Adverse events (AEs) were monitored from the time of informed consent to the end of the study. All AEs were elicited by direct, non-leading questioning or by spontaneous reports.

Safety Population - all enrolled subjects. One subject received placebo in the DB study and two treatment cycles of Dysport in the OL study. However, as both Dysport treatments were outside of the ranges specified (≤7.5 U/kg in Treatment Cycle 1 and Treatment Cycle 2), the subject was excluded from the analysis (no TEAEs reported for this subject).

Baseline for the 'change from DB baseline' was defined as the baseline of Study 141 for all treatment cycles. The Modified Ashworth Scale (MAS) is a 6-point scale which measures the intensity of muscle tone by measuring the resistance of the muscle to passive lengthening or stretching. The investigator graded muscle tone in the GSC from 0 (no increase in tone) to 4 (affected parts rigid in flexion or extension).

Intent-to-treat (ITT) population where n represents number of subjects with data. For Dysport 10 U/kg and 15 U/kg, the actual administered doses in the GSC of the (most) affected leg were >7.5 to ≤12.5 U/kg and >12.5 to ≤17.5 U/kg, respectively, in the OL study. The analysis excluded subjects with doses ≤7.5 or >17.5 U/kg.

Baseline was defined as the value obtained prior to the first injection in the hamstrings. The MAS is a 6-point scale which measures the intensity of muscle tone by measuring the resistance of the muscle to passive lengthening or stretching. The investigator graded muscle tone in the GSC from 0 (no increase in tone) to 4 (affected parts rigid in flexion or extension).

ITT population where n represents number of subjects with data. For Dysport 5 U/kg and 10 U/kg, the actual administered doses in the hamstring of the (most) affected leg were >3 to ≤7.5 U/kg and >7.5 to ≤12.5 U/kg, respectively. Analysis excluded subjects with doses ≤3 or >12.5 U/kg. Only data from subjects injected in the hamstring are presented.

Baseline was defined as the value obtained prior to the first injection in the upper limb(s). The MAS is a 6-point scale which measures the intensity of muscle tone by measuring the resistance of the muscle to passive lengthening or stretching. The investigator graded muscle tone in the GSC from 0 (no increase in tone) to 4 (affected parts rigid in flexion or extension). No subjects were treated in the upper limb in Treatment Cycle 4.

ITT population where n represents number of subjects with data. Only data from subjects injected in the upper limb muscle groups are presented. TC = Treatment Cycle.

Global assessment of treatment response based on changes since the first injection in the DB study. PGA Scale of the Treatment Response: Global assessment of treatment response was assessed by asking the Investigator the following question: "how would you rate the response to treatment in the subject's lower limb(s) since the first injection in the DB study?" Answers were made on a 9 point rating scale (-4: markedly worse, -3: much worse, -2: worse, -1: slightly worse, 0: no change, +1: slightly improved, +2: improved, +3: much improved, +4: markedly improved).

ITT population where n represents number of subjects with data. For Dysport 10 U/kg and 15 U/kg, the actual administered doses in the GSC of the (most) affected leg were >7.5 to ≤12.5 U/kg and >12.5 to ≤17.5 U/kg, respectively. The analysis excluded subjects with doses ≤7.5 or >17.5 U/kg.

Individual goals were defined prior to treatment in each treatment period. The GAS is a functional scale used to measure progress towards individual therapy goals. Individual goals were defined for each subject by the physician, and the child's parents (caregiver) where applicable, prior to treatment. After treatment in each treatment cycle, the GAS for each goal was rated using a defined scale (-2: Much less than expected outcome, -1: Somewhat less than expected outcome, 0: Expected outcome, 1: Somewhat more than expected outcome, and 2: Much more than expected outcome).

ITT population where n represents number of subjects with data. For Dysport 10 U/kg and 15 U/kg, the actual administered doses in the GSC of the (most) affected leg were >7.5 to ≤12.5 U/kg and >12.5 to ≤17.5 U/kg, respectively. The analysis excluded subjects with doses ≤7.5 or >17.5 U/kg.

The TS was used to measure spasticity in the GSC at the ankle joint of the (most) affected lower limb.

The Investigator assessed muscle reactions of the tested muscle to passive stretch at two velocities: SLOW = V1: as slow as possible (slower than the rate of natural drop of the limb segment under gravity); FAST = V2 (speed of the limb segment falling under gravity) or V3 (as fast as possible - faster than the rate of natural drop of the limb segment under gravity).

The mean change from DB baseline in XV1 at slow speed was derived.

ITT population where n represents number of subjects with data. For Dysport 10 U/kg and 15 U/kg, the actual administered doses in the GSC of the (most) affected leg were >7.5 to ≤12.5 U/kg and >12.5 to ≤17.5 U/kg, respectively. The analysis excluded subjects with doses ≤7.5 or >17.5 U/kg.

The TS was used to measure spasticity in the GSC at the ankle joint of the (most) affected lower limb.

