5-in-5

Comparison of Hamstring Tendon and Patellar Tendon Grafts in Anterior Cruciate Ligament Reconstruction in a Nationwide Population-Based Cohort Study Results From the Danish Registry of Knee Ligament Reconstruction Lene Rahr-Wagner,*yz MD, Theis Muncholm Thillemann,y MD, PhD, Alma Becic Pedersen,z MD, PhD, and Martin Lind,y MD, PhD Investigation performed at the Departments of Clinical Epidemiology and Orthopaedic Surgery, Aarhus University Hospital, Aarhus, Denmark Background: The choice of graft for anterior cruciate ligament reconstruction (ACLR) remains controversial, and despite numerous studies, there is still an ongoing debate on this topic. The 2 most widely used grafts are the hamstring tendon and patellar tendon. Hypothesis: In this study, we hypothesized that the revision rate after primary ACLR is greater when using hamstring tendon grafts compared with patellar tendon grafts. Study Design: Cohort study; Level of evidence, 2. Methods: From the nationwide population-based Danish Knee Ligament Reconstruction Registry, we identified all primary ACLR procedures (n = 13,647) performed in Denmark between July 2005 and December 2011. The end point was revision ACLR. As other end points, we used objective measurements and patient-reported outcome scores. Revision rates and relative risk estimates for revision ACLR were calculated using Cox multiple regression. Results: The use of hamstring tendon grafts increased from 68% in 2005 to 85% in 2011. The cumulative revision rates for hamstring tendon grafts at 1 and 5 years were 0.65% (95% confidence interval [CI], 0.51%-0.82%) and 4.45% (95% CI, 3.94%5.01%), respectively. For patellar tendon grafts, the revision rate was 0.16% (95% CI, 0.05%-0.50%) at 1 year and 3.03% (95% CI, 2.27%-4.05%) at 5 years. The adjusted overall relative risk of revision surgery in the hamstring tendon group compared with the patellar tendon group was 1.41 (95% CI, 1.03-1.92), and the adjusted relative risk of undergoing revision surgery performed after 1 and 5 years was 3.82 (95% CI, 1.20-12.2) and 1.90 (95% CI, 0.43-8.40), respectively. Conclusion: In this population-based study, the use of hamstring tendon grafts in ACLR was associated with an increased risk of revision compared with patellar tendon grafts, in particular during the first year after surgery. These results demonstrate that both hamstring and patellar tendon grafts reveal good results after ACLR and suggest that graft selection should be based on an individual evaluation of patient demands and graft morbidity. Keywords: ACL; revision; patellar tendon; bone–patellar tendon–bone; semitendinosus/gracilis; hamstring; graft choice

between countries is seen. In Denmark and Sweden, in recent years, the HT graft has been used in about 84% of cases.17 In the United States, based on the Multicenter Orthopaedic Outcome Network (MOON) database, the most widely used graft types were doubled HT autografts (44%) and PT autografts (42%), whereas the Norwegian National Knee Ligament Registry (NKLR) reported that the HT graft was used in 60% and the PT graft in 37% of patients.26 There are advantages and disadvantages of each graft, and usually, the choice of graft depends on the surgeon’s

Despite numerous studies comparing grafts in anterior cruciate ligament reconstruction (ACLR), the choice of graft remains controversial. The most frequently used grafts for ACLR are hamstring tendon (HT) and patellar tendon (PT) autografts, although a great difference

The American Journal of Sports Medicine, Vol. 42, No. 2 DOI: 10.1177/0363546513509220 Ó 2013 The Author(s)

278

Vol. 42, No. 2, 2014

personal preferences and an assessment in relation to patient characteristics. An optimal graft is fast healing, is strong, and restores knee joint stability with low morbidity. An animal study suggested that PT grafts heal more rapidly because of bone-to-bone healing compared with bone-to-tendon healing.31 Rapid healing is essential to ACLR, as it allows for earlier and more accelerated rehabilitation and thus a more rapid return to previous sports performance. Therefore, PT grafts have been suggested for younger and more active patients younger than 20 years of age because failure rates seem to be higher in this age group.24,27 Also, use of the PT graft has increased because of easy-to-use interference screw fixation.14,31 However, many studies have shown that ACLR using PT autografts is associated with well-documented donor site morbidity, such as anterior knee pain, pain at kneeling, extensor strength deficit, and patellofemoral osteoarthritis.5,12,21,30,36 Therefore, in recent years, use of the HT graft has been popularized because of its lower rate of donor site morbidity.30 Also, this is supported by randomized controlled trials indicating equal stability and subjective outcomes when comparing HT to PT grafts.5,10,22,28,30 However, the HT graft is associated with deficits in knee flexion strength and internal tibial rotation strength.11,28 An abundance of randomized controlled trials, including a Cochrane review, have reported slight or no differences in postoperative stability and the revision rate after primary ACLR between HT and PT grafts; thus, no consensus has been reached yet.5,10,11,22,28,30 Hence, the objective of this nationwide population-based cohort study was to compare the revision rate and clinical outcomes between HT and PT grafts used in primary ACLR. We hypothesized a lower revision rate for PT grafts compared with HT grafts because of better graft incorporation in PT grafts.

