Evidence explained

Low-Certainty Evidence: What Cochrane Ratings Mean For The Names On This Label

When a supplement page says an ingredient is “clinically studied,” the most useful follow-up question is how confident the people who reviewed those studies were. Cochrane reviews answer that question in a standard vocabulary, and several of the names on the Primal Grow Pro label have been through one. This article explains that vocabulary, then reads what the reviewers actually said about ginkgo, huperzine A and vinpocetine, and notes which of the seven names have no Cochrane review at all.

Sourced from the product label and PubMed-indexed research · Published September 27, 2026 · 13 min read · Evidence explained

Why a reviewer’s confidence is its own fact

There are two different questions you can ask about an ingredient. One is what the trials found. The other is how much weight the trials can bear. Most supplement writing answers only the first, because it is the more marketable one: a study found a benefit, therefore. The second question is the one systematic reviewers spend their careers on, and their answer is often the more informative of the two.

Cochrane is an organisation that publishes systematic reviews of health interventions. In a Cochrane review, two authors independently select trials, extract results and assess each study for risk of bias, and the results are pooled where that makes sense. Newer Cochrane reviews then rate the certainty of the evidence for each outcome using a framework called GRADE. The older reviews discussed below do not report GRADE labels in their abstracts and rely on narrative statements about quality and bias instead, and I will say so where it applies.

That distinction is worth holding onto, because the Primal Grow Pro label names seven ingredients and gives no amount for any of them. With no printed figures to check against a trial, the most valuable thing left is the reviewers’ verdict on the trials themselves. This article is not about whether any ingredient works. It is about what independent reviewers concluded when they lined up the evidence, and what their wording means. The focus here is on certainty ratings and reviewers’ conclusions, not on how large any one trial was or who took part in it.

What the rating words mean

GRADE, short for Grading of Recommendations Assessment, Development and Evaluation, is the framework. Its founding description, published in the Journal of Clinical Epidemiology, specifies four categories, high, moderate, low and very low, that are applied to a body of evidence rather than to individual studies. In a systematic review, the rating reflects confidence that the estimate of effect is correct. Randomized trials start as high-certainty evidence and observational studies start as low, and the rating can then go down for reasons that go beyond the trial’s internal quality (Balshem et al., 2011).

The imprecision, inconsistency, indirectness and publication bias entries summarise the GRADE guideline abstracts cited in the references; the risk-of-bias entry refers to the ginkgo review described below.

Two things follow. First, a rating can be low for reasons that have nothing to do with the ingredient: a small evidence base is rated down for imprecision however good the ingredient is. Second, indirectness matters more than it sounds. A review of people with dementia is indirect evidence for anyone else, because the population is different. Keep that in mind, because every review below studied patients.

Ginkgo in cognitive impairment and dementia

Ginkgo has the newest and most detailed review on the label. Wieland and colleagues published a Cochrane review of ginkgo for cognitive impairment and dementia in 2026. It searched to November 2024 and included 82 randomized trials with 10,613 participants, of which 72 studies with 9,783 participants provided usable data. Only four studies were at low risk of bias in every domain. Key outcomes were assessed at six months, and each estimate carries a GRADE certainty label (Wieland et al., 2026). The results are separated by group, and that separation is the point.

• People with subjective cognitive complaints (three studies, 597 participants). For global clinical status the review rated the evidence very low certainty and found it uncertain whether ginkgo has any effect. One larger three-month study suggested adverse events may be higher with ginkgo.
• Multiple sclerosis with cognitive problems (two studies, 164 participants). Moderate-certainty evidence that ginkgo probably has little or no effect on cognition over three months.
• Mild cognitive impairment (twelve studies, 1,913 participants). Moderate-certainty evidence that ginkgo probably has little or no effect at six months on global status, cognition or activities of daily living. For serious adverse events, high-certainty evidence of little or no difference.
• Dementia (thirteen studies, 3,288 participants). Low-certainty evidence that ginkgo may give small benefits at six months on global status, cognition and daily activities. The heterogeneity statistics for those outcomes were very high, I² of 88, 96 and 91 percent, which is exactly the inconsistency the framework flags.

That is one review, and it contains four different certainty stories. The reviewers’ own summary is that in people with cognitive complaints they are unsure whether ginkgo improves global clinical status, that in mild cognitive impairment it probably has little or no effect, and that in dementia there may be small to moderate benefits (Wieland et al., 2026). Notice how much of that is hedged, and how differently each group is treated.

