Background: The optimum surface roughness of cemented femoral stems used for total hip replacement is a subject of controversy. While rougher surfaces provide stronger cement adhesion, it has been hypothesized that polished, tapered, noncollared stems settle into the cement mantle, providing improved stability. However, the effects of surface finish on the stability of straight, cemented stems tapered only in the coronal plane are not known.

Methods: Using composite model femora, we assessed the initial stability of a straight, cemented femoral stem as a function of surface roughness, the presence or absence of a collar, stem size, and the resultant cement thickness under simulated walking and stair-climbing loads. Otherwise identical stems were manufactured with polished or rough surfaces, with or without a collar, in two different sizes. We isolated these three variables and compared their relative contributions to the motion at the stem-cement interface throughout cyclic loading. We defined three indicators of stability: per-cycle motion, rate of migration, and final migration.

Results: Surface roughness had a greater influence on per-cycle motions than did the presence or absence of a collar or cement thickness. Specifically, in the medial-lateral direction, per-cycle motion of polished stems was 43 µm greater than that of rough stems (p < 0.01). None of the per-cycle motions decreased over the 77,000 load cycles. In contrast, with all stems, the rate of migration decreased over the course of cyclic loading, but the rate of migration of the polished stems was greater than that of the rough stems. Final migrations of the stems over the course of loading were generally distal, medial, and into retroversion. Compared with rough stems, polished stems had 8 to 18 µm more axial migration (p < 0.001), 48 µm more anterior-posterior migration (p < 0.001), and 0.4° more rotational migration (p = 0.01).

Conclusions and Clinical Relevance: The results indicated that, for cemented, straight femoral stems tapered only in the coronal plane, a rough surface offers the advantage of less per-cycle motion. These results may apply to widely used cemented stem designs based on the profile of the original Charnley femoral component, which has approximately parallel anterior and posterior aspects.