Reliability of histological criteria in glycogen storage disease of the liver.

The histological criteria for the diagnosis of the hepatic glycogen storage diseases (GSDs) are well recognized. However, some biopsies do not have the characteristic features peculiar to their type and not all biopsies with GSD changes are confirmed by enzyme analysis. We reviewed the liver biopsies of 59 patients with clinically suspected GSD. The enzyme defects in 31 of 40 patients with GSD morphology were demonstrated by enzyme analysis. We describe the history and histology of the 9 patients with GSD morphology not confirmed by enzyme analysis, present the diagnoses of the 19 patients shown not to have a GSD, and evaluate the reliability of the morphological criteria used to distinguish the types of hepatic GSD. In this study the predictive value of a biopsy with GSD changes was 90%. Mosaicism, the most sensitive criterion in the diagnosis of GSD, is not type-specific. Fibrosis does not reliably distinguish between the GSD types and although nuclear hyperglycogenation and lipid are characteristic of type I GSD, these features are not diagnostic of any particular enzyme deficiency. The lack of morphological specificity implies that a complete enzyme analysis be performed on each biopsy. A normal enzyme analysis does not exclude a GSD and careful long-term follow-up may be necessary.