Endotoxin-induced lung injury in α-galactosylceramide-sensitized mice is caused by failure of interleukin-4 production in lung natural killer T cells.

Administration of bacterial lipopolysaccharide (LPS) known as endotoxin into α-galactosylceramide (α-GalCer)-sensitized mice causes severe lung lesions but few hepatic lesions in lethal shock, and interferon (IFN)-γ is suggested to play a pivotal role in preparation of the lung lesions. In order to clarify the mechanism of how α-GalCer sensitization causes lung lesions exclusively in mice, we examined the differential responsiveness of lungs and livers to α-GalCer sensitization. Although lung and liver natural killer T (NK T) cells both produced IFN-γ in response to α-GalCer, IFN-γ signalling was triggered only in the lungs of α-GalCer-sensitized mice. Lung NK T cells did not produce interleukin (IL)-4 in response to α-GalCer and it did not induce the expression of suppressor of cytokine signalling 1 (SOCS1) in the lungs. Conversely, IL-4 produced by liver NK T cells led to the expression of SOCS1 in the livers of the mice. Neutralization of IL-4 reduced SOCS1 expression in the livers and exacerbated LPS-induced hepatic lesions. IL-10 was produced by liver NK T cells but not lung NK T cells. However, IL-10 was produced constitutively by alveolar epithelial cells in normal lung. Lung NK T cells and liver NK T cells might express CD8 and CD4, respectively. Based on the fact that IL-4 inhibited IFN-γ signalling in the livers of α-GalCer-sensitized mice via SOCS1 expression and signal transducer and activator of transcription 1 (STAT-1) activation, no inhibition of the IFN-γ signalling in the lungs caused LPS-induced lung lesions in α-GalCer-sensitized mice. The detailed mechanism of development of the lung lesions in α-GalCer-sensitized mice is discussed.

Metformin attenuates production of nitric oxide in response to lipopolysaccharide by inhibiting MyD88-independent pathway.

Metformin is reported to ameliorate inflammation in diabetic patients. The effect of metformin on lipopolysaccharide-induced nitric oxide production was studied by using RAW 264.7 macrophage-like cells. The action of metformin was analyzed by dividing lipopolysaccharide signaling into the MyD88-dependent and -independent pathways. Metformin significantly reduced the expression of an inducible type of nitric oxide synthase and inhibited lipopolysaccharide-induced nitric oxide production. On the other hand, metformin did not inhibit lipopolysaccharide-induced tumor necrosis factor-alpha production. The expression levels of interferon-beta protein and mRNA, which is a key molecule in MyD88-independent pathway, were significantly inhibited by metformin. Compound C, a specific AMP-activated protein kinase inhibitor, did not affect the inhibitory action of metformin. Metformin was suggested to inhibit lipopolysaccharide-induced nitric oxide production via inhibition of interferon-beta production in MyD88-independent pathway. Metformin might exhibit an anti- inflammatory action on diabetic complications as well as the antidiabetic action.

Novel mechanism of U18666A-induced tumour necrosis factor-alpha production in RAW 264.7 macrophage cells.

U18666A is a cholesterol transport-inhibiting agent that is used widely to mimic Niemann-Pick type C disease. The effect of U18666A on tumour necrosis factor (TNF)-alpha production in mouse macrophage cell line, RAW 264.7 cells and peritoneal macrophages was examined. U18666A induced TNF-alpha mRNA expression 48 h after the treatment, and TNF-alpha production 48 and 72 h after stimulation in RAW 264.7 cells. U18666A accumulated intracellular free cholesterol in the culture of normal medium but not cholesterol-free medium. U18666A also induced reactive oxygen species (ROS) generation in normal medium but much less in cholesterol-free medium. Anti-oxidant N-acetyl-L-cysteine (NAC) abolished U18666A-induced TNF-alpha production. U18666A led to the phosphorylation of p38 mitogen-activated protein kinase 24 and 48 h after the stimulation and the p38 activation was inhibited in presence of cholesterol-free medium or NAC. A p38 inhibitor reduced U18666A-induced TNF-alpha production. Taken together, U18666A was suggested to induce TNF-alpha production in RAW 264.7 cells via free cholesterol accumulation-mediated ROS generation.

The mechanism of development of acute lung injury in lethal endotoxic shock using alpha-galactosylceramide sensitization.

