Your AHI is under 5. Your sleep app says you're doing great. So why are you still dragging yourself through the day, reaching for caffeine by noon, and wondering whether your PAP machine is actually doing anything? You're not imagining it, and you're not alone. A meaningful subset of treated sleep apnea patients report persistent fatigue despite "normal" AHI numbers.
What Your AHI Actually Counts (and Misses)
Your CPAP's AHI is built from two AASM-defined event types, each with hard thresholds. An apneais a ≥90% drop in airflow lasting at least 10 seconds. Under the AASM 2012 recommended rule, a hypopneais a ≥30% airflow reduction lasting ≥10 seconds with either a ≥3% oxygen desaturation or an arousal; an alternative rule used by some labs and payers requires a ≥4% desaturation and does not count arousals. Everything below those thresholds — including the partial narrowing behind flow limitation and RERAs — is structurally excluded, not overlooked.
A third event type, the RERA (Respiratory Effort-Related Arousal), sits below the hypopnea threshold: a sequence of breaths lasting ≥10 seconds with increasing respiratory effort or flattening of the inspiratory flow shape, leading to an arousal, that does not meet apnea or hypopnea criteria. Add RERAs to AHI and you get the RDI (Respiratory Disturbance Index) — the number your sleep study may have generated, but your CPAP almost never reports. For a full breakdown of RERA scoring, see What Are RERAs?. The broader symptom cluster these patterns can produce is described in What Is UARS?
| Event | Threshold | Duration | What it requires |
|---|---|---|---|
| Apnea | ≥90% airflow drop | ≥10 seconds | Threshold only |
| Hypopnea | ≥30% airflow drop | ≥10 seconds | ≥3% desaturation or an arousal (recommended rule) |
| RERA | Below hypopnea threshold | ≥10 seconds | Rising effort or flattened flow shape, ending in an arousal |
| RDI | — | — | Apneas + hypopneas + RERAs |
Some labs and payers use an alternative hypopnea rule requiring a ≥4% desaturation and not counting arousals.
Why a Normal AHI Doesn't Mean Normal Sleep
The Apnea-Hypopnea Index counts apneas (complete breathing stops) and hypopneas (partial reductions with oxygen drops). If your AHI is 2, it means your airway fully or mostly collapsed about twice per hour. That sounds fine.
But AHI has a major blind spot: flow limitation. This is partial airway narrowing that restricts airflow without triggering a scored event. Your airway narrows enough to flatten your breathing waveform, increase respiratory effort, and fragment your sleep, but not enough for AHI to notice.
Think of it this way: AHI counts when the pipe is almost fully blocked. Flow limitation is when the pipe is half-squeezed all night long. The pipe never fully closes, so AHI stays low, but the effort to breathe through a constricted airway is exhausting your nervous system.
What's Actually Happening While You Sleep
Research from Dr. Avram Gold and others has identified several mechanisms that disrupt sleep quality without showing up in AHI:
RERAs (Respiratory Effort-Related Arousals)
Sequences of flow-limited breaths that end in a micro-arousal. Your brain briefly wakes to restore airflow, then falls back asleep. You don't remember it, but your sleep architecture is fragmented. RERAs are not counted in AHI.
Autonomic Stress Response
Flow limitation can activate your body's fight-or-flight response via the limbic system, even without a cortical arousal. Your heart rate spikes, blood pressure rises, and stress hormones are released, all while you appear to be sleeping normally.
Sleep Architecture Disruption
Even without full arousals, flow limitation can shift your sleep between stages, reducing the deep and REM sleep your body needs for restoration. The result is hours of "sleep" that leaves you unrefreshed.
This cluster of symptoms has a clinical name: Upper Airway Resistance Syndrome (UARS). It's characterised by significant flow limitation and symptoms despite a normal AHI.
