Surgery Plus Rehabilitation Reshapes Cerebral Palsy Care, Review Finds

Cerebral palsy has never been a condition that a single medical specialty could fix. A new narrative review published in Frontiers in Neurology argues that the field’s most important shift is not any one new operation or device, but the deliberate coupling of surgery with rehabilitation — a strategy the authors say has moved from concept to clinical practice, yet still lacks the standardized pathways and high-quality trials needed to prove its full value. The review, written by a multidisciplinary team of orthopedic, rehabilitation and surgical researchers in Kunming, China, synthesizes recent advances across spasticity management, deformity correction, neuromodulation and emerging digital technologies, and is unusually candid about where the evidence runs thin.

Cerebral palsy is a group of motor disorders caused by non-progressive brain injury during development, often accompanied by sensory, cognitive and behavioral impairments, and it remains one of the most common causes of childhood disability. Spastic cerebral palsy, driven by corticospinal tract damage that reduces inhibitory input to spinal motor neurons, accounts for roughly 70 to 80 percent of cases. Dyskinetic forms involve abnormalities in basal ganglia-thalamo-cortical circuits, while ataxic forms are linked primarily to cerebellar injury. The review stresses that this classification is not a diagnostic label but a decision-making tool: spastic subtypes are the primary indication for procedures such as selective dorsal rhizotomy, whereas dyskinetic subtypes are better matched to neuromodulation approaches such as deep brain stimulation combined with rehabilitation.

The central logic of integrated care rests on a division of labor. Surgery — whether selective dorsal rhizotomy, which partially severs over-excited dorsal root afferents to reduce lower-limb spasticity, or orthopedic procedures such as tendon lengthening and osteotomy, which correct joint contractures and bony deformities — creates the anatomical and mechanical conditions for function. Rehabilitation is thought to leverage neuroplasticity to strengthen residual motor capacity, consolidate surgical gains and prevent secondary damage, although the authors note that direct evidence for this in cerebral palsy remains limited. A key assumption is that structural improvements must be captured by timely rehabilitation or they will fade. The review describes a postoperative time window: passive mobilization and bed-based activities after selective dorsal rhizotomy typically begin within the first one to two days, and a critical consolidation phase of three to six months is when strengthening and gait training are considered most effective.

The quantitative picture for selective dorsal rhizotomy is among the strongest in the review. Quantitative syntheses report a pooled mean improvement in GMFM-66 score of 3.29 points and a pooled mean reduction in Modified Ashworth Scale score of 1.98 points, with substantial gains achieved without sacrificing muscle strength or selective motor control. Domain-level data show that gains in ambulatory children concentrate in the standing and walking, running and jumping domains. Surgical candidates are typically children with spastic diplegia or quadriplegia at Gross Motor Function Classification System levels I to III without severe fixed deformities. Single-level laminectomy has largely replaced multi-level approaches, reducing complications such as spinal deformity, and intraoperative neurophysiological monitoring helps surgeons selectively section over-excited nerve rootlets. Yet the authors caution that most studies are single-center and retrospective, with few randomized controlled trials, and long-term follow-up shows some patients reach a plateau or even decline 10 to 15 years after surgery.

Orthopedic surgery occupies a different evidentiary tier. Tendon lengthening reduces musculotendinous tension and can improve ankle dorsiflexion, addressing the equinus gait deviation that is typical in cerebral palsy. But surgical timing remains contested: because rapid skeletal growth can drive deformity recurrence, the traditional practice of operating before the pubertal growth spurt rests more on clinical experience than rigorous evidence, and the review argues timing should be individualized using bone age, growth curves and functional deterioration trends. For children with severe spasticity and fixed deformities, combined pathways may be considered, with selective dorsal rhizotomy performed first to reduce spasticity before contractures and bony deformities are corrected.

