A major concern following hamstring harvest is the risk of subsequent hamstring strain or other injury in the future. Hamstring injuries are common, especially in high-level athletes and those participating in various field sports such as soccer. Such injuries can lead to significant disability and time lost due to injury [5, 6]. One may postulate that harvest of the hamstrings could lead to a change in the overall biomechanics of the patient’s lower extremities and potentially place them at increased risk for future hamstring strains. The current evidence in the literature regarding hamstring strains after harvest is limited, although several studies have addressed this subject.

Scranton et al. studied 120 patients at least 2 years following hamstring harvest for ACL reconstruction with a hamstring pain evaluation form. They noted that 22 % of patients had minimal symptoms or a “twinge” of hamstring pain at some point during their post-ACL rehabilitation but that these symptoms did not limit their rehabilitation or return to athletic function [7].

There have been studies linking previous knee injury or ACL reconstruction with increased risk of hamstring injury. Koulouris et al. reported that athletes with a previous ipsilateral ACL reconstruction were at increased risk for hamstring reinjury (66.6 %) compared with athletes without a previous ACL reconstruction (17.1 %) [8]. These findings were similar to those of Verrall et al., who prospectively evaluated athletes on two Australian rules football clubs. When they compared those who sustained hamstring injury during the season with those who did not, they noted the injured athletes were significantly more likely to have had a prior severe knee injury such as an ACL tear [5]. Finally, a systematic review was performed by de Visser et al. to identify risk factors for reinjury following acute hamstring strains. Aside from grade and size of initial injury, the only other variable found to have a significant association with reinjury was a history of previous ACL reconstruction [9].

While three studies described above do demonstrate negative effects of ACL injury of the risk of subsequent hamstring injury and prolonged recovery, it should be noted that none of the studies demonstrated any influence of autograft type (bone patellar tendon bone or hamstrings). Each noted the risk of hamstring injury to be equal between those who were reconstructed with hamstring and bone patellar tendon bone autograft. This finding suggests that ACL injury rather than hamstring harvest is in fact associated with subsequent hamstring problems.

The technique used to harvest the hamstrings has also been hypothesized to affect donor site morbidity. The Push technique uses tendon strippers to push past the myotendinous junction until the tendon comes loose (see Fig. 7.1). The Cut technique involves a device in which a predetermined length of tendon is harvested by using a small blade to cut the tendon near the myotendinous junction. D’Alessandro et al. performed a randomized clinical trial to assess if there was a difference in hamstring pain, muscle strains, and leg flexion strength between these two techniques. While there was no difference found in the strength or the activity level between the two groups as measured by the Cincinnati sports activity ratings scale, the “Cut” group recorded lower average pain on the visual analog scale of 10 mm compared with over 24 mm (p = 0.0398). There was also a significant difference in the incidence of hamstring strains with 25 % of the “Cut” group reporting at least one strain in the postoperative period compared to 50 % of the “Push” group (p = 0.045). The authors theorized that tendon regeneration following the “Cut” technique would be more robust than that from a stripped muscle belly and could promote regeneration of a more normal, functional tendon [10], but this finding was not clearly demonstrated.

Overall the evidence is very limited with regard to the relationship between hamstring harvest and subsequent hamstring injury. While the studies demonstrate that having a previous ACL reconstruction puts one at risk for hamstring injury, the relationship appears independent of graft type. More studies are needed to distinguish whether graft harvest or the altered biomechanics of the recovering limb is the primary risk factor for this association.

Loss of hamstring strength following harvest for ACL reconstruction is a major concern, spawning numerous studies that compare strength between the harvested limb and the contralateral side. In 1997 Simonian and colleagues studied the effect of hamstring harvest of both the semitendinosus and gracilis tendons with a minimum 3-year follow-up [11]. They looked at size of the muscles and extent of retraction on MRI along with dynamometer testing at 90 and 180°/s. The average hamstring strength of the operative limb was 95.3 % of the contralateral side, which they found to not be significant. The MRI findings showed the average semitendinosus and gracilis insertions to be 26.7 and 47.1 mm more proximal on the operative side. The select patients without evidence of semitendinosus tendon remnant at 10 cm above the joint line showed an average of 10.3 % strength deficit compared to the contralateral side.

In 1998 Ohkoshi et al. took the question a step further and looked at not only the peak torque generated by the harvested hamstrings but also their peak torque angle in flexion. Although the overall peak torque values of the operative and nonoperative knee were not significantly different, the peak torque angle decreased significantly, ranging from 11.7° to 15°. This finding shows the effect of the shortening of the muscle unit, with the shift in the peak torque curve to the left toward the earlier angles of flexion. More than 80 % of their patients showed no secondary peak in the latter half of the curve as was seen on the nonoperative side [12]. Tadokoro et al. verified this finding with a 2-year follow-up study in which he measured peak torque in the sitting position at 90° and prone at 90° and 100°. The isometric peak torque was reduced to 86.2 %, 54.6 %, and 49.1 %, respectively, in those three positions [13].

Similarly, Nakamura et al. noted the side to side torque ratio at 90° of flexion was significantly lower (by more than 10 %) than that in the ratio at the peak torque flexion ankle. Their results show that knee flexor strength of the harvested limb is less restored at deeper knee flexion angles than at the angle where peak torque is generated [14].

