At the Canadian Orthopaedic Association meeting in 1971, Galway and MacIntosh presented their description of the pivot shift, considered by many the first description of this phenomenon [34]. Included in this paper was a brief description of an extra-articular technique, using a strip of ITB routed under the LCL (Fig. 33.2). In 1976, MacIntosh presented the results of 90 cases operated using his technique [35]. Termed the “lateral substitution reconstruction for the anterior cruciate ligament,” this procedure utilized a 20 cm strip of ITB, left attached at Gerdy’s tubercle, which was routed under the LCL, through a subperiosteal tunnel and around the insertion of the lateral intermuscular septum, and finally back under the LCL. A combined intra- and extra-articular variant, with an intra-articular limb created by routing the graft “over the top” and through the knee, was also described. This procedure was later known as the MacIntosh II. A third procedure, the MacIntosh III, used a graft formed from the quadriceps tendon, prepatellar periosteum, and patellar tendon, in continuity and left attached distally. The graft was passed over the top from inside out.

Ellison presented a description of his extra-articular procedure at the American Orthopedic Society for Sports Medicine meeting in 1975, before formally publishing the technique and results of his distal iliotibial band transfer in 1979 [36]. In his technique, the ITB was elevated from Gerdy’s tubercle with a button of bone, routed under the proximal LCL, and reattached at or just anteriorly to Gerdy’s tubercle with a staple. He also advocated plication of the middle third capsular ligament using a double breasted repair beneath the LCL (Fig. 33.3).

Most modern techniques are modifications of the Lemaire and MacIntosh procedures. In general, these techniques do not double the graft back to the tibia and thus use a shorter ITB graft and require a shorter skin incision [37]. The graft may be passed over or under the LCL and affixed to the femur using a variety of methods, including a staple or interference screw.

Neyret has described a technique using a patellar tendon intra-articular graft and a gracilis tendon graft for the extra-articular component [38]. The gracilis is threaded through a drill hole in one of the bone blocks, creating one continuous graft. The patellar tendon graft is passed in an anterograde fashion, locking the gracilis tendon in the femoral tunnel with press fit of the bony block (Fig. 33.4). The two free limbs are then passed deep to the LCL and through either end of a bony tunnel through Gerdy’s tubercle.

The anterior cruciate ligament is the primary restraint to anterior tibial translation. A number of structures contribute to the control of internal tibial rotation at the knee, including the ACL, the anterolateral ligament, the iliotibial band, the lateral meniscus [56], and the medial meniscotibial ligament [57].

Despite current interest in the ALL, to date relatively few biomechanical studies have been published. Kennedy examined the biomechanical properties and failure mechanism of the ALL [55]. They determined a mean maximum load of 175 N and stiffness of 20 N/mm. In 12 specimens, they identified four mechanisms of failure, ligamentous tear at the femoral attachment in four specimens, at the tibial insertion in one, in the mid-substance in four, and by a bony avulsion (i.e., Ségond fracture) in six, although it should be noted that the line of pull in these experiments was nonphysiologic. Regarding function, Dodds determined the ligament to be isometric from 0 to 60° of flexion and to lengthen with internal tibial rotation, strongly suggesting a role in rotational control [20]. Kittl studied the isometry of the native anterolateral structures as well as potential points for the fixation of an extra-articular reconstruction [58]. He found an ALL with an origin posterior and proximal to the LCL to be relatively isometric, while an ALL with a distal and anterior origin was lax approaching extension and unlikely to be effective in controlling the pivot shift [58]. Monaco examined the effect of cutting the ACL and lateral capsular ligament using a navigation system and manually applied forces [59]. His description of division of the lateral capsular ligament would have involved division of the ALL. He found an increase in internal rotation in all knee flexion angles in the ACL-deficient knee following division of the lateral capsular ligament, which was significant at 30° with an increase in internal rotation of 5.5°. Spencer investigated both sectioning and reconstruction of the ALL using navigation and manually applied forces. He measured an increase in internal rotation in extension of 2° after division of the ALL in the ACL-deficient knee while performing a simulated pivot shift [60]. Lording, in a cadaveric experiment using a robotic knee examination device, found division of the ALL in the ACL intact knee increased internal rotation at 30° of knee flexion by 2.4° [61]. However, there was wide variation in the effect of ALL sectioning between specimens, which in some specimens was not significant. Parsons, using a six degree of freedom robot, found the ALL to be the primary restraint to internal rotation at knee flexion angles greater than 35°, with the ACL providing the greatest restraint closer to extension [21]. It should be noted that the ITB was removed from all specimens in this study prior to testing. In contrast to Parsons, Kittl found the ALL played no significant role in internal rotational control [62]. In a similar robotic experiment, he determined the superficial and deep components of the ITB to be the primary restraints to internal rotation from 30 to 90°, with the ACL having a significant contribution at 0° only. Interestingly, the ACL provided no restraint to the pivot shift.

