Mask Off & Large-Leak Events: Why Your CPAP Report Shows Gaps

If you've ever pulled up your CPAP report and seen a stretch where the numbers just look... off — a flat line, a missing chunk, or an AHI that seems suspiciously good — you may be looking at a mask-off or large-leak segment. During those periods, your machine simply can't measure your breathing accurately, which means the events it reports (and the ones it misses) can paint a misleading picture. This guide explains what's happening, why it matters, and how to tell a trustworthy night from a corrupted one.

What mask-off and large-leak events are

Your CPAP works by pushing a steady stream of pressurized air through a sealed mask. It also measures the air flowing in and out of that circuit to estimate your breathing and score events like apneas and hypopneas. That measurement only works when the seal is intact.

Two related situations break that seal:

• Mask-off events. You take the mask off (often without fully waking) or it slips completely off your face. The machine sees a sudden, massive air escape and flags the period as "mask off" or simply stops logging meaningful data.
• Large-leak events. The mask is still on, but air is escaping faster than the device expects — through a poorly fitted cushion, an over-tightened or under-tightened strap, a shifted nasal pillow, or your own mouth leak (air venting out through your lips when you use a nasal mask). This is sometimes called "large leak," "leak alert," or a period above the device's leak threshold.

Every manufacturer defines "too much" leak a little differently, and it helps to know the baseline your machine reports:

Because ResMed subtracts the intentional vent and Philips doesn't, you can't compare their leak numbers head-to-head. For a deeper breakdown of acceptable ranges, see CPAP Leak Rate: What's Acceptable and How to Fix High Leaks.

Why the machine can't score events through a big leak

To call an apnea or hypopnea, your CPAP watches for a drop in airflow — a near-stop (apnea) or a partial reduction (hypopnea) lasting at least 10 seconds. It infers all of this from the pressure and flow inside the sealed circuit.

When a large leak opens up, that math falls apart:

• The flow signal is contaminated. The device can no longer tell how much air went into you versus how much escaped out the leak. A reduction in your airflow (a real event) can be masked by air rushing out the gap, or a leak fluctuation can mimic normal breathing.
• The reference baseline shifts. Event scoring compares current breathing to a recent baseline. A big leak distorts that baseline, so the algorithm loses its yardstick.
• The device protects against false alarms. Rather than guess, most machines suppress or down-weight event scoring during high-leak periods. That's a reasonable design choice — but it means real events during the leak may never be counted.

It's worth remembering that even on a perfectly sealed night, your device's AHI is an estimate. Home CPAP machines have no EEG, so they can't see the arousals that a sleep lab uses to score some hypopneas under the AASM's recommended criteria (a 3% oxygen drop or an arousal). Lab studies often use a 4% desaturation rule instead. So device-reported AHI already differs from lab-scored AHI — and a large leak widens that gap considerably. For the fuller picture, see How to Read Your CPAP Data and CPAP Event Types Decoded.

How this falsely lowers AHI

Here's the counterintuitive part that trips up a lot of people: a leaky night often produces a lower AHI, not a higher one — and that low number is the unreliable one.

AHI is events divided by hours of sleep. During a large leak or mask-off stretch, the machine stops reliably scoring events, so fewer events get counted. But you're still potentially having apneas and hypopneas — they're just invisible to the device. The result:

1. Real events during the leak go uncounted.
2. The event total drops.
3. AHI is calculated against time that includes the corrupted period.
4. You see a falsely reassuring number.

So a large leak doesn't just add noise — it can invalidate or under-report your AHI entirely. This is exactly why a "great" AHI on a night with heavy leak deserves a skeptical eye, while a genuinely sealed night with a low AHI is something to celebrate.

Leak corruption isn't limited to AHI, either. It also degrades flow-limitation data — the subtle airflow-shape changes that signal your airway is narrowing before a full event. Flow limitation is one of the most useful early-warning metrics CPAP offers, and a big leak smears it out. If that metric matters to you (it should — it's the hidden signal behind residual symptoms), see CPAP Flow Limitation: The Hidden Metric Beyond AHI.

A quick framing that pays off over time: a single night is noise; trends over weeks are signal. One leaky night with a fake-low AHI won't ruin your understanding of therapy — but if you don't recognize it as unreliable, it can quietly drag your averages and hide a real problem. Night-to-night swings are normal, and Why Does My AHI Change Night to Night? puts that variation in context.

Spotting unreliable segments in your data

The goal is simple: learn to recognize when a number isn't trustworthy, so you don't act on a falsely low AHI. Here's what to look for.

