Adaptogens explained: what the term means and what it does not promise

You are standing in front of the shelf. Or in front of an advertisement that seems to know you very precisely. There is a word there that sounds like science and still stays pleasantly soft: adaptogen. Underneath it, usually a promise. More calm. More focus. Less cortisol. And somewhere in small print: food supplement.

And then that second of uncertainty. Is this medicine or marketing? Is there real physiology behind it, or just a particularly well built word?

Both. That is exactly what makes it so confusing. The term has a real research history, a comprehensible physiological core and a body of data that is neither zero nor what the packaging suggests. I want to sort these three levels apart here. Not to talk you out of anything. But so that you know what you are actually deciding about.

Where the word comes from, and why that matters

Almost every term in medicine carries the time of its origin around with it. With adaptogen this is particularly clear.

By the common account, the term goes back to the year 1947 in Soviet pharmacology. The pharmacologist Lazarev is said to have coined it for substances meant to raise the non-specific resistance of an organism to strain. Not against a particular disease. But against strain as such. The primary sources for this are Russian-language and hard to access, and the dating comes mostly from later review work (Panossian AG et al., Med Res Rev 2021, source 2).

The context was not the wellness department. It was about people under extreme conditions: soldiers, pilots, miners, later cosmonauts and competitive athletes. What was sought was something to support performance under sustained load, without the drawbacks of classic stimulants.

In 1969 Brekhman and Dardymov formalised the concept in a review article in the Annual Review of Pharmacology. That is where the three criteria stand which to this day turn up in almost every text about adaptogens, mostly without a source.

Why that matters: a term that arose within a performance physiology frame is today sold within a therapeutic frame. That is a quiet shift of meaning that hardly anyone comments on. And now you know why a word with so much history still says so little about the contents of a package.

The three classic criteria and their blind spot

The definition from 1969 sounds very tidy at first. It has three parts.

And here lies the blind spot. Criterion two is hard to test cleanly with the tools of clinical research. A randomised trial tests a substance against placebo on a defined target value. A substance whose direction of effect is supposed to depend on the starting state resists this design. It could look unremarkable in the arithmetic of a mixed group even though something is happening in individuals.

That is no proof that adaptogens do not work. It is an indication that concept and testing method do not fit together perfectly. Holding both at once is scientifically uncomfortable, but honest.

What authorities in Europe say about the word

Here it becomes unexpectedly concrete. Because the term has actually occupied authorities.

The Committee on Herbal Medicinal Products of the European medicines agency published its own reflection paper on the adaptogenic concept in 2008, effective from 8 May 2008. The stated aim: to frame the term and describe how adaptogenic herbal medicinal products are to be assessed in principle.

What did not come out of it: a separate indication called adaptogen. A preparation is not licensed against adaptogen, it is licensed against stress symptoms. The term does appear in the rhodiola monograph, but only as a descriptive addition inside the stress indication, not as a field of application in its own right. The formulations in the official monographs are noticeably more restrained than the word on a package suggests.

And one more thing matters: by far the largest part of adaptogen products in the shops are not medicines at all but food supplements. Different and in practice lower requirements for evidence and standardisation apply to this category. The word on the package says nothing about that.

The physiological core: HPA axis and cortisol

Now to the biology. Because there is a real core, and it is more interesting than the marketing.

Imagine your stress system as a thermostat. The hypothalamus measures, the pituitary passes the signal on, the adrenal cortex produces cortisol. The cortisol reports back upwards and throttles its own production. This control loop is called the HPA axis.

Cortisol is not the villain here. It makes energy available, it slows excessive inflammation, it organises your daily rhythm. High in the morning, low in the evening. A thermostat with a daily profile.

Chronic strain presumably does not primarily disturb the level of this value. It can disturb the quality of regulation. The literature describes that the rise can come too late, the fall too slowly, the feedback blurred. There is no single uniform pattern in this, the findings diverge depending on population and measurement method. It is exactly at this model idea that the thinking behind adaptogens starts.

And now the point that almost everyone skips

In a randomised double-blind trial with 60 stressed adults over 60 days, morning cortisol fell markedly under a standardised ashwagandha extract. At the same time, however, DHEA-S fell as well. That is not a side finding. It is an indication that here no stress hormone was throttled selectively, but that the adrenal cortex as a whole was working differently.

