Introduction

The optimal treatment of the injured posterior cruciate ligament (PCL) has not yet been defined. Although there is consensus in the literature regarding the need for surgical treatment in the presence of other knee ligament injuries,¹,² controversy remains in isolated lesions of the PCL. Some authors advocate the conservative treatment in these cases.³,⁴ Others, however, emphasize that the knee with a PCL injury develops degenerative changes that worsen over time.⁵,⁶ The better understanding of the anatomy and biomechanics of the ligament enabled improvements in the results of PCL reconstructions, but to date no ideal technique has been described.⁷,⁸

Based on anatomical and biomechanical studies, some authors suggest that reconstruction with two bundles and two femoral tunnels better mimics the anatomy and function of the normal knee.¹,⁹ However, laboratory tests that show the superiority of these double-bundle reconstructions¹⁰,¹¹ are biased by the lack of uniformity in the thickness of the grafts used in the two compared groups. In these studies, the authors use a larger volume of graft in the two-bundle reconstruction than in the technique with one bundle. When using grafts of the same thickness in both groups, other authors observed no significant difference between the techniques.¹²,¹³ Upon this, the question that motivated the present work was formulated: the successful restoration of knee laxity achieved with the double-bundle technique is due to the construction of the second bundle or to the greater volume of graft? The objective of the present study was, therefore, to evaluate the biomechanical effect of graft thickness compared with the double-bundle technique in PCL reconstruction in human cadaveric knees.

Methods

The present study was approved by the Ethics in Research Committee of the institution responsible for the study (procedure number 391/02).

The knees used were obtained from nine fresh cadavers from the necropsy service of a large university hospital in Brazil. The average age of the cadavers was 52 years, ranging from 43 to 70 years. A total of seven cadavers were male.

At the necropsy service, a medial arthrotomy along with a vertical posterior arthrotomy were performed for joint inspection, to exclude any specimen presenting ligament injuries or degenerative joint disease. Once the specimens were deemed eligible for the study, the arthrotomies were closed with a 2–0 nylon. The knee was then removed from the cadaver and frozen at −15º C.

Testing Protocol

The mechanical tests were performed using the Kratos K5002 (Kratos Equipamentos Industriais, Cotia, SP, Brazil) electromechanical testing machine with an electronic load cell of 100 kgf connected to a computer equipped with an ADS 2000 (Lynx Tecnologia Eletrônica Ltda., São Paulo, SP, Brazil) data acquisition system. To begin testing, a metal clamp was fixated to the metaphysis and diaphysis of the femur, and another one, to the proximal tibial diaphysis. The clamps remained in this position, firmly fixated, during every test, and even during the surgical procedures. The knee with the clamps was fixated to the testing machine (Fig. 1), which was connected to the computer. For each evaluation, the knee was tested in extension, and at 30º, 60º and 90º of flexion.

Fig. 1 Knee positioned at 90º of flexion, fixated to the testing machine. The tibia remained horizontal, with the anterior margin facing the ground. The machine performed the elevation or descent of the femur in relation to the tibia, which corresponded, respectively, to the movement of the anterior and posterior drawers. 

During the tests, the femur movements in the upper and lower direction in relation to the tibia were made using the Kratos machine at a constant speed of 20 mm/min, corresponding to anterior and posterior drawers. The load point was located at the distal femoral metaphysis. The results of the tests (force applied to the tibia and its translation since the beginning of the test) were immediately transferred to the computer. For each evaluation condition, the specimen was submitted to three consecutive cycles of force application by the testing machine. Initially, the load was applied in the direction of the femur to achieve an anterior tibial translation of 2 mm. Then, it was applied in the reverse direction (posterior drawer) with a force equivalent to 134 N, as described by some authors.¹¹,¹⁵,¹⁶

The laxity of each knee was initially tested with intact (INT) ligaments. Then, the PCL was removed, and each knee was successively subjected to three different types of reconstruction, which are described in detail in the next section. The laxity was tested after each reconstruction. Finally, tests were performed on the knee without the original PCL and without reconstruction (NoPCL) (Fig. 2).

Fig. 2 Sequence of testing conditions. 

