Acupuncture Shows Hints of Benefit for Vascular Dementia, but Evidence Falls Short

A sweeping new analysis of acupuncture research for vascular dementia has delivered a verdict that is as sobering as it is revealing: across sixteen systematic reviews and meta-analyses covering hundreds of randomized controlled trials, most pooled estimates favored acupuncture for improving cognition and daily functioning, yet nearly all of that apparent benefit rests on evidence graded as low or very low certainty. The overview, published in Frontiers in Neuroscience by researchers at Guangzhou University of Chinese Medicine and collaborators in New Zealand and Australia, is among the first disease-focused umbrella reviews to examine every acupuncture modality specifically for vascular dementia, and its findings expose a widening gap between a rapidly growing research literature and the quality of evidence needed to guide clinical practice.

Vascular dementia is the second most common cause of dementia worldwide, accounting for roughly twenty percent of all cases globally and about twenty-six percent among Chinese adults aged sixty and above. It arises when cerebrovascular disease starves the brain of adequate blood flow, producing a cascade of ischemic insults that damage neurons through several interacting mechanisms. Reduced cerebral perfusion impairs the cholinergic system, lowering acetylcholine synthesis and degrading learning and memory. Ischemia also triggers oxidative stress, an imbalance between free radical production and clearance that drives lipid peroxidation and structural neuronal injury. Excessive glutamate release during ischemic episodes overactivates NMDA receptors, flooding neurons with calcium in a process known as excitotoxicity. Meanwhile, activated microglia release inflammatory cytokines such as tumor necrosis factor-alpha and interleukin-1-beta, compounding the damage. Crucially, these insults disrupt activity-dependent synaptic plasticity and impair long-term potentiation, the cellular substrate of learning and memory.

Current treatment options are limited. Pharmacological approaches, including agents to improve cerebral microcirculation such as nimodipine, cholinesterase inhibitors such as donepezil, and NMDA receptor antagonists such as memantine, offer modest efficacy with potential adverse reactions and relatively high costs. Non-pharmacological measures like rehabilitation training and hyperbaric oxygen therapy serve mainly as adjuvants. Against this backdrop, acupuncture has attracted growing clinical interest as a non-pharmacological adjunct, and a substantial body of systematic reviews and meta-analyses has emerged, most concluding that acupuncture improves cognitive function and activities of daily living in vascular dementia patients. But notable heterogeneity among these reviews, in inclusion criteria, intervention details, and outcome measures, has produced inconsistent conclusions and undermined confidence in the existing evidence.

To resolve this uncertainty, the research team conducted a comprehensive overview following the Cochrane Handbook and PRISMA 2020 reporting standards, registered prospectively on PROSPERO. They searched eight biomedical databases from inception to October 2025, including CNKI, Wanfang, VIP, CBM, PubMed, Embase, Web of Science, and the Cochrane Library, alongside gray literature and clinical trial registries. From 277 initially retrieved records, sixteen systematic reviews and meta-analyses of randomized controlled trials met the eligibility criteria. The reviews were assessed for methodological quality using AMSTAR-2, for completeness of reporting against the PRISMA 2020 checklist, and for certainty of evidence using the GRADE system. The team also performed a secondary data mining analysis of the primary trials to characterize meridian and acupoint selection patterns, and calculated the Corrected Covered Area to quantify overlap among the reviews.

The methodological findings were concerning. According to AMSTAR-2, only one of the sixteen reviews was rated high quality, one moderate, six low, and eight critically low, meaning 87.5 percent fell into the low or critically low categories. The dominant critical flaw was the failure to develop and register a protocol before the review commenced, present in fourteen of sixteen reviews. Other recurring weaknesses included incomplete literature search strategies, failure to provide lists of excluded studies with justifications, and inadequate consideration of primary study risk of bias when interpreting results. Completeness of reporting showed similar gaps: protocol and registration was adequately reported in only two of sixteen reviews, and funding information in just nine. These deficiencies increase the risk of selective reporting bias and compromise transparency and reproducibility.

The GRADE assessment was even more striking. Of ninety-seven individual outcomes evaluated across the reviews, only one was rated high quality and one moderate quality. Fourteen outcomes were graded low and eighty-one very low, meaning 97.9 percent of all outcomes carried low or very low certainty. The primary downgrading factor was study limitations, specifically risk of bias in the primary randomized controlled trials, which affected all ninety-seven outcomes. Inadequate allocation concealment and lack of blinding were pervasive, creating high risks of performance and detection bias. Imprecision, driven by wide or very wide confidence intervals, affected 58.8 percent of outcomes, high heterogeneity affected 48.5 percent, and suspected publication bias affected 37.1 percent. Despite these limitations, most pooled estimates favored acupuncture for cognitive and functional outcomes, with reported mean differences for the MMSE ranging from 0.87 to 5.06 points and standardized mean differences from 0.39 to 1.09, though the authors deliberately refrained from applying any generic minimal clinically important difference threshold, since VD-specific MCIDs have not been established.

