Introduction

Reconstruction of the anterior cruciate ligament (ACL) has become an increasingly common orthopaedic procedure. In the United Kingdom, approximately 5000 ACL reconstructions are performed per year¹; in the United States, more than 100,000 procedures are currently performed annually.² This trend is likely to continue with the general population’s increasing pursuit of an active lifestyle. Although long-term functional stability and symptom relief after primary ACL reconstruction exceed 90% in some studies,^3,4 overall clinical failure rates of 10% to 25% have been documented.¹ It is currently estimated that between 3000 and 10,000 U.S. patients and approximately 1000 U.K. patients are candidates for revision ACL surgery annually.¹

Revision ACL surgery is recommended for patients who have instability or reduced activity with pathological laxity after a failed primary ACL reconstruction. The important stages in assessing a patient with a failed ACL reconstruction include a detailed history, patient selection, physical examination and appropriate investigations, choice of graft, surgical technique, and rehabilitation.⁵ Eliciting important relevant history from a patient who is usually apprehensive and has some knowledge of the problem (from general practitioners, physiotherapists, and the Internet) can be difficult. One should spend enough time trying to find out the details of the original injury (e.g., high-velocity trauma may suggest multiligament laxity) and also the patient’s experience of previous treatment as well as his or her knowledge of the problem. A full history of occupational and future recreational activities is mandatory. Often the patient’s expectations are not realistic, and therefore despite achieving knee stability, the revision surgery will not result in a contented patient. Instability and/or pain top the list of patient symptoms. It should be clarified with the patient preoperatively that a reduced activity level and/or excellent natural proprioception may result in a reduction or even an abolition in symptoms of instability without the need for surgery. Revision reconstruction should be offered to patients with symptoms of instability or those who wish to increase their activity level to include manoeuvres involving twisting or a sudden change in direction. In such cases symptoms of instability, if left untreated, will contribute to repeated meniscal and chondral damage, leading to an earlier progression of osteoarthritis (OA). At the same time, the patient should be made aware of the risk of a gradual progression of OA, irrespective of the method of treatment but especially if symptoms of instability are ignored. It should be carefully explained that symptoms of pain are likely to be caused by degenerative disease or a torn meniscus and that a revision ACL reconstruction alone is unlikely to be the answer to this problem. One should also alert the patient to the potential need for bone grafting and thus a staged reconstruction.

The most useful investigations are the plain x-ray series of an anteroposterior (AP) standing radiograph and a lateral x-ray in full extension, together with skyline and Rosenberg views (Fig. 58-1).¹ These will show the original tunnel placement and usually the fixation methods used, tunnel widening,^6,7 osteolysis, and the presence and extent of joint space narrowing. If possible, comparison with previous radiographs will help quantify bone loss (tunnel widening). It is our practice to obtain a magnetic resonance imaging (MRI) scan preoperatively, but only rarely has this provided additional information that has changed the course of management. If no reason for the primary failure can be elicited, vigilance for a missed or complex laxity should be exercised. Careful clinical examination and an examination under anesthesia, which may include fluoroscopic stress views, are useful in finding an additional laxity.

Fig. 58-1 Failed anterior cruciate ligament (ACL) reconstruction. Anteroposterior (A) and lateral (B) views.

Causes Of Failure of Primary Procedure

The causes of recurrent patholaxity after primary ACL reconstruction^6–26 can be broadly divided into four groups: technical errors, failure due to biological factors, failure due to significant trauma, and failure owing to laxity in the secondary restraints.¹⁷ By far the most common cause is error in the surgical technique, with 77% to 95% of all cases of ACL failure attributed to technical error.²⁷ This category includes poor graft selection or harvest, improper tensioning or fixation, and especially incorrect tunnel placement.^9,17,18 More than 70% of technical failures, and thus more than 50% of all ACL failures, can be attributed to malpositioned tunnels.^8,17 Inappropriate positioning of either the tibial or the femoral tunnel results in excessive length changes in the graft as the knee moves through a range of motion, resulting in either a limited range of motion or excessive graft laxity.^8,18 Anterior placement of the femoral tunnel, a common mistake, will result in limited flexion and potential graft failure if full flexion is achieved. Tibial tunnel placement may be somewhat more forgiving, but anterior placement leads to impingement in extension and excessive tension in flexion, whereas posterior placement may cause laxity in flexion.¹⁷ Perhaps the most common error in surgical technique involves the anterior placement of femoral and/or tibial tunnels.¹⁶ Carson et al recently published their review of 90 failed ACL reconstructions and quoted 52% of the failures as being due to surgical technical errors.²⁸ Several studies have shown that a posterior and proximal placement of the intraarticular exit of the femoral tunnel is advisable and involves minimal lengthening of the ACL substitute toward extension.