The Investigator assessed muscle reactions of the tested muscle to passive stretch at two velocities: SLOW = V1: as slow as possible (slower than the rate of natural drop of the limb segment under gravity); FAST = V2 (speed of the limb segment falling under gravity) or V3 (as fast as possible - faster than the rate of natural drop of the limb segment under gravity).

The mean change from DB baseline in XV3 at fast speed was derived.

ITT population where n represents number of subjects with data. For Dysport 10 U/kg and 15 U/kg, the actual administered doses in the GSC of the (most) affected leg were >7.5 to ≤12.5 U/kg and >12.5 to ≤17.5 U/kg, respectively. The analysis excluded subjects with doses ≤7.5 or >17.5 U/kg.

The TS was used to measure spasticity in the GSC at the ankle joint of the (most) affected lower limb.

The Investigator assessed muscle reactions of the tested muscle to passive stretch at two velocities: SLOW = V1: as slow as possible (slower than the rate of natural drop of the limb segment under gravity); FAST = V2 (speed of the limb segment falling under gravity) or V3 (as fast as possible - faster than the rate of natural drop of the limb segment under gravity).

X (threshold) was derived as XV1 at slow speed minus XV3 at fast speed and the mean change from DB baseline was calculated.

ITT population where n represents number of subjects with data. For Dysport 10 U/kg and 15 U/kg, the actual administered doses in the GSC of the (most) affected leg were >7.5 to ≤12.5 U/kg and >12.5 to ≤17.5 U/kg, respectively. The analysis excluded subjects with doses ≤7.5 or >17.5 U/kg.

The mean change from baseline (in the DB study) in Y was derived from the TS.

Y was graded according to the following scale: Grade 0 - no resistance throughout passive movement (best outcome); Grade 1 - slight resistance throughout passive movement; Grade 2 - clear catch at precise angle, interrupting passive movement, followed by release; Grade 3 - fatigable clonus (less than 10 sec when maintaining pressure) occurring at a precise angle, followed by release; Grade 4 - unfatigable clonus (more than 10 seconds when maintaining pressure) occurring at a precise angle; Grade 5 - joint immovable (worst outcome).

Catch without release was graded 0 if XV1=XV3, 'unratable' spasticity otherwise; catch with 'minimal' release was graded 2 if XV3 was consistent and consistently less than XV1.

Angle 0 = position of minimal stretch of the tested muscle. For Grades 0 and 1, spasticity angle X = 0 by definition.

ITT population where n represents number of subjects with data. For Dysport 10 U/kg and 15 U/kg, the actual administered doses in the GSC of the (most) affected leg were >7.5 to ≤12.5 U/kg and >12.5 to ≤17.5 U/kg, respectively. The analysis excluded subjects with doses ≤7.5 or >17.5 U/kg.

The TS was used to measure spasticity in the knee flexors at the knee joint of the (most) affected lower limb.

The Investigator assessed muscle reactions of the tested muscle to passive stretch at two velocities: SLOW = V1: as slow as possible (slower than the rate of natural drop of the limb segment under gravity); FAST = V2 (speed of the limb segment falling under gravity) or V3 (as fast as possible - faster than the rate of natural drop of the limb segment under gravity).

The mean change from baseline (prior to the first injection cycle in the hamstrings) in XV1 at slow speed was derived.

ITT population where n represents number of subjects with data. For Dysport 5 U/kg and 10 U/kg, the actual administered doses in the hamstring of the (most) affected leg were >3 to ≤7.5 U/kg and >7.5 to ≤12.5 U/kg, respectively. Analysis excluded subjects with doses ≤3 or >12.5 U/kg. Only data from subjects injected in the hamstring are presented.

The TS was used to measure spasticity in the knee flexors at the knee joint of the (most) affected lower limb.

The Investigator assessed muscle reactions of the tested muscle to passive stretch at two velocities: SLOW = V1: as slow as possible (slower than the rate of natural drop of the limb segment under gravity); FAST = V2 (speed of the limb segment falling under gravity) or V3 (as fast as possible - faster than the rate of natural drop of the limb segment under gravity).

The mean change from baseline (prior to the first injection cycle in the hamstrings) in XV3 at fast speed was derived.

ITT population where n represents number of subjects with data. For Dysport 5 U/kg and 10 U/kg, the actual administered doses in the hamstring of the (most) affected leg were >3 to ≤7.5 U/kg and >7.5 to ≤12.5 U/kg, respectively. Analysis excluded subjects with doses ≤3 or >12.5 U/kg. Only data from subjects injected in the hamstring are presented.

The TS was used to measure spasticity in the knee flexors at the knee joint of the (most) affected lower limb.

The Investigator assessed muscle reactions of the tested muscle to passive stretch at two velocities: SLOW = V1: as slow as possible (slower than the rate of natural drop of the limb segment under gravity); FAST = V2 (speed of the limb segment falling under gravity) or V3 (as fast as possible - faster than the rate of natural drop of the limb segment under gravity).

