Meniscal screw fixation provides sufficient stability to prevent tears from gapping

1. Introduction 

Meniscal repair by suturing techniques or absorbable implants are getting increasingly popular to avoid total or partial resection of the meniscus. This prevalent use of implants is based on the knowledge that meniscal tears occurring in the peripheral vascularised part can heal (Albrecht-Olsen et al., 1999, Seil and Kohn, 2001). The preservation of the meniscus prevents joint instability and degeneration.

Implants for meniscal fixation aim at regaining the integrity of the torn parts of a meniscal tear to provide optimal healing conditions. To achieve this goal several implant designs such as screws, arrows, staples, sutures and others have been developed (Farng and Sherman, 2004). The mechanical stability of the implant–meniscus interface is limited to pull-out forces between 20N and 140N (Albrecht-Olsen et al., 1997, Barber and Herbert, 2000, Becker et al., 2001, Becker et al., 2002, Boenisch et al., 1999, Borden et al., 2003, Dervin et al., 1997, Dürselen et al., 2003b, Kohn and Siebert, 1989, Rimmer et al., 1995, Song and Lee, 1999, Zantop et al., 2004). However, the successful clinical use of implants (Bohnsack et al., 2003) with low pull-out forces suggests that the in situ forces in the implant–meniscus interface are low. Recent investigations on the gapping behavior of longitudinal meniscal tears showed a maximum gap width of only 1.6mm even under critical external loads such as valgus-external rotation (Dürselen et al., 2003a). However, this gap width is probably too broad for an undisturbed healing process, which justifies the need for a fixation implant. From this maximum gap width and the tensile stiffness of 3N/mm of the implant–meniscus interface obtained from pull-out tests (Dürselen et al., 2003b) it can be roughly estimated that forces less than 10N occur in meniscal fixation implants. One in vitro study on human cadaveric knee joints detected tensile forces in meniscal sutures of less than 10N (Kirsch et al., 1999) supporting this coarse estimation. However, it still lacks proof that gapping of longitudinal meniscal tears is sufficiently hampered, especially by meniscal fixation implants of low pull-out strength.

Therefore the purpose of this study was to investigate the gapping behavior of longitudinal meniscal tears that were fixed by implants of low pull-out force. It is hypothesized that meniscal repair with meniscal screws as an example for a device of low pull-out force significantly reduces tear gapping.

2. Methods 

2.2. Detailed procedures 

The special techniques used in this study have already been described previously (Dürselen et al., 2003a). However, the main aspects of methodology are repeated here to guarantee full understanding.

2.2.1. Specimen preparation 

The knee joint specimens were stored at −25°C and thawed at room temperature. Skin and muscles were removed. The joint capsule was detached from the tibia at its medial dorsal aspect. The medial tibial plateau was then removed and a translucent placeholder was prepared using casting resin (Polyesterglas Resin, Creartec, Weiler, Germany). The articular side of the translucent plateau copy had exactly the same shape as the chondral articular surface of the original tibial plateau. A translucent grid of 1mm grid length was glued onto the articular surface of the placeholder as a scale for the gap width measurement. A 3cm long tear was set from caudal in a distance of 5mm from the peripheral rim in the posterior horn of the medial meniscus (Fig. 1). To achieve a better optical contrast between the two sides of the tear the peripheral part was dyed using methylgreen (Dako Corporation, Carpinteria, CA, USA). Subsequently the placeholder was fixed in the bone cavity of the tibia with two Kirschner 2mm wires. Finally the capsule was reattached to the dorsal tibia with PDS 2-0 sutures. The bony ends of femur and tibia were potted in steel cylinders with methylmethacrylate (Technovit 3040, Kulzer, Werheim, Germany). After testing the knees without meniscus fixation implants the tears were fixed with three meniscal screws each (ClearFix Screw, Mitek Products, Westwood, MA). The screws were placed at a distance of 10mm and their position was marked with red ink (Fig. 1).

View Large Image Download to PowerPoint

Fig. 1. View into the cavity of the tibial plateau. The distal aspect of the medial meniscus is visible. t: tear line; a: peripheral part of the meniscus dyed with methylgreen; b: inner part of the meniscus; three arrows: marked positions of the three meniscal screws.

