Introduction

Anterior cruciate ligament (ACL) reconstruction with hamstring graft (HG) is a commonly performed orthopedic procedure encompassing many reconstructive techniques associated with different fixation devices [1–3]. It is well known that the HG, once implanted in the knee joint, undergoes a biological remodeling process termed “ligamentization.” This process involves both the intra-articular and the intra-tunnel portion of the graft [4–6] and consists in 3 consequential phases, where the HG goes thought a central graft necrosis and a consecutive restructuring process, ending with an ACL-like tissue [7, 8].

As already demonstrated by Zaffagnini et al. [9], the pes tendon insertion is well vascularized and richly innervated; what is more, this neurovascular network continues along the length of the pes tendons. The maintenance of the neurovascular supply may help the newly reconstructed ACL by avoiding the mechanical impairment related to the necrotic phase [9, 10].

The International literature reports some case series dealing with the maintenance of HG tibial insertion, but a recent review on the topic was unable to draw definitive conclusions to validate HG tibial insertion sparing in ACL reconstruction because of the paucity of well-designed randomized clinical trials and because of the difficulties in evaluating the graft ligamentization in the early postoperative period [3, 6, 11].

The adoption of arthroscopic second-look procedures as well as biopsy specimens harvesting raises unacceptable ethical concerns on humans. For this reason, the only noninvasive tool able to evaluate the ligamentization process seems to be the MRI. Recently, Figueroa et al. [12] proposed an MRI score aimed to evaluate graft morphology and integration after ACL reconstruction, encompassing the use of a detached HG.

The aim of the present paper is to report the results of a prospective randomized study comparing clinical and MRI results between two different ACL reconstructive procedures with and without HG tibial insertion preservation.

Materials and methods

Sample size was calculated performing a power analysis. The parameter chosen to define sample size was the MRI Figueroa score [8], which estimates graft ligamentization and integration with a scoring system ranging from 0 to 5. Assuming as relevant 0.5 points of difference between groups with a standard deviation of 0.53 points retrieved from the literature [8], with a confidence interval of 95 % and a statistical power of 0.80, 40 patients were required in order to obtain a statistically significant relevance at p ≤ 0.05. These 40 patients were randomly divided into two groups by means of dedicated software. The inclusion criteria were primary unilateral ACL lesion in patients ranging from 18 to 40 years old. The exclusion criteria were represented by ACL re-rupture, preoperative radiographic signs of arthritis, associated osteochondral lesions graded III or IV according to the ICRS classification system, axial deviation of the injured knee, multidirectional instability.

Patients within the study group underwent an ACL reconstruction using a distally inserted HG with a complete tibial tunnel and a femoral socket [1]. Patients within the control group underwent an ACL reconstruction using a detached ST graft with both a tibial and a femoral socket [13, 14].

The ethics committee at the authors’ institution approved the study, and all patients signed an informed consent to participate in the study.

Clinical assessment and MRI evaluation

Subjective and objective International Knee Documentation Committee (IKDC) score was administered preoperatively and at 3-, 6-, 12- and 24-month follow-up. Return to sport activities was evaluated by means of the Tegner activity scale at 12- and 24-month follow-up.

MRI evaluation was performed at 6-month follow-up and was evaluated by a qualified radiologist in musculoskeletal pathology. Graft ligamentization at the intra-articular portion and bone-tendon graft integration on the tibial side were assessed following the parameters described by Figueroa et al. [12]. These parameters estimate both the ligamentization and the integration of the graft, respectively, investigating the graft signal pattern or intensity and the presence of synovial fluid at tunnel–graft interface. Graft signal intensity was evaluated as hyperintense, isointense and hypointense: 1, 2 and 3 points were assigned, respectively. Synovial fluid presence was reported as positive or negative: 1 point was assigned for positive, and 2 points were assigned for negative.

MRI scans were performed with a 1.5 Tesla whole-body scanner, General Electric Medical Systems, used a Phased Array 8 ch Knee Coil. The following sequences were used:

  1. 1.

