What New Caffeine Research Means for Your Sleep

On June 18, 2026, a study in SLEEP Advances of 37,469 US adults reported that taking in 101–400 mg of caffeine within eight hours of bedtime was linked to higher odds of sleeping less than six hours. In April 2026, a review in Nutrients, publicized in late September, concluded that caffeine shifts sleep toward a lighter, more aroused brain-wave pattern, often detectable even when standard sleep scoring changes only modestly. For readers, official daily limits have not changed and no US agency has issued a new caffeine-and-sleep rule; what the research adds is weight to timing, and a 2023 meta-analysis already puts numbers on it: about nine hours before bed for a typical coffee and about 13 for a standard pre-workout serving.

Key takeaways

The FDA still cites 400 mg a day as an amount not generally associated with negative effects in most adults, and ACOG reaffirmed in 2026 that under 200 mg a day in pregnancy does not appear to be a major factor in miscarriage or preterm birth.
A 2023 meta-analysis of 24 studies found caffeine cut total sleep by about 45 minutes and suggested cutoffs of 8.8 hours before bed for a 107 mg coffee (250 mL, about 8.5 ounces) and 13.2 hours for a 217.5 mg pre-workout serving.
Dose changes the picture: in a 2025 trial of 23 young men who were moderate caffeine users, 100 mg had no measurable effect on sleep even 4 hours before bed, while 400 mg disturbed sleep when taken up to 12 hours before.
The FDA does not require caffeine amounts on food, beverage or supplement labels, and decaf coffee still has about 2–15 mg per 8 ounces.
Pregnancy, oral contraceptives, the antidepressant fluvoxamine and cirrhosis slow caffeine clearance, so a dose lasts longer.

What it means for you

Afternoon coffee drinkers: one regular cup is a moderate dose, and the meta-analysis cutoff for it is 8.8 hours before bed, about 2:12 p.m. for an 11 p.m. bedtime. A small cup later in the day carries less risk than a large one.
Energy drink and pre-workout users: a 16-ounce energy drink can carry up to 328 mg, so late timing and high doses collide; the worked example below covers a 4 p.m. workout.
Shift workers: NIOSH training for nurses warns that coffee near the end of a night shift can still be in the system during the daytime sleep that follows, causing restlessness or waking.
Pregnant people: the 200 mg figure is the one to track, and caffeine clears progressively more slowly as pregnancy advances, as explained in "Why sensitivity varies so much."
People taking fluvoxamine or oral contraceptives: both slow caffeine elimination, so a typical cutoff may be too late.
People with insomnia: moving caffeine earlier is reasonable, but sleep-medicine guidelines advise against relying on sleep-hygiene changes alone for chronic insomnia.
Teens and children: the 2025–2030 Dietary Guidelines tell parents to avoid caffeinated drinks for children aged 5–10 and tell adolescents to significantly limit energy drinks.

What the 2026 research found

The SLEEP Advances analysis used National Health and Nutrition Examination Survey (NHANES) data from 2005–2018, with caffeine estimated from a 24-hour diet recall and sleep from a questionnaire. Compared with no caffeine, 400 mg or more a day carried 1.5 times the adjusted odds of short sleep, while intakes under 400 mg showed no clear difference. Within the eight hours before bed, 101–200 mg carried 1.4 times the odds and 201–400 mg 1.6 times; intake in a 12-hour window showed no significant link after adjustment.

The study is cross-sectional, so it shows an association, not proof that caffeine caused the short sleep. Usual bedtime was asked only in the 2015–2018 survey cycles; for earlier years the authors estimated it from the average for people of the same age and sex, so the eight-hour window is approximate. Self-reported "trouble sleeping" was not linked to caffeine once the authors adjusted for factors such as age, smoking and depression.

The Nutrients review covered 32 studies that recorded brain activity during sleep with EEG (electroencephalography), summarizing them in words rather than pooling the numbers because the study designs differed so much. Its most consistent finding was less slow-wave activity, the brain signal of deep sleep, with more fast activity typical of lighter, more aroused sleep. These measures often picked up disruption when standard sleep-stage scoring changed only modestly. Wroclaw Medical University's release of September 22, 2026 put it more plainly: caffeine may make sleep less restorative even when people fall asleep normally and stay in bed for eight hours.

Finding your own cutoff

Most cutoff advice rests on a few controlled studies. Their doses and samples explain why advice ranges from six to 13 hours.

