Novel weight estimation equation for children with cerebral palsy in low-resource settings: Validation in a population-based cohort.

Abstract

To validate a novel equation to estimate weight from mid-upper arm circumference (MUAC) among children with cerebral palsy (CP) in rural Bangladesh. Children with CP aged 2 to 18 years registered in the Bangladesh CP Register were randomly selected. Data on sociodemographics, Gross Motor Function Classification System level, and anthropometric measurements were extracted. Bland-Altman plots with a 95% agreement limit and Lin's concordance correlation coefficient with 95% confidence intervals (CI) were reported to measure agreement between observed and estimated weight. Percentage error was used to determinate the method's accuracy. There were 497 participants with a mean age at assessment of 9 years (SD 4 years 11 months) (47.7% female). Lin's concordance correlation coefficient between the observed and estimated weights was 0.90 (95% CI 0.89-0.92). Bland-Altman plots showed a reasonable accuracy of the equation in the study cohort. The mean percentage error of the equation was 5.04%. The average difference between observed and estimated weights was -1.02 kg (SD 5.1). The differences between observed and estimated weights were significantly greater among children with weight-for-age, height-for-age, or BMI-for-age z-scores less than or equal to -4. It is possible to predict the weight of children with CP from MUAC with sufficient accuracy. The equation can be used for populations in low-resources and low- and middle-income countries. The equations predict the weight of children with cerebral palsy from their mid-upper arm circumference reasonably accurately. The difference between observed and estimated weights ranged between 0 kg and ± 5 kg in 81.5% of children. Sex and Gross Motor Function Classification System level did not affect the accuracy of the equations. The equations were less accurate for estimating the weight of severely undernourished children.