Adding a small concentration of sufentanil to 0.125% bupivacaine for patient-controlled epidural analgesia during labor improved the quality of analgesia but did not modify the bupivacaine requirement. Reducing the sufentanil concentrations to 0.156 microg/mL decreased the pruritus intensity without reducing analgesia.

Applying the principle and equipment of patient-controlled analgesia, this double-blind randomised study was designed to determine the premedication dose of midazolam and to provide information about the need for preoperative anxiety control. The effects of patient-controlled premedication by i.v. midazolam were compared to those of a 1-hour i.v. infusion of a fixed dose of 4 mg. Two groups of 25 patients were studied prior to ambulatory surgery. Using a visual analogue scale graded from 0 (no anxiety) to 100, anxiety and short-term memory as well as vital signs, vigilance, cognition and orientation were assessed before premedication and then every 20 min until admission to the operating room. In both groups, 40% of the patients were free of anxiety before premedication. Sixty-four percent of patients in the self-pre-medicated group never used the pump. The dose of midazolam was only 0.7 0.2 mg (mean SD) in this group taken as a whole, 0.7, 1.1 mg (mean SD) in the 8 patients who had an anxiety score at least equal to 50 before premedication, and 1.7 0.3 mg (mean SD) in the 9 patients who chose to push the button of the patient-controlled device. Anxiety and short-term memory scores decreased significantly and similarly in the self-premedicated group and in the fixed-dose group. For similar intraoperative anaesthetic requirements, time from the end of anaesthesia for determining whether a patient was ready for admission to the postrecovery lounge was shorter in the self-premedicated group (26 +/- 8 min vs 32 +/- 8 min; mean +/- SD. Reduced doses of midazolam self-administered via a patient-controlled device can result in a relaxed preoperative period and amnesia in ambulatory surgery patients. However, the level of anxiety before premedication was low, which calls into question the legitimacy of the patient-controlled premedication in this kind of population. The expense of the pump was not justified by the small number of patients using it, probably because they were worried by the technical nature and the invasiveness of this technique.

(/Oz and Qs/Qt.

We read the paper by Adams et al. (l), who demonstrate the feasibility of recording motor evoked potentials (MEPs) despite 90% neuromuscular blockade added to isoflurane anesthesia, confirming results previously established by others (2,3). Advantages of such a technique, i.e., the suppression of patients' movements to provide stable operative field (movements less disturbing in spinal cord than in motor cortex stimulation), remain questionable in regard to the reliability of this recording and, consequently, its ability to diagnose spinal cord injury.Intraoperative changes in evoked potentials warn the surgeon of excessive retractor pressure to a degree that is compromising blood supply to the cord. The challenge in the intraoperative monitoring lies in the definition of changes considered as significant, and, from neurophysiologist's point of view, it is clear that depression of the signals by increasing fluorinated anesthetic depth or curarization or both at a crucial point can result in confusing changes. To illustrate, the case of a patient reported by Adams et al. whose MEPs were lost during surgery with postoperative neurologic deficit may indicate that less important lesions could have been missed and that curarization could increase the risk of false-negative results. Data demonstrating that paralysis even in a controlled manner does not decrease the sensibility of MEPs are still missing, and, finally, despite this new report (l), we are not convinced that damages which impair MEP amplitude from, e.g., 50% could be clearly detected on a MEP decreased to 10% of the initial value by partial neuromuscular blockade.Even if it means recording responses of small amplitude after epidural stimulation under neuromuscular blockade, why not use, for instance, "neurogenic" motor evoked potentials, as described by Owen et al. (4)? They are recorded from subdermal needle electrodes in the vicinity of the sciatic nerve at the popliteal fossa; the orthodromic motor component is distinguishable from the antidromic sensory component; they are not affected by neuromuscular blocking drugs; and their reliability has been demonstrated by experimental and clinical data (4,5).