Content » Vol 51, Issue 10

Short communication


Peter J. McAllister, MD1, Svetlana E. Khatkova, MD2, Steven G. Faux, FAFRM (RACP)3,4, Philippe Picaut, PharmD5, Romain Raymond, MSc6 and Jean-Michel Gracies, MD, PhD7

From the 1New England Institute for Neurology and Headache, Stamford, CT, USA, 2Neurology Department, Federal State Hospital, Treatments and Rehabilitation, Center of Ministry of Health and Social Development of Russian Federation, Moscow, Russia, 3St Vincent’s Hospital, New South Wales, Australia, 4University of New South Wales, New South Wales, Australia, 5Ipsen Pharma, Les Ulis, France, 6Ividata, Levallois-Perret, France, and 7EA 7377 BIOTN, Université Paris-Est, Hospital Albert Chenevier-Henri Mondor, Department of Neurorehabilitation, Créteil, France

*This work was presented in part as a poster at the American Academy of Physical Medicine and Rehabilitation in Denver, CO, USA, 12–15 October 2017, and as an oral presentation at the 10th World Congress for Neurorehabilitation in Powai, Mumbai, India, 7–10 February 2018.


Objective: To compare walking speed in patients with spastic hemiparesis who received abobotulinumtoxinA either in the lower limb or simultaneously in both the lower and upper limbs.

Design: Post hoc analysis from a phase 3 study of abobotulinumtoxinA (Dysport®, NCT01251367).

Patients: Adult patients with spastic hemiparesis causing gait dysfunction.

Methods: Comfortable barefoot walking speed over 10 m was evaluated in 127 patients receiving lower limb vs lower and upper limb injections over ≤4 treat-ment cycles; 1,000 U or 1,500 U in lower limb for cycle 1/2; optional upper limb injections from cycle 3 (500 U: upper limb, 1,000 U: lower limb).

Results: Mean (standard deviation; SD) lower limb cycle 3/4 doses were lower in the lower plus upper limb group vs lower limb only (1,000 U (SD 50), 1,000 U (SD 50) vs 1,380 U (SD 210), 1,360 U (SD 220). Baseline comfortable barefoot walking speed was similar between groups. Changes at cycle 3 week 4, in m/s, were: lower and upper limb: +0.063 (SD 0.131); lower limb only: +0.078 (SD 0.114), and cycle 4 week 4: lower and upper limb: +0.086 (SD 0.166); lower limb only: +0.086 (SD 0.123).

Conclusion: Simultaneous lower and upper limb abobotulinumtoxinA treatment does not hamper improvement in walking speed compared with lower limb treatment alone. Thus, physicians may split the 1,500 U abobotulinumtoxinA dose as needed to best treat patients with spastic paresis.

Key words: muscle spasticity; paresis; botulinum toxin, type A; walking speed, stroke, traumatic brain injury.

Accepted Sep 4, 2019; Epub ahead of print Sep 17, 2019

J Rehabil Med 2019; 51: 00–00

Correspondence address: Peter J. McAllister, New England Institute for Neurology and Headache, 30 Buxton Farm Road, Ste. 230, Stamford, CT 06905, USA. E-mail:

Lay Abstract

Spastic hemiparesis is a condition in which the muscles on one side of the body become weak and stiff, often after a stroke. One treatment involves injecting botulinum toxin (e.g. abobotulinumtoxinA) into the affected muscles. Patients may find walking difficult, but repeated injections of abobotulinumtoxinA into the legs have been associated with improved walking speed. How-ever, it is not known whether improvements in walking speed are maintained when the dose of abobotulinumtoxinA is split to treat other affected areas, such as the arms. This study shows that splitting the dose of abobotulinumtoxinA across the arms and legs does not hamper walking speed, compared with injecting into the legs only. This suggests that doctors can split the dose across the arms and legs as needed to best treat the symptoms of spastic hemiparesis.


Spastic hemiparesis can result from stroke and often involves both the lower (LL) and upper limb (UL) (1). Up to 75% of stroke survivors depend on caregivers for activities of daily living, and impaired gait contributes to decreased independence (2). In a global survey of patients living with spasticity, 72% and 44% of respondents reported decreased quality of life and a loss of independence, respectively (2). Improvement in walking is a common goal for patients, with walking speed shown to be a predictor for reintegration into the community and increased probability of survival (3, 4).