Surgical Techniques

Single-Bundle Reconstruction with One 10-mm Quadriceps Tendon (SB)

After the test with the INT PCL, the knee was removed from the testing machine. All surgical techniques were performed through a medial parapatellar arthrotomy and also by a small posterior longitudinal arthrotomy. Thus, the PCL was completely removed, including the meniscofemoral ligaments, when present, but sparing the other knee ligaments. The tibial tunnel was made with a 10-mm diameter, from the antero-medial tibia surface to the center of the removed PCL tibial insertion. The 10-mm anterolateral (AL) femoral tunnel was drilled centered 7 mm from the cartilage edge at the 1-o'clock position, in a direction parallel to the notch roof. The quadriceps tendon graft was inserted into the tibial and femoral tunnels, with the bone block located in the femoral tunnel. It was then fixated to the femur by tying it around a 4.5-mm cortical screw with a washer.

After suturing the articular capsule, the knee was repositioned in the testing machine at 90º of flexion, and subjected to an anterior drawer of 134 N. At this moment, the graft was tensioned (digital tensioner, LIM-41, São Paulo, SP, Brazil) (Fig. 3). When the desired tension (88 N) was achieved, the graft was fixated to the tibia, locking the two threads of polyester no. 5 between two metallic platelets and a screw (Fig. 4).

Fig. 3 Graft tensioning monitored by the dynamometer. 

Fig. 4 Tibial fixation device. The polyester threads from the graft were locked between two platelets. 

Double-Bundle Reconstruction with One 10-mm Quadriceps Tendon and a 7-mm Semitendinosus Tendon (DB)

After testing the single-bundle reconstruction (SB), the knee was removed from the testing machine. The sutures were removed from the arthrotomies, as was the graft from the tibial tunnel. This tunnel was extended in thickness with a 12-mm diameter drill.

The graft was kept in the AL femoral tunnel while the second (posteromedial, PM) tunnel was made. The PM tunnel was centered at the 2:30-o'clock position, 14 mm from the cartilage edge (distal-proximal direction) and 9 mm posteriorly to the center of the AL tunnel (anterior-posterior direction).

This PM femoral tunnel was made with a 7-mm diameter drill. The doubled semitendinosus tendon graft was inserted into the tibial and PM femoral tunnels and fixated to the femur by tying it around a second cortical screw with a washer (4.5 mm in diameter). Then, the quadriceps graft was again inserted into the tibial tunnel in a position anterior to the semitendinosus tendon.

After the arthrotomies were sutured, the knee was repositioned in the testing machine at 90º of flexion, and subjected to an anterior drawer of 134 N. The quadriceps graft was fixated with 88 N of tension. The knee was returned to position 0 and positioned at 0º of flexion. It was again subjected to a 134-N anterior drawer, and the semitendinosus graft was tensioned to 67 N and fixated with a second tibial device (similar to the one used for the quadriceps, Figs. 3 and 4). The knee was then returned to position 0 and submitted to laxity tests at 0º, 30º, 60º and 90º of flexion.

Single-bundle reconstruction with quadriceps plus semitendinosus grafts (SBT)

After testing the double-bundle reconstruction (DB), both grafts were released from the femur and removed from their respective tunnels, except from the tibial tunnel.

The AL femoral tunnel was enlarged in thickness with a 12-mm drill. Both grafts were then inserted into this tunnel and fixated to the femur by tying them to the same bicortical screw previously used to fixate only the quadriceps graft. The cortical screw fixating the semitendinosus tendon in the posteromedial femoral tunnel was then removed.

The tibial insertion of the grafts was loosened, the medial arthrotomy was closed, and the knee was repositioned in the testing machine at 90º of flexion. The specimen was subjected to a 134-N anterior drawer, the grafts were tensioned, both at 90º, the quadriceps with 88 N, and the doubled semitendinosus tendon with 67 N, and fixated to the tibia, each in its own tibial fixation device (as previously described). The knee was then returned to position 0 and tested at 0º, 30º, 60º and 90º of flexion.

Results

Posterior Tibial Translation (PTT)

The data regarding the PTT values (mean, SD, and lower and upper limits of the 95%CIs) of the 5 testing conditions in each of the 4 angles tested are shown in Table 1, and their graphic representation, in Fig. 5. The comparison among the 5 testing conditions at each angle of knee flexion is shown in Tables 2-5.

