Progression from idiopathic portal hypertension to incomplete septal cirrhosis with liver failure requiring liver transplantation

We report the case of a 30-year-old male patient suffering from what was initially thought to be end-stage cryptogenic cirrhosis with portal hypertension and liver failure, who underwent liver transplantation. Histological examination of the surgical specimen showed incomplete septal cirrhosis. At the age of 17 this patient had presented pancytopenia and splenomegaly, which were treated by splenectomy. The surgeon discovered portal hypertension. Re-examination of the wedge liver biopsy taken at this time revealed features of idiopathic portal hypertension. This case clearly shows that incomplete septal cirrhosis may be a late manifestation of idiopathic portal hypertension. The presence of sinusoidal dilatation and peliosis as well as early evidence of fibrosis which are already visible on the initial biopsy and are still present on the late specimen, are indirect evidence of a continuous process which ultimately led to incomplete cirrhosis with liver failure.

Ultrastructure of human Kupffer cells maintained in culture

Sinusoidal cells were isolated by collagenase perfusion and metrizamide gradient centrifugation, from liver resected for partial hepatectomy performed under warm ischemic conditions. Kupffer cells were then separated from this population by centrifugal elutriation. Isolated Kupffer cells showed good viability, with the typical features of Kupffer cells and were engaged in the endocytosis of foreign particles. They showed numerous morphological criteria of activation. However, cultured Kupffer cells were no longer in an activated state. Kupffer cells were preserved in maintenance cultures for 2 weeks. Purity of these cultures was to 93-97%. During culture, Kupffer cells retained their ultrastructural characteristics and were active in the endocytosis of latex beads and opsonized zymosan particles. It is thus possible that partial hepatectomy performed under warm ischemia could provide valuable material for the study of Kupffer cells in vitro.

Transformation of sinusoids into capillaries in a rat model of selenium-induced nodular regenerative hyperplasia: an immunolight and immunoelectron microscopic study

The oral administration of selenium (Se) to young rats induces, over a 2-month period, the formation of nodular regenerative hyperplasia with sinusoidal damage around nodules. Perinodular areas located in zone 1 comprise atrophic hepatocytes and capillarized sinusoids without fibrosis. We used this unique model of capillarization without fibrosis to investigate the temporal relationship between the process of capillarization and changes occurring in the deposition of components of the extracellular matrix. After 2 weeks of intoxication, type III collagen and fibronectin were stable, but laminin and type IV collagen had increased in zone 1, resulting in the formation of septae between portal tracts. Even at 8 weeks, these two components still formed the principal deposits in perinodular zones. Electron microscopy showed already at 1 week in zone 1 that part of the endothelial wall had detached from hepatocytes. Sinusoidal endothelial cells progressively acquired certain of the characteristics of a vascular endothelium, some proliferated, and perisinusoidal cells transformed into myofibroblasts, surrounded by deposits of laminin and type IV collagen. These results indicate that both laminin and type IV collagen are involved in capillarization without fibrosis and in angiogenesis; fibronectin would not seem to play a role.

Detection of cell-adhesion molecules on human liver-associated lymphocytes

Liver-associated lymphocytes (LAL) from human liver are phenotypically and functionally different from peripheral blood lymphocytes (PBL). Phenotypically, they are mainly represented by the CD3+/-CD56+ phenotype and functionally they spontaneously possess lymphokine-activated killer (LAK) activity. In this study we evaluated the expression of cell-adhesion molecules (CAM) which could be involved in LAL contacts with other sinusoidal cells and/or be responsible for the LAK activity. The LAL population was isolated by sinusoidal high-pressure lavage from partial hepatectomies obtained from patients operated on for benign liver disease (n = 6). Surface expression of the beta 2 integrin chains (CD18, CD11a, CD11b, CD11c), as well as that of members of the immunoglobulin superfamily (CD2, CD54, CD56, CD58), were analysed by one or two-colour flow cytometry. Quantitative and qualitative differences were observed in the expression of CAM between LAL and PBL. LAL were characterized by an increase in the percentages of CD11b+, CD54+, CD56+ and CD58+ cells and a decrease in the percentage of CD2+ cells compared to PBL. Fluorescence intensity values for CD2 and CD56 were higher in LAL than in PBL. Moreover, CD11a/CD18 cells presented a bimodal distribution (dim and bright) in both PBL and LAL; whereas these two subpopulations were equally represented in PBL, the number of bright cells was significantly greater (> 80%) in LAL. The increase in CAM expression (percentage of positive cells and intensity of fluorescence) on LAL combined with their increase in natural killer (NK) and LAK activities already reported, support the idea that, at least some, LAL might be, compared to PBL, in an activated state in vivo.

