Background
Migraine, a highly disabling neurological disorder, has a heavy global epidemiological burden: it causes 450,000–1,000,000 Years Lived with Disability (YLDs), accounts for 5–6% of global disease burden,1 and mainly affects 30–39-year-olds, impairing productivity. The Eurolight project found migraine impacts child development, partner relationships, and patients’ work productivity, with higher attack frequency worsening family burden.2 The US CaMEO study noted that high-frequency migraine causes family negative emotions and less social participation, with a greater impact in single-parent families.3,4 Migraine progression increases risks of anxiety, depression, asthma, epilepsy, and stroke.5
Current migraine treatments include acute/preventive medications and non-pharmacological therapies. For acute attacks, triptans are first-line agents, yet approximately 30% of patients are non-responsive; long-term frequent use may induce medication-overuse headache.6 Non-steroidal anti-inflammatory drugs (NSAIDs) carry cardiovascular and gastrointestinal risks. For prevention, calcitonin gene-related peptide (CGRP) monoclonal antibodies reduce monthly migraine days by more than 50% but require parenteral injection.7 Traditional oral drugs have many adverse effects and poor adherence.
Opioids are only for extreme cases due to high dependence risk; ergot derivatives are contraindicated in cardiovascular patients for vasoconstrictive effects.8
Accumulating clinical evidence supports the therapeutic value of acupuncture in migraine management. Zhao et al confirmed that real acupuncture outperforms sham acupuncture in reducing migraine frequency with sustained efficacy.9 Neuroimaging work by Zou et al further revealed that acupuncture reverses aberrant functional connectivity within the default mode network in chronic migraine patients, and that connectivity restoration correlates with immediate Visual Analog Scale (VAS) scores, suggesting a brain-network remodeling mechanism.10
Despite these promising findings, substantial methodological and interventional heterogeneity across studies limits the generalizability of the evidence and impedes the development of a standardized acupuncture prescription. Specifically, acupuncture protocols vary widely in treatment frequency, course duration, and acupoint selection;11–13 control designs remain methodologically contested, with sham acupuncture carrying a placebo effect of up to 35%;14 and outcome measures are inconsistent across trials, reducing the strength and reliability of pooled evidence.11 Consequently, conventional meta-analyses—which pool efficacy estimates across these heterogeneous trials—cannot disentangle which protocol elements drive the observed therapeutic benefits.
A scoping review is well suited to address this gap, as it systematically maps the breadth and distribution of existing evidence rather than quantifying intervention efficacy.15 This approach is particularly appropriate for acupuncture research, where multiple interdependent variables and multidisciplinary mechanistic evidence resist integration by traditional meta-analysis. The present scoping review therefore aims to: synthesize the acupuncture-for-migraine literature and map patterns of acupoint use and combination to inform evidence-based prescription standardization; and identify methodological limitations in current research to guide the design of future rigorous trials.
Methods
Identifying the Research Problem
This scoping review adhered to the PRISMA Extension for Scoping Reviews (PRISMA-ScR). We conducted the literature review in accordance with the method described by Arksey and O’Malley.
It was registered on the OSF, an international database of scoping reviews (https://doi.org/10.17605/OSF.IO/G59N8).
After a preliminary review, the team defined the research question: What is the current status of research on acupuncture for migraine treatment? Specifically, the following sub-questions were addressed: (1) What is the quality of the evaluated evidence? (2) What acupuncture treatment protocols are used in clinical studies?
(3) What indicators are employed to assess the evidence for the efficacy and safety of acupuncture? (4) What are the future research trends?
Search Strategy
We searched the following domestic and international databases, with the search period covering from the inception of each database to January 13, 2026: China National Knowledge Infrastructure (CNKI), WanFang Data Knowledge Service Platform (WanFang), Chinese Biomedical Literature Database (SinoMed), PubMed, EMBASE, Cochrane Library, and Web of Science. A comprehensive search strategy was developed for each database, combining terms such as “migraine”, “migraine disorders”, “migraine disorder”, “Hemicrania”, “acupuncture”, “acupoint”, “acupuncture point”, and “Needling”, with searches conducted in the title, abstract, and keywords fields. Detailed search strategies for different databases are provided in Appendix A.
Eligibility Criteria
The inclusion criteria for studies and literature in this review were as follows: Clinical studies on patients with migraine.Studies that implemented various acupuncture therapies (eg, ordinary acupuncture, electroacupuncture, auricular acupuncture, warm acupuncture, intradermal needling, and bloodletting therapy) as therapeutic interventions for migraine patients, with explicit mention of “acupoints” or “acupuncture points”.Prospective experimental clinical studies, including single-arm trials, open-label RCTs, and double-blind RCTs.
The exclusion criteria were: Clinical studies with unclear or non-validated diagnostic criteria for migraine, or studies not enrolling patients with migraine.Studies in which acupuncture was co-administered with another active therapy in the experimental group; studies without explicit mention of “acupoints” or “acupuncture points”; or studies with unverifiable acupoint nomenclature or location.Non-prospective-experimental study designs, including systematic reviews, scoping reviews, narrative reviews, editorials, letters, commentaries, book chapters, conference proceedings, unpublished abstracts, retrospective observational studies, animal experiments, and in vitro cell studies.Duplicate publications, retracted publications, studies carrying an editorial expression of concern, and articles whose full texts were unavailable or irretrievable despite systematic attempts.Clinical studies with incomplete protocol information (eg, unspecified acupoint selection, grouping methods, or outcome indicators) or incompletely reported statistical results.
Screening and Data Extraction
Two independent researchers (WYX and WYN) screened the retrieved literature to determine the inclusion or exclusion of each article. In case of discrepancies during the screening process, the two researchers first conducted discussions; the final decision on whether to include or exclude a study was made by a third researcher (JYY). The two researchers performed screening independently and conducted two rounds of screening to ensure accuracy.
Preliminary screening was carried out by reviewing titles and abstracts, during which studies focusing on migraine secondary to other disorders were excluded. The second round of screening was completed by reading the full texts of the remaining articles. After screening, the research team developed a standardized Data Extraction Form (DEF).
