A spinal cord injury (SCI) damages the nerve pathways that carry movement, sensation and control of the bladder and bowel below the level of the injury. Trials in this field grade an injury on the American Spinal Injury Association (ASIA) impairment scale, from A, a complete injury, to E, normal function, and report strength as the ASIA motor score. They report independence in daily life on the Functional Independence Measure (FIM) or the modified Barthel index (MBI). Most trials below add acupuncture or electroacupuncture, a current passed through the needles (Electroacupuncture covers the method), to rehabilitation and compare the result with rehabilitation alone.
Cochrane reviews
Cochrane has no review of acupuncture for spinal cord injury. Its 2014 review of non-drug treatments for chronic pain after SCI, by Boldt and colleagues, included 16 trials with 616 participants, 2 of them of acupuncture. None of the 16 was judged at low risk of bias, and the authors reported that the trials of acupuncture, like those of five other treatments, provided no evidence that it reduces chronic pain. Its search ran in March 2011, and Cochrane has published no later version. The 2017 Cochrane review of acupuncture for neuropathic pain by Ju and colleagues included 6 studies with 462 participants, all with peripheral neuropathic pain, so none concerned pain arising from the spinal cord.
For the bowel, the 2024 Cochrane review of conservative, physical and surgical treatments for faecal incontinence and constipation in central neurological disease, by Todd and colleagues, included 25 studies with 1,598 participants. The interventions it lists range from nursing protocols and probiotics to massage, standing, electrical stimulation and transanal irrigation; acupuncture is not among them.
Reviews of motor and neurological recovery
A standardised mean difference (SMD) states an effect in units of the spread between patients, so that trials using different scales can be pooled. GRADE is the scale reviewers use to rate certainty, from high to very low.
| Review | Trials | Comparison | Main finding |
|---|---|---|---|
| Heo and colleagues 2013, Pusan National University | 16; 8 on function, 6 on the bladder, 2 on pain | Acupuncture added to conventional treatment, against that treatment alone; sham in 1 | Better ASIA motor and FIM scores; the 2 pain trials conflicted |
| Ma and colleagues 2015, bias-adjusted, searched to August 2014 | 12 | Randomised trials only | Neurological recovery RR 1.28 (95% CI 1.12 to 1.50); motor score 6.86 points higher (0.41 to 13.31) |
| Xiong and colleagues 2019, network meta-analysis | 22, with 1,644 people | Acupuncture at different body regions, with exercise, against exercise alone | Higher ASIA motor scores with electroacupuncture and exercise than with exercise alone |
| Tan and colleagues 2022, electroacupuncture, searched to September 2020 | 10, with 712 people | In 8, electroacupuncture with rehabilitation against rehabilitation alone | ASIA motor score SMD 0.96 (0.75 to 1.18; 5 trials); FIM 13.46 points higher (3 trials); moderate certainty for motor function and daily living, low for sensation |
Of the 16 trials in Heo’s review, 12 were from mainland China, 2 from Taiwan and 2 from the United States; the authors called for sham-controlled trials for each condition. Ma’s review, from Zhejiang Chinese Medical University and the University of Queensland, found larger effects in trials of fewer than 30 people and in acute injury, judged the studies generally of poor quality, and found publication bias favouring positive studies, so the authors wrote that the benefit was by no means definitive. Xiong’s review, from West China Hospital at Sichuan University, ranked electroacupuncture at the head and back, and at the back and front of the trunk, highest. Tan’s review, from Guangzhou University of Chinese Medicine and Sun Yat-sen University, reported that most of its trials were published in Chinese and that most only used the word “random” without describing a randomisation. Five of its trials used Governor Vessel points, and one needled Huatuojiaji points (EX-B2 in the WHO notation) on both sides of the injured segment with GV14 (Dazhui) and GV04 (Mingmen).
A trial in acute injury
Wong and colleagues at Chang Gung Memorial Hospital in Taiwan reported a trial in 2003 that randomised 100 people with acute traumatic injuries graded ASIA A or B to rehabilitation alone or rehabilitation with electrical stimulation through adhesive surface electrodes at SI03 (Houxi) and BL62 (Shenmai), plus ear acupuncture, starting in the emergency room or soon after surgery. The authors reported that sensory, motor and FIM scores in the acupuncture group improved from admission to discharge and to 1 year, and that a greater share of that group moved up an ASIA grade. There was no sham, and the stimulation at the two body points used skin electrodes rather than needles.
Bladder
A 2016 review by Wang and colleagues at Guang’anmen Hospital in Beijing found 3 trials with 334 people with chronic urinary retention after SCI. Against rehabilitation training alone, the trials that added acupuncture reported 109 mL less urine left in the bladder after voiding (95% CI 62 to 157); against sterile intermittent catheterisation alone, they reported a higher response rate (RR 1.23). The authors drew no definitive conclusion for want of high-quality trials. A 2021 network meta-analysis by He and colleagues pooled 26 trials with 1,652 people, ranked electroacupuncture with moxibustion first for retention, and detected publication bias. For incontinence, a 2025 review by Zhang and colleagues found 15 trials with 1,394 people, all published and conducted in China, comparing electroacupuncture with or without rehabilitation against rehabilitation alone or with sham. Incontinence episodes fell by 1.42 a day (95% CI 0.96 to 1.88) and residual urine by 20 mL, both at low certainty; across outcomes certainty ranged from low to moderate, and none of the trials had a registered protocol.
