DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY ORIGINAL ARTICLE

Severity of cerebral palsy and likelihood of adverse events after

botulinum toxin A injections
CAITLYN M SWINNEY 1 | KAREN BAU 2 | KAREN L OAKLEY BURTON 2 | STEPHEN J O’FLAHERTY 3 |
NATASHA L BEAR 4 | SIMON P PAGET 2
1 University of Sydney, Sydney, NSW; 2 The Children’s Hospital at Westmead, Westmead, NSW; 3 Sydney Children’s Hospital, Randwick, NSW; 4 Child and
Adolescent Health Services, Perth, WA, Australia.
Correspondence to Simon P Paget at Kids Rehab, The Children’s Hospital at Westmead, Locked Bag 4001, Westmead, NSW 2145, Australia. E-mail: simon.paget@health.nsw.gov.au

This article is commented on by Kolaski on pages 440–441 of this issue.

PUBLICATION DATA AIM To determine the incidence of common adverse events after botulinum toxin A (BoNT-
Accepted for publication 15th December A) injections in children with cerebral palsy (CP) and to identify whether the severity of CP
2017. influences the incidence of adverse events.
Published online 16th February 2018. METHOD This was an observational study of patients attending a BoNT-A clinic at a tertiary
paediatric hospital (2010–2014). Data examined included procedural adverse events at the
ABBREVIATION time of injection and at follow-up. Systemic adverse events were defined as lower respiratory
IRR Incidence rate ratio tract illnesses, generalized weakness, dysphagia, and death. Severity of CP was categorized
by the Gross Motor Function Classification System (GMFCS). The relationships between
GMFCS and adverse events were analysed using negative binomial regression models.
RESULTS In total, 591 children underwent 2219 injection episodes. Adverse events were
reported during the procedure (130 [6%] injection episodes) and at follow-up (492 [22%]
injection episodes). There were significantly increased rates of systemic adverse events in
injection episodes involving children in GMFCS level IV (incidence rate ratio [IRR] 3.92 [95%
confidence interval] 1.45–10.57]) and GMFCS level V (IRR 7.37 [95% confidence interval 2.90–
18.73]; p<0.001).
INTERPRETATION Adverse events after BoNT-A injections are common but mostly mild and
self-limiting. Children in GMFCS levels IV and V are at increased risk of systemic adverse
events. The relationship between CP severity and BoNT-A adverse events is complex and
further research is required to better understand this relationship.

Cerebral palsy (CP) is a group of disorders of movement
and posture that cause activity limitation and disability,
and are attributed to a nonprogressive disruption in early
brain development.1 CP is the most common motor dis-
ability in childhood occurring in 2 to 2.5 infants per 1000
live births.2 Spasticity and dystonia (hypertonia) are the
most common types of movement disorder in CP.2 In
addition to motor impairment, children with CP often
have a number of comorbidities, including dysphagia, epi-
lepsy, and cognitive impairment.2,3
Current treatments for hypertonia include physiother-
apy, casting and splinting, oral medications (e.g. baclofen,
benzhexol, levodopa), neurosurgery, and orthopaedic sur-
gery.3 Intramuscular botulinum toxin type A (BoNT-A)
injections decrease hypertonia in the targeted muscle group
by inhibiting the release of acetylcholine at the neuromus-
cular junction and suppressing muscle contraction.4
BoNT-A has a maximal effect by 4 weeks postinjection
and wears off after 3 to 4 months.5
The Gross Motor Function Classification System
(GMFCS) enables classification of children with CP based
on their gross motor function ability.6 In ambulant chil-
dren (GMFCS levels I, II, and III), BoNT-A has been
shown to be effective in improving spastic equinus gait and
walking pattern; hand grasp and upper limb function; and
independence in self-care.4,7–10 Additionally, the treatment
has been shown to be effective in non-ambulant children
(GMFCS levels IV and V) for decreasing pain; ease of
positioning and dressing; and improving hygiene.11–14
Adverse events after BoNT-A administration have been
reported with an incidence of 3% to 23% of injection epi-
sodes, with the majority being transient and mild.15,16 Pre-
viously reported adverse events temporally linked with
BoNT-A injections include pharyngitis, non-specific pain,
respiratory tract infection, vomiting, seizures, and urinary
incontinence.16,17 Children with severe CP (GMFCS levels
IV and V) may be at higher risk of adverse events,18 with
one study reporting adverse events in 5% of injection epi-
sodes in this group.14
Concern has been raised that injected BoNT-A may
spread systemically and trigger adverse events distant from
the site of injection, such as dysphagia, aspiration

