Nine RCTs (n=1,403 patients; range 30 to 359) met the inclusion criteria. Seven trials reported the method of randomisation. Four trials reported adequate allocation concealment. None of the trials reported blinding or standardised criteria for initiation of renal replacement therapy. Four trials reported imbalances between groups at baseline. Four trials used an intention-to-treat analysis; treatment crossover occurred in fewer than 5% to 37.5% of trial participants.
There was no significant difference between continuous renal replacement therapy and intermittent renal replacement therapy in terms of mortality, recovery to renal replacement therapy independence and renal death. For treatment-related complications, continuous renal replacement therapy was associated with significantly fewer episodes of haemodynamic instability (OR 0.66, 95% CI 0.45 to 0.96, I2=47%), but not arrhythmic complications or bleeding episodes. These findings did not change in any of the sensitivity analyses. There was no evidence of publication bias from the Egger’s test or funnel plots.
For the assessment of heterogeneity, only the results for the I2 test were presented.