Preservation of human liver grafts in UW solution. Ultrastructural evidence for endothelial and Kupffer cell activation during cold ischemia and after ischemia-reperfusion

Biopsies taken from 13 human liver grafts at different stages of the transplantation process were used for study of the morphology of sinusoidal cells prior to harvesting (5 biopsies), after preservation in UW solution (10 biopsies), and after complete revascularization (13 biopsies). The mean cold ischemic period was 12 h 30. Immediate follow up was uneventful and the mean peak of post-operative transaminases below 1300 IU/l. Biopsies were perfusion-fixed by the transparenchymal route to ensure satisfactory ultrastructural results. There were no loose sinusoidal endothelial cells in the lumen and no signs of cellular death. Some endothelial cells presented signs of activation at the end of the preservation period, and even more after revascularization, with numerous lucent vacuoles resembling endosomes in the cytoplasm. Kupffer cells also presented signs of activation, particularly after reperfusion. The retraction of endothelial cell processes which formed large gaps during cold ischemia proved to be partly reversible after reperfusion. Signs of endothelial cell damage with gaps and partial rupture of the plasmic membrane were also observed, particularly after revascularization, in areas which contained numerous inflammatory cells adhering to the wall. The Disse space was not generally enlarged and contained no inflammatory cells. The sinusoidal pole of hepatocytes was occasionally damaged with the formation of blebs. These results strongly suggest that any drug or preservation solution that will inhibit endothelial and Kupffer cell activation could be beneficial in the prevention of preservation and reperfusion injury.

Ultrastructure of human liver grafts preserved with UW solution. Comparison between patients with low and high postoperative transaminases levels

We studied the morphology of sinusoidal cells on 21 human liver grafts prior to harvesting, at the end of the preservation period in UW solution, and after complete revascularization. The mean cold ischemic period was 11 h 34 min. Immediate follow-up was uneventful in 20 of these cases; 13 showed a mean peak of postoperative transaminases below 1,300 IU/L (group A), and 7 above 1,500 IU/L (group B). In the case of one patient (group C) steatosis was severe (50%) and there was serious postoperative dysfunction (transaminases 18,000 IU/L). Biopsies were perfusion-fixed by the transparenchymal route to ensure satisfactory ultrastructural results. In group A, some sinusoidal endothelial cells presented signs of activation at the end of the preservation period, and even more after revascularization. Kupffer cells also presented signs of activation particularly after reperfusion. Signs of endothelial cell damage with gaps and partial rupture of the plasmic membrane were also observed, particularly after revascularization in areas which contained numerous inflammatory cells adhering to the wall. The sinusoidal pole of hepatocytes was occasionally damaged, with the formation of blebs. In group B, adhesion of inflammatory cells to the sinusoidal wall was increased. Furthermore, in some areas with endothelial cell damage, neutrophils and platelets infiltrated the Disse space, and hepatocytes were increasingly damaged. In the case of patient C, the most obvious signs after reperfusion were hepatocyte drop out and death but there was no evidence of any concomitant sinusoidal cell damage. It would appear that even in cases where immediate follow-up is eventful, endothelial and Kupffer cells show signs of activation. This can be associated with signs of microcirculatory disturbances as was seen in 4 cases in group B. In the only case of severe steatosis that we studied, the essential sign was death of hepatocytes.

Functional characterization of liver-associated lymphocytes in patients with liver metastasis

BACKGROUND

The liver-associated lymphocytes (LAL) population is mainly composed of cells with natural killer (NK) activity expressing the CD3+/-CD56+ phenotype. No evident difference has been found in the phenotypic data between patients with benign or malignant liver disease. In this study, the cytotoxic pattern of this population has been characterized from patients who underwent an operation for benign or metastatic liver disease.

METHODS

LAL were isolated by sinusoidal high-pressure lavage from partial hepatectomies. Phenotype was characterized by flow cytometry, and cytotoxicity was evaluated by standard 4-hour 51Cr release assays against NK and lymphokine-activated killer (LAK)-sensitive targets.

