Abstract
Bone-cement shrinkage has never been quantified in a stem/cement/femur construct. We observed gaps around femoral stems in transverse sections of stem/cement/femur constructs; a greater proportion of stem/cement (s/c) interface gaps were found around grit blasted sections of stems than satin finished sections. If s/c gap formation were a shrinkage artifact then mantles with few s/c interface gaps must manifest shrinkage elsewhere, at the c/b interface or voids. ‘Mould-gaps’ at a c/b interface have been described previously but not quantified. We analysed the area of gaps at both interfaces. We hypothesised 1) Total gap area was the same for all transverse sections. 2) Satin sections had greater c/b gap areas than grit sections.
Transverse sections of stem/cement/femur constructs were processed to highlight gap areas. Five stems had a satin finish (Ra 0.75 um) and five were proximally grit-blasted (Ra 5.3 um). Sections were coated with matt black spray paint and then polished with emery paper. This process left all interface gaps and voids filled with black paint, which facilitated digital imaging. Gaps were visually identified and measured using Image-Pro. Gap areas for each transverse section were normalised by the area of cement in that section.
Gaps were not evenly distributed; there was obvious localisation at both interfaces. No significant difference found between surface finishes in total gap area ((satin 3.1% ± 1.4):(grit 3.4% ± 1.5)), supporting our first hypothesis. S/c gap areas were significantly greater around grit blasted sections ((satin 0.1% ± 0.4):(grit 1.9% ± 1.7) p< 0.0001). C/b gap areas were significantly greater around satin finished sections ((satin 2.3% ± 1.3):(grit 1.0% ± 0.9) p< 0.0001), supporting our second hypothesis.
Shrinkage can localise into large interface gaps; which must lead to stress concentrations. C/b gaps are potentially benign as they can fill with bone. Cement failure at points of s/c contact would generate debris hindering bone formation.
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.