The data mining analysis of 239 unique randomized controlled trials with complete prescription data revealed clear patterns in acupoint selection. The Governor Vessel was the most frequently selected meridian, accounting for 18.8 percent of total usage, followed by the Bladder Meridian at 10.5 percent and the Stomach Meridian at 9.8 percent. Among specific acupoints, Baihui (GV20) was the most frequently used at 10.9 percent, followed by Sishencong (EX-HN1) at 8.4 percent and Zusanli (ST36) at 6.7 percent, with Fengchi, Shenting, Neiguan, and Shuigou also appearing frequently. The data indicate a preference for cranial and facial points supplemented by distal limb points, consistent with a Local plus Distal treatment approach. The authors emphasize that these patterns are hypothesis-generating rather than validated treatment recommendations, reflecting clinical popularity and practitioner consensus rather than established comparative effectiveness.

The overview also quantified overlap among the sixteen reviews using the Corrected Covered Area, calculating a value of 2.09 percent from 318 primary study occurrences and 242 unique trials. According to established thresholds, this indicates slight overlap, suggesting the reviews relied on largely distinct pools of primary trials, which reduces but does not eliminate concerns about redundancy and double-counting. Notably, the Corrected Covered Area formula accounts for the number of reviews, the total occurrences, and the unique studies, providing a standardized metric for quantifying redundancy in umbrella reviews. The low overlap value here suggests that the sixteen synthesized reviews drew from substantially different trial pools rather than repeatedly recycling the same studies.

The authors identify several distinctive characteristics of acupuncture research compared with other non-pharmacological dementia interventions. Behavioral therapies such as exercise and cognitive training typically struggle with high attrition rates and performance bias because participant blinding is difficult. Acupuncture research, by contrast, suffers from heterogeneity arising from variations in stimulation parameters, acupoint selection, and control conditions, along with a widespread absence of rigorous sham controls. Because very few of the included trials used sham-acupuncture controls, the pooled estimates cannot isolate specific needling effects from non-specific contextual and placebo responses. The authors also note that all sixteen reviews were conducted by Chinese researchers, and the original studies were predominantly performed and published in China, raising questions about generalizability. Cultural familiarity with acupuncture in Chinese populations may amplify non-specific contextual effects, patient expectations, and placebo responses that may not replicate in Western countries where acupuncture is perceived as an unfamiliar alternative.

Looking forward, the authors propose a consensus-based reference acupoint combination for prospective testing in future trials, following a Local Cranial plus Distal Limb approach. The core cranial module would prioritize GV20 and EX-HN1 to awaken the mind through local stimulation, integrated with a systemic support module using ST36 to fortify the spleen and replenish qi, addressing systemic deficits often present in elderly vascular dementia patients. This strategy draws support from Traditional Chinese Medicine theory, where the Governor Vessel enters the brain, and from a recent meta-analysis in Ageing Research Reviews confirming that Governor Vessel acupuncture therapy significantly improves cognitive function and daily living activities in post-stroke cognitive impairment. However, the authors stress that these mechanisms remain external hypotheses warranting direct biomarker investigation in vascular dementia cohorts. Future research should prioritize rigorous, large-scale, sham-controlled randomized controlled trials testing standardized candidate protocols, with appropriate random sequence generation, allocation concealment, blinding of participants and outcome assessors, internationally recognized core outcome sets, and standardized adverse event monitoring. Only through such methodological rectification can the field generate evidence of sufficient quality to support incorporation of acupuncture into international clinical guidelines for vascular dementia.

Subject of Research: Acupuncture for vascular dementia: an overview of systematic reviews and meta-analysis

Article Title: Acupuncture for vascular dementia: an overview of systematic reviews and meta-analysis

Article References: Chen, Y., Fan, Y., Wang, Y., Li, X., Lin, H., Li, X., Tang, X., Xu, N., & Wu, Z. (2026). Acupuncture for vascular dementia: an overview of systematic reviews and meta-analysis. Frontiers in Neuroscience, 20, Article 1796611. https://doi.org/10.3389/fnins.2026.1796611

Image Credits: AI Generated

DOI: 10.3389/fnins.2026.1796611

Keywords: Acupuncture, vascular, dementia, overview, systematic, reviews, meta-analysis, scientific research