The data extraction items included: (1) basic information (authors, country, region, publication year, journal name); (2) methodological quality (randomization method); (3) study characteristics (sample size, mean age of participants, diagnostic criteria, follow-up status); (4) details of acupuncture interventions (acupoints, meridians, treatment course, number of intervention sessions); and (5) outcome indicators and adverse reactions. Two evaluators (WYX and WYN) extracted data independently; any disagreements were resolved through consultation with a third evaluator (JYY). Missing data were documented and excluded from quantitative synthesis.
Subsequently, the entire research team discussed the results to address any controversial issues. Finally, the first author repeatedly reviewed and analyzed the included studies to confirm the final outcomes, including the synthesis and descriptive analysis of study findings. The flow diagram for study selection is shown in Figure 1.
Thereafter, data management software (EndNote 20.6, Microsoft Excel 2024) and graphing software (Chiplot, Review Manager 5.3, Tableau Public 2025.1, IBM SPSS Modeler 18.0, Adobe Illustrator) were used to analyze the results and generate figures. Figure 1PRISMA-ScR flow diagram for study selection.Flowchart sequence: Identification of studies via databases, Records identified from, Records removed before screening, Duplicate records removed, Records screened, Records excluded, Titles or Abstracts excluded, Content unqualified, Reviews, Animal experiment, Conference paper, Doctor of Philosophy and post graduate dissertations, Reports sought for retrieval, Reports not retrieved, Reports assessed for eligibility, Reports excluded, Read full text excluded, Incomplete content, Chinese and English repetition, Studies included in review. Counts shown: Records identified from, n equals 13096; CNKI, n equals 3397; WanFang, n equals 4761; SinoMed, n equals 2017; PubMed, n equals 862; Embase, n equals 77; Cochrane Library, n equals 589; Web of Science, n equals 1393.
Duplicate records removed, n equals 7841. Records screened, n equals 5255.Flowchart sequence: Records identified, duplicates removed, screened, retrieval sought, eligibility assessed.
Methodological Quality Assessment
Two reviewers (WYX and JYY) independently assessed the methodological quality of the included randomized controlled trials (RCTs) using the Cochrane Collaboration’s Risk of Bias (RoB) assessment tool. This RoB tool covers seven key domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, completeness of outcome data, selective reporting of results, and other potential sources of bias. Each domain was rated as “low risk”, “high risk”, or “unclear risk” in accordance with standard criteria.
Results
Search Results
An initial search of seven Chinese and English databases yielded 13,096 articles; 269 RCTs (39 English and 230 Chinese) met the inclusion criteria and were included in our analysis. Detailed characteristics of included studies and intervention prescriptions are respectively provided in Appendixs B and C.
Basic Characteristics of Evidence Sources
Figure 2 illustrates the publication trends of the seven databases from their establishment through January 13, 2026, showing a steady increase in total publications. From 2000 to 2005, the number of publications was small and grew slowly; it increased gradually after 2006, peaked at 21 articles in 2017, then declined slightly, and has experienced minor fluctuations in recent years. Figure 2Annual distribution of publications.A bar graph showing yearly publication counts from 2000 to 2025.
In terms of geographical distribution, 245 studies were conducted in China. Additionally, Italy and Germany each contributed 7 studies; Iran, 4; Brazil, 2; and Turkey, Australia, Thailand, and Cuba, 1 study each. Within China, Beijing Municipality and Guangdong Province contributed the largest number of studies (Figure 3a and b).
Figure 3Geographical distribution of included RCTs by country/region (a) Globally (b) China.A) World map with selected countries shaded to indicate included study locations. China is shaded darkest. Several other countries are shaded lighter, including Italy, Germany, Iran, Brazil, Turkey, Australia, Thailand and Cuba.
B) China map divided into provincial level regions with a continuous legend labeled 0, 5, 10, 15, 20, 25. A star symbol marks Beijing Municipality. Guangdong Province is shaded at the highest end of the legend.
Several coastal and southeastern regions show higher values than many inland and western regions, which include areas shaded at the lowest end of the legend.Two maps showing included study counts globally and across China by province.
Relevant studies have been conducted across various regions of China, covering almost the entire country. The increase in the number of studies in China may be associated with funding support, economic foundation, and healthcare environment—all of which are closely associated with the development of research in this field within China.
Basic Characteristics of the Included RCTs
Types of Migraine
226 RCTs referred to Western medicine migraine classification (Figure 4). 132 studies (58.41%) were of migraine without aura, making it the most common subtype, which is closely related to its high clinical incidence and typical symptoms. 58 studies (25.66%) were of unclassified migraine, due to the heterogeneous diagnostic criteria in some early studies and the complex symptoms of some patients that are difficult to classify into a single type, which reflects a certain ambiguity in current migraine classification diagnosis.
36 studies (15.93%) examined other migraine subtypes, such as chronic migraine and acute migraine, which not only indicate the diversity of migraine etiology and manifestations in clinical practice but also guide the exploration of subtypes in subsequent studies. Figure 4Classification of Migraine in Western Medicine.A doughnut chart showing migraine classification shares from 0.88 percent to 58.41 percent.
43 RCTs identified migraine by traditional Chinese medicine (TCM) syndrome types (Figure 5). 18 studies (41.86%) were of the liver-yang hyperactivity type, consistent with the TCM theory of “liver governing dispersion and regulating qi movement”. Long-term emotional distress leads to stagnant liver qi transforming into fire, or liver-kidney yin deficiency causes liver-yang hyperactivity; both can disturb the clear orifices and trigger headaches.
This syndrome type is common in clinical practice and a key focus of TCM acupuncture and herbal treatment. 25 studies (58.14%) examined other TCM syndrome types, such as blood stasis, phlegm turbidity disturbing the upper type, and qi-blood deficiency. Figure 5Classification of Migraine in Traditional Chinese Medicine (TCM).A doughnut chart with nine Traditional Chinese Medicine syndrome types and percentage labels.
Percent by syndrome type: Liver-Yang Hyperactivity Syndrome, 41.86 percent; Blood Stasis Syndrome, 13.95 percent; Phlegm Turbidity Disturbing Upper Jiao Syndrome, 9.30 percent; Qi and Blood Deficiency Syndrome, 9.30 percent; Wind Pathogen Invading Collaterals Syndrome, 6.98 percent; Kidney Deficiency Syndrome, 6.98 percent; Blood Stasis Obstructing Brain Collaterals Syndrome, 4.65 percent; Shaoyang Meridian Syndrome, 4.65 percent; Liver-Kidney Yin Deficiency Syndrome, 2.33 percent. A legend lists these nine syndrome types.A doughnut chart showing Traditional Chinese Medicine migraine syndrome types, from 2.33 percent to 41.86 percent.
