Skip to content
HabitSeek
Back to results

Avoid late-day caffeine

Avoiding caffeine sufficiently close to bedtime to preserve sleep, with timing adjusted for dose and individual sensitivity.

At a glance

  • Evidence: Avoiding late caffeine helps preserve total sleep time.Preserving total sleep time · Moderate certainty · Randomized intervention
  • What’s uncertain: Dose, timing and caffeine metabolism vary widely; a group-average cutoff is not a reliable personal threshold.
  • First step: Choose a personal caffeine cutoff by experiment rather than by rule: the 8.8 h and 13.2 h pre-bedtime figures are model-derived editorial starting points, not measured thresholds for you.
  • Before you try it: Caffeine late in the day can worsen insomnia and reduce sleep quality, so people with existing sleep problems should be especially careful about timing. All precautions

Takeaway

Caffeine taken late in the day shortens sleep, lowers sleep efficiency and delays sleep onset in controlled trials, and the effect is still detectable many hours before bedtime. The published caffeine-free intervals (8.8 h for coffee and 13.2 h for a pre-workout supplement) are model-derived editorial starting points from one meta-analysis, not measured personal thresholds, and caffeine half-life varies roughly 2-10 h between people, so no single cutoff is a rule everyone can follow.

A small cup of coffeeDiscussed by3 peopleView statements ↓
SleepEnergyFocusUnder 10 min · Easy · Low
Reviewed Oct 3, 2026
Save

Scientific evidence

What the research shows

Outcome assessments based on research, including its uncertainties and limitations.

Preserving total sleep time

Finding: Benefit supportedCertainty: Moderatestudies found a benefit from doing this

Evidence basis: randomised trials

Avoiding late caffeine helps preserve total sleep time.

Population: Healthy adults in controlled caffeine-versus-placebo sleep studies: 24 studies pooled by Gardiner et al. and 956 participants in 22 crossover trials in Chang et al.

Research details

Summary

Two independent meta-analyses of controlled trials agree that caffeine taken close to bedtime shortens objective sleep, so moving the last caffeine earlier protects sleep duration. Gardiner et al. pooled 24 studies and found total sleep time reduced by 45.3 min (the source prints reductions as positive values, 95% CI 29.0 to 61.5 min), while Chang et al. pooled 22 crossover polysomnography trials (956 participants) and reported a mean difference of -34.67 min, with reductions for both high and low doses. The effect was still detectable when caffeine was taken many hours before bedtime.

Why this certainty

Two methodologically similar meta-analyses of controlled designs found directionally identical reductions in sleep duration, but the accessible abstracts lack precision estimates and do not report heterogeneity, and the evidence applies mainly to healthy adult habitual consumers rather than the general population. Downgraded for imprecision and indirectness.

Limitations

Caffeine dose, timing and habitual intake varied widely; participants were mostly young, healthy, habitual consumers; and individual caffeine half-life varies from about 2 to 10 hours, so group means mask wide variability.

Effect estimate

Later caffeine versus placebo shortened total sleep time by 45.3 min (the source prints the reduction with a positive confidence interval, 95% CI 29.0 to 61.5 min; Gardiner 2023, 24 studies) and by a mean difference of -34.67 min (Chang 2025, 22 crossover trials, 956 participants); those are the effects that moving the last caffeine earlier in the day avoids.

Sources for this outcome:

Preserving sleep efficiency and deep sleep

Finding: Benefit supportedCertainty: Moderatestudies found a benefit from doing this

Evidence basis: randomised trials

Avoiding late caffeine helps preserve sleep efficiency and deep sleep.

Population: Healthy adults undergoing polysomnography or objective sleep monitoring after acute caffeine versus placebo: 22 crossover trials (956 participants) in Chang et al. and 24 studies in Gardiner et al.

Research details

Summary

Caffeine consistently worsened objective sleep quality across two meta-analyses. Gardiner et al. reported sleep efficiency reduced by 7%, light sleep (N1) increased by 6.1 min and 1.7%, and deep sleep (N3/N4) reduced by 11.4 min and 1.4%. Chang et al. found sleep efficiency reduced by 4.74% and slow-wave sleep proportion reduced by 1.01%, with the slow-wave effect confined to higher doses.

