity, and other signals. Brain activity allows the study to identify when you are asleep.
The calculation is:
AHI = total apneas and hypopneas ÷ hours of sleep
If a study records 48 events during 6 hours of confirmed sleep, the AHI is 8 events per hour. This is an average. It does not mean an event happened once every 7.5 minutes in an even pattern. Events may cluster during rapid eye movement (REM) sleep or while you are on your back.
What counts as an apnea?
In an adult sleep study, an apnea is scored when airflow falls by at least 90% from baseline for at least 10 seconds. Breathing effort helps classify the event:
- Obstructive apnea: airflow stops while the chest and abdomen continue trying to breathe.
- Central apnea: airflow and breathing effort stop together.
- Mixed apnea: the event begins without breathing effort, then effort returns while airflow remains absent.
These distinctions matter because obstructive and central sleep apnea can have different causes and treatments. A report may list the central apnea index and obstructive apnea index separately even though the total AHI combines respiratory events 1.
What counts as a hypopnea?
A hypopnea is a partial reduction in airflow rather than a near-complete pause. Under the American Academy of Sleep Medicine's recommended adult rule, airflow must fall by at least 30% for at least 10 seconds and be followed by either a 3% or greater oxygen drop or a brief brain arousal 2.
Another rule counts a hypopnea only when the airflow reduction is accompanied by at least a 4% oxygen drop. It does not include arousal-only events. Because the 3% or arousal rule can count events that the 4% rule does not, the same recording can produce different AHI values depending on the scoring rule. Your report should say which rule the lab used 3.
AHI, REI, RDI, and PAP AHI are different
A label on the report can tell you as much as the number itself.
| Measure | Where it usually comes from | What the denominator means |
|---|---|---|
| AHI | Laboratory polysomnogram | Apneas and hypopneas per hour of EEG-confirmed sleep |
| REI | Home sleep apnea test | Respiratory events per hour of monitoring or estimated sleep time |
| RDI | Usually a laboratory sleep study | Apneas, hypopneas, and respiratory effort-related arousals per hour of sleep |
| PAP-reported AHI | CPAP, APAP, or bilevel device | Events estimated from the machine's airflow signal while therapy is running |
Most home sleep apnea tests do not record the brain signals needed to distinguish sleep from wakefulness. If you wear the device for 8 hours but sleep for 6, dividing events by the longer monitoring time can make the REI lower than the AHI would have been. Home tests also generally cannot identify hypopneas that end only in a brain arousal. Signal loss can add more uncertainty 3.
This is why a negative, inconclusive, or technically inadequate home test does not always end the evaluation. The AASM recommends a laboratory polysomnogram when an adult's home test is negative or unclear but the clinical concern remains 3.
The term RDI is not used consistently. In many laboratory reports, it adds respiratory effort-related arousals (RERAs), which are breathing changes that disturb sleep without meeting the scoring criteria for an apnea or hypopnea. Some Medicare coverage material also uses RDI for events divided by recording time on certain home tests 4. If your AHI and RDI are different, check the report's definition or ask which events and recording time were included.
How adult AHI ranges are interpreted
Adult reports commonly group AHI values into these frequency bands 5:
| Adult AHI | Common label |
|---|---|
| Fewer than 5 events per hour | Below the usual OSA threshold |
| 5 to fewer than 15 | Mild |
| 15 to fewer than 30 | Moderate |
| 30 or more | Severe |
These labels describe how often scored events occurred. They do not, by themselves, tell you the depth of oxygen drops, degree of sleep disruption, cause of the events, or your individual health risk.
For adults, obstructive sleep apnea may be diagnosed when obstructive events occur at least 5 times per hour along with symptoms or relevant clinical features. A frequency of at least 15 events per hour can meet diagnostic criteria even without reported symptoms. The diagnosis should still be based on a comprehensive sleep evaluation, not a number viewed in isolation 3.
An AHI below 5 does not prove that sleep-related breathing is healthy. A low overall value can hide a higher REM or back-sleeping AHI, arousal-related breathing disturbances, an unrepresentative night, or underestimation by a home test. Persistent symptoms deserve review even when the headline number looks reassuring.
AHI in children uses pediatric criteria
Adult thresholds should not be applied to children. Pediatric breathing patterns, event scoring, symptoms, and treatment decisions are different.
Pediatric reports often emphasize the obstructive AHI (oAHI), which counts obstructive apneas, mixed apneas, and hypopneas but excludes central apneas. An oAHI of 1 or more event per hour can be clinically significant in a child. Pediatric sources commonly describe values from 1 to fewer than 5 as mild, but severity labels and treatment thresholds are not fully uniform. Symptoms, oxygen and carbon dioxide findings, growth, behavior, underlying conditions, examination findings, and age all influence interpretation 6.
A child's report should be reviewed by the clinician who ordered the test or a pediatric sleep specialist. Do not use an adult chart to decide that a child's result is normal or that care is unnecessary.
Why two people with the same AHI can be affected differently
AHI compresses a full night of physiology into one event count. Several details can change what that count means.
Symptoms and daily function
Daytime sleepiness, fatigue, morning headaches, impaired concentration, mood changes, insomnia symptoms, and witnessed choking help show how sleep disruption affects a person. Symptoms do not rise in a fixed line with AHI. Someone with a lower AHI may be very sleepy, while someone with a higher value may report few symptoms 5.
