Cross-validation of the Beunen–Malina method to predict adult height

Abstract

The purpose of this study was to cross-validate the Beunen–Malina method for non-invasive prediction of adult height. Three hundred and eight boys aged 13, 14, 15 and 16 years from the Madeira Growth Study were observed at annual intervals in 1996, 1997 and 1998 and re-measured 7–8 years later. Height, sitting height and the triceps and subscapular skinfolds were measured; skeletal age was assessed using the Tanner–Whitehouse 2 method. Adult height was measured and predicted using the Beunen–Malina method. Maturity groups were classified using relative skeletal age (skeletal age minus chronological age). Pearson correlations, mean differences and standard errors of estimate (SEE) were calculated. Age-specific correlations between predicted and measured adult height vary between 0.70 and 0.85, while age-specific SEE varies between 3.3 and 4.7 cm. The correlations and SEE are similar to those obtained in the development of the original Beunen–Malina method. The Beunen–Malina method is a valid method to predict adult height in adolescent boys and can be used in European populations or populations from European ancestry. Percentage of predicted adult height is a non-invasive valid method to assess biological maturity.