Table 1 Posterior tibial translation, in millimeters, by testing conditions and angle of knee flexion. Mean, standard deviation (in parentheses) and 95% confidence intervals (in brackets) 

Flexion angle			PTT (mm)
	Intact	Injured	SB	DB	SBT
0°	9.48 (1.81) [8.08-10.86]	12.30 (2.72) [10.21-14.40]	10.46 (2.12) [8.83-12.08]	8.83 (1.60) [7.61-10.06]	9.55 (1.76) [8.19-10.90]
30°	10.98 (1.93) [9.49-12.46]	17.36 (3.24) [14.87-19.85]	13.70 (2.35) [11.89-15.51]	12.00 (1.97) [10.49-13.52]	10.96 (2.63) [8.93-12.98]
60°	8.92 (1.24) [7.97-8.87]	17.87 (3.01) [15.55-20.18]	12.60 (1.97) [11.09-14.12]	11.18 (1.82) [9.78-12.58]	9.87 (2.06) [8.29-11.46]
90°	8.31 (1.26) [7.34-9.28]	19.28 (2.33) [17.49-21.07]	12.83 (2.16) [11.17-14.49]	11.67 (2.22) [9.96-13.37]	10.27 (2.43) [8.40-12.14]

Abbreviations: DB, double-bundle reconstruction with a 10-mm quadriceps tendon and a 7-mm semitendinosus tendon; Injured, injured posterior cruciate ligament; Intact, intact posterior cruciate ligament; PTT, posterior tibial translation; SB, single-bundle reconstruction with a 10-mm quadriceps tendon; SBT, single-bundle reconstruction with a 10-mm quadriceps tendon plus a 7-mm doubled semitendinosus tendon.

Fig. 5 Posterior tibial translation (PTT) of all testing conditions at each flexion angle: intact posterior cruciate ligament (PCL) (blue line); injured PCL (red line); single-bundle reconstruction with a 10-mm quadriceps tendon (green line); double-bundle reconstruction with a 10-mm quadriceps tendon for the anterolateral bundle and a 7-mm doubled semitendinosus tendon for the posteromedial bundle (brown line); single-bundle reconstruction with a 10-mm quadriceps tendon plus a 7-mm doubled semitendinosus tendon (black line). 

Table 2 Comparison of the descriptive levels of the posterior tibial translation at 0^∘ of flexion among the testing conditions 

	Intact	Injured	SB	DB	SBT
Intact	-	x	x	x	x
Injured	p = 0.001	-	x	x	x
SB	p = 0.022	p = 0.003	-	x	x
DB	p = 0.180	p < 0.001	p = 0.001	-	x
SBT	p = 0.803	p = 0.006	p = 0.036	p = 0.183	-

Abbreviations: DB, double-bundle reconstruction with a 10-mm quadriceps tendon and a 7-mm semitendinosus tendon; Injured, injured posterior cruciate ligament; Intact, intact posterior cruciate ligament; SB, single-bundle reconstruction with a 10-mm quadriceps tendon; SBT, single-bundle reconstruction with a 10-mm quadriceps tendon plus a 7-mm doubled semitendinosus tendon.

Table 3 Comparison of the descriptive levels of the posterior tibial translation at 30^∘ of flexion among the testing conditions 

	Intact	Injured	SB	DB	SBT
Intact	-	x	x	x	x
Injured	p < 0.001	-	x	x	x
SB	p < 0.001	p = 0.001	-	x	x
DB	p = 0.085	p < 0.001	p = 0.001	-	x
SBT	p = 0.980	p = 0.001	p = 0.002	p = 0.140	-

Abbreviations: DB, double-bundle reconstruction with a 10-mm quadriceps tendon and a 7-mm semitendinosus tendon; Injured, injured posterior cruciate ligament; Intact, intact posterior cruciate ligament; SB, single-bundle reconstruction with a 10-mm quadriceps tendon; SBT, single-bundle reconstruction with a 10-mm quadriceps tendon plus a 7-mm doubled semitendinosus tendon.

