Cerebral palsy is a static encephalopathy of the immature brain that primarily affects motor function and may result in disturbances in cognition, perception, and sensation, and in other neuromotor deficits1. In the developed world, the incidence of cerebral palsy has been estimated to be between 2 and 2.5 cases per 1,000 live births2,3. The Gross Motor Function Classification System (GMFCS) of cerebral palsy segregates patients on the basis of their ability to perform self-initiated movement4. Children at GMFCS levels 4 and 5 have substantial limitations in self-care and rely on wheelchairs for their mobility. Additionally, patients at GMFCS level 5 have severe limitations in the ability to maintain antigravity trunk control.
Children at GMFCS level 4 or 5 are also at higher risk of developing hip and spine problems compared with those at lower GMFCS levels. In a study of 374 children with cerebral palsy, the prevalence of hip displacement was 0% for those at GMFCS level 1 and 90% for those at GMFCS level 55. In another study of 666 children with cerebral palsy, those at GMFCS level 4 or 5 were found to have an approximate 50% risk of developing moderate or severe scoliosis by the age of 18 years6. In addition to spine and hip involvement, patients at GMFCS level 4 or 5 are more likely to have accompanying impairments. In 1 study, 48 (89%) of 54 patients at GMFCS level 5 had ≥2 accompanying impairments such as a learning disability, epilepsy, and severe visual impairments7.
We hypothesized that patients at GMFCS level 5 are a heterogeneous group that can be meaningfully subdivided according to central neuromotor impairments in feeding and/or swallowing, respiratory function, speech, and cortical stability (seizures). It is our experience that patients at GMFCS level 5 are a heterogeneous group with various levels of disability. Some patients in our practice are at GMFCS level 4 or 5 but have normal or high intelligence and cognitive function. The aim of our study was to subclassify patients at GMFCS level 5 who underwent spinal arthrodesis for deformity correction on the basis of defined central neuromotor impairments and to assess whether the subclassification can be used to predict postoperative complications and changes in health-related quality of life.
Materials and Methods
This study was approved by the institutional review board. The subjects or their caretakers gave informed consent to participate in the study.
Using a prospective cerebral palsy registry, we identified 199 patients at GMFCS level 5 who underwent spinal fusion for deformity correction between June 2008 and October 2013. The mean follow-up time (and standard deviation) for this cohort was 2.4 ± 0.5 years.
The mean age at the time of surgery was 13.8 ± 2.8 years, and 104 patients (52%) were boys (Table I). The mean body mass index was 16.9 ± 4.0 kg/m2. One hundred and seventy-two patients (86%) had spastic cerebral palsy. Before surgery, 58 patients (29%) had open triradiate cartilage.
Central Neuromotor Impairments
Patients were assigned to subgroups on the basis of the total number of preoperative central neuromotor impairments in the following 4 categories: feeding and/or swallowing, respiratory function, cortical stability (seizures), and speech. Of the 199 patients, 124 (62%) had a gastrostomy tube and 4 (2%) had a tracheostomy (Table I). Furthermore, 148 patients (74%) had a history of seizure in the last 12 months before spinal arthrodesis, and 166 patients (83%) were classified as nonverbal (Table I). Seventeen patients (9%) had no central impairment (classified as GMFCS level 5.0) (Figs. 1 and 2). Twenty-eight patients (14%) had 1 impairment (GMFCS level 5.1). Fifty-two patients (26%) had 2 impairments (GMFCS level 5.2). One hundred and two patients (51%) had 3 or 4 impairments (GMFCS level 5.3). Whereas no patient with GMFCS subtype 5.0 had any of the 4 neuromotor impairments, all patients with GMFCS subtype 5.3 had a gastrostomy tube, history of seizures, and nonverbal status (Fig. 3).