RESULTS

In patients with benign liver disease, LAL showed spontaneous high levels of NK activity and LAK activity compared with peripheral blood lymphocytes. In patients with metastatic liver disease, no difference was observed in the levels of NK activity between LAL and peripheral blood, and the level of LAK activity was far lower than that expressed in patients with benign liver disease.

CONCLUSIONS

These results show that the cytotoxic pattern of peripheral blood lymphocytes does not mirror that of LAL. In patients with benign liver disease, LAL are in a state of activation, whereas the decreased level of LAL cytotoxicity in patients with metastatic liver disease suggests that the cytotoxic activity of these cells could be inhibited by the presence of suppressive factors.

Fatal neonatal liver failure and mitochondrial cytopathy: an observation with antenatal ascites

Mitochondrial cytopathies are multisystemic diseases of extremely variable expression caused by a deficiency in oxidative phosphorylation. Only five cases of neonatal liver failure in the context of mitochondrial cytopathy have been reported, with incomplete morphological data of the liver in three. In the case presented here, ascites had been diagnosed prenatally and liver failure was particularly severe (factor V less than 15% with fatal coma the fourth day). Histologically there were incomplete cirrhosis, microvesicular steatosis, major canalicular cholestasis with proliferative neocholangioles, and bile duct thrombi. There were also some iron pigments in the periportal area and partial glycogen depletion. By electron microscopy, mitochondria in numerous hepatocytes appeared abnormal with occasional cristae in a fluffy matrix, some containing dense inclusions. Study of respiratory chain activity showed a defect in cytochrome c oxidase (complex IV), revealed by oxygraphic measurement on fresh muscle biopsy and confirmed by spectrophotometric enzymatic assays performed on muscle and liver homogenates. The association of neonatal liver failure with hyperlactacidemia warrants investigation into a deficiency in oxidative phosphorylation.

Liver Adenomatosis With Granulomas in Two Patients on Long-Term Oral Contraceptives

We report two cases of liver cell adenomatosis associated with granulomatous hepatitis, both developed in white women (52 and 39 years of age) on long-term oral contraceptives (for 18 and 12 years, respectively). The first patient underwent surgery for five hepatic tumors 1-7 cm in diameter; the other patient had a partial segmentectomy for a 4-cm hepatic nodule of the right lobe. In both cases, dissection of the liver showed many other diffuse and smaller nodules. Histologically, all tumors were benign liver cell adenomas, with cellular atypia in the largest tumor and associated in both cases with granulomatous hepatitis, with numerous noncaseating epithelioid and giant cell granulomas located either only in the tumors (case 1) or diffusely in the tumoral and nontumoral hepatic parenchyma (case 2). During follow-up, ultrasound showed new nodular lesions in case 2, whereas in case 1 evolution was uneventful. In estroprogestative-associated liver diseases, adenomas are common, but adenomatosis and granulomas are rare. An association of these latter two conditions would seem to be exceptional. The prognosis for adenomatosis remains uncertain.

Nodular regenerative hyperplasia in the rat induced by a selenium-enriched diet: study of a model

Weaned male rats were fed a 4 ppm selenium diet. Compared after 2 mo with a control group fed a 0.4 ppm diet, the rats' body weights had not significantly decreased and liver function was normal, but portal pressure was 1.8 times higher (p less than 0.05). Liver weight was slightly increased (10.3%; p less than 0.05). All livers had an abnormal appearance. In the less severe cases the surface was only slightly irregular, but in the more severe cases, atrophic micronodular lobes and hypertrophic lobes, with mildly irregular surfaces, were present. On light microscopy, atrophic lobes displayed a peripheral nodular zone with micronodules separated by rows of atrophic hepatocytes without fibrosis, characteristic of nodular regenerative hyperplasia, and a central atrophic zone that was sometimes peliotic. Hypertrophic lobes and livers in the less severe cases had only minor and relatively localized evidence of nodular regenerative hyperplasia; occasional peliosis was seen. In all cases portal veins, hepatic veins and hepatic arteries were normal. By electron microscopy, in nonnodular zones with no obvious evidence of parenchymal atrophy, the endothelial wall showed signs of complete or incomplete capillarization with frequent enlargement of the Disse space. The selenium-enriched diet is a reproducible model of liver nodular regenerative hyperplasia. In this model, damage to the sinusoidal wall could represent the primum movens of microcirculatory disturbances.

