Download pdf - Revision Anterior Cruciate Ligament Surgeryacldoc.net/Articles/Revision ACL Recon.pdfInstructional Course 102 Revision Anterior Cruciate Ligament Surgery Bernard R. Bach, Jr, M.D

Aoaalpruasaufifr

pPgipsfCaott

t1

Hb

1

Instructional Course 102

Revision Anterior Cruciate Ligament Surgery

Bernard R. Bach, Jr, M.D.

jt

nterior cruciate ligament (ACL) reconstructionsurgery is one of the most frequently performed

perative procedures in orthopaedic surgery. Therere estimates that 100,000 new ACL injuries occurnnually. Marked improvements have occurred in theast 15 years with regard to graft selection, tunnellacement, graft fixation, and rehabilitation that haveesulted in predictable outcomes for ACL surgerysing patellar tendon, hamstring, quadriceps tendon,utograft, and allograft tissues. Nevertheless, mosttudies report a clinical failure rate of between 10%nd 15% at short- and intermediate-term follow-ps.1-3 Increasing numbers of ACL reconstructionailures are being seen.4-11 The purpose of this articles to discuss the etiology of failure, approaches to theailed ACL patient, surgical technique issues, andesults of revision ACL surgery.7

The author has been in practice since 1986 and haserformed nearly 1,200 ACL reconstructive surgeries.atellar tendon autograft has been the predominantraft source in over 90% of the primary ACL surger-es, and between 75 and 100 ACL reconstructions areerformed annually. Of note is that my personal revi-ion rate has been less than 1%, although our clinicalollow-up studies would suggest a 10% failure rate.urrently, between 5 and 10 ACL failures are revisednnually; the majority of these patients are referred tour center. Nonirradiated patellar tendon allografts arehe primary graft choice in the majority of these pa-ients.

My revision experience is reflective of the matura-ion of my practice. In the first 5 years of, between986 and 1991, I had minimal experience (n � 2) with

Address correspondence to Bernard R. Bach, Jr., M.D., 1725 W.arrison St, Suite 1063, Chicago, IL 60612, U.S.A. E-mail:[email protected]© 2003 by the Arthroscopy Association of North America0749-8063/03/1910-0103$30.00/0

udoi:10.1016/j.arthro.2003.09.044

4 Arthroscopy: The Journal of Arthroscopic and Related Surger

evision surgery. In the second 5 years, between 1991nd 1996, 19 patients underwent revision reconstruc-ion. In the third 5-year period, 34 patients were re-erred for revision surgery. Over the last 2 years2001-2003) 26 revisions were performed. Througheptember 2003, the author has performed 81 revisionCL reconstructions, 71 of whom were referred to ourractice. More than 80% of these patients underwentonirradiated patellar tendon allograft revision sur-ery (Fig 1), 15% underwent patellar tendon autograftevision, and in 5% of the patients other tissues weresed. The majority of patients referred to our practicead failed patellar tendon autograft surgery. Thirty-hree of 34 of these individuals underwent revisionndoscopically; 19 were initially reconstructed endo-copically and 14 were performed using a 2-incisionrthroscopic technique. This is reflective of the patel-ar tendon being the predominant graft choice in thehicago area. Eight patients had a primary allograft

econstruction, 2 of whom were revised with a patellarendon autograft and 6 were revised with a nonirradi-ted patellar tendon allograft performed endoscopi-ally. Six patients had a primary hamstring recon-truction, 4 of whom were revised with a patellarendon autograft, and 2 with a patellar tendon allo-raft. One individual reconstructed with a patellarendon autograft was revised as a hamstring 2-incisionechnique. Failed extra-articular primary reconstruc-ions and primary repairs with augmentation werexcluded. This underscores the need to have a varietyf options available with regards to revision recon-truction.

