Abstract
Anterior cruciate ligament (ACL) injuries are among the most common complete ligamentous injuries about the knee. 15 Although the exact incidence is unknown, it is estimated that approximately 150 000 to 200 000 ACLs are torn in the United States each year, and that there are approximately 64 000 to 100 000 ACL reconstructions performed in the United States annually.9,14 These injuries often occur in young, athletic patients, most often from noncontact deceleration with a valgus external rotation force.6,9,14
The “gold standard” for ACL reconstruction has been the patellar tendon autograft, and the most commonly used technique for ACL reconstruction in America today is an arthroscopic-assisted, single-incision reconstruction using bone–patellar tendon–bone (BPTB) autograft with metal interference screws.1,8,10 This is most commonly performed at least 3 weeks after initial injury. 10 However, there is now debate if this is the most ideal type of graft to use and which patient populations would benefit from the different graft types. As stated by Baer and Harner, 2 the optimal graft should reproduce the anatomy and biomechanics of the ACL, be incorporated rapidly with strong fixation, and cause low graft-site morbidity. Autografts (BPTB or hamstring tendon [HT]) have potential disadvantages such as donor-site morbidity, weakness of the quadriceps mechanism with BPTB or weakness of knee flexion with HT, patellar fracture, variable graft sizes, and longer operative times. 16 In contrast, allografts eliminate donor-site morbidity, have more consistent graft sizes, and shorter operative times. However, they also have potentially increased cost, slower incorporation times, and the potential for disease transmission.2,3,17
Many factors can play a role in the successful outcome of ACL reconstruction, including graft type, gender, activity level, and even initial injury mechanism.5,13,19,20,22 The aim of this study was to access ACL reconstruction failure rates by graft type overall and between patients over the age of 25 years and those under 25 years. We hypothesize that patients under 25 years of age will have higher failure rates with ACL reconstruction than those over age 25 years, and that in the patients under 25 years of age, BPTB autograft will have the lowest failure rate.
Materials And Methods
Patient Selection
A computerized relational database (Access 2000, Microsoft Inc, Redmond, Washington) was used by the senior author (G.R.B.) to track ACL surgeries. Waiver for institutional review board approval was obtained and all patients gave informed consent for the purpose of research. Detailed forms were used at the initial visit; surgery; and at follow-up visits at 3, 6, 9, 12, 18, and 24 months. An objective examination consisting of Lachman test, pivot-shift test, range of motion, effusion, and KT-1000 arthrometer (MEDmetric, San Diego, California) testing were done. Subjective testing was done utilizing a series of 15 visual analog scales (VAS) and Lysholm scale scores. Inclusion criteria consisted of patients undergoing primary ACL reconstruction with BPTB autografts (BPTB), hamstring (semitendinosus and gracilis) quadruple-looped autograft, or allograft (posterior tibialis or BPTB). Acute injuries (<3 months from the date of surgery) and chronic instability (>3 months from the date of surgery) as well as contact and noncontact injuries were included. Failure of reconstruction was defined as 2+ Lachman (no end point), a positive pivot shift, and KT-1000 arthrometer results greater than 5-mm side-to-side difference. Patients with bilateral ACL injury, previous ligament surgery (ACL repairs, extra-articular reconstruction), associated collateral ligament (medial/lateral collateral ligament, posterior cruciate ligament) injury, double-bundle hamstring, or quadriceps tendon injury were excluded.
From January 2000 through July 2007, 1131 consecutive ACL reconstructions were performed without extra-articular reconstruction by the senior author. An arthroscopically assisted transtibial technique was used. Eight hundred fifty-five procedures were primary reconstructions using BPTB, quadruple-looped hamstring, or allograft tendon (Table 1). Four hundred seventeen (417) patients satisfied the inclusion criteria and obtained the 24-month minimum follow-up. Follow-up visits consisted of a routine examination, study radiographs, and a detailed form. Two hundred twenty-four (224) patients were 25 years and younger (average, 17.77; range 12-25 years), while 193 were over the age of 25 years (average, 39.13; range, 26-59 years). The ACL graft failures are divided by age and graft type in Figure 1. Graft type selection was based on a combined surgeon and patient preference after an in-depth discussion of the potential risks and benefits of each graft.
Patient Selection a
BPTB, bone–patellar tendon–bone; ST, semitendinosus.
Anterior cruciate ligament graft failures.