Lipid-laden perisinusoidal cells in patients with acquired immunodeficiency syndrome

Liver biopsies were taken from ten AIDS patients. Liver architecture was normal in all patients. On 1-micron-thick sections stained with toluidine blue, all ten cases showed lipid overload of perisinusoidal cells (1 massive, 5 moderate and 4 mild) compared to 2/8 in control patients, who had mild lipid overload. Other sinusoidal abnormalities such as hypertrophy of Kupffer cells and inclusions in endothelial cells were also noticed. Some hepatocytes presented evidence of cellular damage. Perisinusoidal cell lipid overload was not associated with hypervitaminosis A. We hypothesize that the abnormal accumulation of lipids in perisinusoidal cells (non-induced by hypervitaminosis A) in patients with AIDS could be due to defective transport of vitamin A from perisinusoidal cells to hepatocytes, and/or from hepatocytes to blood. The cause of the defect is unknown. Since lipid overload occurs in many and diverse conditions (diabetes, cholestasis, primary biliary cirrhosis, etc.), it seems reasonable to propose that the defect is non-specific and limited to functional or structural damage of the liver whether induced by drugs, liver or systemic diseases.

Characterization of liver-associated natural killer cells in patients with liver tumors

The existence of a marginal lymphocyte population in rat liver sinusoids has already been demonstrated using the sinusoidal lavage method. We used the same technique to study the lymphocyte population in human liver obtained ex vivo after partial hepatectomy for benign or malignant tumors and compared it with peripheral and portal blood lymphocyte populations. Percentages of lymphocyte surface phenotypes were evaluated by flow cytometry. The lymphocyte population obtained from human liver is mainly made up of CD56+ (35%) cells. This percentage is three times greater than that found in peripheral and portal blood. Two-color flow cytometry analysis showed that within the CD56+ liver cell population, at least three distinct subsets could be found: (a) CD3+/CD56+/CD16-; (b) CD3-/CD56+/CD16-; and (c) CD3-/CD56+/CD16+. Although these subsets were also present in peripheral and portal blood, the percentage distribution was completely different because most CD56+ cells in peripheral and portal blood belonged to the CD3-/CD56+/CD16+ subset. These results show the existence of a heterogeneous natural killer cell population in human livers with tumors. The functional significance of this heterogeneity still needs to be explained.

Late recurrence of a hepatocellular carcinoma in a patient with incomplete Alagille syndrome

In this study, the case of a patient presenting a second hepatocellular carcinoma 13 years after resection of a first tumor of the same type is reported. In this case, etiological investigations remained negative, but an incomplete form of syndromatic Alagille syndrome with paucity of bile ducts in the nontumoral tissue was detected and associated with nodular regenerative hyperplasia and foci of dysplasia. Malignant transformation in Alagille syndrome seems to be extremely rare. The fact that such tumors evolve very slowly could be an argument for partial hepatectomy and, if necessary, liver transplantation.

Modifications of the sinusoidal barrier in a model of postsinusoidal hypertension in the rat. An ultrastructural study of endothelial cells

Sinusoids and sinusoidal cells were examined by light and electron microscopy, using a rat model of postsinusoidal hypertension. One month after partial ligation of the vena cava (PLVC) above the hepatic veins, subcapsular hemorrhagic areas were visible with proliferation of hepatic veins; in non hemorrhagic areas, sinusoidal congestion was found. Postsinusoidal hypertension led to a significant increase in sinusoidal volume and to major abnormalities of the endothelium such as endothelial processes and pouches with numerous diaphragmed fenestrae; some red blood cells could be seen in these pouches. Endothelial cells sent out processes in between hepatocytes. Complete and incomplete pseudo-neolumens were found near sinusoids. Numerous Kupffer cells were located either in the sinusoidal barrier or infiltrating the Disse space close to extravasated red blood cells and perisinusoidal cell processes. 18 months after PLVC, lesions were much the same except for the presence of red blood cells in the Disse space.

