Human recombinant activated factor VII for upper gastrointestinal bleeding in patients with liver diseases

Background Mortality from upper gastrointestinal bleeding in patients with liver disease is high. Recombinant human activated factor VII (rHuFVIIa) has been suggested for patients with liver disease and upper gastrointestinal bleeding. Objectives To assess the beneficial and harmful effects of rHuFVIIa in patients with liver disease and upper gastrointestinal bleeding. Search methods We searched the Cochrane Hepato-Biliary Group Controlled Trials Register (December 2011), the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library (Issue 4, 2011), MEDLINE (1948 to December 2011), EMBASE (1980 to December 2011), Science Citation Index Expanded (1900 to December 2011), and LILACS (December 2011). We sought additional randomised trials from the reference lists of the trials and reviews identified through the electronic searches. Selection criteria Randomised clinical trials. Data collection and analysis Outcome data from randomised clinical trials were extracted and were presented using random-effects model meta-analyses. Data on the risk of bias in the included trials were also extracted. Main results We included two trials with 493 randomised participants with various Child-Pugh scores. The trials had a low risk of bias. The rHuFVIIa administration did not reduce the risk of mortality within five days (21/288 (7.3%) versus 15/205 (7.3%); risk ratio (RR) 0.88, 95% confidence interval (CI) 0.48 to 1.64, I-2 = 49%) and within 42 days (5/286 (1.7%) versus 36/205 (17.6%); RR 1.01, 95% CI 0.55 to 1.87, I-2 = 55%) when compared with placebo. Trial sequential analysis demonstrated that there is sufficient evidence to exclude that rHuFVIIa decreases mortality by 80%, but there is insufficient evidence to exclude smaller effects. The rHuFVIIa did not increase the risk of adverse events by number of patients (218/297 (74%) and 164/210 (78%); RR 0.94, 95% CI 0.84 to 1.04, I-2 = 1%), serious adverse events by adverse events reported (164/590 (28%) versus 123/443 (28%); RR 0.91, 95% CI 0.75 to 1.11, I-2 = 0%), and thromboembolic adverse events (16/297 (5.4%) versus 14/210 (6.7%); RR 0.80, 95% CI 0.40 to 1.60, I-2 = 0%) when compared with placebo. Authors' conclusions We found no evidence to support or reject the administration of rHuFVIIa for patients with liver disease and upper gastrointestinal bleeding. Further adequately powered randomised clinical trials are needed in order to evaluate the proper role of rHuFVIIa for treating upper gastrointestinal bleeding in patients with liver disease. Although the results are based on trials with low risk of bias, the heterogeneity and the small sample size result in rather large confidence intervals that cannot exclude the possibility that the intervention has some beneficial or harmful effect. Further trials with alow risk of bias are required to make more confident conclusions about the effects of the intervention.