Summary of our confidence in effect estimates and ranking of treatments.

1Dominated by evidence at high or moderate risk of bias.2No convincing evidence for the plausibility of the transitivity assumption.3Predictive intervals for treatment effect include effects that would have different interpretations (there is additionally no convincing evidence for the plausibility of the transitivity assumption).4Confidence intervals include values favouring either treatment.560% of the information is from studies at moderate risk of bias.6Moderate level of heterogeneity, and some evidence of inconsistency in the network.