Article focus

• To identify the optimal interference fit of the tibial component of the cementless Oxford Unicompartmental Knee replacement;

• To assess the relationship between the push-in force (the force required to seat the implant) and the pull-out force (the force required to remove the component) of a cementless implant.

Key messages

• The push-in force progressively increases with increasing interference, while the pull-out force was related in a non-linear fashion to interference;

• A reduced interference fit may both improve primary fixation and decrease the risk of fracture.

Strengths and limitations

• First study measuring the relationship between interference and the push-in and pull-out forces of the tibial component of the cementless OUKR;

• This is a plastic bone study and the results need to be confirmed on cadaveric studies or animal models.

Introduction

Interference fit (or press fit) is the fixation between two parts that is generated by friction after they have been pushed together. For orthopaedic implants, interference is defined as the difference between the size of the implant and the size of the bone cavity into which the implant is pressed, and is measured in millimetres. The optimal interference fit for orthopaedic implants is uncertain, however, a previous biomechanical study has suggested that the optimal interference for a peg in cancellous bone should range between 0.65 mm and 0.8 mm.¹ The optimal interference is likely to be influenced by implant design, friction coefficient of the surfaces, and the skeletal segment in which the implant is used. Insufficient interference would compromise the primary stability of the implant and might cause implant migration, loosening and failure. In contrast, excessive interference increases the assembly load required to implant the component which in turn generates stress in the bone, trabecular damage and increases the risk of fractures.^2,3

In 2004, a cementless version of the Oxford Unicompartmental Knee Replacement (OUKR, Zimmer Biomet, Swindon, UK) was introduced with the aim of reducing the incidence of radiolucent lines and improving the survival of the implant.^4,5 This aim has been achieved with the New Zealand Joint Registry having a ten-year survival of about 96%, which is similar to that of total knee arthroplasty.⁶ The primary stability of the cementless OUKR relies on interference fit. The optimal interference for this specific implant is still unclear. While a radiostereometric analysis study has confirmed the reliability of cementless fixation of the OUKR,⁷ a cadaver study has suggested that lower loads are required to cause a fracture in bone implanted with cementless compared with cemented OUKR tibial components.⁸ As fractures are usually internally fixed and the components are not revised National Registries do not provide useful information on fracture rates. However, anecdotally, some surgeons have suggested that there is a higher incidence of perioperative tibial plateau fractures with cementless fixation. While the current interference of the cementless tibial component is sufficient to ensure reliable stability, it may be unnecessarily high and increase the risk of perioperative tibial plateau fracture.

The aim of this plastic bone study was to identify the optimal interference fit of the cementless tibial component of the OUKR. The force required to seat the component (push-in force), which is related to the risk of fracture, and the force required to remove the component (pull-out force), which is related to primary stability, were measured.

Materials and Methods

A computer numerical control (CNC) machine (Hurco VM1 High Wycombe, United Kingdom) was used to cut slots in blocks of solid polyurethane foam (20 pounds per cubic foot (PCF), (Sawbones, Malmo Sweden), which is the most widely used analogue for trabecular bone.^9,10 A custom-made boss was welded to the upper surface of 12 standard cementless size C Oxford UKR components by the company producing the implant (Zimmer Biomet UK Ltd, Swindon, United Kingdom), to securely fasten the components to the materials testing machine. A Dartec materials testing machine HC10 (Zwick Ltd, Herefordshire, United Kingdom) was used to implant and extract the tibial components, at 0.01 mm/s, whilst measuring the push-in and pull-out forces.

Six components were used and each was implanted in 12 new slots with a range of interference of 0.1 mm to 1.9 mm (slot width ranging from 3.8 mm to 2.0 mm). The length and depth of the slots were constant (39 mm and 14 mm, respectively) and exceeded the dimensions of the keel, to exclude the influence of the impingement on the bottom or with the anterior or posterior ends of the slot. Each component was tested starting from the minimal interference to the maximal interference, to reduce the risk of damaging the porous coating on the components. The experiment was repeated with six further tibial components on a cellular polyurethane foam block (15 PCF, Sawbones, Malmo Sweden), which is a more porous analogue for trabecular bone.