ETIOLOGY OF FAILURE

It is well documented in the literature that the ma-ority of patients, when carefully analyzed, have aechnical component that may contribute to graft fail-

ratf(SApngruhtesalCrtacstgttteos

re. If a reconstruction fails within the first 6 months,

y, Vol 19, No 10 (December, Suppl 1), 2003: pp 14-29

a technical component usually plays a role (C. Harner,personal communication, May 2, 2003). This is gen-erally related to tunnel placement. Common technicalerrors include an anteriorized femoral tunnel (Fig 2), a

vertically oriented femoral tunnel (Fig 3), anteriorizedtibial tunnel (Fig 4), a posteriorized tibial tunnel (Fig5) and, less frequently, a posterior cortical “blow out”(Fig 6). Inadequate graft fixation may play a role infailure with marked femoral screw divergence, os-teopenic bone, or graft construct mismatch resulting in

FIGURE 1. A nonirradiated whole patellar tendon allograft withquadriceps tendon.

FIGURE 2. This lateral view shows the concept of a nonanatomicinitial femoral tunnel with a nonoverlapping secondary tunnel. Thissituation generally will not require hardware removal or primarygrafting. (Reprinted with permission from Bach BR Jr, MazzoccaA, Fox JA. Revision anterior cruciate ligament surgery. In: GranaWA, ed. Orthopaedic knowledge online. Rosemont, IL: AmericanAcademy of Orthopaedic Surgeons, 2003. Available at www.aaos.org/oko. Accessed May 15, 2003.7)

FIGURE 3. Anteroposterior view of right knee shows the conceptof a vertically oriented femoral tunnel (1). The initial tibial tunnelwas been placed in a sagittal plane, thus impacting femoral socketcreation and reorientation of tunnels using the divergent tunnelconcept. Note that there is minimal overlap on the femoral tunnelsand there should be an adequate femoral tube to allow unstagedgrafting. The initial femoral screw (1) can be advanced to precludethe possibility of having to bone graft this defect. On the tibial side,the reoriented tibial tunnel provides adequate intact tube to allowfor fixation. If there is significant overlap with these tunnels, it maynecessitate using a stacked screw fixation to provide distal fixation.(Reprinted with permission from Bach BR Jr, Mazzocca A, FoxJA. Revision anterior cruciate ligament surgery. In: Grana WA, ed.Orthopaedic knowledge online. Rosemont, IL: American Academyof Orthopaedic Surgeons, 2003. Available at www.aaos.org/oko.Accessed May 15, 2003.7)

15REVISION ACL SURGERY

inadequate tibial fixation.12,13 Several patients havepresented to our office with loss of fixation on thefemoral side with intra-articular migration of the fem-oral bone plug (Fig 7).14 Inadequate primary graftsource may be a contributing factor either with anarrow patellar tendon autograft or inadequate con-struct. Recent biomechanical testing has demonstratedthat hamstring autograft tissue with multiple loop con-structs provide excellent strength but fixation, incor-

poration and creep may play factors in potential fail-ure.15-21 Biological issues with regard to hyper-elasticindividuals and graft incorporation failure may con-tribute to failures. Medialization of the tibial tunnelmay result in graft abrasion occurring secondary toimpingement on the posterior cruciate ligament. Lat-eral wall abrasion, roof impingement, or interferencescrew abrasion may contribute to failure. Finally, un-recognized or unaddressed patholaxities may contrib-ute to primary graft failure. These include loss of asecondary restraint, such as a medial meniscectomy,chronic medial collateral ligament laxity, or unrecog-nized posterolateral laxity.22 Finally, the internal bio-logic milieu in an individual with early degenerativejoint disease may create a hostile environment forgraft incorporation and maturation.

FIGURE 4. Lateral view shows a nonoverlapped femoral tunneland an overlapped tibial tunnel. In this situation, to provide ananterior tunnel buttress, stacked screw fixation may be necessary toprovide adequate fixation at the time of revision. (Reprinted withpermission from Bach BR Jr, Mazzocca A, Fox JA. Revisionanterior cruciate ligament surgery. In: Grana WA, ed. Orthopaedicknowledge online. Rosemont, IL: American Academy of Ortho-paedic Surgeons, 2003. Available at www.aaos.org/oko. AccessedMay 15, 2003.7)

