How to Read and Understand Sleep Apnea Test Results

Sleep apnea test results boil down to one central number, the apnea-hypopnea index (AHI), which counts how many times per hour your breathing partially or fully stops during sleep. An AHI below 5 is considered normal, 5 to 14 is mild, 15 to 29 is moderate, and 30 or above is severe. But your report likely contains a dozen other metrics, and understanding what they mean individually and how they relate to each other can change how you think about your diagnosis and treatment options.

The AHI and Why Its Number Can Shift

The AHI is the headline figure on any sleep study report. It adds together two types of breathing disruptions: apneas, where airflow drops by more than 90% for at least 10 seconds, and hypopneas, where airflow drops by at least 30% for at least 10 seconds and is accompanied by either an oxygen dip or a brief brain arousal that fragments your sleep.¹ Divide the total events by the hours you actually slept, and you get the AHI.

Here is where it gets tricky. There is more than one way to score a hypopnea, and the method used can meaningfully change your number. The older standard required a 4% drop in blood oxygen to count an event. The newer standard, now recommended by the American Academy of Sleep Medicine, counts an event if oxygen drops by 3% or if the event triggers a brain arousal.² Switching from the 4% rule to the 3% rule raises the median AHI by roughly 5 to 6 events per hour. In one study, the proportion of patients classified as moderate or severe jumped from about 29% to nearly 74% simply by changing this threshold.³ If you are comparing results from two different labs or two different years, check which scoring rule was used before assuming your condition has worsened or improved.

REI Versus AHI on Home Sleep Tests

If you did a home sleep apnea test rather than an overnight stay in a sleep lab, your report probably shows a respiratory event index (REI) instead of an AHI. The difference matters. A home device cannot tell when you are actually asleep, so it divides the total breathing events by the entire recording time rather than by the time you spent sleeping. If you spent 45 minutes reading in bed before falling asleep, or woke up for a bathroom trip, that awake time dilutes your score. The REI tends to come out lower than the AHI would on the same night for the same person.⁴

This underestimation is consistent across all severity groups. Because the home device uses total recording time instead of total sleep time, the denominator is inflated, which drags the index down.⁵ Manually trimming out the non-sleep periods from the recording can improve accuracy, and some labs do this when reviewing home test data.⁶ The practical takeaway: a borderline-normal REI on a home test does not necessarily rule out clinically meaningful sleep apnea. If your REI is in the low-to-mid single digits but you still feel terrible during the day, a follow-up lab study may tell a different story.

Oxygen Numbers on Your Report

Beyond the AHI, your results will include several oxygen-related metrics. The oxygen desaturation index (ODI) counts how many times per hour your blood oxygen drops by a set amount, usually 3% or 4%. This number often tracks closely with the AHI. One study found about 87% agreement between the two metrics when classifying severity, with a strong statistical correlation.⁷ An ODI above 20 had high sensitivity for identifying severe sleep apnea in that same study, making it a useful screening metric when the full AHI is unavailable.

Your report will also show your lowest oxygen saturation (sometimes called the SpO2 nadir) and your average oxygen saturation during sleep. Healthy oxygen saturation during sleep generally stays above 90%. In one study tracking patients over seven consecutive nights, the median overnight saturation was about 96%, while the average lowest dip across the night was about 91%.⁸ If your nadir drops into the 80s or below, that signals more pronounced desaturation, which carries greater cardiovascular risk and will likely influence treatment urgency.

One wrinkle with home tests: because the ODI is also calculated against recording time rather than sleep time, it can appear lower than it would in a lab setting. In one comparison, the median ODI was about 23 events per hour when calculated across the full recording, but only about 17.5 per hour when sleep periods alone were counted. That gap widened as severity increased and as sleep efficiency dropped.⁹

Obstructive, Central, and Mixed Events

Not all breathing pauses are the same, and your report should break events into categories. Obstructive events happen when your airway physically collapses even though your brain is still sending the signal to breathe. You can usually see this on the polysomnography tracing as continued chest and abdominal effort despite absent airflow. Central events are the opposite: the brain temporarily stops sending the breathing signal altogether, so there is no effort and no airflow. Mixed events start as central and then transition to obstructive. The distinction shapes treatment decisions significantly, because positive airway pressure therapy that works well for obstructive apnea can sometimes worsen central apnea.

