Segment skeletal defects that result from resection of a malignant bone neo
plasm commonly are reconstructed with large segment allografts. Excellent f
unctional results after these reconstructions and significant complications
have been reported. Although it is known that a common complication seen w
ith the use of allografts is allograft fracture, the factors associated wit
h allograft fracture are not entirely clear. In this study, the hypothesis
was examined that allograft reconstructions, which use internal fixation de
vices that penetrate the cortex of the allograft, are associated with an in
creased risk of fracture. Findings from large segment allograft reconstruct
ions in 74 patients with a minimum followup of 36 months were studied. Thes
e 74 patients include 35 patients whose outcomes were reported previously a
nd now have additional followup and 39 patients whose outcomes are being re
ported for the first time. Thirty-one of the 74 (42%) allografts fractured,
and the mean time to fracture was 26 months. When the fixation techniques
resulted in cortical penetration of the allograft, fractures occurred in 27
of the 43 (63%) allografts, and when fixation of the graft to the host bon
e required no cortical penetration, only four of the 31 (13%) allografts fr
actured. Fractures occurred in 12 of 15 (80%) tibial allografts and in only
two of 17 (12%) proximal femoral allografts; however, the anatomic site wa
s not statistically independent of method of fixation because 14 of the 15
tibial grafts had cortical penetration and 15 of the 17 proximal femoral gr
afts had no cortical penetration. The authors' analysis indicates that inte
rnal fixation devices that require cortical penetration are associated with
an increased risk of allograft fracture.