What Causes Sleep Bruxism? The Evidence on Grinding Teeth at Night

An international consensus in 2018 framed bruxism in otherwise healthy people as a behavior rather than a disorder, one that can be a risk factor for outcomes such as tooth wear or jaw pain. Sleep bruxism is the nighttime form: rhythmic or non-rhythmic jaw-muscle activity during sleep that you usually cannot feel happening.

Because it happens outside awareness, its causes have to be studied in the sleep lab and across large populations rather than by asking people why they do it. When researchers do that, no single cause carries the behavior. The most useful way to read the evidence is as several contributing factors that mix differently in different people.

If your teeth are wearing or your jaw hurts, that is a dental matter worth taking to a dentist. The evidence here describes what drives grinding across groups, not a diagnosis of your own case.

The clearest mechanism is a micro-arousal: a brief shift toward lighter sleep with a rise in heart rate and brain activity. In sleep bruxism patients, that activation tends to appear a few seconds before the jaw moves, which is why grinding is described as an oromotor event secondary to arousal. The underlying jaw-muscle pattern, rhythmic masticatory muscle activity, also shows up in most normal sleepers at a lower intensity, so bruxism looks like an amplified version of a common sleep behavior.

A clinical overview places these micro-arousals at roughly 8 to 14 times per hour of sleep and notes that grinding activity peaks in the minutes before REM sleep. Both figures below come from separate sleep-lab studies and use different measures, so read them as two pieces of the same arousal model rather than a single statistic.

In a large Finnish twin study of young adults, additive genetic effects explained 52 percent of the variation in who reports sleep bruxism (95% CI 41 to 62), with non-shared environmental effects explaining the rest. In the same cohort, 8.7 percent experienced bruxism weekly, with no significant difference between men and women.

A heritability figure like this describes variation across a population, not the cause of any one person's grinding. It does explain why grinding often runs in families and why willpower alone rarely settles it. Genes appear to set a susceptibility that sleep arousals and everyday factors then act on.

Certain antidepressants are linked to sleep bruxism. A systematic review of psychotropic medications found higher odds of sleep bruxism for paroxetine, venlafaxine, and duloxetine, while several other SSRIs (citalopram, escitalopram, fluoxetine, mirtazapine, and sertraline) showed no increased odds. The review rated the overall quality of evidence as very low, so these numbers point to a signal worth knowing rather than a settled risk.

A separate cross-sectional study of 807 people found bruxism in 24.3 percent of antidepressant users versus 15.3 percent of controls, with an incidence of antidepressant-associated bruxism of 14.0 percent. If grinding starts soon after a new medication, that is worth raising with the prescriber rather than stopping on your own.

A systematic review of sleep bruxism and everyday substances found that the odds of grinding rise almost two times in people who drink alcohol and more than two times in current smokers, with a weaker link of about 1.5 times for those drinking more than eight cups of coffee a day. The authors rated the evidence as limited, so these are useful for direction and rough size rather than precision.

An umbrella review of 41 systematic reviews adds more contributors that were consistently associated with bruxism: caffeine, tobacco, some psychotropic medications, acid in the oesophagus, and second-hand smoke. Association is not proof of cause, and several of these factors travel together, so untangling any single one is hard.

Stress and anxiety are among the most consistently reported psychosocial correlates of bruxism, which is why they get so much attention. They sit alongside genetics and sleep physiology, though, so they explain part of the pattern rather than all of it. For people whose grinding does track with stress, the state you arrive at bedtime in is the part you can most directly work on, and research suggests settling that daytime arousal may help.

The older idea that a bad bite causes grinding has not held up well. A clinical overview found limited evidence for occlusal factors in the cause of sleep bruxism, which is one reason the field moved toward the arousal and central-nervous-system model. So a page that blames grinding on stress alone, or on the shape of your teeth, would be overstating a single strand.

Sleep bruxism is measured indirectly, which shapes what the evidence can say about cause. A few things drive most of the uncertainty.

Every figure on this page traces to a primary study or a named systematic review, linked in the sources list, and was confirmed against a supporting quote from that source.

The arousal figures come from two separate sleep-lab studies with small matched samples (Kato 2001 and Lavigne 2001). They are consistent with the wider arousal model but should not be read as precise population rates.

The genetic estimate (52 percent of variance) is a heritability-style figure from one large twin cohort of young adults. It describes variation across a population, not the cause of any one person's grinding.

The medication odds ratios come from a systematic review that rated its own evidence as very low quality, and only some agents were associated with sleep bruxism. The substance odds ratios come from a review that rated its evidence as limited, so both are shown as approximate and labeled as associations.

This page focuses on causes and risk factors. Prevalence figures, including awake versus sleep bruxism rates, live on the separate bruxism statistics page and are not repeated here.

We omit any figure attributing a fixed share of bruxism to stress alone, occlusion-based causal claims, and derived counts of affected people, because those either overstate a single strand or could not be verified against a primary source.