Diagnostic Criteria for Migraine
A total of 238 articles reported the application of diagnostic criteria (Table 1). Note that some studies employed more than one diagnostic criterion, yielding a total of 285 criterion instances across these articles. The International Classification of Headache Disorders (ICHD) being the most widely used, appearing in 191 studies.
Among these 191 studies, ICHD-2 was the most frequently used version (n=84, 29.47%), reflecting a discrepancy in which updated ICHD versions were not applied in the included literature despite their release.Table 1Diagnostic Criteria for Migraine (n=285)CharacteristicsStudies No.(%)ICHD-I (1988)5117.89%ICHD-II (2004)8429.47%ICHD-3β (2013)186.32%ICHD-III (2018)269.12%ICHD (without specified edition)124.21%Diagnostic and Efficacy Criteria of TCM Syndromes and Diseases175.96%Guiding Principles for Clinical Research of New Chinese Medicines165.61%Chinese Guidelines for the Diagnosis and Treatment of Migraine155.26%Neurology93.16%Diagnostic and Efficacy Evaluation Criteria for Toufeng62.11%Diagnostic Criteria of Syndromes for Toufeng Disease51.75%Other Diagnostic Criteria (number of uses ≤ 3)269.12%
Among Chinese domestic diagnostic criteria, 17 studies (5.96%) adopted the Diagnostic and Efficacy Criteria of TCM Syndromes and Diseases, underscoring the unique value of TCM in the diagnosis and management of migraine. Additionally, 16 studies (5.61%) applied the Guiding Principles for Clinical Research of New Chinese Medicines, demonstrating the integration of international research experience with local normative standards in domestic studies. Only 26 studies (9.12%) employed alternative diagnostic criteria, suggesting some variation in the application of diagnostic standards for migraine.
Future research should prioritize enhancing the integration and mutual complementarity of different diagnostic criteria to improve the accuracy and consistency of migraine diagnosis.
Participants
Among the 269 included RCTs, the majority (n=164, 60.97%) enrolled 51–100 participants, whereas only 75 studies had more than 100 participants (Table 2). This distribution indicates a scarcity of large-sample, multi-center studies, highlighting the need for larger, more robust trials to enhance the generalizability of findings.Table 2Characteristics of the Included Clinical StudiesCharacteristicsStudies No.(%)Study sample size (n=269)n≤503011.15%50<n≤10016460.97%100<n≤3007126.39%n>30041.49%Mean age of Participants (n=222)11–20 years old10.45%21–30 years old2611.71%31–40 years old12455.86%41–50 years old6529.28%51–60 years old41.80%61–70 years old20.90%
Of the 269 included studies, 222 studies reported participant age data, whereas the remaining 47 studies did not specify age information. Among the 222 studies reporting age data, the 31–40 years age group (n=124, 55.86%) was the most frequently represented, followed by the 41–50 years age group (n=65, 29.28%). Together, these middle-aged cohorts accounted for 85.14% of all age-stratified studies, consistent with the epidemiological profile of migraine, in which incidence peaks in mid-adulthood.
By contrast, younger (≤30 years) and older (≥51 years) age groups were substantially underrepresented, suggesting a critical gap in understanding acupuncture efficacy across the full age spectrum of migraine patients.
Acupuncture Interventions
Among 233 studies, 180 studies (77.25%) employed less than 25 acupuncture sessions, reflecting a preference for short-term, intensive interventions to rapidly relieve migraine symptoms (Table 3).Table 3Intervention Characteristics of the Included Clinical StudiesCharacteristicsStudies No.(%)Number of acupuncture treatment (n=233)≤2518077.25%25 <n ≤ 505121.89%>5020.86%Intervention duration (n=243)≤ 4 Weeks18375.31%4 to 8 Weeks3514.40%≥ 8 Weeks2510.29%Follow-up Period (n=153)≤1 month106.53%1 month2918.95%1.5 months21.30%2 months95.88%3 months4227.45%4 months21.30%6 months4730.72%≥ 1 year127.84%
For intervention duration (n=243), 183 studies (75.31%) had durations <4 weeks, the most common. This is due to prominent symptoms during migraine acute attacks, short-cycle interventions enabling rapid observation of symptom improvement, and a simple study design.
In follow-up assessments (n=153), 30.72% and 27.45% of studies chose 6-month and 3-month endpoints, respectively, with only 7.84% extending follow-up to one year or longer. This demonstrates a focus on short- to medium-term efficacy, underscoring the need for longer-term studies to evaluate the durability of acupuncture effects.
Frequency Analysis of Acupuncture Prescription Components
Data on acupoints, meridians, body regions, and specific points were extracted from 269 RCTs, which included a total of 172 acupoints and tender points (Table 4). The most frequently used acupoints were Fengchi (GB20), Shuaigu (GB8), Taichong (LR3), Taiyang (EX-HN5), Baihui (GV20), Waiguan (TE5), Hegu (LI4), Touwei (ST8), Yanglingquan (GB34), and Zulinqi (GB41).Table 4Frequency of Acupoint Use, Meridian Use and Body Regions UseAcupointsNumber of ArticlesMeridiansTotal NumberBody RegionsTotal NumberGB20159GB538Caput465GB8108TE176Feet/Ankle289LR3105ST149Crus187EX-HN5100GV129Collum185GV2087LR129Arm151TE581LI93Hand131LI480SP81Thigh28ST863BL69Abdomen42GB3449KI47Back38GB4147CV40Chest4SP638PC31Shoulder2KI337LU27Ashi Point36SI12GB4036HT6ST4035ST3634TE327TE2026SP1025TE2323
The top 10 most frequently used meridians are the Gallbladder Meridian (GB), Triple Energizer Meridian (TE), Stomach Meridian (ST), Governor Vessel (GV), Liver Meridian (LR), Large Intestine Meridian (LI), Spleen Meridian (SP), Bladder Meridian (BL), Kidney Meridian (KI), and Conception Vessel (CV).