Treatment Options

Revision ACL surgery is often considered a salvage procedure with very limited goals, distinctly different from those of primary ACL reconstruction.^5,29 Many reports in the literature quote inferior results for revision cases compared with primary reconstruction. However, with many categories of failure, the population of failed ACLs is a diverse group and a difficult subset to study.³⁰

We ask the following questions before deciding on a definitive treatment plan:

Unless the answers to all these questions are positive, we tend to favor a cautious two-stage approach. If revision ACL surgery is staged, then the first stage gives the surgeon an added opportunity to assess the chondral and meniscal pathology and then give the patient a more realistic prognosis.

It is also important to ascertain whether the tibial tunnel of primary surgery will interfere with the correctly placed revision tibial tunnel and the extent of loss of bone stock due to tunnel widening. In addition, one needs to determine whether the hardware needs to be removed and, if so, whether that will further contribute to the loss of bone stock. For the revision graft to function optimally, one needs to ensure that the tunnels are placed in an optimal position in a good-quality bone so that the fixation achieved will be as robust as the primary surgery.

Management of previous tunnel malposition is technically demanding, and different surgeons have used various approaches for dealing with bony defects resulting from incorrectly positioned prior tunnels. In some cases of gross tunnel malposition, a new tunnel may simply be drilled without violating the original tunnel or removing any tunnel hardware. Alternatively, tunnels can be oriented in a divergent pathway that maintains the appropriate articular surface attachment. In many cases, however, new tunnels cannot be drilled without overlapping or breaking into a previous tunnel. Two or more screws can be used to supplement fixation and fill the cavity of an enlarged tunnel. Although this might be useful in limited cases, the fixation achieved tends to be inferior and postoperative rehabilitation may be compromised. Graduated tunnel dilators may allow controlled expansion of a previous tunnel, compacting rather than removing additional bone.¹⁷ In such instances, options include the use of an allograft tendon with an enlarged bony portion, an oversized interference screw, or stacked interference screws.³¹ If the original tunnel is correctly positioned and only slightly larger (3–5 mm) than the new graft, stacking two interference screws may be sufficient to fill the tunnel and secure the graft.^32,33 Battaglia and Miller³⁴ have described use of freeze-dried allograft bone dowels to address bony defects during revision ACL reconstruction. These allografts are readily available and can be easily used to fill deficiencies resulting from previous tunnels or osteolysis. The grafts provide sufficient structural support to allow redrilling of new tunnels through or next to the bone plug. Unfortunately this option implies slower graft incorporation³⁵ and has implications for the rehabilitation regimen.

Some authors have tried to stratify the treatment options by the extent of bony defect. For defects larger than 10 mm, they advocate the use of bone graft and a staged procedure. Although this stratification is useful, it has certain limitations. Accurate preoperative assessment of the tunnel size is difficult and unreliable with plain x-rays. CT is the most accurate method.

After clearing the debris from the earlier drilled tunnel, the resultant defect is almost always larger than 10 mm. Therefore we usually favor a staged reconstruction as a default technique. The first stage involves an EUA, an assessment of chondral and meniscal pathology, the removal of old graft, tunnel curettage, drilling, and bone grafting. The second stage is revision ACL reconstruction after incorporation of bone graft, 3 to 6 months after the first stage, when a CT scan has shown adequate incorporation of the bone graft.

