Effects of higher femoral tunnels on clinical outcomes, MRI, and second-look findings in double-bundle anterior cruciate ligament reconstruction with a minimal 5-year follow-up

Abstract

Background:

To perform anatomical anterior cruciate ligament reconstruction (ACLR), tunnels should be placed relatively higher in the femoral anterior cruciate ligament (ACL) footprint based on the findings of direct and indirect femoral insertion. But the clinical results of higher femoral tunnels (HFT) in double-bundle ACLR (DB-ACLR) remain unclear. The purpose was to investigate the clinical results of HFT and lower femoral tunnels (LFT) in DB-ACLR.

Methods:

From September 2014 to February 2016, 83 patients who underwent DB-ACLR and met the inclusion and exclusion criteria were divided into HFT-ACLR (group 1, n = 37) and LFT-ACLR (group 2, n = 46) according to the position of femoral tunnels. Preoperatively and at the final follow-up, clinical scores were evaluated with International Knee Documentation Committee (IKDC), Tegner activity, and Lysholm score. The stability of the knee was evaluated with KT-2000, Lachman test, and pivot-shift test. Cartilage degeneration grades of the International Cartilage Repair Society (ICRS) were evaluated on magnetic resonance imaging (MRI). Graft tension, continuity, and synovialization were evaluated by second-look arthroscopy. Return-to-sports was assessed at the final follow-up.

Results:

Significantly better improvement were found for KT-2000, Lachman test, and pivot-shift test postoperatively in group 1 (P >0.05). Posterolateral bundles (PL) showed significantly better results in second-look arthroscopy regarding graft tension, continuity, and synovialization (P <0.05), but not in anteromedial bundles in group 1. At the final follow-up, cartilage worsening was observed in groups 1 and 2, but it did not reach a stastistically significant difference (P >0.05). No statistically significant differences were found in IKDC subjective score, Tegner activity, and Lysholm score between the two groups. Higher return-to-sports rate was found in group 1 with 86.8% (32/37) vs. 65.2% (30/46) in group 2 (P = 0.027).

Conclusion:

The HFT-ACLR group showed better stability results, better PL, and higher return-to-sports rate compared to the LFT-ACLR group.

Introduction

Anatomical anterior cruciate ligament reconstruction (ACLR) is getting popular and the restoration of native anatomy is considered as one of the final goals in current ACL surgery.^[1,2,3] To reproduce native ACL kinematics, femoral tunnel placement has been emphasized and is one of the critical points in the surgery.

It is technically more difficult to evaluate the position of the femoral tunnel than to determine the tibial tunnel position, especially the position in double bundle (DB) reconstruction.^[4,5] However, studies on femoral ACL footprint morphology and tunnel position vary according to the measurement methods used.^[6,7] Studies showed that there are two different ACL fibers at the femoral insertion: the direct and indirect fibers.^[8,9] The femoral ACL tunnel should be customized based on the graft type and fixation device to ensure that the graft is positioned to cover the central direct ACL fibers. Iriuchishima et al^[10] reported that the center position of the anteromedial (AM) and posterolateral (PL) bundles in the femoral ACL footprint was significantly different depending on the inclusion or exclusion of the fan-like extension fibers. Biomechanical study showed that ACL fibers located high within the femoral footprint beared more force during stability testing and were more isometric during flexion than low fibers.^[11] Pathare et al^[12] found that the indirect femoral ACL insertion contributes minimally to the restraint of tibial translation and rotation, and femoral tunnel positioning for anatomic ACLR should aim to recreate the biomechanically significant direct insertion in a cadaveric study. Clinically, the femoral tunnels should be oriented to cover the entirety of the central direct ACL fibers,^[7] and be placed relatively shallow and high.^[10,11,13] Currently, limited studies have been reported about the clinical results of making the femoral tunnels on a relatively higher position in DB-ACLR.

The purpose of this study was to compare the clinical results, second-look arthroscopic findings, and magnetic resonance imaging (MRI) findings between femoral tunnels at higher femoral tunnels (HFT) and lower femoral tunnels (LFT) in DB-ACLR at a minimal 5-year follow-up. The hypothesis was that the clinical results, second-look arthroscopic findings, and MRI findings would be different between the two techniques.