Table 4 Comparison of the descriptive levels of the posterior tibial translation at 60^∘ of flexion among the testing conditions 

	Intact	Injured	SB	DB	SBT
Intact	-	x	x	x	x
Injured	p < 0.001	-	x	x	x
SB	p < 0.001	p < 0.001	-	x	x
DB	p = 0.001	p < 0.001	p = 0.003	-	x
SBT	p = 0.116	p < 0.001	p = 0.001	p = 0.005	-

Abbreviations: DB, double-bundle reconstruction with a 10-mm quadriceps tendon and a 7-mm semitendinosus tendon; Injured, injured posterior cruciate ligament; Intact, intact posterior cruciate ligament; SB, single-bundle reconstruction with a 10-mm quadriceps tendon; SBT, single-bundle reconstruction with a 10-mm quadriceps tendon plus a 7-mm doubled semitendinosus tendon.

Table 5 Comparison of the descriptive levels of the posterior tibial translation at 90^∘ of flexion among the testing conditions 

	Intact	Injured	SB	DB	SBT
Intact	-	x	x	x	x
Injured	p < 0.001	-	x	x	x
SB	p < 0.001	p < 0.001	-	x	x
DB	p < 0.001	p < 0.001	p = 0.014	-	x
SBT	p = 0.011	p < 0.001	p = 0.001	p = 0.001	-

Abbreviations: DB, double-bundle reconstruction with a 10-mm quadriceps tendon and a 7-mm semitendinosus tendon; Injured, injured posterior cruciate ligament; Intact, intact posterior cruciate ligament; SB, single-bundle reconstruction with a 10-mm quadriceps tendon; SBT, single-bundle reconstruction with a 10-mm quadriceps tendon plus a 7-mm doubled semitendinosus tendon.

The PTT of the SB technique was statistically higher (worse laxity) than that for the INT knee at all angles tested (p = 0.022 at 0º; and p < 0.001 at 30º, 60º and 90º). The DB technique showed values of PTT that were statistically similar to those of the INT knee at 0º (p = 0.18) and 30º (p = 0.085). The SBT technique was similar to the INT knee at 0º (p = 0.803), 30º (p = 0.98) and 60º (p = 0.116) of flexion.

The SB with a thicker graft (SBT) presented lower PTT values (better laxity) than the DB at 60º (p = 0.005) and 90º (p = 0.001). At 90º, all of the PCL reconstructions showed values of PTT that were statistically different from the INT knee.

Discussion

A recent systematic review¹⁷ of the literature analyzed the results of studies comparing the use of one or two bundles in PCL reconstruction.¹⁷ The review was only able to find 11 clinical studies so far, which shows that there is still a gap in the literature regarding this subject. Some authors report more physiological results and better laxity with the DB.²,¹⁰,¹¹,¹⁸-²⁵ Others, however, could not demonstrate the superiority of one technique over the other (one versus two bundles).¹²,¹³,²⁶-²⁹ This scenario made us wonder if the comparison should be made between one and two bundles or if, rather, there is a problem with graft volume that would be better solved with a thicker bundle (and not with several bundles). We then identified two key points to study in depth: the thickness of the graft and the second femoral tunnel. The influence of these two factors on reconstruction laxity was tested.

Increasing the graft thickness (by adding the 7-mm double semitendinosus tendon to the 10-mm quadriceps graft) in our SBs significantly reduced (p < 0.05) the PTT values at all angles tested. We chose the AL, which is not isometric, positioning for the femoral insertion in our SBs, since the AL bundle of the PCL has more fibers and greater ultimate strength than the PM. The choice was also made because the reconstruction with the femoral tunnel in this position results in greater stability than that achieved with other techniques (PM or isometric reconstructions).¹⁰,³⁰ We reproduced the position of the AL tunnel as performed in the study by Kokron et al¹², which also corresponds to the shallow 1 tunnel in the study by Mannor et al.,³¹ and to the AL tunnel in the studies by Harner et al¹¹ and Bergfeld et al.¹³ The value of 88 N of graft tensioning at 90º of flexion, with an anterior drawer of 134 N, was recommended by Harner et al.¹¹

The SB technique was unable to restore knee laxity at every measured angle. The same result was achieved by Kokron et al¹² with the same graft, and by Harner et al¹¹ with an Achilles tendon graft of 10 mm. In both studies, the same AL bundle failed to restore laxity to the knee at every angle. Different results were obtained by other authors, but with other tensioning and fixation methods.¹⁰,¹¹,¹³,¹⁵