498 DOI: 10.1111/dmcn.13686 © 2018 Mac Keith Press

pneumonia, generalized weakness, and death.10,19 These
concerns were heightened after the deaths of four children
in 2005 after BoNT-A injections,20 and a case report
describing the deterioration in the respiratory and oromo-
tor functions of a 9-year-old male (in GMFCS level V)
with repeat BoNT-A injections.21
In 2008, the US Food and Drug Administration released
a black box warning regarding the potential for systemic
spread of BoNT-A leading to respiratory distress and
death.18,22 In 2011, our centre demonstrated severe adverse
events in 2.2% of BoNT-A injections.16 Another centre
reported an incidence of severe adverse events of 4% (all
in GMFCS level V) in a 16-month prospective study,
which prompted them to discontinue BoNT-A treatment
in those children.23 Higher GMFCS level has been found
to be associated with increased unplanned hospital admis-
sion for respiratory complications after BoNT-A injec-
tions.19 The 2010 international consensus statement
recommends cautious dose selection in patients in GMFCS
level V and any patient with dysphagia or breathing
problems.10
Additional information surrounding the frequency of
specific adverse events after BoNT-A injections is required
to enable clinicians and families to make more informed
decisions regarding the intervention and the risks associ-
ated with it. The aim of our study was to identify the inci-
dence and type of adverse events after BoNT-A injections
in a large cohort of children in all GMFCS levels. We also
aimed to investigate whether there is a relationship
between the incidence of specific types of adverse events or
systemic adverse events and GMFCS level.
METHOD
Setting
The study was a review of data collected prospectively
from a purpose-built BoNT-A database and the medical
records of patients attending the Kids Rehab BoNT-A
clinic, an outpatient clinic at The Children’s Hospital at
Westmead, a large tertiary paediatric hospital in Sydney,
Australia.
Participants
All children with a diagnosis of CP who received BoNT-A
injections between July 2010 and June 2014 were included.
Children with other diagnoses were excluded. Guidelines
were used to ensure integrity of diagnosis where there was
any ambiguity. Children with booked appointments who
were not injected were excluded (Fig. 1).
Procedures
Children undergoing BoNT-A injections at Kids Rehab
first undergo an assessment with a multidisciplinary team,
including a paediatric rehabilitation specialist, physiother-
apist, and occupational therapist, to establish goals and
injection sites. The health status of the child in the
month prior and any comorbidities are documented at
this time.
What this paper adds
• Adverse events reported at the time of botulinum toxin A injection occurred
in 6% of injection episodes.
• Adverse events were reported at follow-up in 22% of injection episodes.
• Children in Gross Motor Function Classification System (GMFCS) levels IV
and V have increased rates of systemic adverse events.
• Children in GMFCS levels IV and V report less local weakness and pain.
BoNT-A injections are performed in the same half-day
session using a standard protocol typically involving
nitrous oxide conscious sedation and topical anaesthetic
cream, with the addition of, or substitution with, oral
midazolam in some cases. General anaesthetic was used in
a minority of cases. During the study time-frame, Botox
(Allergan, Irvine, CA, USA) was used in all cases, diluted
to 100U/2mL using 0.9% sodium chloride and injected
under ultrasound guidance, with dosage according to the
child’s weight, size of muscle(s) injected, and previous
doses used. Standard practice is for Botox total dose to not
exceed 16U/kg (up to a maximum of 400U). Adverse
events observed by the injecting team at the time of the
procedure were documented.
Follow-up was performed either in person or via tele-
phone by one of the clinical team. In two cases follow-up
was carried out by e-mail. At this time, health status and
adverse events were recorded. Follow-up was planned to
occur approximately 30 days postinjection, corresponding
to the time at which maximal effect would be expected.
Data from the pro forma were transferred to and stored
on a purpose-built MS Access database.
Assessment appointments
n=2510
Cancelled:
- No goals/BoNT-A
not indicated n=86
- Health reasons n=32
- Other n=173
Injecting episodes
n=2219
Follow-up not conducted:
- Family unable to be
contacted n=61
(2.8% injecting episodes)
Episodes with follow-up
n=2158 (97.3%)
Figure 1: Inclusion flow chart for injecting episodes included in analysis.
BoNT-A, botulinum toxin type A.
Adverse Event Severity after BoNT-A in CP Caitlyn M Swinney et al. 499
Adverse event reporting or changes after classification at an early age). Children
Adverse events were recorded on a ‘per-injection episode’ had a median of four (interquartile range 2–6; range 1–11)
basis, at two time points: at the time of the procedure (by injection episodes during the study time-frame. The mean
