BACKGROUND This study prospectively evaluated the ability of aspiration to detect intravascular placement of multiple-orifice epidural catheters. METHODS Multiple-orifice, 20-gauge epidural catheters were inserted in 1,029 laboring women. Catheters were observed and aspirated for blood or cerebrospinal fluid before they were tested with 2 ml local anesthetic. If the results of this test were negative (no spinal anesthesia), the authors induced and maintained labor analgesia with a dilute local anesthetic and opioid solution. Patients with bilateral sensory change and effective labor analgesia had a "positive" epidural catheter. Women with unilateral block, inadequate analgesia despite some sensory change or those who delivered before being adequately assessed had "equivocal" catheters. Patients with neither analgesia nor sensory change had "negative" catheters. RESULTS Aspiration and observation identified 60 intravenously placed catheters. Six catheters, which were placed initially in a blood vessel, were withdrawn until aspiration was negative, and then the anesthetic was infused. Four of these catheters were positive and two were still positioned intravascularly. Two other catheters may have been intravenously placed despite negative results of aspiration. The incidence of false-negative results of aspiration was 0 to 2 of 1,085 (upper limit of 95% CI, 0.2% to 0.4%). No patient showed any signs or symptoms of local anesthetic toxicity. CONCLUSIONS Under the conditions of this study, which include using multiple-orifice catheters and dilute solutions of local anesthetic and opioid, aspiration and incremental drug injection alone safeguard against the risks of intravenously positioned local anesthetics. These results should not be extrapolated to other clinical settings without further study.