Definition of knee instability

The IKDC classifies knees that are within 2 mm of the normal contralateral knee by means of KT-1000 or similar testing as “normal.”³⁶ Knees that have greater than 5 mm of difference are classified as “abnormal.” The KT-1000 applies a force of 134N to assess knee laxity.

For the past 15 years, we have used the Westminster cruciometer (University College, London) for laxity measurements. It is a validated tool that applies an 89N force during the laxity measurement. Similar to the KT-1000, this cruciometer has been shown to give a reproducible quantitative evaluation of the Lachman test, and a previous study has shown average displacement of normal knee to be 3.2 mm as compared with 8.4 mm in the ACL deficient knee.³⁷ A further validation of the Westminster cruciometer was done recently by comparing the laxity measurements in normal, ACL deficient, and ACL reconstructed knees. The correlation between the Westminster cruciometer and KT-2000 was found to be excellent (Pearson’s coefficient: 97%). The KT-2000 reading can be obtained using the following equation:

Since the original recommendations by the IKDC, various published results have used slightly different criteria in defining knee stability. In addition to instrumented laxity measurements, the clinically relevant pivot-shift test is widely used. The pivot-shift test has various grades, and one needs to be clear in reporting the grade (1+, glide; 2+, clunk; 3+, subluxation) that is being considered as abnormal. In our practice using the Westminster cruciometer (applying a force of 89N rather than 134N) to assess the knees, we use the following criteria to define normal laxity: The side-to-side difference (SSD) in anterior tibial translation is considered normal if within 2 mm and nearly normal if between 3 and 4 mm. Values of 5 mm and greater are considered unsatisfactory. Overall anterior laxity is considered satisfactory if the SSD is less than 5 mm and the pivot shift is absent or 1+ (glide). In the presence of an SSD greater than 4 mm and/or a pivot shift of 2+ (clunk) or 3+ (subluxation), the anterior laxity is considered unsatisfactory.

Surgical procedure

Stage II

The second stage includes a further EUA, arthroscopy, relevant meniscal and chondral surgery, graft harvest, and revision ACL reconstruction. Our choice of graft is described in Table 58-1.

Table 58-1 Choice of Graft for Revision Surgery

Primary Graft	Revision Graft
Bone–patellar tendon–bone (BPTB)	Four-strand hamstring
Four-strand hamstring	BPTB
Prosthetic	Ipsilateral BPTB graft or four-strand hamstring

The revision procedure itself is similar to any primary procedure, and attention is given to achieve the correct anatomical placement of the tunnels. Because the landmarks are often less distinct than in a primary case, the tibial tunnel should be referenced off the PCL on the medial side of the mid-intercondylar point and the femoral tunnel referenced from the over-the-top position. We have favored the use of proprietary jigs (femoral Puddu guide, Acufex). Perioperative imaging using an image intensifier may occasionally be necessary to ensure optimal tunnel placement (Fig. 58-2).

Fig. 58-2 Radiographs after second-stage revision. Anteroposterior (A) and lateral (B) views.

The correctly placed revision tibial tunnel is drilled. For the femoral tunnel, the technique used at the time of revision surgery is different than that used at the time of primary surgery. If an inside-out technique was used during the primary reconstruction, the femoral tunnel is drilled from outside-in (and vice versa) to ensure the new revision tunnel placement in virgin bone.

