To determine the time to onset of analgesia of rofecoxib based on a patient-level meta-analysis of randomized, placebo-controlled, postoperative oral surgery pain studies. A search on MEDLINE and of Merck data on file was conducted to identify studies that met the inclusion criteria. Meta-analysis inclusion criteria required that patients were treated with a single oral dose of rofecoxib 50 mg when they experienced moderate or severe pain after surgical extraction of > or = 2 third molars; study design involved patient randomization, double-blinding, and matching placebo, and onset data from individual patients were available. The meta-analysis of time to onset also required that studies used the two-stopwatch method. Eleven studies fulfilled the onset criteria and included patients who received a single dose of rofecoxib 50 mg (N = 1220) or placebo (N = 483). These studies were analyzed to determine time to onset of analgesia, time to perceptible pain relief, percentage of patients achieving onset of analgesia, and duration of analgesia. Six of the 11 studies included a nonselective nonsteroidal anti-inflammatory drug (N = 303) and were included in the onset meta-analysis for comparison. The meta-analysis of overall efficacy also required that data on total pain relief scores over 8 hours were available. Over-all effectiveness of analgesia was based on analysis of 13 studies involving 1330 rofecoxib patients and 570 placebo patients on the endpoints of total pain relief scores over 8 hours and patient global assessment of response to therapy at 24 hours. Eight of the 13 studies with a nonselective nonsteroidal anti-inflammatory drug comparator (N = 391) were included for the efficacy meta-analysis. Patient demographics and baseline characteristics were similar across treatment groups in each study. Median time to onset of analgesia for rofecoxib was 34 minutes (95% CI, 31-38 minutes), significantly faster than placebo, which did not achieve onset within the 4 hours the assessment was conducted (P < 0.001). Duration of analgesia for rofecoxib 50 mg was > 24 hours. Rofecoxib achieved a greater mean total pain relief score over 8 hours than placebo (17.4 versus 4.4; P < 0.001) and a greater patient response rate on patient global assessment of response to therapy at 24 hours than placebo (73% versus 16%; P < 0.001). Outcomes were similar between the rofecoxib group and the nonselective nonsteroidal anti-inflammatory drug group. In this meta-analysis of over 1200 rofecoxib-treated patients, a single dose of rofecoxib 50 mg demonstrated a rapid onset of analgesia in approximately half an hour combined with sustained effectiveness, supporting its use as a treatment of acute pain.