Session 1: Sports Orthopaedics 1

Christian Fugl Hansen, Maria Østergaard Madsen, Martin Rathcke, Susan Warming, Martin Lind, Peter Faunø, Michael Rindom Krogsgaard, Karl Bang Christensen
Section for Sports Traumatology M51, Bispebjerg and Frederiksberg Copenhagen University Hospital; Department for Physio- and Ergotherapy, Bispebjerg and Frederiksberg Copenhagen University Hospital; Division of Sports Trauma, Orthopaedic Department, Aarhus University Hospital; Section of Biostatistics, Department of Public Health, University of Copenhagen

Background: The paediatric version of IKDC (Pedi-IKDC) is the most frequently used PROM to evaluate treatment effects in children with ACL injuries. It consists of two subscales, symptoms and sports, but all raw scores are aggregated. It is the primary outcome in two large scale initiatives for the treatment of children with ACL deficiency: the European “Paediatric ACL Monitoring Initiative” (PAMI), and the North American “Pediatric ACL: Understanding Treatment Options” (PLUTO). However, Pedi-IKDC has only been subject to validity assessments with classical test theory models, and not modern test theory (MTT) models, which are preferable.
Aim: To study the structural validity and reliability of the questionnaire Pedi-IKDC in a cohort of children aged 9-15 with an ACL injury, using MTT models Rasch analysis and confirmatory factor analysis (CFA).
Materials and Methods: Data were collected prospectively before surgery and at 1-year follow-up from a nationwide cohort of 535 children with an ACL injury, treated with epiphyseal sparing reconstruction at either Aarhus or Bispebjerg University Hospitals. We evaluated the fit of a CFA model, adjusting models where possible, and confirmed results using Rasch analysis for each subscale and for the aggregated score.
Results: Neither of the subscales of Pedi-IKDC showed acceptable fit to the CFA model. Rasch analysis confirmed the results. It was possible to adjust the subscales, and a much better fit for the symptoms scale was achieved, yet the adjusted sports scale fitted the CFA model only slightly better. Reliability could not be reported due to inadequate model fit.
Interpretation / Conclusion: Pedi-IKDC does not exhibit adequate measurement properties (structural validity) for children with ACL- injury in its current form. Considering that the questionnaire also possesses a low degree of content validity for these children, data obtained by Pedi-IKDC should be interpreted with great caution. Future research should look at (i) why the Pedi- IKDC does not work well, (ii) whether a revised version with better measurement properties can be suggested, and (iii) what the consequences for the measurement of clinical change might be.

Niels Christian Kaldau, Niels Nedergaard, Per Hölmich, Jesper Bencke
Sports Orthopedic Research Center - Copenhagen, Department of Orthopedic Surgery, Amager-Hvidovre Hospital, Copenhagen, Denmark; Human Movement Analysis Laboratory, Department of Orthopedic Surgery, Amager-Hvidovre Hospital, Copenhagen, Denmark; Human Movement Analysis Laboratory, Department of Orthopedic Surgery, Amager-Hvidovre Hospital, Copenhagen, Denmark Sports Orthopedic Research Center - Copenhagen, Department of Orthopedic Surgery, Amager-Hvidovre Hospital, Copenhagen, Denmark

