Understanding Your Sleep Study Results

Published by SleepApneaSelect · Sources checked October 7, 2026

Start with two lines on your sleep study report. One is the AHI. On a home test, that main number may instead be labeled REI or pAHI. The other is the doctor's written conclusion, usually labeled Impression or Interpretation.

AHI stands for apnea-hypopnea index. It is the average number of scored pauses or reductions in breathing per hour of sleep. For adults, these are the usual ranges.

Sleep apnea AHI chart for adults

These are the usual adult AHI ranges (MedlinePlus (opens in a new tab)). Adult REI reports commonly use the same bands, and the WatchPAT maker's sample report (opens in a new tab) applies the 5, 15, and 30 cutoffs to pAHI. The indices are measured differently, as explained below. Exactly 15 counts as moderate. Exactly 30 counts as severe. An AHI under 5 does not rule out every sleep or breathing problem.

That chart is the easy part. Three things trip people up:

• The same night can give two different numbers. A report can show 26 on one line and 14 on another, and both can be correct. See why.
• "Mild" describes how often events happened. It does not describe how tired you feel.
• A number under 5 on a home test does not always close the case. See what to do.

This page explains what each line on a report means and what to ask about it. We explain the words and the standard ranges. We don't read your report, and a label on a chart is not a diagnosis. Your clinician's conclusion is the part that counts.

This page is for adult reports. Children are scored on a different scale. For a child's report, see the NHLBI page on sleep apnea in children (opens in a new tab) and talk with your child's sleep team.

Already talked through your results and have a CPAP prescription? See what to check before buying CPAP.

What an AHI number counts

AHI is a rate, not a total. The lab adds up two kinds of events and divides by the hours you slept.

• An apnea is when airflow stops or nearly stops for at least 10 seconds. The American Academy of Sleep Medicine (AASM) uses a drop of at least 90% in the airflow signal for adult scoring (AASM respiratory scoring rules (opens in a new tab)).
• A hypopnea is when breathing gets shallow for at least 10 seconds. Under the AASM's recommended rule, airflow has to drop by 30% or more, and it has to come with at least a 3-percentage-point oxygen drop or a brief brain-wave arousal (AASM hypopnea rule (opens in a new tab)).

An arousal is a brief change in brain-wave activity during sleep. It does not have to be a full awakening you remember.

Here is the math. Say a report shows 90 events in 6 hours of sleep. 90 ÷ 6 = 15. The AHI is 15.

So if your report says "90 events," that is the total for the night. And "AHI 15" does not mean you fully stopped breathing 15 times an hour. The total includes shallow-breathing events as well as full or near-full pauses.

Does my number mean I have sleep apnea?

Your clinician decides that, using the number plus your symptoms and health history.

The AASM's rule for obstructive sleep apnea in adults has two paths. One is 5 or more predominantly obstructive breathing events per hour with symptoms or related health problems, such as daytime sleepiness, waking up gasping, loud snoring someone else has noticed, or high blood pressure. The other is 15 or more predominantly obstructive events per hour, with or without symptoms. The count can include RERAs on a lab study; the denominator is sleep time for polysomnography (PSG), a full sleep study, and monitoring time for a conventional home test (AASM position statement (opens in a new tab)).

That is why an AHI of 7 can lead to a diagnosis for one person and not for another.

Can "mild" sleep apnea still make you feel awful?

Yes. The chart only counts how often events happened. It does not measure how far your oxygen dropped, how long each event lasted, or how many times your sleep was broken up.

The AASM says its recommended scoring rule counts breathing events linked to brief arousals for this reason. Events that wake the brain matter for people with sleepiness, fatigue, and poor sleep, even when oxygen barely drops (AASM position statement (opens in a new tab)).

If your label says "mild" and you feel wrecked, say so plainly at your visit. Your symptoms are part of the picture.

Read your report in five checks

Reports look different from lab to lab. These five checks work on almost any of them.

One habit will save you a lot of confusion. When you write a number down, keep its exact name and its units. "RDI 12" and "12 minutes below 90%" are very different things.

Find a term on your report

Or jump straight to a section: AHI chart · AHI vs. REI vs. RDI · pAHI and pRDI · Two AHI numbers · Oxygen lines · Central events · REM and back sleeping · Other terms · Confusing results · CPAP machine numbers

Example sleep study report, explained line by line

The two examples below are made up for teaching. The numbers are not from a real patient, and they are not a diagnosis. We built them so the math is easy to follow.