Methods

Patient selection

This retrospective trial was approved by the Ethics Committee of Peking University Third Hospital (No. IRB00006761-2015158). All procedures performed in studies were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

From September 2014, we began to place femoral tunnels at the direct insertion of the femur in DB-ACLRs. We explained the difference between HFT and LFT DB-ACLR to the patients and let them decide to apply HFT or LFT DB-ACLR. Considering a minimum follow-up duration of 5 years, patients from September 2014 to February 2016 were enrolled. All patients signed an informed consent form.

During this time period, 127 primary DB-ACLRs in 127 patients (96 men and 31 women) were performed consecutively by a senior author (Jiakuo Yu). The inclusion criteria were (1) unilateral ACL ruptures with surgery from initial injury less than 6 months; (2) age older than 18 years with closed physes and younger than 45 years. The exclusion criteria were (1) grade 3 or higher degenerative cartilage changes according to the Outerbridge classification; (2) subtotal or total meniscectomy; (3) multiple ligament injuries; (4) previous surgery of the involved knee; (5) two patients lost to follow up; (6) two patients who had revision ACLR. Patients were selected by applying the selection criteria [Figure 1]. Femoral tunnel positions were confirmed by 3 dimesion-computerized tomography (3D-CT) postoperatively.

Figure 1.

Flowchart of ACLR patient selection. ACLR: Anterior cruciate ligament reconstruction; DB-ACLR: Double bundle anterior cruciate ligament reconstruction; MRI: Magnetic resonance imaging.

Surgical procedure

The hamstring autograft tendons were harvested from the affected limb through an oblique 4-cm skin incision over the medial aspect of the proximal tibia. The semitendinosus and gracilis tendons were harvested and used to make the AM and PL bundle grafts, respectively. Triple or quadruple strands of the semitendinosus tendon were used for the AM bundle graft (7–8 mm), and triple or quadruple strands of the gracilis tendon were used for the PL bundle graft (5–6 mm).

Then, the femoral tunnel was placed for the AM bundle and the PL bundle according to the osseous landmarks and arthroscopic findings of the ACL footprint. The lateral intercondylar ridge, the lateral bifurcate ridge, and landmarks of the posterior articular cartilage were used to identify the femoral footprint.^[14,15] For LFT-ACLR, the center of the femoral footprints (including direct and indirect insertion) of both the AM and PL bundles were identified and marked with a thermal device, as described in the previous study [Figure 2A]. The direct fiber insertion point was located on the extension line of the posterior femoral cortex on the medial side of the lateral femoral condyle. For HFT-ACLR, the tunnels were placed relatively shallow and high in the center of direct insertion of the femoral ACL footprint which were just behind the resident ridge [Figure 2B]. Femoral tunnels were created with the transportal technique through the anteromedial portal, using the divergent direction between both AM and PL tunnels to avoid femoral tunnel communication. The tibia tunnel was drilled using a tibial drill guide system (Smith & Nephew Endoscopy, Andover, MA, USA) with a 5-mm to 8-mm diameter reamer based on the individualized tibial insertion of ACL [Figure 3]. Each graft was fixed with an Endobutton (Smith & Nephew Endoscopy) at the femoral site, and fixed with one bio-absorbable interference screw (Smith & Nephew Endoscopy) and a staple at the extension of the knee at the tibial site.

Figure 2.

(A) A 24-year-old male patient underwent LFT in DB-ACLR. (B) A 26-year-old male patient underwent HFT in DB-ACLR. DB-ACLR: Double bundle anterior cruciate ligament reconstruction; HFT: Higher femoral tunnels; LFT: Lower femoral tunnels.

Figure 3.

Representative MRI images of positions of tibial tunnels for double-bundle anterior cruciate ligament reconstruction. MRI: Magnetic resonance image.

Rehabilitation

All patients followed a standardized rehabilitation procedure. On the first day postoperatively, quadriceps sets, straight-leg raises, and prone hangs were started. All of the patients were permitted to walk with crutches and braces but without bearing weight. The range of motion (ROM) was begun from Day 3 postoperatively. Closed kinetic-chain exercises and full weight-bearing were started in the sixth week. Patients progressed to running without braces at 4–6 months. Sporting activities were allowed approximately 6 months after surgery.