direct observation of the injector), and at the time of fol- BoNT-A dose administered was 9.6U/kg (SD 4.6). A small
low-up (by parent or carer report). Systemic adverse events number of children received higher Botox doses than typi-
were defined as any report of generalized weakness, lower cally used: the maximum dose administered for GMFCS
respiratory tract illnesses, dysphagia, or death.10 For all levels I, II, and III was 22.9U/kg and for GMFCS level IV
adverse events requiring treatment or assessment by a med- and V, 21.9U/kg (Botox in all cases).
ical professional, further details including date at onset,
severity and course, treatment, recovery, and associations Adverse events at the time of procedure
were recorded. Medical records were used to clarify any One or more adverse events occurred at the time of the
details missing from the database. procedure at a rate of 5.9 per 100 injection episodes.
There was no statistical evidence of a relationship between
Ethics the overall incidence of procedural adverse events and
The study was granted ethical approval by the Sydney GMFCS level.
Children’s Hospitals Network Human Research Ethics Distress (1.8 per 100 injection episodes), pain (1.7 per
Committee, approval number LNR/14/SCHN/330. 100 injection episodes), and nausea and/or vomiting (1.7
per 100 injection episodes) were the most common types
Statistical analysis of procedural adverse events (Table II). Compared with
The number of participants with adverse events, the num- children in GMFCS level I, children in GMFCS level III
ber of adverse events, and the rate of adverse events per were significantly more likely to show distress during the
100 injections was reported overall and for each GMFCS procedure, and children in GMFCS level V were signifi-
level. The number of adverse events was treated as a count cantly less likely (Table II).
variable and negative binomial regression models were
chosen to manage overdispersion. Zero-inflated models Adverse events at follow-up
were considered but did not offer an advantage over the Follow-up data were obtained from 583 children and 2158
negative binomial model. There were a number of children (97%) of injection episodes. Fifty-eight (10%) of these
who changed GMFCS level during the study and were children changed GMFCS level during the study. Follow-
included as separate entries when this occurred. For this up occurred at a median of 26 days postinjection (in-
reason, a random-effects negative binomial model was cho- terquartile range 21–35d).
sen to account for correlation amongst those children who One or more adverse events were reported at the time of
changed GMFCS level. An exposure variable was included follow-up at a rate of 22.8 per 100 injection episodes
using number of injections per child over the course of the (Table III). There was no statistical evidence of a relation-
study. The GMFCS level was entered into the model as a ship between the incidence of follow-up adverse events and
dummy variable in GMFCS level I as the reference cate- GMFCS level.
gory. To assess the overall contribution of GMFCS the The most common adverse events reported at follow-up
likelihood ratio test was used. Incidence rate ratios (IRRs) were bruising (4.5 per 100 injection episodes), upper respi-
and their 95% confidence intervals (CIs) were produced. ratory tract infections (4.4 per 100 injection episodes),
It was not possible to model bruising at time of injection local weakness (3.7 per 100 injection episodes), pain (3.0
as count data as there were never multiple occurrences per 100 injection episodes), and flu-like illnesses (2.9 per
(only 0 or 1 event per participant). A random-effects logis- 100 injection episodes; Table III). Compared with those
tic regression model was used producing odds ratios (ORs)
and their 95% CIs.
All tests were two-tailed and statistical significance level Table I: Participant details
was set at p<0.05. All statistical analysis was performed Injected Children with
using Stata version 14.1 (StataCorp, College Station, TX, children (n=591) follow-up (n=583)
USA). Participants 328 (55) 321 (55)
GMFCS levela
RESULTS I 162 (25) 161 (25)
II 165 (25) 165 (26)
In total, 591 children and a total of 2219 BoNT-A injec- III 77 (12) 75 (12)
tion episodes were included in the study. Figure 1 IV 112 (17) 109 (17)
describes the inclusion pathway for injections to be consid- V 133 (21) 131 (20)
Dysphagia 178 (30) 169 (29)
ered in the final analysis. Participant demographics are
Gastrostomy 102 (17) 93 (16)
detailed in Table I. Fifty-eight (10%) children changed History of aspiration 54 (9) 54 (9)
GMFCS level during the study time-frame (all changes by pneumonia
one GMFCS level, typically either children who showed Data are n (%). aGMFCS classification changed for some children
gross motor function at the boundaries of GMFCS levels during study. GMFCS, Gross Motor Function Classification System.

