Bacterial infection in patients with advanced cirrhosis: a multicentre prospective study.

AIMS: To evaluate the prevalence, incidence and clinical relevance of bacterial infection in predominantly non-alcoholic cirrhotic patients hospitalised for decompensation. PATIENTS/METHODS: A total of 405 consecutive admissions in 361 patients (249 males and 112 females; 66 Child-Pugh class B and 295 class C) were analysed. Blood, urine, ascitic and pleural fluid cultures were performed within the first 24 hours, during hospitalisation whenever infection was suspected, and again before discharge. RESULTS: Over a one year period, 150 (34%) bacterial infections (89 community- and 61 hospital-acquired) involving urinary tract (41%), ascites (23%), blood (21%) and respiratory tract (17%) were diagnosed. The prevalence of bacterial peritonitis was 12%. Infections were asymptomatic in 69 cases (46%) and 130 (87%) involved a single site. Enteric flora accounted for 62% of infections, Escherichia Coli being the most frequent pathogen (25%). Community-acquired infections were associated with more advanced liver disease (Child-Pugh mean score 10.2+/-2.1 versus 9.5+/-1.9, p<0.05), renal failure (p<0.05), and high white blood cell count (p<0.01). Hospital-acquired infections occurred more frequently in patients admitted for gastrointestinal bleeding (p<0.05). The in-hospital mortality was significantly higher in infected than in non-infected patients (15% versus 7%, p<0.05), and infection emerged as an independent variable affecting survival. Moreover bacterial infection accounted for a significantly prolonged hospital stay. CONCLUSIONS: Bacterial infection, regardless of the aetiology, is a severe complication of decompensated cirrhosis, and, although frequently asymptomatic, accounts for both longer hospital stay and increased mortality.

Role of IgM antibody to hepatitis B core antigen in the diagnosis of hepatitis B exacerbations.

IgM anti-HBc levels were measured by the IMx Core-M Abbott assay in 939 serum samples in order to define a specific and sensitive cut-off value for diagnosis of chronic hepatitis B. The sera used were obtained from 52 chronic HBV patients and 10 HBV carriers with HCV or HDV co-infections and 155 asymptomatic subjects without evidence of liver disease. A Youden index value of 95.4% with 98% sensitivity and 97.4% specificity was obtained for an IMx Index value of 0.204 as cut-off. A one-year follow-up study with monthly tests has shown that quantitative analysis of IgM anti-HBc can serve as a noninvasive tool for monitoring HBV infection, and provides an accurate diagnosis of hepatitis B exacerbations. Significant elevations of IgM anti-HBc levels were associated with hepatitis B exacerbations in 96.2% of the cases but with none of the ALT flare-ups observed in HCV or HDV infected individuals. These results suggest that quantitative analysis of IgM anti-HBc provides the highest degree of confidence in definition of spontaneous and therapy-induced exacerbations or remissions of hepatitis B.

Problems in diagnosing viral hepatitis.

The most reliable method of making a specific aetiological diagnosis of chronic viral hepatitis would be to identify virus specific cytotoxic T lymphocytes responsible for the killing of virus infected hepatocytes in each patient's liver. Unfortunately, this can not be proposed for routine diagnosis and surrogate tests are required. The detection of virus markers, and even of the virus itself, does not imply that liver damage is caused by virus infection. Indirect markers of the host's antiviral immunoresponse have to be used to confirm more specifically the diagnosis of viral hepatitis. IgM antibodies against viral antigens implicated in the elimination of the virus seem to be suitable alternative candidates. Significant changes in the serum values of viraemia and aminotransferases occur within a few days, while a significant variation in liver histology takes much longer. Only the kinetics of the highly variable parameters can be used for an appropriate study of the relationship between viraemia, antiviral immunoresponse, and liver cell necrosis. Quantitative and dynamic analyses of hepatitis virus markers seem the most suitable and reliable methods of monitoring the patients eligible for antiviral treatment and identifying the most appropriate time to start this.

The spontaneous bacterial peritonitis in cirrhotic patients. To a new gold standard.