Clinical evaluations

All of these patients were monitored with a defined research protocol after the operation. Subjective evaluation included preoperative and postoperative International Knee Documentation Committee (IKDC), Lysholm, and Tegner scores. Objective evaluation included ROM and side-to-side difference on KT-2000 arthrometer at 30 lbs and 30˚ of flexion. The manual Lachman test was estimated by an independent examiner with the patient’s knee in 30˚ of flexion and was graded as 0 (–1 to 2 mm), 1 (3–5 mm), 2 (6–10 mm), or 3 (>10 mm) compared with the uninjured, contralateral knee.^[16] The pivot-shift test was clinically graded as 0 (negative), 1 (glide), 2 (clunk), and 3 (locking) according to IKDC guideline.^[17] Return-to-sport status was investigated. All the physical examinations were performed by the senior author (Jiakuo Yu) who had more than 20 years of sports medicine practice and was blinded to group. All clinical assessments (IKDC, Lysholm, and Tegner scores) were performed by one blinded and independent examiner who had more than 10 years of sports medicine practice.

3D-CT evaluation and classification of the patients

All patients underwent 3D-CT within 3 days after the operation in our hospital. The femoral tunnel positions were measured using the digital radiography system (PACS, GE, Milwaukee, Wisconsin, USA) with the built-in digital rule. The tunnel position was described in terms of the quadrant method on the 3D-CT reconstruction image described in previous studies.^[6,18] The positions of femoral tunnels were evaluated by two surgeons who were blind to surgical technique, clinical, and other radiographic outcomes based on 3D-CT to assess the appropriateness of classification. The tunnel positions were determined in the high-low position, perpendicular to the Blumensaat line, and the deep-shallow position, parallel to the Blumensaat line.

Evaluation of cartilage injury by MRI preoperatively and postoperatively

A standard institutional MRI examination with a 1.5-T superconducting magnet (GE Signa HDx, Milwaukee, Wisconsin, USA) was performed. The MRI protocols included coronal, sagittal, and axial sequences. Each sequence included the T1- and T2-weighted phases (echotime, 25–30 ms; repetition time, 4000–6000 ms; slice thickness, 3.5 mm with no gap). Preoperative MRI was performed within 3 months before surgery, and MRI images were obtained at the latest follow-up for each subject. The evaluation scale for cartilage injury was completely in accordance with the International Cartilage Repair Society (ICRS) grade, and the most seriously affected part was included in the statistical analysis as categorical data.^[19] The two observers determining cartilage injury were blinded to clinical and other radiographic outcomes.

Second-look arthroscopy

Second-look arthroscopy was performed at least 2 years after ACLR. These patients came for the metal staple removal because of tenderness caused by the staple, which could also be seen at kneeling. Graft continuity was graded as follows: no tear, superficial tear, or substantial tear.^[20,21] Graft tension was graded as taut, mildly lax, and lax by probing at 20˚–90˚ of knee flexion. The graft was considered taut if it was as tense as the normal ACL throughout the previously mentioned range; mildly lax if it had less tension with redundancy; and lax if there was an obvious loss of tension.^[21] Synovial coverage over the grafts was graded as good (coverage >80% around graft), fair (coverage ≥50%), and poor (coverage <50%).^[22] The evaluations were performed by two surgeons (one consultant surgeon and one attending surgeon) at second-look arthroscopy, where the conclusion was only achieved at the agreement of both. If several different grades were found, scoring was based on the most severe one.

Statistical analysis

Reliability of the measurements was assessed by examining intra- and inter-observer reliability using the intra-class correlation coefficient (ICC). The ICCs for intra-observer and inter-observer reliability were >0.8. Pre- and postoperative demographic data, pre- and post-operative clinical scores, second-look arthroscopic findings, and postoperative MRI findings and measurements from the two groups were compared by use of two-sample t-tests, paired t-tests, Mann–Whitney tests, Wilcoxon signed rank tests, Pearson chi-squared tests, and Fisher’s exact tests, as appropriate. Significance was set at P <0.05. Statistical analysis was performed with SPSS software (version 24.0; SPSS, Chicago, IL, USA) for windows.