The caput, feet/ankle, crus, collum, arm and hand are commonly used regions.
Analysis of Prescription Association Rules
The collected data were imported into IBM SPSS Modeller 18.0 software, and association analysis was conducted using the Apriori algorithm. Here, the minimum support was set to 0.1 (ie, the combination appeared in ≥10% of the studies), and the minimum confidence was set to 0.7 (ie, when acupoint A was used, the probability of acupoint B appearing simultaneously was ≥70%).
For the association rule network diagram presented in Figure 6a–c, an association was defined as valid if the support degree reached or exceeded 0.1. Additionally, a positive correlation was identified between line thickness and confidence level, where thicker lines corresponded to higher confidence levels (≥0.7). For meridians, the correlation between GB and TE is most prominent, indicating a higher likelihood of combined application than other meridians.
Similarly, at the acupoint level, the correlation between Fengchi (GB20) and Shuaigu (GB8) is strongest with the highest co-occurrence frequency. The Apriori algorithm intuitively identifies relationships between acupoints, meridians, and body regions. Figure 6Overview of acupuncture intervention points.
(a) Association network of acupoints for migraine. (b) Association network of meridians for migraine. (c) Association network of body regions for migraine.
Data source: 269 RCTs included in this study.A) Network diagram with labeled nodes arranged around a circle and connected by many straight lines. Labels around the perimeter read, GV20, GB8, GB41, GB40, GB34, GB20, EX -HN5, TE5, TE3, TE23, TE20, ST8, ST40, ST36, SP6, SP10, LR3, LI4, KI3. Many lines connect between nodes, with several thicker connections converging near GB20 and GB8.
B) Network diagram with labeled nodes connected by straight lines. Labels around the perimeter read, KI, HT, GV, GB, CV, BL, TE, ST, SP, SI, PC, LU, LR, LI. Many lines connect across the network, with a notably thicker connection between GB and TE.
C) Network diagram with labeled nodes connected by straight lines. Labels around the perimeter read, Caput, Back, Arm, Abdomen, Thigh, Shoulder, Hand, Feet /Ankle, Crus, Collum, Chest. Many lines connect across the network, with several thicker connections converging near Caput and Collum.A diagram showing three association networks for migraine, acupoints, meridians and body regions used in migraine treatment.
By summarizing prescription information and integrating high-frequency analysis with correlation analysis, the core acupuncture prescriptions for migraine treatment include basic acupoint combinations such as Fengchi (GB20)–Shuaigu (GB8) (n=92), Fengchi (GB20)–Taiyang (EX-HN5) (n=83), Fengchi (GB20)–Taichong (LR3) (n=77), Fengchi (GB20)-Baihui (GV20) (n=72). Additionally, the core meridians are GB–TE (n=109) and GB–LR (n=102). As for body regions, the head (caput)–neck(collum) (n=150) is the most frequently combined region, followed by head (caput)–feet/ankle (n=135).
These patterns describe commonly investigated acupuncture approaches and should not be interpreted as evidence of comparative efficacy, optimal point selection or validated biological mechanisms.
Adverse Events
Adverse events were reported in 27 of the included studies, with a total of 188 adverse events documented across the cohort. Among the 188 adverse events, pain (n=50, 26.60%), bleeding (n=35, 18.62%), and subcutaneous hematoma (n=31, 16.49%) were the most frequently reported events (Table 5). These reactions are primarily attributed to variability in needling technique, individual patient sensitivity, and minor iatrogenic tissue trauma.Table 5Adverse EventsType of Adverse ReactionNumber of Patients (n=188)Proportion (%)Pain5026.60%Bleeding3518.62%Subcutaneous hematoma3116.49%Mild adverse reactions168.51%Swelling126.38%Local ecchymosis115.85%Sweating94.79%Fainting during acupuncture73.72%Dizziness52.66%Gastrointestinal discomfort42.12%Nausea31.60%Insomnia21.06%Mild muscle spasm10.53%Leg weakness10.53%Lethargy10.53%
Mild systemic adverse reactions (n=16, 8.51%) and local inflammatory responses, including swelling (n=12, 6.38%) and ecchymosis (n=11, 5.85%), were also observed, alongside less common events such as sweating (n=9, 4.79%) and acupuncture-related fainting (n=7, 3.72%).
Rare adverse events (≤5 cases) included dizziness, gastrointestinal discomfort, nausea, insomnia, and transient neuromuscular symptoms, which resolved spontaneously without intervention.
All reported adverse events were transient and reversible, with no severe complications observed. However, given inconsistent reporting and limited follow-up, these findings should be interpreted cautiously. Standardized adverse-event reporting is recommended for future trials.
Comparative Analyses and Statistical Investigations of Acupuncture-Related Aspects
Comparisons of Acupuncture and Non-Acupuncture Therapies
Among comparisons between acupuncture and non-acupuncture therapies, acupuncture versus pharmacotherapy accounts for the largest proportion (Figure 7), indicating that direct comparisons are a primary research focus. Other subgroups are relatively small: modified acupuncture versus pharmacotherapy, while acupuncture combined with pharmacotherapy against pharmacotherapy remain underrepresented. Figure 7Comparisons of acupuncture and non-acupuncture therapies.A doughnut chart showing proportions of acupuncture versus non-acupuncture therapy comparisons, 1.90 percent to 69.52 percent.Abbreviations: Acup, Conventional acupuncture therapy; Sham-Acup, Sham acupuncture therapies; M-Acup, Modified acupuncture therapy; CM, Conventional medicine; TCM, Traditional Chinese medicine.
Comparison Between Special Acupuncture Techniques and Conventional Acupuncture Techniques
Among various characteristic acupuncture techniques, penetrating needling accounts for 28.85%, close to one-third of the total, and dominates in application, far exceeding other techniques (Figure 8). This technique enables simultaneous stimulation of multiple acupoints along involved meridians, enhancing de qi sensation and potential therapeutic efficacy for migraine involving complex meridian pathways. Mind regulating needling ranks second, with a proportion of 25.00%.
Drawing on the TCM perspective that “the brain houses the spirit,” it is traditionally intended to address the emotional and psychological dimensions of migraine, factors increasingly recognized as important triggers and modulators of attacks. The remaining 11 techniques, such as contralateral needling, deep needling and balance acupuncture are applied at low frequencies and are primarily used as supplementary options in specific scenarios. Figure 8Special Acupuncture Techniques.A doughnut chart showing the percentage distribution of 13 specialized needling techniques across the included RCTs.
Sorted from most to least frequent: Penetrating Needling 28.85 percent, Mind Regulating Needling 25.00 percent, Contralateral Needling 9.62 percent, Deep Needling 5.77 percent, Balance Acupuncture 5.77 percent, Chifeng Yingyuan Needling 5.77 percent, Longhu Jiaozhan Needling 3.85 percent, Warming-dredging Needling 3.85 percent, Linggui BaFa 3.85 percent, Jieqi Tongjing Needling 1.92 percent, Ziwu Liuzhu Needling 1.92 percent, Tongxuan Needling 1.92 percent, and Canggui Tanxue Needling 1.92 percent. The top two techniques, Penetrating Needling and Mind Regulating Needling, together account for 53.85 percent of the total.A doughnut chart showing special acupuncture techniques by percent, ranging from 1.92 percent to 28.85 percent.
Comparison of Different Acupuncture Factors
Among studies comparing acupoint-related factors in migraine acupuncture, 24 focused on the influence of meridians. Of these, 22 investigated Shaoyang Meridian acupoints (Foot-Shaoyang Gallbladder Meridian, Hand-Shaoyang Triple Energizer Meridian) vs general or non-meridian, non-specific acupoints, while only 2 involved Jueyin Meridian acupoints. Evidently, Shaoyang Meridian acupoints are the primary focus of the study.
This finding is consistent with previous statistics indicating that the Foot-Shaoyang Gallbladder and Hand-Shaoyang Triple Energizer Meridians are most frequently used, reflecting the clinical guiding significance of TCM theory: “meridians treat disorders in the regions they traverse”. Migraine is largely associated with the Shaoyang Meridian pathway (lateral head), and its acupoints more precisely resolve head meridian stagnation and regulate qi and blood, thereby making them a preferred choice for controlled studies. In contrast, studies on Jueyin Meridian, general, or non-meridian, non-specific acupoints are relatively scarce.
This is partly because the Jueyin Meridian distribution has lower compatibility with common migraine tender points, and partly because current research tends to verify the targeted analgesic advantages of meridian-specific acupoints by comparing Shaoyang Meridian acupoints with non-specific ones.
Statistical Analysis of the Proportion of Each Acupuncture Method in RCTs
Among acupuncture techniques for migraine, diversity and uneven distribution are notable (Figure 9). Conventional acupuncture accounts for 61.33%, the most mainstream therapy. Widely used clinically due to its simple operation and precise acupoint selection based on meridian theory, it alleviates migraine by stimulating specific acupoints to dredge meridians and regulate qi and blood.
Electroacupuncture (EA) accounts for 12.89%. Integrating electrical stimulation with conventional acupuncture, EA enhances stimulation intensity and duration, regulates neurological function more efficiently, and improves analgesia. Other therapies, including blood-letting therapy (5.78%), scalp acupuncture (4.44%), auricular acupuncture (2.22%), acupoint injection (1.78%), moxibustion (1.78%), fire needling (0.89%), and small needle knife (0.89%), account for relatively low proportions.
Blood-letting resolves stasis by releasing a small amount of blood; scalp acupuncture targets cortical functional areas; acupoint injection combines meridian stimulation with pharmacological effects. However, due to complex operation and narrow indications, their clinical application is more limited than conventional acupuncture and EA. Nevertheless, they provide diversified options for acupuncture-based migraine treatment, selectable based on patients’ conditions and clinicians’ experience.
Figure 9Acupuncture Methods.A doughnut chart showing the percentage distribution of 21 acupuncture modalities used for migraine across the included RCTs. Sorted from most to least frequent: Conventional Acupuncture 61.33 percent, Electroacupuncture 12.89 percent, Blood-letting Therapy 5.78 percent, Scalp Acupuncture 4.44 percent, Auricular Acupuncture 2.22 percent, Acupoint Injection 1.78 percent, Moxibustion 1.78 percent, Auricular Acupressure 1.33 percent, Embedded Needling 1.33 percent, Plum-blossom Needle 0.89 percent, Fire Needling 0.89 percent, Small Needle Knife 0.89 percent, Forearm Skin Acupuncture 0.89 percent, Auricular Comprehensive Therapy 0.44 percent, Cheek Acupuncture 0.44 percent, Eye Acupuncture 0.44 percent, Internal Heat Needling 0.44 percent, Tongue Acupuncture 0.44 percent, Hand Four Needles 0.44 percent, Abdominal Acupuncture 0.44 percent, and Pestle Needle 0.44 percent. Conventional Acupuncture alone accounts for over 60 percent of all modalities, with Electroacupuncture a distant second at 12.89 percent.A donut chart showing migraine acupuncture method shares from 0.44 percent to 61.33 percent.
Outcome Measures
The outcome measures were categorized into four main types (Figure 10): Figure 10Outcome measures.A horizontal bar graph with title, Outcome measures. X axis tick labels 0 to 180 in steps of 20. Y axis categories: Total Effective Rate, Effective Rate, Incidence of Adverse Reactions, Cure Rate, Ineffective Rate, Recurrence Rate, VAS Score, Migraine Syndrome Score, Improvement Score of Accompanying Symptoms, MSQ Score, Traditional Chinese Medicine Syndrome Score, Headache Index, Anxiety/Depression Score, Number of Headache Episodes, Number of Headache Days, Frequency of Headache Episodes, Cerebral Blood Flow Velocity, Dosage of Analgesic Drugs Used, Serum CGRP Level.A horizontal bar graph showing values for outcome measures across multiple clinical and symptom score categories.
For objective measures of migraine attack frequency and duration, migraine duration (70 times) and number of attacks (69 times) were most used, directly quantifying single-episode duration and cycle attack frequency to provide objective migraine pattern data.
For comprehensive assessment, VAS scores (114 times) were most used to assess dynamic pain intensity, followed by migraine severity scores (32 times) and Migraine-Specific Quality of Life Questionnaire (MSQ) scores (24 times) for symptom severity and quality-of-life evaluation, respectively.