MATERIALS AND METHODS Study Setting Denmark has a population of 5.5 million people. The National Health Service provides tax-supported health care to all Danish residents, allowing free access to hospital care at medical, surgical, and psychiatric departments as well as general practitioner visits. Patients with acute medical conditions are admitted for specialist treatment at public hospitals. Private hospitals are also accessible in Denmark, and they also have reimbursement agreements with the Danish State.

Data Sources The Danish Knee Ligament Reconstruction Register (DKRR). The DKRR is a nationwide population-based

Registry-Based Comparison of Graft Types in ACLR

279

clinical database that was established on July 1, 2005 with the purpose of improving the monitoring and quality of both primary and revision ACLR in Denmark.23,34 All private (n = 27) and public (n = 24) hospitals report to this register, and registration is compulsory according to Declaration Number 459 of June 2006.9 The rate of registrations in the DKRR has been more than 85% in the past 3 years.32,35 Detailed preoperative, intraoperative, and 1-year follow-up data are recorded by the operating surgeon using a standardized form and a secure Internet portal.23 Furthermore, patients independently report subjective scores on knee function using the Knee injury and Osteoarthritis Outcome Score (KOOS)37 and the Tegner functional score.41 These data are web-recorded by the patient before surgery and 1 year after surgery. The KOOS ranges from 0 to 100, and the Tegner score ranges from 1 to 10, with higher scores representing better results. The KOOS4 can also be calculated as a validated average of 4 KOOS subscales: quality of life, sport, pain, and symptoms.13 The Civil Registration System (CRS). A unique 10-digit personal identification number is given to all Danish citizens at the date of birth. The CRS records information on changes in the vital status of all Danish citizens including changes in address, date of emigration, and date of death since 1968.33 Because this personal identification number is consistent through all Danish registries, precise individual-level data linkage between all Danish registries is possible. The CRS was also used to obtain complete follow-up information on all patients.

Study Population In total, we identified 13,760 primary ACLR procedures using either the HT or the PT graft in 13,565 patients in the period from July 2005 to December 2011. We excluded 22 patients who had a status date before the operation date, which is the end of the study period (December 31, 2011), and 1 patient who had the revision surgery date registered before the primary surgery date. In 40 operated knees, the patients lived in Greenland, and in 50 operated knees, the patients were from other countries who had visited Denmark; these 2 groups could not be properly followed up and were therefore excluded. Thus, 13,647 primary ACLR procedures were included in the final analysis.

Exposure In this study, we investigated the results of the use of either 4-stranded semitendinosus/gracilis grafts, defined as HT grafts, or PT grafts on our outcome measures. In total, we identified 14,755 operated knees from the

*Address correspondence to Lene Rahr-Wagner, MD, Department of Clinical Epidemiology, Aarhus University Hospital, Olof Palmes Alle 43-45, 8200 Aarhus N, Denmark (e-mail: [email protected]). y Department of Orthopaedic Surgery, Aarhus University Hospital, Aarhus, Denmark. z Department of Clinical Epidemiology, Aarhus University Hospital, Aarhus, Denmark. One or more of the authors has declared the following potential conflict of interest or source of funding: This study received financial support from the Danish Rheumatism Association, the Elisabeth and Karl Ejnar Nis-Hanssens Scholarship, and the Aase and Ejnar Danielsens Foundation.

280

Rahr-Wagner et al

DKRR: 11,676 HT grafts and 1971 PT grafts. The remaining 1108 grafts used were 4-stranded semitendinosus grafts alone (1.7%), iliotibial tract grafts (1.6%), doublebundle HT grafts (1.4%), 2-stranded semitendinosus grafts alone (1.2%), quadriceps tendon grafts (0.9%), allografts (0.2%), and others (0.5%). In this study we only included ACLR procedures using the HT or the PT graft; hence, our final exposed group was 13,647 reconstructed knees.

