A big caffeine dose 12 hours before bed can still cut deep sleep

Key takeaways

• In an Australian trial published in 2025, 400 mg of caffeine taken 12 hours before bed cut about 21 minutes of deep sleep, while 100 mg taken 4 hours before bed had no significant effect⁴.
• A 2023 review of 24 studies found caffeine cut sleep by 45 minutes on average and suggested having a typical coffee at least 8.8 hours before bed².
• People often miss the damage: in a 2013 U.S. study and the 2025 trial, sleepers’ own reports missed losses that sleep monitors recorded¹⁴.
• Estimates of caffeine’s half-life in adults run from about 3 to 7 hours, and pregnancy, birth control pills, some medicines and liver disease can stretch it much longer⁷⁸¹⁵.
• The popular advice to wait 90 minutes after waking for coffee hasn’t been tested in trials, and a later first cup can push your last cup later, too⁹.

Twelve healthy sleepers in a 2013 Detroit study followed a strict routine: three pills a day, taken six hours, three hours and right before bedtime¹. On three of the four test days, one pill held 400 milligrams of caffeine, roughly two to three cups of coffee. The rest were dummies¹.

When the caffeine came six hours before bed, an in-home sleep monitor logged more than an hour of lost sleep¹. The sleepers’ own diaries hinted they hadn’t noticed¹.

That finding has held up. In an Australian trial published in 2025, a large dose taken a full 12 hours before bed still trimmed deep sleep, and the sleepers’ own ratings missed it⁴.

In a 2025 survey for the American Academy of Sleep Medicine (AASM), 56% of U.S. adults said they use caffeine to stay alert¹⁴. Using it well comes down to two numbers: how much, and how many hours before bed.

Caffeine doesn’t add energy. It hides your sleepiness

Through the day, a chemical called adenosine builds up in your brain and helps make you sleepy. Caffeine resembles adenosine and binds to the same receptors, which muffles that sleepy signal until the caffeine is gone¹⁵. Our article on sleep pressure explains the system in more detail.

Caffeine gets into your blood fast. Food Standards Australia New Zealand, a government food-safety agency, reported in 2025 that the body absorbs nearly all of a dose within about 45 minutes, with blood levels peaking around 30 minutes after you swallow it⁷. Chocolate is slower, peaking 1.5 to 2 hours after you eat it⁷.

Getting rid of caffeine takes much longer. The half-life, the time the body needs to clear half a dose, is usually 3 to 7 hours in adults, according to the Australian agency⁷. Other sources differ: slides by a University of Illinois Chicago nurse researcher, posted for a 2020 National Academies event, give a narrower 3 to 4 hours¹⁵. The liver does nearly all the work with an enzyme called CYP1A2⁷⁸.

This is what that means. Picture someone with a 5-hour half-life, the middle of that range, who drinks a coffee with 200 mg at 2 p.m. About 100 mg is still in the body at 7 p.m. and about 50 mg at midnight.

That’s our arithmetic, not a measured case, but it explains why an afternoon cup can linger. The National Heart, Lung, and Blood Institute (NHLBI) says caffeine’s effects can take 6 to 8 hours to wear off¹³.

Three studies put the cutoff earlier than most people think

The 2013 study came from the Henry Ford Sleep Disorders and Research Center and Wayne State University¹. All three timings, even the dose six hours out, significantly disrupted sleep. From that, the team offered a general rule of thumb: no caffeine after 5 p.m.¹.

In 2023, a team led by Carissa Gardiner at Australian Catholic University pooled 24 studies². On average, caffeine cut total sleep by 45 minutes and lowered sleep efficiency, the share of time in bed spent asleep, by 7%². It added 9 minutes to the time needed to fall asleep and cut about 11 minutes of deep sleep, while light sleep grew².

From those data, the team estimated cutoffs. A coffee with about 107 mg should come at least 8.8 hours before bed, and a standard 217.5 mg serving of pre-workout powder at least 13.2 hours before². The authors called these starting points for each person to adjust³.

In a trial published in 2025, the same group tested dose and timing together in a randomized crossover design, in which each person tried every condition in random order⁴. Twenty-three healthy men, average age about 25 and used to moderate caffeine, took 100 mg, 400 mg or a dummy capsule at 12, 8 or 4 hours before bed⁴.

Brain-wave monitors measured their sleep at home⁴. The university funded the trial, according to its registration, and the authors reported no conflicts of interest⁴⁵.