Why this certainty

Two meta-analyses agree on direction and approximate magnitude for sleep efficiency and deep-sleep reduction, but the slow-wave sleep effect was small, appeared only in the high-dose subgroup without a statistically significant between-group difference or a significant dose meta-regression, and precision estimates were unavailable. Certainty downgraded for imprecision and inconsistency across sleep-stage measures.

Limitations

Sleep-stage scoring, dose, timing and habitual intake differed across studies; Chang et al. found no statistically significant dose or age subgroup differences, no significant linear meta-regression with dose or age, and heterogeneity was high. The sleep-stage estimates also come from a source that prints reductions as positive values, so the minus signs here record direction rather than the source's own notation.

Effect estimate

Caffeine versus control: sleep efficiency -7.0 percentage points (95% CI -10.0 to -4.0; Gardiner 2023); N1 sleep +6.1 min (95% CI 2.3 to 9.9); N3/N4 sleep -11.4 min (95% CI -18.5 to -4.3).

Sources for this outcome:

Avoiding delayed sleep onset and night waking

Finding: Benefit supportedCertainty: Limitedstudies found a benefit from doing this

Evidence basis: randomised trials

Caffeine within about nine hours of bedtime delays sleep onset and increases night waking, so an earlier last dose protects sleep.

Population: Healthy adults in controlled caffeine-versus-placebo sleep studies (24 studies in Gardiner et al.; 22 crossover trials in Chang et al.), plus a home-based trial of 400 mg caffeine given 0, 3 or 6 h before habitual bedtime (Drake et al.).

Research details

Summary

Caffeine lengthened the time taken to fall asleep and increased wake after sleep onset, and the suggested caffeine-free interval before bed is long. Gardiner et al. reported sleep onset latency +9 min and wake after sleep onset +12 min, and modelled that coffee should be consumed at least 8.8 h and a pre-workout supplement at least 13.2 h before bedtime. In a home-based crossover trial of 12 completers, 400 mg caffeine taken 6 h before habitual bedtime significantly reduced objectively monitored total sleep time, by 41 min, although the self-reported reduction in total sleep time (p = 0.08) and the increase in reported time to fall asleep (p = 0.06) approached but did not reach significance. The authors recommended avoiding substantial caffeine for at least 6 h before bedtime.

Why this certainty

Latency prolongation is consistent across two meta-analyses, but the practically important timing thresholds rest on modelled estimates from a single review and one high-dose home trial, and no confidence intervals were available in the retrieved abstracts. Relevance to typical low-dose consumers is therefore uncertain; rated Limited.

Limitations

The 8.8 h and 13.2 h thresholds are model-derived from one meta-analysis and depend on the specific dose and product used; the home trial had only 12 completers, used a high 400 mg dose and relied on self-report plus a portable monitor, and its diary measures at 6 h only approached significance while the objective measure was significant; dose and age subgroup differences were not statistically significant in Chang et al.; half-life varies 2-10 h between individuals.

Effect estimate

Sleep onset latency +9 min (Gardiner 2023) and mean difference +8.35 min (Chang 2025); wake after sleep onset +12 min (Gardiner 2023); modelled interval before bedtime 8.8 h for coffee (107 mg per 250 mL) and 13.2 h for a standard 217.5 mg pre-workout supplement (Gardiner 2023).

Sources for this outcome:

Daytime alertness and cognitive/psychomotor performance

Finding: Mixed evidenceCertainty: Limitedstudies disagree, or the effect depends on who is studied

Evidence basis: indirect evidence, not a direct test of this exact habit

No study has tested whether an earlier caffeine cutoff changes daytime alertness; caffeine itself improves vigilance, most clearly after sleep loss.

Population: Adults in controlled acute-dosing studies; Irwin et al. synthesised sleep-deprived or sleep-restricted participants (45 publications, 327 effect estimates); supportive reviews covered athletes, military personnel and habitual caffeine consumers.

Research details

Summary

In people who are sleep-deprived or sleep-restricted, rather than rested desk workers, acute caffeine improves vigilance and reaction time: Irwin et al. pooled 45 publications (327 effect estimates) and found benefits for attention, executive function, reaction time and information processing. Reviewers caution that in habitual consumers much of the apparent benefit may reflect reversal of overnight caffeine withdrawal rather than net enhancement, and effects in rested, non-deprived users are smaller.