Results

Patient demographic data and clinical characteristics

There were 37 patients who received HFT of the DB-ACLR technique (group 1) and 46 patients with LFT of the DB-ACLR technique (group 2). There were 31 patients in group 1 and 35 patients in group 2 who underwent second-look arthroscopy for the metal staple removal. There was no statistically significant difference regarding age, gender, body mass index (BMI), period from injury, meniscus management, and follow-up duration between the two groups [Table 1]. In terms of the diameter of graft, no significant difference was found between groups 1 and 2 with AM: 7.6 ± 0.5 mm (7.0–8.0 mm) vs. 7.5 ± 0.5 mm (7.0–8.0 mm), PL: 5.7 ± 0.5 mm (5.0–6.0 mm) vs. 5.6 ± 0.4 mm (5.0–6.0 mm), respectively. For the length of the graft, there was no significant difference between groups 1 and 2 with AM: 8.0 ± 0.7 cm (7.0–9.0 cm) vs. 8.3 ± 0.6 cm (7.0–9.0 cm), PL: 6.7 ± 0.5 cm (6.0–7.5 cm) vs. 6.4 ± 0.6 cm (6.0–7.5 cm), respectively.

Table 1.

Patient demographic data and clinical characteristics, pre- and intra-operation of ACLR.

Items	Group 1 (n = 37)	Group 2 (n = 46)	Statistic values	P value
Age (years)	26.5 ± 7.8 (18.0–41.0)	28.0 ± 7.5 (18.0–42.0)	–1.060*	0.416
Gender			0.301†	0.583
Female	7	11
Male	30	35
BMI (kg/m2)	24.8 ± 3.3 (18.9–29.4)	23.6 ± 2.9 (19.8–31.5)	–0.990*	0.483
Period from injury to operation (days)	76.3 ± 54.7 (2.0–167.0)	85.2 ± 61.2 (1.0–170.0)	0.222*	0.357
Meniscus status
Lateral meniscus			2.276†	0.320
Intact	26	27
Repaired	2	7
Partial	9	12
Medial meniscus			2.449†	0.294
Intact	18	25
Repaired	4	9
Partial	15	12
Clinical follow-up duration (months)	64.2 ± 6.9 (60.0–72.0)	64.1 ± 5.4 (60.0–72.0)	2.027*	0.474

Data are expressed as n, mean ± standard deviation (range). Group 1: Higher femoral tunnels group; Group 2: Lower femoral tunnels group. ^*t-tests; ^†Pearson chi-squared tests. ACLR: Anterior cruciate ligament reconstruction; BMI: Body mass index.

CT evaluations

Evaluation of tunnel placements was shown in Supplement Table 1, http://links.lww.com/CM9/B821. The center of the AM tunnel was significantly different in depth (t = –5.514, P <0.001) and not different in height (t = –2.506, P = 0.127), and the center of the PL tunnel was significantly different in depth (t = –6.115, P <0.001) and height (t = 4.853, P <0.001) between the two groups. Typical femoral tunnel positions for AM bundle and PL bundle were shown in Figure 2A for LFT and Figure 2B for HFT.

Clinical results

With regard to the stability of the knee joint measured by KT-2000, Lachman test, and pivot-shift test, significantly better results were found in stability tests in group 1 post-operatively [Supplement Table 2, http://links.lww.com/CM9/B821]. No significant differences were found in the IKDC score, Tegner activity, and Lysholm score between the two groups [Supplement Table 3, http://links.lww.com/CM9/B821]. Most of the patients in the two groups had returned to sports, the rate of return-to-sports in group 1 was significantly higher than that in group 2 [Supplement Table 3, http://links.lww.com/CM9/B821]. Representative MRI images after HFT-ACLR were shown in Figure 4.

Figure 4.

MRI findings of double bundle anterior cruciate ligament reconstruction. (A) The MRI findings of the right knee of a male patient showed low signal intensity and good continuity for the anteromedial bundle and posterolateral bundle. (B) The MRI findings of the right knee of a male patient showed intermediate signal intensity for the anteromedial bundle and high signal intensity posterolateral bundle. AM: Anteromedial bundle; MRI: Magnetic resonance imaging; PL: Posterolateral bundle.