Graft Choice: Autograft Versus Allograft

For a long time it has been debated whether an autograft or an allograft should be used for revision ACL reconstruction.^38–41 Allografts have certain advantages. The donor site morbidity is eliminated, which may help during the rehabilitation. When weighed against the total costs of a two-staged ACL reconstruction, their use could be financially justified. However, they do have specific risks. Viral transmission of hepatitis, human immunodeficiency (HIV) virus, or other infection is a concern.⁴² Allografts tend to integrate more slowly than autografts and can cause immunological reactions, which may interfere with the healing process; hence the recommendation of a slower rehabilitation protocol.⁴¹ Furthermore, the sterilization process used may decrease the mechanical properties of the allograft. In addition, this increases the cost. As of March 11, 2002, the Centers for Disease Control and Prevention (CDC) had received 26 reports of bacterial infections from musculoskeletal allografts.⁴³ Because the notification of infection secondary to use of allografts is voluntary, it is likely that its true incidence is underestimated. We agree with the concept that every effort should be made to ensure killing of bacteria and bacterial spores with the help of available technologies. These added risks of using an allograft have led us to refrain from their routine use. An allograft should only be considered when host material is scarce. This is sometimes the case in patients with multi-ligament laxity. One can use ipsilateral and contralateral bone–patellar tendon–bone (BPTB) as well as hamstring tendons but may still have to use allografts for a multidirectional laxity involving the ACL, PCL, posterolateral corner (PLC), and/or medial collateral ligament (MCL).^44–46

We have always favored the use of autograft rather than allograft. It is not our practice routinely to use the contralateral limb for harvesting the graft, although some surgeons prefer the contralateral limb in the primary or revision setting.⁴⁶ We do not have any experience of using reharvested BPTB or four-strand hamstring (4-SH) graft; reports in the literature of their use and satisfactory clinical outcome are few.⁴⁷

Graft Fixation: Cortical or Apertural

The techniques for graft fixation during the revision procedure are similar to those used in the primary procedure. When dealing with bone–bone fixation, the interference screw,⁴⁸ is our traditional method of fixation, although if the femoral tunnel is a tight fit, Rigidfix (Mitek Products, Ethicon, Edinburgh) is satisfactory. For hamstring fixation, we use IntraFix (Mitek) on the tibial side; for a cortical fixation on the femoral side, a Corin anchor (Corin Group, The Corinium Centre, Cirencester, United Kingdom) is used.

The types of fixation method used on both the femoral and tibial sides play a crucial role in the stability achieved after ACL reconstruction. The fixation devices must be able to withstand early postoperative forces until graft–tunnel healing has occurred. The fixation should facilitate graft tunnel healing, producing a normal histological transition zone between the host bone and the new ligament.^45,49

The fixation methods can be broadly classified as cortical (suspensory) or apertural (intratunnel).⁵⁰ There is a belief that cortical fixation may not perform as well as aperture fixation and that there may be a “bungee effect” causing reduced stability because fixation is farther from the joint, resulting in a longer graft with reduced stiffness. Cyclical elastic stretching under loading can be expected to increase with lengthening of the graft between the points of fixation. Anchoring the graft distant to the joint line may also allow AP movement, described as a “windshield wiper” effect after widening of the tunnel. However, this is debatable. Graft fixation relies on the friction between the graft and the fixation device until the graft integrates with the surrounding host tissues.⁵⁰ Older methods such as in-line staples (tibial side) or simple buttons (femoral side) have a high chance of graft slippage due to their dependence on “simple friction” between the fixation device and a smooth, compressible soft tissue graft. This prevents the grafts being satisfactorily tightened and held. On the other hand, techniques such as Endobutton (Smith & Nephew, Andover, MA), Corin Anchor, and newer tibial devices rely on “complex friction” and thereby ensure better stability.^51–56 Prodromos et al,⁵⁰ in a recent meta-analysis of ACL reconstruction, considered stability after ACL reconstruction as a function of hamstring versus patellar tendon graft and fixation type. The authors concluded that four-strand hamstrings had overall higher stability than BPTB and the graft stability was fixation dependent. Four-strand hamstring grafts with Endobutton femoral fixation and second-generation tibial cortical fixation (belt-buckle staple configuration or interference screws augmented with staples) resulted in higher stability than all other graft/fixation combinations. Therefore either the bungee effect does not exist or, if it does exist, it seems inconsequential. Also, it is likely that the bungee effect is only likely during the early postoperative period before the bone tunnels have healed around the graft, converting cortical to apertural fixation.⁵⁰

Different authors have assessed^{34,49,57–60} the fixation strengths of various femoral and tibial fixation devices used for ACL reconstruction. Harvey et al in their succinct review article considered the different types of fixations used along with results of laboratory testing.⁴⁹ These are summarized in Tables 58-2, 58-3, and 58-4.⁴⁹