Management of spastic paresis often requires treatment of both the LL and UL. However, limited data are available regarding their simultaneous treatment with botulinum neurotoxin type A (BoNT-A), which is known to improve symptoms in spastic paresis. BoNT-A treatment of the UL may impact on walking ability by improving trunk control and posture (5). In preliminary studies, improved gait and faster walking speed were observed in patients with hemiparesis following stroke and in children with cerebral palsy, following BoNT-A treatment of the affected UL (5–7).

AbobotulinumtoxinA (Dysport®, aboBoNT-A) has proven efficacious in the treatment of both LL and UL spasticity when injected separately (8, 9). Repeated aboBoNT-A injections into the LL or UL have been associated with improvement in both active movement and function, with increased walking speed observed following repeated LL injections (10, 11).

We report here the results of a post hoc analysis of data from a phase 3 study of aboBoNT-A (11). The objective of this analysis was to evaluate walking speed in patients with spastic hemiparesis who received aboBoNTA, comparing outcomes in those patients treated solely in the LL with those receiving simultaneous treatment in both the LL and UL.


Study design and treatment

The phase 3, open-label (OL) extension study of aboBoNT-A
(NCT01251367) followed a double-blind (DB) study (NCT01249404), as described previously (11). Briefly, the DB phase was a multicentre, prospective, randomized, placebo-controlled study in adults with chronic hemiparesis who received a single injection of 1,000 U or 1,500 U aboBoNT-A or placebo into both soleus and gastrocnemius muscles and ≥ 1 other (investigator-selected) LL muscle(s).

The OL extension was based on multiple injections of 1,000 U or 1,500 U aboBoNT-A for ≤4 treatment cycles (TCs), undertaken at 52 centres across 11 countries. At TC1 of the OL phase, patients received aboBoNT-A 1,500 U in the LL, except those who experienced treatment-emergent adverse events (AEs) during the DB phase that the investigator considered to pose an unacceptable risk, who instead received 1,000 U. At TC2, 1,000 U or 1,500 U could be administered in the LL (according to the investigator’s clinical judgement). At TC3 and TC4, either 1,000 U or 1,500 U was administered into the LL; patients with co-existing disabling muscle overactivity in the UL could receive simultaneous injections of up to 500 U in the affected UL (as per investigator’s clinical judgement), with the total aboBoNT-A dose not exceeding 1,500 U. The minimum time between TCs was 12 weeks.

This study was conducted in accordance with the Declaration of Helsinki, International Conference on Harmonisation Good Clinical Practice Guidelines, and local regulatory requirements, with approval from relevant independent ethics committee/institutional review boards. Written informed consent was obtained from patients prior to study entry.

This post hoc analysis evaluates the subgroups of patients who were injected with aboBoNT-A at both TC3 and TC4 of the OL phase and received either simultaneous injections of aboBoNT-A
in the LL and UL, or injections in the LL only (Fig. 1).

Fig. 1. Design of the post hoc analysis. Patient numbers decreased across cycles due to injection intervals of >12 weeks, with a maximum treatment duration of 1 year per patient. AboBoNT-A: abobotulinumtoxinA; LL: lower limb; TC: treatment cycle; U: units; UL: upper limb.


Inclusion criteria for the overall study population are described elsewhere (11) and included the following requirements to enter the DB trial: adults (18–80 years) with spastic hemiparesis causing gait dysfunction; comfortable barefoot walking speed (CBWS) 0.1–0.8 m/s without walking aids or orthotics; at least 6 months after a stroke (ischaemic or haemorrhagic) or traumatic brain injury (TBI). To continue into the OL phase, patients had to have completed follow-up visits in the DB trial without any major protocol deviations and/or ongoing adverse events. Key exclusion criteria were: major limitation in passive range of motion at hip, knee or ankle; known sensitivity to any form of botulinum neurotoxin (BoNT) or aboBoNT-A excipients; pregnancy; severe cognitive impairment that interfered with consent provision.

Outcome measures

Duration of the walk in a 10-m walk test (10MWT) was recorded using a stopwatch. Walking speed outcomes, assessed as change in unassisted CBWS between baseline and week 4 of each TC, were compared between patients who received simultaneous injections in the LL+UL during TC3 and TC4 and patients who received LL injections only throughout the study (Fig. 1).

Statistical analysis

Due to the post hoc nature of these analyses, descriptive statistics only were performed. No hypothesis testing or p-values are presented.


Patient disposition and baseline characteristics

A total of 352 patients entered the OL study; 127 were injected at both TC3 and TC4 and were thus included in these analyses (n = 63 received co-injection in LL+UL; n = 64 received injection in LL only, Fig. 1).