Most patients were treated with posterior spinal instrumentation and arthrodesis (Table II). A mean of 16.0 ± 0.8 levels were fused, most frequently from T2 to the pelvis. Of the 199 patients, 63 (32%) received all-pedicle screw constructs; 89 (45%) received constructs with at least 80% pedicle screws; 19 (10%) received hybrid constructs with hooks, screws, and sublaminar wires; and 28 (14%) received unit rod constructs. Various pelvic fixation techniques were used, with iliac screws used most commonly. The mean operative time was 397 ± 130 minutes. The mean estimated blood loss was 1.7 ± 1.2 L. On average, patients lost an estimated 79% ± 65% of their blood volume. The mean length of hospitalization was 12.6 ± 10.9 days.
The major coronal curve was corrected from a mean 84° ± 25° to 31° ± 16° (Table III). Pelvic obliquity (measured with respect to the horizontal plane) was corrected from a mean 23° ± 13° to 9° ± 7°.
Health-Related Quality of Life
Health-related quality of life was measured using the Caregiver Priorities and Child Health Index of Life with Disabilities (CPCHILD) questionnaire8. The CPCHILD instrument consists of 37 items covering 6 domains and is self-administered by parents or caregivers of children with cerebral palsy. Scores for each domain range from 0 (worst) to 100 (best). The mean preoperative CPCHILD total score was highest for patients at GMFCS level 5.0 (mean score, 57.8 ± 10.0) and lowest for those at GMFCS level 5.3 (mean score, 45.1 ± 13.6).
Major complications that occurred during the perioperative period or after discharge were queried. The following complications were considered major after consensus-building among the surgeons in the study group: cardiopulmonary (postoperative placement of chest tube, intubation for ≥2 days, reintubation, placement of a new tracheostomy tube, and perioperative cardiac arrest); gastrointestinal (the need for perioperative or postoperative abdominal surgery, or placement of a new gastrostomy tube); neurologic (motor deficits or paralysis); infectious (postoperative deep wound infection); thromboembolic (perioperative or postoperative deep venous thrombosis or pulmonary embolism); and postoperative death.
The differences in complication rates among the GMFCS level-5 subtypes were analyzed using the chi-square test. The differences in radiographic parameters and CPCHILD scores among various GMFCS level-5 subtypes were analyzed using the analysis-of-variance (ANOVA) test. Significance was set at p < 0.05.
There were no significant differences in patient age at the time of surgery (p = 0.084) or in the proportion of patients who were female (p = 0.949) by GMFCS level-5 subtype. Furthermore, we found no significant difference in body mass index values (p = 0.981) by GMFCS level-5 subtype.
Of the 199 patients, 65 (33%) developed 1 major complication, 7 (3.5%) developed 2 major complications, and 2 (1%) developed 3 major complications each. Complications were as follows: cardiopulmonary (56 patients; 28%), infectious (13 patients; 7%), gastrointestinal (4 patients; 2%), thromboembolic (3 patients; 1.5%), and neurologic (2 patients; 1%). There were 7 deaths over the 2-year follow-up period. The mean time to death was 1.3 ± 1.2 years after surgery.
Complication Rate by Subclassification
The rates of major complications varied significantly by GMFCS level-5 subtype (p = 0.002), with 2 (12%) of 17 patients at level 5.0, 6 (21%) of 28 patients at level 5.1, 16 (31%) of 52 patients at level 5.2, and 50 (49%) of 102 patients at level 5.3 having a major complication (Fig. 4). Of the 7 patients who died, 5 were from level 5.3 and 1 each from levels 5.1 and 5.2. There was only 1 perioperative death, a patient at GMFCS level 5.3 who died 7 days after surgery because of pulmonary complications.
There was no significant difference among the groups with respect to estimated blood loss (p = 0.66) or percentage of blood volume lost (p = 0.72) by GMFCS level-5 subtype. However, patients who developed at least 1 major complication were found to have significantly greater estimated blood loss (2.3 versus 1.4 L; p < 0.001) and a significantly greater percentage of blood volume lost (110% versus 61%; p < 0.001) than patients who did not develop a major complication.
