Histological characteristics of patients with 'cryptogenic' cirrhosis and prior biopsy showing NASH
Lee, Vanessa Daisy, Department of Public Health Sciences, University of Virginia
Department of Public Health Sciences, University of Virginia
Previous serial biopsy studies have demonstrated that the course of nonalcoholic steatohepatitis (NASH) may include a late stage in which steatosis decreases to the point that evidence of antecedent NASH may be absent or evident only by residual markers of fatty liver disease. Epidemiological data including risks factors for NASH, such as obesity, non-insulin dependent diabetes, and hypertriglyceridemia, and post-transplant recurrence suggests that NASH is a common precursor to cryptogenic cirrhosis. In order to expand the existing data and because only a handful of cases exist demonstrating the histological progression of NASH to cryptogenic cirrhosis, we sought to characterize late stage histology in patients with prior biopsies showing unequivocal non-cirrhotic NASH. Methods: The case records at the University of Virginia were reviewed to identify patients who had at least two biopsies wherein a biopsy and clinical evaluation had demonstrated non-cirrhotic NASH, and there existed a second biopsy showing cirrhosis without definitive findings of NASH. Retrieved biopsies were admixed with other NASH, cryptogenic cirrhosis, and hepatitis C specimens to insure adequate blinding. All biopsies were reviewed and scored using the NASH-Clinical Research Network (NASH- CRN) scoring system. Results: Fourteen biopsy specimens from 7 patients meeting criteria were examined, including from one patient who was post-liver transplant for NASH-related cirrhosis. In comparison to 13 hepatitis C cirrhosis specimens, there revealed no significant differences in the grades of macrosteatosis, portal and parenchymal inflammation. No acidophil bodies were identified in NASH or NASH cirrhosis specimens, however were 2 evident in hepatitis C specimens (6 out of 13). Over a median of 6 years, there was a trend towards complete loss of macrosteatosis as NASH progressed to cirrhosis (p=0.0699). Certain features of NASH remained preserved even in bland cirrhosis such as glycogenated nuclei, balloon cells, Mallory bodies, megamitochondria, and pigmented macrophages. Conclusions: Consistent with existing epidemiological data, our results suggest that histological features of late stage NASH with progression to cirrhosis is characterized by minimal residual steatosis and the presence of glycogenated nuclei, Mallory bodies, balloon cells, microgranuloma, large lipogranulomas, megamitochondria, and pigmented microphages. 3 Non-alcoholic steatohepatitis (NASH) is one of the most common of all liver diseases and is increasing in prevalence. This disease has historically been associated with obesity, non-insulin-dependent diabetes mellitus, and hypertriglyceridemia; as further epidemiologic data surfaces, it appears as if this condition also occurs in those with normal body weight or hyperglycemia. 1 It is estimated that 150f people with this disease will develop cirrhosis within 5-10 years. 2, 3 Attributes or predisposing factors leading to disease progression are unknown. Steatosis is usually regarded as an essential finding for the diagnosis of NASH. As the disease progresses to cirrhosis, many of the characteristic and diagnostic findings on biopsy may decrease including loss of fatty infiltration. 2 The loss of hallmark steatosis during disease progression is thought to occur due to altered cellular metabolism in regenerating nodules which are characteristic of cirrhosis. As previous epidemiological studies have suggested, NASH may be an underrecognized cause of cryptogenic cirrhosis. It has been suspected that the growing number of cryptogenic cirrhosis is due to burnt out NASH. In a case series by Ayata et al. 4 , an estimated 331620830atients who had liver transplantation for cryptogenic cirrhosis had an underlying diagnosis of NASH. Cryptogenic cirrhosis accounts for up to 5-15ases of cirrhosis 5, 6 and about 120f all liver transplants. 