Long-term follow-up after anterior cruciate ligament reconstruction using a press-fit quadriceps tendon-patellar bone autograft

The survey retrospectively included all patients operated on with a minimum of 7 years and a maximum of 9 years before the start of the study. Sixty-nine patients had received an arthroscopic reconstruction of the ACL using a QTPB autograft with hardware-free fixation and met the strict inclusion and exclusion criteria (Table 1). All patients were operated on by one senior surgeon (JH). Twenty-seven (39%) of the patients were female and 42 (61%) were male. The average age on operation day was 27 years (range: 14 to 43 years).

Six patients (9%) suffered a re-rupture of the ACL graft during the follow-up period as a result of a new trauma; these patients received a revision ACL reconstruction using a different surgical technique and were therefore excluded from the statistical analysis.

After the retrospective inclusion of the patients, we prospectively performed the collection of the data. A total of 6 patients (9%) did not attend for physical examination but all of them were reached and have completed the subjective surveys on the telephone.

The period between injury and operation was on average 4.3 months (range: 8 days to 15 months). The investigation was carried out after a mean follow-up interval of 7.5 years (range: 7.0 to 8.7) by two experienced orthopedic specialists (AB, NAS) who were not involved in the primary operation.

Sixty-four (93%) patients incurred ACL injuries during sports: football (n = 23, 36%), skiing (n = 15, 24%), handball (n = 12, 18%), basketball (n = 5, 8%), tennis (n = 3, 5%), volleyball (n = 2, 3%), rugby (n = 2, 3%), fistball (n = 1, 1.5%), gymnastics (n = 1, 1.5%). Five patients (7%) were injured during the course of everyday physical activities. Additional lesions were identified with meniscus damage in 24 (35%) patients, cartilage damage in 8 (12%) patients and meniscus damage and cartilage damage in 16 (23%) patients. Two patients (3%) had a combination of ACL rupture, rupture of the medial collateral ligament, tear of the outer meniscus and cartilage damage. Only 19 (27%) patients had an isolated ACL rupture. Since meniscus repair was an exclusion criterion, only patients with partial meniscus resection (n = 34, 50%) or stable meniscus lesions which were not addressed (n = 8, 12%), were included. Interventions on the cartilage were not carried out in the case of articular cartilage damage °I (n = 8, 12%) and °II (n = 15, 22%). Higher grade cartilage damage led to exclusion from the study [19].

ACL reconstruction was performed using a technique in which fixation of the graft and the bony filling for the tunnels can be done without the use of hardware. This is made possible by using oscillating hollow burrs (Additional file 1: Figure SA) and a special asymmetrically shaped spongiosa compactor. When harvesting the graft (Fig. 1 / Additional file 2: Figure SB), the diameter of the cylindrical patellar bone block can thereby be precisely defined to 9.4 mm (Fig. 2). The femoral press-fit anchoring is then done by using the bone block to plug the 8 mm femoral tunnel that has been drilled and slightly dilated (Fig. 3). A hollow burr is also used to create the tibial tunnel so that a spongiosa cylinder can be harvested. The tibial fixation is carried out as a hybrid technique. Blocking the tendon in the tibial tunnel with the harvested bone cylinder (Fig. 4) leads to rapid healing of the bone and the knotting of the threads reinforces the quadriceps tendon over a bone bridge (Additional file 3: Figure SC), ensuring primary stability. A detailed description of the individual steps to the surgical procedure and the standardized rehabilitation were described in 2013 [20].

A general questionnaire surveyed the patients’ data for details of the trauma mechanism and the subjective satisfaction of the patient. Patients specified their satisfaction with the result of the operation on a Visual Analog Scale (VAS) of 0 to 10 [21].