FIGURE 5. This lateral view shows the problem of an excessivelyposteriorized tibial tunnel. It is less problematic when a bone-tendon-bone graft has been used but can be a significant problemwhen soft tissue grafts are used. The revision tibial tunnel has beenanteriorized and has minimal overlap. Primary or staged bonegrafting may be necessary in situations where a soft tissue graft hasbeen used at the time of the initial reconstruction. (Reprinted withpermission from Bach BR Jr, Mazzocca A, Fox JA. Revisionanterior cruciate ligament surgery. In: Grana WA, ed. Orthopaedicknowledge online. Rosemont, IL: American Academy of Ortho-paedic Surgeons, 2003. Available at www.aaos.org/oko. AccessedMay 15, 2003.7)

16 B. R. BACH, JR.

It is critical when assessing patients in the office toattempt to determine the cause of failure. Radiograph-ically, one should be able to determine whether thefemoral or tibial tunnels are inadequately positioned.We have noted a transition from anteriorized femoraltunnels to vertically oriented femoral tunnels as acontributing technical cause of failure (Fig 3). Assurgeons have made a transition to an endoscopictechnique, many have not recognized the importanceof tibial tunnel orientation and how it impacts on

femoral tunnel placement. Although radiographicallysurgeons are much more consistent about placing thefemoral tunnel posterior in contrast to the 2-incisiontechnique, the anteroposterior radiographs frequentlydemonstrate a vertically oriented graft. This may re-sult in what is interpreted as a near normal Lachmantest with a firm endpoint, but unfortunately does notcontrol rotation. Therefore, patients may have lowKT-1000 side-to-side differences, clinical complaintsof instability and demonstrable low-grade pivot-shifttests.23,24 The other technical error that appears to beevolving is a transition from an anteriorized tibialtunnel to a posteriorized tibial tunnel (Fig 7). Onceagain, this is reflective of a transition from a 2-incisiontechnique to an endoscopic technique.

Radiographically, one should attempt to determine

FIGURE 6. This right knee arthroscopic photograph shows that theover-the-top posterior cortex has been violated. The depth andinvolvement of the circumference of the “blow out” may impactwhether this may be salvaged as a pure endoscopic technique ortransitioned to a 2-incision approach. (Reprinted withpermission.25)

FIGURE 7. This sagittal magnetic resonance image shows a loss offixation of the femoral bone plug that contributed to failure in thisindividual. This patient presented for consultation and transfer ofcare, mechanical symptoms, and instability. Nonirradiated allograftwas used for revision surgery.

FIGURE 8. Right knee lateral view shows a nonanatomic over-lapped femoral tunnel. This may be problematic in that there maybe anterior wall insufficiency. Depending on the amount of expan-sion and the amount of overlap, this may necessitate a staged bonegrafting procedure that may be approached with a primary graftingof the femoral tunnel redrilling of the revision tunnel. (Reprintedwith permission from Bach BR Jr, Mazzocca A, Fox JA. Revisionanterior cruciate ligament surgery. In: Grana WA, ed. Orthopaedicknowledge online. Rosemont, IL: American Academy of Ortho-paedic Surgeons, 2003. Available at www.aaos.org/oko. AccessedMay 15, 2003.7)


whether the bone tunnels are anatomic or nonanatomic(Figs 2 and 8). Additionally, one must determinewhether these tunnels are expanded or nonexpandedand whether they are overlapped or nonoverlapped(Fig 9). This may impact decisions for staged revisionsurgery or customized bone grafts. Figure 2 showsnonanatomic nonoverlapped femoral tunnels. In thissituation, revision can easily be performed withoutconcerns for customized graft or staged grafting. Fig-ure 8 shows a nonanatomic overlapped tunnel, whichis more problematic. If this is not recognized preop-eratively, one may place the graft back into the orig-

inal malpositioned tunnel or the graft may have inad-equate graft fixation. In this situation, depending onthe amount of overlap, one may need to primarilygraft the abnormally positioned tunnel and return in astaged fashion to perform the definitive revision. Thesenior author has performed staged procedures infre-quently, whereas some surgeons prefer them (C.Harner, personal communication, May 2, 2003). Ourchoice of using nonirradiated bone-tendon-bone allo-graft provides flexibility that allows us to potentiallycustomize the bone plugs to compensate for potentialgraft tunnel overlap.

