Botulinum toxin type A for subacute/chronic neck pain
Cochrane Database Syst Rev. 2026 May 28;5:CD008626. doi: 10.1002/14651858.CD008626.pub4.
ABSTRACT
RATIONALE: Botulinum toxin type A (BoNT-A) intramuscular injections are commonly administered for neck pain unresponsive to other therapies, but the benefit-harm profile remains uncertain.
OBJECTIVES: To assess the benefits and harms of intramuscular BoNT-A injection for people with subacute to chronic neck pain.
SEARCH METHODS: We searched CENTRAL, MEDLINE, Embase, CINAHL and trial registers, together with reference checking, citation searching and contact with study authors up to 4 March 2026.
ELIGIBILITY CRITERIA: Randomised controlled trials (RCTs), including cross-over, pragmatic, or cluster designs, were included for adults with subacute to chronic neck disorder without radicular findings, with or without cervicogenic headache, and compared BoNT-A to placebo or no-treatment.
OUTCOMES: Our outcomes were pain, function-disability, health-related quality of life, participant-reported treatment success, adverse events and costs at both short-term (< 24 weeks) and long-term (≥ 24 weeks) follow-up.
RISK OF BIAS: We assessed the risk of bias (RoB) using the ROB 2 tool.
SYNTHESIS METHODS: We synthesised results using meta-analyses with random-effects models to calculate mean differences (MD) and risk ratios (RR) with 95% confidence intervals (CIs), or using SWiM (Synthesis Without Meta-analysis) methods. We assessed certainty with GRADE.
INCLUDED STUDIES: This update included 16 trials (855 participants) compared to six (252 participants) in the previous review: 13 parallel RCTs and three cross-over RCTs. One other trial is awaiting classification. Short-term follow-up ranged from 3 to 12 weeks post-treatment, and long-term follow-up was six months. Participants (mean age 45 years; 64% female; pain severity 56 on the Visual Analogue Scale (VAS 0 to 100)) were from ambulatory care settings in North America (five trials), Central America (one trial), Europe (six trials), the Middle East (one trial), Central Asia (one trial), and the East Asia & Pacific region (two trials), from upper (12 trials) or upper-middle (four trials) income economies. Disorder classifications included non-specific neck pain - myofascial (eight trials), whiplash-associated disorder (four trials), and cervicogenic headache (four trials). Overall, 88% of participants had chronic pain. The dose varied from 10 to 400 units over 1 to 10 trigger points. The RoB was high or raised some concern across all outcomes.
SYNTHESIS OF RESULTS: Since 2011, evidence has shifted from no benefit to a potentially slight-to-small short-term benefit for BoNT-A compared with placebo. No trials compared BoNT-A with no treatment, and evidence was limited by bias (no studies with an overall low RoB), imprecision, and potential publication bias. For pain intensity measured on the VAS (scale 0 to 100), BoNT-A may result in a slight-to-small improvement in the short term (MD 5.95 points lower, 95% CI 10.89 lower to 1.02 lower; I2 = 0%; 12 studies, 393 participants; low-certainty evidence; downgraded two levels due to high overall RoB and indirectness), but not in the long term (MD 11.74 points lower, 95% CI 35.63 lower to 12.16 higher; I² = 50%; 2 studies, 43 participants; low-certainty evidence; downgraded two levels due to high RoB and imprecision). Mean pain in the placebo group was 46.66 points compared to 40.71 in the BoNT-A group, representing an absolute improvement of 6% (95% CI 11% better to 1% better). Additionally, there may be a moderate risk reduction in pain events in the short term (RR 0.68, 95% CI 0.56 to 0.83; I² = 0%; 2 studies, 292 participants; low-certainty evidence; downgraded two levels due to high overall RoB and dissemination bias, including potential conflict of interest from funding sources that may overestimate effects). Assuming a 70% risk with placebo, the corresponding risk with BoNT-A is 47%, representing a 23% absolute risk reduction. Five people (95% CI 3 to 9) would need BoNT-A treatment for one person to benefit. For physical function measured on the Neck Disability Index (scale 0 to 100), BoNT-A may have little to no effect in the short term, but the evidence is very uncertain (MD 2.49 points higher, 95% CI 9.72 lower to 14.70 higher; I² = 0%; 3 studies, 76 participants; very low-certainty evidence; downgraded three levels due to high overall RoB, indirectness, and wide confidence intervals). For health-related quality of life (change score), BoNT-A may make little to no difference in the short term (MD 2.30 points higher, 95% CI 12.57 lower to 17.17 higher; 1 study, 28 participants; low-certainty evidence; downgraded two levels due to imprecision). For participant-reported treatment success, BoNT-A may provide a small improvement in the global rating of change. Short-term results: MD 1.16 points lower on a 1 to 5 Verbal Rating Scale (95% CI 1.86 lower to 0.46 lower; 1 study, 31 participants; low-certainty evidence), corresponding to an absolute improvement of 20% (95% CI 10% to 36%). Long-term results were similar: MD 2.00 points lower on a 0 to 10 Patient Global Assessment (95% CI 3.80 lower to 0.20 lower; 1 study, 19 participants; 20% absolute risk reduction, low-certainty evidence). Evidence was downgraded by two levels for RoB and imprecision. No studies evaluated participant satisfaction or serious adverse events. For non-serious adverse events, participants treated with BoNT-A may have a small, negligible increased risk (RR 1.51, 95% CI 1.04 to 2.20; I² = 28%; number-needed-to-treat for an additional harmful outcome (NNTH) 6, 95% CI 4 to 12; 7 studies, 489 participants; low-certainty evidence; downgraded two levels due to RoB and indirectness). Participants treated with BoNT-A have a 16% increased absolute risk of experiencing non-serious adverse events lasting up to three weeks, such as injection soreness, muscle fatigue, and headache.
AUTHORS' CONCLUSIONS: BoNT-A may provide slight-to-small short-term pain relief (6%) and perceived treatment success (20%) for subacute and chronic neck pain, but may offer no long-term physical function or health-related quality of life benefits. The moderate (23%) risk reduction in pain events may be overestimated due to publication bias. Non-serious adverse events are 16% more frequent. Uncertainty remains due to wide confidence intervals, short-term single-dose trials, and dose variation, highlighting the need for larger, high-quality studies with comprehensive adverse event reporting.
FUNDING: This Cochrane review had no dedicated funding.
REGISTRATION: Protocol and previous versions available via doi.org/10.1002/14651858.CD008626, doi.org/10.1002/14651858.CD008626.pub2 and doi.org/10.1002/14651858.CD008626.pub3 (withdrawn).
PMID:42206664 | DOI:10.1002/14651858.CD008626.pub4