CPAP Leak Rate: What It Means and When to Worry

Leak rate measures the volume of air escaping your CPAP circuit that isn't part of intentional, designed airflow. Your machine records it continuously throughout a session and expresses it in litres per minute (L/min).

The key word is unintentional. Every PAP mask has exhaust vents built in — small holes or ports designed to flush exhaled CO₂ out of the mask. This is called intentional leak, and it's supposed to be there. Without it, you'd rebreathe your own exhaled air.

What your machine flags is unintentionalleak: air escaping from places it shouldn't — around the mask cushion, at connection points, or through an open mouth on a nasal-only mask.

This distinction trips people up constantly, so it's worth spelling out clearly.

Modern ResMed machines report both figures. Older machines or third-party devices may only report total leak. Knowing which one you're reading makes a significant difference in how you interpret the number.

The intentional vent leak rate varies by mask model and therapy pressure. A full-face mask at 10 cmH₂O might vent around 24 L/min as a baseline. A nasal pillow mask at the same pressure might vent around 18 L/min. ResMed publishes vent flow curves for each mask, and OSCAR uses these curves to calculate the unintentional (residual) leak from the raw total leak data your machine records.

If OSCAR's leak graph shows a fairly steady baseline with occasional spikes above it, the baseline is your intentional vent leak and the spikes are unintentional leak events.

There's no universal "normal" that applies to every mask and machine combination, but some context helps:

• ResMed's "Large Leak" threshold in myAir is typically unintentional leak ≥ 24 L/min sustained over a meaningful portion of the session.
• OSCAR's leak display shows the 95th percentile (L/min), median, and maximum for the session. A lower 95th percentile is generally better — it means leak events were infrequent and brief.
• A well-sealed session on a full-face mask might show unintentional leak near 0 L/min for most of the night, with small spikes during positional shifts.
• Sessions with significant mouth breathing on a nasal-only mask will show elevated unintentional leak, because exhaled air is escaping through the mouth rather than through the circuit.

What matters most isn't a single number in isolation — it's whether leak rate is consistent and low, and how often it spikes during the night.

The "Large Leak" flag on ResMed machines (and similar alerts on other devices) is a data-quality signal, not a diagnosis.

When unintentional leak is high, the machine's ability to accurately detect apnoeas and hypopnoeas is compromised. Pressure algorithms may respond to air rushing out of the mask in a similar way to a respiratory event, potentially triggering unnecessary pressure adjustments. Some real events may also be harder to detect against the noise of a large leak.

This is why your AHI on a high-leak night should be interpreted with caution. The figure may be less reliable than on a low-leak night — not because your breathing was better or worse, but because the underlying signal quality was lower.

OSCAR is the primary open-source tool for detailed PAP data analysis, and it handles leak data well. Here's how to read it:

1. 1Import your CPAP data — from your SD card or ResMed data folder.
2. 2Open the daily view and look at the "Leak Rate" chart. By default, OSCAR displays unintentional (residual) leak after subtracting the vent flow curve for your detected mask model.
3. 3Check the statistics panel on the left for the 95th percentile, median, and maximum values for the session.
4. 4Overlay with AHI or flow limitation charts — look for correlations between leak rate spikes and event clusters. Clusters of events that coincide with high-leak episodes are worth noting and discussing with your clinician.

One thing to verify: confirm OSCAR has detected the correct mask model. The vent flow curve used to calculate unintentional leak depends on the specific mask. If OSCAR is using the wrong mask profile, the residual leak calculation will be off. Set your mask manually in OSCAR's settings if the auto-detected mask doesn't match what you use.

AirwayLab displays your leak rate data in the analysis view alongside your AHI, flow limitation, and breathing pattern charts. Upload your SD card data at /analyze — your data never leaves your browser.

Because AirwayLab is open-source (GPL-3.0), you can verify exactly how leak rate is calculated and displayed. There's nothing opaque in the analysis.

Also worth reading:

• How to read your CPAP data
• How to export and understand your CPAP data

Leak rate is one input among many. A single high-leak night doesn't necessarily mean something is wrong long-term. Positional shifts, nasal congestion, illness, or even a different pillow can temporarily affect seal quality.

What's more informative is the pattern over time — whether high-leak nights cluster, whether they correlate with worse-feeling mornings, and how leak trends have changed since starting therapy or switching equipment.

This is the kind of pattern your clinician or sleep specialist is best placed to interpret in the context of your full therapy history and clinical picture. Your clinician can review your OSCAR or AirwayLab data at your next appointment if you choose to share it. You can generate a session summary at /analyze to share.