Example 1: an in-lab study with no CPAP

This example uses the AASM recommended hypopnea rule: an oxygen drop of at least 3 percentage points or a brain-wave arousal. All 6 hours of sleep have usable breathing and oxygen data.

What to take from it:

• 90 is the total. 15 is the rate. People mix these up all the time.
• RDI is higher than AHI because this lab's RDI also counts 30 RERAs.
• A lowest oxygen of 86% does not mean oxygen sat at 86% for 12 minutes. The 86% is one low point. The 12 minutes is all the time spent anywhere under 90%.
• None of these numbers picks a treatment. They tell you what to ask about.

Example 2: a home test that reports REI

REI stands for respiratory event index. It works like AHI, with one big difference. Most conventional home tests do not measure sleep with brain waves. They divide by monitoring time instead. That is the recording time after unusable data and any identified wake time are excluded. Some wake time can still be included (AASM home-test position statement (opens in a new tab); study of monitoring-time editing (opens in a new tab)).

Time spent lying awake can enlarge the bottom of the fraction. So an REI can underestimate events per hour of actual sleep. There is no set formula to convert one into the other, so don't try to fix the number yourself with a guess at how long you slept.

If your home report says pAHI or pRDI

Those letters come from a finger-sensor home test called WatchPAT. The "p" refers to peripheral arterial tonometry (PAT), the signal the finger sensor reads.

Itamar Medical, the company that makes it, says the device uses signals from the finger sensor and movement to estimate sleep time and light, deep, and REM sleep (Itamar Medical software manual, sleep-stage and index definitions (opens in a new tab)). Its sample report states that its respiratory indices are calculated on technically valid sleep time, not on the full recording (Itamar sample report (opens in a new tab)).

Two things to keep in mind:

• Those sleep stages are estimates. They are not the same as brain-wave staging in a lab.
• Keep the "p" when you talk about your number. Say "pAHI of 12," not "AHI of 12." It tells the next clinician which kind of test you had.

Two of the home testing services we compare say they use the WatchPAT One. We have not seen their patient reports, so yours may be laid out differently from the maker's sample.

AHI vs. REI vs. RDI: why the numbers differ

Sources: American Thoracic Society guide (opens in a new tab) for AHI and lab RDI; AASM home-test statement (opens in a new tab) for REI; Itamar software manual (opens in a new tab) for pAHI and pRDI; Medicare PAP policy (opens in a new tab) for the Medicare meaning of RDI.

pRDI is not lab RDI. Do not subtract pAHI from pRDI and call the difference a brain-wave-confirmed RERA index. WatchPAT uses its own signals and algorithm; it does not measure brain waves.

Why is my RDI higher than my AHI?

In a lab, RDI adds a third kind of event called a RERA. That stands for respiratory effort-related arousal. It means a sequence of breaths with increased effort or restricted airflow ended in a brief brain-wave arousal, but did not meet the scoring rule for an apnea or hypopnea.

With that definition, and the same scoring rule and sleep period, a lab RDI is equal to or higher than the AHI. In Example 1, AHI is 15 and RDI is 20.

Some people get a diagnosis based on RDI when their AHI is under 5. That is not a mistake. The AASM's diagnosis rule counts RERAs too (AASM position statement (opens in a new tab)). If that is your situation, ask your clinician which number they used and why.

The clock matters as much as the count

Here is an example from the WatchPAT maker's published sample report. The device counted 213 events for pAHI 3%. The recording ran 10 hours 13 minutes. The device estimated 8 hours 20 minutes of sleep, but only 8 hours 2 minutes were technically valid for these respiratory indices.

• 213 ÷ (482 minutes ÷ 60) = 26.5 per hour of technically valid sleep
• 213 ÷ (613 minutes ÷ 60) = 20.8 per hour of recording

Same night, same events. The only thing that changed is the clock. The report itself uses the 26.5. The 20.8 is our own math, to show what dividing by recording time would do. It is not a reported REI or a way to convert this pAHI into another test's result.

Which number do I give another doctor or my insurer?

