What Is AHI on CPAP? Numbers and What They Mean

AHI on a CPAP machine stands for apnea-hypopnea index, a count of how many times per hour your breathing partially or fully stops while you sleep. When you check your CPAP’s data each morning, AHI is the headline number telling you how well therapy controlled your sleep apnea overnight. Clinicians generally want to see that number below 5, but what the number actually captures, how accurately your machine measures it, and what it misses are all worth understanding before you obsess over every decimal point.

What AHI Actually Counts

AHI lumps together two types of breathing disruptions. An apnea is a complete or near-complete pause in airflow lasting at least 10 seconds. A hypopnea is a partial reduction in airflow, also lasting at least 10 seconds, that results in a drop in blood oxygen or a brief awakening. Your machine tallies every event of either type during the hours you were actually using it, divides by those hours, and reports the result as events per hour. If you slept with your CPAP on for 7 hours and it detected 14 events, your AHI for that night is 2.0.

Both apneas and hypopneas can be obstructive (your airway physically collapses or narrows) or central (your brain temporarily stops sending the signal to breathe).¹ Most CPAP machines report a combined AHI but also break it into separate apnea and hypopnea subtotals, and some newer devices further distinguish central from obstructive events. That breakdown matters if your residual events turn out to be central rather than obstructive, because the treatment strategy is different.

The Severity Scale and Where “Normal” Sits

Before you started CPAP, your sleep apnea was graded on the same AHI scale used by sleep labs worldwide.² The standard cutoffs, recommended by the American Academy of Sleep Medicine, are:

• Mild: 5 to 15 events per hour
• Moderate: more than 15 to 30 events per hour
• Severe: more than 30 events per hour

Those thresholds were set based on the combination of respiratory event frequency and the degree of daytime sleepiness patients experienced.³ An AHI under 5 is considered the normal range for adults. On CPAP, the goal is to push your residual AHI down into that under-5 zone, ideally well below it. Many well-treated patients see numbers between 0 and 2 on a typical night. If you are consistently under 5, your therapy is generally considered effective from an AHI standpoint.

How Your CPAP Machine Calculates AHI

In a sleep lab, technicians attach sensors to your chest, face, and finger to measure airflow, breathing effort, and blood oxygen simultaneously. Your CPAP machine has none of that. It estimates events using a single data stream: the airflow signal from the pressure sensor inside the device. Algorithms analyze tiny fluctuations in that flow signal to infer when your breathing slowed or stopped. The machine does not measure oxygen levels and cannot detect brain arousals, so it is making educated guesses based on limited information.

That limitation introduces a gap between what the machine reports and what a lab sleep study would find. A meta-analysis pooling data from multiple CPAP brands found that, on average, CPAP-reported AHI was close to the lab-measured AHI. But the precision around that average was poor, meaning individual readings could be meaningfully off in either direction.⁴ The percentage error was high enough across brands that clinicians are cautioned against relying on the machine number alone for clinical decisions.

One study looking specifically at a single CPAP brand found that the device scored hypopneas at roughly double the rate that manual review by a technician did, while apnea counts were nearly identical. The result was that the machine’s overall AHI read about 1.5 times higher than the manually scored AHI.⁵ This matters because a CPAP showing an AHI of 6 might actually correspond to a lab-scored AHI closer to 4. Conversely, a machine reading of 3 does not guarantee the true value is not somewhat higher.

A real-world study raised an additional concern: roughly 28% of patients whose machine reported a controlled AHI actually had a residual AHI above 5 when tested simultaneously with full polysomnography.⁶ The false reassurance rate was not trivial. If you feel great and your machine shows a low number, this probably doesn’t matter much. But if your AHI looks fine yet you still feel exhausted, the machine might be missing events that a lab study would catch.

Why Different Brands Report Different Numbers

Your CPAP’s AHI is not a universal measurement. It is a proprietary estimate generated by algorithms that differ from manufacturer to manufacturer. Researchers who studied patients switching between two different CPAP brands found a statistically significant difference in the AHI each brand reported, even when treating the same patient under similar conditions.⁷ The difference was large enough to be clinically misleading in some cases. There is currently no industry standard requiring manufacturers to use the same detection rules, thresholds, or scoring criteria.⁸

The practical implication is straightforward: if you switch from one brand to another and your AHI suddenly jumps or drops, do not assume your sleep apnea got worse or better. The change may reflect the algorithm, not your airway. If you are comparing numbers with other CPAP users online, keep in mind that someone using a different brand is not scoring on the same test.

What Drives Your AHI Up on CPAP

Seeing an AHI of 8 or 12 on a morning you expected something low can be frustrating. Several things can push the number up, and not all of them mean your therapy failed.

Mask leaks are one of the most common culprits. When air escapes around the seal, the machine’s flow sensor sees erratic signals that it can misinterpret as respiratory events. A bench study simulating sleep apnea events while adding unintentional air leaks found that some CPAP devices essentially lost the ability to detect obstructive apneas and hypopneas accurately once leaks were present.⁹ Different devices responded differently to the same leak, so the effect depends on both your mask fit and your machine model. If your leak data is high on the same night your AHI spiked, the leak is the first thing to investigate.

