If you've been diagnosed with sleep apnea, you probably know your AHI number by heart. Maybe your sleep physician told you anything under 5 is "normal." Maybe your PAP app shows you a score of 2.3 and gives you a thumbs up. But here's the uncomfortable truth that sleep medicine is slowly coming to terms with: AHI is a deeply flawed metric, and relying on it alone may be leaving your sleep problems undertreated.
What AHI Actually Measures (and Doesn't)
The Apnea-Hypopnea Index counts two types of events per hour of sleep: apneas (complete breathing cessation for at least 10 seconds) and hypopneas (partial airflow reduction with accompanying oxygen desaturation). It's simple, standardized, and has been the cornerstone of sleep apnea diagnosis and treatment monitoring since the 1990s.
But simplicity comes at a cost. Here's what AHI doesn't capture:
Event Duration
A 10-second apnea and a 60-second apnea count exactly the same. Yet the physiological impact — oxygen drop, cardiac stress, arousal intensity — is vastly different.
Oxygen Severity
An event causing a 3% desaturation and one causing a 15% desaturation are both just "one event." The latter is far more dangerous to cardiovascular health.
Flow Limitation
Partial airway narrowing that causes arousals without meeting hypopnea criteria (called RERAs) is completely invisible to AHI. Entire syndromes exist around this.
Event Clustering
30 events evenly spread across the night vs. 30 events packed into 2 hours of REM sleep produce the same AHI but very different clinical outcomes.
The Evidence Against AHI-Only Treatment
The limitations of AHI aren't just theoretical. A growing body of research has demonstrated real clinical consequences:
A landmark 2016 study in CHEST Journal analyzed over 6,000 patients and found that AHI alone was a poor predictor of cardiovascular outcomes. Metrics like hypoxic burden (total oxygen desaturation load) and arousal frequency were significantly better at predicting heart disease risk.
The SAVE trial — one of the largest randomized controlled trials of PAP therapy — showed no reduction in cardiovascular events despite successful AHI reduction. Many researchers believe this is because AHI reduction alone doesn't address the full spectrum of breathing disturbances.
Meanwhile, studies on Upper Airway Resistance Syndrome (UARS) have shown that patients with normal AHI but significant flow limitation can experience the same degree of daytime sleepiness and cognitive impairment as patients with moderate obstructive sleep apnea.
Metrics That Actually Matter
Sleep medicine is gradually moving toward a more nuanced set of metrics. Here are the most promising ones — several of which you can already derive from your PAP SD card data:
Glasgow Index (Flow Shape Score)
Quantifies the overall distortion of your breathing waveform on a 0-1 scale. Unlike AHI, it captures the continuous spectrum of airway narrowing — from subtle flattening to severe obstruction. A score below 0.15 generally indicates well-treated therapy; above 0.3 suggests significant residual flow limitation.
Estimated RERA Index
By analyzing patterns of progressive flow limitation followed by sudden flow restoration (suggesting arousal), it's possible to estimate RERA events from PAP data. Combined with AHI, this approximates the more comprehensive RDI.
Hypoxic Burden
Rather than just counting desaturation events, hypoxic burden measures the total area under the oxygen desaturation curve. This captures both the depth and duration of oxygen drops, providing a far better marker of cardiovascular risk.
Breath-to-Breath Variability
Healthy breathing during sleep is rhythmic and stable. High variability in breath amplitude, timing, or shape can indicate periodic breathing patterns, unstable airway control, or treatment inadequacy — none of which AHI captures.
The "AHI 2 But Still Tired" Patient
If this is you, you're not imagining things and you're not alone. The sleep medicine community has a name for your situation: residual excessive daytime sleepiness (REDS) on PAP therapy. It's estimated to affect 15-30% of treated patients.
Common causes that AHI won't reveal:
- Subclinical flow limitation — enough to fragment sleep without triggering scored events
- Periodic breathing patterns — cyclical variations in breath amplitude that suggest ventilatory instability
- REM-clustered events — most of your events happening during dream sleep, which is critical for memory and mood
- Mask-related arousals — brief awakenings from leak or pressure discomfort that don't register as respiratory events
Taking a Deeper Look at Your Therapy
The gap between what your PAP machine reports and what's actually happening during your sleep is significant. But you have more data than you think. Your ResMed SD card contains breath-by-breath flow waveforms from every night — far more detailed than the summary statistics shown in the myAir app.
With the right analysis tools, this data can reveal:
- Whether flow limitation is present despite a low AHI
- How your breathing patterns change across the night
- Whether pressure or EPR adjustments might help
- Objective data to bring to your next sleep clinic visit
References
Azarbarzin et al. (2019). "The Hypoxic Burden of Sleep Apnoea Predicts Cardiovascular Disease-Related Mortality." European Heart Journal, 40(14), 1149-1157.
McEvoy et al. (2016). "CPAP for Prevention of Cardiovascular Events in Obstructive Sleep Apnea (SAVE Trial)." New England Journal of Medicine, 375(10), 919-931.
Guilleminault et al. (2006). "Upper Airway Resistance Syndrome: A Long-Term Outcome Study." Journal of Psychiatric Research, 40(3), 273-279.
Punjabi et al. (2008). "Sleep-disordered breathing and mortality: A prospective cohort study." PLoS Medicine, 5(8), e173.
Go Beyond AHI with AirwayLab
Upload your ResMed SD card to see Glasgow Index, flow limitation scores, RERA estimates, and trends your machine's app won't show you. Free, open-source, and 100% private.