Tibial fixation is commonly considered more problematic than femoral fixation because forces on the ACL substitute are parallel to the tibial drill hole,^60,61 the bone quality of the tibial metaphysis is inferior to that of the femur,^61,62 and the four-tailed end of the hamstring tendon graft that is fixed to the tibia is more difficult to secure. The WasherLoc secures the graft at the external tibial aperture, and the tandem spiked washers have an even longer working length because they are placed completely outside the tibial tunnel. The IntraFix may be considered a semi-aperture fixation because the 30-mm plastic sheath extrudes distally from the entrance of the tibial drill hole and thus, in a normal tibial tunnel of 35 to 45 mm in length, leaves 5 to 15 mm of free graft within the proximal opening (aperture) of the drill hole. Finally, interference screws can be considered truly anatomical (apertural) fixations because they can be advanced to the internal tibial tunnel orifice.

When interference screws are used in the tibia, they are inserted from the outside-in, producing forces that are counter to the direction of the tension on the graft, as opposed to the femoral side, where the screw is placed from the inside-out, thus wedging the graft during screw insertion. The strength of fixation of interference screws is influenced by several variables, such as the density of the bone,⁶¹ the insertion torque,⁶³ the geometry^55,64,65 and material of the screw,⁶⁶ and the length and diameter of the screw.⁶⁷ Different considerations may be important in the fixation of hamstring and BPTB grafts. Increasing the diameter of the screw increases the fixation of the hamstring by a press-fit mechanism that crushes the surrounding cancellous bone. However, poor engagement of the thread into the tendon may make this less important and length of the screw more so. Engagement of the thread into a corticocancellous BPTB block gives good fixation, which is influenced more by the changes in the diameter of the screw and less so by changes in the length (Fig. 58–3).⁴⁹

Fig. 58-3 Scan of a failed anterior cruciate ligament (ACL) reconstruction demonstrating a huge tibial defect.

Postoperative rehabilitation

All patients start knee flexion on the first day after surgery, and resting with the heel supported is encouraged to achieve full hyperextension. Patients are encouraged to perform static quadriceps exercises to prevent a quadriceps lag, and ice therapy is used regularly to ensure an early reduction in swelling. The patients are mobilized on the first or second day after surgery with elbow crutches, which are used until a good gait pattern has been achieved. The regimen is continued at home with emphasis in the first 2 weeks on achieving full hyperextension, flexion past 90 degrees, a reduction in swelling using ice elevation, rest, non–weight-bearing exercises, and minimal walking. The rehabilitation program is continued as an outpatient with a series of graduated mobilizing, strengthening (isometric, closed chain, and [later] some open chain), and dynamic stability exercises. Running is started from 10 weeks or when a “quiet” knee (i.e., minimal pain and swelling) has been achieved. Rehabilitation programs are individually tailored to include sports-specific training, and patients can return to contact sports from 6 months. Patients with meniscal or chondral deficiency may be advised to avoid high-impact training and activities.

Brown and Carson² suggested that an accelerated rehabilitation program⁶⁸ for revision ACL reconstruction is not appropriate due to weaker initial graft fixation. We have found that this is not necessary, as using a two-stage technique ensures that there is good quality bone around the tunnels and initial graft fixation is as secure as in the primary reconstruction. We have followed the same rehabilitation program for both primary and revision ACL patients and have not found any significant difference between the objective and subjective laxity assessment at follow-up between the primary and revision ACL reconstruction.

Our experience with a two-stage revision Anterior Cruciate Ligament reconstruction

From 1991 to 2003 the senior author performed 75 revision ACL reconstructions. Of these, nine were performed as a single-stage revision, and 11 were performed in patients with multi-ligamentous laxity needing attention. The remaining 55 patients underwent revision ACL reconstruction using a two-stage technique with bone grafting of the tibial tunnel (Fig. 58-4). Of these 55 patients, 49 had a minimum follow-up of 3 years or more, and we compared this group of patients with a matched cohort of 49 patients with primary ACL reconstruction. These results were recently published in the American Journal of Sports Medicine.¹ The salient findings of this study are summarized here.