The finding of a role for the ITB in the control of internal tibial rotation is important but not new. Fetto was able to induce a pivot shift by division of the ITB in an ACL intact knee [14]. Jakob noted increased internal rotation but a paradoxical decrease in the pivot shift after division of Kaplan’s fibers, reflecting the complex and multifactorial nature of these rotational abnormalities [26]. When he released the ITB distally by osteotomy of Gerdy’s tubercle, the pivot shift became so marked in the ACL-deficient knee that the subluxation did not reduce before 60° of flexion. Gadikota, in a robotic study investigating the effect of increasing ITB load, found that internal rotation was significantly reduced between 20 and 30° of knee flexion with an ITB load of 50 N and from 15 to 30° with a load of 100 N [63]. Lording measured an increase in internal rotation of 2.6° in the ACL intact knee after division of the ITB at Gerdy’s tubercle, slightly greater than that noted for the ALL and consistent across specimens [61].

Despite many advances in the evolution of anterior cruciate ligament surgery, failure remains an issue. While failure rates as high as 24 % have been reported [64], recent large-scale cohort studies, systematic reviews, and registry reports would suggest a rate between 3.5 % and 7 % [65–67]. There is no universally accepted and applied definition of failure, however, and studies reporting failure rates using hard end-points such as revision reconstruction likely underestimate the true clinical burden. Pain, stiffness, ongoing instability, and an inability to return to sports may all signify a failed procedure, particularly as they relate to patient satisfaction.

It is now well understood that intra-articular ACL reconstruction does not restore normal knee biomechanics with regard to rotational control [1–5] and that this in turn has a negative impact on patient outcomes [6–8]. In an effort to better restore normal kinematics, various modifications of intra-articular techniques have been used.

In the double-bundle technique, the posterolateral bundle is intended to better restrain internal rotation and the pivot shift [68]. While time-zero biomechanical testing has suggested this technique offers superior rotational control than single-bundle techniques [69–71], clinical superiority has not been demonstrated [72–74].

In “anatomical” single-bundle reconstruction, the femoral tunnel is placed in the footprint of the native ACL, rather than the more vertical position seen in traditional techniques. This creates greater graft obliquity [75], which should theoretically better resist rotation and improve stability, although the results of biomechanical studies have been mixed [76–79]. While “anatomical” single-bundle techniques have demonstrated improved patient-reported outcomes compared to traditional techniques [80, 81], these more oblique grafts are subjected to higher in situ forces than more vertical grafts [82], which may lead to a higher graft failure rate [83].

Regardless of graft obliquity, the ACL is poorly positioned to resist internal rotation and probably contributes relatively little to rotational stability [62, 84]. As outlined above, recent anatomic and biomechanical data support the role of the anterolateral peripheral structures in rotational control. It is likely that damage to these peripheral restraints contributes to the variation in clinical laxity seen in the ACL-deficient knee [85, 86] and that failure to address these associated injuries contributes to residual rotatory laxity and poor outcomes [87]. In this light, repair or reconstruction of these structures could be considered more “anatomical” than intra-articular reconstruction alone.

