Why Your CPAP AHI Doesn't Match Your Sleep Study

Quick answer: why don’t the two AHI numbers match?

Your sleep study’s AHI and your CPAP machine’s nightly AHI are not the same measurement wearing the same name. The lab scores events using brain-wave (EEG) data to confirm you were genuinely asleep, plus oxygen and effort sensors (source: NIH – StatPearls). Your CPAP has none of that: it estimates events from airflow and pressure alone, using an algorithm its manufacturer wrote and validated on its own terms, not the sleep field’s standardized scoring rules (source: PLOS ONE). That is why a lab score of 32 and a machine average of 1.4 can both be honestly reported numbers describing two different things.

Key takeaways

Your CPAP has no EEG. It cannot tell whether you are actually asleep — it can only see airflow and pressure (source: NIH – StatPearls).
Manufacturers score events with their own proprietary algorithm, not the AASM’s clinical scoring rules used in a lab (source: PLOS ONE).
The two numbers agree loosely, not exactly. One real-world comparison found device and lab AHI only weakly correlated night for night, though broad agreement (AHI at or under 5 vs. above 5) was much better (source: PLOS ONE).
Devices tend to under-count, not over-count. The same study found the device caught only 39% of the events the lab identified (sensitivity), while correctly clearing 81% of the nights the lab also considered clear (specificity) (source: PLOS ONE).
Hypopneas are the harder call. Apnea counts tend to agree reasonably well between devices and labs; hypopnea counts, which depend on smaller airflow drops, agree far less (source: PLOS ONE).
None of this makes the device number useless. It is a good trend line and the basis for the adherence data Medicare and most clinicians actually use (source: CMS – LCD L33718).

This guide is written by the CPAP Community editorial team. We are not clinicians, so every clinical statement below is linked to its primary source. Only your own sleep clinician can interpret your device data or decide whether you need a repeat study; we can explain what the numbers mean.

What your sleep study actually measured

As we cover in sleep studies and your AHI result, explained, a sleep study scores two kinds of event: an apnea, a drop in airflow of 90% or more lasting at least 10 seconds, and a hypopnea, a smaller drop of at least 30% that also has to bring either a 3% oxygen desaturation or an arousal visible on the EEG (source: NIH – StatPearls) (source: AASM). The EEG channel is what makes an in-lab AHI a true AHI — events per hour of confirmed sleep, not per hour of time in bed.

What your CPAP machine measures instead

A consumer CPAP or APAP has no EEG, and most do not have a built-in pulse oximeter either (source: NIH – StatPearls). It can only see two channels: airflow, inferred from the motor’s pressure and flow sensors, and the pressure it is currently delivering. From that alone it has to guess, night after night, which dips in airflow are real breathing events and which are you rolling over, talking in your sleep, or the mask shifting.

To do that, each manufacturer writes its own detection algorithm, tunes it against its own validation data, and keeps the details proprietary. That algorithm is not the AASM’s scoring manual translated into software — it is a different, unpublished method aimed at roughly the same target (source: PLOS ONE).

How far apart the two numbers can really be

One real-world comparison of positive airway pressure device AHI against same-night polysomnography found the two numbers correlated only weakly across the full range of values (Spearman’s r = 0.13) (source: PLOS ONE). Categorising results more loosely — AHI at or under 5 versus above 5 — the device and the lab agreed on which side of that line a patient fell 77% of the time, which still leaves roughly one in four people whose device reading and lab reading told a different story about whether their sleep apnea looked controlled (source: PLOS ONE). The device caught only 39% of the residual events the lab identified (sensitivity), while correctly clearing 81% of the nights the lab also considered clear (specificity) (source: PLOS ONE) — a pattern that leans toward the device under-counting rather than over-counting.

Earlier research narrows down where the disagreement concentrates: apnea counts — the more dramatic, easier-to-detect event — agree reasonably well between devices and labs, while hypopnea counts, which depend on smaller airflow drops and require inferring an oxygen or arousal response the device usually cannot measure directly, agree far less well (source: PLOS ONE).

What can throw the number off further

Mask leak. A heavy air leak around the cushion changes the pressure and flow signal the device is reading, which can suppress real events or, less often, register events that were not really there.
Central vs. obstructive events. Telling an obstructive event (the airway is blocked) apart from a central one (the brain briefly stops signalling you to breathe) usually needs effort-belt data your CPAP does not have; most consumer devices report a single combined number.
Time in bed vs. actual sleep. This is the same problem the home sleep test has compared to a lab study — the device estimates sleep time from your breathing pattern, it does not confirm it with a brain-wave signal.
Clustering. A device’s nightly average can look fine while events are concentrated on your back or during REM sleep, which a single number will not show you.