7 NASH is often identified as probable cause due to the presence of known risk factors, overall clinical picture, and the exclusion of other recognizable etiologies. The prevalence of diabetes and obesity in cryptogenic cirrhosis is similar to that of patients with NASH. 8-10 4 Despite this, there are few reports documenting the natural histological progression of NASH to cirrhosis. Patients with prior biopsy-proven NASH seldom undergo repeat biopsy if they present with cirrhosis as the diagnosis of advanced NASH is often assumed, and patients presenting only with advanced cirrhosis often lack prior histological data to prove the existence of antecedent NASH. The lack of pathognomonic clinical characteristics, serologic markers, and histological features in cirrhosis contribute to the difficulties in the diagnosing NASH as the underlying etiology. In order to expand the existing data and because only a handful of cases exist which demonstrate the histological progression of NASH to 'cryptogenic' cirrhosis, we sought to characterize late stage histology in patients with prior biopsies showing unequivocal non-cirrhotic NASH. Methods We reviewed the electronic medical records at the Digestive Health Center, University of Virginia Health Center to identify patients who have the clinical diagnosis of NASH and had at least two liver biopsies, wherein a biopsy and clinical evaluation had demonstrated non-cirrhotic NASH and there existed a subsequent histology specimen showing cirrhosis. Seven eligible patients were identified who met these criteria. For patients who had had multiple liver biopsies, the earliest liver biopsy specimen noting NASH and the earliest liver biopsy identifying cirrhosis or stage 4 fibrosis were used for evaluation. The retrieved NASH and NASH cirrhosis histological specimens were 5 originally obtained from 1992 to 2007; these biopsy specimens were admixed with other specimens obtained from twelve NASH and cryptogenic cirrhosis patients to insure adequate blinding. Thirteen control histological specimens were chosen from patients with hepatitis C who had undergone orthotopic liver transplantation at the University of Virginia between June 2006 and April 2007. The reviewed material consisted of 51 total specimens of which 17 were operative wedge biopsy specimens and 34 were percutaneous needle biopsy specimens. All specimens were fixed in formalin and stained with hematoxylin and eosin. All specimens were reviewed and scored blindly by a hepatopathologist (DEK) using the NASH- Clinical Research Network (NASH-CRN) scoring system. 11 This scoring system was developed in coordination with National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) to address the spectrum of lesions found in nonalcoholic fatty liver disease. It is comprised of 14 histological features that are scored semi-quantitatively. All samples were scored using the NASH-CRN categorical scoring system along 14 histological features. Comparisons between histological feature scores were performed between the early NASH to the subsequent NASH cirrhosis specimens and the NASH cirrhosis to those from hepatitis C cirrhosis specimens. Clinical and demographic information were collected via review of the University of Virginia electronic medical record (Carecast/Centricity, GE Healthcare, United Kingdom). Co-morbid diseases and medication use were noted to be present if documented in the clinical chart at any time prior to liver biopsy noting cirrhosis. Fisher's exact test was used in the statistical comparison the proportion of each histological feature identified in the early NASH, 6 NASH cirrhosis, and hepatitis C groups. Signed rank test was used in the comparison of weight and body mass index (BMI) in the early NASH and NASH cirrhosis groups. All statistical analysis was performed using SAS version 9.1 (SAS Institute Inc., Cary, NC, USA). The protocol for this study was reviewed and approved by the University of Virginia Institutional Review Board. Results A total of 20 patients were included in this study. Seven patients were identified as having clinical and histological diagnosis of NASH who had subsequent biopsy with histological features of cirrhosis, and thirteen patients with hepatitis C cirrhosis undergoing orthotopic liver transplantation were included in the study. Demographic and co-morbid characteristics of each group are noted on Table 1. Only one out of seven NASH patients was male in comparison to 9 of 13 hepatitis C patients (p=0.0573). One patient in both the NASH and hepatitis C groups had previously undergone liver transplantation. Notably, there was a significant proportion of diabetes and hyperlipidemia in the NASH population in comparison to the hepatitis C group (p=0.0223; p=0.0072). As a consequence, a greater number of NASH subjects received pharmaceutical treatment with metformin (p=0.0307), troglitazone (p=0.0072), and statin (p=0.0307) for their diabetes and hyperlipidemia. There was no significant difference in the use of insulin (p=0.5868). 