Fig. 58-4 Scan of the proximal tibia 6 months after bone grafting the tibial tunnel.

The average age of the patients at the time of revision ACL reconstruction was 35.4 years (range 26–42 years). None of the patients was lost to follow-up in this study. The mean follow-up was 6 years (range 3–11 years). Preoperatively, all knees had positive Lachman and pivot-shift tests. The pivot shift was graded 1+ (glide) in four knees, 2+ (clunk) in 34 knees, and 3+ (subluxation) in 11 knees. This improved to pivot-shift grades of 0 in 43 knees, 1+ in five knees, and 2+ (clunk) in one knee. The mean laxity measurement (SSD) using Westminster cruciometer was 1.36 mm (standard deviation [SD]: 1.11), and this was not significantly different from the primary reconstructions (mean 1.2 mm; SD: 1.5). In one patient from the revision ACL group, the graft stretched out (the patient suffers from generalized ligamentous laxity) 4 years after the revision surgery, and the patient is awaiting rerevision surgery.

Technical error was the most common reason for graft failure (femoral, 28 cases; tibial, 20 cases; both femoral and tibial, four cases). Tunnel enlargement was seen in all the cases. The mean tibial tunnel measurements were 13.7 mm (SD: 2.5 mm) on AP and 13.9 mm (SD: 2.3 mm) on lateral radiographs.

The mean IKDC subjective and objective scores were lower for the revision group compared with the primary group. On analysis of the subjective scores, the main differences noticed between the two groups were in the pain level and the activity level. On analysis of the objective IKDC scores, main differences were noticed in passive motion deficit and finding of crepitus in various compartments.

None of the revision ACL group patients in our study returned to original level of activity (pre–ACL injury). This can be explained by the presence of associated meniscal and chondral pathology and should form an important part of the counseling offered to revision ACL surgery patients prior to their operation.

Patients who have a prosthetic ligament as the primary graft and are undergoing subsequent revision surgery merit separate discussion. In all cases, the first-stage revision surgery was more demanding and time consuming, as the synthetic graft evoked a lot of synovitis and scarring within the knee. Extra care was needed to identify the PCL before clearing the intercondylar notch. The extent of tunnel enlargement was also more pronounced in these cases.

Review of literature

Noyes and Barber-Westin reported on 55 patients who had a revision ACL reconstruction with a BPTB autograft.³³ The failure rate, which was determined in a fashion similar to that in their revision allograft report, was 24%. Of these 13 patients, six had a reharvested patellar tendon autograft. This was a heterogenous group, and the authors stratified the group of 55 knees into those who had an ACL reconstruction only, those who required a staged high tibial osteotomy, and those who required a concurrent ligament reconstructive procedure. The group requiring only ACL reconstruction had a failure rate of 16% (5/32 knees). Earlier, the same research group had also published their results of revision ACL surgery with the use of a BPTB allograft.³⁹ They noted an incidence of 33% failure at a mean of 42 months. In this series, the allografts were obtained from tissue banks certified by the American Association of Tissue Banks (AATB) and were fresh-frozen at the time of procurement. The grafts had been sterilized with 25,000 gray of gamma irradiation. This amount of low-dose irradiation probably does not alter the mechanical properties of the graft. In fact, AATB has advocated use of low-dose irradiation for many years and several authors have used it for ACL allografts to improve the protection against bacterial contamination.^42,43 Noyes et al advocate that allograft should not be considered as the first choice of graft for revision surgery. If no autograft is available for revision surgery, they advise augmentation of the allograft with the lateral extraarticular iliotibial band procedure to reduce the high failure rate associated with the use of allograft.³³

