In a randomized trial, we compared topical anesthesia by a lidocaine-prilocaine cream (EMLA; Laboratorie ASTRA, Manterre, France) with subcutaneous local lidocaine infiltration for radial artery cannulation. Patients included 538 adults scheduled for coronary angiography using a radial approach. EMLA was applied 2 h before radial cannulation, and lidocaine infiltration was performed 5 min before cannulation. The primary end point was pain as assessed by a verbal numerical scale (0 = no pain, 10 = extreme pain). Pain was less severe in the EMLA group than in the lidocaine infiltration group (Score of 2 vs 7; P = 0.0001). Additional lidocaine infiltration was required significantly less frequently in the EMLA group (relative risk 0.19). The failure rate of cannulation was significantly lower in the EMLA group (relative risk 0.38), and insertion time was shorter (4 versus 6 min). We conclude that EMLA, compared with lidocaine infiltration, reduces pain associated with radial artery cannulation and improves the success rate of the procedure. Routine application of EMLA should be performed in awake patients 2 h before radial artery cannulation. CONCLUSIONS In a randomized trial, we compared topical anesthesia by a lidocaine-prilocaine cream (EMLA) with subcutaneous local lidocaine infiltration for radial artery cannulation in 538 adults patients. EMLA reduced pain associated with radial artery cannulation and improved the success rate of the procedure.