Of 282 consecutive ascites prospectively collected in 54 months, Spontaneous Bacterial Peritonitis (SBP) was diagnosed in 8.5% of the cases, "probable" SBP in 31.1%, Bacteriascites (BA) in 3.5% and Sterile Ascites (SA) (negative ascitic fluid culture with PMN less than 250/mm3) in 74.8%. Escherichia Coli (41.6%) and Staphylococcus Epidermidis (60%) were the most frequent pathogens isolated in patients with SBP and BA, respectively. With regards to in-hospital mortality, 18% of patients with BA and 50% with SBP died; the mortality seemed to be related to the degree of hepatic and renal damage, to a higher peripheral and ascitic WBC concentration and to a lower pH of ascitic fluid (FA). When the comparative analysis was applied to the four groups of ascites, a different distribution of clinical signs and biohumoral parameters appeared. As a matter of fact, abdominal pain, fever and rebound tenderness resulted significantly more frequent in SBP and "probable" SBP. Furthermore, the mean values of peripheral and ascitic WBC concentration, of serum creatinine and of ALT were statistically higher in SBP and "probable" SBP than in SA and BA groups. The strict relationship, both symptomatologic and biochemical, between SA and BA on the one hand and between "probable" SBP and SBP on the other, prompted us to conclude that "probable" SBP and SBP represent different patterns of the same disease. Therefore, the subclassification in the four groups outlined above would not be in accordance with the clinical practice and could give rise to the physician's confusion and uncertainty.

Fibronectin, cholesterol and triglycerides ascitic fluid concentration in the prediction of malignancy.

There have been few trials comparing the efficacy of determinations of cholesterol, fibronectin and triglycerides for diagnosis of malignant ascites. In this study we measured these in 200 ascitic fluids from 93 cirrhotic patients (Group A), 47 hepatocellular-carcinoma patients (Group B), 60 extra-hepatic tumour patients (Group C), 44 of them with malignant cells (Group Cpos) and 16 without (Group Cneg). Anova one-way and the Bonferroni test for multiple comparisons showed that fibronectin and cholesterol were significantly higher in the ascitic fluids of patients of group C than of groups A and B (mean +/- ESM) (Cholesterol in A: 27.2 +/- 2.8; in B 23.5 +/- 1.5; in C: 68.6 +/- 5.3 mg/dl. Fibronectin in A: 32.7 +/- 2.8; in B 31.3 +/- 2.6; in C 127.7 +/- 11.1 mg/l). Both were significantly higher in Group Cpos than in Group Cneg (Cholesterol in Cneg: 41.2 +/- 6.7; in Cpos: 78.6 +/- 6.2 mg/dl. Fibronectin in Cneg: 55.0 +/- 11.2; in Cpos 154 +/- 12.3 mg/dl). We found no differences between cirrhotic ascites and malignant ascites due to primary liver hepatocellular-carcinoma. No difference at all in triglycerides were detected. With the Receiver-Operating Characteristic (ROC) curve, cholesterol had the best Youden Index (57%) at a cut-off of 32 mg/dl (sensitivity 78.3%, specificity 79.3% at this level); the best Youden Index (64%) for fibronectin had a cut-off of 60 mg/dl (sensitivity 65%, specificity 89.3%). Triglycerides appeared to be a great deal less effective as a diagnostic marker, with their best Youden Index (23%) at a cut-off 32 mg/dl (sensitivity 66.7%, specificity 56.4%).(ABSTRACT TRUNCATED AT 250 WORDS)

gamma-Glutamyl-transpeptidase in diabetics: a case control study.

Several investigators have reported high levels of gamma-glutamyl-transpeptidase (GGT) in the diabetic population. Therefore, we undertook a study to see the prevalence of 'isolated' high GGT in a large population of diabetics without chronic liver disease (CLD), as compared to an age- and sex-matched control group of non-diabetic subjects without CLD, and the role of extrahepatic factors in 'isolated' high GGT, as possible etiopathogenetic causes. We selected 351 diabetics with normal hepatologic screening, without echographic abnormalities of the hepatic parenchyma or the biliary tract. Age, duration and therapy of diabetes, body mass index (BMI), alcohol consumption, glycosylated hemoglobin (HbA1), and the presence of hepatitis B virus (HBV) were studied to see if they are related to high GGT. The control group included 260 age- and sex-matched non-diabetic subjects. We did not find any significant difference between diabetics and the control group in the prevalence of high GGT (mean: 17.5% vs. 23%; women: 16% vs. 14.5%). Multiple regression analysis showed that alcohol consumption plays the major role in the high GGT of both men and women.

