One hundred and thirty-one RCTs were included in the review. There were discrepancies between the tables and text that indicated lower doses than those reported below for meloxicam (3.75 mg/day), celecoxib (80 mg/day), etoricoxib (5 mg/day), and valdecoxib (1 mg/day). Statements on study quality were not supported by average scores in all cases.
Etodolac
Only two of 29 RCTs (n=5,775) were considered to be good quality. Etolodac (600 mg/day to 1,000 mg/day) was found to be equally effective and demonstrated equivalent or superior GI tolerability when compared to other non-selective NSAIDs. Significantly fewer clinical UGI events were reported, RR 0.32 (95% CI: 0.15, 0.71). No MIs were reported.
Meloxicam
In 16 RCTs (n=22,886), overall trial quality was considered to be moderate (mean score 3). Meloxicam (range 7.5 mg/day to 22.5 mg/day) was found to be of inferior or equivalent efficacy and demonstrated superior GI tolerability when compared to other non-selective NSAIDs. Significantly fewer clinical UGI events were reported, RR 0.53 (95% CI: 0.29, 0.97). There was insufficient data to assess MI risk.
Celecoxib
In 37 RCTs (median sample size 655 patients), overall trial quality was considered good, but withdrawals were high (20 per cent to 50 per cent). Celecoxib (200 mg/day to 800 mg/day) was found to be equally effective and of superior GI tolerability when compared to other non-selective NSAIDs. Significantly fewer clinical UGI events, RR 0.55 (95% CI: 0.40, 0.76) and complicated UGI events, RR 0.57 (95% CI: 0.35, 0.95) were reported. There was a significantly higher risk of MI, RR 1.77 (95% CI: 1.00, 3.11). The results of sub-group analyses were inconclusive.
Rofecoxib
In 23 RCTs (n=26,406), 21 trials were considered to be of good quality. Rofecoxib (12.5 mg/day to 50 mg/day) was found to be equally effective and had superior GI tolerability when compared with other non-selective NSAIDs. Significantly fewer clinical UGI events RR 0.43 (95% CI: 0.32, 0.57) and complicated UGI events, RR 0.40 (95% CI: 0.23, 0.70) were reported. There was a significantly higher risk of MI, RR 2.92 (95% CI: 1.36, 6.28). The results of sub-group analyses were inconclusive.
Etoricoxib
In seven RCTs (n=approximately 3,461), overall trial quality was judged to be moderate to good (median score 4). Etoricoxib (60 mg/day to 120 mg/day) was equally effective and of equivalent of superior GI tolerability when compared to other non-selective NSAIDs.
Valdecoxib
In 11 RCTs (n=9,293), overall trial quality was considered to be good. Valdecoxib (10 mg/day 80 mg/day) was equally effective and had equivalent or superior GI tolerability when compared to other non-selective NSAIDs. Significantly fewer complicated UGI events, RR 0.43 (95% CI: 0.19, 0.97) and lower risk of MI (although only six events were reported), RR 0.25 (95% CI: 0.06, 1.00) were reported.
Lumiracoxib
In 15 RCTs (median sample size 893), overall trial quality was judged to be good. Lumiracoxib (100 mg/day to 1,200 mg/day) appeared to be equally effective and of significantly superior GI tolerability than other non-selective NSAIDS. Significantly fewer clinical UGI events, RR 0.47 (95% CI: 0.37, 0.61) and complicated UGI events, RR 0.34 (95% CI: 0.23, 0.52) were reported. The results of sub-group analyses were inconclusive.
Direct COX-2 comparisons
In 14 RCTs (n=approximately 7,679), overall trial quality was considered to be good. Trials compared rofecoxib (12.5 mg/day to 25 mg/day) with celecoxib (200 mg/day) or valdecoxib (10 mg/day) or lumiracoxib (200-400 mg/day) and celecoxib (200-400 mg/day) with lumiracoxib (200 mg/day to 800 mg/day). All comparisons were equally effective and of equal tolerability. No comparisons of UGI events and MI were possible.