Background: Achilles tendon (AT) rupture is common among recreational male badminton players. Observations indicate that AT ruptures often occur in the transition from landing to forward acceleration following a forehand stroke on the rear court. It is an observation that many recreational players use a different landing technique from elite players, and the hypothesis is that this result in higher AT forces and increased risk of sustaining an AT rupture
Aim: To investigate if recreational players load the AT more when they use their original landing technique compared to an adapted landing technique used by elite players
Materials and Methods: Ten recreational male badminton players (age: 28.1 ± 6.3 years., height: 182.7 ± 5.9 cm, weight: 79.7 ± 10.5 kg) attended a single test session, where they performed 5 forehand rear court strokes with their usual technique jumping straight backwards with their landing foot in a neutral position, and 5 forehand strokes adopting the technique of elite players landing with the rear foot perpendicular to the direction of movement. AT force, of the landing leg opposite to the players’ racket arm, was calculated from 3D motion analysis. Paired t-test was used to evaluate differences between the two jump conditions for the recreational group, with an alpha level at 0.05
Results: The players landed with significantly more externally rotated foot in the adjusted landings (78.2 ± 10.0 degrees, p < 0.001) vs. the original landings (22.4 ± 21 degrees) without compromising performance parameters such as jump height and forward velocity. The peak AT force was reduced in the adjusted landings (50.1 ± 14.2 N/kg, 3960 ± 1181 N, p = 0.005) vs. the original landings (67.7 ± 18.9 N/kg, 5278 ± 1227 N) with Cohen´s d effect size of 1.17
Interpretation / Conclusion: The loading of the AT was markedly reduced from the original landing technique of the recreational players when mimicking the elite players landing technique with higher external rotation of the foot, despite no change in the functional performance parameters. These findings indicate that recreational players may reduce the high loads on their AT by adopting the landing technique of elite players, and potentially reduce the risk of sustaining an AT rupture

Bo Bregenhof, Per Aagaard, Nis Nissen, Mark Creaby, Jonas Bloch Thorlund, Carsten Jensen, Trine Torfig, Anders Holsgaard-Larsen
1Department of Orthopaedics and Traumatology, Odense University Hospital, Orthopaedic Research Unit, Department of Clinical Research, University of Southern Denmark. 2Department of Sports Science and Clinical Biomechanics, University of Southern Denmark; 3Department of Orthopaedics, Lillebaelt Hospital, Kolding; 4School of Behavioural & Health Sciences, Australian Catholic University; 5Health Innovation Centre of Southern Denmark; 6Department of Radiology, Odense University Hospital;

Background: The ACL remain one of the most injured ligaments of the knee, and usually involve ACL reconstruction (ACLR) with autograft harvesting. Common techniques involve using hamstring (HS) tendon autograft harvesting, and nine to twelve months of rehabilitation is typically needed before returning to sport. Despite rehabilitation, risk of reduced muscle strength is highly pronounced and persistent deficits in maximal knee flexor muscle strength have previously been reported.
Aim: To investigate the effect of targeted exercise on knee-muscle strength and joint function in ACLR participants with persistent hamstring muscle deficiency 12–24 months post-surgery.
Materials and Methods: A prospective, superiority, randomized controlled trial with parallel groups, balanced randomization (1:1) and blinded outcome assessment (level of evidence: II). Participants with ACLR (hamstring autograft) and persistent hamstring muscle deficiency were recruited 1- 2 years post-surgery and randomised to either 12-weeks supervised progressive strength and neuromuscular training (SNG), or home-based, weightbearing low-intensity exercises (CON). Primary outcome was between-group change in maximal isometric knee flexor muscle strength at 12-weeks follow-up. Secondary outcomes included measures of objective strength and subjective function.
Results: Fifty-one participants (45% women, 27 ± 6 years) were randomized, with data from 88% of participants being available at 12-weeks follow-up. A superior effect in knee flexor muscle strength (0.18 Nm/kg [95% CI 0.07; 0.29] p = 0.002) for SNG versus CON was observed. Furthermore, secondary between-group effects emerged in favour of SNG: KOOS Pain (4.56 [95% CI 0.43; 8.69] p = 0.031) and KOOS activity of Daily Living Function (4.71 [95% CI 1.20; 8.22] p = 0.010). No superior between-group effects were observed for other measures of objective strength and subjective function.
Interpretation / Conclusion: Twelve weeks of supervised progressive strength training in ACLR participants with persistent knee muscle deficiency demonstrated superior effects on knee flexor muscle strength and joint function compared with homebased exercise. However, it may be questioned whether observed benefits are clinically meaningful.

Malte Schmücker, Jørgen Haraszuk, Per Hölmich, Kristoffer W. Barfod
Sports Orthopedic Research Center – Copenhagen (SORC-C), Department of Orthopedic Surgery, Copenhagen University Hospital Amager-Hvidovre, Denmark.

