The Runner's High Isn't Caused by Endorphins. In 63 Runners, an Opioid Blocker Changed Nothing While the Body's Own Cannabis-Like Chemicals Rose

The Story Everyone Knows

Ask a runner why the last miles feel so good and you will hear one word: endorphins. The story dates to the early 1980s, when researchers first measured a rise in blood beta-endorphin after long runs, and it soon entered everyday language. You run hard, your brain releases its own opioids, and the euphoria follows.

It had a flaw that sat in plain sight, because endorphins are large, water-soluble peptides that cross the blood-brain barrier poorly, and the measurements that launched the theory showed them rising in the bloodstream rather than in the brain, which makes them a strange candidate for a feeling that lives entirely in the head. Several early studies also found no correlation between blood endorphin levels and mood after exercise, yet the theory survived anyway, because nobody had a better molecule to offer.

The Experiment That Broke It

The cleaner test came from Johannes Fuss and colleagues in Hamburg, whose logic was simple: if endorphins cause the high, blocking the receptors they use should kill it, so they recruited 63 healthy adults who ran regularly, gave half naltrexone, an opioid-receptor blocker, and half a placebo, with neither the runners nor the researchers knowing who got which pill. Each person then ran for 45 minutes at moderate intensity on a treadmill and, in a separate session, walked for 45 minutes.

The results left little room for the old story. Running raised self-rated euphoria and lowered anxiety, measured on a human elevated plus-maze, and it did so identically in both groups. The runners who took naltrexone felt the same lift as those who took the placebo. If endorphins were doing the work, the blocked group should have felt nothing. They felt everything.

Blood samples pointed at the replacement: running raised plasma levels of anandamide and 2-arachidonoylglycerol, while walking did not, and these are the body's own cannabis-like signaling molecules, small and fat-soluble unlike endorphins, which lets them slip across the blood-brain barrier and bind to the same CB1 receptors that THC activates.

A Molecule Named After Bliss

Anandamide takes its name from ananda, the Sanskrit word for bliss. Earlier mouse work showed why it matters: in a 2015 study in the Proceedings of the National Academy of Sciences, wheel running raised endocannabinoids in mice and reduced both anxiety and pain sensitivity. Removing CB1 receptors from forebrain GABAergic neurons erased the calm, blocking cannabinoid receptors erased the pain relief, and opioid blockers changed nothing.

The mouse data hint at how deep the system runs. Mice lacking CB1 receptors voluntarily run 30 to 40 percent less than normal mice, which suggests the system drives the motivation to run, not just the reward. Remove the opioid system and the high survives; remove the cannabinoid system and it does not.

The 2026 Marathon Check

A September 2026 field study put the chemistry on a real race course: nineteen trained runners ran a marathon around Lake Baldeney in Essen, Germany, giving blood before the start, after every 14 kilometers, at the finish, and again after 45 minutes of rest, and walked the same route for the same time on a separate day. Anandamide climbed steadily through the first half, through the wall, and through the final kilometers, and it was still elevated 45 minutes after the finish, while walking barely moved it. Thirty-six ultramarathoners running 100, 160, or 230 kilometers showed the same pattern.

Only 12 of the 36 ultramarathoners, one in three, reported feeling a runner's high during the race. Surveys of recreational runners put the lifetime figure near 70 percent. The gap between those two numbers is the point. A third of people running 100 to 230 kilometers felt nothing of the kind while doing it, even as their anandamide climbed. The chemistry rises reliably; the feeling does not.

The Strongest Counterargument

The case for endorphins is not dead, and the strongest version of it comes from brain imaging: in a 2008 PET study, researchers scanned athletes after a two-hour run and found opioid release in frontolimbic brain regions, and the release tracked with the runners' euphoria, which is direct evidence of opioids acting inside the brain during running rather than just floating in the bloodstream. If endorphins were irrelevant, that correlation would be hard to explain. The naltrexone experiment also tested only two features of the high, euphoria and anxiety, leaving pain relief and sedation untested in humans.

A gap remains, because nobody has yet proven that endocannabinoids cause the high in humans, and the decisive experiment cannot be run: the only drug that blocks CB1 receptors in people, rimonabant, was withdrawn after psychiatric side effects, so researchers cannot do the mirror image of the naltrexone test, and a 2023 systematic review of 21 human trials concluded that the endocannabinoid rise after acute exercise is robust, with 14 of 17 studies detecting it, but that causation in humans is still unproven. The blood measurements are indirect too, since anandamide was measured in plasma, not in the brain, and no PET study has yet imaged endocannabinoid activity during exercise. Endorphins lost their crown, endocannabinoids are the leading candidate, and the case is not closed.

What We Didn’t Prove

This trial does not prove that anandamide causes your good mood after a run. It proves that opioid signaling is not required for the euphoria and anxiety relief after 45 minutes of moderate running. The volunteers were healthy regular runners, so the study says little about beginners. The running happened on a laboratory treadmill at a fixed 45-minute dose, so nobody knows whether 20 minutes or hard intervals produce the same chemistry. The high itself was self-reported, and self-reports of a subtle state are easy to overread. No independent team has repeated the naltrexone experiment, though two older, smaller opioid-blockade studies from 1982 and 1986 reached the same conclusion.

What You Can Do

If you run for the mood lift, aim for the tested dose: about 45 minutes of continuous, moderate effort, hard enough that talking takes work. A gentle stroll did not raise endocannabinoids in the trial, and a 20-minute jog has not been tested.

Second, stop chasing the high on every run: only a third of ultramarathoners felt it during a race, so treat it as a bonus, not the goal. The mood benefits of regular exercise show up in hundreds of trials whether or not any single session produces euphoria.

Third, keep the chemistry in perspective: your body makes its own cannabis-like molecules during sustained exercise, subtler than anything from a dispensary, a legal and free mood shift that arrives after 45 minutes of honest effort and fades within the hour. If low mood or anxiety persists, exercise is a tool, not a treatment, and a clinician is the right next step.