500 Developmental Medicine & Child Neurology 2018, 60: 498–504

Table II: Adverse events occurring during the procedure

GMFCS Participants (n) Adverse Rate (per 100

Outcome level with adverse events (%) events (n) injections) IRR (95% CI)a

Adverse event I 24 (15) 33 6.5 1.00 (reference)

occurred during II 19 (12) 23 4.3 0.67 (0.38–1.20)
an injection III 24 (31) 30 11.0 1.72 (0.97–3.06)
episode (≥1) IV 17 (15) 23 5.5 0.86 (0.48–1.56)
V 17 (13) 21 4.3 0.68 (0.37–1.24)
Total 101 (16) 130 5.9
Distress I 10 (6) 10 2.0 1.00 (reference)
II 9 (5) 10 1.9 0.95 (0.40–2.29)
III 12 (14) 13 4.8 2.43 (1.07–5.55)
IV 5 (4) 5 1.2 0.61 (0.21–1.77)
V 2 (1) 2 0.4 0.21 (0.05–0.96)
Total 38 (5) 40 1.8
Pain I 12 (7) 15 2.9 1.00 (reference)
II 5 (3) 5 0.9 0.31 (0.11–0.91)
III 6 (8) 7 2.6 0.85 (0.32–2.28)
IV 5 (5) 5 1.2 0.41 (0.14–1.21)
V 4 (3) 5 1.0 0.35 (0.12–1.02)
Total 32 (5) 37 1.7
Nausea I 6 (4) 6 1.2 1.00 (reference)
and/or II 3 (2) 5 0.9 0.80 (0.21–3.02)
vomiting III 6 (8) 7 2.6 2.31 (0.62–8.71)
IV 8 (7) 13 3.1 2.60 (0.81–8.30)
V 6 (5) 6 1.2 1.11 (0.31–4.04)
Total 29 (5) 37 1.7
Bruising I 6 (4) 6 1.2 1.00 (reference)b
II 2 (1) 2 0.4 0.32 (0.06–1.57)b
III 4 (5) 4 1.5 1.25 (0.35–4.47)b
IV 0 (0) 0 0.7 –
V 3 (3) 3 0.7 0.52 (0.13–2.10)b
Total 15 (2) 15 0.7
Other I 11 (7) 12 2.4 1.00 (reference)
II 12 (7) 13 2.4 1.03 (0.47–2.26)
III 12 (16) 13 4.8 2.03 (0.92–4.44)
IV 7 (6) 7 1.7 0.71 (0.28–1.80)
V 8 (6) 8 1.7 0.70 (0.29–1.72)
Total 50 (7) 53 2.4

Bold denotes statistical significance (p<0.05). aIncident rate ratio (IRR), unless otherwise specified. bOdds ratio (OR) determined using logis-
tic regression. GMFCS, Gross Motor Function Classification System.

involving children in GMFCS level I, there were signifi-
cantly lower rates of local weakness after injection episodes
involving children in GMFCS level V (IRR 0.16; p=0.001)
and significantly lower rates of pain after injection episodes
involving children in GMFCS levels IV and V (IRR
GMFCS level IV: 0.38; GMFCS level V: 0.04 [p<0.001])
(Table III).
Systemic adverse events
Adverse events suggesting systemic spread of BoNT-A
occurred at a rate of 3.6 per 100 injection episodes. Com-
pared with children in GMFCS level I, there was a signifi-
cantly increased rate of systemic adverse events in injection
episodes involving children in GMFCS levels IV (IRR
3.92) and V (IRR 7.37; p<0.001). Rates of generalized
weakness were significantly increased in children in
GMFCS level V (IRR 11.99; p=0.015). Rates of worsened
dysphagia were significantly increased in children in
GMFCS level IV (IRR 5.39; p=0.036) and rates of lower
respiratory tract illnesses were significantly increased in
injection episodes involving children in GMFCS level V
(IRR 12.22; p<0.001; Table III). One child (aged 11.8y,
GMFCS level V) died at home 28 days after BoNT-A; this
child received BoNT-A as part of palliative treatment for
severe intractable dystonia. The death was expected and
the case was not referred to the coroner. A further 17 chil-
dren died during the study time-frame but at times more
remote from BoNT-A injection (median 212d, interquar-
tile range 110–473d).
Medical advice for systemic adverse events was sought
after 42 (2%) injection episodes. Eighteen (3%) children
were hospitalized in the month after BoNT-A injection,
including two with life-threatening illnesses. One child
(aged 8.8y, GMFCS level V) was admitted to hospital,
including a 5-day intensive care admission for adenovirus-
positive lower respiratory tract infection 14 days after
BoNT-A. One child (aged 10.3y, GMFCS level IV) was
admitted for an unexplained episode of decreased level of
consciousness 30 days after BoNT-A in the context of pre-
vious episodes without preceding BoNT-A injection.
DISCUSSION
Our results represent the largest study to date examining
adverse events after BoNT-A injections in children with

Adverse Event Severity after BoNT-A in CP Caitlyn M Swinney et al. 501

Table III: Adverse events reported at follow-up of 2158 injection episodes

Participants (n) with Adverse Rate (per 100

Outcome GMFCS level adverse event (%) event (n) injections) IRR (95% CI)a

Adverse event I 74 (46) 103 20.9 1.00 (reference)

