Pediatr Radiol (1989) 19:100-103 Pediatric Radiology © Springer-Verlag 1989 Pancreatic venous samplings in infants and children with primary hyperinsulinism* F. Brunelle, V. Negre, M.O. Barth, C. N. Fekete, P. Czernichow, J. M. Saudubray, F. Kuntz, T. Tach and D. Lallemand Department of Radiology, Hopital Enfants Malades, Paris, France Abstract. The authors present 19 cases of hyperinsu- linism in children worked up with selective pancre- atic venous samplings (PVS). Focal lesions were found in 7, diffuse secretion in 8 and normal insulin levels in 4. In three patients with focal hypersecre- tion less extensive surgery could be performed and confirmed the presence of focal lesions in two. These preliminary results are encouraging and PVS seems to be a valuable technic for detection of focal lesions in the pancreas of children with hyperinsulinism. an adenoma by splenic and portal blood sampling has been reported [15], to our knowledge no case of preoperative localisation has ever been reported in children. Pancreatic venous sampling (PVS) is well estab- lished method for preoperative localisation of pan- creatic secreting tumors in adults [16-18]. We wish to report our experience with PVS in 19 infants and children presenting with primary hyperinsulinism. Hyperinsulinism is the main cause for hypoglycemia in infants during the first months of life [1]. The diagnosis is made biologically when low levels of blood glucose are associated with normal or high levels of blood insulin. The treatment is difficult, based on dextrose in- fusions and inhibition of insulin secretion with Diazoxide, or more recently with glucagon, somato- statine or steroids [2, 3]. In case of resistance to the medical treatment, surgical pancreatectomy is performed to prevent neurologic sequelae, a common complication of neonatal severe hyperinsulinism [4-7]. The pancreatic lesions responsible for the hyper- insulinism are various: hyperplasia of the islets of Langherans, adenomatosis, nesidioblastosis and true adenomas [8-14]. Such lesions are diagnosed in retrospect on the pancreatic specimen after surgery. Preoperative di- agnosis and localisation of a focal lesion would change the surgical technique and would allow surgery to be more accurate. Although one case of preoperative localisation of * Presented at the ESPR meeting Montreux 1988. Selected for publication by an International Group of the ESPR Material and method We examined 19 children aged from 1 to 17 with proven hyper- insulinism. Ten of these are below 1 and there is a strong male predominance (14 boys, 5 girls). All patients had normal ultrasound study of the pancreas. Pancreatic venous samplings All drugs stopped 24 hours before the examination. Dextrose in- fusion is precisely adapted to the needs. Under general anesthesia a catheter is placed in an hepatic vein using a fight femoral vein approach. In small infants, a wedge injection is done in an hepatic vein causing retrograde opacification of the portal system. The knowledge of the anatomical location of the portal vein facilitates the hepatic puncture (Fig. 1). The liver is punctured on the mid axillary line with a 18 or 20 G catheter needle (Roche). A specially designed catheter is placed through this into the portal system. Multiple samplings are per- formed in the splenic, superior mesenteric, inferior mesenteric and portal veins as well as in pancreatic collaterals (Figs. 2, 3, 4). As many as 20 samples are taken in a single child. Samples from the hepatic veins are taken at regular intervals as well, to check the peripheral levels of insulin. Phlebography is performed in any collateral sampled to allow a complete and precise reconstruction of the venous anatomy of the pancreas, as variations are known to be frequent [19]. Results of insulin and glucose levels are mapped on a pre- printed drawing of the portal venous system (Fig. 5). Arteriography is not routinely performed, but may be of value in children older than 10.