There is also an earlier Cochrane review of the same question, published in 2009. It included 36 trials, noted that many of the early ones used unsatisfactory methods and were small, said publication bias could not be excluded, and concluded that the evidence of predictable, clinically significant benefit for people with dementia or cognitive impairment was inconsistent and unreliable, while ginkgo appeared safe with no excess side effects compared with placebo (Birks and Grimley Evans, 2009). Comparing the two is instructive: more trials, a formal certainty rating, and a more finely grained answer, but a broadly similar caution.

What does any of this say about a capsule for healthy adult men? Very little directly. Every group in that review had a cognitive complaint or a diagnosis. That is indirectness, in GRADE terms, and it is the honest limit on carrying these results to anyone else. The label’s own claim is about mental sharpness in everyday life, which none of the reviews measured.

Ginkgo in leg-artery walking distance

The label lists circulation among its four supports, and ginkgo is the name that carries it. The circulation question has its own Cochrane review, on ginkgo for intermittent claudication, the leg pain from narrowed arteries that appears on walking. The 2013 update included 14 trials with 739 participants. Eleven of them, with 477 participants, compared ginkgo with placebo and measured absolute claudication distance. At the end of treatment the distance increased by an overall effect of 3.57 kilocalories of treadmill work, with a confidence interval running from −0.10 to 7.23 (P = 0.06). The authors translated that into an increase of just 64.5 metres on a flat treadmill at 3.2 km/h, with an interval of −1.8 to 130.7 metres. They added that publication bias leading to missing negative trials was likely to have inflated the effect, and concluded that there is no evidence that ginkgo has a clinically significant benefit for patients with peripheral arterial disease (Nicolaï et al., 2013).

Map that onto the rating vocabulary and two reasons to lose confidence stand out. The interval crosses zero, which is imprecision: the data are compatible with a small benefit and with none. And the authors themselves flag likely publication bias. The abstract does not give a GRADE label, and I will not invent one. What it gives is the reviewers’ reasons, and they are the same reasons GRADE lists. This is also the circulation evidence that the site’s article on the four supports discusses in terms of which ingredient carries which support.

Huperzine A

Huperzine A has two Cochrane reviews, both of older vintage, and they read very differently from each other.

The review for Alzheimer’s disease, published in 2008, included six trials with 454 patients and noted that the methodological quality of most was not high. Against placebo, huperzine A showed benefit on several scales. For example, the mean difference on the Mini-Mental State Examination was 2.81 points (95% CI 1.87 to 3.76), and on the ADAS-Cog at 12 weeks it was 2.51 (1.74 to 3.28). On two other measures it was not superior to placebo. Adverse events were mild with no significant difference from control groups. But the authors’ conclusion was cautious: only one study was of adequate quality and size, and there was therefore inadequate evidence to make any recommendation about its use (Li et al., 2008).

That paragraph is a good lesson in why effect sizes are not the same as certainty. The numbers on their own look encouraging. The reviewers still declined to recommend, because the studies were mostly of low methodological quality, which is a risk-of-bias problem, and because small trials give fragile estimates. They also had to compare trials that used different scales, which is one more source of uncertainty.

The review for vascular dementia, published in 2009, is starker. It found only one small trial, with 14 participants, and no significant benefit on the Mini-Mental State Examination (difference 2.40, 95% CI −4.78 to 9.58). The authors pointed out that the confidence intervals were wide and included both clinically significant benefits and clinically significant harms, and concluded there was no convincing evidence that huperzine A is of value in vascular dementia (Hao et al., 2009). That is imprecision in its purest form: an interval so wide it cannot tell you which way the truth points.

Neither review studied a healthy population, and neither can be matched to a label that prints no huperzine amount. The site’s piece on why huperzine is measured in micrograms covers the dose side of that story.

Vinpocetine

The Cochrane review of vinpocetine for cognitive impairment and dementia, from 2003, is the oldest here and the most explicit about its limits. It included three studies with a total of 583 people with dementia. All the identified studies had been performed before the 1990s and used varying terms and criteria for cognitive decline. The results showed benefit with vinpocetine at 30 mg and 60 mg a day compared with placebo, but the number of patients treated for six months or more was small, and only one study extended treatment to a year. Adverse effects were inconsistently reported, and intention-to-treat data were not available for any of the trials. The reviewers concluded that the evidence for a beneficial effect in dementia was inconclusive and did not support clinical use, and that the drug seemed to have few adverse effects at the doses studied (Szatmari and Whitehouse, 2003).

