The goals of this study were to develop a new intramedullary, rotation-stable locking device and evaluate it biomechanically and in vivo for maintenance of a critical size osteotomy gap in a model of conscious pseudarthrosis. In standardized osteotomized rat femora (5 mm osteotomy gap) two different rotation-and axial-stable locking devices (group pS þ cS) were tested in vitro with respect to biomechanics and compared to a control group without an additional locking device (K; n ¼ 6 for each group). For in vivo studies, 27 male Sprague Dawley rats (250-300 g) underwent a femoral defect osteotomy of critical size and were stabilized by one of the three methods (n ¼ 9 for each group). All groups were examined radiologically postoperatively, after 14 days, and after 12 weeks. In vitro testing revealed higher compression and torsional rigidities for the two locking devices (p < 0.05) compared to the control group (compression rigidity: pS ¼ 103.6 AE 13.2; cS ¼ 91.3 AE 10.9; K ¼ 52.8 AE 8.4 N/mm; torsional rigidity: pS ¼ 5.9 AE 0.9; cS ¼ 4.3 AE 1.4; K ¼ 0.4 AE 0.1 Nmm/8). In vivo, group K and pS exhibited up to two thirds wire dislocation and reduction of the osteotomy gap, while dislocation was less frequent in the cS group. Thus, the locking device with compression of the wire showed advantages in rotational and axial stability for a critically sized defect, though the osteotomy gap could not be maintained in all cases over the 12-week period. Nevertheless, our data corroborate the necessity of an internal fixation device with sufficient axial and rotational stability. ß