The aim of this prospective, longitudinal, controlled study is to describe the long-term safety and efficacy of growth hormone (GH) administration on craniofacial morphology in boys with short stature. Forty-six boys, who started GH treatment at the Department of Paediatrics Göteborg Paediatric Growth Research Centre, were consecutively included in the study. Twenty-five boys were classified as growth hormone-deficient (GHD) and 21 as idiopathic short stature (ISS). The patients were injected with 33 (n=31) or 67 (n=15) microg GH/kg body weight/day. The mean age at the start of treatment was 11.8 years [standard deviation (SD) 1.7]. To assess craniofacial growth, standard lateral cephalometric radiographs were obtained at the start of GH treatment, annually during 4 years, and at the end of GH treatment or when growth was less than 1 cm/year. The mean follow-up period was 6.4 years (SD 1.4). Growth changes were compared with boys from a semi-longitudinal reference group of 130 healthy subjects, 7-21 years of age. t-tests for independent and paired samples and multiple regression analysis were applied. Age- and gender-specific standard deviation scores for the cephalometric variables were calculated. Repeated measures analysis of variance was used to identify significant covariates over time, such as low/high GH dose and GHD/ISS and orthodontic treatment. During the study period, eight (out of 40) boys were treated with fixed orthodontic appliances, three with functional appliances (activators), and three with other appliances (plates and lingual arches). During GH treatment period, an overall enhancement in growth of the facial skeleton was observed in boys with short stature. The changes induced by GH yielded a more prognathic growth pattern, a more anterior position of the jaws in relation to the cranial base, and increased anterior rotation of the mandible. The mandibular corpus length and anterior face height of the GH-treated boys were greater at the end of the study compared with the boys in the reference group. No differences in growth response were noted either between the GHD and ISS boys or between those treated with either 33 (low dose) or 67 (high dose) microg GH/kg body weight/day. The only change that remained significantly correlated with orthodontic treatment was the alteration in mandibular ramus height, showing a larger change in the boys who had not undergone orthodontic therapy. The findings of this study demonstrate that GH treatment has a favourable influence on the craniofacial growth pattern of boys with short stature without acromegalic features.