According to univariate logistic regression analysis, the odds of developing a major complication were higher in patients with a gastrostomy tube (odds ratio [OR] = 3.2; 95% confidence interval [CI] = 1.7 to 6.2 [p < 0.001]; pseudo-R2, 0.05); nonverbal status (OR = 3.1; 95% CI = 1.2 to 8.0 [p = 0.017]; pseudo-R2, 0.03); and a history of seizures (OR = 2.3; 95% CI = 1.1 to 4.8 [p = 0.021]; pseudo-R2, 0.02). Only 4 patients had a tracheostomy, and 2 of them developed a major complication (a complication rate of 50% compared with 37% in patients without a tracheostomy). However, this difference was not significant (p = 0.6).
Radiographic Changes by Subclassification
No significant differences were found among the GMFCS level-5 subgroups on ANOVA with respect to the preoperative major coronal curve (p = 0.987), postoperative major coronal curve (p = 0.545), magnitude of correction of the major coronal curve (p = 0.938), preoperative pelvic obliquity from horizontal (p = 0.754), postoperative pelvic obliquity from horizontal (p = 0.838), and magnitude of correction of pelvic obliquity (p = 0.541).
CPCHILD Scores by Subclassification
Preoperatively, there was significant variation among GMFCS level-5 subgroups in the CPCHILD communication and social interaction domain, health domain, and total scores (Table IV). At the time of the final follow-up, there was also significant variation among GMFCS level-5 subgroups in the CPCHILD communication and social interaction domain, health domain, overall quality-of-life domain, and total scores (Table IV).
However, no significant differences were found in the magnitude of improvement from the preoperative to the final follow-up evaluation among the GMFCS level-5 subgroups in the communication and social interaction domain (p = 0.496), health domain (p = 0.9588), overall quality-of-life domain (p = 0.680), and total score (p = 0.597).
Whereas patients with cerebral palsy primarily experience disorders of movement, they may also experience various neuromotor impairments of central or autonomic nervous system functions9, which may have substantial implications for the child and family. In this study, we found that patients with cerebral palsy at GMFCS level 5 can be subclassified on the basis of preoperative central neuromotor impairments, which can predict postoperative complications and health-related quality-of-life outcomes after spinal arthrodesis.
The GMFCS is a classification system for cerebral palsy that was developed to categorize patients on the basis of their self-initiated movement abilities4, and it is excellent at identifying patients who are at high risk for scoliosis and hip dysplasia (patients at GMFCS levels 4 and 5). Although the GMFCS does not take into account additional neuromotor impairments, patients at GMFCS level 5 are considered to be the most disabled and are likely to have a higher frequency of additional neuromotor impairments. In a study of 301 children with cerebral palsy, 108 patients were at GMFCS level 1, and 7% of them had a history of seizures in the last 12 months, 6% had severe auditory impairment, 4% had severe visual impairment, and 1% were nonverbal10. In comparison, of the 39 patients at GMFCS level 5, 82% were nonverbal, 44% had a history of seizures in the last 12 months, 33% had severe visual impairment, and 21% had severe auditory impairment10. In the current study, 9 of the 199 patients at GMFCS level 5 had no cognitive impairments.
In the current study, we selected 4 comorbidities that we believed were markers of neuromotor impairment: a gastrostomy tube (as an indicator of swallowing function), a tracheostomy tube (as an indicator of airway protection and respiratory drive), a history of seizures (as an indicator of cortical stability), and nonverbal status (as an indicator of apraxia and dysarthria). Interestingly, a tracheostomy tube was used in only 2% of the population. Although this may have been attributable to selection bias, in our experience, tracheostomy is rare even in patients at GMFCS level 5. Some patients who required a tracheostomy or were dependent on mechanical ventilators might have been considered medically unfit to undergo spinal arthrodesis.