Give the exact name, the value, the type of test, and the scoring rule if it's shown. Better yet, hand over the full signed report. Doctors and insurers can use different definitions for the same letters, so the name matters as much as the number.

Why your report may show two AHI numbers (3% and 4%)

Some reports list the main number twice, such as "AHI 3%" and "AHI 4%." Both can be correct. They use different rules for what counts as a hypopnea.

The AASM recommends the first rule and allows the second, which its Version 3 update calls "optional." Centers must state which rule they used (AASM scoring rules (opens in a new tab); Version 3 update (opens in a new tab)). Medicare uses the 4% rule (CMS (opens in a new tab)).

For the same signals and sleep period, the 4% rule counts the same number of events or fewer. The result can be lower; it does not have to be. A fall from 96% to 92% is a 4-percentage-point drop.

Back to Example 1. Under the recommended rule, 66 hypopneas counted. Now suppose only 42 of them had an oxygen drop of at least 4 percentage points.

That is a made-up example. Now one from the manufacturer's published sample.

The maker's own sample report shows the same thing. Itamar Medical's sample WatchPAT report lists both scores for one night (Itamar sample report (opens in a new tab)):

Same report. Same test. One score lands in the moderate range. The other lands in the mild range. The report applies different counting rules to the same test data.

Two cautions:

• There is no fixed gap between the two numbers. It is different for every person and every night.
• "3%" on a home report does not prove that brain-wave arousals were measured. Most home devices can't see brain waves. Read the method note on your report.

The question to ask: "Which rule did you use for my diagnosis, and which number will go to my insurance plan?"

Oxygen results: lowest oxygen, ODI, and time below 90%

The oxygen lines tell you things the AHI can't. Two people with the same AHI can have very different oxygen results.

Read these lines together. The average can't tell you how long a low reading lasted. The lowest point can't tell you how much of the night was affected.

In Example 1, the sleeper spent 12 minutes below 90%. That is about 3.3% of 360 minutes of sleep. It does not mean 12 dips, and it does not mean 3.3 dips per hour.

What if my lowest oxygen looks scary?

First, ask if the reading was checked. A finger sensor gives an estimate, not an exact measurement. The FDA says cold skin, poor circulation, skin tone, skin thickness, tobacco use, and nail polish can all throw a reading off (FDA (opens in a new tab)). An artifact can produce a false low point; a published WatchPAT study showed why the oxygen trace needs review (Plante and Rumble, 2023 (opens in a new tab)).

Second, ask how long oxygen stayed low. A brief dip and a long stretch are different stories.

We didn't find a simple chart for "safe" oxygen numbers in the guidelines we checked. What matters depends on your heart and lung health, so this is a question for your clinician.

Long stretches of low oxygen can also come from problems other than a blocked airway. When there is significant heart or lung disease, or suspected sleep-related hypoventilation (not breathing enough), the AASM recommends an in-lab study rather than a routine home apnea test (AASM diagnostic guideline (opens in a new tab)).

If you have serious or worsening symptoms right now, get medical help. Don't wait for someone to explain a number from a past night.

Obstructive, central, and mixed events

Sources: AASM respiratory scoring rules (opens in a new tab) and the Itamar manual's pAHIc definition (opens in a new tab).

Seeing some central events on a report does not, by itself, mean you have central sleep apnea. In Example 1, there are 6 central apneas in 6 hours, or 1 per hour.

A large share of central events can change the diagnosis and care plan, but the share alone does not establish a diagnosis. The AASM's home-testing pathway is intended for uncomplicated adults; suspected non-obstructive breathing problems need a clinician to choose the right evaluation (AASM diagnostic guideline (opens in a new tab)). Don't try to diagnose it from one line or one flag on a home report.

Ask: "Were any of my events central, and does that change the plan?"

Our site covers obstructive sleep apnea. If your report points to central or complex sleep apnea, start with a sleep clinician.

REM AHI and supine AHI: why part of the night can look worse

Many reports split the AHI by sleep stage and by body position. "Supine" means on your back.

Each slice of the night has its own clock. REM AHI is the events during REM sleep, divided by the hours of REM sleep. Here is Example 1 again, split two ways.

Notice that the whole-night number is not the average of the two rows. Averaging 30 and 12 gives 21, which is wrong. You add up the events and divide by all the hours: 90 ÷ 6 = 15.

