Caffeine is the most studied legal supplement in endurance sport, and it has a clear working range: 3–6 mg per kilogram, starting at the lower end. In September another idea did the rounds again in biohacking channels: take not caffeine itself but paraxanthine — a “stimulant without the side effects”. Let us look at what the substance is and what it showed in the two studies in which it was given to athletes.

What paraxanthine is

In the liver, caffeine (1,3,7-trimethylxanthine) is broken down by the enzyme CYP1A2 into three metabolites. About 70–72% is converted into paraxanthine (1,7-dimethylxanthine); the rest becomes theobromine and theophylline. In other words, much of what “works” after a cup of coffee is already paraxanthine.

The supplement makers' arguments go like this:

it stays in the blood for a shorter time — in the classic study by Lelo and colleagues (British Journal of Clinical Pharmacology, 1986), in six healthy men the half-life of caffeine was 4.1 hours and that of paraxanthine 3.1 hours;
it does not depend on how fast CYP1A2 works, which varies genetically between people;
in laboratory and animal models it causes less anxiety and toxicity and binds more strongly to adenosine receptors.

Note the figures: the difference in half-life is about an hour, not “several-fold”. In the same study, the clearance of paraxanthine and caffeine was almost identical.

Study 1: 12 runners and a ten-kilometre time trial

Choongsung Yoo and colleagues from Texas A&M University (Journal of the International Society of Sports Nutrition, 2024) gave 12 trained runners (11 men, 1 woman, VO₂max around 52 ml/kg/min) the following across four visits: placebo, 200 mg of caffeine, 200 mg of paraxanthine, or both at 200 mg each. This was followed by a 10 km time trial on a treadmill at race pace, with cognitive tests before and after.

Running: no difference. Finishing time (48.4 minutes on average), heart rate, perceived exertion, lactate and a subsequent 30-second sprint on a cycle ergometer did not differ between any of the conditions.

Cognition: a partial plus. After the run on paraxanthine, the share of correct answers in a card-sorting test rose by 6.8% relative to the pre-run level, and reaction time was 23% faster than on placebo. Paraxanthine did not differ from caffeine on reaction time, and combining the two substances added nothing.

An important detail: the study was funded by the manufacturer of patented paraxanthine, and three co-authors are its employees and named on its patent applications. According to the authors, data collection and analysis were carried out without their involvement.

Study 2: 14 rowers and an evening two-kilometre test

The study by Bingol Diedhiou and colleagues (JISSN, 2026) was carried out without external funding. 14 student rowers (aged 22, with a habitual intake of 335 mg of caffeine a day) completed a 2000 m rowing ergometer test at 19:00–20:00 — deliberately in the evening, so that sleep could be assessed as well. They took placebo, 200 mg of caffeine, 200 mg of paraxanthine, or 200 + 200 mg, 45 minutes before the start.

Caffeine + paraxanthine: 405.9 s vs 409.2 s on placebo — 3.3 s faster (p = 0.044, small effect, d = 0.30), with mean power of 335 W vs 327 W.
Caffeine alone: 406.6 s — in the same direction, but not statistically convincing (p = 0.087).
Paraxanthine alone: no clear improvement.

Sleep was assessed the next morning with a single self-rated scale and a daytime sleepiness scale. The combination noticeably worsened sleep and increased daytime sleepiness; caffeine alone pointed the same way; with paraxanthine alone the picture was milder.

The catch in the “synergy”

The authors see the better result with the combination as a sign that the two substances interact. But that condition involved a double dose of methylxanthines — 400 mg versus 200 mg in the other conditions — and some of the caffeine is itself converted into paraxanthine. Blood levels were not measured, so “synergy” cannot be told apart from “simply more stimulant”, as the authors honestly acknowledge.

Other limitations: 200 mg is about 2.6 mg/kg, the lower end of caffeine's working range; men only; sleep was assessed by questionnaire, without actigraphy.

How to apply this

Race morning — caffeine remains the main tool: 3 mg/kg about an hour beforehand, then adjust to tolerance. There is no evidence that paraxanthine makes you faster over a race distance.
Evening intervals and poor sleep after caffeine — paraxanthine is worth a try, but expect alertness, not extra watts. A simpler option with the same aim is to lower your caffeine dose or take it earlier.
Do not combine them for “synergy” — in the rowing study, the effect of the combination cannot be separated from the double dose.
The “slow CYP1A2” argument is still theoretical: none of the sports studies split participants by genotype.
Test it on yourself, as with any supplement: the same benchmark session, the same time of day, and a record of how you feel and how you sleep.

The bottom line

Paraxanthine is the main metabolite of caffeine (about 70%); it stays in the blood roughly an hour less (3.1 vs 4.1 hours).
12 runners, 10 km: time, heart rate and lactate showed no difference from placebo; attention after the run showed a moderate plus; the study was funded by the manufacturer.
14 rowers, 2000 m: paraxanthine alone did not make them faster; the combination with caffeine gave 3.3 s, but at a double dose of stimulant.
The real potential advantage is being slightly gentler on sleep after an evening session. For speed over a race distance, caffeine remains the standard.

Sources: Bingol Diedhiou A., Yildirim U.C., Ozdenk S. et al. “Comparative effects of caffeine and paraxanthine on rowing performance and sleep quality: a randomized crossover study”. Journal of the International Society of Sports Nutrition, 2026;23(1):2650339. DOI: 10.1080/15502783.2026.2650339 · Yoo C., Xing D., Gonzalez D.E. et al. “Paraxanthine provides greater improvement in cognitive function than caffeine after performing a 10-km run”. Journal of the International Society of Sports Nutrition, 2024;21(1):2352779. DOI: 10.1080/15502783.2024.2352779 · Lelo A., Birkett D.J., Robson R.A., Miners J.O. “Comparative pharmacokinetics of caffeine and its primary demethylated metabolites paraxanthine, theobromine and theophylline in man”. British Journal of Clinical Pharmacology, 1986;22(2):177–182. DOI: 10.1111/j.1365-2125.1986.tb05246.x · Guest N.S., VanDusseldorp T.A., Nelson M.T. et al. “International society of sports nutrition position stand: caffeine and exercise performance”. Journal of the International Society of Sports Nutrition, 2021;18:1. DOI: 10.1186/s12970-020-00383-4