Introduction

Anterior cruciate ligament (ACL) reconstruction has become less invasive, tending to preserve bone stock as well as tendons and to improve fixation1.

Although a significant number of potential grafts for ACL reconstruction have been reported, the hamstring autografts are among the most commonly used.

The most common preparation technique using hamstring is the doubled semitendinosus (ST) and gracilis tendons (G). However, to decrease the morbidity associated to hamstring harvesting, a quadrupled ST graft construct has been developed in the last few years2,3. This technique has the advantage of preserving the gracilis, potentially improving postoperative hamstring strength4, besides providing a larger diameter than ST-G graft.

Recently, Silva et al.5 published a new quadruple ST graft construct using two cortical button and adjustable loops. The purpose of this study was to evaluate its biomechanical properties.

It was hypothesized that this quadruple ST graft construct with double cortical suspension provides a higher stiffness and resistance fixation along with a lower elongation after cycling.

Materials and Methods

Specimen Preparation

The study, in compliance with the Helsinki Declaration, was approved by the institutional ethics committee (DIFD – Department of Investigation, Formation and Documentation), South Section and Board of Directors of the National Institute of Legal Medicine and Forensic Sciences. It was confirmed that all the cadaver bodies used, were not included in the National Register of Non Donors (RENNDA form), which avoids the need for informed consent form.

A total of 15 fresh human cadaver semitendinosus tendons were harvested (age: 21–62) at the first 36 hours after death time.

Grafts were wrapped in a wet dressing (NaCl 0.9%) on sterile boxes, stored at 10 °C while waiting for the biomechanical analysis, performed up to one hour after harvesting. Tendons length was defined as 240 mm in order to achieve a constant folded graft length of approximately 60 mm.

Graft Preparation

Both free ends of the semitendinosus were stitched over a length of 30 mm, using No. 2 ExpressBraid™ suture (Zimmer Biomet, Warsaw, IN). Two adjustable-length loop cortical button devices were used: ToggleLoc™ (Zimmer Biomet, Warsaw, IN), for the femur, and ToggleLoc XL Inline™ (Zimmer Biomet, Warsaw, IN), for the tibia. The tendon graft was symmetrically folded over the tibial cortical button loop and the doubled graft was passed through the femoral cortical button loop and once again symmetrically folded. Both grafts free ends are tied together with 3 knots over the tibial button loop. Then, the same sutures are tied over the graft itself using 4 knots5.

Finally, the graft is reinforced to the tibial side with a cerclage suture and a buried knot6.

To simulate the common tunnel lengths on the femur and tibia we assumed a constant femoral side length of 40 mm (including graft’s 15 mm) and a tibial side length of 50 mm (including graft’s 20 mm), adjusting both ToggleLoc™ loops, in order to achieve these values.

Biomechanical Testing

Tests were conducted in a servo-hydraulic device (retrofitted Instron 8800 plus) with a loading cell of 10.0 kN (class 0.5). The graft construction ends (ToggleLoc™) were then placed in custom designed jigs which were rigidly fixed to the machine grips as shown in Fig. 1a.

Figure 1
figure 1

Test sequence (sample 904): (a) step 1; (b) step 7; (c) after test is completed.

The test protocol consisted in several steps from pretensioning to tensile test until failure finally took place (Table 1).

Table 1 Test Protocol.

Displacement and load were recorded by using the BlueHill Instron native software7. The elongation (millimeter) was determined as the difference between the displacement at a certain load and the initial load. For the load protocol, elongation during pre-tensioning (step 2) - Pre-tensioning elongation- and after 1000 cycles (step 5) - cycling elongation - were measured.

Pre-tensioning elongation represents the construct tensioning in traction table before graft implantation and cycling elongation represents the construct elongation after graft fixation followed by 1000 gait cycles.

A final tensile test (step 7) was done to evaluate both the ultimate load failure (ULF) and stiffness. Stiffness (N/mm) was then calculated as the load-elongation’s slope for the first part of the curve, i.e., until a load of 300 N. The final test set-up can be depicted in Fig. 1a.

The mechanism of failure was recorded (Fig. 1c).

Statistical Analysis

Statistical analysis was performed using SPSS Statistics software, version 24 (IBM Corporation, Armonk, NY). The Shapiro–Wilk test was used to check whether the data had a normal distribution. Paired t-tests were used to compare groups of two values obtained by the same individuals in different circumstances. Data descriptions include means, 95% confidence intervals, standard deviations, first and third quartiles, median, maximum and minimum values. The significance level was set at α = 0.05.

Test power calculation was obtained using R software (R Core Team (2017))8,9.

Results

Study power

The estimate of the correlation (ρ) between elongation after pre-tensioning at 300 N and after 1000 cycles, is given by the sample correlation (r) obtained from the sample, which was equal to 0.7.

Assuming that the standard deviations of each elongation (pre-tensioning at 300 N and after 1000 cycles) are equal, i.e., making the general assumption for the homogeneity of variance we consider that σ300 = σ1000 = σ where σ is estimated by the sample standard deviation s = 1.8 mm. We can then use the following expression to obtain the standard deviation for the differences (diff) in elongations between the pre-tensioning at 300 N and after 1000 cycles (Eq. 1).

$${\sigma }_{diff}=\sigma \sqrt{2(1-\rho )}$$
(1)

Consequently, we obtain an estimate of σdiff equal to 1.394 mm.