The sporting activity level of the patients was surveyed before the injury and at the time of the follow-up investigation using the Tegner Score of 0 to 10. Level 0 corresponds to an inability to work, level 5 corresponds to the ability to do more intensive physical activities such as cross-country skiing, cycling and jogging two times a week, whereas level 10 corresponds to the ability to play football on a national or international level [22]. The Lysholm Score served to evaluate the subjectively perceived impairment of knee-joint function. It is scored using a point system that evaluates the knee joint in eight everyday stress situations. By adding the points a maximum of 100 points can be achieved. A score of 91–100 points refers to a very good knee function, 84–90 points is good, 65–83 points is moderate and 0–64 points refers to a bad knee function [23]. The Form of the IKDC Score 2000 (International Knee Documentation Committee) for subjective assessment of the knee in everyday and sporting activity was also assessed. It includes seven questions about the symptoms and two questions on activities and function of the knee joint. The result is a point value in which category A (90–100 points) represents a normal knee function, category B (80–89 points) is nearly normal, category C (70–79 points) is abnormal and category D (0–69 points) represents a strongly abnormal knee function [24].

The clinical investigation was carried out using the Form for investigation of the knee for the IKDC Score 2000 [24]. The instrumental measurement of the anterior-posterior stability was carried out using the KT-1000™ Knee Ligament Arthrometer® manufactured by MEDmetric® Corporation (http://​www.​medicalproductgu​ide.​com/​companies/​1364/​medmetric_​corp, MEDmetric® Corporation, 7542 Trade Street, San Diego, CA 92121; Patent no. 4,583,555) and a tension force of 134 N. The clinical investigation was supplemented by thigh circumference measurements and the Single-Leg Triple-Hop Test [25]. Two measurements of the thigh circumference were done, 10 cm and 20 cm proximal to the joint line. Järvelä and colleagues showed a highly significant correlation (p < 0.001) between thigh measurement and the isokinetic strength test for knee extension [26].

Osteoarthritis stage was evaluated according to Kellgren-Lawrence classification [27]. Assessment of the tunnels involved X-rays of the knee taken immediately postoperatively in three planes (exercise X-ray according to Rosenberg in the posterior-anterior beam path at 25° knee bend, lateral view at 30° flexion standing and 1 patella-tangential view). The same views of X-rays were taken on follow-up and were compared with the postoperative X-rays.

We further performed a subgroup analysis in order to investigate whether different clinical variables had an impact on the postoperative results. For this purpose, the patient collective was divided into two groups and analyzed statistically. The divisions were made according to sex (Male / Female), by age (< 30 years old on operation day / ≥ 30 years old on operation day), by Body Mass Index (< 25 / ≥ 25), period between injury and surgery (< 3 months ≥3 months), Tegner score before injury (< Level 8 / ≥ Level 8), accompanying injuries in the form of meniscal injuries (presence of meniscal injury / no meniscal injury), cartilage damage (presence of cartilage damage / no cartilage damage), difference in KT-1000 measurement (≤ 2 / > 2), result of the pivot shift test (no pivoting / pivoting), measurement of the thigh circumference (thigh circumference of operated leg ≥ thigh circumference of contralateral side) / thigh circumference of operated side < thigh circumference of contralateral side), and postoperative complications (patients without cyclops or other complications / patients with cyclops or other complications).

There were no intraoperative complications. One patient (1.5%) had a wound infection postoperatively which was treated conservatively with antibiotics. One patient (1.5%) had a deep vein thrombosis which was treated with anticoagulant therapy. Seven patients (12%) experienced limited knee extension postoperatively. A consequent magnetic resonance imaging (MRI) scan identified the development of a Cyclops in all 7 cases. An arthroscopic operation with resection of the Cyclops and a notchplasty was carried out in all cases after approximately 6 months. Thereafter all patients achieved full extension of the knee joint. The 7-year results of the patients with complications differed neither in the subjective nor the objective parameters from the patients without complications.

On Follow up, 42 patients (74%) were very satisfied (VAS 9–10), 11 patients (19%) were satisfied (VAS 6–8), 3 patients (5%) were moderately satisfied (VAS 3–5) and 1 patient (2%) was dissatisfied (VAS 0–2).