The results of our study indicated that, in patients at GMFCS level 5, central neuromotor impairments were not associated with differences in radiographic characteristics such as preoperative major coronal curve or pelvic obliquity. However, neuromotor impairments were significantly associated with the development of major complications. Patients at GMFCS level 5.3 had a rate of major complications of 49%, which was 4 times higher than that for patients at GMFCS level 5.0.
The CPCHILD questionnaire measures the perspectives of caregivers on the health, quality of life, and ease of providing care for children with cerebral palsy. In our study, there was a mean improvement in CPCHILD scores of 7 points (from 48 to 55 points) over the 2-year follow-up period for the entire GMFCS cohort. We found that patients with all GMFCS level-5 subtypes experienced improvement in their CPCHILD total scores after spinal arthrodesis. No significant difference was found in the magnitude of CPCHILD improvement achieved by children with the GMFCS level-5 subtype. When we examined specific components of the CPCHILD for the entire GMFCS level-5 cohort, we found that the greatest reported improvements were (1) in the positioning, transfers, and mobility domain (mean improvement, 9.4 points); (2) in the comfort and emotions domain (mean improvement, 8.9 points); and (3) in the quality-of-life domain (mean improvement, 7.2 points).
As would be expected, we found that, as the number of central neuromotor impairments increased, the preoperative CPCHILD communication and social interaction and health scores decreased significantly. This makes sense considering that 83% of patients in our cohort were classified as nonverbal. Further, there was a trend toward decreased preoperative quality-of-life domain scores from GMFCS level 5.0 to 5.3. The mean total score of 57.8 for the GMFCS level-5.0 subgroup in our study was similar to the reference score of 56.3 for the GMFCS level-4 group reported by Narayanan et al.11. Similarly, the mean total score of 45.1 for the GMFCS level-5.3 subgroup in our study was similar to the reference score of 44.4 for the GMFCS level-5 group reported by Narayanan et al.11.
Blood loss can be associated with development of complications, and a prior study has suggested that children with cerebral palsy may be at a higher risk of blood loss during spinal arthrodesis than patients with other diagnoses12. In the current study, the mean estimated blood loss was 1.7 L, representing a mean loss of 79% of circulating blood volume. These data are consistent with a recent study that found that approximately 40% of children with cerebral palsy lose their entire circulating blood volume during spinal arthrodesis13. Furthermore, in the current study, all patients were at GMFCS level 5. This is important to note because a prior study has suggested that lower body weight may be associated with greater blood loss during spinal arthrodesis14.
Our study has several limitations. First, the markers of central neuromotor impairments were selected on the basis of the data collected for our prospective registry. There may be other important markers of central neuromotor impairment such as cognitive impairment that could potentially help to stratify complications and health-related quality-of-life outcomes that we did not include. Another limitation is that our mean follow-up period was 2.4 years. Longer follow-up is crucial to understanding the clinical course of this population. Finally, we did not include patients at GMFCS level 3 or 4 in this study because their numbers were small and they had few central neuromotor impairments. On the basis of the current literature, it is unclear whether patients at GMFCS levels 3 and 4 are at increased risk of complications after spinal arthrodesis. Future research should focus on these patients. Despite these shortcomings, to our knowledge, this is the largest series in the literature on patients with cerebral palsy at GMFCS level 5 who underwent spinal arthrodesis for deformity correction.
GMFCS level 5 is a heterogeneous category of patients with cerebral palsy that can be meaningfully subdivided on the basis of central neuromotor impairments. Patients with central neuromotor impairments are at higher risk of developing postoperative complications after spinal arthrodesis and have lower health-related quality-of-life outcomes. These data can be used for improved prediction of postoperative complications and to better counsel patients and families on the expected clinical course.
NOTE: The authors thank Dr. Louis Okafor for his illustration (Fig. 1) of neuromotor deficits in patients at GMFCS level 5.