It is important to recognize that, at the time of theindex reconstruction, the posterior cortical wall (i.e.,the over-the-top position) may have been purposely orinadvertently blown out (Fig 10). Some surgeons pre-fer to purposely blow out the posterior cortical edge sothat they can maximize posterior placement of their

FIGURE 9. This right knee lateral view shows tunnel expansion ofboth the tibial and femoral tunnel. In this situation, a staged bonegrafting procedure would be recommended. (Reprinted with per-mission from Bach BR Jr, Mazzocca A, Fox JA. Revision anteriorcruciate ligament surgery. In: Grana WA, ed. Orthopaedic knowl-edge online. Rosemont, IL: American Academy of OrthopaedicSurgeons, 2003. Available at www.aaos.org/oko. Accessed May15, 2003.7)

FIGURE 10. This right knee lateral view shows a posterior corticalwall “blow out” at the time of index reconstruction. The orientationof the femoral tunnel using divergent tunnel concept may allow forrevision endoscopic or 2-incision revision reconstruction. (Re-printed with permission from Bach BR Jr, Mazzocca A, Fox JA.Revision anterior cruciate ligament surgery. In: Grana WA, ed.Orthopaedic knowledge online. Rosemont, IL: American Academyof Orthopaedic Surgeons, 2003. Available at www.aaos.org/oko.Accessed May 15, 2003.7)

18 B. R. BACH, JR.

graft. This has been proposed by some hamstringadvocates (P. Fowler, personal communication,March 2003). One should also be cognizant of previ-ous Gore-Tex prosthetic graft reconstructions (W.L.Gore, Tempe, AZ). These were performed with anover-the-top 2-incision or arthrotomy technique andthe over-the-top position was purposely grooved. Inthese situations, one may not be able to radiographi-cally discern that the posterior cortical wall has beenviolated. If, at the time of revision surgery, the pos-terior cortical wall is determined to be incompetent,there are several technical options (Fig 11). Theseinclude over-the-top fixation, creation of an endo-scopic femoral socket, and using extra-cortical fixa-tion such as an EndoButton (Arthrex, Naples, FL) or

conversion to a standard 2-incision approach with analtered orientation of outside-in femoral tunnel dril-ling.

Osseous tunnel expansion can present its own set oftechnical challenges (Fig 9). Most often, this can beappreciated radiographically, but in specific situations,imaging studies, either magnetic resonance imagingor, preferably, computed tomographic scanning, isadvisable. These include the use of soft tissue grafts asit appears that expansion occurs more commonly withthe use of hamstring constructs and certainly morecommonly with an Achilles tendon allograft or pros-thetic implants such as a Gore-Tex graft or a ligamentaugmentation device (LAD; 3M, Minneapolis, MN).In any patients who had a Gore-Tex graft we wouldconsider obtaining further imaging to better delineatethe magnitude of tunnel expansion. Revision surgeryfollowing Gore-Tex graft usage or an LAD createsparticularly adherent tissue to the osseous walls and,therefore, during the course of removal of this tissuecan result in further expansion and violation of thetunnels. Index reconstructions performed with theseconstructs may require primary bone grafting withsecondarily staged revision reconstruction (Fig 12).

An important principle is the “divergent tunnel”concept (Figs 3 and 13).13,25 It is important to recog-nize the intra-articular anatomic site as the ACL canbe approached from several different orientations. Forexample, radiographically, the angle of femoral tunneldrilling differs with an endoscopic technique despiteidentical tissue placement (Fig 14). As many endo-scopic failures may result in a more vertically orientedfemoral tunnel, one can approach the revision with a2-incision arthroscopic approach that results in a moredivergent tunnel and an intact femoral tube, whichwould allow rigid graft fixation (Fig 15).

The divergent tunnel concept can be applied toeither the femoral or tibial tunnel. For example, thetibial tunnel is frequently placed in a more sagittalorientation (Fig 3). To achieve proper orientation ofthe femoral tunnel, one can use a more medial en-trance site on the tibial metaphysis and once againachieve a nearly intact tibial osseous tube for graftfixation.