The average value for the Lysholm Score at the follow-up examination was 92 ± 10.6 points (range: 60–100). Forty-two patients (74%) rated their knee function as very good (91–100 points), 5 (9%) had good knee function (84–90 points), 7 (12%) had moderate knee function (65–83 points) and 3 (5%) had poor knee function (≤ 64 points).

The average value of the scores for subjective assessment of knee-joint function for everyday activities and in sport (IKDC 2000 Form) was 90 ± 9.9 points (range: 47–100). Thirty-seven patients (65%) of the patients had normal knee function (90–100 points). Thirteen (23%) patients had almost normal knee function (80–89 points), 5 (9%) patients had abnormal knee function (70–79 points) and 2 (3%) patients had significantly abnormal knee function (≤69 points). The evaluation of question 9 of the subjective IKDC 2000 Form revealed no difficulties in kneeling or squatting in any of the patients.

Table 2 shows the results of clinical investigations using palpation and instruments based on the IKDC 2000 Form for clinical investigation of the knee. The posterior-anterior translation was measured using the KT-1000 Arthrometer and yielded an average value of 6.7 ± 1.7 mm (range: 2 to 11 mm) for the unoperated knee and an average value for the operated knee of 7.5 ± 1.7 mm (range: 3 to 12 mm). This increased sagittal translation of the operated knee from an average of 0.9 ± 1.3 mm (range: -2 mm to 4 mm) was statistically significant (p = 0.004). The circumference measured for the thigh 10 cm proximal to the base of the patella gave an average value for the operated knee of 46 ± 4.2 cm (range: 39-58 cm) and 47 ± 4.7 cm (range: 39-65 cm) for the unoperated knee. Twenty cm proximal to the patella yielded an average of 55 ± 4.7 cm (range: 45-69 cm) for measurement of the operated knee and 56 ± 4.7 cm (range: 47-75 cm) for the unoperated knee. However, the lower circumference for the operated thigh was not significant for both values (p = 0.408). In the Single-leg Triple-Hop Test, the hopping distance achieved in three consecutive single-leg hops was measured and the side-to-side difference was determined. Two patients did not want to take part in this test. The average hop distance for the operated leg was 416 ± 91 cm (range: 225-635 cm) and for the unoperated leg was 427 ± 93 cm (range: 200-635 cm). The difference was not statistically significant (p = 0.570). Furthermore, only one patient (1%) reported hypaesthesia (sensitivity to touch) at the donor site, none reported anterior knee pain. Meniscus tests were abnormal in 1 patient. This patient also had the worst functional results.

On follow-up, three patients were pregnant. As a result, the radiological examination could only be carried out in 54 (78%) patients. The X-rays were taken directly postoperatively and on follow-up for these patients. Assessment of osteoarthritis according to Kellgren-Lawrence showed deterioration by one stage in two patients (Table 2). The assessment of the drill tunnel diameter provided radiological evidence in all patients for complete bony filling of the femoral tunnel and almost complete bony filling of the tibial tunnel. No tunnel-widening occurred in any patients. None of the patients experienced a dislocation of the femoral bone block or dislocation of the tibial spongiosa cylinder.

The following differences were revealed in the subgroup analysis:

Older patients had a greater loss of activity level in the Tegner Score than younger people (p = 0.013). Patients with a longer waiting time between injury and surgery achieved a poorer score in the Pivot-Shift Test than patients with a shorter waiting time (p = 0.017). Patients who had meniscus injuries achieved a poorer score in the KT-1000 stabilization measurement than patients without a meniscus injury (p = 0.033) and their difference in the thigh circumference was greater (p = 0.028). Patients who had cartilage damage had a knee flexion deficit more frequently than patients without cartilage damage (p = 0.009). No significant differences in outcome could be identified in regards to the following parameters: sex (p = 0.74), Body mass index (p = 0.073), Tegner level of sporting activity before the injury (p = 0.334), and complications (p = 0.366).