X (threshold) was derived as XV1 at slow speed minus XV3 at fast speed and the mean change from baseline (prior to the first injection cycle in the hamstrings) was calculated.

ITT population where n represents number of subjects with data. For Dysport 5 U/kg and 10 U/kg, the actual administered doses in the hamstring of the (most) affected leg were >3 to ≤7.5 U/kg and >7.5 to ≤12.5 U/kg, respectively. Analysis excluded subjects with doses ≤3 or >12.5 U/kg. Only data from subjects injected in the hamstring are presented.

The mean change from baseline (prior to the first injection cycle in the hamstrings) in Y was derived from the TS.

Y was graded according to the following scale: Grade 0 - no resistance throughout passive movement (best outcome); Grade 1 - slight resistance throughout passive movement; Grade 2 - clear catch at precise angle, interrupting passive movement, followed by release; Grade 3 - fatigable clonus (less than 10 sec when maintaining pressure) occurring at a precise angle, followed by release; Grade 4 - unfatigable clonus (more than 10 seconds when maintaining pressure) occurring at a precise angle; Grade 5 - joint immovable (worst outcome).

Catch without release was graded 0 if XV1=XV3, 'unratable' spasticity otherwise; catch with 'minimal' release was graded 2 if XV3 was consistent and consistently less than XV1.

Angle 0 = position of minimal stretch of the tested muscle. For Grades 0 and 1, spasticity angle X = 0 by definition.

ITT population where n represents number of subjects with data. For Dysport 5 U/kg and 10 U/kg, the actual administered doses in the hamstring of the (most) affected leg were >3 to ≤7.5 U/kg and >7.5 to ≤12.5 U/kg, respectively. Analysis excluded subjects with doses ≤3 or >12.5 U/kg. Only data from subjects injected in the hamstring are presented.

The OGS is a measurement tool used to objectively quantify positive and negative features (impairments) of the upper motor neurone syndrome. The OGS is useful when children are too young or insufficiently cooperative for instrumented gait analysis. It is based on the Physicians Rating Scale but has some modifications to improve its sensitivity to detect changes following administration of Botulinum Toxin Type A (BTX-A).

The OGS total score was calculated as the sum of the individual question scores for Questions 1 to 7, with the highest possible score being 20. The parameters collected were: knee position in midstance, initial foot contact, foot contact at midstance, timing of heel raise, hindfoot at midstance, base of support and gait assistive devices. Higher scores indicate better gait.

The mean change from baseline (in the DB study) in the OGS total score of the (most) affected leg was derived.

ITT population where n represents number of subjects with data. For Dysport 10 U/kg and 15 U/kg, the actual administered doses in the GSC of the (most) affected leg were >7.5 to ≤12.5 U/kg and >12.5 to ≤17.5 U/kg, respectively. The analysis excluded subjects with doses ≤7.5 or >17.5 U/kg.

The PedsQL has a disease specific CP module that is relevant to the study population and complements the core modules.

The 35-item questionnaire encompassed 7 scales including (1) daily activities (2) school activities (3) movement and balance (4) pain and hurt (5) fatigue (6) eating activities and (7) speech and communication. A 5-point scale was utilised for parent proxy-report: 0 = never a problem; 1 = almost never a problem; 2 = sometimes a problem; 3 = often a problem; 4 = almost always a problem.

Each CP score was calculated as follows: (1) Individual item scores were reversed and transformed from a 0-4 scale to a 0-100 scale by assigning 0=100, 1=75, 2=50, 3=25 and 4=0; (2) Each scale score was calculated as the sum of the transformed individual item scores, divided by the number of non-missing items. Higher scores indicated better quality of life (fewer symptoms or problems).

A scale score was only calculated if at least 50% of the associated items were non-missing.

ITT population where n represents number of subjects with data. For Treatment Cycle 4, Week 12 was not included in the study design. Subjects who completed the study or withdrew were counted as missing at subsequent visits.

The PedsQL is a validated quality of life questionnaire, designed for children from 2 to 18 years of age. The Generic Core Scale covers four multidimensional scales including physical, emotional, social and school aspects, with three summary scales of total scale score, physical health summary score and psychosocial health summary score.

Each generic core scale was calculated as follows: (1) Individual item scores were reversed and transformed from a 0-4 scale to a 0-100 scale by assigning 0=100, 1=75, 2=50, 3=25 and 4=0; (2) Each scale score was calculated as the sum of the transformed individual item scores, divided by the number of non-missing items. Higher scores indicated better quality of life (fewer symptoms or problems).

A scale score was only calculated if at least 50% of the associated items were non-missing.

ITT population where n represents number of subjects with data. For Treatment Cycle 4, Week 12 was not included in the study design. Subjects who completed the study or withdrew were counted as missing at subsequent visits. PHS = pyschosocial health summary; W12 = Week 12.