Double vision and ocular motor nerve palsy

Double vision, which doctors call diplopia, follows when one of the three nerves that move an eye stops working and the two eyes no longer point at the same spot. The American Academy of Ophthalmology’s EyeWiki pages describe the oculomotor nerve, the third cranial nerve, as supplying four of the six eye muscles, the muscle that lifts the upper eyelid and the muscle that constricts the pupil, and the abducens nerve, the sixth, as supplying the lateral rectus, the muscle that turns the eye outward. The third of the three, the trochlear or fourth cranial nerve, gets little attention in the sources on this page. The same pages list the causes as diabetic or hypertensive damage to the nerve’s small blood vessels, head injury, compression by an aneurysm or tumour, infection and drug toxicity, state that most microvascular abducens palsies recover under observation within 3 to 6 months, and describe the standard management of the double vision as prisms, covering one eye, botulinum toxin or surgery once the deviation has been stable for 6 months. One systematic review covers acupuncture for any of these palsies. It pooled 18 trials, most of them from China, and rated the certainty of every result low or very low; the trials published since are small and either add needling to drugs or compare two needling schemes.

Is there a Cochrane review?

No. The Cochrane Library holds no review of acupuncture for oculomotor, trochlear or abducens nerve palsy, for ophthalmoplegia (the general term for weakness of the eye muscles) or for diplopia, and a search of PubMed and Europe PMC in October 2026 found one completed systematic review, described below, and two protocols that have not reported.

The 2022 systematic review

The 2022 review by Zhang, Cui and colleagues in Complementary Therapies in Medicine searched PubMed, the Cochrane Library, Medline and three Chinese databases for randomised trials that compared acupuncture alone with no treatment, sham acupuncture or another active therapy, or acupuncture added to a therapy with that therapy alone. It included 18 trials with 1,150 participants, which its abstract describes as comparing acupuncture with other therapies, and assessed them with the Cochrane risk of bias tool and GRADE, the system that rates how far a pooled result can be trusted. Its full text sits behind the publisher’s paywall; this page reports the abstract.

The clinical efficiency rate is the share of people a trial classes as improved; the abstract does not say how the trials defined it. The abstract prints the confidence interval for the diplopia score as 1.39 to 0.77, which does not contain its own point estimate, so the table gives the estimate alone. The authors rate the GRADE certainty of the evidence as low or very low, note that most of the trials were run in China, and conclude that most of the included studies reported acupuncture ahead of its control.

Two review protocols without results

A protocol is the published plan for a review: what will be searched, which trials will count and which outcomes will be pooled, registered in advance so that the finished review can be checked against it. Two have been published for this condition and neither has reported. Ji, Qin and colleagues registered a review of acupuncture for ophthalmoplegia with PROSPERO, the international register of systematic reviews, as CRD42018091536 and published the protocol in Medicine in 2018. It planned to search to 31 March 2018 and to take the distance the eye can move, the size of the palpebral fissure and the degree of strabismus (misalignment of the eyes) as its primary outcomes. Huang, Miao and colleagues published a protocol in the same journal in 2022, registered as CRD42021266447, for a review of acupuncture combined with tuina, Chinese manual therapy applied to the body surface and to acupuncture points, for oculomotor paralysis; its primary outcomes were to be diplopia scores, the pupil’s response to light and the size of the palpebral fissure. The protocol states that in a third of cases the cause is unknown and names aneurysm, tumour, head injury, infection and diabetic neuropathy among the rest. A search of PubMed and Europe PMC in October 2026 found no completed review under either registration.

Trials since the review

The trials published after the review’s search are small and compare regimens rather than needling with no needling. Fahmy and colleagues 2025, physiotherapists in Cairo writing in Physiotherapy Research International, randomised 40 eyes with diplopia from oculomotor or abducens palsy to four weeks of electroacupuncture, needling with a current passed between the needles, plus standard medication, or to the medication alone, and measured the deviation with the Hess screen test, a chart on which the patient marks where a light appears to each eye. The abstract reports a post-test difference in favour of the electroacupuncture group, with p values printed as 0.38 and 0.003, and names no points. Electroacupuncture describes the technique.

