Section 3 -

Potential Differentials Michael Torbenson

1. Introduction We have all had liver biopsies from time to time where the liver tissue showed no or minimal changes despite a clinical history of persistently abnormal liver enzymes. In a typical sort of case, viral hepatitis has been excluded clinically, there is no history of patient medications, and the ANA and other autoimmune markers are negative or minimally elevated. Examination of the biopsy may show no inflammation, or only mild non-specific inflammation, and no evidence for fatty liver disease. What other entities should be considered in this setting? While the potential differential is very long indeed we will consider 4 specific entities (Table 1). Table 1: A Sample of Diagnoses to Consider in the Setting of Persistent Liver Enzyme Elevations and a Nearly Normal Liver Biopsy (after drug effect, viral hepatitis, autoimmune hepatitis, and fatty liver disease have been excluded). Possible Diagnoses Alpha 1 anti trypsin deficiency Amyloid Celiac disease Crohn's disease of the small bowel Glycogen hepatopathy Glycogen psuedoground glass Hemochromatosis hepatoportalsclerosis Hypervitaminosis A Nodular regenerative hyperplasia Small bowel bacterial overgrowth Thyroid disease Wilson's disease 1. Hypervitaminosis A Case: A 53 year old woman with no significant medical history and taking no current medications had persistent elevations in her ALT and AST (53 and 45 IU/L respectively). She had traveled to Europe approximately 8 months ago but had no other travel history. Viral hepatitis was ruled out serologically. Her ANA was negative. Liver biopsy showed stellate cells with enlarged lipid-laden cytoplasm and a diagnosis of hypervitaminosis A was made. Upon further questioning, she indicated that she had starting taking a herbal supplement while in Europe and had continued upon her return. The precise formulation of the supplement was not clear, but it had a cranberry base. Her enzymes were normal 6 months after she stopped taking the supplement. Excess intake of vitamin A has been a recognized cause of liver disease for over 50 years. However, the enlarged and lipid laden stellate cells can be easily overlooked and the clinical information of excess vitamin A intake is almost always lacking. Thus, a high index of suspicious is usually necessary to make the diagnosis. It seems likely that many cases are missed, as the frequency of hypervitaminosis A diagnosed by liver biopsy has been reported as 1.1% in a retrospective study from a tertiary care medical center. [1] Hypervitaminosis A has been associated with many severe clinical symptoms (Table 2) as well as with liver fibrosis and cirrhosis. Table 2. Clinical Manifestiations and Causes of Hypervitaminosis A Sampling of Clinical Manifestations Causes unexplained fevers unexplained elevations in liver enzymes anorexia headache ascites pleural effusions hepatomegaly bone pain muscle pain Excess vitamin A in nutrient and vitamin supplements Diets unusually enriched in yellow-orange vegetables Topical Retin-A Histologically, the stellate cells are typically enlarged and appear increased in numbers, though whether a true hyperplasia is present is not always clear. The stellate cell cytoplasm is filled with small lipid vacuoles giving the cytoplasm a bubbly appearance and has been described as "stellate-cell lipidosis". [1] The location of the stellate cells and the shape of their nuclei usually permits one to easily distinguish stellate cell lipidosis from fatty liver disease. The stellate cells can be focally or diffusely distributed and may or may not have a zonal distribution. No good immunostain is currently available; smooth muscle actin is not helpful in my experience. An important point to remember: Normal serum levels of vitamin A DO NOT exclude vitamin A toxicity. This is because large proportions of vitamin A circulate as esters bound to plasma proteins and will not be evident in the serum. [2] 2. Glycogenic Hepatopathy Case: A 15 year old girl with type I DM presents for liver biopsy. She has persistent mild elevations in her ALT and AST and an ultrasound shows mild fatty change. A liver biopsy shows diffuse pale hepatocytes with cytoplasmic rarefaction and accentuation of the cell membranes—somewhat resembling plant cells Numerous hepatocytes exhibit glycogenated nuclei, but no fatty liver is seen. Abundant cytoplasmic glycogen deposits are seen on PAS stain and a diagnosis of glycogenic hepatopathy (GH) is made. Glycogenic hepatopathy (GH) results from excess accumulation of glycogen in the liver. GH was first documented as a component of Mauriac's syndrome in 