The thickness of the keel of 12 standard, size ‘C’ cementless tibial components was measured by two independent observers (SC and SM) with a digital calliper (Sealey, Bury St. Edmunds, United Kingdom). Ten measurements were performed in ten different positions on the keel following a standardized sequence (Fig. 1), to consider the irregularity and non-homogeneous distribution of the coating. The thickness of each keel was estimated by the mean of the ten measurements.

Fig. 1

The interference is calculated as the difference between the size of the implant (x) and the size of the bone cavity into which it is pressed (y) (interference = x–y).

Interference was defined as the difference in millimetres between the mean width of the keel and the nominal width of the tibial slot (Fig. 1).

Results

Push-in and pull-out measurement

In the solid polyurethane blocks (20 PCF), the push-in force progressively increased with increasing interference (Fig. 2). With the maximum interference tested (1.9 mm) the push in force was 1852 N (Standard deviation (sd) 106). The pull-out force was related in a non-linear fashion to interference. The highest pull-out force (380 N, sd 24) was obtained with an interference of 0.7 mm, which required a push-in force of 925 N (sd 65). The pull-out force decreased with higher interference.

Fig. 2

Mean push-in (PI) and pull-out (PO) value with varying interference for the solid testing material (20 PCF) (95% confidence intervals).

For an interference ranging from 0.9 mm and 1.3 mm, which is the current clinical range based on the design of the implant and surgical instruments, the push-in force varied between 1254 N and 1456 N and the pull-out force between 167 N and 317 N. In comparison, interferences of 0.7 mm, 0.6 mm or 0.5 mm had push-in forces that were reduced by up to 45% (p < 0.001) and comparable or superior pull-out forces (p > 0.05). The pull-out force progressively decreased with interference of 0.4 mm and less.

The one-way ANOVA revealed a significant difference for different values of interference both for the push-in and the pull-out measurements (p < 0.001). The results of the post hoc analysis are reported in Table II.

Table II.

Results of the Bonferroni post hoc analysis. Significance was set as < 0.01. There was a significant difference in push-in force between the current interference and what we are investigating to be the optimal range of interference. There was no difference in the pull-out between the 0.5, 0.6 and 0.7 slot. In contrast, the interference values in the ‘optimal range’ showed a superior pull-out force compared with an interference of 1.1 mm and 1.3 mm

Push-in force
			Current range
	Interference	0.9 mm	1.1 mm	1.3 mm
	0.5 mm	< 0.01	< 0.01	< 0.01
Optimal range	0.6 mm	< 0.01	< 0.01	< 0.01
	0.7 mm	< 0.01	< 0.01	< 0.01
Pull-out force
			Current range
	Interference	0.9 mm	1.1 mm	1.3 mm
	0.5 mm	1	< 0.01	< 0.01
Optimal range	0.6 mm	1	< 0.01	< 0.01
	0.7 mm	0.1	< 0.01	< 0.01

With the more porous bone analogue, although the forces were lower, the relationship between interference and push in and pull out force were similar (Fig. 3). Again, the pull-out force decreased with higher interference. The pull-out force presented almost constant values for an interference ranging from 0.5 mm to 1.9 mm.

Fig. 3

Mean push-in (PI) and pull-out (PO) value with varying interference for the cellular testing material (15 PCF) (95% confidence intervals).

With the maximum interference tested (1.9 mm) the push in force was 1123 N (sd 242). The highest pull-out force was around 200 N, obtained with for all the value of interference ranging from 0.7 mm to 1.3 mm.

For the current range of interference, the push-in force varied between 595 N and 835 N and the pull-out force between 196 and 199 N. Compared with this, interferences of 0.7 mm, 0.6 mm or 0.5 mm had push-in forces that were reduced by up to 47% (p < 0.001) and comparable pull-out forces (p > 0.05). Also in the cellular block, the pull-out force progressively decreased with interference of 0.4 mm and less. The results of the post hoc analysis are reported in Table III.

Table III.