Two trials stand out from that pool. Cheng and colleagues reported in 1998 on 80 patients given an intermittent catheterisation programme, half also given electroacupuncture at CV03 (Zhongji), CV04 (Guanyuan) and both BL32 (Ciliao). Among the 60 who reached balanced voiding, those with upper motor neuron lesions got there in 57 days with electroacupuncture against 85 without; complete injuries with marked detrusor-sphincter dyssynergia, where bladder and sphincter contract against each other, or with a bladder that did not contract at all, were not affected. Gu and colleagues ran a three-arm trial in 2015 with 107 patients with urinary retention: clean intermittent catheterisation alone, with electroacupuncture, or with sham acupuncture. The catheterisation-only and sham arms did not differ at 1 month, and more patients reached bladder balance with electroacupuncture than in either. Urinary retention reports the trials in retention from other causes.
Bowel
No randomised trial of acupuncture for neurogenic bowel dysfunction after SCI turned up in this search. A 2026 single-centre study by Liao and colleagues let 106 patients choose between anorectal biofeedback alone and biofeedback with electroacupuncture at the sacral Baliao points. At 6 weeks the neurogenic bowel dysfunction score fell 5.2 points with the combination and 2.9 with biofeedback alone, an adjusted difference of 2.1 points. The authors wrote that the design does not establish an effect specific to electroacupuncture and that randomised trials with attention-matched controls are needed.
Neuropathic and shoulder pain
A 2022 review by He and colleagues at Zhejiang Chinese Medical University found 6 trials with 286 people, 5 from China and 1 from the United States, each adding acupuncture to conventional therapy. Pain severity was lower (SMD −1.40, 95% CI −2.23 to −0.57), publication bias was observed, and the authors concluded that a role for acupuncture remains to be determined. One included trial needled GV20 (Baihui), LI04 (Hegu) and tender points.
The trials with an active or sham control are small. At the Karolinska Institutet in Stockholm, Norrbrink and Lundeberg assigned 30 people with neuropathic pain after SCI to acupuncture or massage twice a week for 6 weeks; 8 of 15 given acupuncture and 9 given massage reported improvement at the end, and 6 and 1 at 2 months. At the Kessler rehabilitation centre in New Jersey, Dyson-Hudson and colleagues randomised 17 wheelchair users with chronic shoulder pain to acupuncture or shallow needling away from acupuncture points; pain fell 66% and 43%, a difference the trial could not distinguish from chance. Estores and colleagues randomised 24 people with below-level neuropathic pain to 8 weekly sessions of Battlefield Acupuncture, a ten-needle ear protocol, or a waiting list. Pain fell 2.92 points against 1.13 on a 0 to 10 scale (P = 0.065), with pain at baseline higher in the acupuncture group. Reading an acupuncture trial explains why a waiting list and a sham answer different questions.
What guidelines say
The CanPain SCI 2021 update, the Canadian guideline for neuropathic pain after SCI, published in Spinal Cord in 2022 and replacing the 2016 treatment recommendations, lists acupuncture among treatments requiring further research. Its working group cited the lack of a standardised protocol and the waiting-list control in the Battlefield trial; 56% voted to recommend acupuncture, short of the 75% the guideline requires.
The Consortium for Spinal Cord Medicine’s Evaluation and Management of Pain After Spinal Cord Injury, published by Paralyzed Veterans of America (PVA) in 2026 from evidence available in 2025, makes a conditional recommendation (5B.4) that additional non-drug options can be considered after careful discussion of their limited evidence. It lists acupuncture among therapies with limited evidence of positive effect, on the strength of 1 low-quality randomised trial of fewer than 50 people. The Consortium’s bladder management guideline of August 2006, still on PVA’s guideline list, names acupuncture only among the interventions its evidence tables recorded and makes no recommendation about it.
NICE’s CG148 on urinary incontinence in neurological disease, published on 8 August 2012 and last updated on 2 October 2023, does not mention acupuncture. For people with spinal cord disease and overactive bladder symptoms it recommends antimuscarinic drugs, then botulinum toxin injections into the bladder wall when those fail or are not tolerated.
What the reviews could not settle
The reviews could not separate the needling from the extra treatment time around it, because almost every pooled trial added acupuncture to rehabilitation and compared it with rehabilitation alone. They could not say whether the effects hold in larger trials: Ma found the largest effects in trials of fewer than 30 people and in acute injury, and three of the reviews detected publication bias. Nor could they draw on trials from more than a few settings, since the pooled trials come mostly from China. For pain, Cochrane’s review has not been updated since 2014, and the one sham-controlled pain trial in these reviews enrolled 17 people. For the bowel there are no randomised trials to pool.
AcuiQ’s spinal cord injury page lists the protocols individual studies prescribed, each with its citation, and Stroke recovery reports the corresponding evidence after stroke.