Another issue that needs to be considered preoper-atively is pre-existing hardware. The surgeon mustdetermine whether the hardware must be removed orwhether it can be potentially by-passed (Fig 16).26

Hardware removal is not mandatory. Nevertheless,one may approach the procedure with the hopes ofbypassing the hardware and determine intraopera-tively that it needs to be removed. The surgeon must

FIGURE 11. In the event of posterior cortical wall insufficiency,EndoButton fixation is an excellent alternative with or withoutinterference screw aperture fixation. (Reprinted with permissionfrom Bach BR Jr, Mazzocca A, Fox JA. Revision anterior cruciateligament surgery. In: Grana WA, ed. Orthopaedic knowledge on-line. Rosemont, IL: American Academy of Orthopaedic Surgeons,2003. Available at www.aaos.org/oko. Accessed May 15, 2003.7)


FIGURE 12. (A) A significant tunnel expansion, (B) staged bonegrafting, and (C) redrilling of the tibial tunnel are shown. Revisionshould be deferred for at least 4 months after grafting. (Reprintedwith permission from Bach BR Jr, Mazzocca A, Fox JA. Revisionanterior cruciate ligament surgery. In: Grana WA, ed. Orthopaedicknowledge online. Rosemont, IL: American Academy of Ortho-paedic Surgeons, 2003. Available at www.aaos.org/oko. AccessedMay 15, 2003.7)

20 B. R. BACH, JR.

also give consideration for the potential problem of astripped screw and have a game plan for hardwareremoval. It is important to recognize that there are avariety of interference screws that are commerciallyavailable that have different dimension hexagonalscrew head recesses (Fig 17). In addition, one com-mercially available screw requires a threaded inserter

and extractor (Instrument Makar, Okemos, MI).26

When performing revision ACL surgery, one needs tohave a complete set of screwdrivers available as wellas easy-in and easy-out screw extractors. There arecommercially available revision ACL hardware re-moval sets available. Staple removal along the medialtibial metaphyseal cortex may be problematic. Insome circumstances, this may result in considerablecortical bone violation, which potentially may impactgraft fixation. This may necessitate augmentation fix-ation at the time of revision surgery, may alter weight-bearing status, or may necessitate a staged bone graft-ing and reconstruction procedure.

FIGURE 13. This right knee flexed-knee notch view shows theconcept of the divergent cone, which can be applied either to anendoscopic or 2-incision technique. (Reprinted with permission.13)

FIGURE 14. Three-dimensional computed tomography cadavericreconstruction of endoscopic and 2-incision tunnels that have beendrilled and filled with barium sulfate. It shows the differences intheir orientation by the common intra-articular placement. (Re-printed with permission.25)

FIGURE 15. In the situation of a failed endoscopic reconstruction,one can either revise endoscopically or convert to a 2-incisiontechnique. In part, this may be related to how much tunnel expan-sion or overlap may exist. Conversely, if the initial procedure wasa 2-incision technique that failed, one could make a transition to arevision endoscopic technique. (Reprinted with permission fromBach BR Jr, Mazzocca A, Fox JA. Revision anterior cruciateligament surgery. In: Grana WA, ed. Orthopaedic knowledge on-line. Rosemont, IL: American Academy of Orthopaedic Surgeons,2003. Available at www.aaos.org/oko. Accessed May 15, 2003.7)


Preoperatively, patient education is extremely im-portant. Often, these patients are frustrated becausethey have had a failed procedure that did not meettheir expectations. It is critical that this be approachedas a salvage type of procedure and that the surgeonand patient’s expectations be tempered. Our clinicalresults certainly reflect this with a recent review of ourrevision ACL experience. (Fox J, et al. Revision ACLreconstruction with nonirradiated patellar tendon allo-graft using immediate weightbearing and an acceler-ated rehabilitation. Presented at Annual Meeting ofArthroscopy Association of North America, Phoenix,AZ, April 2003, submitted for publication). This ap-plies to the patient’s recovery, rehabilitation, potentialweight-bearing status, and time to return to sporting

activities, as well as, potentially, the types of sportingactivities allowed.

