Acupotomy treats soft tissue with a needle whose tip is a small flat blade, so that the practitioner can cut or loosen tissue under the skin rather than only puncture it. A 2019 overview of systematic reviews by Kwon, Yoon and Lee describes it as a modern type of acupuncture that uses a blade-needle with a flat surgical scalpel at its tip. A 2025 review of Korean clinical studies by Bae and colleagues says the blade is used to dissect, cut or detach adhesions in soft tissue, chiefly for chronic pain that follows tissue damage. This page reports where the method came from, how the instrument differs from an acupuncture needle and from a surgeon’s percutaneous release, what the systematic reviews report and what is known of its harms.
Names and origin
The method has no settled name. A 2019 terminology study by Yoon and colleagues searched English, Chinese and Korean databases up to July 2017 and counted 112 English, 1,129 Chinese and 44 Korean papers. “Needle knife” was the commonest term in the English literature and “acupotomy” in the Chinese and Korean literature; “miniscalpel needle”, “miniscalpel acupuncture” and “needle scalpel” each appeared in under 10% of papers. The authors called for a consensus on one term.
A 2023 survey of Korean medicine doctors by Jun and colleagues states that Zhu Hanzhang created acupotomy in 1976, combining two of the classical needles, the sharp-edged needle and the sword-like needle (bongchim and pichim in their Korean readings). Bae and colleagues trace it to a long, thick, spear-like needle among the nine classical needles of the Huangdi Neijing, which was used to lance abscesses and let blood, and say that Zhu modernised it in 1976 into an instrument with a handle, a body and a blade.
How is acupotomy different from acupuncture?
The instrument is larger and the aim is mechanical. Bae and colleagues describe a knife-shaped tip that is wider and thicker than a fine acupuncture needle, and a 2024 Delphi study by Jun and colleagues calls it a thicker needle with a knife built into the tip. The 2023 survey adds that the needle can go deep, reaching muscle and the periosteum, the membrane covering bone, and in some conditions the joint cavities and bursae. Dry needling, by contrast, uses the fine filiform needle of acupuncture.
The targets are also chosen differently. Bae and colleagues write that acupotomy combines acupuncture with modern anatomy and does not necessarily follow acupoints and meridians: in practice it is applied where the practitioner’s palpation finds pain, which often corresponds to ashi points, the tender spots needled wherever they are found, and to the muscles, ligaments and other soft tissue around them. In the 22 Korean lumbar studies they reviewed, the needle went into acupoints in 10 and into ashi or tender points in 8. They also note that needle length, diameter, blade width and thickness, insertion depth and the number of needle passes all change how strong the stimulus is.
Acupotomy and percutaneous release
A surgical procedure on the same structure is percutaneous A1 pulley release for trigger finger, a condition in which a thickened band of tissue at the base of the finger, the A1 pulley, catches the flexor tendon. A 2025 systematic review by Wen and colleagues describes the surgical options: open release, the reference standard, cuts the pulley through an incision, while percutaneous release cuts it with a small instrument passed through the skin. The review reports that percutaneous release is as effective as open release, less invasive and less prone to infection, with nerve damage as a possible complication. Hand surgeons offer either after corticosteroid injection has failed.
A 2023 meta-analysis of ultrasound-guided acupotomy for trigger finger by Liang and colleagues states that acupotomy treats the condition by a mechanism similar to percutaneous A1 pulley release, and argues that ultrasound lets the practitioner locate the pulley and avoid the nerves and vessels beside it. Trigger finger reports the reviews of acupuncture and related methods for that condition.