Some biomechanical data is available to assess the effect of extra-articular augmentation on rotational control. Draganich studied the effect of both an isolated extra-articular reconstruction and a combined approach in a cadaveric model [88]. The isolated lateral procedure was found to over-constrain tibial internal rotation; however, when the lateral procedure was performed after an intra-articular reconstruction and care was taken not to tension the tenodesis with the knee in external rotation, both anterior translation and rotation were restored to that of the intact knee. In an in vivo study using intraoperative navigation, Monaco demonstrated reduced internal rotation after the augmentation of an intra-articular graft with a lateral extra-articular reconstruction [89]. This combination also showed improved rotational control compared to a double-bundle technique. Spencer examined the effect of an anatomical ALL reconstruction and a modified Lemaire extra-articular procedure in the ACL-deficient knee [60]. The anatomical ALL reconstruction, based on the landmarks of Claes [15], was ineffective in controlling internal rotation or anterior translation in an early phase pivot shift test, supporting the isometry findings of Kittl [58]. With the modified Lemaire reconstruction, however, there was a trend toward reduced internal rotation and a significant reduction in anterior translation.

One concern regarding extra-articular procedures is that this improved internal rotational control comes at the expense of over-constraint of the lateral compartment. While early studies of isolated procedures would support this [45–47], these techniques often called for graft fixation with the knee in maximal external rotation, and this finding was not borne out by Draganich for combined reconstructions [88]. In Kittl’s recent study, graft passage deep to the LCL and attaching proximal to the lateral femoral condyle demonstrated near isometric behavior [58]. Both the MacIntosh and modified Lemaire demonstrated favorable length change behavior. Similarly, there is no evidence that lateral extra-articular procedures cause increased lateral compartment osteoarthritis. Zaffagnini, in a randomized trial comparing patellar tendon, four-strand hamstring, and Marcacci’s combined technique, noted no differences in radiological outcomes at 5 years [90].

Lateral extra-articular procedures may also work synergistically to reduce the failure rate for ACL reconstructions. Terry described the ACL and the capsulo-osseous layer of the ITB as forming an “inverted U” behind the lateral femoral condyle, supporting the condyle and preventing posterior subluxation on the stabilized tibia [85]. Draganich demonstrated load sharing between intra-articular and extra-articular reconstructions in a cadaveric model [88]. Similarly, Engebretsen found that an iliotibial tenodesis reduced the forces seen in an ACL graft by 43 % [47]. These studies suggest that the addition of an extra-articular procedure could shield an intra-articular reconstruction from excessive forces during the healing phase, potentially protecting it from early stretch or fixation failure and reducing the rate of re-injury in the long term. This may be of particular importance for more grafts likely to see higher forces, such as more oblique single-bundle grafts and patients after medial meniscectomy [91], as well as patient groups at higher risk of failure, such as younger and female patients [66, 92, 93].

The first combined procedures were performed soon after the emergence of lateral extra-articular techniques. Some, such as the MacIntosh II, were inherently combined procedures, while others involved the augmentation of an intra-articular reconstruction with a separate lateral procedure.

The early results for combined procedures were encouraging. Bertoia reported good or excellent results in 31 of 34 knees using the MacIntosh lateral substitution over-the-top repair (MacIntosh II), with the pivot shift abolished in 91 % [94]. Zarins and Rowe described a modification of MacIntosh’s over-the-top procedure, with an ITB graft passing from outside in supplemented by the addition of a distally based semitendinosus graft passing from inside out [95]. Eighty-eight of 100 patients reported good or excellent satisfaction with the procedure, with pivot shift reduced to grade 0 or 1+ in 91.

Augmentation procedures also showed promising results. Dejour studied 251 cases operated with a patellar tendon intra-articular reconstruction augmented with the Lemaire procedure [96]. Eighty-three percent had good or excellent functional results, although the pivot shift was described as equivocal in 24 %. Rackemann reported the results of 714 knees treated with a medial third patellar tendon reconstruction augmented with a MacIntosh procedure [97]. At 6 years, results were satisfactory in 93 %, with only one positive pivot shift.

