Conventionally, the standard test for detection of antibodies against botulinum toxin (BT-A) has been the mouse lethality assay (MLA). Because this test has a number of disadvantages, a novel mouse protection assay (MPA) was recently introduced. We sought to compare the results of both tests. Forty-three samples from 38 patients with cervical dystonia and complete or partial subjective BT-A therapy failure underwent simultaneous MPA and MLA testing. Twenty-seven samples showed concordant results in both tests. Eleven of them were MPA- and MLA-positive and 16 MPA- and MLA-negative, resulting in a significant association of the dichotomous test results (Fisher exact test, p <0.01). Sixteen samples showed discordant results. All of those were MPA-positive and MLA-negative. This excess of MPA-positive results was also significant (Wilcoxon signed-rank test, p <0.001). Of the patients with MPA-positive samples, 62% had complete and 38% had partial therapy failure. Of the patients with MLA-positive samples, 90% had complete and 10% had partial therapy failure. MPA and MLA results show significant association. Statistical analysis and predominance of partial therapy failure in MPA-positive patients demonstrate higher sensitivity of MPA. With its methodologic advantages, its test parameter being more relevant to BT-A therapy, and its higher sensitivity, the MPA appears to be superior to the MLA.