In order to detect a difference of about 1 mm between elongations after 300 N and 1000 cycles, when using a sample of size 15 (n = 15) with 0.05 significance level and assuming that the standard deviation of the differences is 1.39 mm, the obtained power was 84%.(a)

Graft

Mean quadruple ST graft final diameter was 9.7 mm (8–11 mm (Table 2)).

Table 2 Graft Characteristics.

Pre-tensioning

Mean total graft construct’s elongation after pre-tensioning at 300 N was 12.8 mm (SD: 1.59 (Table 2)), with mean graft elongation of 4.1 mm (SD: 1.10 (Table 2)). The latest corresponding to 32% of total graft construct’s elongation.

Cyclic loading

Mean cycling elongation (1000th cycle) was 0.44 mm (SD: 0.24), considered significant (p < 0,001), comprising a confidence range between 0.31 and 0.57 mm (Table 3).

Table 3 Biomechanical testing summary.

All specimens survived the cyclic loading testing (load-elongation depicted in Fig. 2).

Figure 2
figure 2

Load–elongation plot of a test (sample 961D) with the 7 steps. For the cyclic step only 3 cycles were shown (5, 30 and 1000th).

Load-to-failure test

Mean ULF and stiffness values were 849.46 N (SD:114.57) and 221.49 N/mm (SD:25.16), respectively (Table 3).

All failures occurred in the graft. In 13 out of 15, the rupture was at the femoral side folded graft (Fig. 1c). In the remaining two cases, one failed on the tibial side in the transition between the graft and suture and the other occurred after loosening of the 4 tied knots on the tibial side over the ST.

Discussion

The most important finding of the present study was that a quadruple ST graft construct with two cortical suspension devices provided a higher stiffness and resistance fixation along with a lower elongation after cycling.

Elongation of the ACL graft construct should be kept to a minimum in order to avoid postoperative laxity and the risk of early ACL reconstruction failure.

With the development of adjustable-length loop cortical button devices, a new graft preparation technique using a cortical button graft fixation for the femur and tibia was introduced as described previously10. However, the effect of this preparation method on graft elongation is so far not known.

Proper pre-tensioning of graft construct is of extreme importance11 to allow a better accommodation of all the interfaces (graft, sutures and adjustable loops12,13,14) and thus avoiding a higher elongation after graft implantation. There is no consensus regarding how much load is actually needed for this pre-tensioning. Considering that the common walking loads acting on the graft are estimated to be around 298 N15, in the current study pre-tensioning was undertaken at 300 N load, in an attempt to reproduce physiological loads.

The elongation of the graft construct at this stage was 12.8 mm substantially higher than the one mentioned by Mayr et al.16. After pre-tensioning under 50 N, the elongation obtained in three different quadruple ST graft constructs ranged in-between 1.6 mm and 2.1 mm. Similar findings were reported by Petre10 referring a 2.45 mm elongation after pre-conditioning between 10 and 50 N. However, the used loads in those studies were 30 to 6 times lower (10–50 N) than the ones we applied.

In fact, greater elongation after pre-tensioning under high loads and low elongation after cycling, proves the importance of pre-tensioning in this kind of graft construct.

Another concern topic is elongation after cycling, corresponding to the elapsed time after graft implantation. Ideally, it should be lower than 3 mm difference (side to side), considered by some authors as the limit above which failure is usually due to laxity17,18.

In the present study, cumulative cyclic displacement determined after 1000 cycles was 0.4 mm and lower than the previous published data. According to Mayr et al.16 cyclic elongation after 1000 cycles ranged between 6.1 mm and 7.0 mm whereas Petre et al.10, mentioned higher cyclic elongation values (3.34 mm) and similar data were described by Barrow et al.11. These results are clearly higher than the ones obtained in our study (0.4 mm), suggesting that elongation after cycling is dependent on the selected pre-tensioning loads.

Stiffness and ultimate load failure of the studied graft construct were also evaluated. All the tested 15 constructs had mean values of stiffness (221.49 N/mm) and ULF (849.46 N) comparable or slightly higher than those published in the literature16.

There are several limitations to the current study.

First, the axial direction of the applied load, differs from clinical conditions6,15,19. This fact leads to higher stresses on the graft bending points over the adjustable loops. However, peak loads used in this study were similar to the forces experienced by the ACL during early rehabilitation19,20.

Secondly, it was not possible to hydrate the graft at the bending points (Fig. 1b), in opposition to the “in vivo” conditions where graft is permanently hydrated, increasing graft fragility and decreasing the ULF.

Finally, it is important to emphasize that this in vitro study focuses solely on graft construct biomechanics and does not include any biological factors.

Nonetheless, in this study, we used fresh human cadaver grafts, minimizing the bias that can be associated to frozen human or animal grafts.

The clinical relevance of this study comes from the fact that this graft construct technique may allow a safe rehabilitation with a low risk of elongation.

Conclusion

Quadruple ST graft construct using two cortical buttons with adjustable loops, showed a higher stiffness and resistance with a very low elongation after cycling.

Pre-tensioning the graft construct under high loads, seems to prevent from additional significant displacement after cycling, assuming a critical role in its preparation.