This is a retrospective analysis of 2 methods of fixation of the femoral component in 86 consecutive revision arthroplasties, for which all clinical and radiographic data were recorded prospectively. There were 56 cemented revisions using precoated femoral components followed for 2 to 8 years (mean, 4 years) and 30 uncemented, proximally porous-coated femoral revisions followed for 2 to 6 years (mean, 4 years). Of the 56 cemented hips, 31 (55%) had a good or excellent clinical result. Rerevision has been performed in 10 hips, and revision of 3 loose femoral components is pending (23%). Radiographic review of 56 femoral components showed that 16 (29%) had probable or definite loosening. Of 30 cementless hips, 24 (80%) had a good or excellent clinical result. There was radiographic bone ingrowth in 22 of 30 hips (73%). Seven hips (23%) had nonprogressive subsidence, and 3 hips (10%) had progressive subsidence or loosening. Rerevision has been performed in only 2 hips (7%). The high rate of loosening (29%) and rerevision (23%) at a mean follow-up of only 4 years suggests that a precoated femoral component may place increased stress at the already damaged bone-cement interface.