Need for common internal controls when assessing the relative efficacy of pharmacologic agents using a meta-analytic approach: case study of cyclooxygenase 2-selective inhibitors for the treatment of osteoarthritis.

Abstract

OBJECTIVE:

To evaluate the role of common internal controls in a meta-analysis of the relative efficacy of cyclooxygenase 2-selective inhibitors (coxibs) in the treatment of osteoarthritis (OA).

METHODS:

A systematic search of Medline and US Food and Drug Administration electronic databases was performed to identify randomized, placebo-controlled clinical trials of coxibs (etoricoxib, celecoxib, rofecoxib, valdecoxib) in patients with hip and/or knee OA. The effect size for coxibs and common active internal controls (nonsteroidal antiinflammatory drugs [NSAIDs], naproxen) were determined by the mean changes from baseline in Western Ontario and McMaster Universities Osteoarthritis Index pain subscores as compared with placebo.

RESULTS:

The effect size for all coxib groups combined (0.44) indicated greater efficacy as compared with placebo, but significant heterogeneity (P < 0.0001) was observed. Rofecoxib at dosages of 12.5 mg/day and 25 mg/day and etoricoxib at a dosage of 60 mg/day had similar effect sizes (0.68 and 0.73, respectively), but these effect sizes were comparatively greater than those for both celecoxib at dosages of 200 mg/day and 100 mg twice daily or valdecoxib at a dosage of 10 mg/day (0.26 and 0.16, respectively). The effect sizes for NSAIDs or naproxen versus placebo, as determined using data from rofecoxib/etoricoxib trials, were consistently higher than the effect sizes derived from trials of celecoxib/valdecoxib. Significant heterogeneity was present in the overall effect size for NSAIDs (P = 0.007) and naproxen (P = 0.04) groups based on data available from all coxib trials.

CONCLUSION:

Coxibs and common active internal controls showed larger effect sizes versus placebo in the rofecoxib/etoricoxib trials than in the celecoxib/valdecoxib trials. These findings suggest systematic differences among published coxib trials and emphasize the need for direct-comparison trials. In the absence of such trials, common internal controls should be assessed when performing indirect meta-analytic comparisons.

• Authors: Chin Lee, Elke Hunsche, Robert Balshaw, Sheldon X Kong, Thomas J Schnitzer
• Journal: Arthritis and rheumatism (Arthritis Rheum) Vol. 53 Issue 4 Pg. 510-8 (Aug 15 2005) ISSN: 0004-3591 [Print] United States
• PMID: 16082648 (Publication Type: Journal Article, Research Support, N.I.H., Extramural, Research Support, U.S. Gov't, P.H.S.)

Chemical References

• Anti-Inflammatory Agents, Non-Steroidal
• Cyclooxygenase Inhibitors

Topics

• Anti-Inflammatory Agents, Non-Steroidal (therapeutic use)
• Cyclooxygenase Inhibitors (therapeutic use)
• Humans
• Meta-Analysis as Topic
• Osteoarthritis (drug therapy)