What Is a Good AHI? Making Sense of Your Sleep Apnea Number
AHI is the number your diagnosis, your severity, and often your insurance coverage hangs on. Here is what it measures, what the bands mean, and where it misleads.
Your sleep study came back with a number. Someone said “moderate” and moved on. Then you got a machine that shows you a different number every morning, and nobody explained the relationship between the two.
Let’s fix that.
What the number counts
AHI stands for apnea-hypopnea index. It is a rate, not a total: the number of breathing events per hour of sleep.
Two kinds of event get counted.
An apnea is a near-total stop — airflow essentially ceases for at least ten seconds. A hypopnea is a partial collapse: airflow drops substantially for at least ten seconds and is accompanied by either an oxygen desaturation or an arousal from sleep.
Add them up, divide by hours of sleep, and you have your AHI. Someone with 300 events across six hours of sleep has an AHI of 50.
The severity bands
For adults, the conventional bands are:
| AHI | Category |
|---|---|
| Under 5 | Normal |
| 5 to under 15 | Mild |
| 15 to under 30 | Moderate |
| 30 and above | Severe |
These are widely used, but they are a convention rather than a law of nature. Someone with an AHI of 14 and profound daytime sleepiness is in worse shape than someone with an AHI of 17 who feels fine. The bands are a shorthand, not a verdict on your health.
Where the number becomes bureaucratic
The bands also decide who pays. Medicare’s coverage rules for PAP devices set out two routes to qualifying: an AHI or RDI of 15 or more events per hour with at least 30 recorded events; or an AHI of 5 to 14 with at least 10 events plus a documented comorbidity — excessive daytime sleepiness, cognitive impairment, mood disorder, insomnia, hypertension, ischemic heart disease, or a history of stroke1 .
That second route is worth knowing about. Plenty of people land in the mild band and assume that means no coverage. If you have hypertension, or documented daytime sleepiness, the picture may be different — and it is your clinician’s documentation that determines it. Most private insurers write similar rules, often modelled directly on Medicare’s.
What “good” looks like once you’re treated
The usual target on therapy is an AHI under 5. Many people on a well-fitting mask at the right pressure end up between 0.5 and 2.
A few things are worth knowing about that number.
It bounces around, and that’s normal. A single bad night — a cold, a glass of wine, sleeping on your back, a leaking mask — can double your AHI. What matters is the trend across a couple of weeks, not last night.
A rising trend is a real signal. If your nightly AHI has been drifting up over a month, something has changed: a worn cushion, weight change, nasal congestion, or genuine progression. That is worth a conversation.
A very low AHI with terrible sleep is still a problem. If your machine says 1.2 and you feel wrecked, believe how you feel. Leaks, an inappropriate pressure, a different sleep disorder, or something entirely unrelated could be responsible. A good number is not permission to be dismissed.
Why your machine’s number differs from your sleep study’s
They are not measuring the same thing, and they never will match exactly.
A sleep study watches you directly: airflow at the nose and mouth, chest and abdominal effort, blood oxygen, and EEG to establish when you were actually asleep. It knows the difference between an hour in bed and an hour asleep.
Your machine has one sensor — the airflow through the hose. It infers events from changes in that flow. Crucially, it usually cannot tell whether you were asleep. If you lie awake for forty minutes with the mask on, that time is generally counted in the denominator, which drags your reported AHI down.
Machines also classify events into obstructive, central, and unknown categories using algorithms that vary by manufacturer. It is a good estimate. It is not a scored polysomnogram, and it is not a diagnosis. More on interpreting it in how to read your CPAP data.
The other numbers worth knowing
RDI — respiratory disturbance index — includes everything the AHI counts plus respiratory effort-related arousals: events where you didn’t quite meet the hypopnea threshold but woke up anyway. RDI is always equal to or higher than AHI. Medicare’s rules treat AHI and RDI as alternatives for qualifying1 .
ODI — oxygen desaturation index — counts how often your blood oxygen dropped by a set amount per hour. Two people with the same AHI can have very different ODIs, and the oxygen picture is arguably the more relevant one for cardiovascular risk. If your report includes a nadir (lowest) oxygen saturation and time spent below 90%, those are worth reading.
Time in position. Many people are dramatically worse on their back. If your study shows an AHI of 45 supine and 8 on your side, positional therapy becomes a genuinely interesting conversation.
The honest summary
AHI is the number everyone uses because it is simple, standardised enough to compare, and easy to measure. It is also a blunt instrument: it treats a ten-second event and a ninety-second event as equal, it says nothing about oxygen, and it is scored inconsistently across labs.
Use it. Track it. Just don’t let it be the only thing anyone looks at — including you.
Common questions
What is a good AHI on CPAP?
Is an AHI of 5 bad?
Why is my AHI higher some nights?
Can my CPAP machine's AHI be wrong?
What AHI qualifies for a CPAP machine?
References
Every clinical claim above links to one of these. Where the evidence is mixed or thin, we say so in the text rather than picking the source that sounds most confident.