The first comparative study of intra- and extra-articular reconstruction versus intra-articular reconstruction alone was published by Jensen in 1983 [98]. In this retrospective study, he found the combined procedure group showed the most marked reduction in anterolateral laxity. Subsequent studies, however, challenged the superiority of combined procedures. Strum reported no benefit of combined procedures over isolated intra-articular reconstructions, stressing the importance of a well-performed intra-articular procedure [99]. O’Brien found no difference in clinical stability for those treated with a central third patellar tendon intra-articular graft with or without the addition of a lateral extra-articular sling procedure; however, 40 % of the extra-articular group had chronic pain or swelling associated with the additional procedure [100]. In the first English language, randomized, prospective study, Anderson compared patellar tendon, hamstring, and hamstring combined with lateral extra-articular procedures and found no benefit to the addition of the extra-articular reconstruction [101].

By this stage, lateral extra-articular reconstructions had been largely abandoned, although a number of centers, particularly in Europe, continued to utilize the technique. Some long-term case series are available from these institutions.

In Lyon, France, both Lerat and Neyret have published long-term results for combined procedures. Lerat reported the results for 138 patients at a mean follow-up of 11.7 years [102]. Patients were treated with a “MacInJones” procedure, in which an intra-articular patellar tendon graft was augmented by an extra-articular reconstruction performed with a strip of quadriceps tendon in continuity with the patellar tendon graft. International Knee Documentation Committee (IKDC) functional results were good or excellent in 60 %. The pivot shift was negative in 66 %, grade 1+ in 30 %, and grade 2+ in 4 %. There were 12 graft failures. Pernin and Neyret reported the long-term outcomes of 100 patients treated by Henri Dejour with a patellar tendon intra-articular reconstruction and a modified Lemaire procedure, at a mean follow-up of 24.5 years and with particular respect to the risk factors for the development of osteoarthritis [103]. The intra-articular reconstruction was performed in an open fashion through an anteromedial arthrotomy. Seventy-four percent reported their outcome to be good or excellent, with IKDC assessment normal or near normal in 46 %. The pivot shift was negative in 77 %, with 17 % having a moderate pivot (2+) and 6 % a gross pivot (3+). Radiographically, the percentage of knees without degenerative changes was stable from 11.5 years (41 %) to 24.5 years (39 %); however, among those with degenerative changes, the proportion with severe osteoarthritis increased from 10 % to 27 %. Both medial meniscectomy and medial articular cartilage lesions at the time of surgery were predictive of the development of osteoarthritis, as were increased age at operation and increased delay between injury and surgery. Residual laxity was not found to correlate with the radiological outcome; however, only anterior translation and not rotatory laxity was assessed.

In Italy, Marcacci described a technique not dissimilar to the hamstring arm of the Zarins-Rowe procedure [104]. The gracilis and semitendinosus tendons are harvested with their tibial insertions maintained. The graft is passed through a tibial bone tunnel and then over the top of the lateral condyle from inside out. The tendons are affixed in a groove on the lateral femur with two staples, and the remaining graft passed deep to the LCL and attached at Gerdy’s tubercle. At 11-year follow-up in 54 knees in high-level sports participants, 90.7 % achieved good or excellent International Knee Documentation Committee (IKDC) scores, with three knees showing a slight residual pivot shift [105]. No increase in osteoarthritis was noted for this combined procedure compared to historical controls.

A number of small, randomized studies comparing combined versus isolated intra-articular reconstruction have been published [90, 101, 106–111]. These are summarized in Table 33.1. Ait Si Selmi evaluated the outcomes of 120 patients randomized to receive either a patellar tendon intra-articular reconstruction or Neyret’s combined procedure [106]. The combined group showed improved satisfaction, IKDC subjective scores, and improved control of the pivot shift, although these differences did not reach statistical significance. The study of Giraud reports the medium-term results of 63 patients treated with either the “MacInJones” or a patellar tendon reconstruction at 7 years [108]. While there was a trend toward improved IKDC scores and a reduction in pivot shift seen in the extra-articular group, this did not reach statistical significance in this small trial.