Background: It has been indicated that anterior cruciate ligament reconstruction (ACLR) with quadriceps tendon (QT) graft has a higher risk of revision compared to hamstring tendon (HT) graft.
Aim: To investigate if ACLR with QT had higher risk of graft failure, revision ACLR or re- operation compared to HT in a high-volume center. We hypothesized that there would be no between group difference.
Materials and Methods: This was a registry study with review of medical records. Our study cohort consist of patients with primary ACLR using either QT or HT performed at Copenhagen University Hospital Hvidovre from January 2015 to December 2018. The cohort was identified from the Danish Knee Ligament Recon- struction Registry and linked to the Danish National Patient Registry to identify all hospital contacts post-ACLR. The outcome variables were graft failure (re-rupture or >3mm side difference in A-P laxity), revision ACLR, re-operation due to cyclops, re- operation due to meniscal injury and re- operation due to any reason. Also, A-P laxity and pivot-shift were assessed at 1 year. Using Kaplan-Meier estimates, the rates of events were evaluated at 2 years and comparison performed with Cox regression analysis.
Results: 475 subjects (HT=252, QT=223) were included. The risk of graft failure at 2 years was 9.4% for QT and 11.1% for HT (p= .46). Respectively, the risk of revision ACLR was 2.3% and 1.6% (p= .66), the risk of re- operation due to cyclops was 5.0% and 2.4% (p= .13), and the risk of re-operation due to meniscal injury was 4.3% and 7.1% (p= .16). The risk of re-operation due to any reason was 20.5% and 23.6% (p= .37). At 1-year follow-up A-P laxity was 1.4 mm for QT and 1.5 mm for HT (p= .35), and the proportion of patients with a positive pivot- shift was 29% for both groups.
Interpretation / Conclusion: QT and HT yield similar rates of graft failure, revision ACLR and re-operation at two years follow-up after ACLR. Graft failure was found in 9-11%. QT showed a non- statistically trend of higher risk for re- operation due to cyclops, and HT a non- statistically trend of higher risk for re- operation due to meniscal injury.

Bo Bregenhof , Anders Holsgaard-Larsen, Jonas Bloch Thorlund, Carsten Jensen, Maria Thorning, Per Aagaard, Nis Nissen, Carsten Bogh Juhl
1 Department of Orthopaedics and Traumatology, Odense University Hospital, Denmark, and Orthopaedic Research Unit, Department of Clinical Research, University of Southern Denmark; 2 Department of Sports Science and Clinical Biomechanics, University of Southern Denmark; 3 Department of Physiotherapy and Occupational Therapy, Copenhagen University Hospital, Herlev and Gentofte; 4 Department of Orthopaedics, Lillebaelt Hospital; 5 Department of Regional Health Research, University of Southern Denmark; 6 Department of Neurology, Odense University Hospital; 7 Research Unit for General Practice, Department of Public Health, University of Southern Denmark; 8 Department of Molecular Medicine, University of Southern Denmark

Background: Regaining adequate quadriceps and hamstrings muscle strength after Anterior Cruciate Ligament Reconstruction (ACLR) is important, as leg muscle strength, after ACLR is positively associated with functional performance.
Aim: We conducted a Systematic review and meta-analysis to investigate differences in knee extensor and flexor muscle strength following ACLR with quadriceps tendon/bone-patellar tendon-bone (QT/BPTB) vs. hamstring (HS) grafts.
Materials and Methods: We searched the electronic databases MEDLINE, EMBASE and CENTRAL for Randomized Controlled Trials (RCTs) comparing muscle strength after ACLR with QT/BPTB vs. HS graft. Differences in muscle strength between graft types and limb asymmetry index (LSI) at 6, 12, 24 and 48+ months postoperatively were combined in a random effects meta-analysis. Risk of bias was assessed using RoB2.
Results: Twenty-seven RCTs were included, of which eighteen were judged as having a high risk of bias. The differences in knee extensor muscle strength between ACLR patients allocated to QT/BPTB vs. HS grafts were (in favor of the HS graft) 9.7% (95% CI: 7.0; 12.4) at 6 months and 5.2% (95% CI: 2.7; 7.8) at 12 months, resolving over time to 1.2% (95% CI: -0.9; 3.2) at 24 months and 2.2% (95% CI: -1.0; 5.4;) at 48+ months. The differences in knee flexor strength between graft types were (in favor of QT/BPTB grafts) 6.2% (95% CI: 2.9; 9.5) at 6 months post-surgery and 7.7% (95% CI: 4.2; 11.2) at 12 months, resolving to 2.2% (95% CI: -1.7; 6.0) at 24 months and 5.1% (95% CI: -1.4; 11.6) at 48+ months post-surgery. Between-graft difference in knee extension LSI was in favor of HS graft at 12 months (4.9% [95% CI: 0.1; 9.8; p = 0.04]). Notably, no difference in LSI of knee flexion between grafts were observed.
Interpretation / Conclusion: Graft-specific differences in knee extensor and flexor strength were observed postoperatively. Differences in knee extensor strength LSI in favor of HS grafts were found, but no difference in knee flexor strength LSI. Loss of muscle strength after ACLR appeared to depend on graft choice, underlining the importance of site-specific postoperative rehabilitation.

