It has recently been reported that morphine given in low doses intra-articularly can produce significant analgesia in patients undergoing arthroscopic knee joint surgery. Data are lacking on the effect of other opioids using a local approach for drug delivery. We studied the analgesic effect of intra-articular opioids in 70 patients, divided into 7 groups, subjected to arthroscopic knee surgery in general anesthesia. The dimension of the study was based on a power of 0.8 to detect a 25% difference in pain intensity between those receiving opioids locally versus systemically (alpha = 0.05 and beta = 0.20). Following surgery, but before terminating anesthesia, the patients received one of the following combinations: 1 mg morphine intra-articularly (i.art.) + saline intramuscularly (i.m.), 10 mg pethidine i.art + saline i.m., or 10 micrograms fentanyl i.art + saline i.m. In three additional groups the three opioids were given i.m. and saline given i.art. An additional control group received saline i.art. + i.m. We did not find any significant difference between the groups considering postoperative pain intensity, need for analgesics or considering time to standing/walking or to discharge, analysing each opioid independently. There was, however, a tendency for pethidine i.art. to produce the lowest pain scores both at rest and during movement (P = 0.06). If analysing the results with regards to if opioids were given intra-articularly or systemically, not considering the type of opioid given, we did however, find a significantly lower total sum of pain scores at movement following local administration (P < 0.05). No specific side-effects were detected. We conclude that pethidine given intra-articularly merits further investigation with respect to postoperative analgesia following the activation of peripheral opioid mechanisms.