Assessing the accuracy of common pediatric age-based weight estimation formulae.

Background: Many of the common equations for weight estimation in children were either introduced before the widespread prevalence of childhood obesity or have not been assessed in overweight/obese children. Therefore, we assessed the accuracy of 3 common age-based weight estimation formulae (Advanced Pediatric Life Support, Luscombe, and Theron) for predicting the weight of children undergoing elective, noncardiac operations. We also developed and validated a new age-based weight estimation formula.

Methods: We used preoperative anthropometric and clinical data on 13,933 children aged 2 to 12 years to evaluate the performance of 3 pediatric age-based weight estimation formulae. Ability of the formulae to predict measured weights was assessed in a derivation cohort (75% randomly selected from the study sample). We also developed and validated a new age-based formula (the Michigan formula) that could be used to estimate the weight of contemporary American children.

Results: Among the 10,488 children in the derivation cohort, 31.8% were overweight or obese while 55.7% were boys. The accuracy of the formulae varied considerably. The Luscombe formula demonstrated the lowest mean bias of 3.4 kg (95% confidence interval, 3.2-3.5 kg) and 89.7% of estimates within 10% of measured weight. Our derived linear regression equation the "Michigan Formula" demonstrated the highest accuracy compared with the existing formulae with a bias of 4.6 kg (95% confidence interval, = 4.36-4.84 kg) and 92% of estimates within 10% of measured weights.

Conclusions: Accuracies of current weight estimation formulae varied greatly. Our derived equation (Michigan formula: weight (kg) = 3 x age (yr) + 10) demonstrated high accuracy when compared with existing formulae and may be more applicable for estimating the weight of contemporary American children.