Every supplement brand cites studies. We do too. But citing a study and being supported by evidence are not the same thing, and the gap between them is where most of this industry lives.
This is the checklist we use when we read a paper about one of our own ingredients. We are publishing it partly because it is genuinely useful, and partly because it is the fairest way to be judged: if you apply these questions to our pages, you will find the caveats we have already written into them.
Fair warning — every cautionary example below comes from research on red clover, sage, horsetail or flaxseed. These are the plants in our own product. We think that is the only honest way to write this article.
Part 1: Was there a real comparison?
Was there a control group at all?
Sage’s reputation in menopause rests heavily on one number: a 64% reduction in hot flash score over eight weeks, from a 2011 study of 71 women.
The number is real. The problem is the design. That study was open-label and uncontrolled — no placebo group, no blinding, and everyone knew they were taking sage. It was also conducted with the involvement of the manufacturer, whose employee was the study’s first author.
Without a comparison group, you cannot tell whether the improvement came from the product, from time passing, from the attention of being in a study, or from the simple expectation of getting better. A single-arm study can tell you a product is tolerable. It cannot tell you it works.
How large is the placebo effect in this condition?
This is the question almost nobody asks, and in menopause research it is decisive.
The Cochrane review of phytoestrogens for menopausal symptoms found that placebo groups improved by anywhere from 1% to 59%. Read that range again. In the phase III flaxseed trial, the placebo group’s hot flash score fell by 3.5 points. In a trial of S-equol, the placebo arm improved by 34.5%.
Hot flashes are self-reported, they fluctuate naturally, and they respond powerfully to expectation. So when an uncontrolled study reports a 64% improvement, a substantial part of that — possibly most of it — is the response you would have seen from a sugar pill.
Any percentage quoted without a placebo comparison is close to meaningless in this field.
Was it blinded, and who exactly was blinded?
“Double-blind” means neither the participants nor the researchers know who got what. “Single-blind” usually means only the participants are kept in the dark — and sometimes not even that.
A 2024 study of flaxseed in 145 perimenopausal women reported large improvements across three menopause questionnaires. It was single-blind, and every outcome was self-reported. Its placebo group improved by under 8% — far below the range Cochrane documented. When a trial’s placebo response is implausibly small, that is usually a signal about the study’s blinding, not about the product’s power.
Where outcomes are subjective — symptoms, mood, quality of life — blinding is not a technicality. It is the whole experiment.
Compared to placebo, or compared to itself?
This is the most common sleight of hand in supplement literature, and once you see it you cannot unsee it.
A 2012 trial tested a flaxseed lignan extract, flaxseed meal and a placebo in 90 women. Measured against their own starting scores, both flaxseed groups improved significantly (p = 0.007 and p = 0.005) while the placebo group did not (p = 0.082). That looks decisive.
But when the groups were properly compared against each other, the difference was not significant (p = 0.084).
Improvement against your own baseline is called a within-group comparison, and it captures everything — the product, the placebo effect, natural variation, seasonal change. The only test that isolates the product is the between-group comparison. If a brand quotes “significantly improved from baseline” without telling you how the placebo group did, you are being shown the wrong number.
Part 2: Does the result mean what the summary says?
Do the conclusions match the actual figures?
We assumed, when we started, that a paper’s abstract would fairly represent its own results. Twice in researching four ingredients, that turned out not to be true.
A 2023 meta-analysis of sage for hot flashes states in its abstract that the effect was significant compared with placebo. The confidence interval it reports for hot flash frequency is −2.37 to +0.14 — a range that crosses zero, meaning the result did not reach statistical significance by the conventional standard.
A 2016 meta-analysis of flaxseed concludes in favor of a beneficial effect on vasomotor symptoms. Its own pooled figure for hot flash severity is −0.203, with a confidence interval of −0.474 to +0.067 and p = 0.141. Again, not significant.
Neither paper is fraudulent. But both are widely cited by their optimistic conclusion sentence rather than their numbers. If a confidence interval crosses zero, the study did not demonstrate an effect, whatever the abstract says.
Was this the question the researchers set out to answer?
Trials are designed to test one main question — the primary endpoint. Everything else is exploratory, and the more comparisons you run, the more likely one turns up positive by chance.
In the main silicon trial in women with low bone density, lumbar spine bone density did not change significantly. The femoral neck result that gets quoted came from a subgroup analysis performed after the fact. Similarly, red clover’s most-cited meta-analysis is often summarized as showing benefit at doses of 80 mg a day or more — but that particular subgroup did not reach significance (p = 0.12) in the analysis itself.