From the perspective of clinical psychoneuroimmunology I look at four levels at the same time here. The nervous system, because the HPA axis begins in the brain and not in the adrenal gland. The hormonal system, because cortisol, DHEA and the thyroid axis hang together. Metabolism, because cortisol mobilises blood sugar and strong blood sugar swings can therefore act back on the stress axis. And the immune system, because cortisol is one of its central brakes. Whoever looks at only one of these levels rarely sees the whole picture in stress. And now you know why a single cortisol value tells so little.

Hormesis: the explanatory model that works against its own advertising

Do you know the principle from strength training? The stimulus does not make you stronger. The stimulus first makes you weaker. You become stronger in the recovery afterwards.

Exactly this pattern has a name: hormesis. A biphasic dose-response relationship. A small strain provokes an adaptation. A large one provokes damage. The curve is not a straight line but an upside down U.

And exactly this model is drawn on as an explanation in adaptogen research. A large review describes adaptogens explicitly as mild stress mimetics that at low dose switch on adaptive signalling pathways so that the organism may cope better with stronger stress.

If this model holds, then something follows from it that stands rather in the way of the advertising.

In hormesis research exactly this point is emphasised: it is worth investigating not only broader dose ranges but also varying the temporal dimension systematically, that is, the timing, number, frequency and duration of the stimuli. That is a fairly precise description of what stands on no supplement label.

The data plant by plant

Now the part you were probably looking for. I go through the six best known. Not in order of popularity, but by study situation.

Ashwagandha, Withania somnifera

The best studied plant of the group, at least over the past fifteen years. Several meta-analyses on stress, anxiety and sleep. The signal points consistently in one direction.

But look at the spread. In a meta-analysis on anxiety and insomnia from five randomised trials with 254 people, the Hamilton anxiety score fell by 5.96 points compared with placebo. The heterogeneity between the studies was at an I squared of 98 percent. That is a very high value. It means that the studies differ strongly from one another and that an averaged value is only of limited informative value.

Rhodiola rosea, roseroot

The plant with the longest European research line and one of the best known individual studies. In a randomised trial with 60 people with stress-related fatigue syndrome over 28 days, several attention parameters improved more clearly under the standardised extract SHR-5 than under placebo, as did a burnout score. The cortisol response to waking in the morning also differed significantly between the groups.

In fairness to this: that study comes from the environment of the manufacturer of this standardised extract, and it is disclosed in the paper. That does not make it wrong, but it belongs in the same scale in which I also place the manufacturer proximity of other work further below.

And then the view from above. A systematic review screened 206 pieces of work and found eleven includable studies. Two out of six studies on physical fatigue came out positive, three out of five on mental fatigue. All included work carried either a high or an unclear risk of bias.

Eleutherococcus senticosus, Siberian ginseng

Widely used in the USSR, classified in Europe as a traditional medicinal product for symptoms of asthenia. The best controlled study on it is soberingly honest.

96 people with persistent fatigue without an identifiable cause were randomised, 76 supplied data after two months. Fatigue improved markedly in both groups. Between verum and placebo there was no statistically significant difference in the overall sample. Signals appeared only in subgroups, explicitly without adjustment for multiple testing.

Panax ginseng

Here two meta-analyses show how strongly the question shapes the result. An evaluation across 19 randomised trials found no significant effect on fatigue severity in the overall analysis. In subgroups, significant effects appeared, for instance with ginseng combination formulas and with chronic fatigue, and more clearly with general, non-disease-bound fatigue. The effect sizes were small throughout. A second meta-analysis across twelve studies with 1298 people found a statistically significant, likewise small effect for disease-related fatigue.

Schisandra chinensis

The plant with the most detailed historical documentation and the thinnest modern one. An extensive review summarises the Russian research: schisandra was recognised as an adaptogen in the official medicine of the USSR as early as the beginning of the 1960s and taken into the State Pharmacopoeia. Described are numerous animal experimental findings on stress protection and endurance as well as clinical observations in asthenia.

