Loss of tension during graft fixation [58, 75], the effects of cyclic loading [8, 13, 15, 36, 57, 61], and graft preconditioning [13, 26, 36, 61] are important factors that affect the outcome of ACL reconstruction surgery. Using a porcine cadaver model, Yoshihara et al. [75] showed that the residual loads following graft fixation were significantly different among fixation methods. They found that the maximum initial graft tension when applied manually by the surgeon resulted in mean graft tension values equal to 116, 54, and 25 N after interference screw, post, and button techniques, respectively. Whether the tension was objectively measured or applied by the surgeon, which have been shown to be equivalent [70], would not change these findings. In a human cadaver study, an initial graft tension of 80 N was applied to soft tissue grafts fixed with interference screws with no preconditioning, cyclic preconditioning or isometric preconditioning applied to the graft prior to fixation [58]. A steady decrease of approximately 60 % occurred over 60 min. Likewise, Howard et al. measured the length change of bone-patellar tendon-bone graft during 4-min preconditioning and reported the graft length increased from 43.6 to 49.6 mm (a 14 % increase) [34].

Several investigators have shown that once the graft is fixed in the knee, anterior-posterior knee laxity increases with subsequent cycling. It has been reported that the tension in bone-patellar tendon-bone grafts dropped by 46 % at full extension after 1,500 cycles and that anterior-posterior knee laxity increased by 100 % after only 500 cycles [8]. This cyclic creep phenomenon is due in part to the viscoelastic behavior of ligament tissue [24], which appears to get worse in the days following ligament reconstruction, possibly due to enzymatic digestion [14]. Therefore, preconditioning the graft prior to implantation has been recommended to reduce stress relaxation and/or cyclic creep and to preserve graft tension following fixation.

Translational animal models provide researchers the opportunity to evaluate graft healing in response to different treatment strategies including the effects of initial graft healing. The first landmark animal study of initial graft tension was performed by Yoshiya et al. [76]. Using the canine model, ACL reconstruction surgery was performed in both knees using the medial one-third of the patellar tendon as a graft. The graft in one knee was fixed under an initial graft tension of 1 N, while the other was fixed with an initial graft tension of 39 N. The investigators found that the high-tensioned grafts exhibited poor vascularity and focal degeneration.

In an effort to control the many different variables that could affect graft healing in an animal model, Katsuragi et al. then designed a bilateral canine model in which they first devitalized the ACLs via freezing and then cored out the tibial insertions [38]. In the right knee, an initial “graft” tension of 20 N was applied to the distal bone block while the bone block in the left knee was anatomically reduced. Significant reductions in the tensile strength and tangent modulus of the grafts were found in the grafts tensioned to 20 N after 12 weeks of healing. Histology revealed focal degeneration in the grafts tensioned to 20 N. These studies demonstrate that minimal tension should be applied to the graft materials during graft fixation in ACL reconstruction and that overtensioning may be deleterious to healing in this highly controlled setting.

Several subsequent animal studies of initial graft tension have been reported. Using initial graft tensions of 1 N, 7.5 N, and 17.5 N with patellar tendon grafts, it was shown that the high-tension grafts were superior both biomechanically and histologically to the low-tension grafts after 32 weeks of healing in the rabbit model [41]. To the contrary, a goat study comparing lax bone-patellar tendon-bone grafts to those tensioned to 44 N found that anterior-posterior knee laxity values between the two groups were not significantly different after 2 weeks of healing [19] and that the failure properties measured after 6 months were not significantly different [20]. These results were supported by another goat study in which patellar tendon grafts were tensioned to 5 N and 35 N [1]. While there were significant differences between the two initial graft tensions of 5 and 35 N at Time Zero, no significant differences in anterior-posterior knee laxity or graft failure properties were found after 6 weeks of healing. A recent rat study [31] comparing graft tensions of 2 N and 4 N also found no significant differences in the failure properties after 6 weeks of graft healing, though they presented qualitative histological evidence that the grafts of the high initial graft tension group was of higher integrity.

From the above review, the interpretations of the animal model results are conflicting with some studies showing that low tension is better and that high tension is superior,and the majority showing that there are no differences between high and low initial graft tensions. Unfortunately the use of animal models to evaluate the effects of initial graft tension is limited as a model for the human condition due to differences in joint anatomy, ranges of joint motion, and limb alignment [62] and because quadrupeds are more dependent on the ACL than humans [12]. Also, it is difficult to precisely control other surgical parameters, such as graft positioning, and postoperative rehabilitation, factors that will also affect graft integrity and outcomes. The novel model proposed by Katsuragi et al. was designed to eliminate some of these confounding variables [38], and their results clearly show that initial graft tension is an important factor to understand. While animal models provide insight into the effects of initial graft tension on graft healing, clinical studies are required to validate the findings of animal models and to ultimately determine which factors will ultimately affect clinical practice.

