The respiratory and analgesic effects of i.v. meperidine, tramadol and their correlation with plasma concentrations of meperidine, tramadol and O-demethyltramadol were determined. Forty-eight patients after total hip or knee replacement were randomly distributed into 3 groups (n = 16 each). At the time of analgesia request, they received in a double-blind manner, i.v. single doses of 100 mg meperidine, 100 mg tramadol, or saline. Thirty minutes after treatment, patients who requested additional analgesia were rescued with 75 mg diclofenac and morphine as required. Patients were evaluated at the time of analgesia request and at set intervals during 4 h. Meperidine induced sedation (p < 0.05), respiratory depression (tidal volume, p < 0.047; respiratory rate, p < 0.004; % O2 Sat, p < 0.036), and hypercapnia (PaCO2, p < 0.002). Incidence of nausea and vomiting was higher with tramadol (p < 0.02). For the first 30 min, meperidine produced lower pain intensity scores than tramadol or saline (p < 0.05). At this time, 14/16 patients on saline, 8/16 on meperidine and 11/16 on tramadol were rescued. Onset for meperidine analgesia was 10 min and > 30 min for tramadol. Both opioids produced similar degree of analgesia in patients who were not rescued. A negative correlation (r = -0.99) between analgesia and tramadol concentrations and a poor positive correlation (r = +0.54) with O-demethyltramadol (a metabolite of tramadol) was observed. Pain intensity differences correlated negatively with meperidine plasma concentrations during the first 30 min (r = -0.97) and positively thereafter (r = +0.92). In the present study, meperidine and tramadol produced comparable analgesia, with a different time course profile, but meperidine induced sedation and respiratory depression while tramadol did not.