Preventive Strategies for Contrast-Induced Acute Kidney Injury in Patients Undergoing Percutaneous Coronary Procedures: Evidence From a Hierarchical Bayesian Network Meta-Analysis of 124 Trials and 28 240 Patients.

BACKGROUND: The effectiveness of currently available effective preventive strategies for contrast-induced acute kidney injury (CIAKI) is a matter of debate.

METHODS AND RESULTS: We performed a Bayesian random-effects network meta-analysis of 124 trials (28 240 patients) comparing a total of 10 strategies: saline, statin, N-acetylcysteine (NAC), sodium bicarbonate (NaHCO3 ), NAC+NaHCO3 , ascorbic acid, xanthine, dopaminergic agent, peripheral ischemic preconditioning, and natriuretic peptide. Compared with saline, the risk of CIAKI was reduced by using statin (odds ratio [OR], 0.42; 95% credible interval [CrI], 0.26-0.67), xanthine (OR, 0.32; 95% CrI, 0.17-0.57), ischemic preconditioning (OR, 0.48; 95% CrI, 0.26-0.87), NAC+NaHCO3 (OR, 0.50; 95% CrI, 0.33-0.76), NAC (OR, 0.68; 95% CrI, 0.55-0.84), and NaHCO3 (OR, 0.66; 95% CrI, 0.47-0.90). The benefit of statin therapy was consistent across multiple sensitivity analyses, whereas the efficacy of all the other strategies was questioned by restricting the analysis to high-quality trials. Overall, high heterogeneity was observed for comparisons involving xanthine and ischemic preconditioning, although the impact of NAC and xanthine was probably influenced by publication bias/small-study effect. Hydration alone was the least effective preventive strategy for CIAKI. Meta-regressions did not reveal significant associations with baseline creatinine and contrast volume. In patients with diabetes mellitus, no strategy was found to reduce the incidence of CIAKI.

CONCLUSIONS: In patients undergoing percutaneous coronary procedures, statin administration is associated with a marked and consistent reduction in the risk of CIAKI compared with saline. Although xanthine, NAC, NaHCO3 , NAC+NaHCO3 , ischemic preconditioning, and natriuretic peptide may have nephroprotective effects, these results were not consistent across multiple sensitivity analyses.