The useful answer before personalization
Adaptogen marketing has invented “cortisol profiles”: exhausted, wired, anxious. A herb for every profile, and often a bottle to match. No validated test sorts you into these boxes, and of the five herbs being sold, only one has a serious body of evidence for stress.
There is no validated test that turns your symptoms into four “biological stress types” and then automatically assigns a herb to each. Supposedly high evening cortisol, low HRV or morning fatigue cannot diagnose an “overactive” or “exhausted” hormonal axis.
Defensible personalization rests on something else: the main symptom, the population studied, individual risks and a reversible trial. By these criteria, ashwagandha is the best documented for short-term perceived stress. Rhodiola remains an uncertain option for mental fatigue. Ginseng shows, at best, a small signal in certain forms of fatigue. Cordyceps and tulsi lack sufficiently direct evidence to be first-line choices for stress.
Before considering a supplement, persistent fatigue warrants ruling out insufficient sleep, sleep apnea, anemia, thyroid disorders, depression, infection, medication effects and an energy deficit. Calling it “adrenal burnout” can delay the right diagnosis. See a doctor for persistent fatigue or worsening mood.
“Adaptogen” is not a level of evidence
The historical term describes a nonspecific increase in resistance to stress. It is neither a uniform regulatory category nor a guarantee that a herb “raises what is low and lowers what is high.”
HPA, HSP70, Nrf2, AMPK and monoaminergic pathways are frequently invoked. They may explain a hypothesis, but they do not predict the outcome of an oral extract in humans. A change in cortisol is a biomarker, not proof that someone functions better or avoids disease.
Two extracts of the same species can differ in plant part, solvent and concentration. Results for a patented product therefore do not automatically apply to a generic powder.
Ashwagandha: the most favorable signal, with real limitations
A systematic review and meta-analysis of nine randomized trials involving 558 participants found improvements in perceived stress, anxiety and cortisol compared with placebo. Reported adverse effects were mostly mild to moderate. However, the authors emphasize the lack of long-term safety data.
This finding makes ashwagandha a plausible option for an adult with perceived stress, after ruling out a condition requiring treatment. It does not show that it permanently “recalibrates” the HPA axis, increases HRV or suits every form of anxiety.
The trials cover different formulations and durations. Promoting KSM-66 extract “for calm” and Shoden extract “for sleep” as though a comparative algorithm existed is misleading. No robust head-to-head study establishes this ranking.
Ashwagandha requires particular caution in the following situations:
- pregnancy or breastfeeding
- thyroid disease or thyroid medication
- autoimmune disease, transplantation or immunosuppression
- liver disease or a history of drug-induced hepatitis
- sedatives, anticonvulsants or multiple medications
Rare cases of liver injury have been reported. Our analysis of supplements associated with liver injury · FR explains which symptoms mean you should stop. The NCCIH safety fact sheet also notes that safety is documented only in the short term. Stop if you develop jaundice, dark urine, widespread itching or unusual fatigue.
Rhodiola: possible help with mental fatigue, low confidence
Rhodiola rosea is often presented as stimulating and assigned to “low-cortisol” profiles. This category has no clinical validation. A systematic review of 11 trials found conflicting results: some trials were positive for physical or mental fatigue, but all had a risk of bias or reporting shortcomings.
Rhodiola may therefore be tried as a second-line option for fatigue linked to a period of overload, not to treat unexplained chronic fatigue. Claims that it inhibits MAO sufficiently to “boost dopamine” in users are not supported by applicable clinical measurements.
Agitation, palpitations or worsening sleep are reasonable reasons to stop. Caution is warranted with bipolar disorder, unstable anxiety and psychotropic medication, because interaction data remain incomplete.
Ginseng: small effects in certain types of fatigue
Ginseng encompasses several species and preparations. A meta-analysis of 19 randomized trials found no significant reduction in fatigue in the overall analysis. Some subgroups showed a small effect, particularly for formulas or certain types of chronic fatigue, with more rigorous trials needed.
Ginseng is therefore not a “burnout recovery remedy.” It must not delay the assessment of persistent fatigue. It can also cause insomnia, headaches or digestive problems and interact with medications, particularly those affecting blood glucose or blood clotting.
Cordyceps and tulsi: why I am not ranking them
Cordyceps has mainly been studied for exercise, with varied preparations and results. This does not demonstrate a treatment for psychological stress. Cordyceps militaris also raises a separate regulatory issue in the European Union, depending on the preparation.
Tulsi has preliminary human trials, but not a sufficiently robust evidence base to assign it to a profile of compulsive sugar eating or metabolic stress. A mechanism involving NF-kB or blood glucose does not prove a clinical reduction in stress.
Not ranking these products does not mean concluding that they are inactive. It means refusing to claim a level of precision the trials cannot support.
A realistic decision matrix
| Situation | Potentially reasonable option | Confidence | What needs to be ruled out |
|---|---|---|---|
| Short-term perceived stress | Ashwagandha, after checking risks | Low to moderate | Severe anxiety disorder, thyroid disease, pregnancy, liver disease |
| Mental fatigue linked to identified overload | Rhodiola as a second-line option | Low | Anemia, sleep debt, depression, bipolar disorder |
| Nonspecific fatigue | Ginseng has sometimes been studied | Low | Medical causes and medications |
| Chronic insomnia | No adaptogen as a first-line option | High for this rejection | Sleep apnea, restless legs, mood disorder |
| Self-reported “low cortisol” | No herb-selection algorithm | High for this rejection | Endocrine disease |
A trial protocol that limits self-deception
If a health professional confirms that a trial is compatible with your health, use just one herb and a preparation that matches a human trial as closely as possible. Do not transfer milligram doses between an extract and a powder.
Proceed as follows:
- Define just one goal, such as your perceived stress score or fatigue at 3 p.m.
- Measure it for one week without a supplement.
- Start at the lowest dose indicated for the chosen extract.
- Keep caffeine, sleep and training as stable as possible.
- Reassess after six to eight weeks, a common duration in stress trials.
- Stop sooner if sleep, mood, digestion or symptoms worsen.
HRV can complement your journal, but it varies with alcohol, infection, training and sleep. An increase does not prove that the herb has “repaired” the HPA axis.
Why stacking and cycling are not solutions
Combining ashwagandha, rhodiola, magnesium and theanine multiplies the variables. If you feel better, you do not know which one helped. If you develop a liver reaction or agitation, investigating becomes harder.
Cycling five days out of seven is not a recommendation derived from comparative trials. No solid evidence shows that it prevents anhedonia or “resensitizes receptors.” Better protection means stopping when the goal is not met and not indefinitely extending an intervention whose long-term safety is poorly understood.
The Verdict
Ashwagandha has the best signal, but that is not a blank check. Rhodiola may be worth discussing for mental fatigue with an identifiable context, with low confidence. Ginseng has no convincing overall effect on fatigue. None can be chosen based on a cortisol archetype or used as a standalone treatment for burnout.
Real personalization is less spectacular: rule out a medical cause, choose a studied extract, introduce just one variable, measure a meaningful outcome and stop if the benefit is not clear.
