TECHNICAL CONSIDERATIONS

The majority of the studies used “aperture fixation” with interference screws for graft fixation on both the femoral and tibial sides of the joint. Metal screws were the preferred fixation for both grafts in the earlier studies, with a trend to bioabsorbable screws for fixation of soft-tissue grafts in more recent investigations. Two studies used extracortical fixation with either plate or endobutton (Smith & Nephew, Andover, Massachusetts, USA) fixation on the femoral side for hamstring grafts.^17,18 Maletis and coworkers¹⁹ used bioabsorbable screws on both the femoral and tibial sides; this is the only study to use identical fixation for both grafts. None of the studies specified the tension at which the grafts were fixed. The flexion angle of the knee at the time of fixation was inconsistently reported and varied from hyperextension to up to 70 degrees of flexion. Rehabilitation protocols were similar, as were follow-up schedules.

JOINT LAXITY

All reconstructions were performed endoscopically through single tibial incisions. Instrumented laxity measurements utilizing the KT-1000 arthrometer were common to all 6 studies; however, the forces at which the knees were tested (89 vs. 134 N vs. maximum manual) and the method of reporting (side-to-side difference vs. proportion of patients with mild/moderate/gross laxity) varied. Despite these differences, no statistically significant differences in terms of instrumented laxity between the 2 grafts were found in any of the individual studies. A trend existed toward increased side-to-side laxity with hamstring tendon grafts in all 6 studies; however, the differences were often small (<3 mm) and of uncertain clinical significance.

CLINICAL OUTCOMES

Strength and range of motion were inconsistently reported among the studies. Although early follow-up (<1 year) often demonstrated strength differences, no consistent trends were noted across the studies for either graft in terms of flexion and extension strength at 2 years after reconstruction. Maletis and coworkers¹⁹ found persistent flexion weakness in the hamstring group and persistent quadriceps weakness in the patellar tendon group, but the deficits were small and of uncertain clinical significance. Similarly, although extension deficits were seen early in the BPTB group in several studies, the groups usually equilibrated at 2-year follow-up. In 2 studies, the extension deficit persisted.^17,20

Two studies demonstrated a statistically significant proportion of patients in the hamstring group with an abnormal pivot shift examination at final follow-up.^17,19 The severity of the pivot shift (i.e., a “glide” vs. an overt subluxation) was specified in only 1 study, and no patients demonstrated more than a grade 1 (glide).¹⁹ No differences were found between the grafts with respect to the one-leg hop test in the studies that reported this outcome.

FUNCTIONAL OUTCOMES

None of the investigations used a validated measure of health-related quality of life. The majority of the studies reported activity ratings and clinical composite measures that included some index of patient satisfaction. One study included the 36-Item Short Form Health Survey (SF-36), a generic measure of health-related quality of life, as part of the postoperative assessment.¹⁹ A significant increase in physical function occurred between years 1 and 2 after ACL reconstruction in both groups. No differences were demonstrated between the grafts at 2 years.

The Lysholm score is a 100-point scoring system for examining a patient’s knee-specific symptoms including mechanical locking, instability, pain, swelling, stair climbing, and squatting.²² Lysholm scores were used to assess outcome in 5 of the 6 studies. Significant improvement was found from preoperative to postoperative scores in all of these studies. No significant differences were noted in the Lysholm scores between the 2 grafts. Similarly, Tegner scores, a patient-reported measure of activity level, improved from preoperative to postoperative assessments in five studies and did not show a statistical difference between the grafts at final assessment. In 1 study, more patients were able to return to their preoperative activity level with a BPTB graft.¹⁹

Knee-specific composite scores were frequently used to assess outcomes. These scores combine objective physician-assessed clinical data with subjective functional data obtained from patients. No differences were seen between the grafts in any of the studies utilizing the Cincinnati knee score or the International Knee Documentation Committee (IKDC) scores (both composite and subjective components).

COMPLICATIONS

Anterior knee pain is an important outcome after ACL reconstruction. Anterior knee pain was variably assessed among the group of studies, with some investigators reporting the incidence of anterior knee pain and others providing patient-reported scores or “knee walking/kneeling” ability. In all the investigations, knee pain was significantly greater after patellar tendon grafts were used. The incidence of knee pain before reconstruction was inconsistently reported.

Graft failure, rerupture, or both were poorly defined among the five studies. Twelve failures (5 BPTB, 11 hamstring) were reported before the 2-year final follow-up examination for an overall rate of 3.3%. No difference was reported in the rate of early failure between the two grafts.

