Do testosterone boosters work? What the evidence actually says

This is the question underneath every other question on this blog. Here is what the pooled evidence says, what it says about, and how to tell an honest product claim from a borrowed one.

What the reviews found

Two pieces of work define the state of play. A systematic review asked directly whether products marketed as testosterone boosters raise serum total testosterone and concluded that most of the ingredients involved lacked supporting human evidence (Morgado et al., Int J Impot Res, 2024). Separately, an analysis of the fifty best-selling online testosterone-boosting supplements found that only a minority of their listed ingredients had human data behind the testosterone claim, and that some products supplied vitamins and minerals above recommended upper intake levels (Balasubramanian et al., J Sex Med, 2019).

For erectile function specifically, a systematic review and meta-analysis of randomised trials of herbal dietary supplements found a small evidence base of limited quality (Ho et al., J Tradit Complement Med, 2026). That is not the same as “nothing works”. It is “we do not have the studies that would let anyone claim it does”.

Where the evidence is genuinely positive — and what for

Three findings survive scrutiny, and none of them is a testosterone claim.

Ashwagandha and stress. A randomised, double-blind, placebo-controlled trial reported reduced stress and anxiety scores over 60 days (Chandrasekhar et al., Indian J Psychol Med, 2012), and a meta-analysis found a small but significant improvement in sleep measures (Cheah et al., PLoS One, 2021).

Maca and desire. The most-cited randomised trial reported improved self-reported sexual desire from week eight, explicitly with no change in serum testosterone or oestradiol (Gonzales et al., Andrologia, 2002). A systematic review judged the maca evidence suggestive but methodologically limited (Shin et al., BMC Complement Altern Med, 2010).

Arginine and erectile-function scores. A meta-analysis of arginine supplements for erectile dysfunction reported improvement against placebo, with small samples and varied preparations (Rhim et al., J Sex Med, 2019).

Read as a set, they describe a category that may modestly influence stress, sleep and self-reported desire, and that has almost nothing to say about hormone levels. The detail behind the first two is in what the maca and ashwagandha trials actually found.

The four gaps between a trial and a jar

Every one of those gaps applies to a finished multi-ingredient product, which is why a formula that has never been tested as a formula cannot inherit the evidence of its parts. The MaxForce page says this about itself, which is the correct thing to say: the research concerns the individual ingredients, and the gummy has no clinical trial of its own.

What placebo has to do with it

Sexual and energy outcomes are self-reported, expectation-sensitive and highly responsive to attention. That is not a reason to dismiss a man’s experience — feeling better is the point — but it is the reason a placebo arm is non-negotiable when you want to know whether the compound did it. The uncontrolled studies in this field are the ones producing the most quotable numbers, and they are the ones least able to support a causal claim.

How to audit a product page in three minutes

1. Find the study behind the boldest claim. If the claim has no citation, the answer is already in.
2. Check the species and the population. Rodent studies and cell-culture studies are common in this category’s reference lists and cannot support a human claim.
3. Compare the trial dose with the label amount. This is where most claims fail. A gram-scale trial dose behind a milligram-scale label figure is not evidence for the product.
4. Check the outcome measured. A cortisol change is not an erection, and a questionnaire score is not a hormone level.
5. Check who paid. Not disqualifying, but funding and authorship belong in your reading of a single small trial.

Run that on any page in this category, including ours, and most of the argument settles itself. The label-side version of the same audit is in what a men’s vitality label must tell you.

So is it worth taking?

That depends on what you expect. If you expect a hormonal effect, the evidence does not support the purchase. If you want a low-risk daily habit that may support the stress and sleep side, at a cost you have calculated per day rather than per jar, and you have already dealt with sleep, weight and training, it is a defensible small addition. What it is not is a substitute for a diagnosis: the symptoms that need one are listed in signs of low testosterone and when to see a doctor, and the timeline for judging any of it is in how long a men’s vitality supplement takes to work.

What would change this answer

It is worth being explicit about what evidence would move the conclusion, because “the evidence is weak” is sometimes read as “the evidence is closed”. Three things would change it. A randomised, placebo-controlled trial of the finished formula rather than its parts, in men resembling the people who buy it. A per-ingredient breakdown published on the label, so the amounts could be compared with the amounts that were trialled. And replication: a single positive trial funded by a manufacturer is a hypothesis, not a finding.

None of those is unreasonable, and a few ingredients in this category have partly met them. That is precisely why the honest verdict differs by ingredient rather than applying to the category as a whole — the calibration exercise is what the tongkat ali research actually shows, where a genuinely interesting trial sits next to a headline claim it does not support.

Common mistakes when reading a product’s evidence page

1. Counting citations instead of reading them. A long reference list is easy to assemble. Check whether any of it tested the ingredient at the amount in the jar.
2. Accepting a mechanism as a result. “Supports nitric-oxide production” describes a pathway, not an outcome anybody measured.
3. Mistaking animal work for human evidence. Rodent data is where most of this category’s citations live, and it does not transfer.
4. Ignoring the comparator. An improvement with no placebo arm is not an improvement over placebo, which is the only comparison that matters.
5. Overlooking the free intervention. For the outcome most buyers actually want, the randomised evidence sits with training and sleep, as the stamina and cardio-fitness overlap lays out.

Frequently asked questions