ESPE Abstracts

Background and Aim: Adult height predictions by the manual Bayley and Pinneau (BP) method in tall boys have shown large confidence intervals up to a bone age of 15 years. The adult height (AH) BoneXpert prediction method, which is based on an automated bone age reading, has not been evaluated in tall adolescents. We therefore compared the bias between the manual BP method and the BoneXpert (BE) method in tall male Flemish adolescents.

Patients and method: In 20 untreated young adult (age > 19 years) men, who had been evaluated for non-pathological tall (height SDS > 2) stature during adolescence at three Flemish University hospitals, AH was obtained by a recent measurement at the hospital or by self-measurement. X-rays of the left hand and wrist, which had been obtained digitally at initial evaluation, were scored again using the Greulich and Pyle atlas by the same experienced examiner (JDS) as well as by the BE software (version 3.0). AH was compared with the BP and BE AH (version 2.23) height prediction methods.

Results: At initial assessment, standing height SDS ranged between 2 and 5.2 (median 3.0). Chronological age varied between 11.0 and 16.0 years (median 13.3 years), while GP bone age and BE bone age varied between 12 and 17 and 11.84 and 16.76 years respectively, resulting in a non-significant (median (range)) difference of 0.05 (-0.83 and 0.79) years. BE advancement ranged between – 1.2 and 1.9 years (median 0.43 years). Measured AH varied between 191 and 208 cm (median 197 cm) and correlated significantly with the BP (r = 0.582; P = 0.007) and the BE (r = 0.774; P < 0.0005) AH predictions. Bias was + 1.8 cm for the BP method and -0.3 cm for BE method, while the limits of agreement were respectively – 6.9 and + 10.4 cm and -5.5 and + 4.9 cm. The BE bias was not significant correlated with age, bone age, bone age advancement and height SDS at the moment of prediction, while the BP bias correlated negatively with age (r =- 0.462; P < 0.05) and bone age (r = -0.609; P < 0. 0005).

Conclusion: In tall adolescent males aged between 11 to 16 years and with moderate bone age advancement, the BE AH prediction method performs better than the classic BP method, given its higher correlation with AH, smaller limits of agreement and lower bias.