What Do Events Mean on a CPAP Machine?

Events on a CPAP machine are breathing disturbances the device detects and records while you sleep. Each time your airflow drops, pauses, or takes on an abnormal pattern, the machine logs it as an “event” and tallies those events into a summary number, usually displayed as an AHI (apnea-hypopnea index) on your machine’s screen or companion app. That number tells you roughly how many times per hour your breathing was disrupted despite the pressurized air. For most users, it is the single most important piece of feedback the machine provides, but the details behind it are more nuanced than the headline number suggests.

The AHI and What It Counts

The AHI your CPAP reports is a count of apneas and hypopneas per hour of use. An apnea is a near-complete pause in airflow, typically lasting at least ten seconds. A hypopnea is a partial reduction in airflow, usually scored when breathing drops by about 30% or more and is followed by a few larger recovery breaths.¹ The machine adds these together and divides by total hours of recorded use to give you a per-hour rate.

Most clinicians and CPAP software use a threshold of five events per hour as the cutoff for “well controlled.” An AHI at or below five means your therapy is keeping most breathing disturbances in check. Above five, something may need adjustment. That said, an AHI of 2 is not meaningfully different from an AHI of 3 on a night-to-night basis. The number bounces around depending on sleep position, nasal congestion, and how much time you actually spent asleep versus just lying in bed with the mask on.

Types of Events Your Machine Reports

Most modern CPAP devices break events into several categories, and understanding what each one means can help you and your sleep clinician figure out what is actually going on.

• Obstructive apneas (OA): Your airway physically collapses or narrows enough to block airflow while your body keeps trying to breathe. The machine infers this by detecting that airflow stops but effort continues, usually by analyzing pressure fluctuations in the tubing.
• Central apneas (CA): Your brain temporarily stops sending the signal to breathe, so airflow ceases and there is no respiratory effort at all. The machine distinguishes these from obstructive events by checking whether the airway appears open or closed during the pause.
• Hypopneas (H): Breathing does not stop completely but drops substantially for at least ten seconds. These are the most common events on a typical CPAP report.
• Flow limitations (FL): The inspiratory airflow curve flattens rather than maintaining its normal rounded shape. This suggests the airway is partially narrowing but not enough to meet the threshold for a scored hypopnea. Auto-adjusting machines use flow limitation data to decide when to bump up pressure.

The distinction between obstructive and central apneas matters because the treatment strategies differ. A current-generation device detects an obstructed airway correctly in about 87% of cases when compared to a simultaneous in-lab sleep study.² That is good but not perfect, and the researchers who tested it cautioned that device-detected airway status should not be equated with a formal diagnosis of a specific respiratory event type.

How the Machine Detects Events

Your CPAP does not have the sensors a sleep lab uses. In a formal polysomnography (PSG) study, technicians monitor brain waves, eye movements, muscle tone, chest and abdominal effort belts, oxygen levels, and airflow through a nasal cannula. Your CPAP has exactly one primary input: the airflow signal it estimates from the pressure and flow in its tubing circuit. Everything else is inferred from that single stream of data.

When the machine senses that airflow has dropped below a set threshold for a set duration, it scores an event. When it senses flattening of the inspiratory flow contour, it flags flow limitation. Early research showed that the shape of the inspiratory flow curve reliably reflects upper-airway resistance: a rounded contour means the airway is wide open, while a plateau or flattened shape signals partial narrowing.³ This same principle is what auto-CPAP devices use to decide in real time whether to raise or lower pressure.⁴

To distinguish obstructive from central apneas, some machines apply small pressure oscillations through the tubing and measure how the airway responds. If the airway is open, the oscillations pass through easily; if it is collapsed, they meet resistance. This technique, borrowed from a research method called forced oscillation, allows the device to infer whether you were trying to breathe against a blocked airway or simply not breathing at all.⁵

How Accurate Is Your Machine’s AHI?

This is where things get interesting and a bit messy. The AHI your CPAP displays is not the same metric as the AHI measured during a sleep study, even though they share a name. Manufacturers use their own proprietary algorithms and their own definitions for what counts as an apnea or hypopnea, and those definitions differ from the standard scoring rules used in sleep labs and even from each other.⁶ A meta-analysis examining CPAP-reported event detection confirmed this lack of uniformity, finding that scoring criteria varied both among manufacturers and among the sleep lab studies used for comparison.⁷

In practical terms, the correlation between device AHI and PSG AHI ranges from very good to surprisingly poor depending on the study and the device. One study of 50 patients found a strong correlation of 0.97 between the two measures.⁸ A larger study of 163 patients, however, found no statistically significant correlation and noted a substantial rate of false positives, where the device flagged an AHI above five but the simultaneous sleep study showed the patient was actually well controlled.⁹ In that study, about 81% of patients whose sleep study AHI was under five also had a device AHI under five, so the machine was fairly good at confirming good control. But when the sleep study showed an AHI above five, the device only caught it about 39% of the time.

