Eighteen prospective controlled trials (n=3,090) were included: nine full papers and nine published as an abstract only. The overall quality of the included studies was described as low: 33% reported allocation concealment and information on blinding was provided in 61% of trials; 73% of these did not use blinding.
There was a statistically significant reduction in incidence of CI-AKI with use of sodium bicarbonate compared to control (RR 0.66, 95% CI 0.45 to 0.95; 18 trials) in the overall analysis. When studies were subgrouped by the type of contrast procedure, there was a statistically significant benefit with sodium bicarbonate in the studies of coronary procedures only (RR 0.65, 95% CI 0.42 to 1.00; 10 trials) and emergency coronary procedures (RR 0.13, 95% CI 0.04 to 0.42; two trials), but not in the pooling of studies of coronary and non-coronary procedures (RR0.68, 95% CI: 0.30 to 1.52; eight trials). There was moderate heterogeneity (>50%) in all the analyses except of acute coronary procedures. Other subgroup analyses based on a small number of studies are reported in the paper.
There was no statistically significant difference between sodium bicarbonate and control in the need for renal replacement therapy (11 trials) or in-hospital mortality (five trials).
Visual inspection of a funnel plot suggested publication bias when only data from full-papers were included, but not when all the included data (full papers and abstracts only) were considered.