Article type
Year
Abstract
Objective: To determine the relative efficacy of non-steroidal anti-inflammatory drugs (NSAlDs) when used in the management of osteoarthritis (OA) of the knee.
Methods: A modified Cochrane Collaboration search strategy was used to identify RCTs in MEDLINE for 1966-95 and EMBASE for 1980-95. The quality of each study was scored by a standardised method. The efficacy for pain reduction and the Peto Odds Ratio (OR) for *withdrawals due to lack of efficacy* were calculated. These were the only outcome measures common to all trials.
Results: Of 1151 trials identified by the search strategy, 22 involved OA at the knee only and 16 of these fulfilled the inclusion criteria. The study methodological design quality was generally poor, with a median score of 3 (of a maximum of 8). Three studies compared etodolac with piroxicam. Patients receiving etodolac were less likely to withdraw due to lack of efficacy but the dose of etodolac was greater than the corresponding dose of piroxicam. For etodolac versus diclofenac or naproxen, there were no clear differences between treatments. Two studies showed a statistical difference in patient global assessment of condition, which favoured the trial NSAID. In both cases the trial NSAID was etodolac, used in doses approximately 25-44% greater than the reference NSAID. A further 2 studies showed a significant difference in pain relief between NSAIDs. In both cases the trial NSAID was etodolac used in higher doses than the reference.
Discussion: In spite of the large number of publications in this area, there are few RCTs. Furthermore, most RCTs comparing two or more NSAIDs suffer from substantial design weaknesses. This review found no substantial evidence related to efficacy, to distinguish between equivalent recommended doses of NSAIDS. Had studies employed appropriate doses of comparator drugs most would have been sufficiently powerful to detect clinically important differences in efficacy. As differences in efficacy between NSAIDS have not been recorded the selection of an NSAID for prescription for OA knee should be based upon relative safety, patient acceptability and cost.
Methods: A modified Cochrane Collaboration search strategy was used to identify RCTs in MEDLINE for 1966-95 and EMBASE for 1980-95. The quality of each study was scored by a standardised method. The efficacy for pain reduction and the Peto Odds Ratio (OR) for *withdrawals due to lack of efficacy* were calculated. These were the only outcome measures common to all trials.
Results: Of 1151 trials identified by the search strategy, 22 involved OA at the knee only and 16 of these fulfilled the inclusion criteria. The study methodological design quality was generally poor, with a median score of 3 (of a maximum of 8). Three studies compared etodolac with piroxicam. Patients receiving etodolac were less likely to withdraw due to lack of efficacy but the dose of etodolac was greater than the corresponding dose of piroxicam. For etodolac versus diclofenac or naproxen, there were no clear differences between treatments. Two studies showed a statistical difference in patient global assessment of condition, which favoured the trial NSAID. In both cases the trial NSAID was etodolac, used in doses approximately 25-44% greater than the reference NSAID. A further 2 studies showed a significant difference in pain relief between NSAIDs. In both cases the trial NSAID was etodolac used in higher doses than the reference.
Discussion: In spite of the large number of publications in this area, there are few RCTs. Furthermore, most RCTs comparing two or more NSAIDs suffer from substantial design weaknesses. This review found no substantial evidence related to efficacy, to distinguish between equivalent recommended doses of NSAIDS. Had studies employed appropriate doses of comparator drugs most would have been sufficiently powerful to detect clinically important differences in efficacy. As differences in efficacy between NSAIDS have not been recorded the selection of an NSAID for prescription for OA knee should be based upon relative safety, patient acceptability and cost.