Broken and retained needles

A needle can stay in the body after acupuncture in three ways: a practitioner forgets to remove a whole needle at the end of a session; a needle breaks and part of it stays under the skin; or a practitioner leaves fragments in the tissue on purpose, a practice described in reports from Japan and Korea. A thin steel fragment can then move through the tissue, and case reports have found fragments in the spinal canal, the heart and the kidney, sometimes decades after the treatment. This page reports how often incident surveys record retained needles, what the case reports describe, what guidance gives as the causes of breakage, what the international needle standard covers, and what the reports say about getting a fragment out. The single-use rule that governs needles today is on Infection and clean needles.

How often it is reported

A review of patient safety incidents reported to the UK’s National Patient Safety Agency between 2009 and 2011 analysed 325 that involved acupuncture: retained needles made up 31% of them, dizziness 30% and loss of consciousness 19%. The reviewers classed 95% of all the incidents as causing low or no harm. Their summary does not say how many of the retained needles were whole needles left in after a session and how many were fragments.

Two prospective surveys published in 2001 asked British practitioners to report events as they happened. Adrian White and colleagues collected 31,822 consultations by 78 doctors and physiotherapists and recorded 5 needles lost or forgotten, and 2 patients forgotten in the treatment room. Hugh MacPherson and colleagues collected 34,407 treatments by 574 members of the British Acupuncture Council and recorded 2 forgotten needles; they classed both as avoidable errors.

A Japanese teaching clinic measured forgotten needles directly. Yamashita and Tsukayama report 87 incidents involving 154 needles over four years at Tsukuba College of Technology Clinic, 73 of them noticed and removed before the session ended. Needles were forgotten mainly in the legs, the head and the back, where a towel, hair or clothing hid them. Reminders to count needles did not reduce the number; an incident reporting and feedback system did.

Breakage appears in the case-report reviews. Of 479 adverse events in the Chinese-language literature from 1980 to 2009, a 2010 review by Zhang and colleagues found one broken needle: a lump in the abdomen that turned out to be a fragment broken off 15 years earlier. A 2001 review of the Japanese case literature from 1987 to 1999 found 124 adverse events, and attributed 48 of them to needle breakage: 26 needles embedded on purpose and 16 broken by accident.

Why needles break

The World Health Organization’s 1999 Guidelines on basic training and safety in acupuncture list the causes as poor manufacture, erosion where the shaft meets the handle, a strong muscle spasm or sudden movement by the patient, pulling out a stuck or bent needle the wrong way, and prolonged use of galvanic (direct) current. The guidelines name the junction of shaft and handle as the part apt to break, and for that reason advise keeping a quarter to a third of the shaft above the skin. A needle that bends during insertion, they say, should be withdrawn and replaced, and a needle with a bent or eroded shaft or a hooked or blunt tip should be discarded before use.

Several of the serious cases followed self-treatment. A 55-year-old woman in Japan was needling the back of her own neck with a disposable 4 cm steel needle when the joint between needle and handle came apart and the whole needle went in. A 37-year-old Japanese man bought needles online, broke some in his neck, and kept treating himself after his doctor told him to see a spinal surgeon. Of 26 patients with injuries to the cervical spinal cord or brain from needles that Miyamoto and colleagues collected in 2010, 5 had needled themselves.

Electrical stimulation is the other cause studied. A 2010 laboratory study from the Korea Institute of Oriental Medicine stimulated stainless steel needles in rats’ leg muscles and found corrosion on their surface under the microscope; the risk was lower with bipolar, brief, low-current stimulation than with monopolar, prolonged, high-current stimulation. A 2014 Chinese study ran three common electroacupuncture waveforms through needles in a rabbit for 12 hours and found no corrosion, while direct current through needles in salt water corroded them severely; its authors suggest that some reported accidents come from the quality of the device or from improper handling. How electroacupuncture is delivered is on Electroacupuncture.

