Four trials with a total of 1,730 participants were included. The funnel plot suggested that there may have been some publication bias. Three of the trials had a low risk of bias (all quality criteria fulfilled) and one had a moderate risk of bias (unclear allocation concealment).
No significant difference in the incidence of post-ERCP pancreatitis was found in the meta-analysis when comparing allopurinol with placebo (RR 0.86, 95% CI: 0.42, 1.77, p=0.68). The number of patients developing post-ERCP pancreatitis was 78 in the allopurinol groups and 83 in the placebo groups. There was evidence of significant statistical heterogeneity (Χ2 p=0.006, I2=75.8%). Of the four included RCTs, one found a significant effect in favour of allopurinol (possibly linked to some demographic differences between the comparison groups) and the other three did not.
A prophylactic effect of allopurinol was not found in any of the subgroups examined (single centre versus multicentre, high versus low dose of allopurinol, high versus low pre-treatment risk of post-ERCP pancreatitis and high versus low/moderate risk of bias).