Outcomes The primary outcome was revision ACLR, defined as another ACLR performed in the same knee as the primary ACLR. The follow-up period started on the date of the primary ACLR and ended on the date of revision ACLR if revision occurred, at the time of death, or on the status date, which is the end of the study period (December 31, 2011), whichever came first. Furthermore, we described the cause of revision surgery. The secondary outcomes were parameters of objective knee stability in terms of instrumented side-to-side differences (eg, Rolimeter [Aircast, Boca Raton, Florida] or KT-1000 arthrometer [MEDmetric Corp, San Diego, California]) and pivot-shift scores. The pivot-shift test is a dynamic and passive test of the knee that measures rotational stability of the ACL. The pivot-shift test is graded by a 4-point scale from normal (0), glide (1), clunk (2), and gross (3).19 The pivotshift data were divided into one group with negative pivotshift results (normal) (n = 4674) and the other group with positive pivot-shift results (glide, clunk, and gross) (n = 918). The instrumented side-to-side difference measures the difference in sagittal stability between the operated knee and the healthy knee at the 1-year control visit. The side-to-side difference was measured manually as the maximal translation at 25° of flexion using either the KT-1000 arthrometer or Rollimeter. Patients were categorized as having a difference of \2 mm (n = 4096) or .2 mm (n = 790). Only patients with no prior ACLR of the contralateral knee were included in this analysis. Hence, 428 knees were excluded from this analysis. Finally, we used patient-reported outcomes, the KOOS and the Tegner score, at 1 year postoperatively, if reported. The KOOS and Tegner score are validated, subjective patient-reported outcomes, calculated according to published standards.37,41 The KOOS4 is a validated patientreported outcome computed from the 4 most responsive KOOS subscores13: symptoms, pain, sport, and quality of life. Preoperative and postoperative KOOS and Tegner scores were available in 4516 of 13,647 patients (33%) and in 3614 of 13,647 patients (26%), respectively.

Confounding Factors We obtained data at the time of surgery from the DKRR on sex, age (20 and .20 years), cartilage damage .1 cm2 (no/yes or missing), operated meniscal damage (yes/no or missing), prior surgery of the knee (yes or no), and activity leading to the primary ACL rupture (sport vs nonsport). All these variables were used as confounders in, or estimates of, the relative risk (RR).

The American Journal of Sports Medicine

Statistics By the Kaplan-Meier method, we estimated graft survival and the cumulative revision probability at 1 and 5 years’ follow-up in the HT and PT groups. We used Cox regression analyses to compare the revision risk after primary ACLR among patients with HT and PT grafts. We computed the hazard ratios as a measure for RR with a 95% confidence interval (95% CI) for patients with HT grafts compared with PT grafts, while adjusting for potentially confounding factors as mentioned above. The assumption of the Cox regression model was assessed with the use of log-log plots and Schoenfeld residuals and was found suitable. Further, using the logistic regression analysis by adjusting for potential confounders, we calculated the odds ratio of having a positive pivot-shift test result in the HT group compared with the PT group and the risk of having more than a 2-mm side-to-side difference between the operated and healthy knee in the HT group compared with the PT group. For the side-to-side difference, we excluded 41 patients (82 knees) who had both knees operated on as well as registered side-to-side differences and 347 knees that had a previous ACLR performed on the contralateral knee. The mean values of the KOOS and Tegner score preoperatively and 1 year postoperatively for the patients in the HT and PT groups were compared using the Student t test if data were normally distributed based on Q-Q plots; otherwise, the Wilcoxon rank-sum test was used. All reported P values have a significant value of .05. All statistical analyses were computed using Stata version 12 (StataCorp, College Station, Texas). The study was approved by the Danish Data Protection Agency.

RESULTS Patient characteristics are outlined in Table 1. The most commonly used graft types in the DKRR besides HT and PT autografts are 4-stranded semitendinosus grafts (1.7%) and iliotibial tract autografts (1.6%). Allografts were only used in 0.20% of the reconstructions. The inclusion criterion for this study was the use of either the HT graft or the PT graft. The use of HT grafts increased from 68% of all graft types in 2005 to 85% of all graft types in 2011. The mean follow-up was 3.01 years (95% CI, 2.99-3.09). In the HT group, 312 of 11,676 knees underwent revision, whereas 47 of 1971 knees underwent revision in the PT group. The Kaplan-Meier cumulative revision rates of primary ACLR for HT grafts were 0.65% (95% CI, 0.51%0.82%) at 1 year and 4.45% (95% CI, 3.94%-5.01%) at 5 years. For PT grafts, the cumulative revision rates were 0.16% (95% CI, 0.05%-0.50%) at 1 year and 3.03% (95% CI, 2.27%-4.05%) at 5 years (Figure 1). The crude overall RR for revision in the HT group compared with the PT group was 1.50 (95% CI, 1.11-2.04). The overall adjusted RR for revision was 1.41 (95% CI, 1.031.92), and the adjusted RRs for revision after 1 and 5 years were 3.82 (95% CI, 1.20-12.20) and 1.90 (95% CI, 0.438.40), respectively.

