Study Criteria

All studies met three criteria for inclusion, as follows:

1 Minimum 2-year follow-up

2 Application of at least 30 pounds or maximum manual testing force

3 Stratified presentation of stability data, not just averages

Statistical Methods

Meta-analytic methods were used to compare the groups. Weighted means for normal and abnormal stability were generated as follows: For each treatment group, the proportion of individuals for an outcome event was determined by adding the number of events that occurred through all studies and dividing by the number of the patients in all the studies. The number of patients for any single-center, but not multi-center, study was capped at 100 to avoid disproportionate influence of any given study. In any study, if the number of patients for a given outcome was not recorded, the study was eliminated from the analysis of that outcome. For example, with one exception, if a study did not record the number of patients who have less than 2 mm of difference, then the study was not included in the <2 mm comparison.²^–⁶⁶

Results

Table 69-1 shows all studies broken down by graft and then by fixation type for all graft types.

Table 69-1 Clinical Series Divided by Graft and Fixation Method

Only About Half of Reconstructed Knees Achieve Stability Symmetrical with the Other Knee

The goal of ACL surgery is to restore the preinjury level of stability, which should be the same as that of the other knee. True symmetry is achieved when the side-to-side difference (SSD) between the knees is 0. Measurement error increases this criterion to 1 mm. Thus we propose a SSD of 1 mm as defining knee stability symmetry. The IKDC “normal” criterion of up to a 2-mm difference may be satisfactory, but it is not truly normal. Indeed, a 2-mm SSD is what is commonly seen with partially torn ACLs.⁶⁷ When the 1-mm criterion is applied, we see the following: For all autografts, about 30% have greater than 2-mm SSD.⁶⁸ The remaining 70% fall into four categories: 2 mm, 1 mm, 0, or less than 0. If we assume that one-fourth of the 70% falls into each of these four categories, then it is reasonable to estimate that one-fourth of 70%, or 18%, are exactly 2 mm different. Adding this 18% (exactly 2 mm) to the 30% (greater than 2 mm) would mean that 48% of the reconstructed population has a 2-mm or greater SSD. This leaves about 52% with 1 mm or less SSD (i.e., true symmetry with the other knee). Thus roughly one-half of the autograft ACLRs, in the hands of the experienced knee surgeons who are the authors of these studies, have stability that is either equivalent to a partially torn ACL or worse. The allograft data⁶⁸ show significantly lower stability rates (see Table 69-1).

Table 69-1 presents the raw data for stability from all the studies. The principal areas of interest are the “normal” and “abnormal” stability columns. Abnormal stability in most cases is equivalent to graft failure. The primary table subdivision is by graft type. These are four-strand hamstring (4HS) autograft, two-strand hamstring (2HS) autograft, BPTB autograft, and quadriceps tendon autograft and allograft. The secondary subdivision is by graft subgroup and by fixation type. Subdividing by fixation groups is possible to do with the autografts because of the large number of studies. It is only possible with the allografts to break out a BPTB/interference subgroup because of the smaller number of studies.

Stability Rates by Graft Type

The principal findings were as follows:

1 The 4HS graft group normal stability rate of 77% was significantly higher than the BPTB rate of 66% (P < 0.001). The 4HS abnormal rate of 4.2% was significantly lower than the BPTB rate of 5.9% (P = 0.029).

2 The autograft normal stability rate of 71% was significantly higher than the allograft normal stability rate of 59% (P <0.001). The autograft abnormal stability rate of 5.2% was significantly lower than the allograft abnormal stability rate of 15% (P <0.001).

3 The BPTB autograft normal stability rate of 66% was significantly higher than the BPTB allograft normal stability rate of 56% (P <0.001). The BPTB autograft abnormal stability rate of 5.9% was significantly lower than the BPTB allograft abnormal stability rate of 17% (P <0.001).

4 Radiated allografts had significantly lower stability rates than nonirradiated allografts. Normal and abnormal rates for radiated were 32% and 41%, respectively, versus 63% and 12% for nonirradiated grafts (P <0.001).

5 The two-strand hamstring normal stability rate of 53% and abnormal stability rate of 13% were both significantly worse than the rates for both 4HS and BPTB (P <0.001).

6 There is only one published quadriceps tendon study with 2-year follow-up and stratified stability rates. The normal stability rate was 75%, and the abnormal rate was 6%.

Stability Rate by Graft Fixation Subgroups

Stability rates for these subgroups are as follows:

1 The 4HS-EB2 subgroup, which used an Endobutton on the femur and second-generation fixation on the tibia, had the highest stability rates of any graft fixation subgroup. Its normal and abnormal stability rates were 77% and 1.7%, respectively. By comparison, the 4HS group with interference screws on both the tibia and femur, 4HS-2IS, had normal and abnormal rates of 75% and 5.4%, respectively.

2 For BPTB, interference screw fixation had slightly higher stability rates than noninterference screw fixation: normal and abnormal rates of 68% and 5.0%, respectively, for interference screws versus 63% and 7.4% for noninterference screws. However, some of the overall highest stability rates were obtained by noninterference screw fixation methods.

Aperture Versus Nonaperture Fixation

For soft tissue grafts, there was no stability advantage for aperture fixation. Indeed, as just described, the 4HS-EB2 group, which had entirely nonaperture cortical fixation, had the highest stability rates of any graft fixation subgroup. Thus the so-called “bungee effect” would appear to be nonexistent. This is not surprising, as studies have shown that fixation on rigid cortical bone enhances stiffness much more than the greater length of the construct diminishes it.⁶⁹ Also, variations in fixation stiffness are eliminated once the graft heals into the bone tunnel because the fixation is no longer load bearing. What really matters is how much the graft may elongate during this healing period from either slippage or plastic deformation. Neither of these is related to fixation stiffness.

Four-Strand Hamstring Versus Bone–Patellar Tendon–Bone Stability Rates

Because BPTB has long been considered the “gold standard” for ACLR. it may be surprising to some that 4HS was found to have higher stability rates. However, as the 4HS graft is a significantly stronger graft (see Chapter 10), the excellent 4HS stability rates do make sense if modern fixation is used. Prior analyses have seemed to show that hamstring grafts had lower stability rates than BPTB.^70,⁷¹ However, these studies commingled the much-lower-stability 2HS studies with the higher-stability 4HS studies. If the 2HS studies are removed, it turns out that there was no stability advantage to BPTB in those studies by comparison with 4HS. Also, many of the highest-stability 4HS series were published after those studies. They were thus not included in those analyses but contribute significantly to the higher 4HS stability rates found here.