The mechanism underlying acute lung injury in lethal endotoxic shock induced by administration of lipopolysaccharide (LPS) into alpha-galactosylceramide (alpha-GalCer)-sensitized mice was studied. Sensitization with alpha-GalCer resulted in the increase of natural killer T (NK T) cells and the production of interferon (IFN)-gamma in the lung. The IFN-gamma that was produced induced expression of adhesion molecules, especially vascular cell adhesion molecule-1 (VCAM-1), on vascular endothelial cells in the lung. Anti-IFN-gamma antibody inhibited significantly the VCAM-1 expression in alpha-GalCer-sensitized mice. Very late activating antigen-4-positive cells, as the counterpart of VCAM-1, accumulated in the lung. Anti-VCAM-1 antibody prevented LPS-mediated lethal shock in alpha-GalCer-sensitized mice. The administration of LPS into alpha-GalCer-sensitized mice caused local production of excessive proinflammatory mediators, such as tumour necrosis factor (TNF)-alpha, interleukin (IL)-1beta, IL-6 and nitric oxide. LPS caused microvascular leakage of proteins and cells into bronchoalveolar lavage fluid. Taken together, sensitization with alpha-GalCer was suggested to induce the expression of VCAM-1 via IFN-gamma produced by NK T cells and recruit a number of inflammatory cells into the lung. Further, LPS was suggested to lead to the production of excessive proinflammatory mediators, the elevation of pulmonary permeability and cell death. The putative mechanism of acute lung injury in LPS-mediated lethal shock using alpha-GalCer sensitization is discussed.

Lipopolysaccharide augments the in vivo lethal action of doxorubicin against mice via hepatic damage.

The effect of lipopolysaccharide (LPS) on the in vivo lethal action of doxorubicin (DOX) against mice was studied. DOX killed LPS-pretreated mice much earlier than untreated mice, and exhibited a stronger toxic action against LPS-pretreated mice. DOX-induced lethality in LPS-pretreated mice was due to severe hepatic damage, but there were no significant lesions in the heart, kidney and lung. Hepatic lesions were accompanied by caspase 3-positive cells and fragmented DNA-positive cells, suggesting the involvement of apoptosis. DOX induced the production of a high level of interferon (IFN)-gamma and tumour necrosis factor (TNF)-alpha in LPS-pretreated mice, but not in non-treated mice. The DOX-induced lethality was prevented significantly by anti-IFN-gamma antibody, but not anti-TNF-alpha antibody. Administration of recombinant IFN-gamma in place of LPS augmented definitively the DOX-induced lethality. LPS augmented the DOX-induced lethality in TNF-alpha-deficient mice. Taken together, LPS was suggested to enhance DOX-induced IFN-gamma production and augment the in vivo lethal action via hepatic damage.

Lipopolysaccharide and interferon-gamma enhance Fas-mediated cell death in mouse vascular endothelial cells via augmentation of Fas expression.

The effect of interferon (IFN)-gamma and/or lipopolysaccharide (LPS) on Fas-mediated cell death with anti-Fas agonistic antibody in vascular endothelial cells was examined using a mouse END-D cell line. Anti-Fas agonistic antibody exhibited cytotoxic actions on END-D cells. Fas-mediated cell death was enhanced by LPS or IFN-gamma. The combination of IFN-gamma and LPS significantly enhanced cell death compared to IFN-gamma or LPS alone. IFN-gamma and LPS augmented cell surface expression of Fas, but not tumour necrosis factor (TNF) receptor 1. Inhibitors of p38 mitogen-activated protein kinase (MAPK) prevented augmentation of Fas expression in IFN-gamma and LPS-treated END-D cells. IFN-gamma and LPS-treated END-D cells did not become susceptible to TNF-alpha or nitric oxide-mediated cytotoxicity. IFN-gamma and LPS thus appear to augment selectively Fas expression via activation of p38 MAPK and enhance Fas-mediated cell death in END-D cells. Furthermore, administration of IFN-gamma and LPS into mice induced in vivo expression of Fas on vascular endothelial cells and Fas ligand (FasL) on peripheral blood leucocytes. The relationship between enhancement of Fas-mediated cell death by IFN-gamma and LPS and the development of vascular endothelial injury is discussed.

Effect of weight loss using formula diet on renal function in obese patients with diabetic nephropathy.