2.2.2. Knee joint motion and external loads 

The knee joints were mounted into a knee motion and loading simulator that allowed unconstrained joint motion (Dürselen et al., 1995, Dürselen et al., 2001) (Fig. 2). The knees were passively flexed and extended by an electric motor between 30° and 140° joint position, with 30° flexion as full extension for a porcine knee joint. Three flexion–extension cycles were applied for each test. A constant anterior shear force of 10N was applied to the tibia throughout all tests to keep the knee joint in a defined reproducible position. The flexion–extension cycles were performed with both 30N and 200N axial joint load, which was introduced in the direction of the tibial axis. Seven different load situations were tested during the flexion–extension cycles: without any external moment, internal and external tibial rotation moment (1Nm), varus and valgus moment (5Nm), combined valgus (5Nm) and external rotation (1Nm) moment, and combined varus (5Nm) and internal rotation (1Nm) moment. The applied moments were constant during the flexion–extension cycles. All tests were repeated after implantation of the meniscal screws.

View Large Image Download to PowerPoint

Fig. 2. Porcine knee joint mounted into the knee joint motion and loading simulator.

2.2.3. Data assessment 

An arthroscope (Karl Storz, Tuttlingen, Germany) with an attached video camera (Arthrex, Karlsfeld, Germany) was placed on the dorsal side of the tibia so that the lens was located underneath the translucent placeholder enabling the observation of the meniscus (Fig. 2). The end of a light pipe of a cold light source was aligned parallel to the arthroscope and delivered the necessary light. The video signal of the camera was recorded by a computer. The meniscal tears were observed during the flexion–extension cycles and still pictures were taken at the joint positions, where maximum gapping occurred (Fig. 3). By image analysis of the still pictures the gap width was calculated by relating it to the scale that was glued to the surface of the translucent placeholder. After fixation of the meniscal tears with screws the maximum gap width was measured at seven different locations along the tear length (Fig. 4). A reliability test revealed an accuracy of the gap width measurement of ±0.25mm.

View Large Image Download to PowerPoint

Fig. 3. Arthroscopic view from distal through the translucent placeholder: the arrows indicate the gap of the meniscal tear. Two Kirschner wires (R) fixed the placeholder to the bone.

View Large Image Download to PowerPoint

Fig. 4. Locations of tear gap measurements: (1) posterior horn 1; (2) screw 1; (3) posterior horn 2; (4) screw 2; (5) pars intermedia; (6) screw 3; (7) anterior horn.

2.2.4. Statistics 

All results were displayed as median values and maximum/minimum bars. The effect of screw fixation on the maximum gap width of the meniscal tears was statistically tested by a paired Wilcoxon-Signed-Rank-Test. The level of significance was set to P=0.05.

3. Results 

The fixation of the meniscal tears with three ClearFix Screws led to a significant decrease of the maximum gap that occurred during flexion–extension cycles (P<0.05). This was the case for all applied joint moments and for both 30N and 200N axial joint load (Table 1, Fig. 5). In case of a 30N axial load the mean maximum gaps were reduced by 43.1%. The minimum reduction was 32.1% under an external tibial rotation moment and the maximum reduction was found 61.5% under a varus moment. In case of a 200N axial joint load the mean maximum gaps were reduced by 47.5%. Here the screws minimally narrowed the maximum tear gaps by 37.4% under an internal tibial rotation moment and maximally by 59.2% under an external tibial rotation moment. The absolute maximum gap widths before fixation measured 1.4mm (median) under both an external tibial rotation moment and a varus moment in case of a 30N axial joint load. Under a 200N joint load maximum gap widths of 1.6mm were found under a combined valgus and external tibial rotation moment. After fixation with three ClearFix Screws the mean maximum gaps were reduced to 0.6–1.1mm throughout all load cases (P<0.05). Increasing the axial force from 30N to 200N did not result in broader gaps.