On the rehabilitation side, the review covers physical, occupational and speech therapy alongside newer technologies. No universal minimum effective therapy dose has been established, and a Bayesian network dose-response meta-analysis confirms that exercise dose and motor skill improvement follow a nonlinear relationship; higher intensity generally yields better functional outcomes, but optimal dosing remains undefined. Constraint-induced movement therapy is a key intervention for hemiplegic cerebral palsy, and speech impairment affects up to 89 percent of children with the condition, while dysphagia affects more than half, compromising nutrition and, in severe cases, respiratory safety — problems the authors say belong in routine integrated assessment rather than being treated as separate issues. Meta-analyses of randomized trials indicate robotic-assisted gait training significantly improves motor function, with a standardized mean difference of 0.89, and gait outcomes, with a standardized mean difference of 0.62, in children with cerebral palsy.

Neuromodulation presents a sharper contrast between promise and proof. Deep brain stimulation targeting the globus pallidus internus, with the cerebellar dentate nucleus emerging as a novel target, has shown value for dystonia in dyskinetic cerebral palsy — a meta-analysis reported a mean improvement of 23.6 percent in Burke-Fahn-Marsden Dystonia Rating Scale motor scores — but the review classifies it as experimental, supported mostly by small case series without large controlled trials. Transcranial magnetic stimulation and transcranial direct current stimulation are being explored as adjuncts that may heighten motor cortex plasticity before training, though effects are heterogeneous and optimal protocols undefined. Spinal cord stimulation has shown preliminary potential to reduce lower-limb muscle tone in selected patients. Proposed synergies, such as rehabilitation timed to a stimulation-induced window of heightened cortical excitability, remain theoretically appealing but evidence-scarce.

The review is equally frank about systemic gaps. Outcome measurement is fragmented: GMFM-66 is the internationally accepted gold standard for motor function, complemented by three-dimensional gait analysis, the 10-meter walk test, the six-minute walk test, the Timed Up and Go test and the CPQoL-Child quality-of-life instrument, but no dedicated tool exists to quantify the interaction between surgery and rehabilitation, so existing measures capture only single-dimension changes. Multidisciplinary teams spanning neurosurgery, orthopedics, rehabilitation medicine and therapy disciplines are widely recommended, yet the evidence for their advantage over single-specialty care is largely descriptive. Prehabilitation — preoperative strength training, range-of-motion work and psychological preparation — is described as an emerging strategy with limited direct evidence rather than an established standard. Most studies follow patients for fewer than five years, leaving the durability of functional gains into adulthood largely unknown.

Looking forward, the authors identify three transformative directions: integrating early intervention, which capitalizes on the peak window of infant neuroplasticity and is associated with reduced risk of joint contractures, into full-cycle integrated care; regenerative medicine, where preliminary randomized evidence suggests mesenchymal stem cell transplantation combined with rehabilitation may improve gross motor function in some children but remains investigational at a proof-of-concept stage; and artificial intelligence, where machine learning applied to gait and electromyography data could shift decisions from experience-driven to data-driven, supporting individualized training protocols and connected care pathways. The review’s closing message is a paradox worth holding onto: mechanistic reasoning and observational cohorts support synergy between the scalpel and the therapy gym, but direct randomized evidence that combined treatment exceeds the sum of its parts is still sparse — and closing that gap, through registries, unified outcome measures and long-term follow-up, is now the field’s central task.

Subject of Research: A narrative review of integrated surgical and rehabilitative treatment for cerebral palsy

Article Title: Advances in integrated surgical and rehabilitation treatment for cerebral palsy

Article References: Zhang, Z., Li, L., Li, K., Yu, C., Du, Q., Hu, Z., Yu, X., Zhang, J., Hu, J., & Zhou, Y. (2026). Advances in integrated surgical and rehabilitation treatment for cerebral palsy. Frontiers in Neurology, 17, Article 1986453. https://doi.org/10.3389/fneur.2026.1986453

Image Credits: AI Generated

DOI: 10.3389/fneur.2026.1986453

Keywords: cerebral palsy, selective dorsal rhizotomy, orthopedic surgery, rehabilitation, neuromodulation, deep brain stimulation, robotic gait training, spasticity, GMFM-66, multidisciplinary team, artificial intelligence, Frontiers in Neurology