OBJECTIVE: To evaluate the effect and safety of treatment with low-calorie formula diet on renal function and proteinuria in obese patients with diabetic nephropathy. DESIGN: Prospective study on safety and efficacy of a 4-week low-calorie (11-19 kcal/kg/day) normal-protein (0.9-1.2 g/kg/day) diet partly supplemented with formula diet. SUBJECTS: In all, 22 obese patients with diabetic nephropathy (BMI: 30.4+/-5.3 kg/m(2), HbA1c: 7.1+/-1.4%, serum creatinine: 172.4+/-57.5 micromol/l, urinary protein: 3.3+/-2.6 g/day). RESULTS: The mean body weight decreased by 6.2+/-3.0 kg. The mean systolic blood pressure, creatinine, blood urea nitrogen, urinary protein, and 8-hydroxydeoxyguanosine decreased significantly by 7.5+/-12.7 mmHg, 41.6+/-23.9 micromol/l, 1.50+/-1.61 mmol/l, 1.8+/-1.7 g/day, and 3.1+/-3.6 ng/mg creatinine, respectively. No patient had increased serum creatinine and urinary protein. Mean creatinine clearance (40.6+/-17.9 to 46.1+/-14.6 ml/s/1.73 m(2)) and serum albumin showed no significant changes. Delta serum creatinine and Delta urinary protein correlated with Delta body weight (r=0.62 and 0.49, respectively) and Delta visceral fat area (r=0.58 and 0.58, respectively), but did not correlate with Delta systolic blood pressure, Delta fasting blood glucose and Delta subcutaneous fat area. CONCLUSION: These results suggested that weight reduction using formula diet might improve renal function and proteinuria safely for a short term in obese patients with diabetic nephropathy.

Thoracoscopic enucleation of esophageal stromal tumor.

Gastrointestinal stromal tumor is a rare entity, especially in the esophagus. We report a patient with a stromal tumor of the esophagus who underwent a thoracoscopic enucleation of the tumor. The patient was a 61-year-old man complaining of slight dysphagia. A submucosal tumor of the middle thoracic esophagus was found endoscopically. The tumor was approximately 4.0 cm in diameter measured by endoscopic ultrasonography. On 17 May 2001, thoracoscopic enucleation of the esophageal tumor was performed using a Kodama Di-suction. The Kodama Di-suction was useful for the thoracoscopic enucleation of the submucosal tumor of the esophagus, acting as both a dissector and a sucker. The patient's course was uneventful after surgery. Histopathologically the esophageal tumor revealed a high cellularity, consisting of spindle cells, and the tumor cells were immunohistochemically positive for CD34 and c-kit protein, but not for a-smooth muscle actin or S-100 protein. From these findings, the esophageal submucosal tumor was diagnosed as gastrointestinal stromal tumor, distinguished from leiomyoma.

Food residue at endoscopy in patients who have previously undergone distal gastrectomy: risk factors and patient preparation.

BACKGROUND AND STUDY AIMS: Food residue is often seen in the gastric remnant after partial gastrectomy, making it difficult to diagnose early cancer in the residual stomach. The aims of this study were to clarify the risk factors for the accumulation of food residue, and to study methods of preparation for endoscopy in patients who had undergone distal gastrectomy. PATIENTS AND METHODS: 374 endoscopic examinations of patients who had undergone distal gastrectomy for gastric cancer were compared with 2168 endoscopic examinations in patients without a history of gastrectomy. Relationships between the presence of food residue and a number of clinical factors, including patient preparation, were evaluated by univariate and multivariate analyses. RESULTS: Food residue in the gastric remnant was observed in 70 examinations (18.7 %), a significantly higher proportion than that found in control patients (0.3 %). From multivariate analysis, underlying diseases (endocrine, metabolic, or connective tissue disease), Billroth type I reconstruction, and postoperative gastric retention were found to be independent risk factors for the accumulation of food residue. Diet preparation (a liquid diet plus aclatonium napadisilate) significantly decreased the incidence of food residue. CONCLUSIONS: Our diet preparation method can be recommended as a preparation for upper gastrointestinal endoscopy in patients who have undergone distal gastrectomy, especially in patients with additional risk factors.

Expression of basic fibroblast growth factor and its receptor-1 in cardiac myxoma.