Investigation performed at the Departments of Orthopaedic Surgery and Anesthesiology, The Johns Hopkins University, Baltimore, Maryland; Nemours/Alfred I. duPont Hospital for Children, Wilmington, Delaware; Department of Orthopaedic Surgery, Shriners Hospitals for Children, Philadelphia, Pennsylvania; Department of Orthopedics, Rady Children’s Hospital of San Diego, San Diego, California; Department of Orthopaedic Surgery, University of Virginia, Charlottesville, Virginia; Setting Scoliosis Straight Foundation, San Diego, California; and Division of Orthopaedic Surgery, The Hospital for Sick Children, University of Toronto, Toronto, Ontario, Canada
Disclosure: Funding was received from DePuy Synthes Spine and K2M to the Setting Scoliosis Straight Foundation to support the cerebral palsy scoliosis research conducted by the Harms Study Group. On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial relationship in the biomedical arena outside the submitted work; “yes” to indicate that the author had a patent and/or copyright, planned, pending, or issued, broadly relevant to this work; and “yes” to indicate that the author had other relationships or activities that could be perceived to influence, or have the potential to influence, what was written in this work.
1. Rosenbaum P, Paneth N, Leviton A, Goldstein M, Bax M, Damiano D, Dan B, Jacobsson B. A report: the definition and classification of cerebral palsy April 2006. Dev Med Child Neurol Suppl. 2007 ;109(S109):8–14.
2. Bhushan V, Paneth N, Kiely JL. Impact of improved survival of very low birth weight infants on recent secular trends in the prevalence of cerebral palsy. Pediatrics. 1993 ;91(6):1094–100.
3. Nelson KB, Ellenberg JH. Epidemiology of cerebral palsy. Adv Neurol. 1978;19:421–35.
4. 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(4):214–23.
5. Soo B, Howard JJ, Boyd RN, Reid SM, Lanigan A, Wolfe R, Reddihough D, Graham HK. Hip displacement in cerebral palsy. J Bone Joint Surg Am. 2006 ;88(1):121–9.
6. Persson-Bunke M, Hägglund G, Lauge-Pedersen H, Wagner P, Westbom L. Scoliosis in a total population of children with cerebral palsy. Spine (Phila Pa 1976). 2012 ;37(12):E708–13.
7. Himmelmann K, Beckung E, Hagberg G, Uvebrant P. Gross and fine motor function and accompanying impairments in cerebral palsy. Dev Med Child Neurol. 2006 ;48(6):417–23.
8. Narayanan UG, Fehlings D, Weir S, Knights S, Kiran S, Campbell K. Initial development and validation of the Caregiver Priorities and Child Health Index of Life with Disabilities (CPCHILD). Dev Med Child Neurol. 2006 ;48(10):804–12.
9. Bax M, Goldstein M, Rosenbaum P, Leviton A, Paneth N, Dan B, Jacobsson B, Damiano D; Executive Committee for the Definition of Cerebral Palsy. Proposed definition and classification of cerebral palsy, April 2005. Dev Med Child Neurol. 2005 ;47(8):571–6.
10. Shevell MI, Dagenais L, Hall N; REPACQ Consortium. Comorbidities in cerebral palsy and their relationship to neurologic subtype and GMFCS level. Neurology. 2009 ;72(24):2090–6.
12. Jain A, Njoku DB, Sponseller PD. Does patient diagnosis predict blood loss during posterior spinal fusion in children? Spine (Phila Pa 1976). 2012 ;37(19):1683–7.
13. Jain A, Sponseller PD, Shah SA, Yaszay B, Njoku DB, Miyanji F, Newton PO, Bastrom TP, Marks MC; Harms Study Group. Incidence of and risk factors for loss of 1 blood volume during spinal fusion surgery in patients with cerebral palsy. J Pediatr Orthop. 2016 . [Epub ahead of print].
14. Jain A, Sponseller PD, Newton PO, Shah SA, Cahill PJ, Njoku DB, Betz RR, Samdani AF, Bastrom TP, Marks MC; Harms Study Group. Smaller body size increases the percentage of blood volume lost during posterior spinal arthrodesis. J Bone Joint Surg Am. 2015 ;97(6):507–11.