Study	Dose and timing	Result
Drake et al., 2013 (12 healthy sleepers)	400 mg at bedtime, 3 or 6 hours before	All three disturbed sleep; at 6 hours, measured sleep fell by more than one hour, per the AASM release
Gardiner et al., 2023 (meta-analysis, 24 studies)	Varied	About 45 minutes less sleep, 7% lower sleep efficiency, 9 minutes longer to fall asleep, 11.4 minutes less deep sleep; cutoffs of 8.8 hours (107 mg per 250 mL coffee) and 13.2 hours (217.5 mg pre-workout)
Gardiner et al., 2025 (23 men, about 25 years old, usually taking under 300 mg a day)	100 or 400 mg at 12, 8 or 4 hours before bed	100 mg: no significant effect at any time; 400 mg: slower sleep onset within 12 hours, more broken sleep within 8 hours, perceived quality down 34% at 4 hours

The 6-hour gap common in sleep-hygiene advice is what the 2013 study tested, with a large 400 mg dose; its authors said the results supported avoiding substantial caffeine for a minimum of six hours before bed. The later work shows the needed gap grows with the dose.

The calculation itself is a subtraction:

latest caffeine time = usual bedtime − cutoff hours for that dose

cutoff hours: about 8.8 for a 107 mg coffee (250 mL),
              about 13.2 for a 217.5 mg pre-workout serving
              (Gardiner et al., 2023)

Illustrative example. Someone with an 11 p.m. bedtime takes a 217.5 mg pre-workout scoop at 4 p.m. That is 7 hours before bed. The meta-analysis cutoff for that dose is 13.2 hours, which puts the latest time at 9:48 a.m., so the 4 p.m. scoop is about 6.2 hours too late. It also falls inside the 8-hour window the 2026 survey linked to short sleep.

A 107 mg coffee has a latest time of 2:12 p.m. for the same bedtime, so even a coffee at 4 p.m. would miss its cutoff by 1.8 hours. The 2025 trial, by contrast, found no significant effect from 100 mg taken as late as 4 hours before bed in young men who were moderate caffeine users, so a 100 mg dose 7 hours before bed is unlikely to produce a measurable effect in that group.

The two results differ because the meta-analysis cutoffs are estimates set to avoid any loss of total sleep across many studies, while the trial tested one narrow group. A late dose near 100 mg looks like a small risk for many adults. From about 200 to 400 mg, the evidence lines up more closely: the survey linked 101–400 mg in the last 8 hours to short sleep, the meta-analysis put the cutoff for 217.5 mg at 13.2 hours, and the trial's 400 mg dose disturbed sleep when taken up to 12 hours before bed. For more than 400 mg in that window, the survey found no clear link either way (odds ratio 1.0, with a wide 95% confidence interval of 0.4–2.2).

Limits that have not changed

The FDA's consumer guidance (current as of August 2024) cites 400 mg a day for most adults as an amount not generally associated with negative effects, and adds that sensitivity and clearance vary widely. It is not a sleep threshold: the trials above show disruption from single late doses well below it.

For pregnancy, ACOG Committee Opinion 462, first issued in August 2010 and reaffirmed in 2026, states that under 200 mg a day does not appear to be a major factor in miscarriage or preterm birth, while the link with growth restriction remains undetermined. The FDA suggests that people who are pregnant, trying to conceive or breastfeeding discuss caffeine with their health care provider.

For young people, the Dietary Guidelines for Americans, 2025–2030, released on January 7, 2026, advise avoiding caffeinated drinks for ages 5–10 and significantly limiting energy drinks for ages 11–18. The document sets no caffeine limit for adults. The American Academy of Pediatrics' 2011 clinical report concluded that caffeine and other stimulants in energy drinks have no place in the diet of children and adolescents.

The American Academy of Sleep Medicine's healthy sleep habits page (updated August 2020) gives the simplest version: avoid caffeine in the afternoon or evening.

Why sensitivity varies so much

Caffeine is broken down mainly in the liver by an enzyme called CYP1A2. The FDA-approved caffeine citrate label puts the adult half-life, the time for blood levels to fall by half, at about 5 hours and notes that drugs such as cimetidine and ketoconazole slow elimination. Averages hide a wide range: a European Food Safety Authority expert presentation from 2015 put the adult half-life at roughly 2–8 hours.

