L-TheanineAmino Acid

• AttentionLikely helps

  A 2026 systematic review and meta-analysis of 31 randomized trials, 1,168 participants, found this the most consistently supported acute effect [6].
  
  A single 200 mg dose taken 30–60 minutes before cognitive testing improved choice reaction time against placebo, at SMD 0.51 (95% CI 0.25–0.77) [6]. ^[6]

• Acute stressPossibly helps

  The same meta-analysis put the acute stress-reduction effect at SMD 0.31, and described it as largely influenced by studies at high risk of bias [6].
  
  No significant effect on fatigue was found across the trial set [6]. ^[6]

• AnxietyPossibly helps

  This is the most-marketed use and the weakest-supported one in these sources.
  
  Across the 31-trial meta-analysis, anxiety effects were inconsistent and non-significant, with one exception: a single trial in psychotic-spectrum anxiety at 400 mg/day for 8 weeks, SMD 0.54 [6].
  
  Drugs.com cites a pooled analysis of 3 studies, 78 people, where 200 mg produced no anxiety effect 40–70 minutes after dosing [2]. ^[6]

• Anxiety — why the trials might disagreePossibly helps

  A 4-week randomized crossover trial of 200 mg/day in 30 healthy Japanese adults did find a significant trait-anxiety reduction, at p=0.006 [7].
  
  The acute studies dosed once; this one dosed daily for a month [6][7].
  
  Duration of use and outcome timing may account for the disagreement, rather than the compound behaving differently. ^[7]

• Sleep qualityPossibly helps

  That same 4-week trial lowered Pittsburgh Sleep Quality Index scores versus placebo, at p=0.013 [7].
  
  A 2025 systematic review of 13 sleep trials in 550 people concluded 200–450 mg/day looks safe and effective for sleep quality [3].
  
  Its authors work for a company that sells L-theanine-containing products, which is worth weighing against the conclusion [3]. ^[3]

• Mood in healthy adultsPossibly helps

  Excluding one statistical outlier, the 31-trial meta-analysis found a significant reduction in depressive symptoms after a single dose in healthy people, SMD 0.69 (95% CI 0.13–1.25) [6].
  
  The review authors flagged it as needing confirmation in larger, higher-quality trials in clinical populations [6].
  
  The 4-week trial separately found lower depression-scale scores (p=0.019), better verbal fluency (p=0.001) and better executive function (p=0.031) [7]. ^[6]

• A 28-day branded-ingredient trialPossibly helps

  A 28-day randomized double-blind trial of a branded L-theanine at 400 mg/day in 30 adults with moderate stress reported lower perceived stress and less light sleep [8].
  
  Sleep quality improved, and Stroop reaction time improved by about 33% by day 28, with no clinically meaningful safety signals [8].
  
  One author is an employee of the company marketing that ingredient — a disclosed financial conflict worth weighing against an otherwise well-controlled design [8]. ^[8]

• Schizophrenia adjunctPossibly helps

  Small trials added 250–400 mg/day for 8 weeks on top of ongoing antipsychotic medication [2].
  
  PANSS positive and general psychopathology scores improved, and one trial also reported better sleep-quality scores [2]. ^[2]

• Suntheanine (enzymatic)The form used in most trials, and the GRAS filing designed to yield pure L-isomer with no D-theanine

  Elemental content

  Made by reacting food-grade L-glutamine with ethylamine using glutaminase; GRAS from February 2007, GRN 209 [4]

  Absorption

  In a randomized human crossover trial, 100 mg in capsule form peaked at about 24.3 µmol/L, plus or minus 5.7, roughly 0.8 hours after intake [12].
  
  The same 100 mg drunk as green tea peaked at 26.5 µmol/L, plus or minus 5.2 [12].

  Best for

  The form used in most trials, and the GRAS filing designed to yield pure L-isomer with no D-theanine [4][12]

• Chemically synthesizedA different manufacturing route to the same molecule on paper

  Elemental content

  A separate ingredient from Zhejiang Tianrui Chemical; GRAS from July 2014, GRN 501 [5]

  Absorption

  No published head-to-head human pharmacokinetic comparison against the enzymatic ingredient was found in this research pass.
  
  That is a genuine evidence gap, not a finding either way.

  Best for

  A different manufacturing route to the same molecule on paper [5]

• Green tea leaf extractWhole-leaf sourcing, at the cost of the caffeine that comes with it

  Elemental content

  Extracted from tea, which is naturally about 98.2% L-isomer with only trace D-theanine [12]

  Absorption

  Drugs.com puts peak plasma at roughly 32–50 minutes and half-life at about 58–74 minutes, and calls capsule and green-tea pharmacokinetics comparable [2].
  
  Half-life estimates vary modestly across the sources reviewed here, which likely reflects dose and population rather than a real dispute.

