Can sleep apnea kill you? An honest answer

A pause in breathing during sleep is almost never instantaneous death. The risks that show up in the literature are slower and more ordinary: crashing a car because you are sleepy, and, in observational cohorts, higher all-cause mortality among people with severe untreated sleep-disordered breathing. Those are reasons to diagnose and treat. They are not reasons to lie awake counting inspirations. This article stays with studies we opened: crash meta-analysis and AASM/NHTSA drowsy-driving documents, plus Young’s Wisconsin cohort and Punjabi’s Sleep Heart Health Study.

Does an apnea kill you in the moment it happens?

Almost never. An obstructive apnea is a temporary airway closure. The brain usually arouses, the airway reopens, and breathing resumes, hundreds of times a night in severe disease. That cycle is miserable physiology (hypoxia, blood-pressure swings, sleep fragmentation). It is not a reliable mechanism for dying in the bed during a single event. People do die in their sleep from many causes. Pinning a specific nocturnal death on one scored apnea, without an arrhythmia record or an autopsy context, is not how those studies are built.

The AHA 2021 scientific statement notes that OSA is associated with a spectrum of rhythm disturbances and that long pauses and bradycardia are common on sleep studies. Association is not the same as “this apnea will stop your heart.” If you have unexplained syncope, high-grade block, or a defibrillator, that is a cardiology conversation that can include a sleep study. It is still not a reason to describe routine OSA as a nightly near-death experience.

The scare version of this question flourishes because the word “apnea” sounds like a code. In clinic it is a scored event lasting at least 10 seconds. Oxygen often falls, the pulse may slow or surge, and then an arousal reopens the airway. That loop can repeat through the night without ever becoming a terminal event. Treat the disease. Do not mythologize the epoch.

What is the crash risk?

This is the harm that is easiest to explain and the one you can do something about tonight. Tregear and colleagues’ systematic review and meta-analysis in the Journal of Clinical Sleep Medicine concluded that people with OSA are at increased risk of motor-vehicle crash. The mean crash-rate ratio, they judged, likely falls between 1.21 and 4.89. They noted the finding was somewhat stronger for non-commercial drivers because few studies enrolled commercial drivers specifically. Predictors that may matter include BMI, AHI, oxygen saturation, and possibly daytime sleepiness. Study quality was often low: retrospective designs, incomplete adjustment, self-reported crashes. As a worked example, they wrote that if a driver’s underlying crash risk were 0.08 crashes per person-year, OSA would put the expected range around 0.10 to 0.39. That is a rate, not a prophecy for your next commute, and it is why sleepy driving is the part of this topic you can change tonight.

NHTSA’s clinician fact sheet on sleep disorders, drawing on its Driver Fitness Medical Guidelines, states that drivers with OSA appear to be at particular risk of drowsy-driving crashes, that sleep fragmentation produces excessive daytime sleepiness and cognitive dysfunction, and that evidence indicates OSA increases crash risk. The same sheet says CPAP is the treatment demonstrated to reduce crash risk, and that once started it needs to continue if the person intends to keep driving. Those NHTSA documents are older (the guidelines cited are 2009-2010). The direction of the finding has not been reversed by later work. The AASM transportation task force, commenting during FMCSA’s 2016 apnea ANPRM, likewise treated moderate-to-severe OSA as a contributor to crash risk among commercial operators and treated CPAP as effective at reducing that risk. That is still not a claim that every untreated apnea ends a life. It is a claim that untreated sleepiness ends trips.

The American Academy of Sleep Medicine’s 2014 drowsy-driving health advisory treats drowsiness as similar to alcohol in reducing alertness, delaying reaction time, and impairing decisions. It asks states to teach drowsy driving in driver’s education and asks transportation companies to screen for sleep disorders including OSA. That is public-health language, not a mortality statistic. If you are fighting sleep on I-270, the intervention is to stop driving, not to finish the errand.

What did the Wisconsin Sleep Cohort find about death?