7 Histology Comparison of NASH to NASH Cirrhosis Biopsy We compared the histological findings of an earlier non-cirrhotic biopsy to a later biopsy noting cirrhosis in patients diagnosed with NASH. Seven patients were identified in this group who had paired liver specimens. All but one subject were women, and the subjects' ages ranged from 29 to 58 years (mean 47.857 years, median 55 years). All NASH patients were obese at first biopsy, BMI ranged from 32.2 to 42.5. The mean weight and BMI were 106.693 kg ± 24.433 and 38.114 ± 3.709 at initial liver biopsy. Six years was the median time period between biopsy documenting early NASH with fibrosis stage of 2 (3 subjects) or 3 (4 subjects) to cirrhosis. During this time only one patient did not gain weight, however the median change in weight and BMI were 8.10 kg and 2.85 (p=0.0781, p=0.0781). Table 2 lists this information. In Table 3, the comparison of NASH-CRN histological features for the early NASH and NASH cirrhosis groups are listed. There was a trend towards significant loss of macrosteatosis as NASH progressed to cirrhosis. When histologically classified as either having or not having macrosteatosis, 4 out of 7 subjects had complete loss of steatosis in the cirrhosis biopsy specimen compared to the earlier NASH biopsy where all 7 subjects had macrosteatosis (p=0.0699). There was no significant change in microsteatosis. All early NASH and cirrhosis specimens had balloon cells; indeed the majority of subjects had prominent ballooning evident in early NASH (5 out of 7) and continued to have prominent ballooning in cirrhosis (5 out of 7).8 Inflammation in the portal area increased as subjects progressed from early NASH to cirrhosis (p=0.0291) whereas parenchymal inflammation remained stable. Other histological features, such as microgranulomas, large lipogranulomas, pigmented macrophages, megamitochondria, Mallory bodies, and glycogenated nuclei, were noted in both early NASH and cirrhosis specimens. There were no significant changes in the presence or absence of these features in the cirrhosis samples compared to the early NASH specimens. However in this series, no acidophil bodies were identified in either NASH or subsequent cirrhosis groups. Histology Comparison of NASH Cirrhosis to Hepatitis C Cirrhosis Biopsy In this comparison, histological features as noted within the NASH-CRN score were compared between seven of NASH cirrhosis and thirteen of hepatitis C specimens. A summary of the results are noted on Table 4. When comparing the presence and absence of macrosteatosis and comparing the grade of macrosteatosis, there were no significant statistical difference between the NASH cirrhosis and hepatitis C groups (p=0.2898, p=0.3544). In 3 hepatitis C specimens, there was no observable macrosteatosis. There were also no significant differences in the presence or absence of microsteatosis, pigmented macrophages, microgranuloma (p=0.1219), lipogranuloma and glycogenated nuclei (p=0.1736). Also noted was the lack of statistical differences in the grade of portal inflammation, as all NASH cirrhosis and hepatitis C specimens were graded 1 or 2 (p=0.3500), and parenchymal inflammation (p=0.5097). The presence and absence of the histological features of megamitochondria and Mallory bodies were both 9 statistically significant (p=0.0307 and p=0.0223). The presence or absence and the grade of balloon cells were also significantly difference between NASH cirrhosis and hepatitis C cirrhosis specimens (p=0.0147 and p=0.0320). Although not statistically significant, there was a trend towards significance in the presence or absence of acidophil bodies. No acidophil bodies were identified within the NASH cirrhosis (0 out of 7) specimens, however they were noted in 6 out of 13 hepatitis C cirrhosis specimens (p=0.0515). Discussion Epidemiologic data strongly supports the progression of NASH to cryptogenic cirrhosis, however few histological studies are available that clearly document this progression. Review of the literature identifies only six patient examples documenting the progression from fibrosis to cirrhosis. 