Height and Extremity-Length Prediction for Healthy Children Using Age-Based Versus Peak Height Velocity Timing-Based Multipliers.

Background The age-based multiplier method described by Paley et al. markedly simplifies height and limb length predictions but may not adequately accommodate children's maturational differences. Multipliers can be derived relative to any maturity measure. This study compares Paley age-based multipliers with those based on peak height velocity (PHV) timing. Methods In a longitudinal cohort of healthy children (66 male and 70 female), actual adult heights and limb lengths were compared with the measurements predicted using the Paley multipliers and multipliers developed relative to PHV timing. The age-based multipliers (adult divided by current) in our series were compared with those reported by Paley et al. to ensure that there were no systematic differences between the series. Absolute differences between the actual and predicted adult heights and limb lengths and the standard deviations of those differences were compared between the 2 methods. Results The average age-based multipliers in our series were nearly identical to those reported by Paley et al. The differences between the predicted and actual adult values showed wide ranges when either the Paley or the PHV multipliers were used during infancy. The Paley method performed better than the PHV method throughout pre-growth-spurt childhood. The PHV-timing-derived multipliers became superior as children entered their growth spurt, whereas the performance of the age-based multipliers worsened. In adolescence, the maximum standard deviation for adult-height-prediction errors with use of the Paley multipliers occurred at the age of 13.5 years for boys and 11.5 years for girls and was 7.0 cm for boys and 5.6 cm for girls. For limb lengths, the maximum standard deviations occurred 6 months earlier and were 3.9 cm for boys and 3.2 cm for girls. The maximum standard deviation for the height prediction error with the age-based method occurred at the average time of PHV for the population. The PHV method became better than the Paley method just before growth-spurt initiation, at age 8 in girls and 11 in boys. Conclusions The age-based multipliers described by Paley et al. are superior to PHV-timing-based multipliers prior to the adolescent growth spurt for predicting height. They become less predictive, with wide standard deviations, as children enter their growth spurts, and PHV-derived multipliers become superior. The Paley height multipliers should be used before the age of 8 years in girls and 11 years in boys. After this, PHV-derived multipliers are superior for height and limb length prediction. In practice, these predictions are currently made using skeletal maturity, which is closely related to PHV during adolescence.