Vascular dementia is the second most common form of dementia after Alzheimer's, affecting an estimated 8.5 million people worldwide, and it has no treatment that halts or reverses its course. So when a new systematic review in Frontiers in Neurology reports that a traditional Chinese herbal approach was associated with better cognitive scores across 38 randomized trials, the headline practically writes itself.
The more interesting story is what it takes to read those results honestly. This is one of those papers where the methods section is more valuable than the conclusion — and where understanding why requires knowing a few things about how clinical evidence is built.
What the researchers actually did
Zhou and colleagues (2026) set out to evaluate kidney-tonifying and blood-activating therapy — abbreviated KTBAT — for vascular dementia. In the traditional framework, vascular dementia is associated with "kidney deficiency" (a depletion of the body's foundational reserves) and "blood stasis" (impaired microcirculation), so formulas that combine kidney-tonifying herbs with blood-activating herbs are a classical treatment strategy. Classic formulas in this family include Buyang Huanwu Tang and Dihuang Yinzi.
The review's conduct is genuinely solid for this literature:
- Pre-registered protocol (PROSPERO CRD420251272344) and PRISMA 2020 reporting standards.
- Eight databases searched from inception through January 2026, including Chinese-language sources (CNKI, Wanfang, VIP, Sino-Med) alongside PubMed, Embase, Web of Science, and Cochrane.
- RCTs only: 38 trials, 3,357 adults aged 40 and older, diagnosed with vascular dementia.
- Risk of bias assessed with the Cochrane RoB 1 tool.
One important operational detail: the review pooled interventions by principle, not by formula. Decoctions, granules, capsules, tablets, and patent medicines were all eligible as long as they contained both kidney-tonifying and blood-activating components. That means the analysis measures the overall effect of a therapeutic principle — not the effect of any single formula you could buy.
What they found
The pooled numbers, taken at face value, are striking:
- Clinical effectiveness rate — the proportion of participants meeting each trial's own definition of "effective" — was about 25% higher with the herbal approach than with Western medicine alone, whether the herbs were used alone (RR 1.26) or added on (RR 1.25).
- Cognitive test scores improved. On the Mini-Mental State Examination (MMSE, a 30-point scale), the herbal groups averaged about 3.7 points higher than controls. The Hasegawa Dementia Scale showed a similar advantage of about 4 points. The Clinical Dementia Rating scale also favored the herbs, but only three small trials reported it.
- Neurotrophic factors — blood markers like BDNF and NGF that support neuron survival — were higher in the herbal groups, though the BDNF estimate shrank substantially when one influential trial was removed.
- Adverse events did not differ significantly between groups in the six trials that could be pooled, and no serious events or deaths were reported.
Effects were observed during treatment periods of one to six months. That's the paper. Now here's the part the headline skips.
What the numbers actually mean — and what they don't
"Clinical effectiveness" is not a standardized outcome. Each of the 38 trials defined effectiveness by its own criteria. A pooled risk ratio of 1.25 means participants were more likely to meet their own trial's bar — not that they improved by a fixed, clinically anchored amount. This is a legitimate but soft outcome, and it's worth less weight than the standardized cognitive scales.
A 3.7-point MMSE gain is real but modest. On a 30-point scale, in patients whose scores are already impaired, a few points can matter to daily life — but it is not a restoration of cognition, and it was measured during treatment, with no follow-up to tell us whether the gain persisted after the herbs were stopped. The review tells us nothing about whether the therapy changes the disease's course.
The biomarkers are hints, not proof. Blood levels of BDNF or NGF are intermediate endpoints. The reviewers themselves stress that a shift in a circulating protein does not demonstrate neuronal protection or clinical benefit. Treating these as mechanism evidence would be overreading.
Subgroup patterns are observational. The analysis suggests capsules and decoctions were associated with larger gains than granules, and that longer treatment showed larger differences — but these are comparisons between trials, not randomized comparisons. They cannot be read as dosing guidance.
The four caveats that should travel with the headline
-
The entire evidence base is geographically and methodologically narrow. All 38 trials were conducted in China, published in Chinese, generally small and single-center. Not one trial provided sufficient detail on allocation concealment, and only three clearly described blinding. In trials of a subjectively scored outcome like a cognitive test, incomplete blinding is exactly the kind of flaw that inflates effect sizes.
-
Publication bias is probable. The funnel plots for the main outcomes were asymmetric — the classic fingerprint of small-study effects, where small positive trials get published and small negative ones disappear. The true average effect is very likely smaller than the pooled estimate.
-
"The herbs" in this headline is not one thing. Ingredient sets, combinations, dosages, and dosage forms differed substantially across the 38 trials. There is no standardized product behind these numbers, and no basis for picking a specific formula from them.
-
The safety evidence is thin. Of 38 trials, 21 did not mention adverse events at all. Only nine reported them, and three crossover trials couldn't be pooled. "No evidence of increased adverse events" in six small, short trials is not the same as "evidence of no harm" — uncommon or delayed harms would be invisible in this dataset.
The bottom line
This is a well-conducted synthesis of a weak literature — and that combination is more useful than it sounds. It does not crown a treatment. What it does is map exactly where a signal exists, on which outcomes, and how much skepticism each number deserves. That is precisely the information a field needs to design the trial that actually settles the question: large, multicenter, properly blinded with a matched placebo, a standardized formula, diverse populations, and long-term follow-up on cognition and daily function — which is exactly what the authors call for.
For readers, the practical guidance is simple and should not be skipped: do not assemble a formula from a headline. Blood-activating herbs can interact with anticoagulant and antiplatelet medications, product quality varies enormously, and none of this evidence tells us what any specific product does. Anyone considering anything in this family of therapies should discuss it with a qualified clinician who knows their full medication list.
Evidence, at its best, doesn't tell you what to believe. It tells you how much you can afford to hope — and what would have to happen before that hope becomes a recommendation. This paper does exactly that, if you read past the headline.
If you are interested in how the classical Chinese medical tradition thinks about health more broadly — its ideas of balance, prevention, and the body's patterns — our book Huangdi Neijing: A Modern Introduction offers a beginner-friendly guide to the foundational text of TCM.
This article is for educational purposes only and is not medical advice. If you or a family member are experiencing cognitive symptoms, please consult a qualified healthcare professional.
References
- Zhou M, Che M, Zhang C, Yang X, Liu X, Liu Z, Huang J, He H. Efficacy and safety of traditional Chinese medicine kidney-tonifying and blood-activating therapy in the treatment of vascular dementia: a systematic review and meta-analysis. Frontiers in Neurology, 2026, 17, 1901245. https://doi.org/10.3389/fneur.2026.1901245
- PROSPERO registration CRD420251272344. https://www.crd.york.ac.uk/PROSPERO/view/CRD420251272344