1-3, 12, 13 These were isolated cases describing NASH at various stages of fibrosis (mostly stage 2 or 3) that progressed to cirrhosis over a time span of 1.5 to 7 years. In our study, we evaluated seven cases of NASH cirrhosis and substantially expanded the published literature in describing the histologic progression of NASH to NASH cirrhosis. The principal goal of this study is to characterize late stage histology of NASH in comparison to hepatitis C cirrhosis and earlier stages of NASH. This study confirms the long suspected trend of the macrovesicular steatosis loss as fibrosis stage increases. Four of the seven NASH patients had almost complete loss of 10 steatosis on the NASH cirrhosis biopsy. Powell et al. 2 reported on similar loss of macrosteatosis as one patient progressed from fibrosis to cirrhosis in 8 years. Two possible mechanisms for the loss of fatty infiltration as NASH progresses in fibrosis stages have been described. One hypothesis suggests that the loss of hepatic fat may be due to altered and reduced hepatic blood flow. 14 Wanless and Shoita 15 present a four-step model regarding the pathogenesis of NASH and progression to cirrhosis. It is proposed that in the case of progressive steatohepatitis, increasing sinusoidal and vein damage occurs as further steatosis and inflammation distort microcirculation. 16 These veins are obliterated, healed, remodeled, and then become collapsed; with continued damage, the accumulation of these lesions leads to cirrhosis. An alternative, but related, hypothesis suggests that there is limited blood flow in hepatic sinusoids clogged with fatty infiltration; as sinusoidal blood flow slows, blood is shunted along alternative routes via portosystemic shunts. These blood flow changes subsequently lead to impaired insulin, glucose, and lipoprotein delivery. 17 Within this study, parenchymal inflammation remained stable while increased portal inflammation developed with progression to cirrhosis. This is contrary to the results presented by Powell et al. 2 and Lee et al. 3 in which cases of NASH cirrhosis had either stable or decreasing amounts of overall inflammation. Review of the NASH histologic studies generally indicate lobular inflammation being a distinct trait of NASH. Lobular inflammation remained evident in both NASH and cirrhosis specimens, however increased grade of portal inflammation was apparent with progression to cirrhosis specimens. Despite this increased grading, no NASH cirrhosis specimens were described 11 to have more than mild portal inflammation. Mild portal inflammation has been described in NASH but rarely noted to be marked. 18 Whereas macrosteatosis and lobular inflammation have been hallmark traits in the histological diagnosis of NASH in non-cirrhosis patients, these two traits may not be as useful in the evaluation cryptogenic cirrhosis suspected to be secondary to NASH. Other histological features noted in non-cirrhosis NASH remain identifiable in NASH cirrhosis. Megamitochondria, pigmented macrophages, large lipogranuloma, microgranuloma, glycogenated nuclei, and Mallory bodies remained consistently present within both NASH specimens and NASH cirrhosis specimens. In spite of the loss of macrosteatosis, the "cryptogenic" or NASH cirrhosis specimens had identifiable features to aid in distinction from our control group of hepatitis C cirrhosis. The NASH cirrhosis and hepatitis C cirrhosis groups did not have significant differences in macrosteatosis, or portal and parenchymal inflammation. In this series, histologic features as described in the NASH-CRN scoring system were distinct and significantly differed from these two groups. Features, such as balloon cells, Mallory bodies, and megamitochondria, were observed in NASH cirrhosis specimens and may aid in the identification of NASH a potential etiology for other cases of bland cirrhosis. The blinded pathologist in this study identified all NASH and NASH cirrhosis specimens as probable or definite cases of NASH. The NASH subjects of this study were chosen on the basis of known paired biopsy specimens; cirrhosis specimens were, in general, identified as bland cirrhosis or cirrhosis with features of steatohepatitis while the patients were undergoing clinical evaluation. The identification of NASH-related liver 12 disease, as either probable or definite, by the hepatopathologist involved in the study was correct in all cases. In clinical practice, a pathologist specializing in liver histology and pathology is often unavailable. In these cases, identification of these others features would be beneficial in the correct diagnosis of NASH-related cirrhosis from bland cirrhosis. Evaluation of clinical and histologic features that predict risk of disease