Results of the Bonferroni post hoc analysis. Significance was set as < 0.01. There was a significant difference in push-in force between the current interference and the ‘optimal’ range of interference. In contrast, there was no difference in the pull-out forces

Push-in force
			Current range
	Interference	0.9 mm	1.1 mm	1.3 mm
	0.5 mm	< 0.01	< 0.01	< 0.01
Optimal range	0.6 mm	0.11	< 0.01	< 0.01
	0.7 mm	1	< 0.01	< 0.01
Pull-out force
			Current range
	Interference	0.9 mm	1.1 mm	1.3 mm
	0.5 mm	1	0.96	1
Optimal range	0.6 mm	1	1	1
	0.7 mm	1	1	1

Discussion

This study suggests that decreasing the interference fit of the keel of the cementless OUKR would significantly decrease the push-in force required to seat the component, without diminishing the pull-out force and therefore ensuring the same level of primary stability.

There was an almost linear correlation between the nominal interference and the push-in force; the higher the interference the higher the push-in force. In contrast, there was a non-linear correlation between interference and the pull-out force. On the 20 PCF solid polyurethane blocks, the pull-out force was approximately constant for interference ranging from 1.1 mm to 1.9 mm, with a median value of 167 N. It increased for values of interference ranging from 0.5 mm to 0.9 mm, reaching the peak value of 380 N for a nominal interference of 0.7 mm, and then progressively decreased for values of interference below 0.4 mm.

The interference that is achieved with the current implant design ranges from 0.9 mm to 1.3 mm. The variability in interference is caused by the irregularity in the geometries of both the porous coating and the bone slot. According to the results of this study, the optimal nominal interference should be between 0.5 mm and 0.7 mm. Compared with the current range (0.9 mm to 1.3 mm), this ‘Optimal range’ produces an equal or superior pull-out force, with a reduction in the push in force up to 45%. A very similar pattern of results was obtained using the cellular polyurethane blocks, which has different mechanical properties and friction coefficient.

The results are consistent with those reported by Berahamani et al¹ who demonstrated, in a cadaveric study assessing the initial stability of pegs, that beyond a certain interference the fixation strength does not increase. These findings were explained by excessive stresses during insertion that would cause damage and permanent deformation to the bone and limit additional stability.^1,11 The permanent deformation leads to an actual interference fit that is less than the nominal interference.^2,12 The permanent deformation was higher for higher nominal interference. The difference between the nominal interference and the effective interference that resulted was minimal for an interference of 0.3 mm, and increased with increased interference. The actual interference was around 30% of nominal interference for an interference of 0.9 mm.²

During the insertion of an implant with an interference fit into bone, forces normal to the bone surfaces are generated. These forces will tend to split the bone and can cause fractures.^2,13 A biomechanical study by Bishop et al¹² demonstrated that whilst the push-in force is significantly higher than the pull-out force, the normal forces on the bone are roughly equivalent to the push-in force.

The results of this study show that a reduced nominal interference would require a lower push-in force to seat the tibial component of the cementless OUKR, yet would achieve a similar or higher pull-out force. Potentially the push-in force could be nearly halved so the normal forces that would tend to split the bone would be nearly halved. In addition, there would be less damage to trabecular bone and the impaction required to implant the component would be reduced. By limiting these problems, a lower nominal interference could reduce the risk of perioperative tibial condyle fracture, without compromising primary fixation.^14,15

This is the first study to identify the current and the optimal range of interference for the tibial component of cementless OUKR and assess its correlation with push-in and pull-out forces. Despite its clinical relevance, there is a lack of published data on this topic. The study has some limitations. First, tests were carried out on artificial bone. Even though the Sawbones blocks used are meant to have properties similar to that of cancellous bone, their friction coefficient will inevitably be different from human bone. However, a second testing material, having a different modulus and structure were tested and found to have similar results. Future research could compare the optimal interference with current interference in animal or cadaveric bone, although neither of these provides a good representation of osteoarthritic bone.

The implants were tested in dry conditions, while bone is covered by fluid and blood. However, a previous study comparing dry and wet Sawbones blocks as experimental models for the testing of press-fit implants has reported that the dry or wet condition has limited effect on the dynamic friction coefficient of plasma sprayed surfaces.¹⁶

In a full assessment of primary stability, micromotion at the bone-implant interface must be assessed. This was not carried out and represents another limitation of the current study. This aspect could be investigated in future studies. Finally, we could not formally correlate the risk of fracture with the nominal interference.

In conclusion, this study suggests that decreasing the interference fit of the keel of the cementless OUKR reduces the push-in force without decreasing the pull-out force. A lower interference fit should therefore decrease the risk of tibial plateau fracture without compromising fixation.