Operative Procedure
After a routine 10-minute preparation and draping of the ACL-deficient knee, a thorough examination under anesthesia and a complete diagnostic arthroscopy were performed. Any associated injuries were addressed. No associated ligament procedures were included in this study. Central meniscus (white-white) tears were partially resected while peripheral tears (red-red/red-white) were repaired using an inside-out technique. A guide was used to place a tibial guide pin into the ACL footprint of the tibia. Fluoroscopic lateral examination was done in extension to confirm no notch impingement was present. If the pin was anterior to the intracondylar roof, it was removed and replaced to a better position. Once the guide pin position was deemed to be correct, it was overreamed with a reamer 2 mm smaller than the graft size. Progressively larger reamers were used until the graft size was reached. These reamers were angled slightly more posterior toward the femoral tunnel position. A 5-mm offset guide was then used through this tibial tunnel and placed over the top of the posterior notch. The guide pin that was placed here was also checked by fluoroscopic examination, to confirm that the pin was well posterior but did not violate the posterior cortex. Reaming of the femoral tunnel began with a small reamer (6 mm or 7 mm) and then worked down the lateral wall with progressively larger reamers to the desired size. This allowed the tunnel to be in the more anatomic position, following along the natural ACL footprint. A Beath pin was then passed through the lateral femoral cortex and the graft was pulled into position.
Interference screws were used to secure the femoral bone block through the anterior medial portal and anterior to the bone block in the tibia when using BPTB. EndoButtons (Smith & Nephew Endoscopy, Andover, Massachusetts) were used on the femoral lateral cortex when soft tissue grafts were used. An absorbable aperture or metal screw was used anterior to the graft on the femur and a similar screw was used concentrically in the tibia.
A compression dressing was applied and a hemovac was used for 23 hours. For the first few days, emphasis was placed on passive extension and swelling prevention. Full weightbearing began between 2 and 4 weeks when the patient was comfortable. Associated procedures, such as a meniscus repair, did not change postoperative protocol. Progressive quadriceps and hamstring rehabilitation began at approximately 4 weeks. When the patients had equal quadriceps size and passed sport-specific tests, they were allowed to return to sports at 6 months (at the earliest). Most patients returned to their previous Tegner activity scale score level, although some did not return to their previous activities.
Statistical Analysis
Tests of association between nominal scales were made using the χ2 statistic. An independent t test was used to test for mean differences between groups. An alpha level of .05 was used to determine statistical significance.
Results
Patients 25 years and younger had a significantly higher failure rate (16.5%) compared with patients over the age of 25 years (8.3%) (P = .012). We found a significant association between graft type and failure among patients aged 25 years and younger (P = .021). Comparisons in this age group indicated that both allograft (29.2%) and semitendinosus/gracilis grafts (25.0%) resulted in significantly higher failure rates than BPTB grafts (11.8%) (P = .024 and P = .036, respectively). No significant difference when comparing allograft failure rate (29.2%) to semitendinosus/gracilis failure rate (25.0%) was noted in the age group 25 years and younger. No significant association was demonstrated between graft type and failure among patients older than 25 years of age (P = .150). Analysis of the demonstrated failures revealed that 31 patients aged 25 years and younger (83.8%) had a reinjury episode with instability, while 6 (16.2%) were defined as failed at assessment by definition (KT-1000 arthrometer >5 mm, + Lachman, and + pivot shift). The older-than-25 age group had 4 patients (25%) who experienced a reinjury event with instability, while 12 (75.0%) had a failure by definition. The reinjury rate was significantly higher in the 25-and-younger group compared with the group older than 25 (P < .001). Sixty-six percent of all failures had a reinjury event and 34% failed without such an episode.
When comparing mean Tegner activity scale scores, the average Tegner activity level score for the 25-and-younger age group was 5.73 (range, 2-9) and the average Tegner activity level score was 4.28 (range, 2-7) for the group of patients older than 25 years (P = .005). Patients 25 years and younger in the failure group (5.43) also had a significantly higher mean Tegner activity scale score than the failure group of patients older than 25 years (4.31) (P = .012).
Discussion
Many factors play a role in the outcome of ACL reconstruction, including graft type, gender, activity level, and even initial injury mechanism.5,12,19,20 Several studies have considered age and/or activity level as an important factor affecting outcome.5,7,13,21
Our study demonstrated that recurrent instability after ACL reconstruction is more common in a younger and more active population. However, there are not many studies that assess ACL reconstruction failures in this manner— breaking out a younger age group. Shelbourne et al, 21 who looked at subsequent ACL injury (in either knee) after initial ACL reconstruction using autograft BPTB, found that younger patients had a higher rate of injury after ACL reconstruction. Those younger than 18 years of age had an injury rate of 17.4% versus 6.7% for 18- to 25-year-olds and 3.9% for those older than 25 years. When their patients are divided into age groups as in our study (<25 years and >25 years) looking only at reinjury of reconstructed ACL, the findings are no less dramatic, with those younger than 25 years nearly 6 times more likely to have a graft failure (6.2% versus 1.1%). 21
Lyman et al 13 reported a retrospective analysis of all ACL reconstructions performed in New York from 1998 to 2006. Although this did not distinguish by Current Procedural Terminology (CPT) codes between primary and revision ACL reconstructions and could not specify the laterality (left or right side) of the primary or revision surgeries involved, they found that patients younger than 40 years, and especially those younger than 20 years of age, were more likely to undergo another ACL reconstruction of either knee within 1 year of their initial ACL surgery. They speculated that this could be attributable to higher postoperative activity levels or poorer compliance with the postoperative rehabilitation program and activity restrictions. 13