For core measures assessing clinical efficacy and safety, the total effective rate (179 times) was the most used indicator of overall efficacy, supplemented by the effective rate (75 times) to simplify efficacy presentation and adverse reaction incidence (61 times) to assess treatment safety.
For measures related to physiological mechanisms and adjunctive interventions, cerebral blood flow velocity (17 times) and analgesic dosage (14 times) were relatively common—the former assessing cerebrovascular regulatory mechanisms and the latter reflecting reduced medication dependence.
Quality Assessment of the Included Studies
The methodological quality of the 269 included randomized controlled trials was evaluated using the Cochrane Risk of Bias tool (Review Manager 5.3) and visualized (Figure 11). Seven key bias domains were assessed, including random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. The visualization results showed that the quality of included studies was generally suboptimal: a considerable proportion of studies had unclear risk of bias in allocation concealment and blinding (due to the inherent characteristics of acupuncture that make complete blinding difficult), while partial studies had high risk in random sequence generation and selective reporting.
Only a small number of studies achieved low risk of bias in all seven domains, indicating deficiencies in the methodological design of most existing clinical studies on acupuncture for migraine. Figure 11Cochrane bias risk analysis diagram.A stacked horizontal bar graph showing Cochrane Risk of Bias domain ratings across trials.
Discussion
Clinical Context of Acupuncture for Migraine Prevention
The clinical positioning of acupuncture should be informed by external efficacy reviews and guidelines rather than the descriptive findings of this scoping review. The 2016 Cochrane review12 (22 trials, 4,985 participants) found reduced headache frequency versus no acupuncture and a small benefit over sham acupuncture, with effectiveness at least comparable to prophylactic drugs but fewer adverse effects; overall evidence was moderate quality. Recent 2025 meta-analyses16 reinforce a nuanced view: acupuncture may serve as an adjunctive or alternative option when conventional prophylaxis is unsuitable, but evidence does not support replacing established pharmacological prevention.
Under NICE CG150, amended in June 2025,17 acupuncture is a preventive option when propranolol, topiramate and amitriptyline are ineffective, poorly tolerated or unsuitable on safety grounds. The suggested course comprises no more than ten sessions across five to eight weeks, with treatment decisions informed by individual preferences and clinical circumstances. A 2025 systematic review of 25 international headache guidelines18 found that, among the 18 guidelines recommending acupuncture, 77.8% lacked procedural details, only 22.2% specified treatment frequency, and 16.7% described needling techniques; methodological quality was suboptimal, with the applicability domain scoring lowest (20.3%).
These sources support a context-specific discussion of acupuncture rather than a universal claim of therapeutic superiority.
The present review contributes a map of research characteristics, intervention patterns and reporting gaps; it does not independently estimate comparative efficacy or safety. Frequently investigated acupoints should therefore be interpreted as patterns of published research rather than validated optimal prescriptions. Furthermore, because combined interventions were excluded, our findings cannot establish the additional benefit of acupuncture when used alongside pharmacotherapy.
The principal implications are the need for more transparent trial methods, reproducible intervention reporting and better harmonization of outcomes.
Global and Regional Disparities in Acupuncture Research for Migraine
From a geographical perspective, acupuncture-for-migraine research shows marked global imbalance: of 269 included studies, 245 originated from China, whereas Germany, Italy, and Iran contributed 7, 7, and 4 studies, respectively. This concentration reflects acupuncture’s cultural origins in China, the maturity of its TCM clinical system, and sustained policy support.19
Geographical concentration also raises concerns beyond generalizability. Methodological research has long documented country-related patterns in published results: Vickers et al found that all acupuncture trials from China, Japan, Hong Kong, and Taiwan reported positive findings, suggesting disproportionate non-publication of negative trials in these regions.20 A related but distinct issue is selective outcome reporting—meta-epidemiological studies of acupuncture RCTs have found that only a minority provide valid prospective registrations, with discrepancies between registered and published outcomes common.21 Geographical concentration alone does not formally demonstrate publication bias, and this review did not quantify it. Nevertheless, unpublished negative findings and selective outcome reporting cannot be excluded, limiting interpretation of the mapped evidence.
In Western countries, alternative needling modalities—most notably trigger point dry needling—have been developed within anatomical and neurophysiological frameworks rather than traditional meridian theory, and are predominantly delivered by physiotherapists.22,23 Because dry needling does not incorporate acupoint selection or TCM diagnostic reasoning, it falls outside the operational definition of acupuncture in this review; its widespread Western adoption may partially account for the relative paucity of traditional acupuncture studies from these regions.
Methodological Quality Assessment of RCTs
The included trials span 2000–2025, during which migraine diagnostic standards evolved from ICHD-1 and ICHD-2 toward ICHD-3. Trials published from 2015 onward predominantly adopted ICHD-3 or ICHD-3β (32/132, 24.2%), indicating greater diagnostic homogeneity in later trials. Reporting of adverse events (13.6% vs 6.6%) and intervention duration (96.2% vs 84.7%) also improved, consistent with broader CONSORT and STRICTA adoption.
Nevertheless, the overall methodological quality of the included RCTs remains deficient, undermining the evidence base. First, reporting of randomization and allocation concealment is inadequate: few studies explicitly described random sequence generation, and most did not report allocation concealment, raising selection bias risk. Second, blinding is challenging and insufficient: practitioner blinding is inherently infeasible because therapists must know acupoint locations and needling techniques;24,25 participant blinding is compromised by the invasive nature of needling and the distinct “de qi” sensation (aching, heaviness, or distension), which may enable participants to identify group allocation.24 Third, sample sizes and follow-up are limited: in the 269 RCTs, 60.97% of studies enrolled 50–100 participants and only 1.49% exceeded 300, restricting external validity; 43.12% did not report follow-up duration and only 4.46% included 1-year follow-up, precluding assessment of long-term efficacy and recurrence prevention.
Witt et al noted that most acupuncture studies lack long-term follow-up data—only 12% exceed 1 year—and safety reports are mostly limited to short-term adverse events, with insufficient monitoring of rare events such as pneumothorax and infection.26
Patterns of Acupoint Selection and Their Interpretation
In TCM theory, migraine is traditionally categorized as toufeng, naofeng and toutong.27 According to Suwen, its pathogenesis is traditionally attributed to external pathogenic wind disrupting meridian harmony in the head, which is thought to underpin the acupuncture principle of “dredging meridians and regulating qi to harmonize blood” and the concept of treating upper disorders from lower regions.
