RESUMEN
To assess the long-term safety of administering growth hormone (GH) in patients with GH deficiency due to treatment for childhood medulloblastoma and primitive neuroectodermal tumor (PNET). Data were retrospectively retrieved on children receiving GH supplementation, assessing their disease-free and overall survival outcomes and risk of secondary malignancies using Kaplan-Meier and Cox models. Overall 65 children were consecutively collected from May 1981 to April 2013. All patients had undergone craniospinal irradiation (total dose 18-39 Gy), and subsequently received GH for a median (interquartile range, IQR) of 81 (50.6-114.9) months. At a median (IQR) of 122.4 months (74.4-149.5) after the end of their adjuvant cancer treatment, two patients (3 %) experienced recurrent disease and 8 (12.3 %) developed secondary malignancies, all but one of them (an osteosarcoma) related to radiation exposure and occurring within the radiation fields. There was no apparent correlation between the administration of GH replacement therapy (or its duration) and primary tumor relapse or the onset of secondary malignancies [HR: 1.01 (95 % CI: 0.98, 1.03) for every additional 12 months of GH supplementation; p = 0.36). At univariate analysis, the large cell or anaplastic medulloblastoma subtype, metastases and myeloablative chemotherapy correlated with a higher risk of secondary malignancies (p < 0.1), but multivariate analysis failed to identify any factors independently associated with this risk. Our data supports once more the safety of long-term GH replacement therapy in children treated for medulloblastoma/PNET, previously reported in larger data sets. The neurooncology community now need to warrant large-scale meta-analyses or international prospective trials in order to consolidate our knowledge of factors other than GH, such as genetic predisposition, high-grade/metastatic disease, high-dose chemotherapy and era of treatment, in promoting the occurrence of secondary malignancies.