[Primary hepatocarcinoma. Epidemiologic case-control study in the province of Bergamo]. / Epatocarcinoma primitivo. Indagine epidemiologica caso-controllo nella provincia di Bergamo.

This is the first epidemiological study of Hepatocellular Carcinoma (HCC) in the province of Bergamo, an area well-known to have a high incidence of HBsAg (9.1%) and chronic alcoholic liver disease. 72 cases of HCC (60 male, 12 female) al from the province of Bergamo and encountered in 1980-84 were subjected to an epidemiological case-control study. Analysis of the results confirmed the role of certain known Risk Factors (RF) with a prevalence of male sex (83.3%), age (mean age 63), association with live cirrhosis (79.2%) and HBsAg+ (31.9%). Such findings are in line with the Italian average found in previous studies. No difference between the sexes was found in these RF except for alcoholic abuse which was significantly higher in the males (53.3%, p less than 0.05). The case-control correlation analysis revealed no difference in the prevalence of alcohol addiction and previous HBV infection (HBV-Ab+) between the HCC (with or without cirrhosis) and the various control groups (Group A: patients with no liver pathology. Group B: patients with cirrhosis of the liver). HBsAg+ was significantly higher among HCC patients without cirrhosis (46.6%, p less than 0.001), but there was no significant difference between HCC + cirrhosis and cirrhosis alone (88%). The difference between "expected" and "observed" HCC-HBsAg+ was highly significant (p less than 0.001). The overall Relative Risk (RR) of HCC for the RF-HBsAg+ was 4.6. When divided into subgroups this gave: RR = 1 for patients with cirrhosis of the liver. RR = 11.5 for patients with no liver pathology. These data confirm the importance of current HBV infection (HBsAg+), in the province, though the presence or absence of cirrhosis probably influences its significance. The approximate incidence of HCC is 9.7%/100,000/year. Considering the limitations of data on a small monocentric study in a limited area (USSL no. 30) the figure is probably underestimated and the real incidence probably higher.

Role of hepatitis B virus infection in chronic liver disease of diabetic patients: a case-control study.

In the literature there is no agreement on the prevalence of chronic liver disease (CLD) and the role of hepatitis B virus (HBV) infection in diabetics. We undertook an epidemiological case-control study of the prevalence of CLD and HBV infection in 394 diabetics and 265 healthy subjects from Seriate and Como. The results did not show any significant differences between: 1) the prevalence of CLD in the diabetic population and in controls (4.8% vs 4.5%); 2) the prevalence of HBV infection in diabetics (HBsAg+: 8.3%; HBVAb+: 55.8%) and in controls (HBsAg+: 8.6%; HBVAb+: 54.7%); 3) the prevalence of HBV infection in diabetics with CLD (HBsAg+: 21%; HBVAb+: 52.6%) and in controls with CLD (HBsAg+: 16.6%; HBVAb+: 50%); 4) the prevalence of HBV infection in diabetics with and without CLD (HBsAg+: 21% vs 7.7%; HBVAb+: 52.6% vs 56%); 5) the prevalence of HBV infection in diabetics treated by injection and orally (HBsAg+: 6.9% vs 8.6%; HBVAb+: 58.3% vs 55.2%). The relative risk of CLD for the factor HBsAg+ was 3.2 in the diabetic population vs 1.4 in controls. In view of the presence of antidelta antibodies (HDVAb) in 25% of HBsAg+ diabetics with CLD and the lack of HBV markers in 26.3% of diabetics with CLD, we assume that other viruses (Delta, nonA-nonB) may play roles. Probably the interaction of all possible etiopathogenetic factors (alcohol, viruses, glycometabolic derangement) is determinant for CLD in diabetics.