at follow up (≥1) II 87 (53) 138 26.5 1.23 (0.94–1.61)
III 39 (52) 60 22.9 1.17 (0.84–1.62)
IV 55 (51) 82 20.0 1.00 (0.74–1.35)
V 75 (57) 109 23.0 1.07 (0.80–1.42)
Total 330 (52) 492 22.8
Systemic adverse event I 5 (3) 6 1.2 1.00 (reference)
II 7 (4) 8 1.5 1.27 (0.42–3.85)
III 5 (7) 5 1.9 1.68 (0.48–5.94)
IV 13 (12) 19 4.6 3.92 (1.45–10.57)
V 30 (23) 39 8.2 7.37 (2.90–18.73)
Total 60 (9) 77 3.6 LR stat p<0.001
Local weakness I 18 (11) 20 4.1 1.00 (reference)
II 30 (18) 37 7.1 1.75 (0.97–3.14)
III 10 (13) 13 5.0 1.18 (0.55–2.50)
IV 6 (6) 7 1.7 0.42 (0.17–1.02)
V 3 (2) 3 0.6 0.16 (0.05–0.54)
Total 67 (11) 80 3.7 LR test p=0.001
Urinary problems I 6 (4) 9 1.8 1.00 (reference)
II 7 (4) 12 2.3 1.15 (0.39–3.40)
III 4 (5) 4 1.5 1.21 (0.34–4.29)
IV 3 (3) 6 1.5 0.62 (0.16–2.47)
V 2 (2) 2 0.4 0.34 (0.07–1.66)
Total 25 (3) 33 1.5
Pain I 21 (13) 27 5.5 1.00 (reference)
II 19 (12) 22 4.2 0.82 (0.44–1.54)
III 6 (8) 6 2.3 0.42 (0.16–1.09)
IV 8 (7) 8 2.0 0.38 (0.16–0.88)
V 1 (1) 1 0.2 0.04 (0.01–0.30)
Total 55 (9) 64 3.0 LR test p<0.001
Flu-like illness I 14 (9) 14 2.8 1.00 (reference)
II 12 (7) 15 2.9 0.98 (0.46–2.12)
III 7 (9) 8 3.1 1.07 (0.42–2.67)
IV 12 (11) 13 3.2 1.13 (0.51–2.52)
V 12 (9) 13 2.7 0.95 (0.43–2.12)
Total 57 (9) 63 2.9
Bowel problems I 3 (2) 3 0.6 1.00 (reference)
II 11 (7) 13 2.5 3.64 (1.02–13.03)
III 2 (3) 3 1.1 1.32 (0.22–7.86)
IV 3 (3) 5 1.2 1.25 (0.25–6.18)
V 6 (5) 6 1.3 2.08 (0.52–8.33)
Total 25 (4) 30 1.4
URTI I 14 (9) 15 3.0 1.00 (reference)
II 20 (12) 25 4.8 1.52 (0.77–3.01)
III 13 (17) 13 5.0 1.67 (0.76–3.68)
IV 14 (13) 18 4.4 1.42 (0.68–2.96)
V 19 (15) 23 4.9 1.57 (0.79–3.12)
Total 80 (13) 94 4.4
Skin problems I 2 (1) 2 0.4 NA
II 2 (1) 2 0.4
III 0 (0) 0 0
IV 0 (0) 0 0
V 0 (0) 0 0
Total 4 (1) 4
Exacerbation of epilepsy I 3 (2) 4 0.8 1.00 (reference)
II 6 (4) 7 1.3 1.64 (0.43–6.21)
III 0 (0) 0 0 No adverse event
IV 4 (5) 4 1.5 1.19 (0.27–5.33)
V 5 (5) 5 1.2 1.28 (0.31–5.28)
Total 18 (3) 20 0.9
Later bruising I 18 (11) 20 4.1 1.00 (reference)
II 20 (12) 25 4.8 1.17 (0.63–2.15)
III 13 (17) 18 6.9 1.70 (0.87–3.33)
IV 15 (14) 15 3.7 0.92 (0.46–1.84)
V 19 (15) 19 4.0 1.01 (0.53–1.95)
Total 86 (13) 97 4.5

502 Developmental Medicine & Child Neurology 2018, 60: 498–504

Table III: Continued

Participants (n) with Adverse Rate (per 100

Outcome GMFCS level adverse event (%) event (n) injections) IRR (95% CI)a

Later nausea and/or vomiting I 4 (3) 4 0.8 1.00 (reference)

II 9 (6) 10 1.9 2.35 (0.73–7.59)
III 7 (9) 8 3.1 3.77 (1.12–12.72)
IV 5 (5) 5 1.2 1.51 (0.40–5.68)
V 10 (8) 10 2.1 2.63 (0.81–8.50)
Total 35 (6) 37 1.7
Generalized weakness I 1 (1) 1 0.2 1.00 (reference)
II 2 (1) 2 0.4 1.91 (0.15–24.19)
III 0 (0) 0 0.0 No adverse event
IV 2 (3) 2 0.8 2.66 (0.20–34.71)
V 7 (6) 11 2.7 11.99 (1.31–109.84)
Total 12 (2) 16 0.7 LR test p=0.015
Dysphagia I 2 (1) 3 0.6 1.00 (reference)
II 3 (2) 3 0.6 1.04 (0.19–5.68)
III 4 (5) 4 1.5 2.86 (0.54–15.03)
IV 9 (8) 13 3.2 5.39 (1.32–22.07)
V 8 (6) 10 2.1 3.87 (0.93–16.11)
Total 27 (4) 33 1.5 LR test p=0.036
LRTI I 2 (1) 2 0.4 1.00 (reference)
II 3 (2) 3 0.6 1.40 (0.23–8.49)
III 1 (1) 1 0.4 0.95 (0.08–10.56)
IV 4 (4) 4 1.0 2.42 (0.44–13.37)
V 20 (15) 23 4.9 12.22 (2.83–52.66)
Total 30 (5) 33 1.5 LR test p<0.001
Death I 0 (0) 0 0 NA
II 0 (0) 0 0
III 0 (0) 0 0
IV 0 (0) 0 0
V 1 (1) 1 0.2
Total 1 (0) 1 0.04

Bold denotes statistical significance (p<0.05). aIncident rate ratio (IRR), unless otherwise specified. GMFCS, Gross Motor Function Classifi-
cation System; CI, confidence interval; LR, likelihood ratio; URTI, upper respiratory tract infection; NA, not applicable; LRTI, lower respira-
tory tract infection.