Why this matters to you:

• A high REM or back-sleeping number does not replace the whole-night number. It adds detail.
• If you barely slept on your back during the test, or had very little REM, the whole-night number may not show your usual night. That is a fair thing to ask about.
• A blank REM row can mean little or no REM was recorded. It does not have to mean zero events.

The NHLBI notes that some providers suggest side sleeping to help keep the airway open (NHLBI (opens in a new tab)). Whether that is enough for you is your clinician's call, not something to read off this table.

Other numbers on a sleep study report

Sources: Shrivastava and colleagues, 2014 (opens in a new tab) for sleep timing and stage terms; ATS sleep-study guide (opens in a new tab) for split-night and titration studies; AASM movement-disorder guideline (opens in a new tab) for why a leg-movement count alone is not a diagnosis; WatchPAT sample report (opens in a new tab) for its hypoxic-burden labels.

What if your results are negative, confusing, or don't match how you feel?

A home test came back negative or unclear

A home test can miss things. Conventional tests without brain-wave monitoring cannot directly score arousals, so they can miss hypopneas that cause an arousal without an oxygen drop. Many do not measure actual sleep time.

The AASM is clear on the next step when a home test was used to investigate suspected obstructive sleep apnea. If that test is negative, unclear, or didn't record enough, it recommends an in-lab study next (AASM guideline summary (opens in a new tab)). Buying a second home test is not the default.

You barely slept, or a sensor came off

Ask how many hours of good data the test captured. The AASM guideline specifies at least 4 hours of usable oxygen and airflow data during a recording attempt that covers the usual sleep period. It also specifies the required sensors, including an alternative PAT setup. Four hours on the clock alone does not prove a test was adequate; ask whether your test met the requirements for that device (AASM diagnostic guideline (opens in a new tab)).

Also ask if a person looked at the recording. The AASM says diagnosis and treatment decisions should not rest only on a home test's automatic score. The raw data must be reviewed by a physician who is board-certified in sleep medicine or supervised by one (AASM position statement on home tests (opens in a new tab)).

A rough night does not make a test worthless. Don't throw out a result just because you remember sleeping badly, and don't redo the math with your own guess at sleep time.

Two tests gave very different numbers

It happens, and there are ordinary reasons for it:

1. Different counting rules (3% or 4%).
2. Different clocks (sleep time or recording time).
3. A different night. Time on your back and how much REM you got can change the number.
4. Different equipment. A home device can't see everything a lab can.
5. A computer score on one test and a human-reviewed score on the other.

A lab night isn't a normal night, either. People often sleep worse in a strange bed with wires on. Doctors call this the first-night effect (Shrivastava and colleagues, 2014 (opens in a new tab)). The AASM also says a second in-lab study can be considered when the first one is negative and the doctor still suspects sleep apnea (AASM guideline summary (opens in a new tab)).

Bring both reports to one clinician and ask them to explain the gap.

The label says "mild" but you feel terrible

Check the exact name of the number and the rule behind it. Then take the mismatch to your clinician. Be specific about the symptoms you are having. Try: "My AHI is 7 on the 4% rule, but I'm exhausted. What else on this report could explain that?"

Sleep study AHI vs. the AHI on your CPAP machine

These are three different numbers. Don't compare them as if they were one.

"My machine says my AHI is 2. Am I cured?" That number cannot tell you. It means the machine estimated about 2 events per hour of use while treating you. It says nothing about your breathing without it.

Machine numbers are also not the same as lab numbers. The American Thoracic Society says each CPAP maker defines events in its own way, and those leftover-event numbers are harder to interpret than hours of use (ATS statement, 2013 (opens in a new tab)).

Don't stop treatment or change your settings based on the machine's number or on this page. Ask the clinician who manages your treatment what they watch to judge whether it is working.

What happens next after your sleep study

The next step is your clinician's conclusion and a talk about the plan. The AHI alone doesn't choose a treatment, a machine, or a pressure setting.

Positive airway pressure (PAP), which includes CPAP, is the most common treatment. It is not the only one. The NHLBI also lists lifestyle changes, oral devices fitted by a dentist, a weight-loss medicine used with diet and physical activity for moderate-to-severe obstructive sleep apnea in adults with obesity, mouth and face muscle therapy, and surgery (NHLBI (opens in a new tab); FDA medicine approval (opens in a new tab)). Ask which ones fit you.

