To biomechanically evaluate the stability of a diaphyseal anchored, cementless stem in presence of a proximal periprosthetic femoral medial wall defect compared to the stability of the same stem in an intact femur. Twenty-two paired human cadaveric femora were pairwise assigned either to a fracture group, featuring a proximal medial wall defect involving 40% of the stems medial anchorage distance, or a control group with native specimens. The specimens were tested under a monotonically increasing cyclic axial loading protocol. Load, cycles, and multiples of the respective body weight at implant loosening was measured. Mean initial stiffness was 2243.9 ± 467.9 N/mm for the intact group and 2190.1 ± 474.8 N/mm for the fracture group. Mean load to loosening in the intact group was 3210.5 ± 1073.2 N and 2543.6 ± 576.4 N in the fracture group, with statistical significance. Mean cycles to loosening in the intact group were 27104.9 ± 10731.7 and 20431.5 ± 5763.7 in the fracture group, with statistical significance. Mean multiples of the resulting body weight at loosening in the intact group was 548.3 ± 158.5% and 441.4 ± 104% in the fracture group, with statistical significance. A medial wall defect involving 40% of the medial anchorage distance significantly decreases the axial stability of a diaphyseal anchored stem. However, mechanical failure occurred beyond physiological stress. At loosening rates of about 4 multiples of the body weight in the fracture group, a "safe zone" remains of a 0.5-fold body weight for maximum loads and twofold body weights for average loads.