Does coffee lift your mood or rattle your nerves? Fifteen caffeine trials say it depends, and the anxiety signal showed up mainly above 200 mg

Caffeine is one of the most widely used mood-altering substances on the planet, and most of us have a theory about what it does to our heads. Large observational studies have found that people who drink coffee report less depression, while others have tied heavy caffeine intake to anxiety. The trouble with those studies is that coffee drinkers differ from non-drinkers in many other ways. A systematic review published in Frontiers in Nutrition on 29 October tries to cut through that by pooling only randomized controlled trials, where people were given a measured dose of caffeine or a caffeine-free comparison and then had their mood scored.

The team, from Nanjing Medical University and Zhejiang University in China, searched eight English and Chinese databases up to 16 May 2026 and screened 788 records. Only 15 trials qualified, published between 1994 and 2025 and run in the United States, China, Tunisia, Colombia, Brazil, the United Kingdom, Iran and Sweden. Together they contributed 668 people. The mix was wide: healthy students, elite wrestlers and judokas, schoolchildren given caffeinated cola, postmenopausal women with an overactive bladder, and patients with depression, panic disorder or social anxiety. Caffeine came as capsules, tablets, coffee, tea, energy drinks or spiked orange juice, from 60 mg a day up to a single 480 mg dose, and most trials looked at the hour or two after one dose rather than weeks of drinking.

For depression, seven comparisons were pooled and the answer was no clear difference between caffeine and control. The two trials that hinted at a benefit were both in hospital patients with depression in China: one gave 60 mg caffeine tablets for four weeks, the other a daily coffee with about 95 mg of caffeine for 12 weeks alongside antidepressant treatment, compared with decaf. The authors say trials run next to antidepressant therapy can't be read as caffeine working on its own. In healthy people, single doses made no reliable difference to low mood, and in two longer trials in people without a psychiatric diagnosis the estimates were small and uncertain.

Anxiety was more consistent in direction. Across 16 comparisons from 12 trials, anxiety scores were higher after caffeine than after the control, a standardized difference of 0.51, roughly a moderate shift. But the spread between trials was large, and the range of effects a new study might find ran from a fall in anxiety to a large rise. When the authors split trials by dose, the picture sharpened: at 200 mg or less, about two cups of brewed coffee, there was essentially no change; above 200 mg the pooled rise was 0.97, and that held when they dropped trials whose doses were set by body weight. One parallel trial in graduate students, who drank 710 mL of regular coffee a day for five days (450 mg of caffeine), showed a very large anxiety jump (a standardized difference of 2.30).

The authors are blunt about the limits. Certainty was rated very low for both outcomes, because of small and older trials, unclear reporting of how people were randomised, very different groups and questionnaires, and signs that small studies with striking results were over-represented for anxiety. Most trials measured how people felt within an hour or two of a dose, often a big one taken in a lab, which is not the same as a morning cup taken every day for years. Scores on an anxiety questionnaire going up is also not the same as developing an anxiety disorder.

So the paper does not set a safe or harmful dose and does not change existing intake advice. What it does suggest is that the comforting idea of coffee as a mood booster rests mostly on observational data, and that if you notice jitters or a racing mind after a large dose, that is consistent with what the trials saw. The authors call for longer trials with clearer doses, better records of people's usual caffeine habits, and a look at genes that change how fast caffeine is cleared and how strongly the brain responds to it.