Login

In brief

  • Nothing in a capsule creates energy. Supplements either correct something missing, or borrow alertness you pay back later. Knowing which you are buying is the whole game.
  • The strongest result here is specific and unglamorous: in women with low ferritin but no anaemia, iron cut fatigue scores by 47.7% against 28.8% on placebo.
  • That same trial moved fatigue and did not move mood, quality of life or anxiety. A supplement that fixes one thing does not fix the others.
  • Look at the placebo arm. Fatigue fell nearly 29% on a dummy pill. Anything you try will seem to work, which is exactly why these trials exist.
  • Daily caffeine mostly repays a debt it created. Habitual consumers scored below low consumers before any dose, and how hard it hits you is partly genetic.

The question worth asking first

Fatigue is a symptom, not a condition, and the list of things that cause it is long and mostly treatable: iron deficiency, thyroid disease, sleep apnoea, depression, B12 deficiency, medication side effects, anaemia from a cause nobody has looked for yet.

Not one of those is fixed by an energy supplement, and several get worse if you spend six months on capsules instead of getting them identified. So the question that actually moves the needle is not "what should I take" but "has anyone found out why?"

If your tiredness has lasted more than a few weeks, that is a blood test and a conversation, not a shelf.

Iron, where the real result is

This is the best evidence on the page and the most specific. A randomised controlled trial recruited 198 non-anaemic menstruating women aged 18 to 53 who complained of unexplained fatigue and had a ferritin below 50 µg/L with normal haemoglobin. They took 80 mg of elemental iron daily or placebo for 12 weeks 1.

What the trial shows

Fatigue scores fell by 47.7% on iron against 28.8% on placebo — a difference of 18.9 percentage points, which held up statistically (p = 0.02). Haemoglobin and ferritin both rose 1.

The population matters as much as the result. These women were not anaemic. Their blood counts were normal and their stores were low, which is the stage most likely to be missed, and it is common: roughly 38% of non-pregnant women of reproductive age in high-income countries are iron deficient without anaemia.

What it does not show

The same trial found no significant effect on quality of life, depression or anxiety 1. Correcting the iron fixed the tiredness and left everything else where it was. If you are tired and also low, iron may help the tiredness specifically.

And it is one trial, in one narrowly defined group. It is not evidence for taking iron while tired without knowing your ferritin — which carries risk and no expected benefit.

The number everyone skips

Go back to the placebo arm: fatigue fell 28.8% in the women taking nothing active 1.

That is why personal experience is such a poor guide in this particular area. Start any supplement during a bad stretch and you will very likely feel better within weeks, because fatigue fluctuates, because you expected to, and because you started when things were at their worst. Roughly three-fifths of the improvement in that trial had nothing to do with iron.

Ginseng

Twelve randomised trials in 1,298 patients were pooled for disease-related fatigue 2.

What the trials show

A statistically significant reduction in fatigue, with an effect size of 0.33 — small, but real and consistent across a decent number of trials 2.

What they do not show

The word doing the work is disease-related. These were patients whose fatigue came with an illness, most often cancer. That is a different population from a healthy person who finds afternoons hard, and the result does not transfer automatically.

Coenzyme Q10

Five randomised trials in 474 participants, all of whom had a diagnosis of depression — three of those trials in people whose depression accompanied multiple sclerosis, breast cancer or polycystic ovary syndrome 3.

What the trials show

CoQ10 significantly reduced depressive symptoms against control 3. Given how tightly low mood and exhaustion travel together, that is relevant to anyone whose tiredness has a mood component.

What they do not show

This is a clinical population with a diagnosis, studied in five small trials. CoQ10 is marketed broadly as a cellular energy supplement on the strength of its role in mitochondria, and that mechanism is real, but mechanism is not evidence of benefit in a well person.

Rhodiola, and why it is not here on merit

Rhodiola rosea is the classic fatigue adaptogen, and it is included because people ask about it, not because the evidence earns it a place.