Nicolai Sandau, Stig Brorson
Centre for Evidence-Based Orthopedics, Department of Orthopedic Surgery, Zealand University Hospital Køge

Background: Magnetic resonance imaging (MRI) is the primary image modality for diagnosing soft tissue knee injuries. Correct interpretation of knee MRI can be time-consuming. Deep learning assisted diagnosis in knee MRI has shown potential for improving diagnostic speed and accuracy, but the results have been inconsistent. Recent advances in deep learning methods have resulted in improvements in both performance and interpretation for other medical image diagnosis tasks.
Aim: We aim to study these advances and develop a deep learning model capable of diagnosing anterior cruciate ligament (ACL) and meniscal tears in knee MRI.
Materials and Methods: We used the MRNet dataset consisting of 1250 knee MRI scans from Stanford University Medical Center. Each case has been labeled with the presence or absence of an ACL tear and/or meniscal tear as diagnosed by a radiologist. The dataset is split into a training set (1130 cases) and a validation set (120 cases). We developed a deep learning model based on the EfficientNet architecture and trained it on the training set. The trained model was then used to diagnose the scans in the validation set. Using the diagnostic labels as gold- standard we calculated the sensitivity and specificity. We calculated the area under the receiver operating characteristic curve (ROC) as an overall measure of performance. Lastly, we used Gradient-weighted Class Activation Mapping to visualize which regions of each scan the model used for diagnosis.
Results: The model achieved ROC (95% CI) values of 0.99 (0.97 - 1) for ACL tears and 0.88 (0.82 - 0.94) for meniscal tears. The average ROC was 0.93, the currently highest reported for the MRNet dataset. The sensitivity (95% CI) was 0.87 (0.76 - 0.94) for ACL tears and 0.83 (0.7 - 0.91) for meniscal tears. The specificity (95 CI) was 0.94 (0.85 - 0.98) for ACL tears and 0.76 (0.65 - 0.85) for meniscal tears . The sensitivity and specificity is comparable to previously reported values achieved by trained radiologists on the MRNet dataset.
Interpretation / Conclusion: Our findings indicate that deep learning has the potential to aid radiologists and orthopedic surgeons when diagnosing ACL and meniscal tears in knee MRI.

Pernille Melbye, Per Hviid Gundtoft, Teodor Lien-Iversen, Daniel Barklin Morgan, Jens Christian Pörneki, Bjarke Viberg
Department of Orthopaedic Surgery and Traumatology, Lillebaelt Hospital, University Hospital of Southern Denmark

