Functional Benefits of Multisite Upper Limb Surgery in Children With Unilateral Cerebral Palsy (RETRO-CHIRc)

Background and Rationale Cerebral palsy (CP) is the most common cause of neurological disability in children, with a prevalence of approximately 2-2.5 per 1,000 live births in Europe. It results from irreversible lesions of the developing brain, causing lifelong motor and postural disorders often accompanied by cognitive or sensory difficulties. Hemiplegia accounts for approximately 30% of CP cases and predominantly affects the upper limb. Spasticity and dyskinesia impair unimanual capacities (range of motion, dexterity, fluidity) as well as bimanual performance.

Focal treatment with botulinum toxin injections combined with occupational therapy has demonstrated efficacy for targeted functional goals. However, the functional effects of upper limb surgery in the context of CP remain poorly documented, with existing systematic reviews highlighting the weakness of current evidence.

Neuro-orthopedic upper limb surgery in children with CP is individualized and multisite, combining several procedures tailored to each patient's functional profile and goals. These may include: short adductor of thumb release, flexor carpi ulnaris release or transfer to extensor carpi radialis brevis, finger flexor lengthening, round pronator release, elbow flexor lengthening, wrist arthrodesis, and others. At the Centre des Massues, all patients undergo a pre-surgical evaluation and a multidisciplinary surgical planning meeting (Équipe Médico-Chirurgicale) promoting shared decision-making, followed by specialized neuro-orthopedic surgery and structured post-operative rehabilitation.

A preliminary retrospective analysis of 23 patients at the Centre des Massues demonstrated that this integrated care pathway improved unimanual upper limb function and patient satisfaction. However, the cohort required expansion to better characterize bimanual performance outcomes and to identify functional subgroups with more homogeneous surgical profiles.

Study Design This is a retrospective non-interventional study (Research Not Involving Human Subjects - RNIPH) using a before/after design in which each patient serves as their own control. All data are collected retrospectively from medical records and routine clinical assessments. Longitudinal data collected during standard care up to 3-5 years before surgery and 1-2 years after surgery are used to establish each patient's individual baseline trajectory and compare it to their post-operative evolution. This design controls for the high interindividual variability inherent to this population and renders a separate control group unnecessary.

Data are collected retrospectively from medical records and the existing occupational therapy database by Audrey COMBEY (occupational therapist). The database will then be locked and transmitted for statistical analysis to Julie ROZAIRE (research engineer).

Data Management and Confidentiality Data are pseudonymized using a participant code (inclusion number + 2 initials). The linkage table is maintained separately by the principal investigator and authorized collaborators on the Centre des Massues information system, protected by a secure password. The database is stored in a Microsoft Excel file on the internal server, accessible only to authorized collaborators. No data transfer to subcontractors is planned. Data will be retained for 3 years after study completion and archived for 15 years following the last publication.

Regulatory and Ethical Framework This study meets the definition of an "internal study" under French CNIL regulations, as it is conducted from data collected during routine patient follow-up, by the personnel providing that care, for their exclusive use. No individual participant information is required. The study is exempt from CNIL authorization and from the provisions of the Jardé Law (Law No. 2012-300 of March 5, 2012). The protocol is submitted for evaluation to the Internal Ethics Committee for Research (CIER) of the GHICL on May 19, 2026. In accordance with Article 30 of the GDPR, the study is registered in the data processing activities register of the Centre des Massues.

Statistical Analysis All participants meeting inclusion criteria will be included, representing approximately 70 patients. Descriptive statistics (means, standard deviations, medians, interquartile ranges, frequencies) will be computed for all variables. The primary and secondary outcomes will be analyzed using a before/after approach. For each outcome score, a linear mixed model (LMM) will be applied with fixed effects for period (before/after surgery) and measurement time, and a random patient effect to account for repeated measures correlation. Results will be reported with 95% confidence intervals. Holm correction will be applied systematically to limit type I error inflation. A significance threshold of 5% is used throughout. Statistical analyses will be performed using JASP 0.96.0 by Julie ROZAIRE.

Dissemination Results will be published in a peer-reviewed journal and presented at scientific society meetings (posters and/or oral communications).