Brandsson et al 8 also looked at ACL reconstruction based on age. They had 2 groups (>40 vs 20-24 years of age) and found equal results among the 2 with regard to recurrent instability using BPTB autografts. Their older group was overall more satisfied with reconstruction— likely because of maintenance of activity level as seen with Tegner scores (6 to 5 in the >40 years group and 9 to 6 in the 20-24 years group). They had much lower patient numbers (total of 67). Barber et al 4 looked at outcomes of patients older and younger than 40 years of age with ACL reconstruction using allografts and autografts. Although they had small numbers of patients, they found no difference between the 2 groups, indicating that older patients got the same benefits from ACL reconstruction, and all groups did better than their nonoperatively treated counterparts. 4
We looked at a difference in graft type and ACL reconstruction failure. In our study, there was a significantly lower failure rate with BPTB grafts in patients under 25 years of age when compared with both allografts (11.8% vs 29.2%) and HT grafts (11.8% vs 25%). No statistically significant difference was found between BPTB, hamstrings, and allograft failure rate in those over 25 years of age (4.5% vs 6% vs 13%). Hamstring autografts had significantly lower failure rates in patients over 25 years of age (P = .009) than in those 25 years and younger. Salmon et al 20 reviewed their group of ACL reconstruction patients with a 5-year follow-up and found no difference in recurrent instability rates for BPTB versus hamstring reconstructions (6% vs 7%) and could not find any predictors for retear other than a new injury via a contact mechanism (presumably more active patients). Their follow-up was mainly via phone interview, only reexamining those with a history of possible instability, potentially missing some asymptomatic graft failures. They did not break down patients by age groups. However, they did have more reruptures in those more active patients (8% vs 4% for International Knee Documentation Committee [IKDC] 1/2 vs 3/4 activity levels.)
Herrington et al 11 performed a meta-analysis of 13 randomized controlled trials comparing BPTB and HT grafts. They concluded from these studies that there were no statistically significant differences in return to preinjury activities, strength of muscles, range of motion, IKDC scores, or complication rates between the 2 groups. 11 There was, however, slightly increased pain in the BPTB group and greater stability in the BPTB group in some studies. It was also found that more patients in the HT group required reoperation, although this may have been because of differences in fixation technique. 11
Activity level after ACL reconstruction may also play a part in failure rates. 18 In our study, we found that the younger age group generally achieved a significantly higher postoperative Tegner score than the older group did (5.72 vs 4.28, respectively) (P = .005). Theoretically, a higher activity level postoperatively can have a negative effect on the rate of ACL failure. A significant number of our patients 25 years and younger experienced a reinjury event. This may indicate that the activity level and the graft type were major factors in this group of failures. Perhaps, those grafts were intact and functioning before this episode. This would tend to validate our opinion that activity level is a valid concern in ACL failure.
Borchers et al 7 reviewed 21 patients with failure after primary ACL reconstruction who underwent ACL reconstruction with either gracilis/semitendinosus autograft or tibialis tendon allograft for primary and revision surgeries. Using a Marx activity score, they determined the patients’ activity levels both pre- and postoperatively and found that patients with higher activity level at the time of graft failure and those with allograft reconstructions had increased odds ratios (5.53 and 5.56, respectively) of ACL failure. There were no increased odds of failure based on activity level at the time of initial injury.
The strengths of our study include the fact that all surgeries were performed by 1 surgeon using the same arthroscopically assisted single-bundle transtibial technique for every operation, over the same time period. We had information from a large database with 7 years of data and a minimum of 2 years of patient follow-up. The consistency of technique is important. To our knowledge, this is the first published study to examine ACL failure rates comparing age, activity level, and 3 different graft types.
The newer techniques (ie, anteromedial portal drilling of the femoral tunnel) may improve these results. A similar study should be repeated using the newer technique. We are in the early stages of such a study now.
Weaknesses of our study include the fact that the different types of allografts were not separated and compared individually with other grafts. Also, this was a retrospective review of 1 surgeon’s database and not a blinded, randomized, prospective study. In fact, before choosing which graft type to use, the surgeon had a discussion with the patients and/or their families regarding the hypothesized risks/benefits of each graft type and a decision was made with the patients regarding which graft type they wanted to use. Although this is a good medical practice, it could produce a selection bias when considering which graft type to use regarding a patient’s age as well as their activity level. Anticipated activity level was considered before the surgery.
Conclusion
In conclusion, this study showed that, similar to several previous studies, all graft types used had a higher failure rate in the younger, more active patient population. Younger patients achieved a significantly higher level of activity postoperatively and had significantly higher failure rates than did the older patient population. Of the graft types used in our study, the BPTB grafts had the lowest failure rates in all groups. Using this information, we conclude that the BPTB autograft may be a more appropriate graft type to use in the younger population.
Footnotes
Presented at the 36th annual meeting of the AOSSM, Providence, Rhode Island, July 2010.
The authors declared that they have no conflicts of interest in the authorship and publication of this contribution.