Graft Choice, n (%)

Male sex Age 20 y Prior surgery of the knee Sports activity leading to the tear No cartilage lesion Meniscal treatment

1309 416 588 1590

(66) (21) (30) (81)

1626 (83) 766 (39)

Hamstring Tendon (n = 11,676) 6840 3149 3018 9472

(58) (27) (26) (81)

9244 (79) 4469 (38)

New trauma was ascribed as the most frequent reason for revision surgery and accounted for 41% of revisions in the HT group and 47% in the PT group. This was followed by suboptimal placement of the graft in the femur, which accounted for 16% in the HT group and 19% in the PT group (Table 2). In the HT group, 39% had pivot-shift test results recorded 1 year postoperatively, and in the PT group, 52% had pivotshift test results recorded 1 year postoperatively. Sixteen percent of the knees in the HT group had positive pivot-shift test results, and 19% of the knees in the PT group had positive pivot-shift test results. The HT group was associated with a slightly but significantly lower risk of positive pivot-shift test results compared with the PT group, with an adjusted odds ratio of 0.81 (95% CI, 0.68-0.96). In the HT group, 36% had side-to-side differences recorded 1 year postoperatively, and in the PT group, 44% had side-to-side differences recorded 1 year postoperatively. One year postoperatively, 16% of the knees in the HT group had side-to-side differences of .2 mm, and 19% of the knees in the PT group had side-to-side differences of .2 mm. The odds ratio of having a side-to-side difference of .2 mm in the HT group compared with the PT group was 0.82 (95% CI, 0.68-1.01). Preoperatively, the KOOS and Tegner scores were comparable in the 2 groups (Table 3). One year postoperatively, the Tegner scores were 4.9 (95% CI, 4.9-5.0) for the HT graft and 4.7 (95% CI, 4.6-4.9) for the PT graft (P \ .05). Further, 1 year postoperatively, the KOOS sports subscores were 62.6 (95% CI, 61.7-63.5) for HT grafts and 58.0 (95% CI, 56.2-59.7) for PT grafts (P \ .05).

DISCUSSION This is the first nationwide registry-based cohort study presenting the results comparing the use of HT grafts with PT grafts in primary ACLR. Prior studies relied on small sample sizes and are often based on 1 surgeon or patients in 1 department, which may not present the everyday clinical praxis.2,10 Because the risk for revision after ACLR is relatively low, a large sample size is necessary for a valid estimation of revision risks. A nationwide registry-based study would fulfill these criteria. Until now, such a study has not been conducted.

Hamstring Tendon

95% CI

Patellar Tendon

0

Patellar Tendon (n = 1971)

95% CI

281

.04

Cumulative ACL revisions

TABLE 1 Patient Characteristics

.06

Registry-Based Comparison of Graft Types in ACLR

.02

Vol. 42, No. 2, 2014

0 Number at risk: PT graft (n = 1971) HT graft (n = 11,676)

2 4 Time of follow-up (years) 1622 7214

1072 3434

6 181 552

Figure 1. Kaplan-Meier cumulative revision curve of primary anterior cruciate ligament reconstruction using either a hamstring tendon graft or a patellar tendon graft. TABLE 2 Causes of Revision Surgery Recorded in the Danish Knee Ligament Reconstruction Registry Graft Choice, % Cause New trauma Tunnel widening Suboptimal placement of the graft in the tibia Suboptimal placement of the graft in the femur Infection Unknown reason for instability Other ligament failure Other Totala

Hamstring Tendon

Patellar Tendon

41.0 2.3 8.3

46.8 0 8.5

15.6

19.2

3.6 22.2 4.3 2.7 100

0 12.7 8.5 4.3 100

a

Ten knees had no explanation as to the cause of revision.

In this study, we found a minor difference in the overall risk of revision between HT grafts and PT grafts when adjusting for relevant confounders. However, we found almost 4 times an increased risk of early revision at 1-year follow-up in the HT group compared with the PT group. Objective stability measures showed a minor decreased risk of positive pivot-shift test results in the HT group compared with the PT group. Further, the Tegner scores and KOOS subscores of sport and activities of daily living at 1 year postoperatively were slightly higher in the HT group. The 3.8-fold increased RR of revision at 1 year after ACLR using HT grafts compared with PT grafts is a concern. However, the absolute risk of revision is small: less than 1% for both groups. This short-term increased risk of revision could represent failures that occur because patients return to sport too quickly; hence, a solution may be to prolong the rehabilitation period before returning to sport. This is also supported by the better KOOS

282

Rahr-Wagner et al

The American Journal of Sports Medicine

TABLE 3 Preoperative and 1-Year Postoperative KOOS and Tegner Scores for Primary ACLR in the PT and HT Groupsa HT Group Preoperative KOOS (n = 4516, 33%), n (%) Symptoms Pain ADL Sport Quality of life KOOS4 Preoperative Tegner score (n = 4516, 33%) Postoperative KOOS (n = 3677, 27%), n (%) Symptoms Pain ADL Sport Quality of life KOOS4 Postoperative Tegner score (n = 3677, 27%)

