Abstract
<b><i>Introduction:</i></b> Growth hormone (GH) treatment in children with growth hormone deficiency (GHD), short children born small for gestational age (SGA), and Turner syndrome (TS) is well established. However, a variety of parameters are still under discussion to achieve optimal growth results and efficiency of GH use in real-world treatment. <b><i>Methods:</i></b> German GH-treatment naïve patients of the PATRO Children database were grouped according to their start of treatment into groups of 3 years from 2007 to 2018. Time trends in age, gender, GH dose, height standard deviation score (SDS), first-year growth response, and Index of Responsiveness (IoR) were investigated in children with GHD, short children born SGA, and TS starting GH treatment in the German patient population of the PATRO Children database from 2007 to 2018 to determine specific parameters for GH treatment optimization. <b><i>Results:</i></b> All patient groups started GH treatment at a relatively high chronological age (2007–2009: GHD 8.33 ± 3.19, SGA 7.32 ± 2.52, TS 8.65 ± 4.39) with a slight but not significant trend towards younger therapy start up to 2016–2018 (GHD 8.04 ± 3.36, SGA 6.67 ± 2.65, TS 7.85 ± 3.38). In the GHD and SGA groups, female patients were underrepresented compared to male patients (GHD 32.3%, SGA 43.6%) with no significant change over the 4 time periods. Patients with GHD started GH treatment at a low dose (0.026 mg/kg/day). In SGA and TS patients, GH therapy was started below the registered dose recommendation (30.0 μg/kg/day and 33.7 μg/kg/day, respectively). In the first year of treatment, the mean GH dose was increased moderately (GHD: 30.7, SGA: 35.7, TS: 40.8 μg/kg/day). There was no significant change of GH dosing over time from 2007 to 2018. The IoR was comparable between time-groups for all 3 diagnoses. <b><i>Discussion:</i></b> This study shows potential for improvement of GH treatment results in GHD, SGA, and TS patients in terms of early dose adjustment and younger age at the start of treatment. This is in accordance with important parameters used in prediction models.