What the reviews report
Kwon and colleagues searched 11 databases to February 2019, without language limits, and found 11 systematic reviews covering 69 randomised trials, all on musculoskeletal conditions. They rated the reviews medium to high in methodological quality on AMSTAR, a checklist for appraising systematic reviews. Against other active treatments, acupotomy groups had higher total effective and cure rates, the shares of patients each trial rated as improved or cured, for frozen shoulder, cervical spondylosis, third lumbar transverse process syndrome, trigger finger, knee osteoarthritis and lumbar spinal stenosis. The authors concluded that more high-quality evidence is needed before the procedure can be recommended. Reviews published since then are condition by condition:
| Review | Condition and comparison | Included | What it found |
|---|---|---|---|
| Zhang and colleagues, 2024 | Knee osteoarthritis; overview of reviews published 2018 to 2023 | 9 systematic reviews | On AMSTAR 2, 1 review was of low and 8 of very low methodological quality. Of 35 outcomes graded with GRADE, 2 were moderate certainty, 7 low and 26 very low. |
| Lee and colleagues, 2023 | Knee osteoarthritis; acupotomy against acupuncture | 43 trials | Higher cure rate with acupotomy (odds ratio 2.94, 95% CI 2.36 to 3.65) and more pain relief; most trials at moderate risk of bias. |
| Zhang and colleagues, 2025 | Cervical spondylosis; ultrasound-guided acupotomy against placebo, physical therapy or other treatment | 33 trials | Higher clinical effective rate (risk ratio 1.17), rated high certainty; less pain, with evidence for pain, disability and range of motion rated low or very low. |
| Guan and colleagues, 2026 | Chronic spinal pain; ultrasound-guided acupotomy against controls | 32 trials, 2,669 participants | Less pain (standardised mean difference −0.80) and disability; moderate certainty for pain and response rate, low for other outcomes; every trial was run in China. |
| Liang and colleagues, 2023 | Trigger finger; ultrasound-guided acupotomy against unguided acupotomy, surgery or injection | 15 trials, 988 patients | Higher clinical effectiveness and fewer recurrences (risk ratio 0.14) with ultrasound guidance; few trials reported adverse reactions. |
| Tang and colleagues, 2025 | Carpal tunnel syndrome; ultrasound-guided needle-knife against unguided needle-knife or no intervention | 17 trials, 996 patients | Better pain, wrist function, nerve conduction and cure rate with ultrasound guidance; all trials from China, 6 reported adverse effects. |
| Feng and colleagues, 2024 | Plantar fasciitis; small needle-knife against extracorporeal shock wave therapy | 6 trials | No difference in curative effect or short-term pain; less pain with the needle-knife at mid-term and long-term follow-up; none of the 6 trials reported adverse events. |
Three of these reviews compare acupotomy with acupotomy, guided by ultrasound or not, so they test the guidance rather than the method. Knee osteoarthritis and Plantar heel pain report the reviews of acupuncture for those conditions.
Who practises it
Bae and colleagues write that acupotomy is used in China and in South Korea, mostly for musculoskeletal disorders such as disc herniation, neck pain and knee osteoarthritis. In Korea it is practised by Korean medicine doctors: the Jun survey went to the 185 members of the country’s largest acupotomy society, of whom 107 replied. They named musculoskeletal and connective tissue diseases in 80.8% of the indications they suggested, and their procedures differed considerably from one another. Most said they used no pain relief during the procedure, and 19.1% used a local anaesthetic cream. Respondents asked for imaging devices during treatment, clinical practice guidelines and wider national health insurance coverage. The trials in the reviews above come mostly from China: Guan and Tang report that every trial they included was run there.
Adverse events
In a 2022 pilot study in Korea, Yoon and colleagues followed 28 patients through 258 sessions and 1,185 treated points: 8 sessions (3.11%) were followed by a systemic adverse event and 27 points (2.28%) by a local one, and none was severe. Adverse events were more frequent with wider blades and with more needle passes at each point.
The Jun survey asked doctors which events they had seen: bruising (77.3% of respondents), fatigue (57.7%), pain (51.8%) and haematoma (51.8%); 1.8% reported a severe event. The 2024 Delphi study, in which 13 Korean doctors agreed a checklist for recording harms in research, lists 12 local events, among them infection, nerve damage and pneumothorax (a collapsed lung), and 14 systemic ones, among them fainting and procedural shock. In the 22 Korean lumbar studies, Bae and colleagues found that only 6 addressed safety and 2 reported adverse events, both pain or discomfort; they list bleeding, infection, scarring and nerve damage, including spinal cord injury, as harms the procedure can cause. In the knee trials Lee and colleagues pooled, adverse events were more frequent with acupotomy than with acupuncture, but the interval (odds ratio 1.23, 95% CI 0.42 to 3.60) includes no difference.
Jin and Yin reported in 2023 a Chinese patient with a healthy immune system who developed a fungal infection of the L4/5 disc and vertebrae, caused by Aspergillus terreus, after acupuncture and acupotomy. The vertebrae were destroyed despite antifungal treatment, and the patient needed debridement, decompression and spinal fusion before scans at six months showed the infection treated.
What the sources cannot settle
Kwon and colleagues found reviews that compared acupotomy with other active treatments and measured success mostly as effective and cure rates. The later reviews add more trials but not more certainty: the knee overview rated most of its evidence very low, and Guan and colleagues name the lack of blinding as the main risk of bias and note that trials run only in China may not carry over elsewhere. How often serious harm occurs is not known, because trials seldom record it and the pilot study followed 28 patients; the Korean group behind the pilot and the checklist plans larger prospective studies built on them.