Background: Studies with less than 10 years of follow-up have demonstrated no difference between surgical and non-surgical treatment after an anterior cruciate ligament (ACL) rupture; however, long-term effects remain unclear.
Aim: The aim of this study was to compare the risk of long-term secondary surgical procedures after primary surgical and non- surgical treatment for ACL ruptures.
Materials and Methods: Patients aged 18–35, registered in the Danish National Patient Registry with an ACL rupture between January 1, 1996 and December 31, 2000 were included with approximately 20 years follow-up. The surgically treated group was defined as receiving an ACL reconstruction within 1 year after diagnosis. Major secondary surgical procedures were defined as subsequent ACL surgeries (reconstruction/revision), arthroplasty, deep infection, arthrodesis, or amputation. Minor secondary surgical procedures were defined as meniscal surgery, synovectomy, and manipulation under anaesthesia (MUA). Multivariate regression analysis was performed to assess relative risk (RR), adjusted for age and sex. The results are reported with 95% confidence intervals.
Results: In total, 7,539 patients had an ACL rupture; 1,970 patients were surgically treated. There were 63% males and a mean age of 25 years (24.9-25.1) with no clinically relevant difference between groups. In the surgical group, 5.9% of patients underwent major secondary surgeries compared to 6.1% in the non-surgical group, yielding an adjusted RR of 1.05 (0.85;1.30). The majority (88%) had only one major secondary surgery with no difference between the groups (p=0.171). 44% underwent minor secondary surgeries in the surgical group compared to 49% in the non- surgical group, yielding an adjusted RR of 1.29 (1.20;1.39). For both groups, the majority of the minor reoperations were meniscal surgeries (71%). A total of 37% had more than 1 minor secondary procedure with no difference between the groups (p=0.381).
Interpretation / Conclusion: No significant differences in the rate of major secondary surgical procedures between surgically and non-surgically treated ACL patients, the non-surgical group was associated with a higher risk of minor secondary surgeries.

Bjørn Borsøe Christensen, Anders El-Galaly, Jens Ole Laursen, Martin Lind, Bjørn Christensen
Department of orthopedics, Horsens Regional Hospital; Department of orthopedics, Aarhus University Hospital; Interdisciplinary Orthopedics, Aalborg University Hospital; Department of Orthopedic Surgery, Aalborg University Hospital; Department of Emergency Medicine, Hospital of Southern Jutland

Background: Focal cartilage injuries are debilitating and difficult to treat. Biological cartilage repair procedures are used for patients younger than 40 years, and knee arthroplasties are generally reserved for patients older than 60. Condyle resurfacing implants are well suited for patients in this treatment gap.
Aim: The objective was to investigate the 10-year survival of condyle resurfacing implants in the Danish Knee Arthroplasty Registry.
Materials and Methods: In this retrospective cohort study, patients treated condyle resurfacing implants were followed longitudinally in the Danish Knee Arthroplasty Registry from 1997 to 2020. The primary endpoint was revision surgery. The survival of the condyle resurfacing implants was analyzed by Kaplan Meier method.
Results: 379 condylar implant procedures were retrieved from the Danish Knee Arthroplasty Registry. The mean age and weight of patients were 50 years (SD: 11) and 84 kg (SD: 17). The indications for surgery were: Secondary osteoarthritis (42%), primary osteoarthritis (32%) and osteochondral lesions (20%). Within the follow-up period, 70 (19%) of the implants were revised to arthroplasties. The 1-, 5- and 10-year revision free survival estimation was 0.95 (95% CI: 0.93-0.97), 0.84 (95% CI: 0.80- 0.88) and 0.80 (95% CI: 0.75-0.84). The median time to revision was 2 years.
Interpretation / Conclusion: The 10-year revision free survival rate for condyle resurfacing implants was 80%. Based on the revision rates, this treatment offers a viable alternative to biological cartilage repair methods in middle-aged patients with focal cartilage pathology. Improved patient selection could further improve the implant survival rate. Further studies are needed to investigate this treatment method.

Markus Gadeberg, Allan Cramer, Per Hölmich, Kristoffer Barfod
Sports Orthopedic Research Center - Copenhagen (SORC-C), Department of Orthopedic Surgery, Copenhagen University Hospital, Amager-Hvidovre; Sports Orthopedic Research Center - Copenhagen (SORC-C), Department of Orthopedic Surgery, Copenhagen University Hospital, Amager-Hvidovre; Sports Orthopedic Research Center - Copenhagen (SORC-C), Department of Orthopedic Surgery, Copenhagen University Hospital, Amager-Hvidovre; Sports Orthopedic Research Center - Copenhagen (SORC-C), Department of Orthopedic Surgery, Copenhagen University Hospital, Amager-Hvidovre;

