ESPE Abstracts (2019) 92 P1-251

Central Diabetes Insipidus in Children: Role of GH Antibodies

Flavia Napoli1, Fabiana Pani2, Francesca Gianti3, Natascia Di Iorgi3,1, Giovanni Morana4, Anna Elsa Maria Allegri1, Hanan Farid Mufleh Al_Thiabat3, Annalisa Gallizia3, Daniela Fava3, Chiara Longo3, Camilla Olcese3, Francesco Vinci3, Angela Pistorio5, Patrizio Caturegli2, Mohamad Maghnie3,1


1Clinica Pediatrica, Istituto Giannina Gaslini, Genoa, Italy. 2Division of Immunology, Department of Pathology, The Johns Hopkins School of Medicine, Ross Research Building-Room 656,, 720 Rutland Avenue, Baltimore, MD, 21205, USA. 3Università degli Studi di Genova, Genoa, Italy. 4Neuroradiology Unit, Istituto Giannina Gaslini, Genoa, Italy. 5Servizio di Epidemiologia e Biostatistica, Istituto Giannina Gaslini, Genoa, Italy


Central diabetes insipidus (CDI) in children is caused by brain tumors, Langerhans cell histiocytosis (LCH), trauma, infections, or genetic abnormalities in about 60% of the cases. In the remaining 40%, CDI is idiopathic even after detailed clinical and radiological investigations. Aim of the study was to assess whether measurement of serum antibodies against human growth hormone (GH) could aid in the identification of the etiological factors for CDI.

We followed 65 children with a diagnosis of CDI established at a single centre (Gaslini Hospital, Genoa, Italy) between March 2000 and May 2018. Of them, 36 were females and 29 males. Their diagnosis upon detailed clinical and morphological follow up was idiopathic (No.=31), brain tumor (No.=24), LCH (No.=6), genetic defects (No.=3), or pituitary abscess (No.=1). The median age at diagnosis and median follow up time were 8.9 and 8.5 years, respectively. Brain MRI and studies of anterior pituitary functions were performed in all patients at baseline and on follow up. Serum antibodies against native human GH were measured by a in-house ELISA assay where values below 50arbitrary units per mL are considered negative. In 8 children two serum collections were available and in 4 three, for a total of 81 sera.

Age at diagnosis was significantly higher in patients with brain tumor. Anterior pituitary hormone defects were more frequent in patients with LCH even before specific treatment began(P=0.008), while the association between anterior defects and tumors was significant only after treatment (surgery, radiation or chemotherapy). Patient with normal pituitary stalk at disease onset were more likely to develop GH deficiency during follow up(P=0.043). The median nGH antibody levels were overall similar in the 5 diagnostic categories, being in all cases<50 AU/mL. Interestingly, however, very high GH antibody levels(>200) were seen only in patients with idiopathic CDI. When sequential sera were available, GH antibodies decreased over time. GH antibodies were higher in patients without pituitary stalk thickening (P=0.032), and did not differ according to gender, age at diagnosis, age at time of blood test, disease duration, treatment type, or anterior pituitary hormone defect.

The study confirms the challenge of discovering new etiologic factors in idiopathic CDI, as well as identifying factors that predict a specific etiology. GH antibodies do not seem to associate with specific CDI etiologies, although the highest levels observed in the idiopathic form suggest autoimmunity as a possible contributor to the pathogenesis of this condition.