Systematic review with network meta-analysis: the comparative effectiveness and safety of interventions in patients with overt hepatic encephalopathy.

BACKGROUND

Interventional treatment for overt hepatic encephalopathy (OHE), includes non-absorbable disaccharides, neomycin, rifaximin, L-ornithine-L-aspartate and branched chain amino acids (BCAA). However, the optimum regimen remains inconclusive.

AIM

To compare interventions in terms of patients' adverse events and major clinical outcomes.

METHODS

Literature search of PubMed, Embase, Scopus, and Cochrane Library studies published up to July 31 2014. RCTs of above interventions in OHE patients were included. Network meta-analysis combined direct and indirect evidence to estimate odds ratios (ORs) and mean difference (MD) between treatments and the probabilities of ranking for treatment based on clinical outcomes.

RESULTS

Twenty eligible RCTs were included. When compared with observation, only L-ornithine-L-aspartate (OR 3.71, P < 0.001) and BCAA (OR 3.37, P < 0.001) improved clinical efficacy significantly. However, when L-ornithine-L-aspartate was compared with BCAA, non-absorbable disaccharides and neomycin, there was a trend suggesting that L-ornithine-L-aspartate may be the most effective intervention with respect to clinical improvement (OR 1.10), rifaximin (OR 1.31), non-absorbable disaccharides (OR 2.75), neomycin (OR 2.22). In addition, L-ornithine-L-aspartate (MD -20.18, 95% CI -40.12 to -0.27) provided a significant reduction in blood ammonia concentration compared with observation. Neomycin appeared to be associated with more adverse events in comparison with non-absorbable disaccharides (OR 10.15), rifaximin (OR 17.31), L-ornithine-L-aspartate (OR 3.16) or BCAA (OR 7.69).

CONCLUSIONS

L-ornithine-L-aspartate treatment may show a trend in superiority for clinical efficacy among standard interventions for OHE. Rifaximin shows the greatest reduction in blood ammonia concentration, and treatment with neomycin demonstrates a higher probability in causing adverse effects among the five compared interventions.

© 2015 John Wiley & Sons Ltd.