Histologic findings in persistent hyperinsulinemic hypoglycemia of infancy: Australian experience

Persistent hyperinsulinemic hypoglycemia of infancy (PHHI) is characterized by hyperinsulinism and profound hypoglycemia, with most children requiring pancreatic resection. The histological classification of PHHI is controversial. Most authors acknowledge the existence of focal areas of islet cell proliferation (adenomatosis) in 30%-50% of cases and a diffuse disorganisation of islet architecture, termed 'nesidiodysplasia,' in others. De Lonlay et al. reported that cases with adenomatosis are focal with normal remainder of pancreas and that focal and diffuse disease can be differentiated intraoperatively, on the basis of increased β-cell nuclear size found only in the diffusely abnormal pancreas. We have examined pancreatic histology in a blinded controlled study of PHHI patients. Pancreatic tissue was obtained at autopsy from 60 normal subjects (age 17 weeks gestation to 76 years) and from surgical specimens of 31 PHHI patients. Sections from PHHI subjects (n = 294 blocks) and control sections were stained with hematoxylin and eosin, insulin, glucagon, somatostatin, NSE, cytokeratin 19, and vimentin. Three sections from each PHHI patient were randomly chosen for further analysis. Age-matched control (n = 34) and PHHI sections (n = 66) were examined, with the identity of subjects concealed. A diagnosis of normal histology, adenomatosis, or diffuse nesidiodysplasia was recorded for each section. The presence of large β-cell nuclei (>19 μm), ductuloinsular complexes, and centroacinar cell proliferation was noted. Of a total of 65 subjects examined (34 control and 31 PHHI), 37 subjects were identified as normal on both sections examined. All the control cases were correctly identified as normal and none had large β-cell nuclei or centroacinar cell proliferation. Of 31 PHHI patients, 28 were identified as abnormal, either on the basis of abnormal architecture and/or abnormally large β-cell nuclei. Three patients were identified as normal in both sections. Fifteen of 31 patients had diffuse nesidiodysplasia only. Of 13 patients with areas of adenomatosis, 2 had resection of a nodule with adenomatosis present in most of the tissue removed at surgery. Nine patients had a diagnosis of adenomatosis in one section and a diagnosis of diffuse nesidiodysplasia in the other sections from nonadjacent pancreas. Only 2 of 31 PHHI cases had adenomatosis on one section examined and normal pancreas on the other section examined. Large β-cell nuclei were variably found in PHHI sections. Only 5 of 15 patients with diffuse nesidiodysplasia had large nuclei in both sections examined. Centroacinar cell proliferation was identified in 12 PHHI subjects, 6 with adenomatosis and diffuse nesidiodysplasia and 6 with diffuse changes only. It was patchy in distribution within sections and present in only one section in 7 of the 12 subjects. In summary, we have shown that a blinded observer could differentiate control and PHHI pancreatic tissue. Only 2 of 31 patients (6%) had focal adenomatosis with normal nonadjacent pancreas, the majority (24 of 31) had diffuse nesidiodysplasia affecting the remainder of their pancreas, with 38% (9 of 24) also having areas of adenomatosis. Large β-cell nuclei did not reliably identify those with diffuse disease in this study. There was evidence of significant ductal and centroacinar proliferation in 39% of PHHI cases, which was not observed in any of the controls. We have shown that PHHI subjects have a spectrum of pancreatic histological abnormalities, from no abnormality to diffuse subtle changes to florid adenomatosis. Patients could not be segregated into subtypes for different operative intervention despite the availability of full immunohistochemical staining.