Seventeen studies were included; 8 reports of 9 RCTs (n=43,143), 3 case-control studies (n=102,632) and 6 retrospective cohort studies (n=742,395).
Two RCTs (n=10,003) and one retrospective cohort study (n=36,545) evaluated celecoxib.
One RCT (n=18,325) evaluated lumiracoxib.
Three RCTs (n=12,682) and 2 retrospective cohort studies (n=512,449) evaluated rofecoxib.
Two RCTs (n=2,133) evaluated parecoxib and valdecoxib
Three case-control studies (n=102,632) and 3 retrospective cohort studies (n=193,401) compared celecoxib versus rofecoxib.
Celecoxib.
One RCT (n=7,968) reported no difference in CV events between celecoxib 400 mg and ibuprofen 800 mg thrice daily or diclofenac 75 mg twice daily.
One RCT (n=2,035) reported that high-dose celecoxib (400 mg twice daily) was associated with a significant increase in CV events compared with placebo (RR 3.4, 95% CI: 1.4, 7.8), but reported no significant difference between low-dose celecoxib (200 mg twice daily) and placebo.
One retrospective cohort study (n=36,545) reported no significant difference in CV events between celecoxib and meloxicam.
Rofecoxib. One RCT (n=8,076) reported that naproxen 500 mg twice daily was associated with a significant decrease in the risk of MI compared with rofecoxib 50 mg/day (RR 0.2, 95% CI: 0.1, 0.7).
One report of 2 RCTs (n=2,020) reported mixed results: one RCT (n=1,042) suggested that the risk of CV events was lower with rofecoxib than with nabumetone, but not placebo (0.2% versus 0.4% versus 0%, respectively), while the other RCT (n=978) suggested that CV events were more common with rofecoxib than with nabumetone or placebo (1.5% versus 0.5% versus 0.5%, respectively).
One RCT (n=2,586) reported that rofecoxib was associated with a significant increase in the risk of CV events compared with placebo (RR 1.92, 95% CI: 1.19, 3.11). One retrospective cohort study (n=478,094) reported that users of higher dose rofecoxib (25 mg/day or more) had a non significantly higher risk of developing coronary disease compared with non-users of rofecoxib (RR 1.7, 95% CI: 0.98, 2.95). New users of rofecoxib had a significantly higher risk of a CV event (RR 1.93, 95% CI: 1.09, 3.42). One retrospective cohort study (n=34,355) reported no significant difference in the risk of CV events between rofecoxib and meloxicam.
Parecoxib and valdecoxib.
One RCT (n=462) reported that parecoxib and valdecoxib were associated with an increase in the risk of serious events (including MI, cerebrovascular disease, pulmonary thromboembolism and sternal wound infection) compared with placebo (19% versus 9.9%).
One RCT (n=1,671) reported that parecoxib and valdecoxib were associated with a significant increase in the risk of CV events compared with placebo (RR 1.9, 95% CI: 1.1, 3.2).
Celecoxib versus rofecoxib.
Three case-control studies (n=54,475, n=39,639 and n=8,518) reported that rofecoxib was associated with an increased risk of CV events compared with celecoxib: the OR of acute MI was 1.24 (95% CI: 1.01, 1.46), the RR of MI was 1.59 (95% CI: 1.10, 2.32), and the OR of MI was 2.72 (95% CI: 1.24, 5.95), respectively. Two of the case-control studies reported that higher doses of rofecoxib (25 mg/day or more) were associated with a significantly increased risk of CV events compared with celecoxib.
One retrospective study (n=127,427) reported no significant difference in the risk of MI between celecoxib versus non-NSAID users or between rofecoxib versus non-NSAID users.
One retrospective study (n=59,724) reported that rofecoxib significantly increased the risk of MI compared with no NSAID use (RR 1.24, 95% CI: 1.05, 1.46), but found no significant difference between celecoxib and no NSAID use. One retrospective study (n=6,250) reported no significant difference in the risk of CV events between either rofecoxib or celecoxib and non-naproxen NSAIDs.
Lumiracoxib.
One RCT (n=18,325) reported no significant difference in the risk of CV events between lumiracoxib (400 mg/day) and naproxen (500 mg twice daily).
Etoricoxib.
No published studies that assessed safety were identified.