Weight and Health Outcomes in Cerebral Palsy: A Causality Dilemma

Cerebral palsy (CP), the most common lifelong motor disability, is a multimodal condition with impacts far beyond motor functioning.1 CP motor prognosis, feeding impairments, medical complexity, and rates of coexisting conditions are known to covary with the child’s Gross Motor Function Classification System (GMFCS) level.2 The GMFCS level indicates whether the child ambulates without an assistive device (I-II), ambulates with an assistive device (III), or primarily uses a wheelchair (IV-V).3 Children at different GMFCS levels are known to have different patterns of growth. Children who primarily use wheelchairs for mobility weigh less and are more likely to have feeding impairments than those who independently ambulate.4 These GMFCS-based growth differences are reflected in CP-specific growth charts that provide GMFCS-stratified weight percentiles (GWt%-ile).4 In the long canon of pediatric tradition dedicated to ensuring a child is growing as anticipated,5 including via the use of condition-specific growth charts,6 we are due for the widespread clinical adoption of CP-specific growth charts. However, it has remained unclear whether GWt%-ile are a true representation of a child’s ideal growth trajectory by GMFCS level as opposed to a reflection of the feeding impairments and medical complexity that also vary by GMFCS level.7 We continue to have a causality dilemma.Stevenson et al8 begin to address this in this issue of Pediatrics. They present an analysis of data collected from the North American Growth in Cerebral Palsy Project (NAGCPP) to represent a more geographically diverse population to further validate the CP-specific growth charts developed using a California cohort, specifically in those at GMFCS levels III-V. They additionally show that children at less than the 20th GWt%-ile have increased health care use and decreased participation in the 4 weeks prior to weight measurement. These metrics reached significance for days in the hospital, days the child missed their usual program (school), and total days missed from their usual program and activities for both child and caregiver. Children at less than the 20th GWt%-ile primarily differed from children at greater than 20th GWt%-ile by anthropometric measures (significantly smaller) and feeding impairments (significantly more frequent). Children at less than the 20th GWt%-ile were almost twice as likely to have severe feeding problems defined as “severe difficulties consuming liquids and foods (requiring well moistened foods or thickened fluids or requiring tube feeding).”8 They were also more likely to have symptoms that cause feeding problems (swallowing difficulties, gagging, vomiting) and to take medications to treat such problems (reflux/motility medications, dietary supplements). In fact, controlling for GMFCS, feeding method, and sex, little else differed between children at less than the 20th GWt%-ile and children at greater than the 20th GWt%-ile. Finally, the authors showed that triceps skin-fold measurements and severe feeding problems were independently associated with increased risk of the same health care use and participation measures associated with GWt%-ile.The authors “propose that weight below the 20th percentile for children with CP and GMFCS III, IV, and V could be considered a cutoff for further nutritional assessment.”8 Their data support this proposal but could just as easily support that the primary screen should be for severe feeding problems using a caregiver questionnaire (like the North American Growth questionnaire as done for NAGCPP)9 and that this alone should trigger further nutritional assessment. Furthermore, because the data analyzed reflect health care use and participation in the 4 weeks prior to weight assessment, it is unclear whether the GWt%-ile is predictive of these important health measures or instead is a result of these phenomena.However, Stevenson et al8 successfully argue to focus on what matters. Whether a child has underweight for their GMFCS level or whether they have severe feeding problems, a clinician should consider nutritional assessment. The use of CP-specific growth charts additionally allows tracking changes in GWt%-ile over time, determining the health impacts of significant changes in GWt%-ile over time (increases or decreases) and the health impacts of interventions focused on increasing GWt%-ile in children who have underweight (or decreasing it in children who have overweight). It is possible that the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition may still have the same concerns about CP-specific growth charts as they did in 2017. They “describe growth which is not necessarily ideal as they include many children with other health conditions affecting growth, mostly malnutrition.”7 Even if not tracking “ideal” growth, a potentially unattainable metric in a heterogeneous condition like CP, Stevenson et al8 demonstrate that CP-specific growth charts have value in ideal CP care.