A systematic review of 42 trials reporting equol status found that most positive findings came from underpowered secondary analyses, and that 28 of the 42 were primarily null.
Ask: was this finding the trial’s actual question, or something noticed afterwards?
Did the promising early result survive a proper trial?
This is the pattern that should make anyone cautious, because it has happened repeatedly with our own ingredients.
An uncontrolled 2007 pilot of flaxseed reported a 57% reduction in hot flash score and suggested this was more than placebo would give. The same research group then ran a proper phase III randomized trial in 188 women at a much higher lignan dose — and found no significant difference from placebo. They concluded the results did not support flaxseed lignans for hot flashes.
A 2010 uncontrolled pilot of a flaxseed lignan extract found that 80% of women showed a fall in a breast tissue proliferation marker. The randomized, placebo-controlled follow-up in 180 women, run by the same team, did not meet its primary endpoint — the difference disappeared once there was something to compare against.
Both pilots were honest science, doing exactly what pilots are for. The lesson is what happened next: promising uncontrolled results frequently do not survive randomization. When a brand cites the pilot and not the trial, that is a choice.
Part 3: Does it apply to you, and to this product?
Was it even the same ingredient?
This one is subtle and extremely common.
Much of the “silicon supports bone, hair and skin” evidence used choline-stabilized orthosilicic acid, a manufactured silicon compound — not horsetail. The two behave very differently: in human absorption studies, monomeric silicon forms were absorbed at 43–64% of the dose while polymerized colloidal silica came in at about 1%.
Most flaxseed menopause research used 25–40 g of whole ground seed, a food-sized portion. A lignan extract capsule is a different intervention at a different scale.
Roughly half the research on sage and cognition used Spanish sage essential oil (Salvia lavandulaefolia), not common sage leaf extract (Salvia officinalis) — different species, different preparation.
Same plant family, or same marketing category, does not mean same thing in the body.
Was it the same kind of person as you?
The best human evidence for horsetail is a diuretic trial — conducted in 36 healthy men over four days.
The largest study linking dietary silicon to bone density found the association in men and premenopausal women, but not in postmenopausal women — a qualification that appears in the paper’s own title, and one that matters a great deal for a menopause product.
One of the well-known hair supplement trials specifically excluded women who had used HRT or hormonal contraception in the previous six months — that is, a good portion of the menopausal population.
Was it one ingredient, or a formula?
If a study tests a product containing eight ingredients against an inert placebo, a positive result tells you the formula did something. It cannot tell you which part.
The hair-thinning studies often cited for horsetail tested a marine-protein supplement in which horsetail was one component alongside biotin and zinc — both of which carry their own authorized European claims for hair. The nail research tested a topical lacquer in which the film-forming agent, not the horsetail, is the presumed active ingredient.
To isolate one ingredient, you need the formula with it versus the formula without it. That study is rarely done, because it is expensive and risks an unhelpful answer.
Is the outcome one you would actually notice?
Many trials measure surrogate endpoints — a marker that stands in for the thing you care about. Bone mineral density instead of fractures. A cell-proliferation marker instead of cancer. Vaginal cytology instead of comfort. Arterial stiffness instead of heart attacks.
Surrogates are legitimate research tools and often the only practical option. But a change in a marker is a hypothesis about a benefit, not the benefit itself, and plenty of interventions have moved a marker without ever improving an outcome.
Part 4: Who ran it, and is the product what it claims?
Who paid for it?
Industry funding does not make a study wrong. Much good research is manufacturer-funded, simply because someone has to pay for trials.
But it is context. Of the sage menopause research, the most-quoted study and the best-designed trial were both conducted by the same manufacturer, involving the same employee. When the strongest evidence for an ingredient comes overwhelmingly from the company selling it, independent replication becomes the thing to watch for — and often the thing that is missing.
Is the product in the bottle the product in the study?
A study only transfers to a supplement if the supplement contains what it claims.
When researchers analyzed 33 commercial phytoestrogen supplements, they found “considerable differences in the isoflavone content from that claimed by the manufacturers.” When another group tested 34 retail horsetail products, 26 of them — 76% — carried chemical markers of contamination with a different, toxic Equisetum species. Analyses of commercial sage teas have found up to 20-fold differences in rosmarinic acid between brands.
This is why extract standardization and verified botanical identity are not marketing details. Without them, even perfect evidence for a plant tells you nothing about a particular jar of capsules.