Oxygen burden
AHI counts an event once whether the oxygen change is small or deep. Clinicians may also look at the oxygen nadir, oxygen desaturation index, time spent below a specified oxygen level, and the pattern of drops.
Research on hypoxic burden measures the depth and duration of oxygen drops linked to respiratory events. In two cohorts of middle-aged and older adults, higher hypoxic burden was more consistently associated with cardiovascular mortality than AHI. This observational finding does not turn hypoxic burden into a stand-alone forecast for an individual, but it shows why event frequency cannot represent the whole physiologic load 7.
Event duration and arousals
AHI gives the same count to any event that crosses the scoring threshold. It does not show whether events lasted 10 seconds or much longer, or how strongly the brain and nervous system responded.
Event duration is not a simple "longer is always worse" measure. In a prospective cohort, shorter average respiratory events were associated with higher all-cause mortality after adjustment for AHI and other factors. The authors interpreted shorter events as a possible marker of a lower arousal threshold. The study was observational, so the finding should not be used to predict one person's outcome 8.
The total arousal index and the number of respiratory arousals can also help explain fragmented sleep. AHI alone does not capture the intensity of each arousal or all causes of disrupted sleep 5.
Sleep stage and body position
A total-night average may hide a much higher AHI during REM sleep or while sleeping on your back. Ask whether the report lists REM AHI, non-REM AHI, supine AHI, and non-supine AHI. Also check how much time you actually spent in each stage and position. A high value based on only a few minutes needs cautious interpretation, while very little REM or back sleep can leave uncertainty about how severe apnea becomes under those conditions.
Event type and medical context
The balance of obstructive, central, and mixed events matters. So do heart or lung disease, stroke history, neuromuscular disease, opioid use, and other factors that can affect testing choices, risk, or treatment. An AHI label does not account for these differences 3.
Researchers and sleep specialists therefore describe AHI as useful but incomplete. Measures of oxygen stress, sleep fragmentation, event timing, and symptoms may add information, but no single replacement metric currently covers every aspect of sleep apnea 5.
AHI can change from one night to the next
Sleep position, REM sleep time, alcohol or sedative exposure, nasal congestion, altitude, and ordinary biological variation can change the number and pattern of events. A large community-based study using repeated home measurements found substantial night-to-night variation and diagnostic misclassification when results were based on one night. The sample consisted of people who had bought an under-mattress sleep monitor, so the exact results may not apply to every clinical population. The broader lesson is most relevant near a diagnostic cutoff: one night is a sample, not a permanent personal score 9.
If a result does not fit the symptoms or the observed breathing problems, ask about:
- whether the study captured enough sleep, REM sleep, and back sleeping
- whether the signals were technically adequate
- which hypopnea rule was used
- whether the number is AHI, REI, or RDI
- whether a repeat or laboratory study would change the next decision
What PAP-reported AHI can tell you
CPAP, APAP, and bilevel machines estimate residual breathing events from airflow while the device is running. They do not record the same set of signals as a diagnostic polysomnogram, and machine time can include periods when you are awake. Manufacturers also use different event and leak definitions. For those reasons, a PAP-reported AHI is useful for following trends but is not the same measurement as the diagnostic AHI 10.
A change in the machine's number should be read alongside:
- nightly use and whether the mask stayed on through sleep
- mask or mouth leak
- pressure patterns
- whether events are labeled obstructive or central
- symptoms such as sleepiness, awakenings, or air hunger
- oxygen data, when your clinician considers monitoring necessary
A consistently low PAP-reported AHI is generally more reassuring than a high one, but no single device value should be interpreted without this context. Do not change pressure settings or stop therapy based only on one reading. Contact your sleep clinician if the value remains elevated, changes substantially, central events rise or persist, leak is persistent, or symptoms continue.
Follow-up after treatment starts
Effective follow-up asks more than whether AHI fell. It includes treatment use, comfort, leaks or side effects, symptoms, daytime function, and the condition being treated. The AASM recommends adequate follow-up with troubleshooting and monitoring of objective treatment use and effectiveness after PAP begins and over time 11.
Adults who are doing well on PAP generally do not need routine repeat sleep testing simply to generate a new AHI. Follow-up polysomnography or home testing may be appropriate when symptoms persist or return, after certain non-PAP treatments, after a clinically significant weight change, when sleep-related low oxygen or hypoventilation needs reassessment, when cardiovascular disease changes, or when PAP-generated data are unexplained 12.
Your clinician can help decide which result would answer the practical question. Sometimes that is a machine download, sometimes it is an overnight oxygen assessment, and sometimes it is another sleep study.
Questions to ask about your sleep report
Bring the full report, not only the summary AHI, and ask:
- Is this number an AHI, REI, RDI, oAHI, or PAP estimate?
- Was the study performed in a lab or at home, and what time was used in the denominator?
- Which hypopnea scoring rule was used?
- What were the obstructive, central, REM, and back-sleeping indices?
- How low did oxygen fall, how long was it low, and were the drops linked to respiratory events?
- How often did arousals occur?
- Did the recording include enough sleep in the relevant stages and positions?
- How do my symptoms and medical history change the interpretation?
- What treatment outcome are we following besides the AHI?
AHI is most useful when it starts this conversation. The number tells you how often scored breathing events occurred under the conditions of that test. The rest of the report helps explain what those events did and what to do next.