So what should you actually do with the number

Treat it as a trend line, not a diagnosis. One good or bad night means less than a pattern over several weeks.
A consistently low device AHI is reassuring, not proof of a “cure” — it is one data point your clinician weighs alongside how you actually feel.
A rising trend, or symptoms that persist despite a low device number, is a reason to call your clinic rather than wait it out.
Do not adjust your own pressure based on the app number alone; that is a conversation for your clinician, using your full data and symptoms.
If you are repeatedly worried about your numbers, ask your clinic to review your downloaded data, or ask whether a repeat titration or in-lab study makes sense — that is the only way to get another AASM-scored number.

If you are still in your first few weeks of therapy, this kind of week-to-week noise is normal and expected — see our guide to your first 30 days on CPAP for what a reasonable adjustment period looks like.

What counts as good control on therapy

The lab-based severity thresholds — under 5 normal, 5 to 15 mild, above 15 to 30 moderate, above 30 severe — are useful shorthand, but they were built for an AASM-scored, EEG-confirmed AHI (source: NIH – StatPearls), not your device’s proprietary estimate. Most clinicians read a device AHI persistently under 5 as a good sign, combine it with how you actually feel during the day, and reserve a repeat study for cases where the two disagree (source: NHLBI – CPAP). The number on the app is a useful signal. It is not the same instrument that diagnosed you.

Frequently asked questions about your CPAP’s AHI vs. your sleep study AHI

Why is my CPAP AHI lower than my sleep study AHI?

This is the most common direction of disagreement. Your CPAP has no EEG, so it cannot confirm you were asleep, and it tends to under-count hypopneas in particular because it cannot directly measure the oxygen desaturation or arousal that a lab uses to score them. A real-world comparison found the device caught only about 39% of the events a same-night lab study identified.

Can my CPAP AHI be higher than my sleep study AHI?

It can, though it is less common. Mask leak can distort the signal the device is reading, sleep position or a new pressure setting can genuinely change how many events occur night to night, and if your original study was scored under the stricter 4% hypopnea rule rather than the 3%-or-arousal rule, the study number itself would have been lower to begin with.

Is my CPAP machine's AHI a real diagnosis?

No. It is an estimate from a proprietary algorithm reading airflow and pressure only, useful for tracking trends over time, but it was not scored using the AASM criteria a sleep lab uses and should not be treated as a new diagnosis on its own.

What is a good CPAP AHI number?

Most clinicians look for a device AHI consistently under 5, treated as a trend over several weeks rather than judged on any single night, and weighed alongside your symptoms rather than as a stand-alone verdict.

Should I worry if my device shows an AHI of 0?

A very low device AHI is a reassuring sign, but it is not proof of perfect control, since the device can under-count real events. If you still feel unusually sleepy or unrested despite a low number, it is worth mentioning to your clinician rather than assuming the number rules everything out.

Do different CPAP brands report different AHI numbers for the same person?

Yes, this is expected. Each manufacturer uses its own proprietary detection algorithm rather than a shared standard, so switching machines or brands can shift your reported number even if your actual sleep apnea has not changed.

When should I get a repeat sleep study instead of relying on my device numbers?

Consider raising it with your clinician if you have persistent daytime symptoms despite good device data, a device AHI that keeps rising over weeks without an obvious cause like a leak, or a major weight change since your last study. Your clinician is best placed to decide whether a repeat titration or full study is warranted.

Sources

How we source this guide

This guide is researched and written by the CPAP Community editorial team. We are writers and researchers, not clinicians. Clinical statements are traced to primary sources — the National Heart, Lung, and Blood Institute, the U.S. Food and Drug Administration, the Centers for Medicare & Medicaid Services and the American Academy of Sleep Medicine — and cited in the text so you can check them. We sell no CPAP equipment and take no manufacturer sponsorship. Nothing here is medical advice; see our medical disclaimer and speak to your own sleep clinician before changing your therapy.

Last reviewed by the CPAP Community editorial team: August 2026.

Why Your CPAP Machine’s AHI Doesn’t Match Your Sleep Study