Background: The diagnosis codes for sports injuries in the Danish National Patient Register (DNPR) are frequently used for research, however the validity is unknown.
Aim: The aim was to investigate the validity of diagnosis codes for some of the most common sports related injuries: acute Achilles tendon rupture (DS86.0A), Achilles tendinitis (DM76.6), rupture of anterior cruciate ligament of the knee (ACL) (DS83.5E), dislocation of the patella (DS83.0), traumatic tear of the meniscus (DS83.2) and degenerative meniscal lesion (DM23.2).
Materials and Methods: The study was performed as a registry study in the DNPR. For each diagnosis code, patient records from Copenhagen University Hospital Hvidovre were retrieved from January 1st to December 31st 2017. We considered a positive predictive value (PPV) of 80% or higher to be satisfying.
Results: The population consisted of 85 patients registered with the diagnosis code for acute Achilles tendon rupture, 65 patients with Achilles tendinitis, 73 patients with ACL rupture, 100 patients with dislocation of the patella, 100 patients with traumatic tear of the meniscus, and 100 patients with degenerative meniscal lesion. For acute Achilles tendon rupture the PPV was 98% (95% CI: 92%-100%), for Achilles tendinitis 85% (95% CI: 74%-92%), for ACL rupture 96% (95% CI: 88%-99%) and for dislocation of the patella 96% (95% CI: 90%-99%). Depending on the definition of the diagnoses, the PPVs were 56%-72% for traumatic tear of the meniscus and 53%-77% for degenerative meniscal lesion.
Interpretation / Conclusion: This study documented an acceptable validity allowing for epidemiological research of the diagnosis codes for acute Achilles tendon rupture, Achilles tendinitis, ACL rupture and dislocation of the patella. The diagnosis codes for traumatic tear of the meniscus and degenerative meniscal lesion showed a lower validity, and thus caution should be taken when using these codes.

Randi Gram Rasmussen, Julie Sandell Jacobsen, Birgitte Blaabjerg, Torsten Grønbech Nielsen, Lene Miller Lindberg, Martin Lind
Department of Occupational and Physical Therapy; Aarhus University Hospital; Research Centre for Health and Welfare Technology, Programme for Rehabilitaiton, VIA University College, Aarhus; Research Unit for General Practice, Aarhus; Department of Orthopaedic Surgery, Aarhus University Hospital, Aarhus.

Background: PCL injuries can be treated surgically or with progressive exercises in combination with a PCL support brace. However, larger prospective studies reporting outcome of exercise-related treatment are lacking.
Aim: We aimed to investigate changes in patient-reported and functional outcome of a physiotherapy-led progressive exercise program plus a PCL support brace in patients with an acute injury of the PCL over a 24-months follow-up. Furthermore, to report conversion to surgical reconstruction.
Materials and Methods: In a prospective case-series study, 50 patients were treated with a PCL support brace for 12 weeks and underwent a 16-week physiotherapy-led progressive exercise program. Changes in patient-reported outcome was investigated with the International Knee Documentation Committee Subjective Knee Form (IKDC-SKF) from baseline to 1 and 2 years. Furthermore changes in isometric knee flexion and extension strength was measured from 16 weeks to 1 year. Mean changes were analyzed with a mixed effect model with patients as a random factor and time as a fixed factor.
Results: Seven patients (14 %) converted to PCL reconstruction resulting in 43 patients for 1-year follow-up that completed the combined brace and rehabilitation treatment. Of the patients converting to reconstruction, two patients had an isolated ligament injury and five patients had dislocation of the knee. The IKDC-SKF score at baseline was 35 (SD 9.7) and at 2 years 62 (SD 15). Isometric knee flexion strength of the injured knee increased statistically significantly from 0.93 (SD 0.36) Nm/kg to 1.1 (SD 0.36) Nm/kg, corresponding to an increase of 17%. In contrary isometric knee extension strength of the injured knee did not change (0.10 (-0.022-0.21) Nm/kg, p=0.107
Interpretation / Conclusion: Treatment resulted in a 14% conversion rate to surgical treatment. The treatment demonstrated clinically relevant improvements in patient-reported outcome and an improvement of 17% in flexor strength. Consequently, limited need for conversion to surgical treatment, clinically relevant improvements in subjective outcome and strength after PCL support brace treatment and a progressive exercise program can be expected in patients with an acute PCL injury.