Table 1. Maximum gap (mm) occurring during flexion–extension motion for a 3cm longitudinal meniscal tear before and after fixation with three ClearFix Screws

		3cm tear without fixation		3cm tear fixed with three ClearFix Screws
		30N axial load (mm)	200N axial load (mm)	30N axial load (mm)	200N axial load (mm)
	No external moment	1.3	1.1	0.7	0.8
		(1.1, 1.5)	(1.0, 1.5)	(0.6, 1.0)	(0.0, 0.9)
	External rotation	1.4	1.4	1.1	0.6
		(1.2, 1.6)	(1.2, 1.9)	(0.7, 1.2)	(0.0, 1.1)
	Internal rotation	0.9	0.8	0.6	0.6
		(0.5, 1.1)	(0.7, 1.0)	(0.0, 0.8)	(0.0, 0.8)
	Valgus	1.2	1.2	0.7	0.7
		(0.7, 1.7)	(1.0, 1.7)	(0.3, 0.9)	(0.0, 1.0)
	Valgus+ext. rotation	1.3	1.6	0.8	0.8
		(1.1, 1.8)	(1.3, 1.9)	(0.5, 1.0)	(0.7, 0.9)
	Varus	1.4	1.3	0.8	0.8
		(1.1, 1.6)	(0.9, 1.7)	(0.0, 1.0)	(0.0, 1.1)
	Varus+int. rotation	1.3	1.2	0.8	0.9
		(1.0, 1.7)	(1.0, 1.5)	(0.0, 1.0)	(0.0, 1.0)

All values are medians (min, max), n=5.

View Large Image Download to PowerPoint

Fig. 5. Maximum gap width before and after fixation of a 3cm longitudinal meniscal tear with three ClearFix Screws under 30N and 200N axial load and seven different load cases (no: no external moments, IR: internal rotation moment, ER: external rotation moment, VLG: valgus moment, VAR: varus moment). Note that the values represent medians±max/min. The screw fixation led to a statistically significant reduction of maximum gap in all cases (P<0.05).

Regarding the gapping of the meniscal tears at different locations along the tear it was found that the medians of the tear widths at the location of the screws were zero in all cases (Table 2, position 2, 4, 6 and Fig. 6). However, in between the screws gapping still occurred up to 1.1mm. Similar results were found under 30N and 200N axial joint load.

Table 2. Maximum gapping (mm) of the tears fixed with three ClearFix Screws under 200N axial joint laod

	Moments	Location
		1	2	3	4	5	6	7
		Posterior 1	Screw 1	Posterior 2	Screw 2	Pars intermedia	Screw 3	Anterior
	No	0.0	0.0	0.0	0.0	0.0	0.0	0.5
		(0, 0.8)	(0, 0.9)	(0, 0)	(0, 0)	(0, 0.8)	(0, 0)	(0, 0.7)
	ER	0.3	0.0	0.0	0.0	0.0	0.0	0.5
		(0, 1.1)	(0, 0.7)	(0, 0)	(0, 0)	(0, 0.8)	(0, 0.5)	(0, 0.6)
	IR	0.6	0.0	0.0	0.0	0.0	0.0	0.0
		(0, 0.8)	(0, 0.6)	(0, 0)	(0, 0)	(0, 0.7)	(0, 0)	(0, 0)
	VLG	0.0	0.0	0.0	0.0	0.3	0.0	0.2
		(0, 0.8)	(0, 0.8)	(0, 0)	(0, 0)	(0, 1.0)	(0, 0)	(0, 0.7)
	VLG+ER	0.0	0.0	0.0	0.0	0.0	0.0	0.7
		(0, 0.8)	(0, 0.5)	(0, 0)	(0, 0.9)	(0, 0.8)	(0, 0)	(0, 0.8)
	VAR	0.0	0.0	0.0	0.0	0.5	0.0	0.0
		(0, 1.1)	(0, 0.7)	(0, 0)	(0, 0)	(0, 0.8)	(0, 0)	(0, 0.8)
	VAR+IR	0.0	0.0	0.0	0.0	0.0	0.0	0.0
		(0, 0.9)	(0, 0)	(0, 0.9)	(0, 0)	(0, 1.0)	(0, 0)	(0, 0.7)

All values are medians (min, max), n=5. Locations of the measurements according to Fig. 4. External moments: ER=external tibial rotation, IR=internal tibial rotation, VLG=Valgus, VAR=Varus.