What is missing are modern randomised trials by today's standards. Whoever recommends schisandra rests on tradition and mechanism, not on a current package of studies. That is legitimate, as long as one names it that way.

Tulsi, Ocimum sanctum, holy basil

And here it becomes methodologically most interesting. A systematic review found 24 human studies on tulsi with clinical endpoints, among others on metabolism, cardiovascular health, immune function and cognition. All 24 studies reported favourable clinical results. Meaningful adverse events were reported in no study.

Why these studies are harder to read than they look

I want to stay fair here. No field has perfect studies, and phytotherapy has structural disadvantages that have nothing to do with intent. Plant extracts are multi-substance mixtures, they vary by origin and harvest, and nobody can file a patent on them. Large, expensive studies are therefore rarely funded by anyone.

Still, a sober look at five recurring patterns is worthwhile.

And an important counter-movement: precisely because the data are thin, the reverse conclusion is inadmissible too. Missing evidence is not evidence of ineffectiveness. For eleutherococcus, the European committee explicitly formulated that a possible effect on fatigue and weakness was observed, but that design flaws prevented robust conclusions. That is something different from a negative result. And now you know why I so often say on this topic: we do not know it precisely enough yet.

The risk side that rarely stands on the package

Botanical does not mean harmless. That is not an opinion, it is by now well documented. And because adaptogens are marketed as particularly gentle, this side is especially important.

Liver

On ashwagandha there is a published case series from Iceland and the US network for drug-induced liver injury. Five people developed jaundice with nausea, exhaustion, itching and abdominal discomfort after two to twelve weeks of use. The liver injury was cholestatic or mixed. Itching and elevated bilirubin persisted for five to twenty weeks. No case led to liver failure, and in four people the liver values came back into the normal range within one to five months. In one case rhodiola was named as a possible co-contributor.

Since then further individual cases have been published, among them an acute hepatitis with pronounced jaundice after roughly a year of use as well as a case of acute-on-chronic liver failure with pre-existing liver disease. Reviews on supplement-related liver injury now list ashwagandha regularly among the substances to watch.

Thyroid

Published is the case of a 47-year-old previously healthy man who presented two months after starting ashwagandha with exhaustion, night fever and weight loss. The diagnosis was a painless thyroiditis with thyrotoxicosis. After stopping, symptoms and laboratory values improved.

Fitting with this is a small safety analysis from a randomised trial in people with bipolar disorder. There, thyroxine rose from baseline in all three evaluated people treated with ashwagandha, by 7, 12 and 24 percent. In six out of seven placebo-treated people it fell. The numbers are tiny and the analysis was a secondary safety aim. The authors concluded from this that vigilance regarding overactivity might be appropriate.

Pregnancy, breastfeeding, wish to conceive

Here the answer is short, and it is stricter than many expect. Robust human safety data are missing. At the same time the European monograph on rhodiola is unambiguous at this point: pregnancy and breastfeeding are a contraindication there, not merely an open question. For ashwagandha, a possible harm to the unborn child or a labour-inducing action is discussed in the specialist literature, and the data on this are contested.

As long as that is the case, the rule for this whole group is: no self-experiments in pregnancy, in breastfeeding and with an existing wish to conceive. If you have already taken something, that is no reason to panic, but it is a reason to raise it at your next check-up.

Children and adolescents

And one more group that rarely comes to mind with bars and drinks. For children and adolescents under 18 years of age, use is explicitly not recommended in the European monographs, because robust data are missing. That is not proof of a danger, it is simply untested ground. On a coffee with an adaptogen print, none of that is written anywhere.

Autoimmune diseases

Adaptogens are regularly described as immunomodulating in the mechanism literature. With a healthy immune situation that may be unproblematic. With an autoimmune disease, the direction of this modulation is not predictable, and studies that examine exactly this question in affected people are largely missing. I formulate this deliberately cautiously: that is a reasoned restraint arising from the mechanism, not a demonstrated harm.

Sedatives, sleeping aids, alcohol

For ashwagandha, improvements in sleep onset latency and sleep efficiency have been described in several studies. If a substance can influence sleep, it can also add up with other dampening substances. That can apply to benzodiazepines, to Z-substances, to sedating antidepressants and to alcohol. The first three groups are prescription only and are adjusted deliberately.