    Two-dimensional (2D) Sagittal proton weighted fast spin-echo (DPFSE) with fat saturation (repetition time [TR] 3500 ms; echo time [TE] 43 ms; field of view [FOV] 18; matrix: 320 × 256; slice thickness 3 mm; interslice gap 0.3 mm).

  2. 2.

    2D coronal DPFSE proton density weighted with fat saturation (TR 3500 ms; TE 43 ms; FOV 18; matrix: 320 × 256; slice thickness 3 mm; interslice gap 0.3 mm).

  3. 3.

    2D coronal T1-weighted FSE (TR 500 ms; TE 15 ms; FOV 18; matrix: 320 × 256; slice thickness 3 mm; interslice gap 0.3 mm).

  4. 4.

    2D axial DPFSE proton density weighted with fat saturation (TR 3500 ms; TE 43 ms; FOV 20; matrix: 320 × 256; slice thickness 4 mm; interslice gap 0.4 mm).

  5. 5.

    A three-dimensional (3D) DPFSE “CUBE” (TR 2500 ms; TE 32 ms) proton density weighted and a 3D T2 FSE “CUBE” proton density weighted with fat saturation (TR 2500 ms; TE 80 ms) in order to improve, respectively, the visualization of the intra-articular graft and the presence of synovial fluid at the bone–graft interface. These 3D sequences were performed on a sagittal plane with slice thickness of 0.8 mm, interslice gap 0 mm and high-resolution matrix 384 X 384.

We reformatted in postprocessing phase the 3D images on sagittal and coronal planes oriented along the axis of the graft.

Surgical techniques

Preliminary arthroscopic evaluation was performed by antero-lateral (AL) and antero-medial (AM) portals, under general or peripheral anesthesia, with the use of a tourniquet. Associated meniscal lesions were addressed and treated, and the ACL lesion was confirmed. ACL remnants were carefully debrided.

Study group: HG tibial attachment preservation [1, 13]

Semitendinosus and gracilis tendon were harvested maintaining their tibial insertion. The harvested tendons were then sutured together using 2 nonabsorbable No. 2 stitches.

For the tibial tunnel creation [1], a reamer was inserted from the antero-medial portion of the tibial metaphysis approximately 1 cm medial and 1 cm proximal with respect to the tibial insertion of hamstrings. The reamer emerged at the level of the ACL tibial footprint.

For the femoral socket preparation [13], a second-generation retrograde drill was used (FlipCutter, Arthrex Ltd). The FlipCutter was inserted approximately 1 cm anterior to the posterior border of the iliotibial tract and 1.5 cm proximal to the lateral femoral epicondyle, in order to drill a retrosocket of at least 25 mm, according to the width of the lateral femoral condyle.

The graft (average size 9 mm) was fixed on the lateral femoral condyle with a second-generation cortical suspensory device (TightRope RT; Arthrex Ltd, Naples, Florida, USA). With the knee at 90° of flexion, femoral fixation was achieved pulling the femoral pull suture. With the knee at 30° of flexion, the graft remnant was then fixed with a titanium staple placed at the level of the tibial metaphysis distally with respect to the hamstrings insertion (Fig. 1).

Fig. 1
figure 1

Schematic illustrations showing the creation of a complete tibial tunnel and the drilling of a femoral socket using a second-generation retrograde drill in the study group. Hamstrings tibial insertion is preserved, and the HG is fixed with a second-generation cortical suspensory device (TightRope RT; Arthrex Ltd, Naples, Florida, USA) on the femoral side, while tibial fixation is achieved with a titanium staple

Full size image

Control group: all-inside reconstruction with HG tibial detachment [2, 13]

The semitendinosus tendon was harvested without preserving its tibial insertion. The tendon was then quadrupled, loaded with 2-second generation cortical suspensory devices (ACL TightRope RT; Arthrex Ltd, Naples, Florida, USA) and sutured according to the technique described by Lubowitz et al. [14]. The center of the graft was marked.