View Large Image Download to PowerPoint

Fig. 6. Maximum gap width after fixation of a 3cm longitudinal meniscal tear with three ClearFix Screws at different locations along the meniscus under 200N axial joint load (median values).

4. Discussion 

The purpose of this study was to investigate the gapping behavior of longitudinal meniscal tears that were fixed by implants of low pull-out force. It was hypothesized that meniscal repair with the meniscal screw as an example for a device of low pull-out force significantly reduces tear gapping.

Human knee joint specimens of young donors are rarely available. Therefore the study was performed on porcine knees. It was assumed that the behavior under joint compression is similar in humans and pigs because menisci have the same mechanical functions (Jaspers et al., 1980). Joint compression displaces the meniscus in a peripheral direction leading to an increased circumferential stress. This effect is not species dependent. Tissue properties vary in modulus and permeability between species (Joshi et al., 1995). While permeability of porcine meniscal tissue is almost the same as in humans the compression modulus of porcine menisci was shown to be 20% greater compared to human tissue (Joshi et al., 1995). This could lead to slightly less gapping of a longitudinal tear in our porcine model than it would occur in a human knee. However, the orientation of collagen fibers, which is most important for the behavior of a meniscus under stress is similar in humans, pigs, and dogs (Aspden et al., 1985). A different friction coefficient between the meniscus and the translucent resin compared to articular cartilage could be an issue but we always observed continuous gliding of the meniscus without any arresting of the motion. This let us conclude that possible effects due to differences in friction coefficient can be neglected. The relatively low accuracy of the gap width measurement (±0.25mm) resulted from optical distortion caused by the irregularly shaped translucent copy of the tibial plateau. However, the differences in gap width between untreated and fixed menisci were large enough to state the following conclusions.

Permanent gapping occurred in the untreated longitudinal meniscal tears during flexion–extension cycles under most of the loading conditions tested in the current study. However, fixation with three ClearFix Screws led to a complete adaptation of the tear fragments in the majority of load cases. Seventy-two percent of the 490 gap measurements, which resulted from seven localization and all tested load conditions, revealed a gap width of zero. In these cases the gap width measurement was not prone to the measuring accuracy of ±0.25mm because a gap width of zero could unambiguously be detected. This clearly shows the efficiency of this implant. Only little previously published data is available on the effect of meniscal fixation systems on the gapping behavior of meniscal tears. Ganley et al. (2000) investigated the influence of axial joint load on 3cm meniscal tears in human knee joint specimens in three different flexion positions using computer tomography. To identify the gaps in the CT-scans the authors labeled the meniscus fragments with metal markers. Due to the limited accuracy of their visualization method Ganley et al. provided no information on the reduction of gap width by the suture repair. However, they did not find any significant difference in displacement of the gap boundaries between an axially loaded and unloaded joint for sutured and not sutured menisci. This complies with our finding that an axial joint load does not influence tear gapping. Compression loads seem to stabilize the position of the meniscus rather than to exert forces on it that would detach the tear rims from each other. This seems logical considering the function of the meniscus. Because of its wedge shape the meniscus is displaced to the periphery under compressive joint loads. Naturally, this applies for both the inner and outer fragment of the tear and for all flexion positions. Therefore it is not likely that joint compression broadens a longitudinal tear. Another study showed that the pull-out resistance of repair implants is increased under axial compression, which confirms our findings (Staerke et al., 2004). Ganley et al. (2000) speculated that separation of the tear edges could occur at flexion angles larger than 60°, which they could not test due to experimental restraints. In the current study no such effect was observed throughout the range of motion up to 140° flexion.

Application of external moments, in turn, showed an effect on the tear gap width. Internal tibial rotation created smaller gaps than other moments or combination of moments did. During internal rotation the medial plateau shifts posteriorly. Thus, the medial femoral condyle pushes the anterior horn of the meniscus anteriorly while releasing the posterior horn, and thereby decreasing the gapping. Moments acting on the knee joint lead to forward and backwards shifting of the menisci. This creates load situations in the meniscal tissue that differ from that, which is apparent during axial compression. Shifting the meniscus changes its shape and can induce shear loads, which probably leads to gapping of meniscal tears. Releasing the posterior horn by internal tibial rotation obviously reduces these shear loads and thus the gap width. However, meniscal repair with, e.g., three screws is able to reduce tear gapping even under critical joint loads like valgus-external rotation moments.

