Literature screening process

A total of 18576 studies were retrieved after screening the databases, of which 5967 duplicate articles were excluded. After reviewing the title and abstract, 11705 studies were excluded (irrelevant topics, editorial or letters, review protocols, case reports, and non-English studies). The remaining 904 studies were subjected to full-text screening to exclude the studies without adverse effects reactions, wrong intervention, and non-SR. Finally, 535 studies were included in the analysis (Fig. 1 and Additional file 1: appendix table S1). Additional file 1: Appendix text S2 presents the list of eligible studies.

Fig. 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Flow Diagram of Study Selection Process

Study characteristics

The heatmap illustrated in Fig. 2 shows the country distribution of the first authors, disease types, and the number of included studies. Results indicated that the studies were conducted in 18 countries, with China having the largest number (n = 336), followed by South Korea (n = 90) and the United States (n = 30). The included SRs described 23 disease types classified by ICD-11, of which the most widespread were symptoms, signs or clinical findings (83 SRs), mental, behavioral or neurodevelopmental disorders (67 SRs), and diseases of the nervous system (66 SRs). In addition, the acupuncture techniques involved were electroacupuncture (n = 67), manual acupuncture (n = 47), acupoint catgut embedding (n = 41), dry needling therapy (n = 39), auricular acupuncture (n = 22), acupoint injection (n = 14), scalp acupuncture (n = 9), bee venom acupuncture (n = 9), fire needling therapy (n = 9), battlefield acupuncture (n = 5), ear acupuncture (n = 5), filiform acupuncture (n = 2), wrist-ankle acupuncture (n = 3), laser acupuncture (n = 2), warm needling therapy (n = 2), body acupuncture (n = 1), abdominal acupuncture (n = 1), and eye-acupuncture (n = 1).

Fig. 2

Heatmap for Characteristic of Included Studies

Quality assessment

Results obtained from the AMSTAR-2 [22] quality assessment tool indicated that the methodological quality of most SRs ranged from low to critically low (Additional file 1: Appendix table S2 provides details of the evaluation result of methodological quality, with the single items summarized in Fig. 3). Specifically, 106 SRs exhibited low methodological quality, whereas 379 SRs had critically low quality. Among the remaining 50 SRs, 22 were free of critical flaws and "provided an accurate and comprehensive summary of the results that address the question of interest." With regard to critical items in AMSTAR-2, such as protocol registration, literature search strategy, excluded studies, risk of bias assessment and influence on results, methods for meta-analyses, and assessment and influence of publication bias, only 38% and 20% of the studies reported protocol registration (domain 2) and excluded studies (domain 7), respectively. In other domains, almost all included studies (over 99%) reported PICO (i.e., population, intervention, comparison and outcomes) elements (domain 1). However, the vast majority did not explain the study designs (domain 3, over 98%) nor indicate the source of funding (item 10, over 93%).

Fig. 3

Evaluation of the Methodological Quality with AMSTAR-2

Evidence map of acupuncture-related adverse events

Figure 4 shows that 119 SRs were statistically analyzed for adverse events, 176 SRs reported the occurrence of no adverse events, 120 SRs reported no specific adverse reaction type, 89 SRs showed that none of the included original studies reported adverse reactions, and 260 SRs reported different adverse reactions. Acupuncture-related adverse events were classified as syncope (86 SRs), organ or tissue injury (233 SRs), systemic reactions (113 SRs), infection (19 SRs), and other adverse events (373 SRs). The most common adverse reactions are pain (144 SRs), bleeding/bruising (120 SRs), dizziness (86 SRs), haematoma (70 SRs), digestive system symptoms (46 SRs). Notably, only a few included studies reported adverse effects due to acupuncture therapists.

Fig. 4

Distribution of Acupuncture-related Adverse Effect

The effect size of adverse events caused by acupunctureEvidence of a positive effect

This study found 50 SRs with statistically significant acupuncture-related adverse effects, ranging from high- to critically low-quality. Most studies described general acupuncture or the specific types of acupuncture could not be distinguished, except for one slap acupuncture and one electroacupuncture in the low-quality category, and two fire acupuncture, two electroacupuncture, one auricular acupuncture, and one filiform needle in the critically low-quality category.