Among the included studies, Fengchi (GB20) and Shuaigu (GB8) were the most frequently selected acupoints. Both belong to the Foot-Shaoyang Gallbladder Meridian (GB). Electroacupuncture at Fengchi has been reported to reduce CGRP expression and positive cell counts in the trigeminal ganglion, suggesting a potential anti-inflammatory and analgesic mechanism;28 Shuaigu, the convergence point of the Foot-Taiyang Bladder and Foot-Shaoyang Gallbladder meridians, has been associated with enhanced functional connectivity between the ACC and inferior frontal gyrus.29
The predominance of GB meridian acupoints aligns with traditional acupoint selection rules30 and the TCM concept of “Shaoyang Governing the Pivot”.31 In TCM theory, the Shaoyang Meridian is classified as half-exterior and half-interior and is believed to regulate qi ascent and descent; pivot dysfunction is traditionally thought to cause upward counterflow of stagnant heat, manifesting as distending headache and eye pain.32 Shaoyang meridian acupuncture has also been associated with inhibition of connexin 43 activation in astrocytes of the trigeminal nucleus caudalis and reduced pro-inflammatory factor release,33 as well as enhanced prefrontal functional connectivity.10
Furthermore, the Hand-Shaoyang and Foot-Shaoyang meridians course along the lateral head, corresponding to typical migraine pain sites, which is consistent with the TCM principle of treating disorders along the meridian pathway.
The TCM concept of “treating upper disorders from lower regions” is traditionally applied to migraine, as stated in Lingshu: “For diseases in the head, select acupoints on the feet.” Taichong (LR3) exemplifies this approach, with electroacupuncture at LR3 reported to alleviate migraine via inhibition of CGRP and BDNF expression.34 However, such distal acupoints remain underutilized as auxiliary points, likely due to clinical preference for local needling.
Current acupuncture research on migraine exhibits path dependence in acupoint and meridian selection, with limited systematic analysis of synergistic effects across combined meridians. Future studies should expand meridian selection dimensions, integrate TCM syndrome differentiation with neuroregulatory mechanisms, and explore targeted effects of different acupoint–meridian combinations on migraine subtypes.
Heterogeneity in Outcome Measures
The included studies employed diverse outcome measures in acupuncture research for migraine, without unified standards and with substantial heterogeneity. The “total effective rate” is widely used but its ambiguous criteria fail to objectively reflect therapeutic effects, increasing reporting bias risk. VAS scores are most commonly used for pain intensity,35 but included studies varied greatly in assessment time points and recording methods, hindering standardized comparisons.
Headache duration and attack frequency are also commonly used, but their definitions are inconsistent and some studies lack clear data collection protocols.
Patient-centered outcome reporting is insufficient: only 24 studies used the IHS-recommended MSQ. Additionally, two key indicators—analgesic use (for evaluating preventive effects) and serum CGRP levels (for mechanistic exploration)—were rarely included. Diverse assessment scales further reduce cross-study comparability,36 and the systematic inclusion of acupuncture-specific indicators is lacking, limiting mechanistic research depth.
To address this heterogeneity, we propose candidate outcome domains and measures for future consensus development (Table 6), informed by the present review findings.Table 6Candidate Outcome Domains and Measures for Future Consensus DevelopmentIndicator TypeSpecific IndicatorStandardized Measurement ProtocolSource of BasisObjective Efficacy IndicatorsMigraine Attack FrequencyUse an electronic headache diary to record the number of migraine attacks continuously for 4 weeksInternational Headache Society (IHS) The International Classification of Headache Disorders, 3rd edition (DOI: 10.1177/0333102417738202); Chinese Guidelines for the Diagnosis and Treatment of Migraine (2022 Edition) (DOI: 10.3969/j.issn.1006-9852.2022.12.001)Migraine DurationFor each attack, record the “time from onset to complete relief of symptoms (unit: hours)” and continue recording for 4 weeksInternational Headache Society (IHS) The International Classification of Headache Disorders, 3rd edition (DOI: 10.1177/0333102417738202); Chinese Guidelines for the Diagnosis and Treatment of Migraine (2022 Edition) (DOI: 10.3969/j.issn.1006-9852.2022.12.001)Migraine DaysContinuously record the number of days with migraine symptoms within 4 weeks (a day is counted if there is an attack on that day)Chinese Guidelines for the Diagnosis and Treatment of Migraine (2022 Edition) (DOI: 10.3969/j.issn.1006-9852.2022.12.001); Clinical Guideline on Acupuncture and Moxibustion—Migraine (WFAS 007.8-2023)Total Number of Migraine AttacksAccurately record the total number of independent attacks that meet the migraine diagnosis criteria within 4 weeks using a paper or electronic diaryClinical Guideline on Acupuncture and Moxibustion—Migraine (WFAS 007.8-2023)Analgesic ConsumptionRecord the number of uses and cumulative dose of migraine-relieving drugs (triptans, NSAIDs, etc.) within 4 weeks, and calculate the average daily consumptionChinese Practice Guidelines for Diagnosis and Treatment of Migraine (1st Edition, Chinese Society of Neurology) (DOI: 10.3760/cma.j.cn113694-20230222-00118); International Headache Society Global Practice Recommendations for the Acute Pharmacological Treatment of Migraine (DOI: 10.1177/03331024241252666)Subjective Symptom AssessmentVisual Analog Scale (VAS)Adopt a 0–10 point scale (0 = no pain, 10 = most severe pain) to assess headache severity before treatment, 30 minutes after each treatment, and 4 weeks after treatmentExpert Consensus on the Application of Pain Evaluation Questionnaires in China (2020) (DOI: 10.3760/cma.j.cn101379-20190915-00075)Migraine Score (Symptom Score)Calculate the composite score using “attack frequency × average duration × average VAS score”, and calculate once before treatment and once 4 weeks after treatmentChinese Guidelines for the Diagnosis and Treatment of Migraine (Patient Version) (DOI: 10.3969/j.issn.1006-9852.2025.05.001)Improvement Rate of Accompanying Symptoms (nausea, vomiting, photophobia, phonophobia, etc.)Adopt a 4-point rating scale (none, mild, moderate, severe) to assess accompanying symptoms, calculate the improvement ratio after treatment compared with before treatment, and assess once before treatment and once 4 weeks after treatmentChinese Practice Guidelines for Diagnosis