Two common justifications for orienting cervical screws in an angled direction is to increase pull-out strength and to allow use of longer screws. This concept is widely taught and has guided implant design. Fixed versus variable angle systems may offer strength advantages. The purpose of our study is to test the influence of screw orientation and plate design on the maximum screw pull-out load. Variable and fixed angle 4.0 x 15 mm and 4.0 x 13 mm self-tapping screws were used to affix a Medtronic Atlantis cervical plate to polyurethane foam bone samples (density 0.160/cm). This synthetic product is a model of osteoporotic cancellous bone. The fixed angle screws can only be placed at 12 degrees convergent to the midline and 12 degrees in the cephalad/caudal ("12 degrees up and in") direction. Three groups were tested: (1) all fixed angle screws, (2) variable angle, all screws 12 degrees up and in, (3) variable angle, all screws 90 degrees to the plate. Plate constructs were pulled off with an Instron DynaMight 8841 servohydrolic machine measuring for maximum screw pull-out force. There was no difference between group 1, fixed angle (288.4 +/- 37.7 N) (mean +/- SD) and 2, variable angle group (297.7 +/- 41.31 N P< or =0.73). There was a significant increase in maximum pull-out force to failure for the construct with all screws at 90 degrees (415.2+/-17.4 N) compared with all screws 12 degrees "up and in" (297.4 +/- 41.3 N, P< or =0.0016). Group 3 done with 13 mm screws, showed a trend toward better pull-out strength, compared to group 2 w/15 mm screws (345.2 +/- 20.5 vs. 297.4 +/- 41.3, P< or =0.06). In this plate pull-out model, screw orientation influences maximum force to failure. When all 4 screws are 90 degrees to the plate the construct has the greatest ability to resist pullout. Fixed angle designs show no advantage over variable angle. These findings are contrary to current teaching.