A systematic review found 11 trials. Two of six in physical fatigue reported benefit. Three of five in mental fatigue reported benefit. And every single included study carried either a high risk of bias or reporting flaws severe enough that its validity could not be assessed. The reviewers' own summary was that the research is contradictory 4.

That is not the same as saying it does nothing. It means that after 11 trials we still cannot tell, which after this long is itself informative.

Caffeine: borrowing, not creating

Caffeine works. It reliably improves alertness and reaction time, which is why it is the most used psychoactive substance on earth and why it turns up in most energy products.

What it does is block the adenosine signal that tells you that you are tired. The tiredness is still there; you have muted the message. That is genuinely useful for a specific morning or a long drive, and it is not a treatment for being tired all the time. Used that way it fragments sleep, which deepens the fatigue it was hiding.

The part that changes how you read your own coffee

Once you drink it daily, most of what caffeine is doing is repairing a deficit it created.

In a double-blind study of 26 children, habitual consumers taking an average of 109 mg a day were compared with low consumers taking 12 mg, both tested after overnight abstinence. Before any treatment was given, the habitual consumers performed worse on the cognitive test than the low consumers did. Caffeine brought them back up 5.

That is the pattern the adult literature describes as well: the psychostimulant benefit shows up mainly in people who have been deprived of caffeine, and what they gain is a return to their own baseline rather than a gain above it. In a randomised double-blind study of 379 adults, the anxiety response to caffeine was reduced in medium and high habitual consumers compared with non-consumers, which is tolerance appearing in the same way 6.

None of this means you should stop. It means the comparison that feels true, between how you feel after coffee and how you feel before it, is not the comparison that tells you what caffeine is worth. The relevant comparison is against someone who never started.

How much, and who should take less

For healthy adults, a comprehensive review of the human evidence put moderate daily intake at up to 400 mg a day, roughly 6 mg per kilogram of body weight, without adverse effects on general toxicity, cardiovascular measures, bone and calcium status, behaviour, cancer incidence or male fertility 7.

The same review identified women of reproductive age and children as groups at risk who need lower intakes and specific advice. A large energy drink and a strong coffee can approach half that daily ceiling between them, which is worth knowing before stacking a pre-workout on top.

Why the same cup affects people so differently

Caffeine is cleared by the enzyme CYP1A2, and a common genetic variant determines how fast. Carriers of the CYP1A2*1F allele are slow metabolisers; people homozygous for *1A are rapid.

In a study of 2,014 people who had a first non-fatal heart attack and 2,014 matched controls, coffee intake was associated with increased risk only in slow metabolisers. At four or more cups a day the odds ratio was 1.64 for slow metabolisers and 0.99 for rapid ones, with a significant gene-by-coffee interaction. Among people under 59, the slow-metaboliser odds ratio at that intake rose to 2.33 8.

About half the population carries the slow variant. Separately, variation in the adenosine receptor gene ADORA2A predicts who gets anxious on caffeine rather than alert 6.

This is the honest explanation for something people notice and then doubt: two people drink the same coffee, one is focused and one is jittery and awake at midnight. That is not imagination or willpower, and it means general advice about caffeine is less useful than paying attention to your own response.

Natural versus synthetic caffeine

Products sold on botanically sourced caffeine, usually from green tea or green coffee bean extract, are priced above synthetic caffeine on the implication that the molecule is somehow different.

It is not. A randomised double-blind crossover trial in 16 healthy men compared 60 mg of caffeine from green coffee bean extract with 60 mg of synthetic USP caffeine. Time to peak concentration was 0.75 hours for the botanical source and 0.63 hours for the synthetic, a difference with no practical consequence 9.

What a botanical source does add is the compounds that travel with it. The green coffee extract delivered measurable chlorogenic acids; green tea brings L-theanine and EGCG. Those are a real reason to prefer one, and they are what to evaluate, rather than the caffeine.

