Table 2

 Characteristics of randomised controlled trials

TrialDesignOADurationMean age (range or SD)Male/femaleBPS (%)Comparison (mg/day)n1n2ES (SE)
OA, osteoarthritis; DB-C, double blind crossover; DB-P, double blind parallel; RCT, randomised controlled trial, no further information for blindness; BPS (%), percentage of baseline pain score relative to maximum pain score on scale: 0%, no pain; 50%, moderate pain; 100%, severe pain; Par, paracetamol; Dic, diclofenac; Cele, celecoxib; Ibu, ibuprofen; Rof, rofecoxib; Nap, naproxen; ES, effect size; SE, standard error; n1, number of patients in the treatment group; n2, number of patients in the control group.
*Only 6 week results were available for the primary outcome (pain at baseline and end point). The 2 year study mainly looked at the withdrawals due to lack of efficacy and side effects.
Paracetamol v placebo
Amadio 19836DB-CKnee6 Weeks64 (43–38)3/22Par (4000)Placebo2525
Zoppi 199527DB-PHip/knee1 Week56 (20–70)23/3765Par (3000)Placebo3030
NSAIDs v paracetamol v placebo
Case 200313DB-PKnee12 Weeks62 (40–75)41/4134Dic (150)Par (4000)25290.36 (0.28)
Par (4000)Placebo29280.09 (0.27)
Pincus 200325DB-CHip/knee6 Weeks63195/32951Cele (200)Par (4000)1781680.14 (0.11)
Par (4000)Placebo1681690.23 (0.11)
NSAIDs v paracetamol
Bradley 199110DP-PKnee4 Weeks56 (11.6)47/13750Ibu (1200)Par (4000)6261−0.04 (0.28)
Ibu (2400)Par (4000)61610.02 (0.18)
Geba 200223DB-PKnee6 Weeks63 (39–91)121/26160Rof (12.5)Par (4000)96940.14 (0.15)
Rof (25)Par (4000)95940.46 (0.15)
Cel (200)Par (4000)97940.16 (0.14)
March 19948n of 1Any6 Weeks64 (38–85)5/20Dic (100)Par (2000)15150.31 (0.37)
Pincus 200124DB-CHip/knee6 Weeks61 (19.6)67/16053Dic (150)Par (4000)1121150.28 (0.13)
Shen 200326RCTKnee3 MonthsRof (25)Par (4000)10100.08 (0.45)
Williams 19939DB-PKnee2 Years*60 (33–85)44/13453Nap (750)Par (2600)73750.32 (0.17)