On the leak graph

• A sustained spike that climbs above your machine's threshold (24 L/min excess leak for ResMed; a higher total-leak figure for Philips) and stays there.
• Long plateaus rather than brief blips — short spikes from a position change are usually harmless.

On the flow rate waveform

• Flat or chaotic stretches that don't look like the rhythmic rise-and-fall of normal breathing.
• A clean, "too quiet" section right where the leak graph spikes — that's often a mask-off period.

Learning to read this waveform is the single most powerful skill for trusting your own data; start with How to Read the CPAP Flow Rate Waveform.

On the event timeline

• Suspicious gaps — long event-free zones that line up exactly with the leak spike.
• Clusters of events bunched right at the edges of a leak period (events the machine caught just before and after losing the seal).

The shortcut: let software flag it for you

Cross-referencing three graphs by eye, night after night, is tedious — and easy to get wrong. This is where automatic analysis earns its keep. SomniCharts marks the unreliable, leak-corrupted segments of your night for you, so you're not tempted to trust a falsely low AHI that the machine couldn't actually measure. It imports ResMed, Philips Respironics (including the encrypted DreamStation 2, which OSCAR and most third-party tools can't read), and Löwenstein prisma data, then explains all of it in plain language — including which parts of the night to trust. That "is this number real?" layer is exactly the kind of thing a raw report leaves you to figure out alone. For how the major tools compare on what they actually detect, see OSCAR vs SleepHQ vs SomniCharts.

Fixing the underlying leak

The good news: fixing the leak restores reliable scoring. Once the seal holds, your machine can measure your breathing again, and your AHI and flow-limitation data become trustworthy. Leak control is the highest-leverage adjustment most CPAP users can make on their own — and unlike pressure, it's squarely in your hands.

Common causes and fixes:

• Worn or wrong-size cushion. Mask cushions degrade; silicone loses its seal over months. Replace on schedule and re-check sizing — a too-large cushion leaks as readily as a too-small one.
• Strap tension. Counterintuitively, over-tightening often creates leaks by deforming the cushion. Aim for "snug, not crushing."
• Mouth leak on a nasal mask. If air escapes through your lips, options include a chin strap, switching to a full-face mask, or addressing nasal congestion that pushes you to mouth-breathe. See CPAP Dry Mouth and Mouth Breathing and CPAP Nasal Congestion and Stuffy Nose.
• Mask style mismatch. Sleepers who move a lot, or who breathe through their mouth, often seal better with a different mask type. CPAP Mask Types and Fit walks through the trade-offs.
• Positional leaks. A mask that seals on your back but breaks when you turn may need a different frame or a softer pillow setup.

For a step-by-step troubleshooting walkthrough, CPAP Mask Leaks & Mouth Leak: Causes, the Data Threshold, and Fixes is the companion guide. This article is part of our broader Troubleshooting & Optimizing CPAP pillar.

One important boundary

Use your leak data to make mechanical fixes — mask, cushion, straps, humidity — freely. But if you've sealed the leak and your AHI is still high, that's a conversation for your provider, not a reason to change your prescribed pressure on your own. The right move is to bring your trends to your sleep clinician so they can interpret them. A residual AHI below 5 events per hour is the widely used benchmark for effective therapy (the AASM defines an "optimal" titration as fewer than 5 events per hour); some clinicians aim lower when it's comfortably achievable, but goals are individualized. If your AHI stays elevated after the leak is fixed, see Why Is My AHI High on CPAP? and Can I Adjust My Own CPAP Pressure? for what the data can — and can't — tell you.

Frequently asked questions

Does a high leak mean my therapy isn't working? Not necessarily — but it means you can't trust the night's data to tell you whether it's working. Fix the leak, then read the next few nights to see your real numbers.

Why is my AHI lower on nights with bad leaks? Because the machine stops reliably scoring events during a large leak, so fewer events get counted. The low number is an artifact, not an improvement.

Is a brief leak spike a problem? Usually not. Short spikes from rolling over or repositioning are normal. It's the sustained periods above your machine's threshold that corrupt your data.

What leak number is too high? On ResMed, excess leak above 24 L/min at the 95th percentile is the common flag. Philips reports total leak (which includes the intentional vent), so its raw numbers run higher and should be read against Philips' own guidance. See CPAP Leak Rate: What's Acceptable.

Should I raise my pressure to stop the leaks? No — raising pressure tends to increase leak, not reduce it, and pressure changes belong with your clinician. Address the seal mechanically first.