What this means for you: if your CPAP says your AHI is low, you can feel reasonably confident that things are going well. If it says your AHI is elevated, it is worth investigating but not necessarily cause for alarm, because the machine may be overcounting. And if you switch CPAP brands, do not be surprised if your AHI shifts noticeably. Research has shown that the reported AHI is brand-specific, and clinicians have been warned that jumping from one manufacturer to another can produce clinically misleading changes in the number.¹⁰

How Leaks Mess With Your Data

Mask leak is the single biggest confounder of CPAP event data. When air escapes around the edges of your mask or through your mouth, the machine loses its ability to accurately read airflow. In some cases, leak-related turbulence can be misinterpreted as respiratory events, inflating your AHI. In others, a large enough leak prevents the machine from detecting real events because the airflow signal is too noisy to analyze.

One recent study found that excluding events that occurred during periods of major leak or during the ramp-up period at the start of the night reduced the number of flagged alerts by about 20%.¹¹ That is a significant chunk of events that may not represent genuine breathing disturbances at all.

Mouth leak specifically is a common problem for nasal mask users. Research has shown that mouth leak episodes are often associated with arousals from sleep and that roughly a third of leak episodes end with a respiratory event.¹² The relationship is bidirectional: leaks can cause arousals that trigger breathing disturbances, and breathing disturbances can cause jaw relaxation that opens the mouth and creates leak. Heated humidification helps reduce the drying effect of mouth leak and blunts some of its impact on delivered airflow.¹³ A chin strap or switching to a full-face mask are the other standard fixes.

When Central Apneas Appear on Your Report

Seeing a handful of central apneas on your nightly data is usually nothing to worry about. A few centrals per hour occur in healthy sleepers as a normal part of sleep-stage transitions. But if your central apnea index climbs above five per hour, or if centrals make up the bulk of your total events, something else may be going on.

Treatment-emergent central sleep apnea (TECSA) is a well-recognized phenomenon where central apneas appear or increase after you start CPAP therapy for obstructive sleep apnea. It shows up in roughly 8% of patients during initial CPAP titration studies.¹⁴ The mechanisms involve ventilatory instability: once the obstructive component is eliminated by CPAP, the brain’s breathing control system overshoots and undershoots before finding a new equilibrium. In most people, TECSA resolves on its own within a few weeks to months of continued use.¹⁵ In a minority, however, it persists and may require a different type of device, such as adaptive servo-ventilation.

Research tracking TECSA trajectories over time found that patients who developed it had higher residual event rates and more leak at 13 weeks, and they were also at greater risk of abandoning therapy altogether.¹⁶ So if your device keeps showing elevated central apneas beyond the first month or two, bringing it up with your sleep clinician is worthwhile rather than just hoping the numbers will drop.

Cheyne-Stokes Respiration as a Warning Sign

Some newer CPAP devices report a pattern called Cheyne-Stokes respiration (CSR), which shows up as a distinctive crescendo-decrescendo cycling of breathing depth. On your data, it may appear as a separate metric or as a periodic breathing percentage. This pattern is worth paying attention to because it can signal heart failure or other cardiovascular problems.

A systematic review found that CSR detected by CPAP was strongly associated with serious cardiac events, with an adjusted odds ratio of 5.74. Longer cycle lengths above about 69 seconds were particularly accurate at discriminating heart-failure-related CSR from other causes, and day-to-day instability in CSR burden was highly predictive of acute heart failure worsening.¹⁷ A prospective study tracking CPAP-treated patients over two years similarly found that more nights with CSR and a higher percentage of CSR were linked to increased cardiovascular events.¹⁸

Before you panic, though, there is a caveat. Device-based CSR detection relies entirely on the airflow signal and can produce false positives. A case report described a patient whose CPAP flagged CSR that turned out to be an artifact caused by nasal obstruction and unstable airflow patterns, not actual periodic breathing. Once the nasal obstruction was treated, the CSR signal disappeared.¹⁹ The takeaway: if your device starts flagging CSR, it is a reason to talk with your doctor, not to self-diagnose heart failure. But it is also not a signal to ignore.

Why Auto-CPAP Machines Do Not All Behave the Same

If you use an auto-adjusting (APAP) machine, the events on your report are shaped not only by your actual breathing but also by how aggressively or conservatively your machine’s algorithm responds to disturbances. A bench test comparing several APAP devices found enormous variation in their behavior. Some devices normalized breathing quickly but overshot the needed pressure, while others failed to eliminate prolonged flow limitation at all. The time to reach maximum pressure ranged from about 4 minutes to 96 minutes, and average pressures varied by as much as 7 cmH₂O across devices responding to the same simulated breathing pattern.²⁰

A head-to-head comparison of two APAP devices in actual patients showed that one left about 14% of patients with a residual AHI above five, while the other left about 62% above that threshold.²¹ Both machines were treating the same patients, with the same disease. The difference was purely algorithmic. This means the event count on your report partly reflects which machine you own, not just how your airway behaved.