At 100 mg, caffeine had no statistically significant effect at any timing, even 4 hours before bed⁴. The 400 mg dose was another story.

Taken before bed	Total sleep	Deep sleep	Awake after first falling asleep	How people rated their sleep
4 hours	51 minutes less	30 minutes less	26 minutes more	34% worse
8 hours	29 minutes less (not statistically clear)	15 minutes less	29 minutes more	No clear change
12 hours	30 minutes less (not statistically clear)	21 minutes less	No clear change	No clear change

What 400 mg of caffeine did compared with a dummy capsule, from the 2025 trial⁴.

The trial had limits: only 23 young men took part, women were excluded, and each condition tested a single dose⁴⁵. A study this size can miss small effects, and it can’t tell you how a 60-year-old woman will sleep.

Your own sense of a good night may miss the damage

In the 2025 trial, people rated their sleep worse only when they took 400 mg 4 hours before bed. At 8 and 12 hours, their ratings didn’t drop, though monitors recorded lost deep sleep⁴. The Detroit sleepers showed the same gap¹.

Natural health puts weight on listening to your body, and for most habits that serves you well. Caffeine is a tricky case, because its toll on sleep can slip past the very sense you’d use to judge it.

So test it the way a careful researcher would. For two weeks, move your last caffeinated drink at least 4 hours earlier and keep everything else the same.

Each morning, jot down how long you think it took to fall asleep, how often you woke, and how rested you feel. Then compare with two ordinary weeks. Consumer trackers are rough at sorting sleep stages, as our tracker article explains, so trust the diary and your midafternoon alertness.

Your liver sets the clock, and many things slow it down

How fast you clear caffeine can vary about 40-fold from one person to the next, the Australian agency reported⁷. Several things can push you toward the slow end:

• Genes. A common variant of the CYP1A2 gene sorts people into faster and slower processors⁷.
• Pregnancy. The half-life stretches to about 10 hours by week 17 and up to 18 hours near term⁷.
• Birth control pills. They slow clearance⁸; the Illinois slides list a 5- to 10-hour half-life for women on the pill¹⁵.
• Quitting smoking. Smoking speeds caffeine clearance, so people who quit may find the same cup lasts longer⁷⁸.
• Liver disease. Clearance drops sharply, most of all in cirrhosis⁸.
• Some medicines. The antidepressant fluvoxamine and the antibiotic norfloxacin slow caffeine’s breakdown⁸. The Food and Drug Administration (FDA) says medical conditions and medicines can raise your sensitivity⁶.

Should you buy a caffeine gene test? The 2025 trial checked two genes tied to caffeine processing and sensitivity and found no clear effect on sleep, though it was too small to settle the question⁴. A 2024 review in the Journal of the International Society of Sports Nutrition (JISSN) found no consensus that the CYP1A2 gene predicts how people respond to caffeine in exercise studies⁹. Your own two-week test is cheaper and tells you more.

Find your last-cup time

These are our suggested cutoffs, based on the 2023 review and 2025 trial described above, not an official guideline. They assume a healthy adult. Find your bedtime, then read across to the latest time for each size of drink.

Your bedtime	Up to about 100 mg (12 oz of black tea, most sodas)	A regular brewed coffee (12 oz has 113–247 mg)	200–400 mg (energy drinks, pre-workout, large coffees)
9 p.m.	5 p.m.	Noon	7:45 a.m.
10 p.m.	6 p.m.	1 p.m.	8:45 a.m.
11 p.m.	7 p.m.	2 p.m.	9:45 a.m.
Midnight	8 p.m.	3 p.m.	10:45 a.m.

The first column uses the 4-hour result for 100 mg⁴. The middle column rounds the 2023 review’s 8.8-hour estimate up to 9 hours². The last column uses its 13.2-hour estimate for a 217.5 mg dose and the 2025 finding that 400 mg could disturb sleep 12 hours later²⁴. Drink sizes come from the FDA⁶.

Then adjust for your body. These adjustments are our suggestions:

• If your coffee is large or strong, use the right-hand column.
• If you’re a slow clearer (pregnant, on birth control pills or fluvoxamine, living with liver disease, or recently quit smoking), move each time earlier or keep caffeine to the morning.
• If you have trouble sleeping, use the right-hand column for every caffeinated drink.
• Count the whole afternoon. Two small drinks a couple of hours apart can add up to one bigger dose.
• Use your diary. If the times above don’t fix your sleep, move earlier again.

