Background: Interfacial gaps between cement and femoral hip implants are a potential source of stem debounding and loosening. We used a CT-scan technique to measure the characteristics of these gaps for two different implant types and two different implantation protocols.

Methods: Using a third generation cementing technique, 22 plastic replicas of straight Charnley-Kerboul stems (Stratec) and 18 replicas of anatomic Lubinus SPII stems (Waldemar-Link) were implanted in 20 pairs of embalmed cadaver femora. In each pair of femora, the same stem type was used. However, at one side a stem with the same size as the broach was used (line-to-line), while at the other, we cemented a stem that was one size smaller (undersized). Based on a validated CT-scan measurement tool[1], we quantified the extent of interfacial gaps, determined their location and measured the cement thickness in areas with and without gaps. Differences in interfacial gaps between both implant types and both implantation techniques were analysed with a general linear model (GLM).

Results: In total 10308 connective CT-images were segmented and analysed. Interfacial gaps were found in every specimen (average: 6.43% of the stem surface), but their extent varied widely between specimens (SD: 8.99%; 0.15% – 31.69%). According to a GLM, the quantity of interfacial gaps did not vary significantly between implant types (Kerboul: 7.92±10.69%; Lubinus: 4.61±6.14%, p=0.246) and between implantation techniques (line-to-line: 7.73±10.24%; undersized: 5.13±7.57%, p=0.416). Irrespective of the implant type and the implantation protocol, flat regions of the stem surface showed significantly more interfacial gaps than corner regions (flat: 6.02%; corner: 4.36%; Chi-square: p< 0.001). Overall, more gaps were found at the anterior surface of the implant (anterior: 7.61%; medial: 4.99%; lateral: 4.46%; posterior: 4.43%; Chi-square: p< 0.001). For Kerboul and line-to-line stems, the extent of interfacial gaps increased consistently from the distal to the proximal stem region. This was not true for Lubinus and undersized stems. The thickness of the cement mantle showed no significant differences in areas with and without interfacial gaps (cement thickness: 3.50 mm in gap regions and 3.45 mm in regions without gaps, paired t-test: p=0.823).

Conclusions: Interfacial gaps between a femoral hip implant and cement are common and were found in all specimens. As these gaps can be extensive (up to 30 % of the stem surface), they could influence implant stability and survival. The interfacial gaps described here were probably due to air that was introduced along the implant during stem insertion. Further studies are needed to confirm that theory, to determine the consequences of these gaps and to search for methods to prevent them.

Interfacial defects between the cement mantle and a hip implant may arise from constrained shrinkage of the cement or from air introduced during insertion of the stem. Shrinkage-induced interfacial porosity consists of small pores randomly located around the stem, whereas introduced interfacial gaps are large, individual and less uniformly distributed areas of stem-cement separation. Using a validated CT-based technique, we investigated the extent, morphology and distribution of interfacial gaps for two types of stem, the Charnley-Kerboul and the Lubinus SPII, and for two techniques of implantation, line-to-line and undersized.

The interfacial gaps were variable and involved a mean of 6.43% (sd 8.99) of the surface of the stem. Neither the type of implant nor the technique of implantation had a significant effect on the regions of the gaps, which occurred more often over the flat areas of the implant than along the corners of the stems, and were more common proximally than distally for Charnley-Kerboul stems cemented line-to-line. Interfacial defects could have a major effect on the stability and survival of the implant.