You did everything right. You wear the mask every night, the app shows an AHI of 3, your machine says therapy is “good” — and you are still exhausted at 3pm. This is one of the most common and most frustrating situations in CPAP therapy, and the explanation is almost always the same: the AHI is not a measure of how well you slept. It is a count of two specific kinds of breathing event, and a low count leaves a great deal of room for poor sleep.
The AHI is not a sleep-quality score
The Apnea-Hypopnea Index counts apneas (airflow essentially stops for ten seconds or more) and hypopneas (airflow drops by a set percentage with a desaturation or arousal), then divides by hours of sleep. That is all it counts. It does not measure how fragmented your sleep was, how much effort you spent breathing, how often you nearly woke, whether you reached deep (N3) and REM sleep, or how oxygenated you stayed between scored events. A machine can drive your AHI to 3 and still leave you sleeping badly — because the things that wrecked your night were never in the count.
There is a second, quieter issue: the AHI your machine reports is its own estimate from airflow and pressure signals, not the EEG-scored AHI from a sleep lab. The two usually track together, but the device cannot see your brain waves, so it cannot tell a genuine event from a moment of wakefulness as reliably as an attended study. We cover that gap in what’s a good AHI on CPAP. For now, take the reported AHI as directional, not gospel.
With that framing, there are four usual reasons you can have a low AHI and still feel terrible.
Cause 1 — residual flow limitation and RERAs
This is the big one, and the most commonly missed. Below the threshold that counts as a hypopnea, your airway can still be partly narrowed — enough that you work harder to breathe and your brain briefly arouses to fix it. These are respiratory effort-related arousals (RERAs), and the underlying airflow restriction is flow limitation. Thirty of these an hour will shred your sleep architecture while your AHI stays low, because none of them meet apnea or hypopnea criteria. Physiologically this is the same picture as upper-airway resistance syndrome (UARS), and it is disproportionately common in slimmer patients, in women, and in people whose original study was scored with conservative hypopnea rules.
The tell is in the data: a flow-limitation graph that stays elevated through the night, even with a low AHI, and an inspiratory flow waveform that looks flattened or “chair-shaped” rather than rounded. If you have never looked at it, that is the first thing to pull up — our explainer on flow-limitation events walks through what it looks like and the “triple peak” patterns people notice on their reports.
An auto-titrating machine helps here, because APAP algorithms are designed to respond to flow limitation by nudging pressure up rather than waiting for a frank event. A unit like the Home Medix HM-CV-20, which runs in APAP mode across 4–20 cmH₂O with EPR, logs a nightly flow-limitation trend you can actually review — so instead of inferring the problem from how you feel, you can see whether airflow restriction is persisting and at what pressures it eases.
Cause 2 — mask leak
A mask that leaks fragments your sleep two ways at once: the noise and air-jet wake you directly, and the leak bleeds off pressure so the therapy under-delivers right when you need it most. Crucially, leak can do all this while the AHI stays low, because the machine is still counting events against the pressure it intended to deliver, not the lower pressure that actually reached your airway.
If your nights are restless and your AHI is fine, the leak graph is the next thing to check — and the number that matters is not the nightly average but the spikes. A perfect seal for six hours and a bad leak for ninety minutes in REM will “average” to an acceptable figure while that 90-minute window quietly destroyed your sleep. We cover what counts as acceptable, and why ResMed and Philips report leak completely differently, in what’s a normal CPAP leak number, and the mechanics of each leak type in CPAP leak types.
Cause 3 — pressure set to the apneas, not the airflow
A fixed CPAP prescription is usually built from the 95th-percentile pressure that suppressed apneas during titration. That pressure can be entirely correct for apneas and still sit a centimetre or two below what it takes to relieve flow limitation during REM or supine sleep. The result is exactly the picture in Cause 1: AHI under 5, flow limitation persisting, you tired.
The fix is not always “more pressure.” Sometimes it is a higher minimum on an auto range so you are not starting each night below the effective floor; sometimes it is a wider range so the algorithm can chase REM-supine events; sometimes the EPR setting is undermining the splint and needs trimming. These are titration questions, covered in CPAP pressure titration, and they belong with your physician and your download data — not with the menu on the machine at 2am.
Cause 4 — it is not your CPAP at all
Sometimes the apnea really is controlled and the tiredness is coming from somewhere else entirely. This is the category people forget, and it is large:
- Short sleep. Six hours of well-treated apnea is still six hours. CPAP cannot manufacture sleep you did not give yourself the time for.
- Periodic limb movements (PLMS), which arouse you dozens of times an hour independently of breathing and are invisible to your CPAP.
- Depression, hypothyroidism, anaemia, and uncontrolled diabetes — all classic, common causes of daytime fatigue that coexist with sleep apnea.
- Alcohol or sedatives, which fragment deep sleep even when breathing is fully supported.
- Circadian misalignment from shift work or an irregular schedule — your apnea can be perfect and your body clock still wrong.
- Other sleep disorders, including narcolepsy and idiopathic hypersomnia, which are sometimes only unmasked once the apnea is treated and the tiredness stubbornly remains.
A patient whose AHI is genuinely under 2, whose leak and flow-limitation graphs are clean, and who is still exhausted needs a fatigue work-up — thyroid panel, ferritin, mood screen, a look at total sleep time and medications — not another pressure change.
A worked example
A 38-year-old woman, BMI 24, diagnosed with “mild” OSA (AHI 8), is put on fixed CPAP at 8 cmH₂O. Her machine reports an AHI of 2 and she feels no better after two months. The summary looks like success. The graphs do not: flow limitation is elevated for most of the night, and the events that remain cluster in REM. She is a textbook flow-limitation/UARS picture — the apneas were never the main driver of her symptoms. Moving to an auto range with a higher floor, which lets the pressure rise to flatten the flow limitation, is what finally helps. Nothing in the headline AHI would ever have told her that.
What to actually do — and what to bring your physician
- Pull the detailed data, not just the app’s summary score. myAir, AirView, Care Orchestrator, or OSCAR all show the graphs — see reading your CPAP report.
- Look at three things, in order: flow limitation, leak, then residual AHI broken down by event type (obstructive vs central vs hypopnea).
- Confirm your total sleep time is genuinely adequate for at least a fortnight before blaming the machine.
- Bring a specific sentence to the appointment. “My AHI is 3 but I feel terrible, and the flow-limitation graph is high all night, worst in REM” is a far more useful thing to say than “the machine says I’m fine but I’m not.” It points your physician straight at the data that matters.
Takeaway
A low AHI tells you the apneas are controlled. It does not tell you that you slept well. When tiredness persists despite good numbers, the answer is almost always in the data the AHI leaves out — flow limitation, RERAs, and leak — or in a cause outside sleep apnea altogether. Read the graphs, not just the headline, and if they are clean, look beyond the CPAP.
Consult your sleep physician before changing pressure or settings; persistent sleepiness on well-controlled therapy is a clinical finding that deserves a proper look, not a self-adjustment. (AASM Practice Guidelines)