Second-look arthroscopy

The mean period from ACL reconstruction to the second-look arthroscopic examination was 32.1 ± 8.7 months (range, 24.0–44.0 months). For PL bundles, graft tension, graft continuity, and synovialization were significantly different (P <0.05) between two groups, and no significant difference was found for AM bundles [Supplement Table 4, http://links.lww.com/CM9/B821]. Representative second-look arthroscopic findings after HFT DB-ACLR were shown in Figure 5.

Figure 5.

Second-look arthroscopic findings of double bundle anterior cructiate ligament reconstruction. (A) Second-look arthroscopic findings of the right knee of a male patient showed no tear, taut graft tension, and good synovial coverage for the AM bundle and PL bundle. (B) Second-look arthroscopic findings of the right knee of a male patient showed good synovial coverage for AM bundle, mild laxity, and fair synovial coverage for the PL bundle. (C) Second-look arthroscopic findings of the left knee of a male patient showed poor synovial coverage for AM bundle and superficial tear and poor synovial coverage for the PL bundle. AM: Anteromedial; PL: Posterolateral.

ICRS grade for cartilage injury based on MRI before ACLR and at final follow-up

Based on preoperative MRI, there was no significant difference in grades of ICRS between groups 1 and 2. At the final follow-up, cartilage worsening was seen in the groups 1 and 2, but it did not reach a statistically significant difference [Supplement Table 5, http://links.lww.com/CM9/B821].

Discussion

The most important finding is that the HFT DB-ACLR group showed better stability results according to the KT-2000, Lachman test, and pivot-shift test, with better PL bundles on second-look arthroscopy. Although no significant differences were found between these two techniques in the IKDC score, Tegner activity, and Lysholm score, higher return-to-sports rate was found in the HFT-ACLR group.

Better results in postoperative graft failure rate and rotational laxity for DB-ACLR were reported in several clinical and biomechanical studies.^[23,24] However, clinical scores and subjective outcomes were not significantly different in other studies.^[25,26,27] Anatomic studies showed that the femoral insertion of ACL comprises direct and indirect fibers, and the direct fiber insertion was located close to the resident ridge in a ribbon-like shape.^[9,10] Biomechanical studies showed that ACL fibers located high within the femoral footprint bear more force during stability testing and are more isometric during flexion than low fibers, and it might be advantageous to create a “higher” femoral tunnel during ACL reconstruction at the lateral intercondylar ridge.^[7,10,11] These findings potentially explain why the traditional DB-ACLR has no significant advantages than single bundle ACL reconstruction, in which the tunnels located in the center of footprints are more inclined to indirect fibers.^[13] In the present study, our results demonstrated that patients in the HFT group had significantly better stability and higher return-to-sports rate in a minimal 5-year follow-up.

The relationship between the degree of knee laxity after ACLR and patient-reported knee function remains unclear. Previous work has generally noted poor correlation between residual anterior laxity after ACLR and patient-reported outcomes.^[28,29] The relationship between residual rotational knee laxity and patient-reported outcomes after ACLR is important but has been studied less frequently. A series study performed by Magnussen et al^[30,31,32] demonstrated that the presence of a residual side-to-side KT-1000 arthrometer difference <6 mm or pivot glide at 2 years after ACLR is not associated with an increased risk of subsequent ipsilateral knee surgery or decreased patient-reported outcomes up to 6 years after ACLR. Conversely, patients exhibiting a difference >6 mm in side-to-side anterior laxity were noted to have significantly decreased patient-reported outcomes at 6 years after ACLR. Recently, Rahardja et al^[33] reported that higher rate of return to sports was found in patients with bone-patellar tendon-bone compared with those with hamstring tendon autograft. However, the Knee injury and Osteoarthritis Outcome Score (KOOS) (pain, symptoms, activities of daily living, and quality of life) was not significantly different.^[33] Although significant differences were noted in femoral tunnel positions, better stability, and higher return-to-sports rate in HFT DB-ACLR, these differences did not appear to be clinically significant in terms of clinical scores (IKDC score, Tegner activity, and Lysholm score) in our series. Mean side-to-side KT-2000 arthrometer difference was 1.2 mm and 2.6 mm in groups 1 and 2, respectively. One possible reason was that current clinical scores might not be subtle enough to make a significant difference.