Needles left in on purpose

Yamashita and Tsukayama describe the Japanese practice: a silver or gold needle is inserted, the exposed part is cut off, and the fragment is pressed further in to stay for good. They write that in some areas of Japan it was not rare to find countless fragments on X-rays of people over 60, and that the Japan Acupuncture and Moxibustion Association recommended curtailing the method in 1976. Yamashita wrote to the CMAJ in 2011 that the Japanese term is maibotsu-shin, that it is no longer performed because of its adverse events, and that the retained needles are as thin as ordinary Japanese needles, 0.14 to 0.20 mm. A 2024 report from a São Paulo hospital calls the same practice okibari, and describes the fragments as up to 1 cm long. In Korea, short gold threads are implanted under the skin for pain; a 2021 report calls the practice widely used in Asian countries. Absorbable threads are a different practice, covered on Thread embedding.

Where fragments have been found

In three of the reports below, the fragments showed up on an X-ray, a mammogram or a CT scan (computed tomography, a cross-sectional X-ray) taken for another reason.

In the Miyamoto series of 26, 15 injuries came from needles embedded on purpose and 11 from needles broken by accident, and 16 patients developed symptoms more than 30 days after the event. The woman whose needle entered the back of her neck had it reach the space beside the medulla oblongata, the lowest part of the brainstem, and pierce the outer wall of the vertebral artery; the authors found six cases, theirs included, of needles migrating inside the skull, all after needling at the back of the neck. The 37-year-old man’s fragment entered the spinal canal between the first and second vertebrae and touched the spinal cord, leaving him with weakness in one leg and loss of sensation on the other side of his body.

What the needle standards test

ISO 17218:2014, from the International Organization for Standardization (ISO), specifies requirements for sterile single-use filiform needles, the solid needles of ordinary acupuncture. ISO confirmed it in 2019 and has a second edition at its final draft stage. The text is sold rather than published. A Brazilian study in 2025 tested nine brands of 0.25 by 30 mm needles with the standard’s mechanical tests, which it names as puncture performance, “intensity” and tip sharpness. Every brand’s tips deformed in the intensity test, and none fell outside the standard’s limits.

A second standard, ISO 20487:2019, sets out a test of corrosion resistance for single-use needles meant for electroacupuncture. It sets no pass or fail criteria, and ISO states that it is not intended to give safety information for clinical practice.

What the reports say about removal

The WHO guidelines set out the steps when a needle breaks during treatment: the patient keeps still so the fragment goes no deeper; a fragment above the skin is pulled out with forceps; one level with the skin is exposed by pressing gently around it; one completely under the skin may reappear if the patient returns to the position they were needled in; and if none of this works, surgery is needed.

Once a fragment has moved, the reports describe surgery guided by imaging. Miyamoto and colleagues found that 21 of their 26 patients had the needle removed surgically and 10 showed signs of recovery, while none of the patients treated without surgery improved; they conclude that needles embedded in the brain should be removed urgently where possible. The Japanese neurosurgeons who removed the needle beside the vertebral artery write that magnetic resonance imaging (MRI) must be avoided before surgery because the needles are usually stainless steel, and planned the operation with CT. The surgeons who removed the fragment from the spinal canal used CT in the operating room to confirm that nothing was left behind, because the needles are thin enough to be hard to see. The Chinese surgeons checked their laparoscopic removal with an X-ray.

Not every fragment is taken out. The Brazilian authors write that there is no consensus: some authors argue that the risk justifies removal, and others observe a fragment in the heart that has not caused bleeding around it. The Korean dermatologists left the gold threads in place and warned their patient that the cellulitis could recur.

What the sources cannot tell you

No study gives a rate of needle breakage. The incident surveys record forgotten needles and the case-report reviews record breakage only when someone wrote it up, and the English incident category does not separate the two. The migration reports describe single patients, and the series that compares surgery with no surgery is a collection of earlier reports rather than a trial. The two corrosion studies were done in animals and disagree about whether ordinary electroacupuncture corrodes needles. ISO’s test limits are not public, so this page names the tests as a study that used them describes them. AcuiQ’s entries record what a study prescribed, not what happened to the needles or the people in it; What AcuiQ does not do sets out that limit.