3726 70.8 71.0 78.2 38.3 39.5 54.9 3.0 2968 77.1 83.6 89.0 62.6 59.6 70.8 4.9

(32) (70.2-71.3) (70.5-71.6) (77.6-78.8) (37.5-39.2) (38.9-40.0) (54.4-55.4) (3.0-3.1) (25) (76.4-77.7) (83.0-84.1) (88.6-89.5) (61.7-63.5) (58.8-60.3) (70.1-70.9) (4.9-5.0)

PT Group 790 71.7 71.7 78.8 38.8 39.3 55.4 3.0 709 78.3 83.4 88.5 58.0 59.0 69.7 4.7

(40) (70.6-72.8) (70.5-72.9) (77.5-80.0) (37.0-40.6) (38.2-40.5) (54.3-56.5) (2.8-3.1) (36) (77.2-79.4) (82.4-84.4) (87.6-89.5) (56.2-59.7) (57.5-60.5) (68.5-70.9) (4.6-4.9)

P Valueb

NS NS NS NS NS NS NS NS NS \.05 \.05 NS NS \.05

a Values are presented as the median (95% confidence interval) unless otherwise indicated. ACLR, anterior cruciate ligament reconstruction; ADL, activities of daily living; HT, hamstring tendon; KOOS, Knee injury and Osteoarthritis Outcome Score; KOOS4, quality of life, sport, pain, and symptoms of the KOOS; NS, not significant; PT, patellar tendon. b P values for the difference in the KOOS and Tegner scores between the 2 groups were calculated using the Wilcoxon rank-sum test and Student t test.

sport subscore in the HT group, as this may indicate that patients feel ready for sports activities. Also, it is well known that PT grafts heal faster because of bone-to-bone healing.31 This may also explain the higher risk of revision in the HT group at 1 year after surgery because this soft tissue graft with bone-to-tendon healing needs more time to obtain maximal strength. Zebis et al42 showed a reduced risk for noncontact ACL injuries in female athletes who had undergone neuromuscular training. They found that it seems essential to have an adequate function of the HT muscles to protect the ACL from injuries. These findings could also support our data, suggesting that patients who undergo ACLR with HT grafts need more focused rehabilitation and a thorough evaluation before they return to sport. Many randomized controlled trials have studied the difference between HT and PT grafts.10,11,22,28 Contrary to our study, many recent studies have shown comparable revision rates between HT and PT grafts.2,5,20,25,40 Supporting our data is a meta-analysis from 2003 in which Freedman et al12 compared different studies using either HT or PT grafts in ACLR with a minimum of 2-year follow-up. Similar to our study, they showed a significantly higher graft failure rate when using HT grafts (4.9%) at a mean follow-up of 34 months compared with PT grafts (1.9%) at a mean followup of 46 months.12 Most of the studies included in this meta-analysis were from the early and mid-1990s. Also, a recent study by Maletis et al29 supports our data of a higher revision rate in the HT group compared with the PT group. Despite this abundance of literature, it has not been possible to show any clear advantage of one method over the other, and it seems that one ideal graft for all patients does not exist. Hence, recently, a Cochrane review that included 19 randomized or quasi-randomized controlled

trials with a minimum of 2 years’ follow-up concluded that there was no evidence to draw definite conclusions on differences between the 2 graft types for long-term follow-up in terms of subjective scores and objective measurements.30 In contrast to our study, they did conclude that the PT group is more likely to have statistically significant stable knees and is associated with more anterior knee pain.30 Knee stability after ACLR is of major importance to return to normal physical activity and to avoid the early onset of osteoarthritis due to an unstable knee.18 Therefore, postoperative stability, measured by the pivot shift and side-to-side difference4 of the knee, is an essential parameter when evaluating outcomes after ACLR. The risk of instability in terms of positive pivot-shift test results was slightly lower in the HT group compared with the PT group, whereas sideto-side instrumented stability measures were comparable between the groups. In our results, we divided pivot shift into 2 groups only (positive and negative). This was done because the pivot-shift test is a very subjective evaluation and because a more specific classification of pivot shift would result in groups with too small numbers of patients and even smaller numbers of outcomes, despite our large sample size. Therefore, we chose this subdivision to have a valid analysis and conclusions. Several studies have reported comparable clinical assessment or instrumented stability findings for different graft types,10,15,20,22,28 with no preferences between PT and HT autografts. Our objective stability results may indicate that the HT group, which also reports higher KOOS sport subscores, may have a higher activity level, which corresponds with our finding of less positive pivot shift in the HT group and thus a possible higher risk of reinjury. Patient-reported outcome measures, the Tegner score and KOOS subscores of sport and activities of daily living, were higher for HT grafts at 1 year postoperatively