CP. We report a rate of adverse events at the time of the
procedure of 5.9 per 100 injection episodes and adverse
events at follow-up of 22.8 per 100 injection episodes. Sys-
temic adverse events were reported at a rate of 3.6 per 100
injection episodes. Our results suggest that children with
severe CP (GMFCS levels IV and V) have a significantly
increased risk of systemic adverse events compared with
children with mild-to-moderate CP (GMFCS levels I, II,
and III).
The rates of procedural and follow-up adverse events we
describe herein are similar to those we have reported previ-
ously.16 Other studies have described a much lower inci-
dence of adverse events.14,15,19 We postulate that these
differences may represent differences in the methodologies
of the studies. The inclusive and non-causal definition of
an adverse event we have applied here and the capture of
adverse events using a formalized assessment may have
contributed to capturing more (minor) adverse events,
which may have been under-reported in retrospective
analysis.
The increased rate of systemic adverse events in children
in GMFCS levels IV and V we describe is supported by
previous reports, which suggest that this group is at an
increased risk of adverse events and of unplanned hospital
admissions after BoNT-A injection.15,19 Similar to previous
reports, the majority of systemic adverse events in our
study were mild, and there was no clear indication that
BoNT-A contributed in the two cases of life-threatening
illness or one death during the follow-up period. We feel
that our results do not support any changes in our BoNT-
A practice in children with severe CP but suggest that clin-
icians and families carefully consider these risks when
deciding on treatment.
The relationship between BoNT-A and adverse events is
complex. Children with severe CP have increased rates of
comorbidities that may predispose them to complications
after BoNT-A injections.24 A recent randomized controlled
trial reported similar rates of moderate-to-severe adverse
events in both BoNT-A treatment and placebo arms.25
What we have described in this study is a temporal associa-
tion between BoNT-A injections and adverse events. For
some adverse events (perhaps many) this relationship may
also be causative; for others, they may be related to the
comorbidities and poor health associated with severe CP
and unrelated to BoNT-A injections.
Children’s baseline functional abilities are also likely to
influence adverse events reporting. In our study, there was
a six-fold increased rate of local weakness and a 25-fold
increased rate of pain in injection episodes involving chil-
dren in GMFCS level I compared with injection episodes

Adverse Event Severity after BoNT-A in CP Caitlyn M Swinney et al. 503

involving children in GMFCS level V. We feel these dif-
ferences are likely to have been influenced by their
increased baseline gross motor function and communica-
tion abilities. Adverse events occurring post BoNT-A are
of varying degrees of importance to clinicians and parents/
guardians. Pain or bruising attributable to the injection
procedure, for example, may represent more acceptable
risks than the risk of generalized weakness, lower respira-
tory tract illnesses, or dysphagia. It is therefore important
to discuss clinically relevant adverse events when discussing
the risks of the intervention with families.
Our study has limitations. No follow-up was available in
3% of the cohort, and the range of follow-up time was
variable. A large interval between the injection episode and
follow-up is likely to have introduced some recall bias.
Relying on parent/guardian reporting of adverse events
also introduces the possibility of reporting bias. We feel it
is likely that we have captured the more serious adverse
events because by their nature systemic adverse events and
adverse events that require medical follow-up are less vul-
nerable to these biases. The use of parent and guardian
reports for adverse events may mean that subjective
experiences of pain and distress were missed in more severe
GMFCS cases, but more intensive monitoring of patient
comfort was not feasible in the present study.
CONCLUSION
This is one of the largest clinical studies to date of adverse
events after BoNT-A injection episodes in children with
CP in all GMFCS levels with follow-up data after injec-
tions in 97%. Our results suggest that while adverse events
after BoNT-A are common, most are minor and self-limit-
ing. Children in GMFCS levels IV and V have an
increased rate of systemic adverse events, making the deci-
sion to administer BoNT-A injections in these children
complex and requiring consideration of risks compared
with benefits. In our department, we have continued to
treat children in GMFCS levels IV and V with BoNT-A,
consistent with the literature supporting the use of BoNT-
A to improve quality of life in these children.10,11,14 Fur-
ther research is required to clarify the relationship between
pre-existing comorbidities in children with CP and the
incidence and severity of adverse events after BoNT-A
injections.

