Does L-Theanine Reduce Stress and Improve Sleep?

Does L-theanine reduce stress and improve sleep?

For sleep, partly. For stress, this trial did not show it. A randomized, double-blind, placebo-controlled crossover trial gave 30 healthy adults 200 mg of L-theanine nightly for 4 weeks, with a 2 week washout before crossing to the other condition. Scores improved while participants were taking L-theanine, but when those changes were compared head to head against placebo, most of the differences did not reach statistical significance.

The comparisons that did beat placebo were three specific sleep components: how long it took to fall asleep, night time sleep disturbances, and use of sleep medication.

What the trial found:

Sleep latency, sleep disturbance, and sleep medication use: significantly better than placebo (p = 0.0499, 0.046, and 0.047)
Overall sleep quality score (PSQI): improved on L-theanine from 9.67 to 8.30 (p = 0.013), but the change versus placebo was only a trend (p = 0.073)
Trait anxiety (STAI-T): improved on L-theanine from 51.9 to 48.5 (p = 0.006), but versus placebo, p = 0.13, not significant
Depression score (SDS): improved on L-theanine from 45.8 to 43.2 (p = 0.019), but versus placebo, p = 0.084, not significant
Verbal fluency: improved on L-theanine (p = 0.001), but versus placebo, p = 0.22, not significant overall
Attention, working memory, motor speed, Trail-Making, and Stroop: no significant effect
Cortisol, immunoglobulin-A, glucose, and HbA1c: unchanged

The distinction that matters here

Almost every popular summary of this study, including the earlier version of this page, reports the improvements that happened while people were taking L-theanine and stops there. That is the wrong number to quote.

In a placebo-controlled trial, the within-arm change tells you very little. People improve over four weeks for all sorts of reasons: regression to the mean, getting used to the questionnaires, the attention that comes with being in a study. The number that isolates the compound is the comparison between conditions, and in this trial most of those comparisons were null.

So “trait anxiety fell from 51.9 to 48.5, p = 0.006” is true and also not evidence that L-theanine reduced anxiety. When the same participants took placebo, the head to head difference came out at p = 0.13.

Where the signal was real

Two findings survive that scrutiny.

Specific sleep components. Sleep latency, sleep disturbances, and sleep medication use each beat placebo, though all three sit close to the 0.05 threshold. The overall sleep quality score did not, and neither did the subscales for sleep duration or efficiency. So the readable claim is narrow: people fell asleep somewhat faster and were disturbed somewhat less, not that their sleep was transformed.

Verbal fluency in lower baseline performers. The authors split the group at the median by baseline cognitive score. In the 15 participants who started lower, verbal fluency and letter fluency both improved significantly against placebo (p = 0.002 for each). In the 15 who started higher, nothing was significant. This is a subgroup analysis in an already small trial, so treat it as a hypothesis rather than a result, but the direction is at least coherent: more room to improve, more measurable improvement.

Dr. Kumar’s Take

I like L-theanine and I think this trial is worth reading, but it is a much smaller and softer result than the supplement industry implies.

Thirty people is a small trial. The crossover design helps, since every participant serves as their own control, and adherence was excellent at 100% with serum L-theanine confirmed to rise. Those are real strengths. What it cannot overcome is that the study was not powered to detect modest between-condition differences, so a p of 0.13 on anxiety does not prove absence of effect either. It genuinely leaves the question open.

What I take from it practically: if your complaint is that you lie awake at the start of the night, 200 mg before bed is a low risk thing to try, and this is the outcome with the best support here. If your complaint is daytime anxiety or stress, this trial does not give you a reason to expect much, and I would not spend money on it for that indication on the strength of this evidence alone.

The cortisol result is worth dwelling on. Cortisol, immunoglobulin-A, glucose, and HbA1c all stayed flat. If L-theanine were meaningfully changing the physiological stress response, cortisol is where I would look first, and it did not move.

One practical detail that gets garbled constantly: in this study the dose was taken once nightly before sleep, not split across the day. If you are trying to reproduce the sleep finding, reproduce the timing.

Study Design

Randomized, double-blind, placebo-controlled crossover trial in 30 healthy adults with no major psychiatric illness, 9 men and 21 women, mean age 48.3 years with a range of 20 to 69. Each participant took 200 mg of L-theanine daily and placebo daily for 4 weeks each, in randomized order, separated by a 2 week washout. Outcomes were the Self-rating Depression Scale, the trait subscale of the State-Trait Anxiety Inventory, the Pittsburgh Sleep Quality Index with its seven components, and a cognitive battery including the Trail-Making Test, the Stroop test, and the Brief Assessment of Cognition in Schizophrenia. Serum cortisol, immunoglobulin-A, glucose, HbA1c, and L-theanine concentration were measured as secondary outcomes.

Limitations

Thirty participants is small, and the study was underpowered for the between-condition comparisons that matter most. Participants skewed female and middle aged, so the results do not necessarily transfer to younger adults or to men. The primary outcomes were self-reported questionnaires, with no polysomnography to confirm the sleep findings objectively. The cognitive subgroup analysis was defined by a median split, which is exploratory. Four weeks tells you nothing about longer term use. The study also used product supplied by a commercial manufacturer of L-theanine, with co-authors affiliated to that company, which is disclosed in the paper and worth factoring in.

FAQs

How much L-theanine did the study use, and when?

200 mg once daily, taken orally before sleep, for 4 weeks. Compliance was 100% and serum L-theanine concentration rose significantly, confirming participants took it.

Does L-theanine actually lower anxiety?

This trial does not establish that. Trait anxiety scores fell while participants took L-theanine, but the comparison against placebo came out at p = 0.13. The authors ran a power calculation before the trial and reported that 30 participants per condition gave better than 80% power for the effect size they expected, so the null result on that particular comparison is not simply a matter of too few people.

Can I get this dose from green tea?

Not realistically. Reaching 200 mg would take many cups, and the amount per cup varies widely by cultivar, leaf grade, and brewing. If you are trying to reproduce this trial, a measured supplement is the only way to know your dose.

Were there any side effects?

None were reported. Compliance was 100% across both conditions, and no adverse events occurred during the study.

Is L-theanine safe with caffeine?

The two occur together naturally in tea and the combination is widely used and well tolerated. This particular trial did not test that combination, so nothing here speaks to it directly.

Bottom Line

A 30 person crossover trial found that 200 mg of L-theanine before bed produced significant improvements against placebo on three specific sleep components: how fast people fell asleep, how much their sleep was disturbed, and their use of sleep medication. Overall sleep quality, anxiety, depression scores, cognition, and cortisol did not separate from placebo. It is a reasonable thing to try for trouble falling asleep and a weak bet for daytime stress.

Read the study

Hidese S, Ogawa S, Ota M, Ishida I, Yasukawa Z, Ozeki M, Kunugi H. Effects of L-Theanine Administration on Stress-Related Symptoms and Cognitive Functions in Healthy Adults: A Randomized Controlled Trial. Nutrients. 2019;11(10):2362.