The questions, in one place
When you next read that a supplement is “clinically proven,” try these:
Was there a control group? How big is the placebo effect in this condition? Was it double-blind? Was the comparison against placebo, or against the participants’ own baseline? Do the numbers match the conclusion? Was this the trial’s main question or a later subgroup? Did an earlier promising result survive a randomized trial? Was it the same ingredient, preparation and dose? Were the participants like you? Was it one ingredient or a whole formula? Is the outcome one you would notice? Who funded it? And is the product standardized and correctly identified?
Very few supplement claims survive all thirteen. Ours do not survive all of them either — which is precisely why our red clover, sage, horsetail and flaxseed lignan pages tell you where the evidence is strong, where it is mixed, and where it runs out entirely.
We would rather you bought from us knowing that than not knowing it.
Food supplements are not medicines, and individual responses vary.
References
- Bommer S, Klein P, Suter A. First time proof of sage’s tolerability and efficacy in menopausal women with hot flushes. Advances in Therapy. 2011. PubMed 21630133
- Lethaby A, et al. Phytoestrogens for menopausal vasomotor symptoms. Cochrane Database of Systematic Reviews. 2013. PubMed 24323914
- Shrivastava R, et al. Effects of flaxseed on perimenopausal symptoms: a single-blind, randomized, placebo-controlled study. Cureus. 2024. PMC11448742
- Colli MC, et al. Evaluation of the efficacy of flaxseed meal and flaxseed extract in reducing menopausal symptoms. J Med Food. 2012. PubMed 22925074
- Moradi M, et al. The effect of Salvia officinalis on hot flashes in postmenopausal women: a systematic review and meta-analysis. Int J Community Based Nurs Midwifery. 2023. PMC10363264
- Ghazanfarpour M, et al. Effects of flaxseed and Hypericum perforatum on hot flush, vaginal atrophy and estrogen-dependent cancers. Avicenna J Phytomed. 2016. PMC4930534
- Spector TD, et al. Choline-stabilized orthosilicic acid supplementation as an adjunct to calcium/vitamin D3 stimulates markers of bone formation in osteopenic females. BMC Musculoskelet Disord. 2008. PMC2442067
- Kanadys W, et al. Evaluation of clinical meaningfulness of red clover extract to relieve hot flushes and menopausal symptoms. Nutrients. 2021. PMC8069620
- Birru RL, et al. Equol status and cardiovascular risk factors: a systematic review. J Nutr Sci. 2016. PMC4976117
- Pruthi S, et al. Pilot evaluation of flaxseed for the management of hot flashes. J Soc Integr Oncol. 2007. PubMed 17761129
- Pruthi S, et al. A phase III, randomized, placebo-controlled, double-blind trial of flaxseed for the treatment of hot flashes: NCCTG N08C7. Menopause. 2012. PubMed 21900849
- Fabian CJ, et al. Reduction in Ki-67 in benign breast tissue of high-risk women with the lignan secoisolariciresinol diglucoside. Cancer Prev Res. 2010. PMC2955777
- Fabian CJ, et al. Randomized phase IIB trial of the lignan secoisolariciresinol diglucoside in premenopausal women at increased risk for breast cancer. Cancer Prev Res. 2020. PMC7335358
- Sripanyakorn S, et al. The comparative absorption of silicon from different foods and food supplements. Br J Nutr. 2009. PMC2744664
- Jugdaohsingh R, et al. Dietary silicon intake is positively associated with bone mineral density in men and premenopausal women of the Framingham Offspring cohort. J Bone Miner Res. 2004. PubMed 14969400
- Carneiro DM, et al. Randomized, double-blind clinical trial to assess the acute diuretic effect of Equisetum arvense in healthy volunteers. Evid Based Complement Alternat Med. 2014. PMC3960516
- Ablon G. A 3-month randomized, double-blind, placebo-controlled study evaluating a marine complex supplement in women with self-perceived thinning hair. Dermatol Res Pract. 2015. PMC4389977
- Setchell KDR, et al. Bioavailability of pure isoflavones in healthy humans and analysis of commercial soy isoflavone supplements. J Nutr. 2001. PubMed 11285356
- Nowak M, et al. Biomarker-based determination of Equiseti herba contamination by Equisetum palustre using HPLC-MS/MS. Planta Medica. 2022. PubMed 34293807
- Walch SG, et al. Antioxidant capacity and polyphenolic composition as quality indicators for aqueous infusions of Salvia officinalis L. Front Pharmacol. 2011. PMC3242359
This page is for general information and is not a substitute for medical advice. Food supplements should not be used as a substitute for a varied, balanced diet and a healthy lifestyle.