progression has been attempted. In a study by Matteoni et al. 19 , it was suggested that patients with liver pathology containing fat accumulation, balloon degeneration, and Mallory hyaline or fibrosis had a higher incidence of cirrhosis or liver-related death compared to those with simple steatosis and lobular inflammation. Correlating with that observation, our series noted balloon degeneration and fibrosis in every initial NASH biopsy that later progressed to cirrhosis. Mallory hyaline was also present in the majority of the NASH patients. Acidophil bodies are histological markers of cell necrosis; however in this series no acidophil bodies were identified in NASH or NASH cirrhosis specimens. This histological feature is often described within NASH patients; in several series, roughly 600f NASH patients at different fibrosis stages had the presence of eosinophilic intracytoplasmic inclusion bodies. 11, 20 Although difficult to reproduce acidophil bodies quantification, Brunt et al. 18 noted a trend of increasing number of acidophil bodies within increasing necroinflammatory grade of NASH. Signs of hepatocyte apoptosis have also been described more recently by Feldstein et al. 21 They identified increased Fas expression and immuno-histological signs of apoptosis using the TUNEL-assay in 13 patients with more severe NASH, thus suggesting that apoptosis is either an indication of disease severity or contributing to the process of further development of inflammation and fibrosis through other cellular mediators. Our findings are contrary to those described by these previous studies; acidophil bodies, a sign of apoptotic necrosis, are not present in this case series of NASH with progressive fibrosis. Given this conflicting data, further study is certainly warranted. NASH with progression to cirrhosis has been noted to occur in 10-150f postliver transplantation patients who had initially received transplantation for cryptogenic cirrhosis. 22, 23 Growing evidence supports the concern that post-liver transplantation NASH represents recurrence of this disease. Contos et al. 22 described the seemingly benign presentation steatohepatitis in 100f patients who had received liver transplantation for cryptogenic cirrhosis. However, this report was limited by a median follow-up of 3.5 years. In our series, one subject developed histological signs of recurrent NASH after receiving liver transplantation for cryptogenic cirrhosis 6 years prior. Progression from steatohepatitis with fibrosis to cirrhosis was confirmed via biopsy over a time span of 8 years. Similarly, Sutedja at al. 24 reported on 2 such patients while following 32 cryptogenic cirrhosis post-transplant patients; cirrhosis reoccurred in these 2 patients within 72 and 84 months following transplant. These cases would support the progressive nature of steatohepatitis and raises concern for the risk of developing recurrent cirrhosis following liver transplantation in NASH and cryptogenic cirrhosis patients. 14 There are several limitations to a study such as this. One weakness is the small sample. However, this series has expanded the published record of documented NASH progression to cirrhosis from 6 1-3, 12, 13 to a total of 11 patients. Another limitation of this study is the inability to correlate histological features with specific serologic markers. Many of these patients were referred to our institution following the diagnosis of NASH; due to difficulties obtaining patient documents prior referral to the University of Virginia, we were unable to consistently pair serologic and laboratory values with pathologic samples for the purpose of this study. Despite these limitations, this study contributes much needed histologic documentation regarding the progression of NASH fibrosis and histologic features of NASH cirrhosis. Finally, further studies in to the pathophysiology and potential areas for treatment of NASH are active and ongoing. Although bland and "cryptogenic" cirrhosis pose difficult diagnostic challenges, there are residual NASH histologic features that remain even in cirrhosis that may aid in a diagnosis of underlying NASH. It is clear that we cannot rely on the presence of significant macrosteatosis as a hallmark trait of NASH in cirrhosis patients. Further studies are much needed to aid in the production of criteria for the diagnosis of NASH cirrhosis.
Note: Abstract extracted from PDF text
MS (Master of Science)
All rights reserved (no additional license for public reuse)