Using a thicker graft (SBT), the PTT values were not significantly higher than in the INT knee at 0º, 30º and 60º of flexion. At 90º, however, this value was statistically higher. Once the insertion of the PCL into the femur is broad, corresponding to 150% of the femoral insertion area of the anterior cruciate ligament (ACL), and covering ~ 32 mm in the femur,³⁰ it seems reasonable that the graft used to reconstruct it should also be thicker than in the ACL reconstruction. We believe the improvement with a thicker graft is due to the greater amount of fibers opposing the tibial posteriorization, which is closer to the normal anatomy. In our opinion, the SBT results would improve if we could fill the entire native PCL femoral footprint with a graft. However, this cannot be accomplished with cylindrical tunnels. We also believe a femoral inlay fixation, which would better reproduce the anatomical femoral insertion of the PCL, could be a good option to obtain an even better laxity.

Regarding the influence of the second femoral tunnel, we observed that it also improved the reconstruction laxity at every tested angle compared with our SB with a thinner graft (quadriceps tendon of 10 mm). With the DB, the PTT showed no statistical differences in relation to the PCL INT knee at 0º and 30º of flexion. At angles of 60º and 90º, however, the PTT values in the DB were significantly higher than those of the PCL INT knee. As the femoral tunnel of our second bundle was positioned in a more posterior region of the anatomical insertion of the PCL, even more posteriorly than the second bundle (PM) described by Harner et al¹¹ and Race and Amis,¹⁰ and the deep tunnel described by Mannor et al,³¹ we believe it offered little resistance to the PTT at these higher angles of flexion. In our view, at these angles, the key role was played by the AL bundle, which was represented by the quadriceps graft of 10 mm, which, as in the SB, was unable to, alone, keep PTT close to the INT status.

As observed in other studies, adding more graft through a second femoral tunnel also decreased tibial posteriorization in our tests.²,¹⁰,¹¹,¹⁸,²⁰,²²,²³,²⁶ We agree, however, with Bergfeld et al,¹³ that this does not enable us to infer the superiority of the DB, since we are comparing techniques with different amounts of graft. Bergfeld et al¹³ and Kokron et al¹² also found no statistical differences between the SB and DB when using grafts with the same thickness in cadaveric isolated PCL injuries.

At this point, the question to be made is: since increasing the thickness of the graft improves the quality of the reconstruction, is it better to operate using a thicker single bundle or to increase the volume of the graft through a second femoral tunnel? To answer this question, we compared the SBT technique with the DB technique. In both situations, we used the same amount of graft (10-mm quadriceps tendon, plus doubled 7-mm semitendinosus tendon).

The site where the semitendinosus graft was added made no difference in terms of PTT with the knee in extension or at 30º of flexion. However, at 60º (p = 0.005) and 90º (p = 0.001), placing the graft in a second extra femoral tunnel significantly increased the PTT values when compared with adding it via the same AL femoral tunnel. In other words, the division of the graft into 2 bundles worsened the reconstruction laxity at 60º and 90º of knee flexion.

In the clinical field, three studies³²-³⁴ corroborate our experimental findings, reporting better results with thicker grafts. This improvement was observed both in the SB and DB techniques. Zhao and Huangfu,³² in a retrospective study comparing the techniques of single bundle reconstructions, concluded that the seven-bundle graft offers more stability than the quadruple graft in isolated PCL lesions. In DBs, Zhao et al³³ published results of 88.9% of normal and 11.1% of nearly normal International Knee Documentation Committee (IKDC) scores, using a thicker (8-fold) graft (quadruple semitendinosus tendon for the AL bundle and quadruple gracilis for the PM bundle). Chen and Gao³⁴ also published results with 8-fold grafts, with 78.9% of normal and 15.8% of nearly normal IKDC scores. Therefore, in our view, there is no reason to perform the DB, which is technically more complex and lengthy, in isolated PCL lesions, since better results can be obtained by adding the graft, not through a second femoral tunnel, but through the AL femoral tunnel, that is, through a reconstruction with a single thicker bundle.

The method used in the present study enabled the performance of all evaluations in each knee. Thus, the results of each technique could be compared among each other and with the biomechanical behavior of the same knee with an INT PCL, excluding any possible bias imposed by the effect of variation among specimens. However, to make it possible, the sequence of the tests could not be randomized, and this might be considered a limitation of the present study.

Conclusions

The results of the present biomechanical study enabled us to conclude that: