The extensive use of invasive procedures and of long-term norfloxacin prophylaxis in the management of cirrhotic patients may have influenced the epidemiology of bacterial infections in cirrhosis. We conducted a prospective evaluation of all bacterial infections diagnosed in patients with cirrhosis in a Liver Unit between April 1998 and April 2000. A total of 405 patients presented 572 bacterial infections in 507 admissions. Spontaneous bacterial peritonitis was the most frequent infection (138 cases). Gram-positive cocci were responsible for 53% of total bacterial infections in the study, being the main bacteria isolated in nosocomial infections (59%). Patients requiring treatment in an intensive care unit and those submitted to invasive procedures presented a higher rate of infections caused by gram-positive cocci (77% vs. 48%, P < .001 and 58% vs. 40%, P < .02, respectively). Fifty percent of culture-positive spontaneous bacterial peritonitis in patients on long-term norfloxacin administration (n = 93) and 16% in patients not receiving this therapy (n = 414) were caused by quinolone-resistant gram-negative bacilli, P = .01. The rate of culture-positive spontaneous bacterial peritonitis caused by trimethoprim-sulfamethoxazole-resistant gram-negative bacilli was also very high in patients on long-term norfloxacin administration (44% vs. 18%, P = .09). In conclusion, infections caused by gram-positive cocci have markedly increased in cirrhosis. This phenomenon may be related to the current high degree of instrumentation of cirrhotic patients. Quinolone-resistant spontaneous bacterial peritonitis constitutes an emergent problem in patients on long-term norfloxacin prophylaxis, with trimethoprim-sulfamethoxazole not being a valid alternative.

Prognostic stratification of patients with alcoholic hepatitis (AH) may improve the clinical management and facilitate clinical trials. We aimed at developing a scoring system capable of providing prognostic stratification of patients with AH. Patients with biopsy-proven AH were prospectively included between 2000 and 2006. The biochemical, clinical, portal hemodynamic and histological parameters were evaluated. A Cox regression model was used for univariate and multivariate analyses. A predictive score was built using variables obtained at admission identified in the multivariate analysis. The resulting score was validated in an independent prospective cohort. In total, 103 patients with biopsy-proven AH were included in the study cohort. Age, serum bilirubin, serum creatinine, and international normalized ratio (INR) independently predicted 90-day mortality. We generated the Age, serum Bilirubin, INR, and serum Creatinine (ABIC) score: (age × 0.1) + (serum bilirubin × 0.08) + (serum creatinine × 0.3) + (INR × 0.8). The area under the curve (AUC) was 0.82. Using the Kaplan-Meier analysis with the cutoff values of 6.71 and 9.0, we identified patients with low, intermediate, and high risk of death at 90 days (100%, 70%, and 25% of survival rate, respectively). Using the same cutoff values, the ABIC score also stratified patients according to their risk of death at 1 yr. These results were validated by a confirmatory cohort (N = 80). The ABIC score is a new tool that allows the stratification of risk of death in patients with AH at 90 days and 1 yr. This score can help improve the management of these patients and also help to perform clinical trials.