A related pattern worth understanding is Cheyne-Stokes respiration, a distinct breathing pattern where central apneas alternate with periods of gradually increasing then decreasing breathing effort in a rhythmic, wave-like cycle.¹⁰ This pattern is strongly associated with heart failure and stroke, so if it shows up on your study, it usually triggers a cardiology referral. Automated screening tools can detect Cheyne-Stokes respiration from airflow signals alone with accuracy above 90%.¹¹

If your report shows mostly obstructive events, you are in the majority of sleep apnea patients. If central events make up a large share, or if the technologist noted a Cheyne-Stokes pattern, your physician will likely take a different diagnostic and treatment path, often including cardiac evaluation.

Positional Data and REM-Related Severity

Many reports break your AHI into different body positions, typically supine (on your back) and non-supine (on your side or stomach). Sleeping on your back consistently produces a higher AHI because gravity pulls the tongue and soft palate backward into the airway.¹² If your supine AHI is dramatically higher than your non-supine AHI, you may have positional sleep apnea, which means simple interventions like a positional therapy device or even a tennis ball sewn into the back of a shirt could make a real difference. In one study of positional therapy, the overall AHI dropped from a median of about 14.5 to about 6 events per hour just by keeping people off their backs.¹³

Your report may also separate events by sleep stage, particularly REM (dream sleep) versus non-REM sleep. The airway muscles relax more deeply during REM, so many people have significantly worse breathing disruptions during that stage. Some patients have what researchers call REM-predominant sleep apnea, where the AHI during REM is elevated but the overall AHI looks mild or moderate. This phenotype deserves attention: emerging data suggest that REM-predominant apnea is linked to a higher risk of hypertension and blood sugar problems.¹⁴ A report with an overall AHI of 12 but a REM AHI of 45 is telling you something different than an AHI of 12 spread evenly across the night.

Arousal Index and Sleep Fragmentation

An arousal is a brief burst of brain activity that partially wakes you, even if you do not remember it the next morning. The arousal index counts how many of these happen per hour. A normal arousal index is generally considered to be fewer than about 10 to 15 per hour. In sleep apnea, each breathing pause typically ends with an arousal that restores muscle tone and reopens the airway. The respiratory arousal index tracks closely with the AHI. One study found a correlation of 0.96 between respiratory arousals and the AHI, making it almost a mirror of breathing severity.¹⁵

What is interesting is that the arousal index does not reliably predict how sleepy you feel during the day. That same study found no significant link between arousal frequency and scores on standard sleepiness questionnaires. This is counterintuitive: you would expect more arousals to mean worse daytime tiredness, but the relationship is messier than that. Some people tolerate heavy sleep fragmentation without much subjective sleepiness, while others feel wrecked after relatively few disruptions. This is one reason sleep specialists look at the full picture rather than any single number.

Why Your Symptoms May Not Match Your AHI

If your AHI comes back high but you feel fine, or comes back low but you are exhausted, you are not imagining things. The correlation between the AHI and self-reported sleepiness, typically measured by the Epworth Sleepiness Scale, is real but weak. One large review of the evidence found inconsistencies between Epworth scores and AHI across multiple studies, noting that while the two metrics tend to move in the same direction, the relationship is not strong enough to predict one from the other reliably.¹⁶ Interestingly, the oxygen desaturation index shows a stronger correlation with daytime sleepiness than the AHI does, suggesting that it is the depth of the oxygen dips, not just their frequency, that drives how tired you feel.¹⁷

This disconnect means two things for you as a patient. First, do not let a “mild” AHI convince you that treatment is unnecessary if you have significant symptoms. Second, do not assume a “severe” AHI means you will definitely feel terrible; some people adapt to substantial sleep disruption without dramatic daytime effects, although the cardiovascular risks remain regardless of how you feel.

Heart Rhythm Findings on a Lab Study

If you had an in-lab polysomnography, the study likely monitored your heart rhythm overnight. Obstructive sleep apnea produces a characteristic cyclical pattern: heart rate slows during the apnea and then speeds up with the arousal that restores breathing. Sleep specialists look for arrhythmias that coincide with these cycles.¹⁸

These are not rare findings. In one population-based study, about 53% of people without significant breathing disorders had at least one type of nocturnal rhythm disturbance, but that figure climbed to over 92% in people with severe sleep apnea. Both atrial and ventricular ectopic beats were significantly more common as severity increased, and after adjusting for age, weight, and other risk factors, the AHI itself remained an independent predictor of nocturnal arrhythmia.¹⁹ If your report mentions atrial fibrillation, frequent premature beats, or pauses in heart rhythm, bring these findings to your cardiologist even if you have no daytime cardiac symptoms.