It was believed that the PL bundle was at higher risk of failure because more tension was applied to the PL graft at full knee extension.^[34,35] Furthermore, unlike AM bundle graft, the PL bundle is not an isometric graft, it underwent greater length change for it slackens as the knee flexes.^[36] The change in length will impair the PL bundle’s total remodeling process.^[36,37] Several studies showed that more tears, poor synovial coverage, and inferior graft maturity of PL bundles were observed compared with AM bundles during second-look arthroscopy and MRI.^[37,38,39] Femoral tunnel positioning for anatomic ACL reconstruction should aim to recreate the biomechanically significant direct insertion.^[12] Forsythe et al^[40] found that the femoral tunnel location for ACL reconstruction with the least amount of length change through the ROM should encompass the direct fibers of the ACL in cadaveric knees. In the present study, better PL bundles in HFT DB-ACLR were found in second-look arthroscopy. These findings could be explained by the biomechanical advantages of the direct insertion of PL. Thus, the PL bundle might have less tension and length change when the knee joint is at full extension.

In the present study, the heights of AM femoral tunnel in the HFT group were higher than those in the LFT group, but they did not reach a statistical difference. During the surgery, the AM tunnels were created to be higher than the traditional AM tunnels and still within the anatomic footprint. That could be the reason why there was no significant difference in terms of AM tunnel height between the HFT and LFT groups.

Increased laxity may increase the stress borne by other intraarticular structures, potentially increasing the risk of subsequent meniscal tears and articular cartilage damage. Such damage may increase the risk of subsequent knee surgery and poorer patient reported outcomes (PROs) in patients with increased residual laxity.^[41,42] Articular cartilage status is an important indicator of ACLR outcomes.^[19,43] ACLR could not completely prevent cartilage lesions, especially in the patellofemoral joint, although it can help restore stability of the knee joint after operation.^[44,45] Culvenor et al^[46] reported the patellofemoral compartment seems to be at particular risk for early osteoarthritis after ACLR, especially in men. It is reported that the median prevalence of patellofemoral joint osteoarthritis at 10–15 years after ACL reconstruction was almost 50%.^[47] Li et al^[48] found that 39% of patients had radiographic osteoarthritis in an average of 7.8 years after single-bundle ACLR. Abnormal patellar rotation and tilt, inflammation, and concomitant damage after ACL injury and reconstruction might make the patellofemoral joint vulnerable to degeneration, which could explain that cartilage lesions were mainly observed on the patellofemoral joint after ACLR.^[47,49] Our results demonstrated that cartilage worsening was seen in both groups, but it did not reach a significant difference at the final follow-up. Longer-term follow-up is needed to clarify the deterioration of cartilage with time in these two groups.

This study had some limitations. The first limitation is that our study was retrospective, which could introduce patient selection bias in group selection, although the patients were in consecutive series. Patients who are selected for DB-ACLR have higher activity-level demands and return-to-sports rate. The second limitation is that some variations in the difference in graft tension at fixation, tibial tunnel locations, and other additional cofactors might exist, although many possible co-factors were excluded when the patients were enrolled. The third limitation is that the sample size was relatively small and further prospective study with more patients is needed. Another limitation is that more sensitive methods like T2 mapping should be used to investigate cartilage changes of different DB-ACLR techniques in further study.

In conclusion, the HFT-ACLR group showed better stability results according to the KT-2000, Lachman test, and pivot-shift test, with better PL on second-look arthroscopy. Although there was no significant difference in clinical scores at a minimal 5-year follow-up between HFT-ACLR and LFT-ACLR groups, higher return-to-sports rate was found in the HFT-ACLR group.

Funding

This work was supported by the Beijing Natural Science Foundation of China (No. J210011), the National Natural Science Foundation of China (No. U22A2051), and the Ministerial Commission of Science and Technology (No. JK-2022-07).

Conflicts of interest

None.