Then pick the row that matches where you are.

If you already have a CPAP prescription, start with what to check before you pay: the exact model, what's in the box, and whether you can return a machine after trying it.

Our current comparison covers one machine, the ResMed AirSense 11 AutoSet, at two sellers. Make sure the model matches your prescription before you order. We aren't suggesting that model based on anyone's results.

Does this result mean insurance will cover CPAP?

Not on its own. Each plan has its own rules, and only your plan can confirm what it will pay.

Medicare's rules are public, and many people are surprised by them:

• Medicare's initial 12-week CPAP coverage pathway uses an AHI or RDI of 15 or more, or 5 to 14 with documented excessive daytime sleepiness, trouble thinking clearly, a mood disorder, insomnia, high blood pressure, heart disease from narrowed arteries, or a past stroke (CMS coverage rule (opens in a new tab)).
• Medicare counts a hypopnea only when oxygen drops by at least 4 percentage points, along with the breathing-reduction and duration requirements (CMS (opens in a new tab)).
• Medicare leaves RERAs out of both AHI and RDI (CGS Medicare (opens in a new tab)).

These are only part of the requirements. Medicare's PAP equipment policy also requires at least 30 qualifying events for the 15-or-higher path, or at least 10 for the 5-to-14 path, along with the specified evaluation, test, instruction, and documentation (CMS PAP equipment policy (opens in a new tab)).

The maker's 26.5 and 14.2 scores show why the scoring rule matters. They do not establish Medicare eligibility. Ask your clinician and equipment supplier which Medicare-defined index and reporting rules apply. Medicare contractor Noridian gives separate rounding instructions for coverage reporting (opens in a new tab); do not use billing rounding to change the clinical AHI chart.

So ask two things: "Which number and which rule will you send to my plan?" and "What records does my plan need?"

Questions to take to your sleep clinician

Take your signed report and this list to the clinician who ordered or read your test. You can copy or print it. No account is needed. Copy and print happen in your browser; the questions are not sent to SleepApneaSelect.

Common questions

How long does it take to get sleep study results?

It depends on the lab or service. We didn't find a standard timeline to quote. Ask two things when you book: when the signed report will be ready, and whether you need a results visit to get it. An automatic summary from a device is not the finished report.

How do I get a copy of my sleep study report?

Ask the clinician or testing service that ordered the test. Ask for the full report, not just a summary letter. Under HIPAA you generally have the right to see and get copies of records held by covered providers. They normally must act on an access request within 30 days. One additional 30-day extension is allowed with written reasons and a completion date. Any copying fee must be reasonable and cost-based (HHS (opens in a new tab); 45 CFR 164.524 (opens in a new tab)). Our prescription and records help section lists what to ask for.

Can I work out my own AHI from the report?

You can check the math. If the report gives the number of apneas and hypopneas and the total sleep time, divide the events by the hours. That shows how the number was built. It does not check that the events were scored correctly, and it does not replace the clinician's reading.

I haven't had a sleep test yet. Where do I start?

A clinician decides if a home test fits your situation. If you are looking into that route, see what to compare before a home sleep test, or go straight to the comparison.

Can SleepApneaSelect look at my report?

No. We explain report terms and compare services. We don't read personal reports, and please don't email us your health records. The clinician who ordered or read your test is the right person to ask.

How we checked this page

We wrote this page from published sources that we checked on October 7, 2026:

• The American Academy of Sleep Medicine's scoring rules, diagnosis rules, and testing guidelines
• Medicare's published CPAP coverage rules
• The American Thoracic Society's guide to reading sleep study reports and its statement on CPAP machine data
• FDA and NHLBI patient pages
• MedlinePlus report guidance and HHS record-access rules
• The sample report published by the maker of the WatchPAT home test

The in-lab and home examples are made up for teaching, and we checked that every number in them adds up. The 26.5 and 14.2 example comes straight from the device maker's public sample report. It is the maker's example, not a patient we know.

No clinician has reviewed this page. We are a comparison publisher, not a medical provider. Here are our editorial standards.

See something wrong or out of date? Send us the source.