Fox et al recently published their results of revision ACL reconstruction using nonirradiated patellar tendon allograft.⁶⁹ Thirty-two of 38 patients (84%) were available for follow-up. The mean patient age was 28 years with a mean follow-up of 4.8 years (range 2.1–12.1). This is a good homogenous group of patients with critical evaluation of results. None of the patients in the series had meniscal allograft surgery, posterolateral reconstructions, high tibial osteotomies, or contralateral ACL deficiency or reconstruction. Of this patient group, 87% were subjectively satisfied, 87% had 0/1+ pivot shift, and 84% had a KT-1000 SSD of less than 3 mm. The authors quote a failure rate of 28% using stringent criteria, namely the presence of a positive pivot-shift test (grade 1/2/3) and/or a KT-1000 result of more than 5 mm SSD. If the criteria used for definition are changed to SSD greater than 5 mm and pivot-shift grade 2 or more, then the failure rate is just 6%. The authors very correctly point out that the criteria used for defining failure after revision surgery are variable, which further complicates the comparison.

The Pittsburgh series of Johnson et al^20,21,48 used irradiated, fresh-frozen allografts. Nine of the 25 patients had a KT-1000 maximum manual difference of greater than 5.5 mm. Eighty percent had grade 0 or 1 Lachman result, and 20% had grade 2 Lachman result; 76% of patients were satisfied with the results.²¹

Uribe et al reported on 54 patients with revision ACL reconstruction using a variety of grafts including ipsilateral patellar tendon autograft, contralateral patellar tendon autograft, allograft patellar tendon, and hamstring autograft. All the patients had an improvement in their objective stability; however, only 54% of the patients returned to their pre–ACL injury activity level.⁷⁰

Battaglia et al³⁴ recently published their experience of revision ACL reconstruction using freeze-dried allograft bone dowels. The advantages of these grafts are their ready availability, the elimination of donor site (iliac crest) morbidity, and their ability to provide sufficient structural support for the new tunnels. Although all these proposed benefits are true, with passage of time the allograft may become resorbed, compromising the stability of the ACL reconstruction.

Our results compare favorably to those published in the literature^{28,33,34,39,69–71} with regard to laxity measurements, and our failure rate is significantly lower. In only one case the graft had failed at 52 months and the patient complained of instability requiring revision, giving a failure rate of 2.04%. In another patient, the cruciometer reading was 5 mm, suggestive of increased laxity. However, this patient is coping well at present and has not had any further surgical intervention. These results have been achieved despite an uncompromised rehabilitation regimen, and we believe that this can be attributed at least partly to the two-stage technique we used, which allows for the consolidation of the bone graft in the tibial tunnel. These figures also represent the “worst-case” scenario, as no patient was lost to follow-up.

This study has certain limitations. In an ideal world, we would have set up a randomized controlled trial comparing the results of a two-stage revision surgery with a one-stage revision surgery to highlight the differences (if any). Although a two-stage surgery provides good bone for placement of the tibial tunnel, it exposes the patient to another surgical intervention. This may indeed have negative effects on the patient’s range of motion and pain after surgery and also may prolong the rehabilitation period. In the period of the past 10 years, the senior surgeon has performed nine revision ACL reconstructions using a single-stage technique. As this is quite a small number, we did not compare the results with one-stage revisions. In our opinion, a single-stage technique should only be used if the placement of the primary tibial tunnel is so incorrect that it will not overlap at all with the correctly placed revision tunnel. This was rarely the case in our series with both tunnels.

In this study, we decided to graft the tibial tunnel, as there was no other way of bypassing the defect in the bone created during the primary surgery. Bone grafting the primary tibial tunnel and ensuring its adequate healing prior to proceeding to the second stage of revision reconstruction ensured that the new tunnel could be drilled through an area with adequate bone stock, which in turn would not compromise the fixation achieved or the postoperative rehabilitation. In cases of femoral tunnels in this series, it was always possible to alter the technique of drilling the femoral tunnel (outside-in versus inside-out), thereby ensuring that the bony defect created during the primary procedure was avoided.

Revision ACL surgery is both a challenging and a rewarding enterprise for the surgeon. It is challenging in that it uses all the surgeon’s experience and communication skills in dealing with these complex cases and patients. It also tests the surgeon’s surgical expertise and knowledge of fixation techniques and biological healing and thus the best predictable outcome for each patient. The treatment process benefits from a mature organizational setup and a competent team. The rewards are self-evident.