Subsequently, both the tibial and the femoral socket were created using an outside–in technique with a retrograde drill (FlipCutter, Arthrex Ltd) [13, 14] as previously described for the creation of the femoral socket in the study group. Once correctly positioned, the graft (average size 9 mm) was fixed both on the femoral and on the tibial side by means of the ACL TightRope RT. With the knee in full extension, the graft was tensioned symmetrically on both sides. After 5 cycles of knee flexion–extension, graft tensioning was checked again (Fig. 2).

Fig. 2
figure 2

Schematic illustrations showing the creation of both tibial and the femoral sockets in a retrograde fashion using a second-generation retrograde drill in the control group. HG is detached from its tibial insertion, and it is fixed on both sides with a second-generation cortical suspensory device (TightRope RT; Arthrex Ltd, Naples, Florida, USA)

Full size image

Postoperative rehabilitation

After surgery, both groups underwent the same rehabilitation program. A rigid extension brace was positioned postoperatively to avoid knee flexion contracture. The day after surgery, all patients began continuous passive motion (CPM): The degree of joint movement allowed on the first day was between 0° and 40°. During the 2 weeks after surgery, patients were allowed to load the limb progressively, using the brace and two forearm crutches, aimed to obtain a range of motion between 0° and 100°. After 2 weeks, full weight bearing was allowed without the brace. From the third to the fifth week, with the help of a physical therapist, all patients began rehabilitation reaching the complete range of motion and performing closed-chain kinetic exercises aimed to increase the quadriceps strength. From the fifth to the ninth week, quadriceps strength was progressively increased performing open-chain kinetic exercises associated with proprioception recovery. Straight running was allowed 3 months after surgery, while return to competitive sports at 6 months after surgery.

Statistical analysis

All continuous data were expressed in terms of the mean and the standard deviation of the mean; the categorical data were expressed as frequency and percentages. The Kolmogorov–Smirnov test was performed to test normality of continuous variables. The Levene test was performed to assess homoscedasticity of the continuous variables. The ANOVA test was performed to assess the differences between groups of continuous, normally distributed and homoscedastic data, and the Mann–Whitney test was used otherwise. The repeated-measures general linear model (GLM) with Sidak test for multiple comparisons was performed to assess the differences at different follow-up times. The Kendall Tau correlation was used to assess correlation between ordinal data. For all tests, p < 0.05 was considered significant. All statistical analysis was performed using SPSS v.19.0 (IBM Corp., Armonk, NY, USA).

Results

Clinical findings

In our series, we enrolled 32 males and 8 females with a mean age of 27.5 (range 18–49). All the patients were available at the final follow-up.

No intra-operative complications were observed. In the study group, we observed 6 lesions of the medial meniscus, 2 lesions of lateral meniscus and 6 bilateral meniscal lesions. In the control group, we experienced 3 lesions of the medial meniscus, 1 lesion of the lateral meniscus and 2 cases of bilateral meniscal involvement. All the meniscal lesions were addressed with partial meniscectomy. One patient in the control group who underwent a medial meniscal suture required further surgery at 8-month follow-up to perform a partial meniscectomy. We experienced one ACL traumatic failure in the control group 6 months after surgery.

At 3-month follow-up, the average IKDC subjective score improved from 54.3 (range 39–66) preoperatively to 60.5 (range 37–82) in the study group (p < 0.05) and from 53.4 (range 41–62) preoperatively to 66.4 (range 43–84) in the control group (p < 0.05). At 6-month follow-up, the mean IKDC subjective score improved to 77.2 (range 61–95) in the study group and to 82.4 (range 56–98) in the control group, while at 12-month follow-up it was 89.6 (range 72–100) in the study group and 93 (range 62–100) in the control group. At the final follow-up of 24 months, the IKDC subjective score was 91.6 (range 75–100) in the study group and 93.9 (range 62–100) in the control group. Improvement of the IKDC score at each follow-up was found to be statistically significant with respect to the preoperative situation in both groups. The control group showed a better improvement in the IKDC subjective score at 3 and 6 months (respectively, p = 0.058 and p = 0.061) compared to the study group, but no stastically significant relationship was found (Fig. 3).

