Why Does My AHI Change Night to Night? Normal Variation Explained

You wake up, check last night's number, and your AHI has jumped. Yesterday it was 1.8; today it says 6.4. Before you panic, take a breath: a single night's AHI is one of the noisiest numbers in all of CPAP therapy, and a one-off spike almost never means your treatment has failed. This guide explains what makes AHI swing from night to night, why one bad reading isn't a crisis, and how to read the trend that actually matters.

First, a quick definition. AHI stands for the Apnea-Hypopnea Index — the average number of apneas (full breathing pauses) and hypopneas (partial reductions) your machine recorded per hour of use. Because it's an average over a single night, it's exquisitely sensitive to whatever happened that one night. (For the full ranges, see What Is a Good AHI on CPAP?.)

Why AHI Varies Night to Night

Night-to-night variation in AHI is completely normal. Your airway, your sleep stages, and your body chemistry are different every night, so the number that summarizes them will be different too. The biggest, most common drivers are:

• Sleep position. Most people have more events on their back (supine) than on their side, because gravity pulls the tongue and soft tissue backward. A night spent mostly supine can double or triple your AHI compared with a side-sleeping night — with no change in your machine or your health.
• Alcohol. Even a couple of drinks in the evening relaxes the airway muscles and suppresses the arousal response, which tends to lengthen and increase events. This is one of the most reliable single-night spikes people see.
• Illness and congestion. A cold, allergies, or a stuffy nose narrows your airway and can force mouth breathing, both of which push events and leak up. Sinus pressure alone can wreck an otherwise good night.
• Sedatives and certain medications. Sleep aids, muscle relaxants, and opioids can deepen muscle relaxation or blunt your breathing drive, raising the count.
• Mask leak. This is the sneaky one. A large leak — air escaping around the mask or through your mouth — can both cause more events (the machine can't hold pressure) and corrupt the number itself. When leak is high, your machine may under-report or invalidate the AHI for stretches of the night, so the figure you see isn't trustworthy. Always glance at your leak rate before you react to an AHI change. (For thresholds and fixes, see CPAP Leak Rate: What's Acceptable and CPAP Mask Leaks & Mouth Leak.)

Other contributors include how much deep and REM sleep you got (REM relaxes the airway more), late heavy meals, weight changes over time, and even how well the mask seated that particular night. None of these mean your therapy is broken — they mean you're a human being whose nights aren't identical.

Why One Bad Night Isn't a Crisis

Here's the mental model that saves a lot of needless worry: a single night's AHI is a data point, not a diagnosis. It's a snapshot of one night, shaped by dozens of variables you can't fully control. Treating it like a report card invites anxiety that, ironically, can make your sleep worse.

Consider what a single high reading could be:

• A back-sleeping night after weeks of side sleeping.
• A glass or two of wine at dinner.
• The first night of a head cold.
• A mask that shifted at 3 a.m. and leaked for two hours.
• Simply a night with more REM sleep.

Any one of these can lift your AHI without any change in your underlying condition or your equipment. Statisticians call this noise — random, expected fluctuation around your true average. The fix is not to chase the number, and it is definitely not to start adjusting your machine. Never raise your own CPAP pressure to "correct" a bad night; higher pressure won't fix a leak or a back-sleeping night, and chasing it can actually provoke new problems (more on that below). Use your data to have an informed conversation with your provider — that's the defensible move. See Can I Adjust My Own CPAP Pressure? for where that line sits.

Reading Trends Instead of Single Nights

If single nights are noise, where's the signal? In the trend — your AHI averaged and viewed over weeks, not one morning at a time.

A useful way to think about it:

When you zoom out, the random spikes flatten and a true pattern emerges. A month where most nights sit comfortably below 5 events per hour — with the occasional spike — is a month of well-controlled therapy, even if two or three individual nights looked alarming. (For context on what counts as well-controlled, the treatment goal for adults on CPAP is generally a residual AHI below 5 events per hour; many clinicians aim lower when it's comfortably achievable, though targets are individualized with your provider.)

This is exactly where SomniCharts earns its keep. Its trend view plots your AHI, leak, and pressure over time and in plain language, so a single ugly night sits in context instead of dominating your attention. SomniCharts imports data from ResMed, Philips Respironics (including the encrypted DreamStation 2), and Löwenstein prisma machines and explains it automatically — which is a big reason people who used to panic over one morning's number end up trusting the pattern instead. To learn what every metric means first, start with How to Read Your CPAP Data.