BACKGROUND: To clarify the significance of basic fibroblast growth factor (bFGF) in angiogenesis or proliferative activity in cardiac myxoma, the expression of bFGF and its receptor (FGFR-1) were immunohistochemically examined. METHODS: Formalin-embedded tissues of cardiac myxomas were obtained by surgical resection from 15 patients and analyzed by immunostaining of bFGF and FGFR-1. The microvessel density was measured in the 15 myxomas using platelet derived endothelial cell adhesion molecule-1. For evaluation of proliferative activity of the cardiac myxomas, proliferating cell nuclear antigen (PCNA) immunostaining was performed, and the PCNA labeling index was measured in each section. RESULTS: bFGF and FGFR-1 were observed in 73.3% and 67.7% of the myxomas, respectively. There was a close correlation between the expression of bFGF and FGFR-1. This co-expression was frequently observed in the myxoma cells around the microvessels appearing as a ring structure. Regarding possible relationships between the expression of bFGF or FGFR-1 and the clinicopathologic features, there were no parameters excluding the macroscopic type of myxoma. The microvessel density in the myxomas with bFGF or FGFR-1 expression was higher than that in myxomas without it. The PCNA labeling index in myxomas with bFGF expression was higher than that in myxomas without it, and the PCNA labeling index tended to be higher in myxomas with FGFR-1 expression than that in myxomas without it. CONCLUSIONS: bFGF and/or FGFR-1 was expressed in some of cardiac myxoma, and may be an important role for tumor angiogenesis and proliferative activity.

Regulation of circulating immune complexes by complement receptor type 1 on erythrocytes in chronic viral liver diseases.

BACKGROUND AND AIM: Complement receptor type 1 (CR1) is a transmembrane protein, and human erythrocyte CR1 (E-CR1) is involved in the transport of circulating immune complexes (IC) from the circulation to the reticuloendothelial system, including the liver and spleen. In chronic viral hepatitis, increased levels of IC containing viral particles and an association with various extrahepatic manifestations have been reported. However, regulatory mechanisms for IC levels are not fully understood. PATIENTS/SUBJECTS AND METHODS: We analysed IC, E-CR1, and quantitative polymorphism of the CR1 gene in 149 patients with chronic viral liver diseases and in 64 normal blood donors using an enzyme linked immunosorbent assay, radioimmunoassay, and polymerase chain reaction-restriction fragment length polymorphism, respectively. We also analysed the effect of CR1 gene polymorphism on IC binding to E-CR1 using molecular methods. RESULTS: E-CR1 levels in patients with chronic hepatitis and chronic viral liver diseases as a whole correlated inversely with increased levels of IC. Moreover, significantly high levels of IC were observed in patients with chronic hepatitis C (CH-C) who were homozygous for the E-CR1 low density allele. We also found low levels of E-CR1 in liver cirrhosis and CH-C but not in CH-B. Low levels of E-CR1 in CH-C were observed, even after considering the polymorphism of the CR1 gene. Finally, we demonstrated CR1 gene polymorphism dependent binding of hepatitis virus containing IC. CONCLUSIONS: Our results emphasise the important role of E-CR1 in clearance of IC from the circulation and the acquired, rather than inherited, decrease in E-CR1 in chronic viral liver diseases, especially of type C.

Can sedation reduce the cardiac stress during gastrointestinal endoscopy? A study with non-invasive automated cardiac flow measurement by color Doppler echocardiography.

BACKGROUND: Upper gastrointestinal endoscopy (UGIE) may cause some cardiac stress. The effect of sedation on hemodynamics during UGIE has not been fully studied, and therefore the aim of this study was to clarify whether or not sedation can reduce cardiac stress dufing UGIE. METHODS: Eight normal male volunteers undergoing UGIE with sedation (0.1 mg/kg of midazolam) and without it (two endoscopies per volunteer in random order) were monitored throughout the procedure by means of electrocardiogram, blood pressure and peripheral oxygen saturation (SpO2). Cardiac output was measured at six points before, during and after endoscopy from automated cardiac flow measurement by color Doppler echocardiography. Serum norepinephrine, epinephrine, dopamine and ACTH concentrations were measured before and after the examination. RESULTS: No significant differences in heart rate, systolic blood pressure, rate-pressure product, cardiac output and left ventricular work index were observed between the sedated and non-sedated groups. SpO2 hardly changed during endoscopy in the non-sedated group, but decreased slightly in the sedated group (P = 0.075). Although all serum catecholamine concentration changes were within normal limits in both groups, after endoscopy only epinephrine concentration was significantly lower in the sedated group than in the non-sedated group (P = 0.0027). CONCLUSIONS: Conscious sedation with midazolam does not reduce the cardiac stress during UGIE.