They also carry their own considerations, and here the intuition runs backwards. A United States Pharmacopeia review of green tea extract found liver injury case reports at EGCG intakes from 140 mg to around 1,000 mg a day, with large differences between individuals in susceptibility and higher absorption when taken on an empty stomach. USP added a cautionary labelling requirement: take with food, and do not use with an existing liver problem 10. Synthetic caffeine carries nothing comparable.

The one combination with decent data

L-theanine, an amino acid from tea, is the companion compound most often paired with caffeine, and the pairing has been tested directly.

In 27 healthy volunteers, 50 mg of caffeine with 100 mg of L-theanine improved both speed and accuracy on an attention-switching task at 60 minutes and reduced susceptibility to distraction in a memory task at 60 and 90 minutes. Caffeine alone improved subjective alertness and accuracy, but not the same combination of measures 11.

Twenty-seven people is a small trial, and the effect is on demanding attention tasks rather than on tiredness. It is a reasonable thing to try at a modest dose, and it is not a fatigue treatment. Theanine's separate evidence for sleep and for stress is covered in the sleep guide and the stress guide.

Bitter orange, the stimulant that replaced ephedra

After ephedra was banned, Citrus aurantium and its active compound p-synephrine took its place in energy and weight-loss products.

A meta-analysis of 18 placebo-controlled human trials found that prolonged use raised systolic blood pressure by 6.37 mmHg and diastolic by 4.33 mmHg, and tended to raise heart rate. Weight loss was not significant and body composition did not change 12.

A measurable cardiovascular effect with no measurable benefit on the outcome it is sold for is a straightforward reason to skip it, and a reason to read the label of anything marketed as a thermogenic.

What was actually studied

WhatWhat improvedIn whom
Iron, 80 mg elemental daily, 12 weeksFatigue, 47.7% vs 28.8% on placeboNon-anaemic women with ferritin under 50 µg/L
GinsengFatigue, small effect (0.33)Patients with disease-related fatigue
CoQ10Depressive symptomsPeople diagnosed with depression, often alongside another illness
RhodiolaContradictory across 11 trialsAll trials at high or unclear risk of bias
CaffeineAlertness, temporarilyEveryone, at the cost of later sleep
Caffeine, dailyMostly restores a deficit it createdHabitual consumers scored below low consumers at baseline
Caffeine, 4+ cups dailyHeart attack odds 1.64 vs 0.99Slow CYP1A2 metabolisers only; 2.33 in under-59s
Caffeine 50 mg + L-theanine 100 mgAttention switching, resistance to distraction27 healthy volunteers; demanding tasks, not fatigue
p-Synephrine (bitter orange)Blood pressure up 6.37/4.33 mmHg, no weight loss18 placebo-controlled trials

What this adds up to

If you are tired all the time, the single most valuable thing available to you is a diagnosis, and for a large group of people — especially menstruating women — that diagnosis is cheap, common and genuinely fixable.

If a test finds low iron, the evidence supports correcting it and expecting the tiredness specifically to improve. If you have an illness causing fatigue, ginseng has a small, real effect worth discussing. If low mood is part of the picture, CoQ10 has evidence in people with a diagnosis.

And if everything comes back normal, that is useful too. It points at sleep, at how much you are actually recovering, and at the things no capsule reaches: hours, light, movement, and whether anything in your life has been quietly costing you more than it used to.

Keep reading

The companion to this page is Are you actually short of anything?, which covers what is worth testing and how to read a result. The iron profile goes further on why a normal blood count does not rule deficiency out.

If sleep is the suspected cause, Sleep: what the trials show covers what the evidence supports. Before trying anything, Did it work? How to tell explains why the 28.8% placebo result above should change how you judge your own experience.

Fatigue lasting more than a few weeks, or arriving with weight loss, breathlessness or fever, needs a doctor rather than a supplement. Iron should not be taken without a confirmed deficiency.

Advanced research

The technical write-ups behind this topic, for readers who want the primary literature rather than the plain-language version.

Back to Top
Added to cart
Your Cart
0
Close

No products in the cart.