There are several clinical studies evaluating the effects of initial graft tension on clinical outcomes after ACL reconstruction. Yasuda et al. compared three different initial tensions at graft fixation (20, 40, and 80 N) with single-bundle ACL reconstruction using autogenous hamstring tendon graft in line with polyester tape. They reported that the postoperative side-to-side difference in anterior knee laxity was significantly less in the 80 N group compared to the 20 N group 2 years or more after surgery [73]. Thus they concluded that relatively high initial graft tensions reduced the postoperative anterior knee laxity after ACL reconstruction. Using patellar tendon autografts, Nicolas et al. reported significant differences in anterior-posterior knee laxity when initial graft tensions of 45 and 90 N were applied and that a graft tension of 45 N was not sufficient for restoring knee stability [56].

In contrast, Yoshiya et al. who reconstructed the ACL with patellar tendon autografts using two different initial tensions (25 and 50 N) [77] and Kim et al. who compared three initial tension levels (78, 117 and 147 N) with autogenous hamstring tendon grafts [39] reported no significant differences in clinical outcomes at final follow-up. Similarly, van Kampen et al. compared clinical outcomes 2 years after patellar tendon ACL reconstruction using initial graft tensions of 20 and 40 N, and they found no significant differences in outcomes between the two initial graft tension levels [71]. They also argued that the initial graft tension of 20 N seemed to be sufficient without the risk of over-constraining the knee joint. In an evidence-based review of the randomized control trials of initial graft tension, it was concluded that “there is no clear trend in terms of statistically significant or clinically relevant differences in terms of the amount of tension to apply to the graft during graft fixation” [5].

More recently, Mae and Shino previously performed 33 isometric Rosenberg bi-socket ACL reconstructions with hamstring tendon graft via three different amount of initial tension of 60, 80, and 100 N as a pilot study, and compared the side-to-side difference with KT Knee Arthrometer at 2-year follow-up among the three initial graft tension groups. The average side-to-side difference was 2.0 mm for 60 N, 1.1 mm for 80 N, and 1.6 mm for 100 N, respectively. While there were no significant differences between the initial tension conditions, they showed that the variation associated with the grafts tensioned to 100 N was the largest among three groups (Fig. 27.2). These results suggest that excessively high initial graft tensions may be unnecessary for improved ACL outcomes.

Alternatively “laxity-based” initial graft tension protocols that restore or modulate anterior-posterior knee laxity at the time of surgery have also been recommended [3, 29, 30]. A “laxity-based” initial graft tension technique would provide benefits because it does not require the use of a tension measuring device, only an assessment of anterior-posterior knee laxity during the graft fixation procedure [29]. When using a laxity matching protocol, the initial graft tension can be adjusted to produce an anterior-posterior knee laxity value that is equal to or less than that of the contralateral normal knee [30], depending on the surgeon’s preference. Laxity could either be objectively measured intraoperatively using an arthrometer [30] or subjectively using the Lachman and drawer test. In a cadaver study, it was determined that the laxity-based approach better restored normal knee laxity than the force-based approach at Time Zero [29]. However, in a recent prospective randomized clinical trial, it was determined that the setting the anterior-posterior knee laxity to be equal to that of the contralateral uninjured knee at the time of surgery resulted in equivalent outcomes when compared to over-constraining anterior-posterior knee laxity by 2 mm [30]. These data suggest that setting the anterior-posterior knee laxity within this laxity window at the time of surgery is a reasonable target.

The standard of tension required to determine the optimal initial tension is not known. However, a review of the literature suggests that a laxity-matched pretension (LMP), which is the graft tension required to obtain the normal anterior-posterior knee laxity in ACL reconstruction, may serve as a useful standard. One of the first cadaver studies evaluating the LMP was performed by Burks and Leland [16]. They measured the LMP for several graft materials in single-bundle ACL reconstruction and reported that the LMP value was 16 N for bone-patellar tendon-bone graft, 38 N for doubled semitendinosus graft, and 61 N for iliotibial band graft. They demonstrated that the required tension varied among graft materials. Other cadaver studies supported the general finding that graft type is a primary factor that must be accounted for when evaluating optimal initial graft tension strategies [2, 25, 37]. Surgical technique, including tunnel position and the number of tunnels, is also an important factor for determining the LMP in ACL reconstruction. The LMP value was 25 N for the conventional isometric Rosenberg technique with twin femoral tunnels and smaller than that using the same technique with a single femoral tunnel (44 N), while the LMP value for anatomic twin-tunnel technique (7.3 N) was smaller than that for the isometric Rosenberg twin femoral tunnel technique (25 N) [45, 46]. Therefore, the optimal graft initial tension should be determined, based on the graft materials and the operative techniques assuming no changes occur postoperatively.