REFERENCES
1. Badawi N, Watson L, Petterson B, et al. What consti-
tutes cerebral palsy? Dev Med Child Neurol 1998; 40:
520–7.
2. Odding E, Roebroeck ME, Stam HJ. The epidemiology
of cerebral palsy: incidence, impairments and risk fac-
tors. Disabil Rehabil 2006; 28: 183–91.
3. Fairhurst C. Cerebral palsy: the whys and hows. Arch
Dis Child Educ Pract Ed 2012; 97: 122–31.
4. Koman LA, Mooney JF 3rd, Smith BP, Walker F, Leon
JM. Botulinum toxin type A neuromuscular blockade in
the treatment of lower extremity spasticity in cerebral
palsy: a randomized, double-blind, placebo-controlled
trial. BOTOX Study Group. J Pediatr Orthop 2000; 20:
108–15.
5. Tilton AH. Evidence-based review of safety and efficacy
in cerebral palsy. Toxicon 2015; 107: 105–8.
6. Palisano R, Rosenbaum P, Walter S, Russell D, Wood
E, Galuppi B. Development and reliability of a system
to classify gross motor function in children with cerebral
palsy. Dev Med Child Neurol 1997; 39: 214–23.
7. Ubhi T, Bhakta BB, Ives HL, Allgar V, Roussounis SH.
Randomised double blind placebo controlled trial of the
effect of botulinum toxin on walking in cerebral palsy.
Arch Dis Child 2000; 83: 481–7.
8. Autti-R€am€
o I, Larsen A, Taimo A, von Wendt L. Man-
agement of the upper limb with botulinum toxin type A
in children with spastic type cerebral palsy and acquired
brain injury: clinical implications. Eur J Neurol 2001; 8
(Suppl. 5): 136–44.
9. Colovic H, Dimitrijevic L, Stankovic I, Nikolic D,
Radovic-Janosevic D. Estimation of botulinum toxin
type A efficacy on spasticity and functional outcome in
children with spastic cerebral palsy. Biomed Pap Med Fac
Univ Palacky Olomouc Czech Repub 2012; 156: 41–7.
10. Love SC, Novak I, Kentish M, et al. Botulinum toxin
assessment, intervention and after-care for lower limb
spasticity in children with cerebral palsy: international
consensus statement. Eur J Neurol 2010; 17(Suppl. 2):
9–37.
11. Copeland L, Edwards P, Thorley M, et al. Botulinum
toxin A for nonambulatory children with cerebral palsy:
a double blind randomized controlled trial. J Pediatr
2014; 165: 140–6. e4
12. Lundy CT, Doherty GM, Fairhurst CB. Botulinum
toxin type A injections can be an effective treatment for
pain in children with hip spasms and cerebral palsy. Dev
Med Child Neurol 2009; 51: 705–10.
13. Pin TW, Elmasry J, Lewis J. Efficacy of botulinum
toxin A in children with cerebral palsy in Gross Motor
Function Classification System levels IV and V: a sys-
tematic review. Dev Med Child Neurol 2013; 55: 304–13.
14. Mesterman R, Gorter JW, Harvey A, et al. Botulinum
toxin type A in children and adolescents with severe
cerebral palsy: a retrospective chart review. J Child Neu-
rol 2014; 29: 210–3.
15. Papavasiliou AS, Nikaina I, Foska K, Bouros P, Mitsou
G, Filiopoulos C. Safety of botulinum toxin A in chil-
dren and adolescents with cerebral palsy in a pragmatic
setting. Toxins 2013; 5: 524–36.
16. O’Flaherty SJ, Janakan V, Morrow AM, Scheinberg
AM, Waugh MC. Adverse events and health status fol-
lowing botulinum toxin type A injections in children
with cerebral palsy. Dev Med Child Neurol 2011; 53:
125–30.
17. Albavera-Hernandez C, Rodrıguez JM, Idrovo AJ. Safety
of botulinum toxin type A among children with spastic-
ity secondary to cerebral palsy: a systematic review of
randomized clinical trials. Clin Rehabil 2009; 23: 394–
407.
18. Narayanan UG. Botulinum toxin: does the black box
warning justify change in practice? Dev Med Child Neurol
2011; 53: 101–2.
19. Naidu K, Smith K, Sheedy M, Adair B, Yu X, Graham
HK. Systemic adverse events following botulinum toxin
A therapy in children with cerebral palsy. Dev Med Child
Neurol 2010; 52: 139–44.
20. Apkon SD, Cassidy D. Safety considerations in the use
of botulinum toxins in children with cerebral palsy. PM
R 2010; 2: 282–4.
21. Howell K, Selber P, Graham HK, Reddihough D. Botu-
linum neurotoxin A: an unusual systemic effect. J Paedi-
atr Child Health 2007; 43: 499–501.
22. Kuehn B. FDA requires black box warnings on labeling
for botulinum toxin products. JAMA 2009; 301: 2316.
23. Blaszczyk I, Foumani NP, Ljungberg C, Wiberg M.
Questionnaire about the adverse events and side effects
following botulinum toxin A treatment in patients with
cerebral palsy. Toxins 2015; 7: 4645–54.
24. Shevell MI, Dagenais L, Hall N; REPACQ Consortium.
Comorbidities in cerebral palsy and their relationship to
neurologic subtype and GMFCS level. Neurology 2009;
72: 2090–6.
25. Edwards P, Sakzewski L, Copeland L, et al. Safety of
botulinum toxin type A for children with nonambulatory
cerebral palsy. Pediatrics 2015; 136: 895–904.