  Best for

  Whole-leaf sourcing, at the cost of the caffeine that comes with it [2][12]

• Loose powder and chewablesNo documented advantage — the capsule-versus-tea comparison is the only one that has actually been run

  Elemental content

  Same molecule, different delivery format

  Absorption

  No published human study compared loose powder against capsules, and none covered chewable formats [12].
  
  Another gap rather than a result.

  Best for

  No documented advantage — the capsule-versus-tea comparison is the only one that has actually been run [12]

• Headache, and an HDL change

  One study among elderly participants recorded more headaches on four daily 250 mg doses [2].
  
  A major depressive disorder adjunct trial at 250 mg/day for 8 weeks measured a small drop in mean HDL cholesterol, -4.4 mg/dL, at p=0.011 [2].
  
  Most published trials are small and report adverse events inconsistently, so isolated-theanine safety data stay thin [2]. ^[2]

• Pregnancy and breastfeeding

  The NIH's LactMed database, revised May 2025, found no direct published information on theanine use during breastfeeding [15].
  
  It calls the amounts naturally present in green tea likely acceptable while nursing, but says amounts from high-dose supplements are not necessarily safe for a breastfed infant [15].
  
  It recommends avoiding theanine supplements while nursing a newborn or preterm infant, noting a roughly 1-hour half-life should clear it from milk in about 3–5 hours [15].
  
  Drugs.com separately states pregnancy and lactation safety and efficacy information is simply lacking [2]. ^[15]

• Those blood-pressure findings do not agree

  A third small crossover trial, 18 people, added 200 mg to a food product and found no acute difference in blood pressure or heart-rate variability over 90 minutes [2].
  
  Three studies, three outcomes, differing by design and stress type — read the effect as plausible but inconsistent [2]. ^[2]

• Documented drug interactions are thin

  Drugs.com's professional monograph states outright that clinical interaction data are none well documented, and that no contraindications are well established [2].
  
  That is a notably more cautious position than the detailed interaction-warning tables published on several consumer health sites [2]. ^[2]

• Children

  The only robust pediatric trial identified is a 6-week randomized double-blind study in 93 boys aged 8–12 with diagnosed ADHD, at 400 mg/day as 200 mg twice daily [16].
  
  It was well tolerated with no significant adverse events, and improved sleep efficiency, awakenings after sleep onset and nocturnal activity [16].
  
  It did not improve sleep latency or total sleep time [16]. No data covers younger children, girls, or non-ADHD populations [16]. ^[16]

• Blood pressure with caffeine and stress

  In one small crossover study, 200 mg muted the heart-rate increase from an acute stress test [2].
  
  In another, 200 mg attenuated systolic and diastolic rises during a psychological — but not a physical — stress test, in high-stress-response adults [2].
  
  A physiological-stress study found only a modest treatment-by-group interaction on blood pressure, F(2,24)=3.438, p=0.049 [14]. ^[2]

Tea (Camellia sinensis and two related Camellia species) is effectively the only dietary source, alongside trace amounts in one wild mushroom reported in a single 1960 paper [2].

Which tea carries the most is genuinely disputed across well-controlled HPLC studies, and we are not picking a winner.

A 2025 study brewing six tea types under controlled conditions, 80°C for 2 minutes, put black and green roughly level at about 12.3 and 12.5 mg per 200 mL [17]. White and oolong came in at 1.8 mg, yellow at 4 mg and Pu-erh at 3.1 mg [17].

An analysis of 37 commercial samples ranked green highest per dry weight at 6.56 mg/g, then white 6.26, oolong 6.09 and black 5.13, with no detectable theanine in pu-erh at all [18].

A 2011 UK study of single prepared cups found the opposite ordering: black tea at 24.2 mg per 200 mL, plus or minus 5.7, against green tea at 7.9 mg, plus or minus 3.8 [19].

All three agree pu-erh is lowest [17][18][19]. Cultivar, growing region, leaf grade, degree of fermentation and brewing practice all move the number [17].

Longer steeping consistently raised extracted theanine, while temperature mattered less for theanine than for caffeine [17][19]. Adding more than 50 mL of milk per 200 mL cup markedly lowered measured theanine in one study [19].

Trial doses of 200–450 mg/day sit far above what any cup delivers, so tea alone would mean many cups a day — at which point the caffeine becomes the practical limit [3].

There is no single validated best time; the trial literature is use-case dependent rather than settled [3].

For sleep, 8 of 13 trials in a 2025 systematic review dosed in the hours immediately before bed [3]. The review's 200–450 mg/day conclusion rests on that timing pattern [3].

Those authors are affiliated with a company selling L-theanine products — a disclosed conflict worth weighing, even against a rigorous PRISMA-based method [3].

For stress and focus, the cited trials dosed in the daytime, tied to an acute stress challenge or taken daily on an ongoing basis [2][8].

Pharmacokinetically, peak plasma arrives at roughly 32–50 minutes, with EEG alpha-wave changes linked to relaxed alertness appearing within about 30 minutes [2].

No source reviewed here tested empty stomach against with food. That one is an evidence gap.