This matters in both directions. An addition can lead to stronger daytime sleepiness, to a tendency to fall and to impaired fitness to drive. And the other way round: never stop a prescribed sleeping aid or sedative on your own and do not reduce it because you are now taking something botanical. Both belong in the same consultation, not in a self-experiment.

Blood sugar lowering medication

For several plants in this group, above all for ginseng and tulsi, effects on blood sugar have been described. The systematic review on tulsi lists metabolic disorders including diabetes and metabolic syndrome explicitly among the endpoints examined. Whoever takes insulin, sulfonylureas or other blood sugar lowering agents can slip lower than planned because of this. That is not a ban. It is a reason to measure more often in the first weeks and to discuss the combination beforehand.

Anticoagulants and ginseng

For ginseng, the interaction with anticoagulants is an established research topic. In an investigation with cell experiments and a rat model, ginsenosides weakened the anticoagulant action of warfarin, among other routes via an altered expression of clotting factors and of cytochrome P450 enzymes in the liver. That is an animal and cell study and cannot be transferred directly to humans. As a reason to talk with the practice treating you, it is more than enough.

One practical note for Germany. The vitamin K antagonist commonly used here is not warfarin but phenprocoumon. Both are prescription only. Whether the signal from the rat study transfers is open. Anyone taking an anticoagulant should therefore not start anything from this group without agreement, and in case of doubt should have the INR checked more closely.

Quality: why the word on the package says the least

Suppose you have informed yourself, you have spoken with your physician, you want to try it. Then comes the question almost nobody talks about: is what the label says even inside the tin?

An investigation of around 40 commercial rhodiola products examined exactly this. About one fifth of the products that declared Rhodiola rosea contained no rosavin, that is the marker substance by which Rhodiola rosea is distinguished from related species. In some products salidroside was missing too. Around 80 percent of the remaining commercial products were below the preparations registered as medicines in rosavin content and appeared blended with other rhodiola species. Rhodiola crenulata is named as a frequent blending partner.

Why sleep, light and recovery stay the foundation

Now comes the part where I am least diplomatic. Because it follows logically from everything above.

If the explanatory model for adaptogens is hormesis, then they need functioning recovery in order to be able to do anything at all. A stimulus without recovery is not training. A stimulus without recovery is simply strain.

That means: whoever chronically sleeps too little, sees no daylight in the morning, hardly moves, eats constantly under time pressure and never has a real break gives their system no opportunity to make an adaptation out of any stimulus. Not out of a botanical one either.

And there is something else that is bigger than any capsule. If you are considering whether an adaptogen might do something for you, then behind that there is usually not the question of a substance at all. Behind it there is a need: you want to be able to respond to strain again without every week knocking you over. You want to come down in the evening. You want to be someone in the morning who can carry a day.

That is not a wish about a symptom. That is a wish for agency. And agency is not a side issue in life, it is fairly precisely what people mean by quality of life.

And now you know why in my practice I almost never start with the question about the preparation. Not because plants would not interest me. But because their own theory says that without a foundation they do not get much to do.

How I place the term today

I write this as one position among several, not as a final word.

As an advertising promise I consider the word adaptogen weak. It suggests a tested category that legally does not exist, and it quietly transfers the trust earned by one plant onto all the others on the same shelf.

As a thinking tool I consider it rather good. It forces a question that is rarely asked out loud in medicine: not which value is too high, but how well can this system still answer and then come to rest again? Capacity to regulate instead of a snapshot.

This question also stands at the beginning of clinical psychoneuroimmunology. And it stays right whether or not a plant is involved in the end.

If after all this you would still like to try something, then that is a completely legitimate decision. It just belongs in a certain order: first the foundation, then a conversation about your pre-existing conditions and your medications, then a deliberate decision about a specific preparation with a comprehensible declaration, and then an honest observation of what actually changes. Not by the laboratory value alone, but by your morning, your afternoon and your sleep.

Frequently asked questions about adaptogens

Where this topic connects

Adaptogens never stand alone. They touch stress physiology, hormonal axes, sleep and nutrient supply all at once.