The hepatic safety and tolerability of the cyclooxygenase-2 selective NSAID celecoxib: Pooled analysis of 41 randomized controlled trials

Objective: To assess the hepatic safety and tolerability of celecoxib versus placebo and three commonly prescribed nonselective nonsteroidal anti-inflammatory drugs (NSAIDs). Research design and methods: This was a retrospective, pooled analysis of a 41-study dataset involving patients with osteoarthritis, rheumatoid arthritis, ankylosing spondylitis, chronic low back pain, and Alzheimer's disease. Criteria for selection of studies were: (1) Randomized, parallel-group design and planned treatment duration of ≥2 weeks (2) ≥1 placebo or NSAID comparator (3) ≥1 arm with celecoxib at total daily dose of ≥200 mg (4) Data available as of October 31, 2004 Data were pooled by treatment and subject from the safety analysis population of included studies. Treatment-emergent hepatobiliary adverse events (AEs) were compared for celecoxib <200 mg/day (943 patients), 200 mg/day (12 008 patients), 400 mg/day (7380 patients), and 800 mg/day (4602 patients); placebo (4057 patients); diclofenac 100-150 mg/day (7639 patients); naproxen 1000 mg/day (2953 patients); and ibuprofen 2400 mg/day (2484 patients). Hepatobiliary laboratory abnormalities were also analyzed. Results: There were no cases of liver failure, treatment-related liver transplant, or treatment-related hepatobiliary death. Incidence of serious hepatic AEs was low, with 13 (0.05%) serious hepatic AEs among 24 933 celecoxib-treated patients, and 16 (0.21%) among 7639 diclofenac-treated patients. No patients receiving celecoxib or any nonselective NSAID met criteria for Hy's rule (alanine aminotransferase [ALT] ≥3 x upper limit of normal [ULN] with bilirubin ≥2 x ULN). The incidence of notable (≥5 x ULN) and severe (≥10 x ULN) ALT elevations was similar for all treatment groups except diclofenac. Significantly fewer hepatobiliary AEs were reported for celecoxib (any dose; 1.11%) than for diclofenac (vs. 4.24%, p < 0.0001); for ibuprofen (vs. 1.53%, p = 0.06) and placebo (vs. 0.89%, p = 0.21) the incidence of AEs was comparable to celecoxib. Limitations: A number of limitations should be considered when evaluating the results: findings were limited by the quality and reporting of the studies selected; difficulty in estimating the incidence of AEs due to the low frequency of events; acetaminophen not included as an active comparator. Conclusions: In this pooled analysis, the incidence of hepatic AEs in patients treated with celecoxib was similar to that for both placebo-treated patients and patients treated with ibuprofen or naproxen, but lower than for diclofenac. © 2009 Informa UK Ltd. All rights reserved.