Coffee, tea and the cost of alertness
In this investigation
The useful decision starts with the next night
I am interested in coffee and tea, and I own 200 mg caffeine capsules. Actual capsule use and a regular drink habit remain unreported. For the current daily shop, the useful question is whether I enjoy a drink and how it fits alertness, training and that night’s sleep.
An optional prepared morning drink is a manageable taste experiment with my limited storage. Its type, size and time belong in the log alongside any capsules actually taken. The caffeine guide develops dose and sleep timing; the food overview covers brewing, matcha, iron timing and extracts. This follow-up examines the newer claims that could otherwise turn an enjoyable drink into a supposed requirement.
Coffee days can change behavior as well as alertness
In CRAVE, 100 adults, mean age 39, received randomized instructions to consume caffeinated coffee or avoid caffeine over 14 days. Coffee days had an adjusted 36 fewer minutes of sleep and 1,058 more steps. These were secondary findings with confidence intervals unadjusted for multiple comparisons. Sleep was estimated by Fitbit in 81 participants; step data were available for 86. The primary outcome, premature atrial contractions, did not significantly differ; ventricular premature beats were higher in a secondary analysis. Marcus et al., 2023
The trial was open-label, allowed varying coffee intake and could include withdrawal effects on avoidance days. It neither isolated caffeine from the drink nor established an optimal cup or hour. UCSF and NIH funded it; Dexcom donated glucose monitors. Its practical contribution is a reason to observe sleep alongside daytime benefit. Additional steps do not establish longevity, and electrical beats are not a measured clinical cardiac event. CRAVE methods and limitations
The apparently different DECAF result concerns another endpoint: recurrence after cardioversion in 200 previous coffee drinkers with atrial fibrillation or related flutter, mean age 69. Six-month recurrence was 47% with regular coffee versus 64% with abstinence from all coffee and other caffeine; the primary hazard ratio was 0.61 (95% CI 0.42–0.89). This open-label trial used clinically detected recurrence and nonuniform monitoring. It informs a selected patient question, not a coffee prescription for a healthy young man. Wong et al., published online 2025
Long-term associations do not choose my daily dose
A 2026 cohort analysis followed 131,821 US health professionals and documented 11,033 dementia cases over a median 36.8 years. Higher caffeinated-coffee and tea intake were associated with lower dementia risk: highest versus lowest intake hazard ratios were 0.82 (95% CI 0.76–0.89) and 0.86 (0.83–0.90). Decaffeinated coffee lacked a comparable significant association. Objective cognitive testing occurred only in the women’s cohort: some measures favored intake, while caffeinated coffee’s global-cognition association was not significant. Zhang et al., 2026
Repeated dietary assessments strengthen this observational work, but residual confounding, early disease changing drinking habits and dementia misclassification remain possible. The cup ranges on its dose-response curves are observed patterns. Neither those curves nor the decaffeinated comparison establishes that caffeine prevents dementia or defines my dose at age 19–30. The study did not identify particular tea types or brewing methods. Original methods and limitations
A 2025 timing cohort associated morning coffee with lower all-cause mortality in 40,725 NHANES adults: HR 0.84 (95% CI 0.74–0.95) versus non-drinking. A separate 1,463-person sample validated the drinking-pattern classification, not the mortality result. Timing was recalled on dietary survey days rather than randomized, and shift work or other lifestyle differences could contribute. Earlier use may be a sensible sleep choice; this study does not prove that moving coffee to the morning extends life. Wang et al., 2025
Theanine is an unresolved addition, not a reason to buy a stack
A 2025 counterbalanced crossover study analyzed 37 of 40 enrolled sleep-deprived adults aged 22–30. Capsules containing 200 mg theanine plus 160 mg caffeine improved selected traffic-scene attention measures versus placebo, including a 38 ms greater reaction-time improvement. There was no caffeine-only arm, so theanine’s additional contribution remains unknown. Multiple outcomes were tested without adjustment. The authors reported university/foundation funding and no conflicts. This was a computer task, not actual driving, recovered sleep or an ordinary tea drink. Nawarathna et al., 2025
The current registry describes alternating treatment order, labels allocation non-randomized and also uses “random counter-balanced” wording. Its several primary outcomes prevent treating this as a clean single-primary randomized result. These inconsistencies limit confidence in the allocation and outcome hierarchy; they do not erase the observed task result. Trial registration
An earlier four-condition crossover trial in 24 young adults did compare separate ingredients: 75 mg caffeine, 50 mg theanine, both and placebo. The combination showed no positive behavioral effect at those exposures, despite altered cerebral hemodynamics. Within-measure pairwise follow-ups used Bonferroni correction. This dose-specific result does not refute every combination, but it leaves an ordinary tea’s special cognitive advantage unestablished. Dodd et al., 2015
What changes the next shopping trip?
There is no required drink addition here. If I want to try coffee or tea, one enjoyable prepared drink earlier in the day is enough to start observing the experience. It can fit a daily purchase without depending on refrigerator space or buying several powders. A caffeine-free option can supply a similar evening ritual; its ingredients still matter, and it need not be sold as a sleep treatment.
The useful log is small: drink and preparation, approximate amount, time, capsules actually taken, daytime alertness and that night’s sleep. Taste and perceived sleep are observations; neither supplies a clinical outcome. If an enjoyable drink repeatedly coincides with worse sleep, that is a more immediate reason to adjust amount or timing than a distant disease-prevention headline is to increase it. No exact personal cutoff has been established.