Point injection (穴位注射, xuewei zhushe), also called aquapuncture, and in Korea pharmacopuncture when the liquid is a herbal extract, uses a hypodermic syringe to inject a small amount of liquid into an acupuncture point instead of, or as well as, inserting a needle. The liquid may be a vitamin, saline, sterile water, a local anaesthetic, a conventional drug, a herbal extract or bee venom. A 2016 overview in the QJM by Sha and colleagues found the method applied to a wide range of conditions, mostly painful ones, on poor clinical evidence, and wrote that reported adverse effects made standardising the procedure urgent. This page reports what is injected and who may inject it, what reviews found, what their comparisons can and cannot show, and the adverse events reported.
What is injected
The trials in the reviews below injected:
- Vitamins: vitamin B1, and mecobalamin, a form of vitamin B12.
- Conventional drugs: neostigmine and metoclopramide to restart the bowel after surgery, and anisodamine against vomiting.
- Saline, mostly as the control.
- Herbal extracts. In Korea, pharmacopuncture is defined in a 2026 trial by Lee and colleagues as injecting herbal extracts into acupoints, muscles or nerves, and the authors describe it as widely used in Korean clinics.
- Bee venom, injected into points as bee venom acupuncture.
Who may inject
Who may inject is set by each jurisdiction’s law. Washington State is one that sets the terms out for acupuncturists: its Department of Health defines point injection therapy as injection under the skin, into muscle or into the skin of acupuncture points, ashi points, trigger points and other points, and limits the substances to saline, sterile water, herbs, minerals, vitamins in liquid form, and homeopathic and nutritional substances, with local anaesthetics such as lidocaine and procaine allowed to reduce the pain of injection. Controlled drugs and steroids are excluded. A practitioner needs at least 24 hours of training, 8 of them clinical, including 2 hours on the use of intramuscular adrenaline (epinephrine), local anaesthetics and oxygen. The rules elsewhere differ, and Choosing a practitioner covers how acupuncture is regulated country by country.
What the reviews report
A point injection trial tests two things at once: the substance and the place it goes. A trial that compares a drug injected at a point with usual care and no drug shows the effect of the drug at least as much as the effect of the point. The comparison that isolates the point gives the same drug, in the same dose, by another route or at a non-point site, and only some trials make it.
| Review | Condition and substance | Trials (participants) | Comparison | What it found |
|---|---|---|---|---|
| Xie and colleagues, 2020 | Chronic non-specific low back pain, various | 13 (1,381) | Various controls | Insufficient evidence that injection reduced pain (2 trials pooled); a better “effective rate”, a composite outcome; very low-certainty evidence (GRADE). |
| Wang and colleagues, 2015 | Preventing bowel paralysis after surgery, injection at ST36 (Zusanli) | 30 (2,967) | Usual care; intramuscular injection elsewhere | Earlier passing of wind with neostigmine (15 trials), vitamin B1 (5) and metoclopramide (3) at ST36 than with usual care; vitamin B1 at ST36 also ahead of the same vitamin injected into muscle (4 trials, about 17 hours sooner). All trials at high risk of bias. |
| Zhang and colleagues, 2023 | Diabetic peripheral neuropathy, mecobalamin | 10 (927) | The same drug by other routes | A higher clinical effective rate with injection at points (risk ratio 1.27), with similar safety. |
| Sung and colleagues, 2025 | Musculoskeletal pain, bee venom | 20 | Saline injection, with or without acupuncture | Lower pain scores than saline injection (2 trials, 85 patients); no anaphylaxis in the eight trials that reported on safety. The authors note few trials, small samples and wide variation in dose. |
Two trials show the distinction in practice. In a 2023 trial by Xue and colleagues, 90 women having weight-loss surgery received anisodamine or saline injected into ST36 on both legs: 42.4% vomited in the first 24 hours with anisodamine against 72.4% with saline, and nausea did not differ. Saline at the same point is the control, so the trial tests the drug, delivered at a point. In Lee’s 2026 trial in Korea, 96 patients with lumbar spinal stenosis received herbal pharmacopuncture twice a week for 12 weeks or conventional usual care; pain on a 10-point scale fell 2.7 points more with pharmacopuncture, and no serious adverse events occurred. The comparison there is with a different treatment altogether.
AcuiQ’s postoperative ileus, chronic low back pain and diabetic peripheral neuropathy pages list the protocols individual studies prescribed, each with its citation.
Adverse events
The adverse events of point injection include those of the substance injected. Bee venom is the most studied:
- A 2015 systematic review by Park and colleagues of 145 studies found a median of 28.9% of patients with an adverse event in audit studies of bee venom immunotherapy. In randomised trials, bee venom acupuncture was 3.61 times as likely as a saline injection to cause one. The authors called adverse events frequent and asked that practitioners’ education and qualifications be ensured.
- A 2022 review by Ko and colleagues of 49 studies put the rate of anaphylaxis, a severe whole-body allergic reaction, at 0.045% of patients given bee venom acupuncture, about 1 in 2,200. Women had a higher rate than men, and the rate was lower in studies published since 2010 than before 2000.
For other substances the trials say less. Wang’s review of ST36 injection called safety inconclusive because few trials reported adverse events, and Sha’s overview cites reported adverse effects as the reason the procedure needs standardising. Infection and clean needles covers skin preparation and the infections reported after acupuncture, and After a treatment lists the symptoms that call for a doctor.
How AcuiQ records it
Each protocol on AcuiQ records how the points were treated, from a fixed list of modalities, and an injection study is recorded as “Point injection”, with its sessions and course length. Reading a protocol explains each field. The entry lists the points; what was injected, and in what dose, is in the cited paper, and a reader comparing two injection protocols at the same point may be comparing two different drugs.
What the sources cannot settle
Most reviews pool trials of different substances under one heading, and most trials compare an injection at a point with care that includes no injection, so they cannot separate the point from the substance. The trials that make that separation, such as the vitamin B1 and mecobalamin comparisons above, are few, and Wang’s review rated all of its trials at high risk of bias. Safety is counted well only for bee venom; for vitamins, drugs and herbal extracts injected at points, the reviews found too few trials reporting adverse events to give a rate.