For meniscal repair it is of interest, which forces refixation implants need to resist. From the gap width found in this study and the stiffness of the implant–meniscus interface measured in some pull-out studies the forces acting in the implant can be estimated. Assuming stiffness values of 3.1–4.3N/mm (Becker et al., 2001, Dürselen et al., 2003b) and multiplying it with the maximum gap width of 0.9mm found after repair at the location of a ClearFix Screw results in maximum forces acting in the interface of 2.8–3.9N. This is far below the ultimate pull-out force of 14–32N that this implant exhibited in various in vitro tests (Asik and Sener, 2002, Barber and Herbert, 2000, Becker et al., 2001, Dürselen et al., 2003b). The in vitro study of Kirsch et al. (1999), who measured holding forces of meniscal sutures lower than 10N, confirms this estimation of forces acting in the implant–meniscus interface. Therefore it seems unlikely that longitudinal tears that were repaired with implants of relatively low pull-out force are in danger of pull-out failure even under joint loads like tibial rotation and/or varus–valgus moments. We therefore conclude that the parameter pull-out force is not a critical factor for meniscal repair devices.

An important issue in meniscal repair is the question under which condition meniscal tears can heal. Besides the undeniable fact that tears heal best in the peripheral vascularized red zone it is not known at which gap width healing still occurs. In the present study the distance between the meniscal screws was 10mm and the distance between the tear ends and the adjacent screws was approximately 5mm. Taking into consideration that tear gapping at the locations of the screws was zero in almost any case, the screws literally divided the 30mm tear in two tears of 10mm and two tears of 5mm length. Clinically it has been observed that longitudinal tears smaller than 10mm do not need surgical treatment due to the ability of spontaneous healing (Ihara et al., 1994). This supports the hypothesis that stabilizing a 30mm tear with three screws can provide conditions for successful healing.

From a pure biomechanical point of view the results of this study support the possibility of early rehabilitation under full weight bearing after meniscal repair with implants or sutures. Recent investigations in a goat model revealed worse results for all-inside techniques using resorbable implants compared to suturing (Miller et al., 2004). However, this was caused rather by cartilage damage due to implant interference with the articular surfaces than by instability of the implant–meniscus interface. Hence the aspect of possible cartilage damage may let the surgeon restrain from the uncritical use of any implant that is available on the market. In the current study we accidentally positioned one of the screws not deep enough into the meniscal tissue so that it stitched out of the meniscal surface by about 0.5mm. After the test, which included about 50 flexion–extension cycles, we opened the knee joint and found massive cartilage damage produced by scratching of the screw head over the femoral condyle. Of course this happened because of a wrong implantation. But looking at other implants one must realize that some of the available devices are designed in a way that they must stitch out of the meniscal tissue. For example, there are implants, whose heads lay between the meniscal surface and the cartilaginous femoral condyle (Meniscal Fastener (Mitek), Meniscal Stapler (Arthrotek), Biostinger (Linvatec), Meniscus Arrow (Bionx Implants)). Some of these devices are even made of polylactic acid with a long degradation time and no chance of a rapid material softening, which would reduce the danger of cartilage damage by the implant. High rates of chondral lesions have also been reported recently (Miller et al., 2004, Otte et al., 2002). This important aspect should be taken into account when selecting an implant for meniscal repair. Therefore implants without a head that is exposed to the articular surface can be recommended.

5. Conclusion 

The current study confirmed the hypothesis that meniscal screws, which show low pull-out strength, significantly reduce gapping of longitudinal tears. This applies for various joint loads including critical loads like valgus-external rotation. We therefore conclude that fixation implants of low pull-out strength like the ClearFix Screw are not in danger of failure in a normal rehabilitation regimen. However, meniscal implants are not designed for traumatic situations. Excessive loads can of course lead to failure of the refixed meniscal tear. In case of low patient compliance a knee rotation limiting brace prescribed for a limited time can provide more safety for the healing meniscus.