Hidden caffeine: where the milligrams hide

Brewed coffee alone can vary more than twofold⁶.

Drink	Typical caffeine
Brewed coffee, 12 oz	113–247 mg6
Black tea, 12 oz	71 mg6
Green tea, 12 oz	37 mg6
Caffeinated soda, 12 oz	23–83 mg6
Energy drink, 16 oz	54–328 mg6
Pre-workout powder, standard serving	about 218 mg, the serving size the 2023 review used2
Decaf coffee or tea, 8 oz	2–15 mg6

Caffeine also turns up in places you might not check:

• Chocolate. Labels list chocolate by name, not as “caffeine,” and its caffeine peaks later than coffee’s⁶⁷.
• Bars, ice cream and gum. The FDA says caffeine can be added to energy and protein bars, ice cream and chewing gum⁶.
• Supplements and over-the-counter medicines can contain it, too⁶.
• Plant extracts. Energy drinks often stack caffeine from guarana, yerba mate and green tea extract, and some single cans hold up to 350 mg¹². The FDA says your body handles naturally occurring caffeine no differently from added caffeine⁶.
• Café drinks. No law makes restaurants disclose a drink’s caffeine, so ask the staff⁶.

The “wait 90 minutes” rule is a theory nobody has tested

A trend that spread mostly on social media says to wait 30 to 120 minutes after waking before your first coffee, to avoid an afternoon crash⁹. The usual explanation involves cortisol, a hormone that rises in the morning, and adenosine.

The 2024 JISSN review examined those claims⁹. Caffeine can raise cortisol, but the response seems blunted in regular users and may vanish at intakes of 300 to 600 mg a day⁹. The authors found no support for the adenosine argument and no evidence that morning caffeine causes a crash or that waiting prevents one⁹.

To be even-handed, note who wrote it. Several authors have received money from supplement companies, and the society the lead author runs gets grants from companies that sell caffeine sports products⁹.

Our verdict: not enough evidence. If waiting suits you, there’s no known harm in it. Just don’t let a late first cup drag your last cup into the late afternoon. The JISSN authors point out that delaying would mostly push the need for an afternoon dose later⁹.

Daily limits for adults, teens and pregnancy

Adults. For most adults, the FDA cites 400 mg a day, about two to three 12-ounce cups of coffee, as an amount not generally linked with negative effects⁶. That’s a daily ceiling, not a timing rule: in the 2025 trial, 400 mg in one dose still disturbed sleep 12 hours later⁴.

Teens. The FDA notes that the American Academy of Pediatrics advises against energy drinks for children and teens⁶. The Utah Poison Control Center reported in 2025 that U.S. poison centers had seen a 24% rise in recent years in calls about children and teens who got too much caffeine from energy drinks¹². The 2025 Australia-New Zealand safety review notes that more than 1.4 mg per kilogram of body weight a day has been linked with disturbed sleep in children⁷. For a 110-pound (50-kilogram) teen, that’s about 70 mg, by our arithmetic, less than many energy drinks.

Pregnancy. The American College of Obstetricians and Gynecologists (ACOG) says under 200 mg a day doesn’t appear to raise the risk of miscarriage or preterm birth¹⁰¹¹. Some researchers argue for none at all, including a 2020 review in BMJ Evidence-Based Medicine and the JISSN authors⁹¹¹. Because the half-life can reach 18 hours near term, timing matters more⁷. Talk with your prenatal care provider⁶.

Cutting back. Stopping suddenly can bring headaches, irritability and fatigue, and caffeine withdrawal is a recognized diagnosis in the DSM-5, the psychiatric diagnostic manual⁹. The FDA calls withdrawal unpleasant rather than dangerous and suggests tapering slowly⁶. One gentle approach, our suggestion, is to move your last cup an hour earlier every few days, or switch the afternoon cup to decaf.