Septic joint infections were rare with five reported from the six investigations (1%). Three cases resolved with arthroscopic lavage, administration of intravenous antibiotics, and retention of the original graft. In two cases, a staged revision reconstruction was necessary. No difference was seen in the rate of deep infection for the two grafts. Superficial infections were inconsistently reported, as were the rates of deep venous thrombosis, pretibial numbness, and arthro-fibrosis.

VALIDITY/QUALITY ASSESSMENT

Despite the fact that these investigations were all randomized studies, from an epidemiologic perspective, the quality of the trials was variable. All six studies were relatively small. Only two studies specified a primary outcome of interest and performed an a priori sample size calculation accordingly. Loss to follow-up was poorly reported and not accounted for in any of the analyses. Neither post hoc power calculations nor calculation of confidence intervals was used in any of the studies, all of which failed to show major differences between the two grafts.

DISCUSSION

Utilizing only Level I, strictly randomized data, researchers found no major differences between the two grafts with respect to laxity, patient-assessed outcomes, or subjective scores (Table 93-1). The finding of a persistent pivot shift in a greater proportion of patients with hamstring reconstruction in 2 studies is an important observation. Both anterior knee pain and kneeling pain were experienced by more patients after BPTB grafts in all studies. The trials were international, including 4 Scandinavian, 2 American, and 1 Australian trial. The study group was young, active, and predominantly male. This would be consistent with the average patient requesting ACL reconstruction in the North American population.

In summarizing existing evidence, limitations are introduced by the heterogeneity and quality of the individual studies. In terms of heterogeneity, although the individual study designs and sample populations were similar—all were randomized trials conducted in young, active, predominantly male populations—the technical aspects of graft harvest, fixation, and tensioning were variable and may have introduced potentially important differences in graft survival and function at 2 years. Variations in epidemiologic quality among the individual trials might have led to overestimation or underestimation of treatment effects. Publication bias might have played a role in the original literature search because only 1 foreign language study was identified.⁸

Although this review has provided some insight into autograft selection, many unanswered questions remain regarding ACL reconstruction. There is a complete lack of data in the literature to guide decision making as to the optimal timing of ACL reconstruction. No Level I investigations compare acute ACL reconstruction (performed within 3 months of injury) with chronic reconstructions. Furthermore, in the Level I investigations that were included in this review, no attempts were made to stratify randomization based on prognostic indicators such as concomitant cartilage, ligament or meniscal injury, the presence of ligamentous hypermobility, or body mass index. Such stratification could provide important prognostic information regarding these factors and guide future research initiatives.

Several previous reviews have been previously published comparing these 2 grafts with inconsistent findings. A “review of reviews” was undertaken by Poolman and coworkers²³ to investigate the reasons for the discordant findings. Eleven such reviews were identified in that investigation, with 3 favoring the patellar tendon graft for stability and 6 favoring the hamstring graft to prevent anterior knee pain. The quality of the individual reviews was highly variable and believed to be the most significant reason for the differences in their findings.

Limitations are imposed on this analysis because of inconsistent reporting of the outcome measures across the investigations and the lack of validated, patient-based measures of function. Although we suspect from these data that neither graft has a propensity for failure outright, a definitive conclusion on superiority awaits functional data at longer-term follow-up. The finding of a trend toward increased incidence of anterior knee pain with patellar tendon grafts should be interpreted in light of the fact that few studies record whether this pain preceded the reconstruction. Furthermore, the association of the increase in laxity (as assessed by the pivot shift examination) seen with hamstring grafts and patient function or sport participation is unknown.

The breadth of research on ACL reconstruction is expanding and evolving. The literature is plentiful, but its impact is weakened by underpowered investigations. A place exists for a large, multicenter trial to definitively explore this question with long-term functional outcomes; however, further small investigations are unnecessary and, we would suggest, an inefficient use of time and resource.

This brief narrative review is intended to provide the surgeon with up-to-date data, which can be used to appropriately educate his or her patients in the consent process before ACL reconstruction. Furthermore, this work represents the beginnings of a more extensive analysis that will appear in the Cochrane database. A wealth of data can be derived from this sample of Level I randomized trials on ACL reconstruction. Further inference may be possible with the raw data from each study and/or a formal meta-analysis as per the Cochrane database methodology. The Cochrane review will attempt to provide overall estimates of treatment effect via statistical pooling of study data. Definitive information, however, can come only from a large randomized trial with long-term functional data as a primary end point. Table 93-2 provides a summary of recommendations.

TABLE 93-2 Summary of Recommendations