Two trials in Acupuncture Research (Zhen Ci Yan Jiu) tested point selection by Biaoben Genjie theory, which the trials describe as the distribution, connection and gathering of channel qi in the body, and which picks points at four named sites on a channel. Wang and colleagues 2025 randomised 62 people with oculomotor nerve palsy, all on conventional medical treatment, to that scheme or to routine acupuncture, six courses of six daily sessions, and reported lower diplopia and strabismus degrees in the scheme group and an effective rate of 29 of 31 against 26 of 31. Guan and colleagues 2026 ran the same design in 60 people with abducens nerve palsy. Its scheme group was needled at TB23 (Sizhukong), TB03 (Zhongzhu), GB44 (Zuqiaoyin) and GB01 (Tongziliao), with GB20 (Fengchi), XH05 (Taiyang), XH07 (Qiuhou) and GB37 (Guangming), and the trial reported an effective rate of 29 of 30 against 26 of 30 with routine acupuncture. Because both arms were needled, neither trial measures what acupuncture adds to medical treatment.

Earlier trials in Chinese Acupuncture and Moxibustion (Zhongguo Zhen Jiu) compared needling with drugs. Liu and colleagues 2020 randomised 46 people with oculomotor nerve palsy after a stroke to mecobalamin injections with oral citicoline, or to the same drugs plus four weeks of needling at ST03 (Juliao) and GB04 (Hanyan) on the unaffected side, BL01 (Jingming), ST02 (Sibai), GB14 (Yangbai), BL02 (Cuanzhu), Shangming, TB23 and GB01 on the affected side, and LI11 (Quchi), LI06 (Pianli), TB05 (Waiguan), LI04 (Hegu) and GB20 on both sides. It reported a wider eye opening, a smaller pupil and more movement in three of the muscles in the needled group, and an effective rate of 19 of 23 against 15 of 23. Zhang and colleagues 2009 allocated 100 people with paralytic strabismus to needling at BL01, GB01 and Shangming or to oral methylcobalamin and vitamin B1, and reported an effective rate of 95% against 86%, with the needled group ahead for the muscles the oculomotor and abducens nerves supply and no difference for the superior oblique muscle.

Needling near the eye

Several of these points lie on the rim of the orbit. A 2023 review in the Korean Journal of Ophthalmology by Lee, Wu and Hwang collected 14 case reports, covering 15 eyes, of serious injury after acupuncture around the eye published between 1982 and 2020; ten eyes were perforated, and six of those needed vitrectomy, an operation to remove the gel inside the eye. The review counts six classical points in the periorbital area, BL01, BL02, GB01, ST01, ST02 and TB23, and lists ophthalmoplegia among the conditions for which needling there has been reported. A 2026 case report in the American Journal of Emergency Medicine describes orbital compartment syndrome, bleeding behind the eye that raised the pressure in the orbit to 79 mmHg, 90 minutes after periocular acupuncture; the eye saw nothing on arrival, emergency decompression within two hours restored vision to 20/25 by three months, and imaging at six months showed permanent thinning of the retina’s inner layers.

What guidelines say

The guideline nearest this subject is the American Academy of Ophthalmology’s Adult Strabismus Preferred Practice Pattern, written in 2019 and published in Ophthalmology in January 2020, which covers misalignment of the eyes in adults, including that left by nerve palsies. The Academy’s page for it, which carries the abstract and highlighted recommendations, does not mention acupuncture; this page could not open the full text. No later edition is indexed in Europe PMC as of October 2026. NICE’s Clinical Knowledge Summaries, the topic guides for English general practice, list Bell’s palsy and squint in children but no topic for diplopia or ocular motor nerve palsy. Clinical guidelines sets out other bodies’ positions on acupuncture.

What the reviews could not settle

The one review pooled trials whose comparators its abstract calls other therapies, and it reports no result against sham acupuncture, so the pooled estimates cannot separate needling from the attention that comes with it. Its effective rate is a category whose definition the abstract does not give, its certainty is low or very low on every outcome, and most of its trials were run in one country. The trials since have randomised 40 to 62 people each and compared one needling scheme with another, or needling plus drugs with drugs, never needling with none. Because the EyeWiki pages describe most microvascular palsies as recovering on their own within months, a trial without an untreated or sham arm cannot show that acupuncture shortened that recovery. The two registered reviews have not reported, and nothing in the sources here covers trochlear nerve palsy on its own.

AcuiQ’s diplopia, oculomotor nerve palsy, abducens nerve palsy, paralytic strabismus and eyelid ptosis pages list the protocols individual studies prescribed, each with its citation. Bell’s palsy covers palsy of the seventh cranial nerve, which moves the face, and hemifacial spasm its overactivity.