ABSTRACT
Background: In patients with growth hormone (GH) deficiency, the prediction of adult height before initiation of GH treatment can be helpful to guide clinicians and families. However, data regarding the effectiveness of prediction methods in such patients are limited.Objective: We aimed to investigate the accuracy of the three most used adult height prediction methods [Bayley-Pinneau (BP), Roche-Wainer-Thissen (RWT), and Tanner-Whitehouse 2 (TW2)] by comparing their results with the near-adult height (NAH) data of children treated with GH.Methods: A single-center retrospective study was conducted including patients treated with somatotropin due to GH deficiency. Bone age radiographs were reread by three authors. Adult height predictions were made using BP, RWT, and TW2 methods for each patient.Results: Forty-nine patients with GH deficiency [median age at diagnosis 10.8 (9.2-12.0) years, 63.3% girls, 69.4% prepubertal] were included. Median differences between predicted adult height (PAH) and NAH standard deviation (SD) scores were -0.5, 0.0, and 0.3 for BP, RWT, and TW2 methods, respectively. The rates of PAH within ±1 SD score of NAH were 54.7%, 62.3%, and 77.4% for BP, TW2, and RWT methods, respectively. RWT was the most accurate method in girls, however, it showed a similar efficiency with TW2 in prepubertal patients or those with delayed bone age between 1-2 years, independent of gender.Conclusions: We found that RWT and TW2 methods may be preferable rather than the BP method for predicting adult height in patients with a diagnosis of GH deficiency.