and Treatment of Migraine (1st Edition, Chinese Society of Neurology) (DOI: 10.3760/cma.j.cn113694-20230222-00118)TCM Syndrome ScoreScore the main symptoms (headache) and secondary symptoms (dizziness, irritability, tongue coating, pulse condition, etc.) of “migraine (headache disorder)” in accordance with the Criteria for the Diagnosis and Efficacy Evaluation of TCM Diseases and Syndromes, and assess once before treatment and once 4 weeks after treatmentGeneral Revision and Development Principle for Specification of Diagnosis and Therapeutic Effect Evaluation of Diseases and Syndromes in Traditional Chinese Medicine (ZY/T 10-2024)Anxiety/Depression Score (GAD-7, PHQ-9)Use the Generalized Anxiety Disorder Scale (GAD-7) to assess anxiety and the Patient Health Questionnaire Depression Scale (PHQ-9) to assess depression, and assess once before treatment and once 4 weeks after treatmentChinese Expert Consensus on the Diagnosis and Treatment of Comorbid Migraine and Depressive Disorder (DOI: 10.12114/j.issn.1007-9572.2025.0216Patient-Reported Outcomes (PROs)Migraine Disability Assessment Scale (MIDAS)Assess headache-related disability over the past 3 months using the 5-item MIDAS questionnaire. Disability is graded as: Grade I (0–5, minimal), Grade II (6–10, mild), Grade III (11–20, moderate), Grade IV (≥21, severe).Migraine Disability Assessment (MIDAS) QuestionnaireMigraine-Specific Quality of Life Questionnaire (MSQ)Assess migraine-specific quality of life over the past 4 weeks using the MSQ version 2.1, a 14-item questionnaire with three domains: Role Function-Restrictive (RFR, 7 items); Role Function-Preventive (RFP, 4 items); Emotional Function (EF, 3 items); domain scores are rescaled to 0–100, with higher scores indicating better quality of life.Migraine-Specific Quality of Life Questionnaire Chinese Version 2.1TCM-Based Patient-Reported Outcome Scale for MigraineBased on TCM theory, assess dimensions such as pain, life interference, mood, and TCM symptoms, and assess once before treatment and once 4 weeks after treatmentGeneral Revision and Development Principle for Specification of Diagnosis and Therapeutic Effect Evaluation of Diseases and Syndromes in Traditional Chinese Medicine (ZY/T 10-2024)Safety IndicatorsIncidence of Adverse ReactionsRecord adverse reactions by category in accordance with the STRICTA Guidelines: ①Local (pain/bleeding/hematoma/infection at the acupuncture site, etc.); ② Systemic (nausea/palpitation/allergies, etc.); follow up for 6 months to monitor the recurrence rateSTRICTA 2010 Statement and 2015 Update; Record specification on adverse even in clinical trials of acupuncture and moxibustion (T/CAAM 0002-2022). Handling methods for acupuncture abnormal conditions (GB/T33415-2016)Incidence of Acupuncture SyncopeRecord the number of cases with acupuncture syncope manifestations such as “dizziness, vertigo, nausea, pale complexion, sweating, cold limbs” during treatment, and calculate the incidenceRecord Specification on Adverse Events in Clinical Trials of Acupuncture and Moxibustion (T/CAAM 0002-2022);Incidence of Local HematomaRecord the number of cases with subcutaneous hematoma (diameter ≥1 cm) at the acupuncture site, and calculate the incidenceSpecification of Formulation and Evaluation for the Clinical Practice Guideline of Acupuncture and Moxibustion (GB/T 40972-2021)
We emphasize that this is a preliminary proposal generated from scoping review evidence, rather than a formally validated or consensus-endorsed Core Outcome Set (COS). It is intended to inform subsequent formal development rather than serve as a finalized recommendation. Future formal development should follow COMET methodology, including interdisciplinary team establishment, systematic literature review, Delphi consensus rounds, and a consensus meeting with anonymous voting, followed by publication and a 3–5-year dynamic update mechanism.
The proposed preliminary COS comprises four categories: objective indicators, subjective indicators, patient-reported outcomes and safety indicators. If validated through formal consensus processes in future work, this preliminary COS has the potential to enhance the comparability and clinical relevance of evidence in acupuncture research for migraine.
Limitations
This scoping review has certain limitations. First, only published literature in Chinese and English databases was searched, excluding grey literature such as theses and conference abstracts, which may omit potential evidence and compromise the completeness of the overall evidence. Second, as a medical technique with distinct Chinese characteristics, acupuncture is widely used and accepted in China, resulting in most included studies being conducted domestically and insufficient geographical representativeness.
Third, to assess the independent effect of acupuncture, the review included only single-acupuncture interventions and excluded acupuncture combined with medication. Although this facilitates exploration of acupuncture’s independent mechanisms and efficacy, such combined regimens are common in clinical practice, thereby limiting the applicability of the conclusions. Future studies should supplement evidence on acupuncture combined with medication to better meet clinical needs.
Conclusion
This scoping review maps the clinical research landscape of acupuncture for migraine across 269 RCTs, predominantly from China. The evidence reveals a consistent acupoint selection pattern—Fengchi (GB20), Shuaigu (GB8), and Taichong (LR3) are most frequently used, primarily along the Foot-Shaoyang Gallbladder Meridian—aligning with the traditional TCM concept of “Shaoyang as the pivot” and trigeminal nerve distribution, though the clinical significance of these associations remains to be established.
Simultaneously, this study identified common methodological issues in current research, classified outcome measures, clarified sources of heterogeneity, and linked TCM theory to clinical evidence, thereby guiding acupuncture prescription standardization and future research.
Nevertheless, the evidence base remains constrained by geographical concentration, small samples, short follow-ups, and suboptimal methodological reporting—particularly in randomization, allocation concealment, and blinding. The absence of a harmonized core outcome set further limits cross-study comparability.
Future research should prioritize multi-center, adequately powered RCTs with standardized protocols and transparent reporting, alongside formal consensus development of a core outcome set. Strengthening the evidence base in these areas will facilitate acupuncture’s integration into international migraine management guidelines.