Vol. 42, No. 2, 2014

compared with PT grafts; however, the difference is small and clinically not important. Different aspects influence the strength of the graft, such as graft size, tunnel position, bone density, and graft fixation.25,27 As previously mentioned, PT grafts have the advantage of bone-to-bone healing in contrast to HT grafts, which do not have a bone block and need tendon-to-bone healing and thus prolonged time for graft incorporation.8,16 Furthermore, most surgeons’ standardized screw fixation for PT grafts often uses interference screws,7 which is in contrast to the great variety of fixation methods used for the HT graft. Hence, a lot of factors are a concern when optimizing graft fixation. Our results do not necessarily support PT grafts to be superior to HT grafts but rather that clinicians should continue to select the graft type according to patient-reported factors including graft morbidity, patient activity level (ie, high-performance athletes), instability of the knee, and those with jobs demanding kneeling. Further, a prolonged rehabilitation period should be considered for athletes treated with an HT graft.

Strength and Limitations This study was based on data from a national clinical registry. Thus, there are several strengths and limitations. The fact that the DKRR is a large national database is a clear strength for the quality of the data. Also, because of the unrestricted and free access to health care in Denmark, the DKRR provides an unselected study population. Also, data are readily available, and the DKRR gives access to a large amount of data that otherwise would not be possible to obtain, reduces the risk of bias, and minimizes the cost of research.39 In addition, clinical databases make timely and early dissemination of information on specific clinical issues possible. Further, the DKRR has the potential for extensive linkage on an individual level to other important databases because of the unique personal identification number assigned to all Danish citizens, which enables the possibility of individual measurements. Further, the cohort design provides the researcher with the possibility of adjusting for relevant confounders. Also, data from a national clinical registry encounter several limitations. One problem is the completeness of the data and patients’ compliance of online subjective patient registrations. The rate of registration in the database was greater than 85% for the past 3 years, which we consider as acceptable.35 However, the low number of patients reporting preoperative and postoperative patient-reported outcome scores is a concern. This could lead to information bias if missing data on patient-reported outcomes are associated with both graft choices and later revision. Because the data collection is prospective and registration of primary ACLR is independent of registration of later revision, the risk of information bias is very limited. Also, because there is no reason to believe that patients completing the subjective scores are fundamentally different than those lost to follow-up, the risk of bias is considered to be low. Nevertheless, a validity study has been performed recently, showing no difference in patient-reported outcome scores between responders and nonresponders in the DKRR.35 We therefore

Registry-Based Comparison of Graft Types in ACLR

283

consider our subjective data as valid despite the limited completeness of data. A substantial number of data were missing at the 1-year postoperative clinical control visit. The challenge of obtaining a high percentage of patients at clinical follow-up is substantial when dealing with a national cohort. Again, because of the prospective collection of data, it is unlikely that a lack of these data is associated with both graft choice and later revision. Using revision ACLR as the outcome in this study is looking at only 1 definition of failure after ACLR. Some patients are not willing to undergo revision surgery, accepting a reduction in their activity level and chronic knee instability. In other cases, surgeons do not find patients suitable for revision surgery. Thus, the revision rate may be a conservative measure for the real number of ACL patients with treatment failure. However, there is no reason to believe that patient and surgeon willingness for revision differs in the HT and PT groups. In this multivariate analysis, we included important confounders that have previously been associated with the risk of revision surgery.1,3,6,12 We chose not to include fixation technique as a confounding factor because graft choice is associated with fixation and not the other way around. Thus, fixation is part of the causal path between the graft (exposure) and revision (outcome) and should not be included in the model as a confounder.38 Although we adjusted for a number of potential confounding factors, our study, like all observational studies, may suffer from unmeasured and residual confounding. For example, data on graft size, smoking habit, alcohol consumption, medication use, sports activity during the follow-up period, and occupation were not available in the DKRR.

CONCLUSION This is the first nationwide registry-based cohort study of more than 13,000 ACL knees to present results comparing HT and PT grafts in primary ACLR. This study demonstrated a minimally higher overall revision rate using HT grafts. However, HT grafts were associated with a substantially increased risk of early revision at 1-year follow-up. Thus, the recovery and training period before returning to sport may be of special importance for patients treated with HT autografts in primary ACLR. We conclude that both HT and PT grafts reveal good results after primary ACLR and recommend that graft selection should be based on an individual evaluation of patient demands and graft morbidity.

ACKNOWLEDGMENT The authors acknowledge the help and support of Frank Mehnert at the Department of Clinical Epidemiology and the kind secretarial help of Anne Haagen Hjelm at the Department of Clinical Epidemiology.

REFERENCES 1. Ageberg E, Forssblad M, Herbertsson P, et al. Sex differences in patient-reported outcomes after anterior cruciate ligament

284

2.