Evidence of a negative effect

Among the acupuncture-related adverse events, 76 were found to have no statistical difference when acupuncture alone or combined with other interventions was compared with other interventions. There was no significant difference in the adverse effects described in 44 studies that were rated as high quality (n = 11), moderate quality (n = 10) and low quality (n = 23) by the AMSTAR-2 tool, all of which used traditional acupuncture or the specific acupuncture type was not described. The remaining 32 SRs were of critically low quality, including dry needling (n = 1), fire needle (n = 1), scalp acupuncture (n = 1), wrist ankle needle (n = 1), electroacupuncture (n = 2), acupoint injection (n = 2), and the rest acupuncture.

Evidence of unclear effects

Moreover, results showed that more than 90 percent of the adverse reactions were only qualitatively described, and no statistical difference was reported. Forty-three adverse effects were described in a high-quality study that compared acupuncture with other interventions. In moderate quality studies, there were several types of acupuncture techniques, including electroacupuncture, filiform needle, manual acupuncture, auricular acupuncture, and dry needle, compared to other interventions. Low-quality studies compared acupoint catgut embedding, acupoint injection, ear acupuncture, scalp acupuncture, warm acupuncture, and fire needle. In moderate- and low-qualities, manual acupuncture was the most common method compared to other interventions. Importantly, critically low-quality studies explored all acupuncture types, with acupoint catgut embedding being the most frequently used among the minority acupuncture types.

Acupuncture-related adverse effectsSyncope

Acupuncture-related syncope and fainting, a reflex induced by vagal excitation and a common adverse event during acupuncture [24], were reported in 86 SRs (7.04%). This may be attributed to excessive stress and fear of acupuncture in patients, especially those new to acupuncture, frail, seriously ill, with excessive blood loss, with a history of dizziness from acupuncture, or in an improper posture. For doctors, most incidents of needle sickness were caused by the intensity of acupuncture. However, most studies demonstrated that acupuncture disease could be improved by rest and symptomatic treatment, with only a few severe cases.

Organ or tissue injury

A total of 233 SRs reported tissue or organ injury caused by acupuncture, mainly including pain (11.79%, 144 SRs), bleeding or bruising (9.83%, 120 SRs), haematoma (5.73%, 70 SRs), headache (2.21%, 27 SRs), pruritus (1.88%, 23 SRs), erythema (1.56%, 19 SRs), neuromuscular disease (1.31%, 16 SRs), aggravation of symptoms (1.15%, 14 SRs), skin allergy reaction (1.15%, 14 SRs), minor swelling (1.06%, 13 SRs), numbness (1.06%, 13 SRs), induration (0.90%, 11 SRs), palpitations (0.82%, 10 SRs), menoxenia (0.41%, 5 SRs), mild edema (0.25%, 3 SRs), blisters (0.25%, 3 SRs), and motor disorders (0.16%, 2 SRs).

Pain was the most frequently reported tissue or organ injury, with 144 studies. It was evident that improper acupuncture operation could lead to local chronic pain, chronic swelling pain, and other adverse reactions.

Local bleeding during acupuncture was reported in 120 studies. It is worth noting that bleeding during acupuncture is relatively common because many blood vessels are under the skin. Although acupuncturists avoid large and thick blood vessels when performing acupuncture, the tiny capillaries are difficult to see with the naked eye, and thus the needle may break the small blood vessel resulting in bleeding. However, patients only need to wipe the blood with a cotton swab in time and press it to relieve pain.

Heamatoma, a swelling and pain caused by bleeding under the skin at the acupuncture site, was reported in 70 studies. Given that the needle tip is bent with a hook, it can damage the flesh or stab a blood vessel. After withdrawing the needle, the acupuncture site was swollen and painful, and the skin was blue-purple. Generally, it is not necessary to deal with a small amount of subcutaneous hemorrhage and local bruising because it can subside on its own. However, cold compresses can be applied to stop the bleeding in instances where the local swelling and pain are severe, the bruising area is large, and the activity function is affected. In addition, hot compresses or light massage can be applied in the local area to promote dissipation and absorption of local blood stasis. To avoid heamatoma adverse effects, acupuncturists or professionals should carefully check the needles, familiarize themselves with the anatomical parts of the human body, avoid puncturing blood vessels, and press the needle hole with a sterile dry cotton ball immediately after the needle is withdrawn.