Those doses are trial doses of a compound that has been used as a drug, and the label prints no vinpocetine amount at all. What is worth taking from the review is the structure of the doubt. Old trials with inconsistent definitions, small numbers at longer durations, missing intention-to-treat data: each of those would today be a reason to rate the certainty down. The reviewers reached, in 2003 prose, roughly the place the modern framework would send them.

The names with no review, or the wrong one

One of the seven names has no Cochrane review that I could find, and three more have reviews that answer a different question from the one this label raises. Here is what I found by searching PubMed for Cochrane Database of Systematic Reviews entries under each name. A title search can miss a review that mentions an ingredient only in its text, so treat this as what I found rather than an exhaustive census.

• Bacopa Monnieri: none found. A title search of the Cochrane journal for bacopa returned nothing. Bacopa does have non-Cochrane meta-analyses. One pooled nine trials of at least 12 weeks with 518 subjects and reported low risk of bias by the Cochrane tool (Kongkeaw et al., 2014), and a 2026 network meta-analysis included 29 trials (Tiemtad et al., 2026). Neither abstract reports GRADE certainty ratings, so I will not attach one. This is why the site has a separate article on what “standardized bacopa” means when no bacoside figure is printed. Lack of a Cochrane review is not a verdict; it means the ingredient has not been through that particular process.
• St. John’s Wort: a review, for depression. Linde and colleagues’ Cochrane review covered 29 trials with 5,489 patients with major depression. It reported marked heterogeneity in the placebo-controlled trials: the nine larger trials gave a response rate ratio of 1.28 (1.10 to 1.49) while the nine smaller ones gave 1.87 (1.22 to 2.87), and trials from German-speaking countries reported more favourable findings. In the reviewers’ words, the association of country of origin and precision with effect size complicates the interpretation (Linde et al., 2008). It says nothing about vitality, energy or circulation, so for this label’s claims it is indirect.
• L-Carnitine: reviews for other conditions, and one very relevant one. Cochrane reviews of carnitine and its derivatives exist for dialysis, diabetic neuropathy and claudication, among others. The one closest to a healthy user is a 2017 review of L-carnitine for cognitive enhancement in people without cognitive impairment. Only two trials were eligible, all participants were young adults, the risk of bias was unclear, and the evidence was of very low quality for all reported outcomes. The authors were unable to draw any conclusions about efficacy or safety (Chen et al., 2017). It is a Cochrane review that concludes the evidence is not there yet.
• L-Glutamine: reviews for other populations. Cochrane reviews of glutamine cover critically ill adults, preterm infants and Crohn’s disease among others (Tao et al., 2014, for example, on critically ill adults). None that I found addresses glutamine for energy in healthy men.

All seven names in one table

The 2009 ginkgo entry is the earlier review of the same question as the 2026 one. Nothing in this table is a rating of the capsule.

How to read a rating without over-reading it

It would be easy to walk away from a page like this thinking that low certainty means the ingredients are useless. That is not what the reviewers said, and it is worth being clear about what they did say.

• Low certainty is not the same as no effect. It means the reviewers are not confident about the estimate. For dementia, ginkgo’s estimate was of small benefits, held with low certainty. That is a different statement from “ginkgo does nothing.”
• Low certainty is not the same as harm. In the ginkgo review, the certainty for serious adverse events in mild cognitive impairment was actually high, and it showed little or no difference. Uncertainty about benefit and confidence about safety can sit side by side.
• A rating belongs to a population. The reviews above are about people with dementia, mild cognitive impairment, claudication or depression. Their ratings do not transfer to healthy men, and that is not a defect in the reviews; it is what indirectness means.
• A rating belongs to a preparation and dose. The trials tested particular extracts and doses. A label that prints no amount cannot be matched against them, so the reviews describe those products, not this capsule.
• Absence of a review is information about process, not about the ingredient. Bacopa has no Cochrane review in what I found; that does not make it worse than an ingredient that has one with a discouraging conclusion.

The practical use of all this is to change the questions you ask a product page. Instead of “is it studied,” ask: in whom, at what dose, with what certainty, and did anyone independent line the studies up? For most of the names on this label the truthful answer is that independent reviewers looked at other people, used other doses, and reached cautious conclusions. For the claim the label leads with, everyday male vitality, the site’s article on what that phrase can mean works through how far any of this can reach.

References