The Scoring Rules Are Not Universal Either

Even in sleep labs, what counts as a hypopnea depends on which scoring rule is applied. The American Academy of Sleep Medicine offers two definitions: one requires a 4% or greater oxygen desaturation, while the other requires only a 3% desaturation or an arousal from sleep.²² The more sensitive rule naturally produces a higher event count for the same night of breathing. Your CPAP manufacturer chose one approach or another (or invented their own variation), and that choice is baked into every number the machine reports.

This lack of standardization is one reason sleep specialists tend to look at trends over time rather than obsessing over any single night’s number. If your AHI has been 2-3 for months and suddenly jumps to 8-9 for several nights running, that shift is meaningful regardless of which scoring definition the machine uses. A one-night spike to 7 that drops back to 2 the next morning is almost certainly noise, possibly a night of unusually poor mask fit, nasal congestion, or sleeping on your back more than usual.

What to Do When Your Events Are High

The American Academy of Sleep Medicine recommends ongoing monitoring of objective efficacy and usage data, with troubleshooting as needed, both after starting therapy and throughout treatment.²³ In practice, that means checking your data regularly and acting on persistent changes, not just looking at the number once a year.

If your AHI starts creeping up, a reasonable troubleshooting sequence looks like this:

• Check leak first: A loose mask, a worn-out cushion, or mouth breathing with a nasal mask are the most common culprits. High leak rates can inflate your event numbers artificially and also reduce the machine’s ability to maintain effective pressure.
• Look at the event breakdown: If events are mostly obstructive apneas or hypopneas, your pressure may need to go up or your auto-CPAP range may need widening. If events are mostly central, review whether you have treatment-emergent centrals that might resolve with time or need a different device.
• Consider positional factors: Some people have significantly worse breathing in the supine position. Research has found that people with positional sleep apnea tend to need slightly lower CPAP pressures overall but still see more events when sleeping on their back.²⁴
• Look for new symptoms: Persistent CSR or a rising central apnea index in someone with heart disease risk factors warrants a cardiology conversation. New-onset congestion or weight gain can change your pressure needs.

Engagement with your data has a real payoff. Studies on patient engagement technology for CPAP have found that active monitoring and feedback is associated with meaningfully higher adherence rates compared to standard care.²⁵ Paying attention to your events is not just satisfying a curiosity. It is part of making the therapy work.

Alcohol, Medications, and Other Factors People Ask About

A common worry is that drinking alcohol before bed will wreck your CPAP data. The evidence is more reassuring than most people expect. A controlled study of patients with mild-to-severe obstructive sleep apnea found that a moderate dose of vodka had little effect on breathing or oxygen levels during sleep and no effect on the pressure required to keep the airway open.²⁶ Heavy drinking is a different story and was not tested in that study, but a glass or two is unlikely to blow up your event count if your pressure settings are appropriate.

Sedating medications, on the other hand, can increase the frequency of central events by depressing respiratory drive. Weight gain raises the event count by narrowing the airway and increasing the pressure needed to keep it open. Nasal congestion from allergies or illness reduces airflow through the nose and can trigger both mouth leak and flow limitation events. All of these show up in your data as an elevated AHI, and all of them have different solutions, which is why looking at the event breakdown rather than just the headline number matters.

Using CPAP Software for Detailed Review

The summary screen on your machine or app gives you a single AHI and sometimes a breakdown by event type. Third-party software tools, most famously the open-source program OSCAR, can import the raw data from your machine’s SD card and display individual events on a timeline alongside pressure, leak, and flow data. This level of detail is not necessary for everyone, but it is useful when something looks off in the summary numbers.

For example, if your AHI is 6 but the detailed data shows that all the events cluster in a 20-minute window early in the night during a period of high leak, the problem is probably a mask issue at sleep onset rather than a therapy failure. Similarly, seeing periodic breathing patterns in the flow data that correlate with a CSR flag can help distinguish genuine periodic breathing from artifact caused by congestion or mask fit issues. The AASM’s guidance on follow-up monitoring implicitly supports this kind of granular review by emphasizing both objective efficacy data and troubleshooting as core components of ongoing care.

The one thing to keep in mind with detailed data review is that it is easy to over-interpret. Every healthy person has brief breathing pauses during sleep, and no CPAP machine achieves a zero event rate every night. A consistent pattern over many nights means something. A single weird night usually does not.