Headache, primarily due to subarachnoid hemorrhage, was reported in 27 SRs. Results revealed that the acupoints associated with headaches were Fengchi (GB20), Yamen (GV16), Anmian (EX-HN14), and Yiming (EXHN13).

Furthermore, 23 SRs reported observation of pruritus events during the interventions. The remaining 80 SRs reported mild and transient adverse events, including erythema (1.56%, 19 SRs), neuromuscular disease (1.31%, 16 SRs), aggravation of symptoms (1.15%, 14 SRs), skin allergy reactions (1.15%, 14 SRs), swelling (1.06%, 13 SRs), numbness (1.06%, 13 SRs), induration (0.90%, 11 SRs), palpitations (0.82%, 10 SRs), blood pressure disorders (0.82%, 10 SRs), menoxenia (0.41%, 5 SRs), edema (0.25%, 3 SRs), blisters (0.25%, 3 SRs), insomnia (0.25%, 3 SRs), motor disorders (0.16%, 2 SRs), and other adverse events (1.72%, 21 SRs). It should be noted that no serious adverse events were associated with the acupuncture treatments.

Systemic reactions

Acupuncture-related systemic reactions, including digestive system symptoms, tiredness, discomfort, psychological disorder, fever, and heat/sweating, were reported in 113SRs.

Among them, digestive system symptoms were reported in 46 SRs (3.77%). The common adverse effects associated with digestive symptoms included nausea/vomiting, loss of appetite, dry mouth, constipation, diarrhea, stomach, dyspepsia, and heartburn.

The reported tiredness (3.28%, 40 SRs), discomforts (2.54%, 31 SRs), and psychological disorder (2.46%, 30 SRs) adverse effects may be attributed to improvement of the viscera function after acupuncture, blood adjusting well, short fatigue and discomforts after constantly improved body function, improved metabolism, and improved immunity and physical fitness. However, results suggested that blood biochemical and related inspection diagnosis should be done before treatment for patients with severe fatigue. Moreover, the patient's complexion was generally dark after treatment, leading to a psychological disorder. Therefore, it is recommended that acupuncture treatment should first consider personal conditions and should not be carried out frequently.

Fevers and heat/sweating were reported in 10 SRs (0.82%) and 8 SRs (0.66%), respectively. Although some patients developed a fever after acupuncture, it was infrequent. Patients may suffer from mild colds and low fever after acupuncture because the body is exposed and may not be well protected during the process. On the other hand, fever after acupuncture may be due to hyperthermia during acupuncture, and is characterized by a slight local or systemic heat sensation after the treatment, which is a normal phenomenon.

Infection

Infection, one of the most common acupuncture adverse effects, was reported in 19 studies (1.56%). All patients recovered after appropriate treatment. Infection is mainly due to unsterilized needles, repeated use of needles, or contact with clothing at the needlepoint. Acupuncture points on the head are most commonly infected because the hair makes sterility standards challenging to enforce. One study reported that acupuncture techniques for relieving toothache could cause facial abscesses [25]. Other infections may be associated with the patient's blood glucose level [26]. However, acupuncture-related infection rates have been dwindling in recent years due to the popularity of health consciousness and the concept of disinfection.

Other adverse events

A total of 373 articles reported adverse reactions that were neither due to trauma nor infection, such as needle sticking, broken needle, and bent needle. These effects may be associated with a poor sense of responsibility of the doctors, improper twisting, improper patient positioning during the treatment process, and poor needle quality [26]. In addition, 176 studies did not report any acupuncture-related adverse events, 120 studies did not specify which adverse events were involved, and 89 studies stated that no adverse events were reported in the original study.