We used computer-controlled individually tailored infusions to study relationships between plasma drug concentration and opioid effects, and to evaluate the therapeutic margins of alfentanil, fentanyl and morphine in human subjects. In order to compare the 3 drugs, we infused each opioid to 3 different steady-state target plasma concentrations during separate 8 h test periods so that concentration-effect curves could be defined for each opioid and subject. Dental electrical stimulation produced a consistent degree of baseline experimental pain, and we measured the influence of increasing plasma opioid concentrations on pain intensity and the magnitude of pain-related evoked potentials. We also quantified ventilatory function and subjective side-effects during baseline (no drug), at the 3 target plasma concentrations with each drug. Finally, we measured actual plasma opioid concentrations during each phase of the infusion period. This procedure allowed us to calculate for each opioid the plasma concentration required to produce a 50% decrease in reported pain intensity and evoked potential amplitude (IC50). Subsequent calculation of side-effect magnitudes at the analgesic IC50s permitted direct comparisons of therapeutic margins between alfentanil, fentanyl and morphine. We found a robust relationship between plasma drug concentration and analgesic, ventilatory, and subjective-effect magnitudes for each opioid in this study. We conclude that the magnitudes of individual side-effects associated with equianalgesic, steady-state plasma concentrations of these 3 mu receptor-selective opioids do not differ across drugs.