The imputability of neutralizing antibodies (NAB) in secondary non-responsiveness (SnR) to treatment by botulinum toxin (BTX) injections is still debated for limb spasticity. This systematic literature review aimed to determine the prevalence of NAB after BTX injections for limb spasticity, analyze their determinants, and their causal role in SnR.

Medline, Cochrane and Embase databases were consulted for the period 1990–2017. Two independent reviewers extracted the data and assessed the quality of the studies with a specific scale (PRIMA and STROBE guidelines). Since the techniques used to detect NAB did not influence the results, we calculated the sensitivity and specificity of NAB+ to reveal SnR.

We analyzed 13 articles published between 2002 and 2017, the quality of which was satisfactory (mean score 18/28a.u.). They were 5 randomized controlled trials, 5 interventional and 3 observational studies. NAB detection was the primary criterion for 4 and a secondary criteria for 9. On a total of 1201 participants (91% stroke), 1201 serums were tested after BTX injections. NAB prevalence was estimated between 1% and 2% according to various scenario, and did not significantly differ between the 3 BTX-A formulations. NAB production seemed favored by long duration therapy with high doses and short interval between injections. NAB revealed SnR with poor sensitivity (56%) but extremely high specificity (100%). No consensual criteria were used to diagnose non-responsiveness to BTX treatment.

This study is the first systematic review to specifically address the question of the BTX immunogenicity for limb spasticity. NAB prevalence is much lower than that reported for cervical dystonia. Consensual criteria must be defined to diagnose non-responsiveness to BTX treatment. If immunogenicity is not the most common cause of non-responsiveness to treatment, NAB should be more often sought in SnR persons for whom there is no other cause explaining treatment inefficacy.