Amendment 1: -Corrected typographical error regarding the maximum dose allowed in the 3 U/kg group to read “not to exceed 100 U” -Clarified that in the primary efficacy analyses, if a pairwise comparison in MAS-B was not statistically significant, the corresponding pairwise comparison in CGI by Physician would not be considered statistically significant regardless of the actual p value -Added clarification of when dose effectiveness would be concluded to address comments by US FDA (United States Food and Drug Administration) during the Special Protocol Assessment review -Clarified the ideal order in the timing of the CGI by Physician and spasticity assessments -Revised the question for the Patient-reported Onset of Spasticity Symptom Relief assessment from “have you noticed any treatment effect” to “have you noticed any effect” to minimize potential bias

Amendment 2: -Added distinction between US FDA and non-US FDA clinical hypotheses, measures, and analyses -Specified that the C-SSRS (Columbia Suicide Severity Rating Scale) was to be performed as a safety measure for participants ≥ 6 years of age at Day 1, and provided description of scale, data handling, and reference information -Added participant-reported benefit of injection question -Modified Exclusion regarding seizure frequency for exclusion -Modified Exclusion regarding vulnerable respiratory state -Added Exclusion to exclude participants with significant risk of suicide from treatment -Added a statement that participants may have stayed in the study even if a prohibited medication was administered -Clarified that school-based therapy, if relevant per local legislation, was permitted during the study -Added “or EMG” to the devices for muscle localization techniques -The original multiple testing procedure (Fisher’s Protected Testing procedure) was changed to a gatekeeping procedure to control type I error rate -Pairwise comparison for high dose (BOTOX 6 U/kg) versus low dose (BOTOX 3 U/kg) was removed to incorporate US FDA recommendation -Overall test for among-group comparison was removed since the gatekeeping procedure for the 2 pairwise comparisons was sufficient -Clarified that CGI by Physician was removed for non-US FDA primary analyses and was added for non-US FDA analyses as a secondary measure -Removed subgroup efficacy analyses by type of anesthesia -Revised sample size calculation to base the calculation on a 2-sample t-test -Removed passive range of motion bullet to indicate that assessment of passive range of motion could be performed as part of the MTS using the angle of slow stretch (R2) -Clarified that MAS-B and MTS for finger flexors were to be analyzed for the subgroup of participants with wrist identified as the principal muscle group

Amendment 3: -Number of participants and sample size calculations were revised such that the estimated number of participants needed to complete study decreased from 351 to 213 based on adjusted treatment differences from upper limb studies -The intent-to-treat (ITT) population was replaced with the mITT population based on US FDA recommendation -Text describing sensitivity analyses was edited to spell out the covariate and factors in the MMRM model -Edited text and figure to specify that the principal muscle group should have been the wrist in participants with the same MAS-B scores in the elbow and the wrist at baseline -Added a responder status based on +1 score of CGI by Physician -The sensitivity analyses of MAS-B and CGI were changed to use the MI method for missing values instead of observed cases; sensitivity analyses using last observation carried forward (LOCF) were removed -The primary MAS-B analysis and US FDA coprimary MAS-B and CGI analyses were changed to use MMRM with observed data; ANCOVA with MI and observed data were used as sensitivity analyses -Subgroup analyses of AEs were deleted -Changed the multiple testing procedure (gatekeeping procedure) to the Hochberg procedure for the coprimary analysis for US FDA