The apnea-hypopnea index (AHI) is calculated by dividing the total number of apneas and hypopneas during sleep by the total hours of sleep. If you had 60 breathing disruptions over 6 hours of sleep, your AHI would be 10, meaning an average of 10 events per hour. The formula is simple, but the way each breathing event gets counted involves specific clinical criteria that directly affect your score.
The Basic Formula
AHI = (total apneas + total hypopneas) ÷ total sleep time in hours. Every pause in breathing and every episode of shallow breathing gets tallied, then divided by how long you actually slept. The result is a per-hour average that serves as the primary number used to diagnose and classify obstructive sleep apnea.
The denominator matters more than you might expect. In a lab sleep study (polysomnography), technicians measure your actual sleep time using brain wave sensors, so periods when you were lying awake don’t count. Home sleep tests can’t measure brain waves, so they use total recording time instead. Because that recording time includes any minutes you spent awake in bed, home tests tend to underestimate AHI by about 15%. A lower score doesn’t necessarily mean fewer breathing problems; it may just mean the math used a larger denominator.
What Counts as an Apnea
An apnea is a complete stop in airflow lasting at least 10 seconds. Sensors placed near your nose detect whether air is moving. If airflow drops to zero for 10 seconds or longer, that’s one apnea event added to the count. These can happen dozens or even hundreds of times per night in severe cases, often without the sleeper being aware.
What Counts as a Hypopnea
Hypopneas are partial blockages: you’re still breathing, but not enough air is getting through. The current scoring standard from the American Academy of Sleep Medicine requires all three of these criteria to be met for an event to count as a hypopnea:
- Airflow reduction: Breathing drops by at least 30% from your baseline.
- Duration: The reduction lasts at least 10 seconds.
- Oxygen drop or arousal: Your blood oxygen level falls by at least 3% from where it was before the event, or the event causes a brief awakening (arousal) detected on brain wave monitoring.
These criteria have changed over the years, which is worth knowing if you’re comparing results from different time periods. An older standard required a 4% oxygen drop and didn’t count arousals. Under that stricter rule, many shallow-breathing events went unscored, producing lower AHI numbers for the same night of sleep. The current 3%-or-arousal rule catches more events, so an AHI measured today may be higher than one measured a decade ago even if your breathing hasn’t changed.
Severity Ranges
Once your AHI is calculated, it falls into one of four categories used by most sleep specialists:
- Normal: fewer than 5 events per hour
- Mild: 5 to fewer than 15 events per hour
- Moderate: 15 to fewer than 30 events per hour
- Severe: 30 or more events per hour
These cutoffs were formalized in 1999, largely based on data from the Wisconsin Sleep Cohort showing that hypertension risk climbs substantially around an AHI of 30. The boundaries between mild and moderate, however, have less data behind them. Research has found a positive correlation between AHI and cardiovascular disease but hasn’t identified a clear threshold where risk suddenly jumps. Many sleep specialists now treat AHI as a continuous variable rather than rigidly sorting patients into categories, combining it with symptoms, oxygen levels, and other factors to guide treatment decisions.
Lab Study vs. Home Sleep Test
A lab polysomnography measures AHI with more precision because it tracks brain activity, eye movement, and muscle tone to determine exactly when you’re asleep. That gives a true total sleep time for the denominator. Home sleep tests use a simpler device that monitors airflow, chest movement, and blood oxygen but can’t tell whether you’re asleep or just lying still. The denominator becomes total recording time, which inflates the number you’re dividing by and pushes AHI down.
One study comparing the two methods found an average AHI of 26.5 on lab polysomnography versus 23.8 on home testing for the same patients. The gap widens as sleep apnea gets more severe. Interestingly, lab studies can also skew results in the other direction: being hooked up to equipment tends to keep people on their backs more than they’d normally sleep. Supine time during polysomnography increases by roughly 56% compared to sleeping at home, and back-sleeping worsens airway collapse. So lab studies may overestimate severity in people whose apnea is position-dependent.
AHI vs. RDI
You may see another metric on your sleep report called the respiratory disturbance index (RDI). AHI counts only apneas and hypopneas. RDI includes those same events plus a third category: respiratory effort-related arousals, which are episodes where your breathing doesn’t drop enough to qualify as a hypopnea but still disrupts your sleep. RDI is always equal to or higher than AHI for the same night. Some insurance companies and diagnostic guidelines reference one or the other, so it’s useful to know which number you’re looking at when reading your results.
Why Your AHI Can Vary Night to Night
AHI is a snapshot of one night, and several factors can shift it. Sleeping on your back increases airway collapse compared to side-sleeping. Alcohol relaxes throat muscles and can raise your count significantly. Nasal congestion, altitude, and even how deeply you sleep all play a role. Sleep stage matters too: breathing disruptions are typically more frequent during REM sleep, when muscle tone drops to its lowest. A night with more REM time can produce a higher AHI than a night with less, even though your underlying anatomy hasn’t changed.
This variability is one reason clinicians look at the full picture rather than anchoring solely to a single AHI number. Your oxygen saturation pattern, how long events last, how much your oxygen dips during each event, and whether you have daytime symptoms like excessive sleepiness all factor into how your results are interpreted and what treatment looks like.