Young, Finn, Peppard, and colleagues published an 18-year mortality follow-up of 1,522 adults from a population-based Wisconsin sample, not a sleep-clinic roster. Sleep-disordered breathing was measured with laboratory polysomnography. Cut points at 5, 15, and 30 events per hour defined mild, moderate, and severe disease. Eighty deaths occurred in the analysis they emphasized.

The adjusted hazard ratio for all-cause mortality, severe versus no SDB, was 3.0 (95 percent CI 1.4 to 6.3), adjusted for age, sex, and BMI. After excluding 126 people who had used CPAP, that hazard ratio was 3.8 (1.6 to 9.0), and the cardiovascular-mortality hazard ratio was 5.2 (1.4 to 19.2). Mild and moderate categories had death rates that looked higher but were not statistically significant in the adjusted models they trusted. The authors were careful: residual confounding is possible, the severe group was small, and this is still observational.

Read the CPAP-exclusion analysis as a clue, not as a randomized trial. People who used CPAP may have differed in adherence to other care. Wisconsin cannot prove that treating OSA lengthens life. It can show that severe, mostly untreated SDB in a community sample tracked with higher mortality over 18 years. The authors also tested whether sleepiness changed the mortality association and did not find a statistically significant interaction: severe SDB tracked with higher death rates in both “sleepy” and “not sleepy” strata. That is another reason not to wait for spectacular daytime dozing before taking a high AHI seriously, and another reason not to equate a quiet afternoon with immortality.

What about Punjabi and the Sleep Heart Health Study?

Punjabi, Caffo, Goodwin, and colleagues followed 6,441 Sleep Heart Health Study participants for a mean of 8.2 years; 1,047 died. Sleep-disordered breathing was measured with in-home polysomnography. Fully adjusted hazard ratios for all-cause mortality versus AHI under 5 were 0.93 (0.80 to 1.08) for mild (AHI 5.0 to 14.9), 1.17 (0.97 to 1.42) for moderate (15.0 to 29.9), and 1.46 (1.14 to 1.86) for severe (AHI 30 or more). The severe-disease signal was statistically clear in men aged 40 to 70 (hazard ratio 2.09, 1.31 to 3.33). Intermittent hypoxemia, not sleep fragmentation, associated independently with death.

Put Wisconsin and Sleep Heart Health next to each other and the pattern is consistent enough to take seriously: severe disease, not mild disease, carries the observational mortality association, and it is stronger in middle-aged men in SHHS. The hazard ratios are not identical (3-fold in Wisconsin’s severe group versus 1.46 in SHHS fully adjusted). Different samples, different follow-up, different adjustment. Neither study is a reason to tell a person with AHI 8 that they have a death sentence. Both are a reason not to shrug at AHI 40 in a 55-year-old who refuses a study.

SAVE, discussed in the heart article, then reminds us that giving CPAP to a less-sleepy secondary-prevention cohort at 3.3 hours a night did not change a cardiovascular composite. Observational mortality and randomized treatment effects are allowed to disagree. An honest clinic says so.

How should you hold these facts without panic?

Treat sleepiness as an immediate safety problem. If you are drowsy at the wheel, you are in the Tregear/NHTSA story, not the 18-year hazard-ratio story. Do not drive. Seek treatment for sleep apnea if the history and a study support it. CPAP therapy is the intervention with the crash-risk evidence NHTSA cites, and it is the AASM’s first-line treatment for sleepy adults with OSA.

Treat severe untreated disease as a chronic risk marker, like severe untreated hypertension is a chronic risk marker. You address it because the observational signal is there and because symptoms and crash risk are there. You do not address it because a website used the word “killer.”

If a family member died in their sleep and you are here because of that, say so. We can talk about what a sleep study can and cannot explain. We cannot reconstruct that night from an AHI printed a decade later.

What is the next step in Westerville?

A consult and, when indicated, a test. The appointments page explains how to get on the schedule. Call (614) 898-9340. If you are sleepy driving to the visit, have someone else drive. That is the part of this topic that is not theoretical.