Mask type also plays a role. Switching from a nasal mask or nasal pillows to a full-face (oronasal) mask can change your residual AHI. Research has shown that oronasal masks sometimes require higher pressure to achieve the same level of control, and in some patients, residual AHI climbs despite the pressure increase.¹⁰ Nasal pillows, by contrast, tend to produce less air leak and fewer sleep-disrupting side effects.¹¹

Then there are central apneas, which CPAP does not prevent. Some people develop what is known as treatment-emergent central sleep apnea, where central events appear or persist after obstructive events are brought under control.¹² About 3.5% of CPAP patients show central apnea at some point in the first few months of therapy, and for the majority of those, the central events resolve on their own within weeks.¹³ For a smaller group, the central events persist. When machine learning models were built to predict which patients would have high residual AHI on CPAP, the central apnea index turned out to be the single most important factor. In other words, stubborn central events are the biggest driver of treatment-resistant high AHI readings.

Sleep Position and REM Sleep

Your AHI can vary substantially within the same night depending on your body position and sleep stage. Sleeping on your back tends to make the airway more collapsible, and many people with sleep apnea have far more events in the supine position than on their side. A study of 100 men with sleep apnea found that a substantial portion of those who showed positional dependency during lighter sleep stages lost that positional advantage during REM sleep, the deep dreaming stage when muscle tone drops further.¹⁴ About 41% of patients who had strongly positional apnea during non-REM sleep became non-positional during REM.

This explains why some nights register higher AHI than others even with the same CPAP settings. If you have more REM sleep one night, or spend more time on your back, your residual events can climb. A single elevated reading does not necessarily mean anything changed with your equipment or your condition.

Night-to-Night Variability and What It Signals

Speaking of fluctuations, AHI is inherently variable from one night to the next. Some people see readings that bounce between 1 and 6 in the same week. A certain amount of this is noise from the measurement process itself. But the pattern of variability may carry its own health information. Research using large datasets of nightly CPAP data found that high night-to-night variability in AHI was independently associated with uncontrolled high blood pressure, even after accounting for the average severity of the sleep apnea.¹⁵ In other words, someone whose AHI swings widely from night to night may face greater cardiovascular risk than someone whose AHI is steady at the same average level.

For day-to-day monitoring, it makes more sense to watch your trend over a week or a month than to react to any single morning’s number. Most CPAP apps show a rolling average, which smooths out the noise. If that trend starts creeping upward over weeks, that is worth bringing to your sleep doctor. A one-night spike after a stuffy nose or a few glasses of wine is not cause for alarm.

Does Alcohol Make AHI Worse on CPAP?

Many CPAP users assume that drinking alcohol before bed will raise their AHI, and there is a common belief that pressure settings need to increase on drinking nights. The evidence is more nuanced than that assumption suggests. A controlled study looking at moderate alcohol intake in patients ranging from mild to severe sleep apnea found little to no effect on AHI, oxygen saturation, or the pressure needed to control events, either for the full night or just the first two hours after drinking.¹⁶ That study used moderate doses, so heavy drinking may tell a different story. But the idea that a couple of drinks will wreck your CPAP data does not hold up well in the controlled evidence.

What AHI Doesn’t Tell You

AHI counts events, but it does not measure the severity of each event. A three-second dip in oxygen after a brief hypopnea is tallied the same as a 30-second plunge to dangerously low levels. Two people can have identical AHI numbers while experiencing very different levels of physiological stress.

Researchers have increasingly looked at a metric called hypoxic burden, which captures the total amount of oxygen desaturation caused by breathing events rather than simply counting them. Studies found that hypoxic burden predicted cardiovascular death independently of AHI: people in the highest hypoxic burden groups had roughly two to two-and-a-half times the risk of dying from cardiovascular disease, even after adjusting for AHI.¹⁷ A separate analysis showed that this same hypoxic burden metric predicted incident heart failure in men, whereas AHI alone did not.¹⁸

AHI also says nothing about sleep fragmentation. Each breathing event can trigger a brief arousal, a micro-awakening that disrupts sleep architecture without fully waking you. The arousal index, a count of these micro-awakenings per hour, correlates with AHI but explains only about 30% of the variance.¹⁹ Two people with the same AHI can have markedly different arousal patterns and therefore different levels of daytime sleepiness. Sleep disruption and the resulting daytime sleepiness are driven by these recurrent arousals, not by the event count itself.²⁰

Your CPAP machine does not report hypoxic burden, arousal index, or sleep staging. It gives you AHI because that is what it can estimate from airflow data alone. The number is useful as a quick check that therapy is keeping your airway open, but it is a narrow window into what is actually happening during your sleep.

When to Worry and When to Let It Go

A useful rule of thumb for interpreting your morning AHI: consistent averages under 5 with occasional spikes are normal and expected. A trend that stays above 5 over multiple weeks warrants a conversation with your sleep physician, especially if accompanied by returning symptoms like morning headaches, unrefreshing sleep, or excessive daytime tiredness. Check your leak data first, since high leaks are the simplest fix and the most common source of inflated readings.

If your leak data is clean and your AHI is still elevated, the next question is whether the events are obstructive or central. Most CPAP apps and data-reading software distinguish the two. Persistent central events may require a change in therapy mode or device type, and that is a clinical decision, not something to troubleshoot by adjusting your mask at home. For elevated obstructive events, pressure adjustment, weight changes, and positional strategies are all on the table.

One underappreciated point: a low AHI does not automatically mean your treatment is optimized. If you still feel unrested despite excellent numbers, the problem may involve arousals, sleep fragmentation, or other conditions like periodic limb movements or insomnia, none of which show up in your CPAP data. The AHI on your nightstand screen is a useful barometer of airway patency. It was never designed to be a complete report card on your sleep quality, and treating it as one can send you chasing a number while the real issue lies elsewhere.