Pregnancy: the same EFSA material says half-life rises progressively during pregnancy, up to about 16 hours, three to four times longer than outside pregnancy.
Oral contraceptives and smoking: a 2022 analysis of published caffeine pharmacokinetic data found smoking consistently speeds caffeine clearance and oral contraceptives consistently slow it.
Medicines: the same analysis found most drugs had no effect on caffeine levels, with exceptions including the antidepressant fluvoxamine and the antibiotic norfloxacin, where it says caution is warranted.
Liver disease: the largest reductions in clearance appear in cirrhosis and track how impaired the liver is.
Genes: the 2026 Nutrients review lists variation in the adenosine receptor gene ADORA2A among the factors that changed how strongly caffeine affected sleep EEG.

For shift workers, NIOSH training for nurses gives a half-life of 5 to 6 hours and warns that coffee near the end of a night shift can disturb the sleep that follows. For people with chronic insomnia, the AASM's 2021 guideline strongly recommends cognitive behavioral therapy for insomnia (CBT-I) and suggests clinicians not use sleep hygiene, which includes caffeine rules, as a stand-alone treatment.

Caffeine you may not be counting

Under FDA rules, packaged foods must list caffeine in the ingredients when it is added on its own, but the amount is voluntary, and caffeine from an ingredient such as chocolate does not appear as "caffeine" at all. Restaurants and coffee shops are not required to disclose it. The FDA notes that most US energy drinks do state their total caffeine voluntarily.

Source	Serving	Caffeine
Brewed coffee (FDA)	12 fl oz	113–247 mg
Energy drink (FDA)	16 fl oz	54–328 mg
Black tea / green tea (FDA)	12 fl oz	71 mg / 37 mg
Caffeinated soft drink (FDA)	12 fl oz	23–83 mg
Decaf coffee (FDA)	8 fl oz	2–15 mg
Dark chocolate, 70–85% cacao (USDA)	1.5 oz (about 43 g)	About 34 mg
Excedrin Migraine (label dose)	2 caplets	130 mg

The chocolate figure is calculated from the 80 mg per 100 g listed in USDA FoodData Central for 70–85% dark chocolate. The Excedrin Migraine label lists 65 mg of caffeine per caplet, warns that a dose holds about as much caffeine as a cup of coffee, and lists sleeplessness among the effects of too much caffeine. The FDA also lists energy bars, gum, ice cream and supplements among products that may contain added caffeine.

Frequently asked questions

Is caffeine from tea or chocolate gentler than added caffeine?

No. The FDA states that the body handles naturally occurring caffeine the same way as caffeine added to a product, with no difference in safety. What differs is the dose: green tea typically has far less per cup than brewed coffee.

Can decaf be an evening drink?

For most people, yes, with a caveat. The FDA puts decaf coffee at about 2–15 mg per 8 ounces, below the 100 mg dose that showed no significant sleep effect in the 2025 trial. Several large mugs add up, and for people who clear caffeine slowly, such as during pregnancy, even a small dose stays in the body longer.

Can caffeine hurt sleep without any feeling of being wired?

Yes. In the 2013 study, participants' own reports suggested they were unaware of the sleep disturbance that the monitor recorded. The 2025 trial authors also noted a gap between measured and perceived sleep, concluding that people may have difficulty judging how caffeine affects their sleep.

How much caffeine is dangerous?

The FDA estimates that toxic effects such as seizures can occur with rapid intake of about 1,200 mg, which is less than half a teaspoon of pure powdered caffeine. It warns that pure and highly concentrated caffeine products can cause serious harm, including death.

Does caffeine before a nap make sense on night shifts?

For staying alert at work, it can. NIOSH training for nurses says a brief nap of about 30 minutes can increase alertness during a night shift, and that caffeine taken immediately before such a nap gives the benefits of both. Its caffeine module adds that caffeine takes about 30 minutes to work, with a range of 15–45 minutes, so the effect tends to arrive around the time a short nap ends. This applies to short naps on shift, not to the main sleep period afterward.

A one-week caffeine check

This article is general education, not medical advice. A doctor or pharmacist is the right person to ask if sleep problems last for weeks, if caffeine causes a racing heart or anxiety, or if a prescription might slow caffeine clearance.

For seven days, write down every caffeinated item with its time and an estimated amount in milligrams, including pain relievers, pre-workout powders, tea, soda and dark chocolate; use the label where one is given and the FDA ranges above where it is not.
Mark the latest time anything above about 100 mg was taken and compare it with the bedtime cutoff: about 9 hours for a regular coffee, at least 13 hours for a dose around 200 mg; the meta-analysis gave no figure for larger doses.
Add up each day's total and check it against 400 mg, or 200 mg in pregnancy.
Move the largest late dose earlier first, rather than cutting everything at once, and note any change in time to fall asleep and morning alertness over the following week.