Baseline demographic characteristics were similar between the LL+UL and LL only treatment groups, including cause of spasticity (stroke or traumatic brain injury (TBI)), time since causal event, and whether naïve to any form of BoNT administered to the affected LL (Table I).

Table I. Baseline demographics and characteristics of patients included in the post hoc analysis

Injection doses are shown in Table II. Mean (standard deviation; SD) doses in LL at TC3 and TC4 were lower in patients injected in LL+UL vs LL only (1,006.7 U (SD 57.2), 1,001.1 U (SD 58.3) vs 1,381.8 U (SD 215.4), 1,367.2 U (SD 222.6)). Mean doses in the UL at TC3 and TC4 were 486.9 U (SD 56.1) and 491.0 U (SD 44.5), respectively. Overall LL dose ranges received at TC3 and TC4, respectively, were 937.5–1,500 U and 1,000–1,500 U in LL only and 1,300–1,500 U and 1,000–1,500 U in LL+UL. The most frequently injected UL muscles in TC3 and TC4, respectively, were the flexor digitorum superficialis (61.9% vs 63.5%), the flexor digitorum profundus (52.4% vs 44.4%), and the flexor carpi radialis (42.9% vs 41.3%).

Table II. Doses (units) of aboBoNT-A administered


At baseline, unassisted CBWS (mean) was similar between patients injected in the LL+UL (0.419 m/s (SD 0.195), n = 63) and LL only (0.420 m/s (SD 0.199), n = 64). In the subgroup of patients who received LL injections only at TC3 and TC4, a consistently greater change from baseline in CBWS was observed at week 4 of the DB study through to OL TC2 (mean change, m/s, DB to TC2: 0.072 (SD 0.084) to 0.080 (SD 0.132)), compared with patients who received co-injection into the UL at TC3 and TC4 (DB to TC2: 0.027 (SD 0.095) to 0.060 (SD 0.135); Fig. 2).

At the first TC to include UL injections (TC3), both subgroups had continued improvements from baseline at week 4 (mean change: LL+UL: 0.063 (SD 0.131); LL only: 0.078 (SD 0.114)). Further improvements were observed at week 4 of TC4 in both subgroups (LL+UL: 0.086 (SD 0.166); LL only: 0.086 (SD 0.123)). These values equate to improvements of 20.8% and 19.8% from baseline in unassisted CBWS in the LL+UL and LL group, respectively.

Fig. 2. Mean change in comfortable barefoot walking speed (CBWS) at week 4 of each treatment cycle (TC) for patients who received lower limb (LL)+ upper limb (UL) injections or LL only injections at both TC3 and TC4 (doses combined). CBWS was assessed using the 10-m walk test (10MWT). DB: double-blind.


In patients with spastic hemiparesis treated with
aboBoNT-A, walking speed improved from the first injection and continued with repeated treatment through-out the OL phase. For patients requiring simultaneous treatment of muscle overactivity in the LL and UL, the total aboBoNT-A dose of 1,500 U split between LL and UL provided similar improvements in walking speed to those observed in patients injected in the LL only.

Retrospectively, the observed mean change in CBWS during the DB phase was markedly different between groups, with patients in the LL+UL group displaying less improvement than the group who did not require UL injections and received LL injections only (Fig. 2). In patients requiring concomitant UL injection, muscle overactivity in the UL may have had an impact on gait (5); thus in the LL+UL group, LL injections alone may not have been sufficient to overcome this overactivity. This may have explained the smaller improvements in CBWS compared with the LL only group prior to TC3. When UL injection became available, further improvements in CBWS were observed in this group, eventually reaching the same level as the LL only group by TC4.

Treatment of muscle overactivity in the UL with BoNT-A to improve gait has been associated with walking speed improvement in previous, smaller studies in patients with hemiparesis due to upper motor neurone syndrome or following a stroke (5, 12). However, these studies focused on specific UL muscles that were expected to improve walking speed by improving UL symmetry of movement in gait and speed, whereas this study included all UL muscles. This may explain why patients with simultaneous injections showed less improvement than those injected only in the LL, and could suggest that specific UL muscles need to be targeted to improve walking speed and gait. This analysis provides initial data on split dosing of aboBoNT-A when injecting LL and UL simultaneously, and on the muscles injected, for an efficacious response. Importantly, data show that a 1,500 U dose of aboBoNT-A
can be either injected entirely into the LL, or split between the UL and LL, as needed by the physician to enable best treatment of each patients’ unique presentation of spastic paresis and treatment goals.