Fig. 3
figure 3

Improvement of the subjective International Knee Documentation Committee (IKDC) score at 3-, 6-, 12- and 24-month follow-up in both the study and control groups

Full size image

The ACL failure in the control group was excluded from the objective analysis at 6-month follow-up. The IKDC objective evaluation performed at 6-month follow-up was normal (A) in 11 cases and nearly normal (B) in 9 cases in the study group while resulted normal (A) in 16 cases, nearly normal (B) in 3 cases in the control group. At 12-month follow-up, the IKDC objective score in the study group resulted normal (A) in 14 cases, nearly normal (B) in 6 cases, while it was normal (A) in 17 cases and nearly normal (B) in 2 cases in the control group. At the final follow-up of 24 months, the IKDC objective score in the study group was normal (A) in 17 cases and nearly normal (B) in 3 cases, while it was normal (A) in 17 cases and nearly normal (B) in 2 cases in the control group. Among the patients rated as B according to the IKDC objective score, stabilometric parameters (Lachman and pivot shift tests) resulted normal (A) in all the patients of both groups at 6-, 12- and 24-month follow-up evaluation.

The mean preoperative Tegner scale value was 2.8 (range 1–3) in the study group and 3.0 (range 1–4) in the control group. The Tegner score was 6.3 (range 4–7) and 6.4 (range 4–7), respectively, at 12 and 24 months after surgery in the study group, while in the control group the Tegner score was 6.6 both at 12- and 24-month follow-up. Regarding the activity level evaluated with the Tegner score, significant differences were found according to the scores preoperatively and postoperatively (p < .05) in each of the groups.

Regarding the resumption of sport activities at final follow-up, in the study group 17 patients resumed sport activity at preinjury level and 3 patients resumed at a lower level. In the control group, 16 patients resumed sport at preinjury level, while 3 patients resumed the same activity at a lower level.

MRI findings

The first parameter studied was the graft ligamentization by evaluating graft signal intensity. Hypointense signals were found in 20 and 5 % of the cases, respectively, in the study group and in the control group, while isointense signals were found in 65 % of the cases in the study group and in 55 % of the cases in the control group. Hyperintense signals were found in 15 and 40 % of the cases, respectively, in the study group and in the control group. At 6-month follow-up, the mean ligamentization score was 2.1 ± 0.6 in the study group and 1.7 ± 0.6 in the control group (p < 0.05). In detail regarding the ligamentization score, 11 patients gained 1 point (3 in the study group and 8 in the control group), 24 patients reached 2 points (13 in the study group and 11 in the control group), and 5 patients had 3 points (4 in the study group and 1 in the control group). A statistically significant improvement regarding the morphology of the intra-articular portion of the graft was observed in the study group (Tau = 0.313, p = 0.024).

The second parameter studied was the graft integration at the level of the tibial tunnel evaluated by the presence or absence of synovial fluid at the tendon–bone interface. Absence of synovial fluid at the interface was founded in 55 % of the cases in the study group against the 50 % of the cases observed in the control group. The corresponding mean scores were 1.6 ± 0.5 in the study group and 1.5 ± 0.5 in the control group.

The sum of the mean values of both the parameters analyzed was 3.7 in the study group and 3.2 in the control group. Good ligamentization (3–5 points) was found in 90 and 75 % of the cases, respectively, in the study and control group (Figs. 4, 5).

Fig. 4
figure 4

Sagittal MRI images of a 33-year-old patient (study group) showing the absence of synovial fluid at the bone–tendon graft interface and the hypointense signal of the intra-articular portion of the graft

Full size image
Fig. 5
figure 5

Sagittal MRI images of a 30-year-old patient (control group) showing the absence of synovial fluid at the bone–tendon graft interface and an inhomogeneous signal with focal areas of hyperintensity of the intra-articular portion of the graft

