Introduction: Scant amount of information is available on mechanical properties of composite specimens of old and new cement. In previous studies evaluating this, old samples were only few days old, unlike clinical situation, where the old cement is a few years old.

We evaluated short-term mechanical properties of composite specimens and compared these with new uniform specimens.

Material and Methods: Uniform and composite specimens were fabricated and were tested for bending, tensile and shear strength. Seventeen beams and eight cylindrical specimens fabricated earlier (median age 11.8 years) using same moulds were available to form composite specimens. Specimens were stored in saline at 37 °C for 6 weeks before testing.

Results: Bending tests: Load and bending stress for new specimen was 82.9N and 49.5MPa as compared with 74.3N and 40.3MPa for composite specimens. 4 composite specimens failed though old cement, 3 through junction and 1 through new cement. There was no statistical difference in maximum load (p, 0.3) or stress (P, 0.06) between uniform and composite specimens.

Tensile tests: Load and tensile stress for new specimen was 941.5N and 29.5MPa as compared with 726.9N and 22.1MPa. There was difference in the load and stress of uniform specimens as compared with composite specimens.

Shear tests: Load and shear stress for new specimen was 2692.9N and 34.5MPa as compared with 2009.9N and 25.3MPa. There was significant difference in load as well as stress in uniform specimens as compared with composite specimens.

Discussion: This study demonstrates that composite specimens fail at 89.6% of bending load, 77.2% of tensile and 74.6% of shear load as compared with uniform new cement specimens. Of more importance is the fact that only four of these composite specimens (23.5%) failed at the junction and the rest thirteen failed either through old cement (64.7%) or through new cement (11.8%).

Controversy exists with regard to the thickness of cement mantles that are necessary around the femoral components of cemented total hip arthroplasties. Conventional teaching, based on bench-top or computor models and theoretical analyses, as well as post-mortem & follow-up studies, suggests that the cement mantle should be complete and not less than 2–3mm in thickness. Mantles that are less than this are held to be at risk from mechanical failure in the long term; if they are incomplete, focal lysis may occur and progress to aseptic loosening.

However, long term experience with a number of French cemented femoral components suggests that these conventions may be erroneous. These French femoral components include the Charnley-Kerboull (stainless steel) and the Ceraver Osteal (Ti6Al4V) stems, in both of which the underlying design principle is that the stem should completely fill the femoral canal, the cement then being used purely to fill the gaps. Such a design philosophy implies that the cement mantles will be very thin, and since both of these stems are straight and the femoral medullary canal is not, the mantles may not only be thin, but also in places incomplete.

Conventional teaching would suggest that any stem utilising mantles of this type would fail from a combination of focal lysis and cement fracture. Yet the long term results of both of these stems have been outstandingly good, with extremely low levels of aseptic loosening and endosteal lysis, irrespective of the bearing combinations being used. Both these stems have a surface finish of Ra < 0.1 microns. A third French design, the Fare stem, manufactured from Ti6Al4V and based on the same principles, was associated with bad results when manufactured with a rough (> 1.5 microns) surface, and appreciably better results after the surface roughness was changed to < 0.1 microns.

These findings, that constitute the ‘French Paradox’, have profound implications for the mechanical behaviour of cement in the femur and for the mechanisms that underlie stem failure from loosening.