Differences and relationships of thymidine phosphorylase expression in tumor-associated macrophages and cancer cells in squamous cell carcinoma of the esophagus.

Thymidine phosphorylase (TP), which has been shown to be identical to platelet-derived endothelial cell growth factor, is expressed in tumor-associated macrophages (TAMs) as well as cancer cells. The aim of this study was to clarify the differences or relationships of TP expression in TAMs and cancer cells in esophageal squamous cell carcinoma (SCC). Tissues samples were taken from 56 patients with esophageal SCC after curative surgery. The expression of TP in TAMs or SCC cells was examined using a monoclonal antibody to TP (clone 654-1). Microvessels in SCC that stained positively for Factor VIII-related antigen were counted (microvessel density, MVD). Macrophages in SCC that stained positively for CD68 antigen were counted (monocytic count). Ki-67 antigen was immunostained with MIB-1, terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate biotin nick end labeling was performed, and Ki-67 labeling index (LI) and apoptotic index were calculated. The expression of TP in stromal cells and cancer cells was observed in 43 (76.8%) and 33 patients (58.9%), respectively. There were significant correlations between TP expression in stromal cells (TAMs) as well as in cancer cells and venous invasion, distant metastasis, or MVD. There was a correlation between TP expression in cancer cells and lymph node metastasis, and there were correlations between TP expression in TAMs and monocytic count or Ki-67 LI; however, there was no correlation between TP expression in TAMs and lymph node metastasis. On the other hand, in SCCs with TP expression in both TAMs and cancer cells, higher frequencies of venous invasion and distant metastasis, higher MVD and lower apoptotic index were observed than in other SCCs. The 5-year survival rate in patients with TP expression in both TAMs and cancer cells was poorer than that in patients with TP expression in neither TAMs and cancer cell. In conclusion, these results suggest that co-expression of TP in TAMs and cancer cells is strongly associated with angiogenic promotion and distant metastasis. However, other effects of TP, such as promotion of tumor growth and lymph node metastasis, may be different depending on whether these are expressed in TAMs or cancer cells in esophageal SCCs. Patients with coexpression of TP in TAMs and cancer cells may be associated with a poor prognosis.

A case of hyperplastic polyposis of the colon with adenocarcinomas in hyperplastic polyps after long-term follow-up.

A 66-year-old woman had been receiving follow-up since 1990 for hyperplastic polyposis, which remained unchanged endoscopically and radiologically. In 1999, a small (28 x 22 mm) superficial adenocarcinoma was detected in the ascending colon. Histologically, this was a hyperplastic polyp containing a well-differentiated adenocarcinoma invading into the submucosa. A review of the English and Japanese literature identified 32 cases of "hyperplastic polyposis." In about half of the 32 cases, an adenocarcinoma was also found amongst the hyperplastic polyps. Half of the adenocarcinomas were located in the right colon. Although hyperplastic polyposis is uncommon, it warrants regular surveillance, as it appears to be associated with an increased risk of colorectal cancer.

The inhibitory action of quercetin on lipopolysaccharide-induced nitric oxide production in RAW 264.7 macrophage cells.

The effect of quercetin on lipopolysaccharide (LPS)-induced nitric oxide (NO) production was studied. Quercetin pretreatment significantly inhibited NO production in an LPS-stimulated RAW 264.7 murine macrophage cell line. Post-treatment with quercetin partially inhibited NO production. The inhibitory action of quercetin was due to neither the cytotoxic action nor altered LPS binding. The expression of inducible-type NO synthase (iNOS) was markedly down-regulated by quercetin. Quercetin suppressed the release of free nuclear factor (NF)-kappaB by preventing degradation of IkappaB-alpha and IkappaB-beta. Moreover, quercetin blocked the phosphorylation of extracellular signal regulated kinase 1/2 (Erk 1/2), p38, and c-Jun NH2-terminal kinase/stress-activated protein kinase (JNK/SAPK) and, further, the activity of tyrosine kinases in LPS-stimulated RAW cells. Quercetin also inhibited interferon (IFN)-gamma-induced NO production. Taken together, these results indicate that the inhibitory action of quercetin on NO production in LPS- and/or IFN-gamma-stimulated macrophages might be due to abrogation of iNOS protein induction by impairment of a series of intracellular signal pathways.

