The effects of poorly/non-absorbable antibiotics on hepatic venous pressure gradient (HVPG) are debated.
To analyze the effects of rifaximin or norfloxacin on HVPG and on markers of bacterial translocation and proinflammatory cytokines.
We performed a systematic search of randomized clinical trials (RCTs) involving patients with cirrhosis and portal hypertension, assessing the effect of rifaximin or norfloxacin vs control on HVPG. Pooled analyses were based on random-effects models, heterogeneity was assessed by Cochran\'s Q, I2 statistic and subgroup analyses.
Five studies (215 patients) were included. Risk of bias was high in three. We found no significant differences using antibiotics versus control. The summary mean difference in HVPG was of -0.55 mmHg (95%CI:-1.52, 0.42; P = 0.27), with moderate heterogeneity (P = 0.15; I2 = 40%). RCTs with longer therapy (60-90 days) used non-selective-beta-blockers (NSBB) in both antibiotics and control arms. Subgroup analysis showed a significantly greater reduction in HVPG in the combination arm over controls (mean difference -1.46 mmHg [95%CI: -2.63, -0.28; P = 0.01]) with no heterogeneity (P = 0.46; I2 = 0%). Serum lipopolysaccharide-binding protein (LBP) significantly decreased with antibiotics, but with high heterogeneity (P < 0.001; I2 = 92%).
Rifaximin or norfloxacin did not significantly reduce HVPG in patients with cirrhosis and portal hypertension. Studies using antibiotic for longer periods on top of NSBB showed a significant decrease in HVPG.

OBJECTIVE: Liver stiffness measurement (LSM), using elastography, can independently predict outcomes of patients with chronic liver diseases (CLDs). However, there is much variation in reporting and consistency of findings. We performed a systematic review and meta-analysis to evaluate the association between LSM and outcomes of patients with CLDs.
METHODS: We performed a systematic review of the literature, through February 2013, for studies that followed up patients with CLDs prospectively for at least 6 months and reported the association between baseline LSM and subsequent development of decompensated cirrhosis or hepatocellular carcinoma (HCC), as well as mortality. Summary relative risk (RR) estimates per unit of LSM and 95% confidence intervals (CIs) were estimated using the random effects model.
RESULTS: Our final analysis included 17 studies, reporting on 7058 patients with CLDs. Baseline LSM was associated significantly with risk of hepatic decompensation (6 studies; RR, 1.07; 95% CI, 1.03-1.11), HCC (9 studies; RR, 1.11; 95% CI, 1.05-1.18), death (5 studies; RR, 1.22; 95% CI, 1.05-1.43), or a composite of these outcomes (7 studies; RR, 1.32; 95% CI, 1.16-1.51). We observed considerable heterogeneity among studies-primarily in the magnitude of effect, rather than the direction of effect. This heterogeneity could not be explained by variations in study locations, etiologies and stages of CLD, techniques to measure liver stiffness, adjustment for covariates, or method of imputing relationship in the meta-analysis.
CONCLUSIONS: Based on a meta-analysis of cohort studies, the degree of liver stiffness is associated with risk of decompensated cirrhosis, HCC, and death in patients with CLDs. LSM therefore might be used in risk stratification.