The post hoc exploratory nature of this analysis limits the extrapolation of these data, as the OL extension phase of the study was not powered for the statistical comparison of walking speed in patients who received aboBoNT-A either in the LL or both the LL and UL simultaneously (11). In addition, this study enrolled patients who were physically able to walk 10 m unaided; therefore, the results may not be generalizable to patients who walk permanently with orthotics or aids. Despite this, the results presented here provide an additional insight into the treatment of spastic hemiparesis. Further efficacy data is currently being prospectively collected from a larger patient group of adults with hemiparesis after stroke or TBI, following a 1,500 U dose of aboBoNT-A split between LL and UL, in the global ENGAGE study (NCT02969356).


The authors thank all patients involved in the study, as well as their caregivers, care team, investigators and research staff in participating institutions. The authors acknowledge Jovita Balcaitiene, former employee of Ipsen, for contributing to the conception of the analyses.

Funding. The phase 3 study and post hoc analysis were funded by Ipsen.

PM served on the speakers’ bureau for Ipsen, Allergan and Merz, and received grant support from Allergan. SK and SF have no interests to disclose. PP is an employee of Ipsen. RR is an employee of Ividata subcontracted to Ipsen at the time of manuscript development. JMG has served as a consultant for and received research grant support from Allergan, Ipsen and Merz.

  1. Carmo AA, Kleiner AFR, Costa PHLd, Barros RML. Three-dimensional kinematic analysis of upper and lower limb motion during gait of post-stroke patients. Braz J Med Biol Res 2012; 45: 537–545.
    View article    Google Scholar
  2. Barnes M, Kocer S, Murie Fernandez M, Balcaitiene J, Fheodoroff K. An international survey of patients living with spasticity. Disabil Rehabil 2017; 39: 1428–1434.
    View article    Google Scholar
  3. Eng JJ, Tang PF. Gait training strategies to optimize walking ability in people with stroke: a synthesis of the evidence. Expert Rev Neurother 2007; 7: 1417–1436.
    View article    Google Scholar
  4. Bohannon R, Andrews A, Smith M. Rehabilitation goals of patients with hemiplegia. Int J Rehabil Res 1988; 11: 181–183.
    View article    Google Scholar
  5. Esquenazi A, Mayer N, Garreta R. Influence of botulinum toxin type A treatment of elbow flexor spasticity on hemiparetic gait. Am J Phys Med Rehabil 2008; 87: 305–310; quiz 311, 329.
    View article    Google Scholar
  6. Falso M, Galluso R, Malvicini A. Functional influence of botulinum neurotoxin type A treatment (Xeomin(R)) of multifocal upper and lower limb spasticity on chronic hemiparetic gait. Neurol Int 2012; 4: e8.
    View article    Google Scholar
  7. Hefter H, Rosenthal D. Improvement of upper trunk posture during walking in hemiplegic patients after injections of botulinum toxin into the arm. Clin Biomech (Bristol, Avon) 2017; 43: 15–22.
    View article    Google Scholar
  8. Dashtipour K, Chen JJ, Walker HW, Lee MY. Systematic literature review of abobotulinumtoxinA in clinical trials for adult upper limb spasticity. Am J Phys Med Rehabil 2015; 94: 229–238.
    View article    Google Scholar
  9. Dashtipour K, Chen JJ, Walker HW, Lee MY. Systematic literature review of abobotulinumtoxinA in clinical trials for lower limb spasticity. Medicine (Baltimore) 2016; 95: e2468.
    View article    Google Scholar
  10. Gracies JM, O’Dell M, Vecchio M, Hedera P, Kocer S, Rudzinska-Bar M, et al. Effects of repeated abobotulinumtoxinA injections in upper limb spasticity. Muscle Nerve 2018; 57: 245–254.
    View article    Google Scholar
  11. Gracies JM, Esquenazi A, Brashear A, Banach M, Kocer S, Jech R, et al. Efficacy and safety of abobotulinumtoxinA in spastic lower limb: randomized trial and extension. Neurology 2017; 89: 2245–2253.
    View article    Google Scholar
  12. Hirsch MA, Westhoff B, Toole T, Haupenthal S, Krauspe R, Hefter H. Association between botulinum toxin injection into the arm and changes in gait in adults after stroke. Mov Disord 2005; 20: 1014–1020.
    View article    Google Scholar


Do you want to comment on this paper? The comments will show up here and if appropriate the comments will also separately be forwarded to the authors. You need to login/create an account to comment on articles. Click here to login/create an account.