504 Developmental Medicine & Child Neurology 2018, 60: 498–504

 DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY ORIGINAL ARTICLE

RESUMEN


SEVERIDAD DE LA PARALISIS CEREBRAL Y PROBABILIDAD DE EVENTOS ADVERSOS DESPU

ES DE LAS INYECCIONES DE TOXINA
BOTUL
INICA A

OBJETIVO Determinar la incidencia de eventos adversos comunes despue
s de las inyecciones de toxina botul
ınica A (BoNT-A) en
~ os con para
nin
lisis cerebral (PC) e identificar si la severidad en la presentacio
n clinca de la PC influye en la incidencia de eventos
adversos.
M
ETODO Este fue un estudio observacional de pacientes que asist
ıan a una cl
ınica de BoNT-A en un hospital pedia
trico terciario
(2010-2014). Los datos examinados incluyeron los eventos adversos en el momento de la inyeccio
n y en el seguimiento. Los
eventos adversos siste
micos se definieron como enfermedades del tracto respiratorio inferior, debilidad generalizada, disfagia y
muerte. La gravedad de la PC se clasifico
segu
n de la Funcio

n el Sistema de Clasificacio
n Motora Gruesa (GMFCS). Las relaciones
entre el GMFCS y los eventos adversos se analizaron usando modelos de regresio
n binomial negativa.
RESULTADOS En total, 591 nin~ os se sometieron a 2.219 episodios de inyeccio
n. Eventos adversos reportados durante el
procedimiento (130 [6%] episodios de inyeccio
n) y en el seguimiento (492 [22%] episodios de inyeccio
n). Hubo un aumento
significativo en las tasas de eventos adversos siste
micos en los episodios de inyeccio
n en nin
~ os con un nivel IV del GMFCS (tasa
de incidencia [IRR] 3,92 [intervalo de confianza del 95% 1,45-10,57] y GMFCS nivel V (IRR 7,37 [intervalo de confianza del 95% 2,90
-18,73]; p <0,001).
INTERPRETACION
Los eventos adversos despue
s de las inyecciones de BoNT-A son comunes, pero en su mayor
ıa leves y
autolimitados. Los nin ~ os en los niveles IV y V de GMFCS tienen un mayor riesgo de eventos adversos siste
micos. La relacio

n
entre la severidad de PC y los eventos adversos de BoNT-A es compleja y se requieren investigaciones adicionales para
comprender mejor esta relacio
n.

RESUMO

INJEC
SEVERIDADE DA PARALISIA CEREBRAL E PROBABILIDADE DE EVENTOS ADVERSOS APOS ~ DE TOXINA BOTUL
INICA A
ß OES

OBJETIVO Determinar a incide^ncia de efeitos adversos comuns apo
s injecßo~ es de toxina botul
ınica A (BoNT-A) em criancßas com
^ncia de eventos adversos.
paralisia cerebral (PC) e identificar se a severidade da PC influencia a incide
M

ETODO Este foi um estudo observacional de pacientes que frequentavam uma cl
ınica de BoNT-A em um hospital pedia
trico

rio (2010-2014). Os dados examinados inclu
ıram eventos adversos procedurais no momento da injecßa
tercia ~ o e no seguimento.
Eventos adversos siste ^ micos foram definidos como doencßas do trato respirato
rio inferior, fraqueza generalizada, disfagia, e morte.
A severidade da PC foi categorizada pelo Sistema de Classificacßa ~o da Funcßa~o Motora Grossa (GMFCS). A relacßa ~o entre GMFCS e
eventos adversos foi analisada usando modelos de regresssa ~o binomiais negativos.
RESULTADOS No total, 591 criancßas passaram por 2219 episo
dios de injecßa~o. Eventos adversos foram relatados durante o
procedimento (130 [6%] dos episo ~o) e no seguimento (492 [22%] de episo

dios de injecßa
dios de injecßa~o). Houve taxas
significativamente aumentadas de eventos adversos siste ^ micos nos episo
dios de injecßa~ o envolvendo criancßas no n
ıvel GMFCS IV
~o da taxa de incide
(raza ^ncia [RTI] 3,92 [intervalo de confiancßa 95% 1,45-10,57]) e GMFCS V (RTI 7,37 [intervalo de confiancßa 95%
2,90-18,73]; p<0,001).
INTERPRETAC ~ Eventos adversos apo
s injecßo~ es de BoNT-A sa~o comuns, mas principalmente leves e auto-limitantes. Criancßas
ß AO
dos n
ıveis GMFCS IV e V te ^m risco aumentado de eventos adversos siste ^micos. A relacßa ~o entre severidade da PC e eventos
adversos apo
s BoNT-A e
complexa, e futuras pesquisas sa ~o necessa
rias para compreender melhor esta relacßa ~o.