Heavy Caffeine Use May Have a Surprising Effect on Deep Sleep

Coffee has a well-earned reputation for keeping people awake. But what happens when someone drinks enough caffeine every day that their brain begins adapting to it?

A large 2025 study uncovered a curious pattern. People who consumed four or more caffeinated beverages a day slept for less time, yet when researchers measured their brain activity overnight, their non-REM sleep was actually deeper. Just as surprisingly, the heavy caffeine users didn’t report feeling that their sleep was significantly worse.

Researchers from the University of Zurich and Lausanne University Hospital combined genetic information from nearly half a million UK Biobank participants with detailed at-home sleep recordings from the Swiss HypnoLaus cohort. Published in the Journal of Psychopharmacology, the study used several statistical approaches to investigate how habitual caffeine consumption may affect sleep.

The findings don’t mean that drinking more coffee improves sleep. Instead, they suggest something more complicated: the brain may partially compensate for caffeine-related sleep loss by increasing the depth of the sleep it does get.

Whether that compensation is enough to make up for lost sleep is another question entirely.

Study Design and Methods

The HypnoLaus Cohort

The HypnoLaus study is a population-based sleep cohort embedded within the larger CoLaus/PsyCoLaus study in Lausanne, Switzerland. Participants underwent complete polysomnographic recordings at home and had extensive phenotyping for a range of cardiometabolic and psychiatric conditions.

Polysomnography (PSG) – the simultaneous recording of brain electrical activity (EEG), eye movements, muscle tone, and other physiological signals during sleep – is the established reference standard for objective sleep measurement. Conducting PSG at home, as HypnoLaus did, rather than in a controlled laboratory removes the confounding effect of an unfamiliar sleeping environment on results – a meaningful methodological advantage when studying habitual, real-world sleep patterns.

Mendelian Randomization and Causal Matching

The Mendelian randomization approach makes use of the natural random allocation of single-nucleotide polymorphisms – specific genetic variants – that influence caffeine intake, as estimated from the UK Biobank (n = 485,511). Tracking this genetic influence enables estimating causal effects of caffeine on both objective and subjective sleep variables. The researchers applied three distinct Mendelian randomization estimators – MR-Egger, inverse variance weighting, and weighted median – and supplemented them with causal matching, an independent statistical approach that pairs individuals with similar background characteristics across caffeine intake groups. Agreement across all three methods strengthens the confidence that can be placed in the findings.

To further support the causal analysis, participants were matched across caffeine intake groups based on key characteristics of the HypnoLaus dataset, ensuring greater comparability with respect to known confounders. The study compared two groups: those reporting high habitual caffeine intake (four or more caffeinated beverages per day) against those reporting moderate intake (three or fewer per day).

The sleep outcomes measured were both objective and subjective. Objective measures included total sleep time in minutes, sleep latency – the time between lights-out and first occurrence of stage N2 sleep – number of awakenings, REM sleep as a percentage of total sleep time, EEG delta power in NREM sleep (spectral power in the 1 – 4 Hz range), and EEG sigma power in NREM sleep (spectral power in the 12 – 16 Hz range). Subjective measures included the Pittsburgh Sleep Quality Index (PSQI), the Epworth Sleepiness Scale (ESS), and the Morningness-Eveningness Questionnaire (MEQ).

Key Findings: Sleep Duration, Depth, and the Adaptation Hypothesis

Sleep Is Shorter – Substantially So

All statistical models revealed that four or more caffeinated beverages per day shorten total sleep time when compared to fewer caffeine-containing drinks per day, with estimated reductions in sleep length ranging from 11 to 229 minutes. The range is wide – and deliberately so. The different statistical estimators applied to different analytic assumptions yielded different point estimates, with the most conservative methods producing smaller effects and more permissive approaches producing larger ones. The authors treat the full range as informative rather than collapsing it into a single figure. The practical implication is unambiguous in all cases: habitual high caffeine intake costs sleep time.

This is consistent with controlled laboratory evidence. A 2025 randomized clinical crossover trial published in Sleep found that 400 mg of caffeine consumed within 12 hours of bedtime caused significant delays in sleep initiation and alterations to sleep architecture. The Zurich/Lausanne study extends these findings from acute, controlled exposures to habitual, free-living consumption.

NREM Sleep Depth Increases

Consistent with the homeostatic facet of sleep-wake regulation, the shorter sleep in high habitual caffeine consumers was characterized by increased non-rapid-eye movement sleep depth, as measured by all-night electrical brain activity.

NREM sleep – often called slow-wave sleep – is dominated by high-amplitude delta oscillations, the slow, synchronized brain waves that characterize the deepest stage of sleep. The depth of NREM sleep is regulated by sleep homeostasis: the biological system that tracks how long a person has been awake and accumulates sleep pressure accordingly.

Caffeine blocks adenosine receptors in the brain. Adenosine is an endogenous sleep-regulatory substance – it accumulates in the brain during waking hours and drives the mounting pressure to sleep. When caffeine blocks adenosine receptors, it suppresses this pressure signal while adenosine continues to build. Once caffeine clears the system, the accumulated adenosine floods the receptors. In a habitual daily drinker, sleep pressure on any given night may therefore be higher than in a non-caffeine user – and that elevated homeostatic pressure appears to manifest as deeper NREM sleep when sleep does occur.