How Results Guide CPAP Pressure

Once the diagnosis is established, the numbers on your sleep study directly inform treatment. For CPAP therapy, the goal is to find the lowest pressure that eliminates apneas, hypopneas, arousals, and snoring while keeping oxygen saturation above 90%. Clinical guidelines recommend targeting a residual breathing disturbance of fewer than 5 events per hour at the selected pressure.²⁰

Some labs determine this pressure during a second night dedicated entirely to titration. Others use a split-night protocol, where the first portion of the night establishes the diagnosis and the second portion titrates the pressure. Split-night protocols work well for most patients, with one study reporting successful titration in about 77% of cases.²¹ The approach is especially effective in people with an AHI above 20, where the diagnosis is clear early in the night and plenty of recording time remains for pressure adjustment.²²

If you did a home test, an auto-adjusting CPAP machine is often used instead: you wear it for a week or more, and the machine data are used to determine a fixed pressure, typically set at the 90th or 95th percentile of the pressures the machine delivered.²³ Some algorithms also factor in your oxygen dip rate and neck circumference to estimate a starting pressure.²⁴ Either way, the numbers from your diagnostic study are the starting point.

What Consumer Wearables Can and Cannot Tell You

Smartwatches and sleep-tracking rings now report sleep stages, oxygen trends, and even breathing disturbance estimates. It is tempting to treat these as a substitute for a sleep study, but the accuracy gap is still wide. In a large validation study comparing 11 consumer trackers against polysomnography, the best device achieved a sleep-stage agreement score of 0.69 out of 1.0, while the worst hit only 0.26.²⁵ That means even the best consumer device was misclassifying a meaningful chunk of sleep stages.

A separate study comparing three popular devices found that all three were over 95% accurate at detecting sleep versus wakefulness, a relatively easy task. But when it came to distinguishing between light sleep, deep sleep, and REM, accuracy dropped to between 50% and 86% depending on the device and the sleep stage.²⁶ Deep sleep was particularly problematic: one device underestimated it by over 40 minutes per night. These devices are useful as longitudinal trend monitors, showing you general patterns over weeks and months, but they cannot replace a diagnostic sleep study when sleep apnea is suspected. If your watch flags repeated oxygen dips or breathing irregularities, take that as a reason to pursue formal testing rather than a diagnosis in itself.

The Respiratory Disturbance Index

Some reports list an RDI (respiratory disturbance index) alongside or instead of the AHI. The RDI includes everything the AHI counts plus an additional category: respiratory effort-related arousals (RERAs). A RERA is an event where increasing breathing effort disrupts sleep even though the airflow drop does not meet the threshold for a hypopnea.² The RDI will always be equal to or higher than the AHI, because it captures events the AHI misses.

In practice, the clinical significance of RERAs as a separate category has diminished. The newer 3% hypopnea scoring rule now captures many events that would previously have been classified as RERAs, absorbing them into the AHI itself. In a large population-based study, only about 4% of middle-aged participants had a RERA index of 5 or more per hour, and that elevated RERA index was not independently associated with negative health outcomes.²⁷ If you see a gap between your AHI and RDI, it is worth noting but usually does not change the treatment plan.

REM-AHI and Metabolic Risk

The breakdown of your AHI by sleep stage is not just an academic detail. As discussed earlier, REM sleep produces more severe airway collapse in many patients, and emerging research points to real metabolic consequences. In one cross-sectional study of sleep apnea patients, those who also had nonalcoholic fatty liver disease showed a significantly higher REM-AHI (about 54 per hour versus 44 per hour) and higher arousal index compared to those without liver disease. After controlling for confounders, both REM-AHI and arousal index remained independent predictors of fatty liver disease.²⁸ This is a developing area of research, but it reinforces the value of looking at the stage-specific numbers on your report rather than focusing solely on the overall AHI.

If your physician has not commented on the REM-specific data in your results, it is reasonable to ask about it, especially if you have other metabolic risk factors like obesity, insulin resistance, or elevated liver enzymes. The overall AHI is the diagnostic workhorse, but the stage-specific and position-specific breakdowns are where personalized treatment insights often hide.