Electroacupuncture connects pairs of inserted needles to a stimulator that sends a pulsed current between them. The US Council of Colleges of Acupuncture and Oriental Medicine’s Clean Needle Technique Manual (seventh edition, revised 2017) describes it as a replacement for prolonged manual manipulation of the needle. Transcutaneous electrical acupoint stimulation, or TEAS, sends a similar current through pads stuck to the skin over the points, with no needle at all. The settings that distinguish one electrical treatment from another are the frequency, the intensity, the shape of the pulse and the time it runs, and the STRICTA reporting guideline for acupuncture trials asks every trial that uses current to report its strength and frequency.

Frequency: 2 Hz, 100 Hz and dense-disperse

The case for choosing between 2 Hz and 100 Hz comes from the Chinese physiologist Han Jisheng. In a 2004 review of his group’s work, Han reports that stimulation at 2 Hz (two pulses a second) speeds the release of the opioid peptides enkephalin, beta-endorphin and endomorphin, that 100 Hz selectively increases the release of dynorphin, and that alternating the two releases all four at once. He states that clinical studies in chronic pain, including low back pain and diabetic nerve pain, bore this out. The alternating pattern is called dense-disperse. A 2025 commentary by Fan Qu and Rong Zhang adds that Peking University’s version of it pairs each frequency with its own pulse width, 600 microseconds at 2 Hz and 200 at 100 Hz.

Intensity, waveform and needles

The Clean Needle Technique Manual gives the current used in therapeutic electroacupuncture as about 0.5 to 6 milliamperes (mA) and says this level should be safe in a healthy person with no implanted electrical device. It sets the intensity by the patient’s report: the current is turned up slowly, kept just below the level the patient feels as pain, and switched off before the needles come out. It says muscle contraction should be avoided except where contraction is the aim, as in some treatments for palsy, and that a session often runs 20 to 30 minutes. On needles, it calls stainless steel safe with current, advises against silver needles, which are softer and can break down electrically in the body with a toxic reaction, and notes that plastic handles may conduct poorly. It also records that the negative electrode makes the tissue around it alkaline and the positive one acidic, and that what the two poles do in practice has not been worked out.

Devices and their regulation

In the United States, electro-acupuncture stimulators have product code BWK, and the Food and Drug Administration’s classification database lists them as unclassified because they were on the market before the 1976 medical device amendments. At a June 2021 advisory panel meeting, the FDA said it had cleared 21 such devices, each by showing it was substantially equivalent to one already sold, and proposed making them class II devices with special controls: tests of biocompatibility, sterility, electrical and thermal safety and electromagnetic compatibility, and labelling that states the device’s electrical output. The FDA’s own description of the device type covers both needle and skin electrodes, and states its principle as the stimulation of acupuncture points influencing parts of the body that are not next to them. The database still listed the code as unclassified when this page was checked.

TEAS: current through the skin

Qu and Zhang define TEAS as stimulation of acupuncture points by current from paired electrodes on the skin surface, and trace it to 1990, when Han’s team built a stimulator that combined the technology of transcutaneous electrical nerve stimulation (TENS) with acupuncture point selection and the alternating 2 and 100 Hz pattern. A 2026 survey of trial registries by Yilong Liu and colleagues found 239 trials of TEAS around surgery registered between 2012 and 2025, with the number rising in 2024. A 2026 trial in JAMA Surgery by Zheng and colleagues gave 232 women with moderate to severe nausea and vomiting after thyroid or neck surgery either a TEAS wristband over PC06 (Neiguan) for 2 hours with a saltwater injection, or an inactive wristband with the anti-sickness drug metoclopramide. After 2 hours, 77.6% of the TEAS group and 55.2% of the drug group had a nausea score of 3 or less on the trial’s rating scale, its definition of remission. Nausea after surgery reports the reviews of PC6 stimulation.

How TEAS differs from TENS

Both send current through skin pads, and the difference is where the pads go and why. A TENS trial puts them at or near the pain: that is the inclusion rule of a 2022 meta-analysis by Mark Johnson and colleagues, which included 381 randomised trials with 24,532 participants. Across the 91 trials that compared TENS with a placebo, pain was lower during or straight after TENS, and the authors rated that evidence moderate certainty; adverse events were poorly reported. A TEAS trial places the pads on acupuncture points, which can be far from the complaint, as with the wrist point used for nausea in the trial above.

How electroacupuncture appears in trials

The FDA’s 2021 literature review found 105 relevant human studies published from 2010 to 2020, of which 42 concerned musculoskeletal pain and 17 pain after surgery. The table gives three sham-controlled trials of electroacupuncture with several hundred participants each.

Three large sham-controlled trials of electroacupuncture
TrialParticipantsTreatment and controlMain result
Liu and colleagues, Annals of Internal Medicine, 20161,075 people with chronic severe functional constipation, 15 hospitals in China28 sessions over 8 weeks at traditional points, or sham electroacupuncture at non-pointsComplete spontaneous bowel movements rose by 1.76 a week against 0.87 with sham (difference 0.90, 95% CI 0.74 to 1.10)
Liu and colleagues, JAMA, 2017504 women with stress urinary incontinence, 12 hospitals in China18 sessions over 6 weeks in the lower back, or sham with no skin penetrationUrine leakage on a 1-hour pad test fell by 9.9 g against 2.6 g (difference 7.4 g, 95% CI 4.8 to 10.0)
Tu and colleagues, Arthritis & Rheumatology, 2021480 people with knee osteoarthritisElectroacupuncture, manual acupuncture or sham, 3 times a week for 8 weeksResponse at week 8 in 60.3%, 58.6% and 47.3%; electroacupuncture differed from sham by 13.0 points (97.5% CI 0.2 to 25.9), manual acupuncture by 11.3 (−1.6 to 24.4)

The constipation trial notes that the acupuncturists could not be blinded. AcuiQ’s pages on constipation, stress urinary incontinence and knee osteoarthritis list the protocols individual studies prescribed, each with its citation.

Safety

Of the 105 studies in the FDA’s review, 28 reported on safety and 17 reported adverse events linked to electroacupuncture, which the FDA listed as ranging from mild fainting, bruising and nausea to loss of consciousness and death; its search of its own adverse event database from 1979 to 2020 found no reports under the product code, and it found no recalls. The Clean Needle Technique Manual advises against any circuit that crosses the chest from one side to the other, crosses the spine, or runs near the brainstem or the front of the neck, and against electroacupuncture in anyone with a pacemaker or implanted defibrillator. Electrical stimulation and implants covers implanted devices in full.

What the sources cannot tell you

Han’s findings describe which peptides each frequency releases; they do not show that one frequency serves a given condition better than another. The trials in the table compare electroacupuncture with a sham, and the one that also ran manual acupuncture compared each with sham rather than with each other. Urinary incontinence reports the sham-controlled trials of electroacupuncture over the sacrum for leakage. Trials also differ in intensity, pulse width and session length, and STRICTA’s authors found some of the details their checklist asks for poorly reported in the trials they surveyed.

A protocol entry on AcuiQ names how the study delivered it, and electroacupuncture and electrical stimulation on the skin are separate modalities there; the entry does not record frequency or current. Reading a protocol explains each field.