3.

4.

5.

6.

7.

8. 9.

10.

11.

12.

13.

14.

15.

16.

17. 18.

19. 20.

21.

22.

Rahr-Wagner et al

reconstruction: data from the Swedish knee ligament register. Am J Sports Med. 2010;38(7):1334-1342. Aglietti P, Giron F, Buzzi R, et al. Anterior cruciate ligament reconstruction: bone-patellar tendon-bone compared with double semitendinosus and gracilis tendon grafts. A prospective, randomized clinical trial. J Bone Joint Surg Am. 2004;86(10):2143-2155. Alentorn-Geli E, Lajara F, Samitier G, et al. The transtibial versus the anteromedial portal technique in the arthroscopic bone-patellar tendon-bone anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc. 2010;18(8):1013-1037. Ayeni OR, Chahal M, Tran MN, et al. Pivot shift as an outcome measure for ACL reconstruction: a systematic review. Knee Surg Sports Traumatol Arthrosc. 2012;20(4):767-777. Barenius B, Nordlander M, Ponzer S, et al. Quality of life and clinical outcome after anterior cruciate ligament reconstruction using patellar tendon graft or quadrupled semitendinosus graft: an 8-year follow-up of a randomized controlled trial. Am J Sports Med. 2010;38(8):1533-1541. Barrett AM, Craft JA, Replogle WH, et al. Anterior cruciate ligament graft failure: a comparison of graft type based on age and Tegner activity level. Am J Sports Med. 2011;39(10):2194-2198. Biau DJ, Katsahian S, Kartus J, et al. Patellar tendon versus hamstring tendon autografts for reconstructing the anterior cruciate ligament: a meta-analysis based on individual patient data. Am J Sports Med. 2009;37(12):2470-2478. Blickenstaff KR, Grana WA, Egle D. Analysis of a semitendinosus autograft in a rabbit model. Am J Sports Med. 1997;25(4):554-559. Danish Health and Medicines Authority. Declaration on Approval of Nationwide and Regional Clinical Quality Databases [in Danish]. Statutory Order 459. June 2006. Ejerhed L, Kartus J, Sernert N, et al. Patellar tendon or semitendinosus tendon autografts for anterior cruciate ligament reconstruction? A prospective randomized study with a two-year follow-up. Am J Sports Med. 2003;31(1):19-25. Feller JA, Webster KE. A randomized comparison of patellar tendon and hamstring tendon anterior cruciate ligament reconstruction. Am J Sports Med. 2003;31(4):564-573. Freedman KB, D’Amato MJ, Nedeff DD, et al. Arthroscopic anterior cruciate ligament reconstruction: a metaanalysis comparing patellar tendon and hamstring tendon autografts. Am J Sports Med. 2003;31(1):2-11. Frobell RB, Roos EM, Roos HP, et al. A randomized trial of treatment for acute anterior cruciate ligament tears. N Engl J Med. 2010;363(4):331-342. Fu FH, Bennett CH, Lattermann C, et al. Current trends in anterior cruciate ligament reconstruction, part 1: biology and biomechanics of reconstruction. Am J Sports Med. 1999;27(6):821-830. Gifstad T, Sole A, Strand T, et al. Long-term follow-up of patellar tendon grafts or hamstring tendon grafts in endoscopic ACL reconstructions. Knee Surg Sports Traumatol Arthrosc. 2013;21(3):576-583. Grana WA, Egle DM, Mahnken R, et al. An analysis of autograft fixation after anterior cruciate ligament reconstruction in a rabbit model. Am J Sports Med. 1994;22(3):344-351. Granan LP, Forssblad M, Lind M, et al. The Scandinavian ACL registries 2004-2007: baseline epidemiology. Acta Orthop. 2009;80(5):563-567. Jonsson H, Riklund-Ahlstrom K, Lind J. Positive pivot shift after ACL reconstruction predicts later osteoarthrosis: 63 patients followed 5-9 years after surgery. Acta Orthop Scand. 2004;75(5):594-599. Kopf S, Kauert R, Halfpaap J, et al. A new quantitative method for pivot shift grading. Knee Surg Sports Traumatol Arthrosc. 2012;20(4):718-723. Laxdal G, Kartus J, Hansson L, et al. A prospective randomized comparison of bone-patellar tendon-bone and hamstring grafts for anterior cruciate ligament reconstruction. Arthroscopy. 2005;21(1):34-42. Li S, Chen Y, Lin Z, et al. A systematic review of randomized controlled clinical trials comparing hamstring autografts versus bonepatellar tendon-bone autografts for the reconstruction of the anterior cruciate ligament. Arch Orthop Trauma Surg. 2012;132(9):1287-1297. Liden M, Ejerhed L, Sernert N, et al. Patellar tendon or semitendinosus tendon autografts for anterior cruciate ligament reconstruction:

The American Journal of Sports Medicine

23.

