Does apigenin help you sleep better?
There are preliminary indications that apigenin has some connection to sleep, but clinical evidence in humans is entirely lacking. If you want to sleep better, there are supplements with a stronger evidence base than apigenin; discuss this with your doctor or pharmacist if sleep problems persist.
Apigenin occurs naturally in plants such as chamomile and parsley. People who eat more apigenin sleep better on average, according to an observational study of nearly two thousand Italian adults. But that is an association: people who eat more apigenin may also lead healthier lives in other ways. In people who were overweight or obese, the association was absent altogether. Evidence that apigenin itself is the cause of better sleep is still lacking in humans.
How apigenin acts on brain receptors is more complicated than supplement manufacturers often suggest. Laboratory studies show that apigenin has a dual effect on the brain's main inhibitory receptors (GABA-A, the target of sleeping pills and sedatives). At high concentrations, apigenin actually blocks these receptors, which can be the opposite of promoting sleep. At low concentrations, it may enhance the action of certain sedatives. What this means in humans has not been demonstrated.
The most concrete sleep results come from mouse research. A combination of apigenin and magnesium extended sleep duration in mice by 32 to 44 percent, more than either agent alone. This sounds impressive, but the translation of mouse results to humans is notoriously uncertain. There are no clinical studies confirming this effect in humans.
A narrative review on dietary supplements for sleep names apigenin as a possible candidate, but presents no original clinical data on apigenin alone. Computer models show that apigenin can bind to multiple sleep-related proteins, but those are calculations, not experiments.
Sources: one observational study in humans (n=1936), multiple laboratory studies on mechanism of action, one mouse study (apigenin+magnesium combination), two narrative reviews, one computational modelling study. No randomised studies in humans.