What the evidence says, rated

Claim	What studies found	Evidence level
Caffeine later in the day shortens and lightens sleep	Consistent results across 24 studies and randomized trials24	Strong evidence
A big dose (about 400 mg) can disturb deep sleep 12 hours later	One 23-person randomized trial4	Promising but early
About 100 mg is fine 4 hours before bed	One trial, young men only4	Promising but early
You can tell when caffeine is hurting your sleep	Ratings and diaries missed losses that monitors caught14	Evidence says no
A gene test tells you your safe cutoff	Gene effects unclear; trials too small49	Not enough evidence
Waiting 90 minutes after waking prevents an afternoon crash	No trials; the proposed reasons don’t hold up9	Not enough evidence
Plant caffeine (guarana, yerba mate) is gentler on sleep	FDA says the body treats natural and added caffeine the same6	Evidence says no

Questions readers ask

The bottom line

Caffeine is a useful plant compound, and most adults can enjoy it. The catch is that it lingers for hours and can cost you deep sleep without changing how rested you feel. A big dose late in the morning can still matter at night.

Match the cutoff to the dose: small drinks can come later, a regular coffee belongs before early afternoon, and energy drinks and pre-workout powders belong in the morning, if at all. Move your times earlier if your body clears caffeine slowly, run your own two-week test, and don’t let a delayed first cup push your last one late.

Related reading

• Sleep pressure builds all day, and your body clock decides when it wins
• Short naps restore alertness, while long ones can be a warning sign
• Nootropics run from coffee to drugs the FDA never approved
• Sleep trackers count your hours well, but their sleep stages are rough guesses
• Feeling better after trying something isn’t a test, but you can run one

Sources

1. American Academy of Sleep Medicine. “Late afternoon and early evening caffeine can disrupt sleep at night” (on Drake C, Roehrs T, Shambroom J, Roth T, Journal of Clinical Sleep Medicine, 2013). News release, November 14, 2013. aasm.org
2. Gardiner C, Weakley J, Burke LM, et al. “The effect of caffeine on subsequent sleep: A systematic review and meta-analysis.” Sleep Medicine Reviews, 2023;69:101764 (abstract record). nutrition-evidence.com
3. Petrullo J. “Caffeine and Sleep Review Suggests Consumption Guidelines.” AJMC, July 21, 2023. ajmc.com
4. Gardiner CL, Weakley J, Burke LM, et al. “Dose and timing effects of caffeine on subsequent sleep: a randomized clinical crossover trial.” Sleep, 2025;48(4):zsae230. pmc.ncbi.nlm.nih.gov
5. Australian New Zealand Clinical Trials Registry. “The Dose and Timing Relationship Between Caffeine and Subsequent Sleep Quantity and Quality.” ACTRN12621001625864, updated January 28, 2024. anzctr.org.au
6. U.S. Food and Drug Administration. “Spilling the Beans: How Much Caffeine Is Too Much?” FDA Consumer Update, content current as of August 28, 2024. fda.gov
7. Food Standards Australia New Zealand. “Supporting Document 1: Safety assessment of caffeine, Proposal P1056.” FSANZ, March 4, 2025.
8. Grzegorzewski J, Bartsch F, Köller A, König M. “Pharmacokinetics of Caffeine: A Systematic Analysis of Reported Data for Application in Metabolic Phenotyping and Liver Function Testing.” Frontiers in Pharmacology, 2022. frontiersin.org
9. Antonio J, Newmire DE, Stout JR, et al. “Common questions and misconceptions about caffeine supplementation: what does the scientific evidence really show?” Journal of the International Society of Sports Nutrition, 2024;21(1):2323919 (Nova Southeastern University record). scholars.nova.edu
10. NEJM Journal Watch. “ACOG: Moderate Caffeine Consumption Does Not Cause Miscarriage, Preterm Birth” (on ACOG Committee Opinion No. 462, Obstetrics & Gynecology, August 2010). July 23, 2010. clinician.nejm.org
11. Healio. “No caffeine consumption safe in pregnancy, study finds” (with ACOG response). August 24, 2020, updated August 27, 2020. cm.healio.com
12. Utah Poison Control Center, University of Utah Health. “Over the Limit: The Perils of Too Much Caffeine.” September 17, 2025. poisoncontrol.utah.edu
13. National Heart, Lung, and Blood Institute. “Your Guide to Healthy Sleep.” NIH Publication No. 11-5271, revised August 2011. nhlbi.nih.gov
14. American Academy of Sleep Medicine. “Daytime sleepiness affects daily activities of most adults.” News release, November 17, 2025. aasm.org
15. Thorlton J (University of Illinois Chicago). “Caffeine in Pregnancy and Lactation: What’s New?” Slides posted for a National Academies of Sciences, Engineering, and Medicine event, January 2020.