24.

25.

26.

27.

28.

29. 30.

31.

32.

33. 34.

35.

36.

37.

38.

39.

40.

41. 42.

a prospective, randomized study with a 7-year follow-up. Am J Sports Med. 2007;35(5):740-748. Lind M, Menhert F, Pedersen AB. The first results from the Danish ACL reconstruction registry: epidemiologic and 2 year follow-up results from 5,818 knee ligament reconstructions. Knee Surg Sports Traumatol Arthrosc. 2009;17(2):117-124. Lind M, Menhert F, Pedersen AB. Incidence and outcome after revision anterior cruciate ligament reconstruction: results from the Danish registry for knee ligament reconstructions. Am J Sports Med. 2012;40(7):1551-1557. Magnussen RA, Carey JL, Spindler KP. Does autograft choice determine intermediate-term outcome of ACL reconstruction? Knee Surg Sports Traumatol Arthrosc. 2011;19(3):462-472. Magnussen RA, Granan LP, Dunn WR, et al. Cross-cultural comparison of patients undergoing ACL reconstruction in the United States and Norway. Knee Surg Sports Traumatol Arthrosc. 2010;18(1):98-105. Magnussen RA, Lawrence JT, West RL, et al. Graft size and patient age are predictors of early revision after anterior cruciate ligament reconstruction with hamstring autograft. Arthroscopy. 2012;28(4):526-531. Maletis GB, Cameron SL, Tengan JJ, et al. A prospective randomized study of anterior cruciate ligament reconstruction: a comparison of patellar tendon and quadruple-strand semitendinosus/gracilis tendons fixed with bioabsorbable interference screws. Am J Sports Med. 2007;35(3):384-394. Maletis GB, Incasio MCS, Desmod JL, et al. Reconstruction of the anterior cruciate ligament. Bone Joint J. 2013;95-B(5):623-628. Mohtadi NG, Chan DS, Dainty KN, et al. Patellar tendon versus hamstring tendon autograft for anterior cruciate ligament rupture in adults. Cochrane Database Syst Rev. 2011;9:CD005960. Papageorgiou CD, Ma CB, Abramowitch SD, et al. A multidisciplinary study of the healing of an intraarticular anterior cruciate ligament graft in a goat model. Am J Sports Med. 2001;29(5):620-626. Pedersen AB, Graversen HP, Klareskov B, et al. Dansk Korsba˚nds Rekonstruktions Register: A˚rsrapport 2011. Aarhus, Denmark: SUNTRYK; 2012. Pedersen CB. The Danish Civil Registration System. Scand J Public Health. 2011;39(7 Suppl):22-25. Rahr-Wagner L, Thillemann TM, Pedersen AB, Lind MC. Increased risk of revision after anteromedial compared with transtibial drilling of the femoral tunnel during primary anterior cruciate ligament reconstruction: results from the Danish Knee Ligament Reconstruction Register. Arthroscopy. 2013;29(1):98-105. Rahr-Wagner L, Thillemann TM, Pedersen AB, Lind MC. Validation of 14,500 operated knees registered in the Danish Knee Ligament Reconstruction Register: registration completeness and validity of key variables. Clin Epidemiol. 2013;5(1):219-228. Roe J, Pinczewski LA, Russell VJ, et al. A 7-year follow-up of patellar tendon and hamstring tendon grafts for arthroscopic anterior cruciate ligament reconstruction: differences and similarities. Am J Sports Med. 2005;33(9):1337-1345. Roos EM, Toksvig-Larsen S. Knee injury and Osteoarthritis Outcome Score (KOOS): validation and comparison to the WOMAC in total knee replacement. Health Qual Life Outcomes. 2003;1:17. Rothman KJ, Greenland S, Lash T. Causal diagrams. In: Modern Epidemiology. 3rd ed. Philadelphia: Lippincott Williams & Wilkins; 2008:183-200. Sorensen HT, Sabroe S, Olsen J. A framework for evaluation of secondary data sources for epidemiological research. Int J Epidemiol. 1996;25(2):435-442. Taylor DC, DeBerardino TM, Nelson BJ, et al. Patellar tendon versus hamstring tendon autografts for anterior cruciate ligament reconstruction: a randomized controlled trial using similar femoral and tibial fixation methods. Am J Sports Med. 2009;37(10):1946-1957. Tegner Y, Lysholm J. Rating systems in the evaluation of knee ligament injuries. Clin Orthop Relat Res. 1985;198:43-49. Zebis MK, Bencke J, Andersen LL, et al. The effects of neuromuscular training on knee joint motor control during sidecutting in female elite soccer and handball players. Clin J Sport Med. 2008;18(4):329-337.

For reprints and permission queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav