AbstractObjectiveTo estimate the prevalence of underweight between 1998 and 2006 in Liverpool schoolchildren aged 9–10 years using recently published underweight cut-off points.Design and settingStature and body mass data collected at the LiverpoolSportsLinx project’s fitness testing sessions were used to calculate BMI.SubjectsData were available on 26 782 (n13 637 boys, 13 145 girls) participants.ResultsOverall underweight declined in boys from 10·3 % in 1998–1999 to 6·9 % in 2005–2006, and all sub-classifications of underweight declined, in particular grade 3 underweight, with the most recent prevalence being 0·1 %. In girls, the prevalence of underweight declined from 10·8 % in 1998–1999 to 7·5 % in 2005–2006. The prevalence of all grades of underweight was higher in girls than in boys. Underweight showed a fluctuating pattern across all grades over time for boys and girls, and overall prevalence in 2005–2006 represents over 200 children across the city.ConclusionsUnderweight may have reduced slightly from baseline, but remains a substantial problem in Liverpool, with the prevalence of overall underweight being relatively similar to the prevalence of obesity. The present study highlights the requirement for policy makers and funders to consider both ends of the body mass spectrum when fixing priorities in child health.