The expression of membrane-bound CD14 renders mouse B-1 cells susceptible to LPS.

The surface expression of CD14 on mouse B-1 cells and its role on their response to lipopolysaccharide (LPS) were studied by using the murine TH2.52 B-1 cell line and peritoneal B-1 cells. TH2.52 cells with the B-1 phenotype were found to express membrane-bound CD14. Furthermore, CD14 was expressed on physiological peritoneal CD5+ B-1 cells. The stimulation of CD14-expressing TH2.52 cells with a low concentration of LPS resulted in the activation of nuclear factor (NF)-B and a mitogen-activated protein kinase (MAPK). The LPS-induced NF-B and MAPK activation was markedly inhibited by anti-CD14 antibody. These results suggest that B-1 cells may respond to LPS via membrane-bound CD14.

Mouse B1 cell line responds to lipopolysaccharide via membrane-bound CD14.

The role of membrane-bound CD14 in the response of mouse B1 cell lines to lipopolysaccharide (LPS) was studied. The surface profile of mouse TH2.52 B cells was positive for CD5, IgM, B220, CD11b and F4/80, suggesting that TH2.52 cells carried the typical phenotype of B1 cells. Furthermore, TH2.52 B1 cells were found to express membrane-bound CD14, which plays a critical role in LPS recognition. TH2.52 B1 cells responded to a very low concentration of LPS and exhibited: (i) augmentation of DNA synthesis; (ii) activation of nuclear factor (NF)-kappaB; and (iii) phosphorylation of extracellular signal regulated kinase 1/2 (Erk1/2). They were markedly inhibited by anti-CD14 antibody. Therefore, the expression of membrane-bound CD14 was suggested to provide high sensitivity to LPS for TH2.52 B1 cells.

Flavonoids protect mice from two types of lethal shock induced by endotoxin.

The protective effect of flavonoids on two types of lethal endotoxic shock was studied. A lethal endotoxic shock was induced by administration of lipopolysaccharide (LPS) into D-galactosamine (D-GalN)-sensitized mice and another one was done by administration of a high dose of LPS into normal mice. Pretreatment with a series of flavonoids protected mice from two types of endotoxin lethality. Flavonoid pretreatment reduced the serum tumor necrosis factor-alpha (TNF-alpha) level in mice injected with D-GalN and LPS, but not in mice injected with a high dose of LPS. TNF-alpha-induced lethal shock in D-GalN-sensitized mice was also protected by pretreatment with flavonoids, suggesting that flavonoids augmented the resistance to TNF-alpha lethality. On the other hand, flavonoids reduced the plasma level of lipid peroxides in mice injected with a high dose of LPS, but not in D-GalN-sensitized mice. Taken together, these results indicated that flavonoids might protect mice from two types of endotoxin lethality. The protective mechanism of flavonoids in each endotoxin lethality is discussed.

Hepatic resection for large hepatocellular carcinoma.

BACKGROUND: Long-term survival and prognostic factors after hepatic resection for large hepatocellular carcinoma (HCC) remain to be proved. METHODS: The surgical outcome in 133 consecutive patients with HCC in diameter of > or = 5 cm (large HCC; L group) undergoing hepatic resection was retrospectively clarified and compared with that of 253 patients with HCC in diameter of < 5 cm (small HCC; S group). Postresection prognostic factors were evaluated by univariate and multivariate analysis using Cox's proportional hazards model. RESULTS: The disease-free 3- and 5-year survival rates between L group and S group were 26% versus 42% and 20% versus 25%, respectively (P = 0.0032). The overall 3- and 5-year survival rates between L group and S group were 38% versus 67% and 28% versus 47%, respectively (P < 0.0001). Multivariate analysis revealed that large amount of intraoperative blood transfusion was an independently significant factor of poor disease-free and overall survivals. CONCLUSIONS: Long-term survival in patients with large HCC remains unsatisfactory compared with that in patients with non-large HCC. Restriction of intraoperative blood transfusion may play an important role in the improvement of survival and recurrence in such patients.