When fixing a hamstring tendon graft to the tibia, sutures are typically tied to a fixation post screw with manually applied “maximum” tension. However, the suture-post method includes some indefinite factors: (1) variability between surgeons, (2) risk of loosening or breakage of the sutures during knot tying, and (3) stress relaxation of the graft-suture fixation construct after fixation. Double staple techniques combined with polyester tape and spike washer with a screw for soft tissue grafts are also available to control tension with a tensioner [53]. However, these fixation techniques still run the risk of graft slippage, resulting in a loss of graft tension. With interference screw fixation, it is difficult to control the initial graft tension as the tension changes substantially during screw insertion. The double-spike plate (Meira Co., Nagoya, Japan) was developed for secure graft fixation with the intended tension (Fig. 27.3). Shino et al. reported that the graft tension after fixation with the double-spike plate temporarily increased while the base spikes were hammered in place, but the intended tension was maintained even 5 min after fixation [66]. They showed the high reliability in initial fixation using the Double Spike Plate.

In the situation where the graft is tensioned with a tensioner, the tension is typically measured when the graft is manually pulled by the surgeon and is not adequately referenced to the tibia. While the manual technique is quite simple, the graft tension after fixation is likely to be variable because the position of the tibia relative to the femur is not controlled and because the tension measurement is referenced to the surgeon’s hand which may not be transferred directly to the tibia. In this case, the tension in the graft when it is fixed may immediately decrease after fixation due to the subsequent posterior and proximal translation of tibia. A metal shell boot with tensioners connected to grafts (tensioning boot system; Meira Co., Nagoya, Japan) makes it possible to tension the grafts relative to the tibia, as the boot is fixed to the calf with a bandage (Fig. 27.4). It may be expected that the intended tension will remain in the graft after fixation using this tensioning boot system assuming no changes occur during the healing process.

The initial graft tension does not change the shape of the graft tension versus flexion angle curve during passive flexion extension motion, but only shifts the entire curve up or down [17, 28, 52]. Given that the tension in the ACL graft after fixation is dependent on the knee flexion angle, it becomes clear that the knee flexion angle at which the tension is applied is extremely important. The tension in the ACL when the knee is at full extension is high, drops to a minimum or becomes slack at about 20–30°, and then increases with further flexion [11, 51]. If a graft is tensioned when the knee is at 30° of flexion, the entire tension-flexion curve is shifted upward increasing the tensions across the entire range of motion, particularly high at full extension. This effect of flexion angle on graft force in anatomic graft position can be larger than that in isometric graft position, as the graft length change in the former is larger than that in the latter [69]. In a cadaver study, Bylski-Austrow et al. demonstrated that an increase in knee flexion angle from 0 to 30° when the initial graft tension was applied increased the forces in the ligament across the entire range of motion [17]. As previously mentioned, an excessively large graft tension during range of motion may lead to abnormal tibiofemoral positioning, resulting in cartilage degeneration and a graft tear. This finding was verified by several other investigators [28, 60] and emphasized by Gertel [32]. Therefore, the knee flexion angle at which the initial graft tension is applied is an important parameter that must be designated and/or controlled when performing ACL reconstruction surgery.

There are several limitations inherent to all Time Zero studies that must also be considered [27] when developing recommendations for optimal tension strategies for either single or double-bundle ACL reconstruction procedures. While cadaver experiments and Time Zero human experiments permit the use of tightly controlled experimental protocols and accurate data collection, the conclusions drawn may be limited in clinical relevance. While these studies provide important information of performance at the time of the ACL reconstruction, they do not take into account the changes that may occur during graft healing, such as tunnel enlargement [35, 43], viscoelastic changes [8, 13], graft remodeling [22], histologic degeneration and decreased vascularity [76], and the long-term consequences of cyclic creep even with graft preconditioning [13]. It is important to remember that the graft first undergoes a period of necrosis, followed by cell infiltration, revascularization, and then remodeling [6]. It is very likely that the initial graft tension condition at the time of surgery is not maintained. Translational and clinical studies that include the temporal effects of healing are paramount to determining the relevance of different initial graft tension strategies. While biomechanical and translational studies are important to understand the interactions between the initial parameters, only through carefully controlled, prospective randomized controlled trials will we better understand which initial graft tension parameters really matter.