When a Pattern Actually Matters

Normal variation bounces around a stable average. A pattern moves in one direction and stays there. Here's how to tell them apart and when to reach out to your clinician.

Likely just noise: - One high night surrounded by good ones. - Spikes you can explain (alcohol, a cold, a back-sleeping night, a leaky mask). - An average that returns to baseline within a few days.

Worth a clinician conversation: - A persistent upward trend over two to four weeks, not just one bad night. - A new, steady rise in central (clear-airway) events specifically. - Returning daytime symptoms — sleepiness, morning headaches, unrefreshing sleep — even if the numbers look okay. - High AHI that tracks with consistently high leak (fix the leak first, then re-read the AHI).

The central-event case deserves a note. CPAP is designed to splint the airway open and reliably treats obstructive apneas, but it does not directly correct the unstable breathing drive behind central (clear-airway) apneas. A rise in central events on CPAP can indicate treatment-emergent central sleep apnea (TECSA), sometimes called complex sleep apnea, which shows up in roughly 5–15% of PAP titrations. The reassuring part: it often resolves on its own within weeks to a few months of continued CPAP (spontaneous resolution is reported in about 60–80% of cases). Management may be watchful waiting on CPAP under your clinician's guidance, or — if it persists — switching to BiPAP or ASV. The wrong move is to raise pressure yourself to chase those centrals; higher pressure doesn't fix them and can make them worse. Bring it to your sleep clinician, especially if it persists or comes with new symptoms or medications. See Central Apneas Showing Up on CPAP: Treatment-Emergent CSA Explained and Central vs Obstructive Apnea.

One more honesty note about the number itself: your machine estimates AHI from airflow and pressure. It has no EEG and can't score arousals the way a sleep lab does, so device-reported AHI naturally differs from lab-scored AHI. That's another reason to weight trends over any single decimal. (More on this in How to Read Your CPAP Data.)

How to Track the Real Picture

You don't need to obsess to stay on top of your therapy — you need a sane routine and the right view.

1. Check trends weekly, not nightly. Glancing at last night's number is fine for curiosity, but make decisions from the 7- and 30-day view.
2. Always read leak next to AHI. If a high-AHI night also has high leak, the leak is the story — and the AHI may not even be valid that night. ResMed's excess-leak threshold is about 24 L/min at the 95th percentile; Philips reports total leak, so the baselines differ.
3. Note your context. Jot down alcohol, illness, travel, or back-sleeping nights. A spike you can explain is a spike you can ignore.
4. Watch direction, not data points. Ask "is my average drifting up over weeks?" rather than "was last night bad?"
5. Bring data, not panic, to your provider. A clean trend chart turns a vague worry into a productive conversation. SomniCharts can generate a compliance and trend report you can hand to your doctor.

The bottom line: AHI changes night to night because you change night to night. One bad reading is almost always noise. The signal lives in the trend — and once you start reading the pattern instead of the panic, the day-to-day swings stop running your mornings. For more on tuning the rest of your setup, see the Troubleshooting & Optimizing CPAP hub and Why Is My AHI High on CPAP?.

Frequently Asked Questions

Is it normal for my AHI to double from one night to the next? Yes. Doubling — say from 2 to 4 — is well within normal night-to-night variation, especially with a change in sleep position, a drink, a cold, or a leaky mask. Look at your weekly average before drawing any conclusion.

My AHI spiked to 8 last night. Should I worry? A single night at 8, surrounded by good nights, is almost always noise. Check your leak rate first, think about what was different (alcohol, congestion, back sleeping), and watch the trend. Worry is warranted only if the number stays elevated over weeks.

Why did my AHI spike when nothing changed? Something almost always changed, even if it's invisible: more REM sleep, more time on your back, a slight mask shift, early congestion, or a quiet leak. These don't reflect a problem with your therapy — they reflect a normal, variable night.

How many nights do I need before a high AHI means something? Look for a sustained upward trend across roughly two to four weeks, not a single reading. A persistent rise — especially with returning symptoms or growing central events — is the cue to talk to your sleep clinician.

Does high leak make my AHI unreliable? It can. Large leaks prevent the machine from holding pressure and can cause it to under-report or invalidate the AHI for parts of the night. Always read leak alongside AHI; if leak is high, fix that before reacting to the event count.