Signs Your AHI Might Be Missing the Problem
If several of these resonate, flow limitation may be worth investigating:
- Your AHI is consistently low (under 5), but you never feel rested
- You feel worse in the second half of the night or wake up feeling like you barely slept
- Morning headaches, jaw tension, or a dry mouth that your current settings haven't resolved
- Your Epworth Sleepiness Scale score is normal, but your fatigue is real (ESS measures sleepiness, not fatigue)
- Your sleep physician says "your numbers look great" but you don't feel great
- Brain fog, difficulty concentrating, or cognitive symptoms that PAP hasn't fixed
How to See What AHI Is Hiding
Your PAP machine's SD card contains breath-by-breath flow waveform data from every night. It records far more than what the myAir or DreamMapper app shows you. With the right analysis, this raw data reveals the flow limitation patterns AHI ignores.
Here's what to look for:
Glasgow Index (Breath Shape)
Scores how distorted your breathing waveform is across 9 shape characteristics. A score above 2.0 suggests significant residual flow limitation, even with a low AHI. This is the single most informative metric for detecting undertreated airway resistance.
FL Score (Flow Shape Flattening)
Measures how flat-topped your inspiratory flow shape is. It's an uncalibrated relative index, not a percentage of breaths. The trend across your own nights carries more information than the absolute number; your clinician can help interpret it in context.
NED + RERA Estimate
Detects per-breath flow limitation and identifies RERA-like events, the arousal sequences AHI misses entirely. The estimated RERA Index combined with AHI gives you something closer to the true Respiratory Disturbance Index (RDI).
How AirwayLab Flags a RERA Candidate
AirwayLab's engine looks for runs of 3 to 15 consecutive flow-limited breaths — breaths where the flow waveform shows either a sharp mid-breath drop after the initial peak (NED above 20%) or a plateaued, flattened top (flatness index at or above 0.85). A run qualifies as a RERA candidate if the NED severity climbs across the sequence, if it ends with a clear recovery breath followed by a deeper "sigh" breath, or if NED inside the sequence peaks above 34%.
These are flow-based estimates, not scored RERAs — formal RERA scoring requires the EEG-confirmed arousals of a laboratory sleep study, which SD card data does not contain.
When to Bring This to Your Clinician
Data is most useful when it informs a conversation with your sleep physician. Consider requesting a review if:
- Your flow limitation metrics are consistently elevated (Glasgow above 2.0, FL Score persistently above 70 -- an uncalibrated relative index, so a rising trend matters more than the absolute number -- or high RERA count) despite low AHI
- You notice a pattern of worsening metrics in the second half of the night (the H2 split), which often correlates with REM-related airway narrowing
- Your symptoms persist after 3+ months of compliant PAP use with no improvement
AirwayLab provides detailed reports you can export as PDF, CSV, or a formatted forum post. Objective data makes it easier for your clinician to evaluate whether a pressure adjustment, mode change (e.g. BiPAP), or further investigation is warranted.
Further Reading
Understanding Flow Limitation: What Your PAP Machine Doesn't Tell You — a deeper look at what flow limitation is and why it matters.
Beyond AHI: Why Your Sleep Apnea Score Might Be Misleading You — the research case against relying on AHI alone.
Arousals Don't Tell the Whole Story — why flow limitation may matter more than cortical arousals.
Does Flow Limitation Drive Sleepiness? — evidence linking flow limitation directly to daytime symptoms.
References
Berry et al. (2012). "Rules for Scoring Respiratory Events in Sleep: Update of the 2007 AASM Manual." Journal of Clinical Sleep Medicine, 8(5), 597–619.
A note on self-analysis
AirwayLab helps you understand your PAP data, but it is not a diagnostic tool. Flow limitation analysis from SD card data is an estimate, not a polysomnography-grade measurement. Always discuss therapy changes with your sleep physician. The metrics provided are for educational purposes and to inform clinical conversations.
See What Your AHI Is Missing
Load your ResMed SD card into AirwayLab. Four research-grade engines analyse your flow data for the patterns AHI ignores. Free, built on an open-source engine, and 100% private — your data never leaves your browser.