The researchers interpret this as evidence of adaptation. After prolonged daily use, the brain’s sleep-wake regulatory system appears to partially compensate for chronic adenosine-receptor blockade. Crucially, subjective self-reported measures of sleep quality – including Pittsburgh Sleep Quality Index global scores, Epworth Sleepiness Scale scores, and Morningness-Eveningness Questionnaire scores – did not differ between the high-caffeine and moderate-caffeine groups. High-caffeine consumers slept less by objective measurement, yet rated their sleep quality similarly to those drinking fewer caffeinated beverages. This discrepancy between objective and subjective assessment is a recurring theme in sleep research and has significant implications for how habitual users perceive their own sleep.

What Did Not Change

Not every sleep variable was affected. REM sleep percentage, sleep latency, and number of awakenings were not consistently altered across all statistical estimators. Self-rated chronotype – whether someone is a morning or evening person – also did not differ. The authors note this selectivity: caffeine’s habitual effects appear to be specific to sleep duration and NREM depth rather than a global disruption of sleep architecture. They caution that this does not mean habitual high caffeine use is without cost.

The Broader Context: Population-Level Caffeine Consumption

According to a 2025 study funded by the Institute for the Advancement of Food and Nutrition Sciences, approximately 69% of the U.S. population consumed at least one caffeinated beverage per day. Among caffeine consumers, the mean daily caffeine intake was 210 mg – roughly equivalent to two standard cups of brewed coffee. Coffee is the largest contributor to caffeine intake across all age groups at 69%, followed by carbonated soft drinks at 15.4%.

The U.S. Food and Drug Administration sets its upper safe limit at 400 mg per day for healthy adults – approximately four standard cups of brewed coffee. High habitual consumers, as defined in the Zurich/Lausanne study at four or more caffeinated beverages per day, are therefore operating near or at the regulatory ceiling. Whether that level of consumption represents a meaningful health concern in the context of sleep, beyond the duration reduction already documented, remains an open question that this research does not fully resolve.

Sleep Physiology: Why Duration Still Matters

Even granting the adaptation argument, that deeper NREM sleep may partially compensate for shorter total sleep time, whether that compensation is complete remains unresolved. Sleep duration and sleep depth are not interchangeable.

Both NREM and REM sleep play important roles in normal sleep physiology, including processes involved in memory and learning. Research suggests that slow-wave NREM sleep is particularly important for the consolidation and reorganization of newly learned information, while REM sleep may contribute to later stages of memory processing. REM sleep also becomes more prominent during later sleep cycles, which is one reason total sleep duration still matters. Importantly, however, the Zurich/Lausanne study did not find consistent changes in the percentage of sleep spent in REM across its statistical models.

Research on sleep deprivation consistently documents effects on attention, memory, decision-making, executive function, and emotional reactivity. This matters because deeper NREM sleep should not automatically be interpreted as fully compensating for lost sleep time. The Zurich/Lausanne study did not examine long-term cognitive or health outcomes. What it establishes is evidence of a homeostatic response to habitual caffeine use, not that shorter sleep becomes harmless.

The study’s authors also acknowledge several limitations. Caffeinated beverages consumed per day were self-reported, making the measure susceptible to recall error. Caffeine content wasn’t standardized either, so “four caffeinated beverages” could represent very different doses depending on whether someone drinks filtered coffee, espresso, tea, soda, or energy drinks. Additionally, the HypnoLaus cohort consisted predominantly of middle-aged adults in Lausanne, Switzerland, limiting how broadly the findings can be applied to other ages, populations, and caffeine-consumption patterns.

Caffeine metabolism also varies considerably between individuals, partly because of genetic differences affecting the CYP1A2 enzyme, which plays a major role in metabolizing caffeine. Caffeine’s half-life can vary substantially between people, meaning two people who consume the same amount at the same time may experience very different effects hours later. Population-level findings can therefore obscure important differences in how individuals respond to caffeine.

Deeper Sleep Doesn’t Necessarily Mean Better Sleep

At first glance, the findings seem almost contradictory. People consuming four or more caffeinated beverages per day slept less, but their brains showed greater NREM sleep depth. Meanwhile, they didn’t report substantially worse sleep than people consuming less caffeine.

One explanation is adaptation. Caffeine blocks receptors for adenosine, a chemical involved in building sleep pressure throughout the day. With habitual caffeine exposure, the brain’s sleep-regulating systems may partially compensate, producing deeper NREM sleep once sleep finally occurs.

But compensation isn’t the same thing as cancellation.

The study doesn’t establish that deeper NREM sleep completely makes up for sleeping fewer hours. Sleep duration and sleep depth measure different things, and the researchers didn’t follow participants long enough to determine whether this pattern translates into better or worse long-term health, cognition, or daytime functioning.

There is also enormous variation between individuals. Caffeine content differs dramatically between beverages, people metabolize caffeine at different rates, and the timing of that last coffee may matter just as much as the number of cups consumed. Someone who can drink coffee after dinner and fall asleep easily isn’t necessarily experiencing caffeine the same way as someone who becomes restless after an afternoon cup.

So the takeaway isn’t that four cups of coffee will somehow make your sleep deeper and therefore healthier. It’s that habitual caffeine use may change sleep in ways we don’t necessarily notice ourselves. The brain appears capable of adapting to some of caffeine’s effects, but that adaptation doesn’t make the lost sleep disappear.

For regular coffee drinkers, the more useful question may therefore be less about whether caffeine is “good” or “bad” for sleep and more about something simpler: Are you getting enough sleep, and could the timing or amount of caffeine you’re drinking be quietly cutting into it?

Disclaimer: This information is not intended to be a substitute for professional medical advice, diagnosis, or treatment and is for information only. Always seek the advice of your physician or another qualified health provider with any questions about your medical condition and/or current medication. Do not disregard professional medical advice or delay seeking advice or treatment because of something you have read here.

AI Disclaimer: This article was created with the assistance of AI tools and reviewed by a human editor.