Abstract
There is little information available regarding mechanical aspects of cemented implant loosening and the initiation and development of cement damage. Previous studies have come to a variety of conclusions about the development of cement damage and the relative importance of voids, the stem/cement interface and the cement/bone interface.
Cement micro-cracks and stem/bone micro-motions were quantified for Charnley Cobra stems under “stair-climbing” loads. Six stem/cement/femur constructs were subjected to loads based on estimated body weight for 300 kcycles at 2 Hz; two additional constructs were not loaded. Transverse sections were cut at 10 mm intervals, stained with a fluorescent dye penetrant and examined using epifluorescence stereo-microscopy.
Despite the aggressive loading, all stem/bone micro-motions were small and all stems were “well fixed” at the end of the loading. The only consistent micro-motion was internal rotation but this did not significantly correlate with cement damage (p=0.9). For cyclically loaded constructs mean crack length was 0.49 mm (SD 0.37, range 0.07 to 4.42) and for non-loaded controls mean crack length was 0.25 mm (SD 0.18, range 0.03 to 1.16). Total crack length (46–281 mm) was significantly correlated (R2=0.819, p=0.002) with femoral head load (0 & 1.0–1.8 kN). There was a significantly (p< 0.05) greater proportion of damage at the cement/bone interface (66% ± 9) than at the stem/cement interface (28% ± 8). A small fraction of micro-cracks involved voids (5% ± 5), but these were significantly (p< 0.001) less than the cement/ bone fraction. Micro-cracks in unloaded specimens were evenly distributed axially (R2=0.0002, p=0.95) consistent with the theory that they were induced by cement shrinkage. ANCOVA for total crack density using head load and axial position as covariates showed a significant positive effect for head load (p< 0.0001) and a significant interaction between head load and axial position (p=0.001); under load, micro-crack density increased proximally, and this effect was stronger with increasing head load.
The abstracts were prepared by Nico Verdoschot. Correspondence should be addressed to him at Orthopaedic Research Laboratory, Universitair Medisch Centrum, Orthopaedie / CSS1, Huispost 800, Postbus 9101, 6500 HB Nijmegen, Th. Craanenlaan 7, 6525 GH Nijmegen, The Netherlands.
