Construction of Lambda, Mu, Sigma Values for Determining Mid-Upper Arm Circumference z Scores in US Children Aged 2 Months Through 18 Years

Background: Mid-upper arm circumference (MUAC) has proven highly predictive of morbidity and mortality associated with malnutrition better, in some cases, than other growth indicators, including body mass index (BMI) z scores and weight-for-height z scores. A recent consensus statement recommended the inclusion of MUAC and MUAC z scores in the nutrition assessment of children in the United States; however, the requisite data to permit z score calculations for children aged >5 years have not been published. Objective: This investigation was designed to generate lambda mu sigma (LMS) values to permit the calculation of MUAC z scores in U.S. children 2 months through 18 years of age. Design: Anthropometric data from the Centers for Disease Control and Prevention (CDC) National Health and Nutrition Examination Survey (1999-2012) were used for model development (n = 28,995). Smoothed centiles were constructed and compared with previously described CDC percentiles. Independently collected MUAC data from 2 different U.S. studies were used for external validation (n = 1438). Statistical Analyses: Goodness-of-fit was assessed visually and statistically by examining detrended quantile-quantile plots, Q statistics, and the distribution of z scores. Results: The curves generated in this investigation fit the raw data well with no systematic bias and no sacrifice in fit for children aged <12 months. The curves were consistent with those published by the CDC, and the distribution z scores approximated 0 +/- 1 in all age groups. Conclusions: These LMS values derived in this investigation can be used by clinicians to generate MUAC z scores for U.S. children.