1930, a syndrome consisting of excess glycogen in the hepatocytes, hepatomegaly, abnormal liver enzymes, and other clinical features such as growth retardation and/or dwarfism, delayed puberty, cushingoid features and hypercholesterolemia. Over time, it has been recognized that GH is more often seen in isolation, that is without the full spectrum of changes that define the Mauriac syndrome. GH can be seen in adults or in children with marked or prolonged hyperglycemia who are treated with insulin, usually in the setting of Type I diabetes mellitus. Most individuals will have a history of poor glycemic control, elevated liver transaminases and hepatomegaly. [3] Ultrasound does not readily distinguish fatty liver from glycogen overload, so the patient may have a pre-biopsy clinical impression of fatty liver based on the ultrasound findings. GH demonstrates several key histological findings (1) pale, swollen hepatocytes, (2) no or minimal inflammation, (3) no or minimal spotty necrosis, (4) coincidental fatty change or NASH can be seen but is uncommon (<20% of cases). GH does not appear be a risk factor for developing fibrosis or cirrhosis (in contrast to NASH). Liver transaminases typically return to normal with adequate control of blood sugar levels, even in those cases with marked enzyme elevations. [3] Mechanistically, GH results from excess accumulation of glycogen in hepatocytes. Glucose in the sinusoidal blood is rapidly taken-up by hepatocytes in response to the insulin and this is followed by rapid conversion of the glucose to glycogen, which is then trapped within the liver. [4] Differential: At the time of GH presentation, the clinical differential usually includes NAFLD, from which GH is histologically easily distinguished. The differential may also include glycogen storage disease, as the hepatocytes in both entities are markedly swollen and filled with glycogen. However, clinical parameters, such as the poorly controlled diabetes and response to diabetic control of GH, are key features that distinguish this entity from a glycogen storage disorder. GH is also known to occur following short term high-dose steroid therapy. 3. Glycogen Pseudo-ground Glass Case: A 12 year old boy s/p bone marrow transplant for leukemia has an elevated ALT of 109 and AST of 98 IU/L. A biopsy is performed to rule out GVHD. The biopsy shows no significant inflammation and no evidence for GVHD. However, the zone 3 hepatocytes show patchy inclusions that are strongly PAS positive and resemble HBV ground glass on H&E stain. HBV infection is ruled out clinically and by immunostaining, and a diagnosis of glycogen pseudo-ground glass (GPGG) is made. GPGG is typically seen in immunosuppressed individuals (from HIV, cancer, transplantation, etc) who are on numerous medications. [5] The pseudo-ground glass in GPGG closely mimic ground glass change seen in chronic HBV infection and HBV should be excluded in all cases. The GPCC is strongly PAS positive and variable diastase sensitive. The changes can resolve, but also may persist for years. [5] The clinical significance of glycogen pseudo-ground glass is unclear, but the number of cases with GPGG seems to be increasing in the past several years. Clinically, the presence of significant co-morbid diseases and the numerous medications makes it very difficult to identify a specific drug as the culprit. Likewise, the question come up as to whether the pseudo-ground glass reflects hepatocyte adaptation or hepatocyte injury. In most cases, the indication for biopsy is mild unexplained elevations in transaminases and the glycogen pseudo-ground glass is the only significant histological finding, suggesting it most likely reflects active hepatocellular injury. By analogy, cyanamide pseudo-ground glass, which appears to have many similarities to glycogen pseudo-ground glass, has also been linked in some cases to mildly elevated liver transaminases and hepatomegaly. Differential : Mechanistically, ground glass change in hepatocytes tends to be caused by either smooth endoplasmic reticulum proliferation, abnormal glycogen accumulation, or less commonly protein accumulation. The differential for the pseudo-ground glass change reflects these three general causes and includes drug effects such as cyanamide, Lafora bodies, fibrinogen, and uremia (Table 3). Table 3: Characteristics of Pseudo-ground Glass Change in Hepatocytes Composition of pseudo-ground glass inclusions Immunohistochemical staining results glycogen PAS + diastase sensitive Cyanamide PAS + diastase sensitive Type IV glycogen storage disease PAS + variably diastase sensitive colloidal iron negative LaFora PAS+ diastase resistant colloidal iron positive Fibrinogen PAS - Fibrinogen + C3, C4 positive +/- Uremia PAS + The specific agent(s) that cause GPCC are unclear, but is likely to be drug(s) that inhibit an enzymatic step in glucose metabolism and it is possible that several drugs can have the same effect. It is also unclear why glycogen pseudo-ground glass was more commonly seen in immunosuppressed patients. 4. Celiac Disease. Case: A 25 year old man presents with a history of abdominal bloating, mild intermittent diarrhea and elevated liver enzymes: ALT= 101 IU/L, AST= 78 IU/L AP= 198. He is not taking any medications and viral studies are negative. His ANA is 1:40. His anti-smooth muscle antibodies are negative. The biopsy is performed to "rule out autoimmune hepatitis". The biopsy shows very mild non-specific portal and lobular chronic inflammation. No plasma cells are seen. The liver biopsy also shows very focal and mild macrovesicular steatosis (less than 1%) and no fibrosis. The histological findings are mild and non-specific and the case is signed out descriptively noting that the differential includes celiac disease. Subsequent serological testing confirms a diagnosis of celiac disease. A gluten free diet led to improved clinical symptoms and resolution of the elevated liver enzymes. Approximately 40% of individuals with celiac disease will have elevated transaminases. [6] They typically average about 60 for ALT and 50 for AST, but can occasionally be seen in the several 100's. [6] Of these, nearly all will normalize their enzyme levels following a gluten free diet. However, the normalization often takes several months and may take up to a year. About 4% of individuals with celiac disease have mild elevations in their ANA and 9% have mild elevations of smooth muscle antibodies. [7] The histology of celiac disease is typically very mild and consist of non-specific portal and lobular chronic inflammation + mild fatty change. Less commonly, the biopsies can show chronic active hepatitis or other unusual changes (Table 4). It is also important to remember that PBC, PSC, and autoimmune hepatitis of the usual sort can occasionally be seen in individuals with celiac disease. Table 4: Histological Findings in Individuals with known Celiac Disease and Elevated Transaminases Histological findings in celiac disease Normal appearing liver (15% of cases) [8] Mild non-specific changes (65% of cases) [8] Chronic portal inflammation + mild interface and lobular inflammation (15%) [8] PSC/PBC/autoimmune hepatitis (1-2%) Nodular regenerative hyperplasia (<1%) [9, 10, 11] Acute hepatitis with extensive necrosis (<1%) [12] Inactive cirrhosis (<1%) Interestingly, there are rare reports of celiac disease associated with liver failure and of improvement in liver functions following a gluten free diet. [12] References Levine PH, Delgado Y, Theise ND, et al. Stellate-cell lipidosis in liver biopsy specimens. Recognition and significance. Am J Clin Pathol 2003; 119:254-8. Miksad R, de Ledinghen V, McDougall C, et al. Hepatic hydrothorax associated with vitamin a toxicity. J Clin Gastroenterol 2002; 34:275-9. Torbenson M, Chen YY, Brunt E, et al. Glycogenic hepatopathy: an underrecognized hepatic complication of diabetes mellitus. Am J Surg Pathol 2006; 30:508-13. Munns CF, McCrossin RB, Thomsett MJ, et al. Hepatic glycogenosis: reversible hepatomegaly in type 1 diabetes. J Paediatr Child Health 2000; 36:449-52. Wisell J, Boitnott J, Haas M, et al. Glycogen Pseudo-ground Glass Change in Hepatocytes. Am J Surg Pathol 2006; In Press Duggan JM, Duggan AE. Systematic review: the liver in coeliac disease. Aliment Pharmacol Ther 2005; 21:515-8. da Rosa Utiyama SR, da Silva Kotze LM, Nisihara RM, et al. Spectrum of autoantibodies in celiac patients and relatives. Dig Dis Sci 2001; 46:2624-30. Jacobsen MB, Fausa O, Elgjo K, et al. Hepatic lesions in adult coeliac disease. Scand J Gastroenterol 1990; 25:656-62. Riestra S, Dominguez F, Rodrigo L. Nodular regenerative hyperplasia of the liver in a patient with celiac disease. J Clin Gastroenterol 2001; 33:323-6. Austin A, Campbell E, Lane P, et al. Nodular regenerative hyperplasia of the liver and coeliac disease: potential role of IgA anticardiolipin antibody. Gut 2004; 53:1032-4. Cancado EL, Medeiros DM, Deguti MM, et al. Celiac disease associated with nodular regenerative hyperplasia, pulmonary abnormalities, and IgA anticardiolipin antibodies. J Clin Gastroenterol 2006; 40:135-9. Ojetti V, Fini L, Zileri Dal Verme L, et al. Acute cryptogenic liver failure in an untreated coeliac patient: a case report. Eur J Gastroenterol Hepatol 2005; 17:1119-21.