Background The arterial access required during most invasive vascular procedures provides a common source of complications and morbidity. This problem has been made worse by recent trends in earlier ambulation and more aggressive antihemostatic drug regimens. Despite these trends, no randomized trials have been reported comparing the 3 most commonly used techniques in achieving hemostasis at the arterial puncture site. Methods A cohort of 400 patients undergoing catheterization laboratory procedures were randomly assigned to 1 of 3 groups of arterial compression: manual compression, mechanical clamp, and pneumatic compression device. Standard requirements of the trial included uniformity in initial compression times, patient instructions, nursing follow-up, and timing of ambulation as well as a structured interview and physical examination at 24 hours. Results Prolonged compression was required in 13% of the manual group, 20% of the clamp group, and 35% of the pneumatic group (P < .0001). In-lab bleeding was more common in the pneumatic group (3%, 4%, and 16%, respectively, P < .0001), as was the need for an alternate compression technique (1%, 1%, and 27%, P < .0001). The groups also differed in respect to mean hematoma size (3.9 cm², 7.8 cm², and 19.8 cm², P = .036) and level of discomfort during compression (1.9, 2.2, and 3.1 on a 1- to 10-point scale, P < .0001). Comparable findings were observed in the subgroup of patients eligible for outpatient procedures. Conclusions Use of the pneumatic compression device leads to longer compression times, greater discomfort, more bleeding, and larger hematomas. Differences between manual compression and the mechanical clamp were more subtle but tend to favor use of the manual technique. (Am Heart J 1999;138:1118-25.)