Suture anchors of various designs are gaining acceptance for open and arthroscopic procedures. The rapid proliferation of these devices challenges those using them to apply objective criteria for device selection. Comparative data on implant security in different settings, modes of failure, and ultimate failure strengths is lacking. This study was undertaken to independently develop such data for an objective comparison of the suture anchors currently available. Using a fresh never-frozen porcine femur model, 10 samples of each of the 14 different anchors tested were inserted into each of the three different test areas; diaphyseal cortex (usually 3- to 4-mm thick), metaphyseal cortex (usually 1- to 2-mm thick), and a cancellous bone "trough". The suture anchors were threaded with 0.018-inch stainless steel wire or, for anchors requiring a more flexible suture, 0.018-inch stainless steel 1 x 7 wire braid. Tensile stress parallel to the axis of insertion was applied at a rate of 12.5 mm/second by an Instron 1321 (Instron Corp, City, State) until failure. Average failure strength was calculated for each anchor at each test area. The anchors tested were the Mitek G2, Mitek G3, Mitek G4 (Mitek Surgical Products, Norwood, MA), Linvatec Revo screw (Linvatec, Largo, FL), Acufex TAG Wedge, Acufex TAG Rod 2 (Acufex Microsurgical, Mansfield, MA), Statak models 1.5, 2.5, 3.5, 5.0, and 5.2 (Zimmer, Warsaw, IN), Arthrex ESP (Arthrex Inc., Naples, FL), Arthrotek Harpoon, and Arthrotek LactoSorb (Arthrotek, Warsaw, IN). The average failure strength of each of these anchors in the diaphyseal cortex, metaphyseal cortex, and cancellous bone is reported.(ABSTRACT TRUNCATED AT 250 WORDS)