Previous studies suggest that recruitment of motor units in the quadriceps was hindered bilaterally after ACL reconstruction [15, 16]. A similar effect may be present in the hamstring muscle group. A study by Konishi et al. used MRI to measure muscle volume of the semitendinosus in controls, uninjured limbs of patients having ACL reconstruction, and the injured limb after harvest at time periods of ≤6 and 12 months after surgery [17]. The results of the study indicated that the strength of the injured knee flexors at ≤6 months were significantly lower at both 60 and 180°/s velocities. At 12 months the strength recovered to more than 90 % of the uninjured knee, which suggested slight residual weakness in the injured group, but the significance dissipated. The total volume of hamstrings was also significantly decreased in the injured compared to uninjured limbs. When calculating the muscle torque per unit volume, the results indicated significantly lower peak torques in both injured and uninjured sides at 12 months after surgery compared to controls. Interestingly there was no difference between the injured and non-injured sides. This suggests that there may be bilateral weakness in the hamstrings similar to reports of the quadriceps. The mechanism is still unknown as to the cause of this pattern, but the authors deduced that it was not likely neurological in nature because they did not show a significant difference at ≤6 months of muscle torque per unit volume compared to controls.

The flexion strength of the hamstrings may be correlated to their ability to regenerate. Choi et al. in 2012 evaluated flexion strength and functional performance based on the hamstring tendon regeneration seen on 2-year follow-up MRI. Isokinetic testing and MRI were performed on patients, and they were placed in three groups: both semitendinosus and gracilis regenerated, only one tendon regenerated, and no tendon regeneration. Their results showed significant differences in flexor deficit between the groups in which no tendon regenerated and the groups where at least one tendon regenerated. Functionally there was a significant difference in the carioca test time between those that had no tendons regenerate and those that had both tendons regenerate [18].

It is important to remember that the hamstrings also function as secondary hip extensors and their harvest may have an impact on hip extension and adduction strength. This role could be important for athletes who participate in sports with significant running at high speeds such as soccer, American football, track, and rugby. Hiemstra et al. briefly reported on an outcome analysis in which they compared 15 patients 1 year removed from ACL reconstruction with hamstring autograft to a control group and found that hip adduction strength was diminished significantly by 35.8–43.7 % when compared to BMI normalized controls. In terms of hip extension, the control group had a naturally significant difference in dominant over nondominant limb strength. That natural difference was lost in patients in who the dominant ACL reconstructed. Hip extension in these limbs was diminished to the level of the nondominant side [19]. Goeghegan and colleagues performed a nonrandomized prospective case control comparing hip extension strength following ACL reconstruction with semitendinosus and gracilis harvest versus bone patellar tendon bone harvest. They found a significant difference at the 3-month mark with the hamstring harvest group having weaker extension than the bone patellar tendon bone group. By the 12-month mark; however, the hamstring group had recovered to a level equal to the bone patellar tendon bone group. They concluded therefore hamstring harvest has no detrimental long-term effect on hip extension compared to bone patellar tendon bone graft harvest [20].

Two common four-strand hamstring autografts are the quadrupled semitendinosus graft (4ST) and the doubled semitendinosus and gracilis graft (2ST-2G). Some would suggest that the gracilis plays an important role in reinforcing the semitendinosus muscle in deep knee flexion and that harvesting the gracilis as well puts the patient at a strength disadvantage. The gracilis and semitendinosus also are active in internal tibial torsion, and there is concern that harvesting both leaves the patient without a muscle to be able to compensate for the loss. The counterargument for harvesting both tendons is that the 4ST graft is at times not long enough to support adequate bone tunnel fixation.

Adern and Webster performed a systematic review on the limited number of studies that compared outcomes of 4ST or 2ST-2G grafts in ACL reconstruction [21]. They found seven studies that compared the knee flexion strength in the two groups. All studies had a harvest from the ipsilateral limb. No differences were found between the two groups in the six of the seven studies in isokinetic peak torque at 12 months, although there was a trend in favor of the 4ST graft. One study obtained torques at 70°, 90°, and 110° knee flexion at 60°/s and 180°/s velocities and noted a significant difference at all angles favoring the 4ST group at the 18-month mark [22]. Two studies recorded active knee flexion range of motion with the hip in full extension, and both studies showed a significant deficit in standing knee flexion in the 2ST-2G grafts [14, 23]. The clinical relevance of the prone isometric difference and the standing knee flexion angle remains to be determined. These tests are performed with the hip in full extension and may place the hamstrings at a length disadvantage if there is any shortening of the muscle after harvest. This position is relatively rarely encountered during most sports, and weakness in this position may be clinically important.

Since the review by Adern and Webster was performed, Yosmaoglu et al. evaluated 46 subjects who underwent ACL reconstruction and divided them into 4ST and 2ST-2G groups [24]. They recorded isokinetic quadriceps and hamstring torque, motor coordination, and anterior tibial translation at 12 months after surgery. There were no side to side differences in quadriceps peak torque, tibial translation, or motor coordination found. The only difference found that was significant was in side to side flexor peak torque at 60°/s favoring the 4ST group.

Recently, Barenius and colleagues used subjective scores, function, strength, and tibial rotation measured by gait analysis as variables for evaluating 20 patients who had ACL reconstruction using either 4ST or 2ST-2G grafts [25]. For the tibial torsion portion, they used a three-dimensional motion analysis system and analyzed rotational laxity during stair descent and pivoting activity. All of the outcome measures including tibial torsion showed no significant differences between the two groups.