Diabetic hepatosclerosis: A novel entity or a rare form of nonalcoholic fatty liver disease?


  • Potential conflict of interest: Dr. Younossi is on the Speakers' Bureau and received grants from Axcan.



This study sought to characterize a noncirrhotic form of hepatic sinusoidal fibrosis in patients with long-standing diabetes mellitus (DM) who underwent liver biopsy. The authors suggest that diabetic hepatosclerosis is distinct from other hepatic manifestations of DM.


Clinical records, detailed light microscopy results, and immunohistochemical preparations of twelve diabetic patients presenting with characteristic histologic findings were reviewed. Liver biopsy slides were evaluated for the degree of perisinusoidal, perivenular, and portal fibrosis, along with other histologic findings. When available, unstained biopsy sections were used for additional immunohistochemical analyses. Medical records were examined for relevant clinical data. Patients were chosen from Saint Louis University School of Medicine and the Armed Forces Institute of Pathology.


All patients presented with a long-standing history of DM and most had evidence of microvascular complications resulting in end-organ damage. Alkaline phosphatase elevation was common. Immunohistochemical staining of liver biopsies showed basement membrane components in a perisinusoidal distribution. Hematoxylin-eosin and trichrome staining of the biopsies revealed dense perisinusoidal fibrosis while lacking other features commonly seen in non-alcoholic steatohepatitis (NASH).


These findings suggest a form of diabetic microangiopathy affecting the liver, described as diabetic hepatosclerosis. Further investigation must be done to understand the pathogenesis and clinical significance of this novel entity.

Harrison SA, Brunt EM, Goodman ZD, Di Bisceglie AM. Diabetic hepatosclerosis: diabetic microangiopathy of the liver. Arch Pathol Lab Med 2006;130:27–32.(Reprinted with permission from Archives of Pathology & Laboratory Medicine. Copyright 2006. College of American Pathologists.)


Hepatic sinusoids consist of delicate capillaries that lack a genuine basement membrane. They are limited by fenestrated endothelial cells and surrounded by a number of other cells. Certain conditions may cause basement membrane formation in hepatic sinusoids, often leading to fibrosis of the surrounding tissue. This fibrosis with basement membrane formation results from hepatic stellate cell (HSC) activation, leading to excess production of extracellular matrix (ECM) similar to the pathologic lesions seen in some types of chronic liver disease.

In addition to chronic liver disease, perisinusoidal fibrosis may be caused by other processes. High levels of blood glucose may damage the sinusoidal endothelial cells via a different mechanism. Hyperglycemia can modify intracellular proteins producing advanced glycation end products (AGEs).1 These AGEs may travel outside the cell, adhering to extracellular molecules critical in the signaling pathways involved between hepatocytes and their surroundings causing abnormal cellular function.2 More specifically, work has shown that AGEs may bind to their receptors causing an upregulation of pro-inflammatory cytokines inducing a fibrogenic cascade.3–7 These hyperglycemia-induced pathways could potentially lead to thickening of sinusoidal capillary walls and perisinusoidal fibrosis as well as basement membrane thickening. Such vascular complications, or diabetic microangiopathies, affect a number of organs contributing to the development of diabetic heart disease, kidney disease, retinopathy, neuropathy and dental disease.8 While the connection between diabetes and these conditions has been established, a similar association between diabetic microangiopathy and liver disease has been poorly characterized. Nevertheless, a few cases of perisinusoidal fibrosis without NASH has previously been described in diabetic patients.9

In an attempt to identify this potential hepatic manifestation of diabetic microangiopathy, Harrison et al. provide an in-depth study of twelve diabetic patients with available liver biopsies. Immunohistochemical, histologic, and clinical data were presented to support their findings. Indeed, in the majority of liver biopsies reviewed, steatosis was not noted, with only one case having thirty percent macrovesicular steatosis. Masson trichrome staining demonstrated dense collagen distribution within the perisinusoidal space, in a fashion similar to that seen in other illnesses associated with diabetic microangiopathy. Additional immunostaining also showed the presence of a basement membrane and basement membrane components such as laminin. No histologic features of cirrhosis were found in any of the biopsies. Review of clinical data also noted elevated levels of alkaline phosphatase (ALP) rather than serum aminotransferases which is usually seen in patients with nonalcoholic fatty liver disease (NAFLD).10 This finding is especially significant because ALP originates mainly from the biliary system. Thus, increased levels of this enzyme indicate that diabetic hepatosclerosis may be associated with nonparenchymal hepatic tissue involvement, as opposed to the direct parenchymal effects associated with NAFLD.

While the histological and immunohistochemical evidence in this study provides some proof for diabetic hepatosclerosis, more work needs to be done to substantiate its existence. Among the weaknesses of the study is its retrospective nature, which may provide a source of bias as well as incomplete data. Lack of detailed medical histories for some of the patients may fail to account for potential confounders. Therefore, one cannot exclude other concomitant factors (i.e., hepatotoxins) as potential causes for the lesions seen in this patient population. In fact, toxins such as benzene and xylene have been reported to cause liver damage somewhat similar to what was seen in this series of patients.11 Additionally, the retrospective analysis could not distinguish between type 1 and type 2 diabetes in these patients. Ascertaining whether this entity is related to hyperglycemia or insulin resistance is critical in shedding light on its pathogenesis. Finally, since approximately 6% of the general population has diabetes, “diabetic hepatosclerosis” must be quite rare to be repeated in only 12 cases from two major pathology centers. Additionally, it would have been interesting to compare this group with a well matched group of patients with NASH to assess for any clinical differences. Nonetheless, the authors present an interesting case series describing a potentially novel liver disease which may be related to diabetes.