Perisinusoidal cell hypertrophy in a patient with acquired immunodeficiency syndrome

A 33-year-old heterosexual white man underwent a liver biopsy for determination of mild elevation of aminotransferase levels (aspartate aminotransferase, two times; alanine aminotransferase, three times). The patient had acquired immunodeficiency syndrome (stage IVC2) with tuberculosis of the lymph nodes. Antibody to hepatitis B surface antigen and antibody to hepatitis B core antigen were positive. Syphillis tests were positive. Liver architecture was normal; sinusoids were dilated with perisinusoidal, centrilobular, and portal fibrosis. On a 1-micron-thick section and under electron microscopy, perisinusoidal cells appeared to be massively loaded with lipids, while endothelial cells contained numerous dense bodies. Some hepatocytes presented evidence of cell damage. Sinusoids were infiltrated by an increased number of lymphocytes and macrophages. This patient who had recently been treated for tuberculosis was not taking extra vitamin A. He had no disease so far reported as being associated with perisinusoidal cell hypertrophy. This case and others are evidence that acquired immunodeficiency syndrome represents another cause of perisinusoidal cell hypertrophy in which there is no documented hypervitaminosis A.

Fine structure of hepatic sinusoids and sinusoidal cells in disease

Liver sinusoids are special capillaries that are limited by fenestrated endothelial cells, without a genuine basement membrane, surrounded by perisinusoidal cells storing vitamin A, and harbouring Kupffer cells and pit cells, resident macrophages, and large granular lymphocytes, respectively. Each nonparenchymal cell and parenchymal cell of the liver interacts with all others and with the extracellular matrix. Therefore, the functional ability of each cell is constantly being modified by the metabolic activity of the others. Human liver biopsies (132), needle or surgical, perfusion-fixed with glutaraldehyde and processed for transmission electron microscopy (TEM), and occasionally for scanning electron microscopy (SEM), were examined. The study included liver diseases (such as alcoholic liver diseases, benign and malignant liver tumors, cholestasis of various origins, fulminant hepatitis, acute rejection after orthotopic liver transplantation, Budd-Chiari syndrome), as well as general or extrahepatic diseases (such as diabetes, hemochromatosis, hypervitaminosis A, various hematological disorders), and normal controls. Ultrastructural abnormalities are described and illustrated under two different headings: 1) elementary lesions of sinusoidal cells (endothelial, Kupffer, perisinusoidal and pit cells), nonsinusoidal cells (in the space of Disse and/or in the lumen), the extracellular matrix; and 2) the major pathological entities including perisinusoidal fibrosis, capillarization of sinusoids, sinusoidal dilatation, and peliosis. In the discussion, an overview of the major abnormalities reported in the literature is presented, and some specific questions regarding 1) perisinusoidal fibrosis in liver with normal histology, 2) the overload of perisinusoidal cells with lipids in non-hypervitaminosis A intoxication and 3) the etiological relationship of sinusoidal dilatation, peliosis, perisinusoidal fibrosis, or sinusoidal tumors with drugs and toxic compounds are discussed. In the event that lesions are not specific to any diagnosis, the knowledge of the ultrastructure of sinusoids is extremely useful from the perspective of the liver as an ecosystem.

Nerves and perisinusoidal cells in human liver

Unmyelinated nerve fibres are visible in the human hepatic lobule. They extend through the Disse space, surrounded by Schwann cell processes, often close to perisinusoidal cell processes. A few bare nerve endings or varicosities are found contiguous to either hepatocytes or perisinusoidal cells. These nerve endings or varicosities contain large and small granular vesicles and small clear vesicles. This heterogeneity probably corresponds to the presence of various neurotransmitters (noradrenaline, acetylcholine, various neuropeptides...). The effect of nerves on perisinusoidal cells has not yet been elucidated. However, the location, shape, morphology and origin of perisinusoidal cells would suggest that they play a role in the hemodynamic regulation of sinusoidal blood flow. It has recently been shown how important non-parenchymal-parenchymal communication is in the action of nerves on glucose release by hepatocytes; the cell to cell communications may also apply to nerves and sinusoidal cells for the hemodynamic regulation of sinusoidal blood flow.

Chance discovery of hepatic fibrosis in patient with asymptomatic hypervitaminosis A

A liver biopsy specimen was obtained from a 50-year-old patient whose clinical and functional liver tests showed no abnormalities but who had for some time a high vitamin A intake (109 X 10(6) IU over four years). Liver architecture was normal. Sinusoids were slightly dilated in zone 2. Perisinusoidal cells were numerous and enlarged. On Sirius red staining, there was mild fibrosis of the central veins, portal tracts, and terminal portal venules and perisinusoidal fibrosis in zone 1 of the acinus. Liver vitamin A level was increased. By electron microscopy, perisinusoidal cells filled with numerous lipid droplets had slightly dilated rough endoplasmic reticulum, numerous minute filament condensations below the plasma membrane, and stellate-shaped processes giving them the appearance of fibroblast-myofibroblast-like cells. Numerous collagen bundles, fibrils, amorphous material, and fragments of basement membrane-like material were identified in Disse's space. Immunocytochemistry showed increased amounts of collagen types I, III, IV, laminin, and fibronectin. This observation suggests that vitamin A per se, and not the cellular damage often seen in hypervitaminosis A, is responsible for fibrosis.

