The goal of this prospective, randomised and double-blind pilot-study was to investigate the analgesic potency and the side-effects of tramadol enantiomers in clinical practice. One hundred patients recovering from orthopaedic surgery with a postoperative pain intensity of more than 50 on a visual analogue scale 0-100 mm (Table 1) were recruited for the study. They were treated in a randomised, double-blind way with a maximal dose of 150 mg i.v.(+)-,(-)-tramadol, racemate, or 15 mg i.v. morphine or saline in the placebo group (5 groups, 20 patients each). The primary criterium of efficacy was the number of responders defined as patients with a pain reduction of at least 20 on VAS after 40 min. In case of pain, responders were allowed to continue with the double-blind drug up to six hours. The non-responders were treated with morphine as the rescue analgesic. The secondary criterium was the incidence and severity of side-effects. Six patients terminated the study prematurely. One patient was excluded because of an allergic reaction to morphine, one patient could not be treated sufficiently with morphine, four were excluded because of protocol violations. There were 8 responders in the (+)-tramadol-,6 in the (-)-tramadol- and 6 in the racemate group, 16* (P < 0.05) in the morphine group, and 5 in the placebo group. Pain intensity after 40 min was reduced by 20 (p < 0.05), 17 (p < 0.05), 17 (p < 0.05), 36 (p < 0.01 vs placebo, p < 0.05 vs (+)-,(-)-tramadol, and racemate group) and 5 mm on the VAS in the (+)-, (-)-, (+/-)-tramadol-, morphine- and placebo-group, respectively. Thirty eight adverse events like nausea, vomiting, PCO2-increase, and urinary retention occurred in 20 patients, most frequently in the (+)-tramadol- and morphine group. Sedation was significantly less profound in the (-)-tramadol group 1-4 h postoperatively. There were no side-effect in the tramadol racemate group. The enantiomers were equal to the racemate in analgesic potency, but inferior by far to morphine. They showed more adverse events and, hence, can not be preferred to the racemate in postoperative pain therapy.