Why this certainty

No cited study randomised an earlier caffeine cutoff and measured alertness, so the comparison this outcome asks about was never tested; what exists is evidence about caffeine itself. Large pooled effects with confidence intervals exist, but only in sleep-deprived or restricted samples; relevance to rested users, the withdrawal-reversal confound, an internally inconsistent published interval and high neurophysiological heterogeneity limit confidence. Downgraded for indirectness and inconsistency.

Limitations

The strongest pooled effects come from sleep-deprived samples, limiting applicability to rested desk workers; one published confidence interval in the retrieved abstract is internally inconsistent (g = 0.86; 95% CI 0.53-0.83); the withdrawal-reversal confound is rarely controlled; the P300 meta-analysis reports high heterogeneity across studies.

Effect estimate

Accuracy on attention tests g = 0.68 (95% CI 0.48-0.88) and executive function g = 0.35 (95% CI 0.15-0.55) after sleep loss (Irwin 2020, 45 publications, 327 effect estimates). The same abstract also reports attention response time g = 0.86 with 95% CI 0.53-0.83, an internally inconsistent interval that needs re-checking.

Sources for this outcome:

What people say

Who discusses this?

Public statements, with positions specific to each source and topic. These views do not determine the scientific evidence assessment.

What the research shows ↑

Practical steps and precautions

How to try it

  1. Choose a personal caffeine cutoff by experiment rather than by rule: the 8.8 h and 13.2 h pre-bedtime figures are model-derived editorial starting points, not measured thresholds for you.
  2. Because caffeine half-life varies roughly 2-10 h between people, move your cutoff earlier if you notice delayed sleep onset, more night waking, or shorter sleep.
  3. If using caffeine for a workout, weigh the possible performance benefit against the sleep cost. Use the smallest amount that meets your purpose and avoid letting a late training session override your sleep cutoff; the performance evidence does not establish a universal caffeine-cycling schedule.
  4. Reduce gradually if you decide to cut back, since abrupt cessation can cause headache and fatigue.

Precautions

  • Caffeine late in the day can worsen insomnia and reduce sleep quality, so people with existing sleep problems should be especially careful about timing.
  • Caffeine clearance varies with age, sex and hormones, liver disease, obesity, smoking, diet and many medications, so sensitivity differs widely between people.
  • Check with a clinician before relying on caffeine if you have a heart condition, an anxiety disorder, or take medication that interacts with caffeine.

References

The effect of caffeine on subsequent sleep: A systematic review and meta-analysis

Gardiner C, Weakley J, Burke LM, Roach GD, Sargent C, Maniar N, Townshend A, Halson SL · Sleep Medicine Reviews · 2023

Systematic review and meta-analysis of studies measuring the effect of caffeine on subsequent night-time sleep (24 studies included)Verified from full text
Open original source

Age- and dose-specific effects of caffeine on sleep: A meta-analysis of controlled crossover trials

Chang YH, Cheng YC, Cheng WJ · Sleep Medicine · 2025

Systematic review and meta-analysis of controlled crossover trials measuring nighttime polysomnography after acute caffeine or placeboVerified from abstract
Open original source

Caffeine Effects on Sleep Taken 0, 3, or 6 Hours before Going to Bed

Drake C, Roehrs T, Shambroom J, Roth T · Journal of Clinical Sleep Medicine · 2013

Placebo-controlled study of a fixed 400 mg caffeine dose administered at 0, 3 and 6 hours before habitual bedtime, with self-reported sleep in the home and a validated portable sleep monitorVerified from abstract
Open original source

Effects of caffeine on sleep quality and daytime functioning

O'Callaghan F, Muurlink O, Reid N · Risk Management and Healthcare Policy · 2018

Integrative (narrative) review of human studies on caffeine, sleep quality and daytime functioningVerified from full text
Open original source

Interindividual Differences in Caffeine Metabolism and Factors Driving Caffeine Consumption

Nehlig A · Pharmacological Reviews · 2018

Narrative pharmacology review of caffeine metabolism, clearance, pharmacokinetics and consumption driversVerified from abstract
Open original source

