Background: Hyperuricemia is a common metabolic disease and has become a public health problem because of its increasing prevalence and association with comorbidities. Allopurinol and febuxostat are recommended as the first-line treatments for hyperuricemia and gout. But cardiovascular safety between febuxostat and allopurinol is still controversial. The purpose of this study is to compare the cardiovascular safety of XOIs and placebo in hyperuricemic patients with or without gout.Methods: PubMed, Embase via OVID, Cochrane Library, CNKI, Wanfang, and VIP were searched from their earliest records to February 8th 2021. ClinicalTrials.gov was also searched for unpublished data. The reference lists of included studies and relevant review articles investigating the cardiovascular safety of XOIs in hyperuricemia patients are screened for potentially eligible studies. Randomized controlled trials (RCTs) evaluating allopurinol (100~900 mg/d), febuxostat (20~120 mg/d), or placebo for hyperuricemia were included. The outcomes were incidence of MACE, non-fatal MI, non-fatal stroke, and cardiovascular death. We conducted a Bayesian random-effects network meta-analysis on the included randomized controlled trials using the Markov Chain Monte Carlo simulation method. The grading of recommendations assessment, development, and evaluation (GRADE) approach was used to assesses the certainty of the evidence.Results: Ten RCTs with 18,004 participants were included. The network estimates showed that there was no significant difference observed among febuxostat, allopurinol, and placebo regarding outcomes. The certainty of the evidence ranged from very low to moderate. The probabilities of rankings and SUCRA showed that compared to placebo, febuxostat, and allopurinol might prevent adverse cardiovascular events.Conclusion: Febuxostat is not associated with increasing risk of adverse cardiovascular events compared to allopurinol; and compared to placebo, whether febuxostat and allopurinol reduce the risk of adverse cardiovascular events remains uncertain.