The increase in the number of liver sinusoidal pit cells in four patients with primary or metastatic cancer of the liver

Four patients with liver carcinoma (case 1: hepatocellular carcinoma; cases 2 and 3: metastases; case 4: adenocarcinoma possibly of hepatic origin) underwent a wedge liver biopsy taken at some distance from the tumor. Liver histology was normal in cases 2 and 3. Sinusoids were dilated in case 4. Fibrosis formed bridges between portal tracts in case 1. In all 4 cases, sinusoids contained lymphocytic cells. By electron microscopy (perfusion-fixation with glutaraldehyde) numerous lymphocytes could be identified as pit cells with characteristic dense granules and occasional rod-cored vesicles. The majority of the pit cells were luminal cells in contact with endothelial or Kupffer cells; some were in the Disse space. It is now accepted that pit cells are resident large granular lymphocytes with natural killer activity. The increase in the number of pit cells in liver carcinoma compared to the number observed in the control group (uncomplicated gallbladder lithiasis) could be hypothetically interpreted as a defense mechanism against tumor extension.

Hepatic fibrosis in patients with idiopathic thrombocytopenic purpura

Three patients showing all the symptoms of idiopathic thrombocytopenic purpura underwent a splenectomy. A wedge liver biopsy revealed hepatic fibrosis around central veins and an increased perisinusoidal network on Sirius red staining. Fibrosis was moderate or mild. Liver histology was otherwise normal, as were liver function tests. Hepatic fibrosis could not be attributed to any known causes. Electron microscopy showed numerous Kupffer cells with intense phagocytic activity, and perisinusoidal cells with some of the characteristics of fibro/myofibroblasts. The mechanisms of fibrosis remain unknown but could be attributed, by analogy to agnogenic myeloid metaplasia, to the massive destruction of platelets liberating PDGF and to increased activation of macrophages.

Perisinusoidal fibrosis and basement membrane-like material in the livers of diabetic patients

A direct correlation exists between collagenization of Disse's space and the presence of diabetic microangiopathy in type I diabetes. To confirm and extend this finding, we studied four liver biopsy samples from two patients with type I diabetes (one with retinopathy) and two patients with type II diabetes (no retinopathy). All had normal or subnormal results on liver function tests and normal liver architecture. Levels of collagen types I, III, and IV, laminin, and fibronectin, as determined by immunocytochemical techniques, appeared increased in all patients. Liver biopsy samples were perfusion fixed for electron microscopy of sinusoids and sinusoidal cells. Numerous and thick collagen bundles could be seen in Disse's space, as could the increase of basement membrane-like material underlying the endothelial cells, perisinusoidal cells, and sinusoidal membrane of hepatocytes. Perisinusoidal cells were active and had abundant rough endoplasmic reticula and thick processes. This preliminary study indicates that collagenization of Disse's space is not specific to a certain type of diabetes. The increase of basement membrane-like material raises the question of whether liver sinusoids are truly different from other capillaries as far as diabetic microangiopathy is concerned.

Liver sinusoids and sinusoidal cells in patients with agnogenic myeloid metaplasia

The origin of sinusoidal portal hypertension often remains unknown in patients with agnogenic myeloid metaplasia. Four consecutive patients with agnogenic myeloid metaplasia had liver biopsies examined under light and electron microscopy. Two of the four had obvious clinical portal hypertension; of these two, only one had portal vein thrombosis. All four cases showed sinusoidal infiltration by myeloid cells (from very mild to obvious) and an increased perisinusoidal collagen network. Under electron microscopy we observed (a) the collagenization of the Disse space, (b) myeloid cells in the lumen and the Disse space, (c) the transformation of perisinusoidal cells into transitional cells, and (d) fragments of basement membrane-like material. It is possible that all these abnormalities, and not only sinusoidal infiltration, contribute to increase vascular resistance, even when there is no clinical evidence of portal hypertension unrelated to vascular thrombosis.