Effects of acute caffeine consumption following sleep loss on cognitive, physical, occupational and driving performance: A systematic review and meta-analysis

Irwin C, Khalesi S, Desbrow B, McCartney D · Neuroscience & Biobehavioral Reviews · 2020

Systematic review and meta-analysis of acute caffeine studies in sleep-deprived or sleep-restricted individuals (45 publications, 327 effect estimates)Verified from abstract
Open original source

Caffeine and Sports Performance: The Conflict between Caffeine Intake to Enhance Performance and Avoiding Caffeine to Ensure Sleep Quality

Silva H, Del Coso J, Pickering C · Sports Medicine · 2025

Narrative review of caffeine supplementation, performance and sleep qualityVerified from abstract
Open original source

Effects of caffeine on P300: A systematic review and meta-analysis

Pinho BB, Mazer P, Barroso B, Pinto JO · International Journal of Psychophysiology · 2026

Systematic review and meta-analysis of empirical studies measuring the effect of caffeine on P300 amplitude and latencyVerified from abstract
Open original source

Community

Discussion

Have you tried this? What helped or made it difficult?

Reader experiences are separate from the scientific evidence. Explore all discussions

No discussions yet.

Be the first to ask a question or share your experience.

Sign in to vote or join the discussion.

Related habits

Explore other habits in a shared category. Each has its own evidence and precautions.

Editorial history and review method

Revision 5 · Approved for publication · Codex (publication approval explicitly authorized by site owner; category migration)

Prepared and checked with AI. Publication approval does not imply independent expert review. Request a correction.

Revision 3, applying the high-risk audit corrections (_report-focus-a.md). Changes: (1) the total-sleep-time effectEstimate keeps Gardiner's own positive-direction confidence interval (95% CI 29.0 to 61.5 min for a reduction) instead of printing a sign-flipped version as though it were the source figure, and the sleep-efficiency limitations now explain that minus signs elsewhere record direction only. (2) The total-sleep-time summary is oriented to the habit - caffeine close to bedtime shortens sleep, so moving the last caffeine earlier protects duration. (3) "Dose-dependent" was removed from the slow-wave characterisation; it now reads "only in the high-dose subgroup without a statistically significant between-group difference or a significant dose meta-regression". (4) The sleep-onset headline no longer claims that earlier cutoffs reduce risk - it describes the exposure and states the practical inference. (5) Drake is now reported accurately: 12 completers, objective total sleep time reduced by 41 min at 6 hours, with the diary measures only approaching significance (p = 0.08 and p = 0.06). (6) The daytime-alertness headline says plainly that no study has tested an earlier cutoff against alertness, and the same point was added to that outcome's certaintyRationale; the pooled benefit is now attributed to sleep-deprived or sleep-restricted participants rather than to people in general. A second reviewer must re-check: (a) the audit recommends updating drake-2013 to accessLevel full_text with the PMC3805807 URL because the added numbers are full-text values - this revision left the source record at "abstract", so that change is outstanding; (b) whether the exposure-framed evidence inside habit-framed findings is acceptable as a house convention, since all four findings keep their Benefit supported/Mixed labels; (c) the Gardiner 2023 full-text values (k/n, the 45.3 min and 7.0 percentage-point figures) remain unverifiable from this seat; (d) the published Comment/Response exchange on Silva 2025 (Sports Med 2026;56(4):1057-1061) is still unassessed. October 3 training follow-up: Clarified exercise performance versus sleep tradeoffs using the existing performance sources; did not adopt an untested cycling rule. Independent editorial approval remains pending. Publication release 2026-10-04: site owner authorized a bounded AI audit and release of this exact revision. See docs/release-audit-2026-10-04.md for source-check findings and limits. No independent health-professional or peer review is claimed. Category migration 2026-10-04: removed the separate goals field; categories: Sleep, Energy, Focus → Sleep, Energy, Focus. Research assessments, sources, guidance, precautions, and evidence review dates are unchanged. This revision awaits publication approval; no new evidence review is claimed. Publication approval 2026-10-04: site owner explicitly accepted all current category-migration revisions for deployment and requested that remaining limitations be retained in docs/content-review-backlog.md. Approval applies to this exact content revision. No new evidence review, independent expert review, or clinical validation is claimed; existing precautions and disclosures remain in effect.