Full size image

Discussion

The enhancement of the ligamentization process is a crucial topic in sports medicine. The reduction of the detrimental effects of graft necrosis in the early postoperative period may permit a more aggressive rehabilitation protocols aimed to improve the functional recovery after ACL reconstruction [1, 3]. Some attempts to enhance this process have been described, such as native ACL remnant preservation and platelet-rich fibrin or bone marrow-derived cell administration at the time of ACL reconstruction [15–17]. However, even though these procedures may play a role in accelerating graft maturity with respect to traditional ACL reconstruction, there is no definitive proof of clinical outcome enhancement in ACL surgery. Some authors pointed out a faster integration of the ACL graft with hamstrings tibial insertion preservation [10, 18]. Sparing hamstring tibial insertion permits to maintain the neurovascular network [9] and to improve proprioception after ACL reconstruction [19]. In addition, the maintenance of the vascular supply plays a crucial role in maintaining tendon viability preventing the phase of graft necrosis, thus leading to a faster ligamentization. This hypothesis has been evaluated in the rabbit model by Papachristou et al. [18] that observed that the semitendinosus tendon autograft retained its viability when harvested without detachment of its peripheral insertion. Based on these results, they concluded that retaining the tibial insertion of the semitendinosus autograft seems to preserve its viability, thus avoiding the avascular necrosis and the subsequent revascularization, which occur with the use of a free tendon autograft.

To our knowledge, the present study is the first trying to evaluate the effects of retaining HG tibial insertion in the humans. The preliminary finding of the present study is that hamstring tibial insertion preservation seems to be able to enhance graft ligamentization with a better morphology of the intra-articular portion of the graft, with respect to a technique which encompasses the detachment of the hamstring tendons from the tibial side.

The literature reported the results of several case series dealing with HG tibial insertion preservation reporting excellent results at mid- and long-term follow-up [6, 10, 20–22]. Among these studies indirect signs of a faster graft ligamentization can be found. In particular, Marcacci et al. [3] showed the absence of osteolysis or tunnel widening after ACL reconstruction with inserted HG at 11-year follow-up. Furthermore, Zaffagnini et al. [22], in their comparative study, showed a lower rate of tunnel enlargement following ACL reconstruction with HG tibial insertion preservation with respect to 4-strand HG reconstruction with tibial detachment.

In the present study, a cohort of patients operated by an all-inside reconstruction represented the control group. The presence of a socket on the tibial side instead of a complete tunnel may be seen as a bias, but this kind of reconstruction was adopted to avoid the employment of an interference screw that may disguise the evaluation of graft integration on the tibial side. The MRI evaluation showed a better morphology of the intra-articular portion of the graft in the study group, deposing for a better ligamentization. No differences between the 2 groups regarding graft integration on the tibial side were observed. This finding may be presumptively related to the absence of the interference screw also in the group with detached tendon with consequent maximization of the tendon bone contact and subsequent better integration.

The clinical results obtained in the study were excellent in both groups. No differences were observed regarding clinical and stabilometric parameters between the 2 groups. A tendency toward a better improvement in the subjective clinical score at 3- and 6-month follow-up was detected in the control group. This tendency was lost at both 12- and 24-month follow-up. This finding, in contrast with the MRI results, may be explained by the characteristics of the all-inside reconstruction. This mini-invasive procedure encompasses a low morbidity for the patient sparing bone on the tibial side and preserving the gracilis tendon resulting in a better early subjective outcome for the patients as pointed out by Lubowitz et al. [14]. The author also reported a statistically significant lower pain in the early postoperative period following the all-inside procedure with respect to traditional ACL reconstruction [23].

The present study has some limitations. First of all, the number of patients enrolled is quite low to draw definitive conclusions about the researched topic even if the sample was calculated performing a power analysis. Secondarily, the MRI evaluation was carried out only at 6 months. An intermediate MRI evaluation performed at 3 months should better highlight the progression of the ligamentization process thus permitting a better understanding of the supposed beneficial effects of HG tibial insertion preservation on graft biology. Nevertheless, a 3-month MRI analysis is frequently affected by artifacts, thus not ensuring a reliable evaluation of graft’s ligamentization and integration. Finally, the employed score described by Figueroa should be integrated with other parameters in order to better define graft morphology.

In conclusion, according to the results obtained, HG insertion preservation seems to have a beneficial effect on graft ligamentization. Further randomized clinical trials with a longer follow-up with higher numbers of patients, more seriate MRI evaluations and more complete MRI scores are required in order to validate these preliminary findings.