GRAFT SELECTION ISSUES

There are a variety of grafts used for revision ACLsurgery. This is dependent on the individual and thesurgeon’s preference. For example, in our practice, ifthe patient has had a primary hamstring failure, wewill discuss with the patient consideration of a patellartendon autograft for revision surgery. If the patient hashad a previous patellar tendon autograft used, ourdiscussion is consideration of a nonirradiated patellartendon allograft or contralateral patellar tendon au-tograft. We have had a difficult time presenting con-tralateral graft usage to our patient population. Manyof these individuals do not want to have their normalknee surgically violated. In general, Achilles tendonallografts, quadriceps tendon autografts or allografts,tibialis anterior allografts, hamstrings, or prostheticgrafts are not used in our practice. Although there aremany advocates of hamstrings for revision surgery,intuitively it makes little sense because many of thesepatients have expanded tunnels and one of the keys tohamstring reconstruction is having appropriate “fit andfill” of the graft and its host tunnel. As previouslydescribed, the majority of our patients who have un-

FIGURE 16. This left knee anteroposterior radiograph shows theinitial hardware from a 2-incision patellar tendon autograft tech-nique as well as the revision hardware. Note that the femoralscrews have been bypassed. On the tibial side, the more proximalscrew was removed for initial tunnel drilling and then reinserted toprovide a buttress.

FIGURE 17. A variety of interference screws are commerciallyavailable. The screw on the upper left, manufactured by InstrumentMakar, has a threaded insert rather than a hexagonal screwdriverrecess. Different manufacturers use different sized hexagonal re-cesses. (Reprinted with permission.26)

22 B. R. BACH, JR.

dergone revision surgery had nonirradiated patellartendon allografts.

ALLOGRAFT CONSIDERATIONS

We have used nonirradiated patellar tendon graftssince 1986. There are issues that are specific to the useof allograft. Specific informed consent is necessary.Patients need to know that the tissues are tested forbacteria, hepatitis, HIV, and Jacob-Creutzfeldt dis-ease. Nevertheless, despite the use of preliminarychain reaction enzyme (PCR) testing, there is not a100% chance of preventing potential disease transmis-sion. Some surgeons are advocates of freeze-driedallograft tissues maintaining that no disease has beenreportedly contracted from a freeze-dried allograft.Cryopreserved allografts or ethylene oxide second-arily sterilized grafts have not been used. We haveused used one tissue bank since 1986 that is a not-for-profit organization. It is important to recognize thatthere can be up to 4 rejections of tissues from differenttissue banks; hence, our preference is to use a not-for-profit tissue bank where there are fewer pressures touse potentially marginal tissue. Nonirradiated ratherthan irradiated grafts are preferred.27,28 Historically, itwas felt that lower level radiation (e.g., 2.5 m/rad)would protect and secondarily sterilize the tissue fromHIV and hepatitis viral particles. Studies have foundthat it takes up to 4 m/rad to effectively kill these viralparticles.12,28 However, lower dosage of radiation ishelpful for surface bacterial contamination. Our con-tention is that irradiated tissues mechanically do notcompare to nonirradiated tissues.27 Another importantconsideration is that of availability and cost. Allo-grafts may in general range between $800 and $2,000.Most insurance companies cover the charges for anallograft, particularly in revision situations. The in-creasing demand for the use of allografts has resultedin some tissue banks providing hemipatellar tendonallografts. The hemipatellar tendon graft is a sagittallysplit patellar tendon and tibial tubercle graft. This hasits own potential problems as effectively the hemipa-tellar tendon graft is longer than the standard counter-part whole patellar tendon.

Another important consideration with allografts isrecognition of length parameters. For example, whenone is performing an allograft on a petite 5�2� womanand receives an allograft from a 6�4� donor, there willbe significant construct mismatch issues that may alsoaffect fixation and subsequent success of the revisionprocedures. Therefore, when we request allografts, wegenerally include the patient’s height and an estima-

tion of the potential soft tissue length that would bedesirable. For example, the majority of patellar ten-dons are approximately 45 � 3mm. For a 6-foot tallmale patient, we generally feel that a 45 to 48 mm softtissue graft construct would be an appropriate length.For someone who is 5�6,� we would like to have a 40to 42 mm construct, and for a patient who is 5�3� orunder, we would prefer to have a soft tissue constructless than 40 mm.

Nonirradiated patellar tendon allografts have beenour preferred graft for several reasons. Many of thesepatients have compromised knees and are willing tomodify some of their activities and yet, nevertheless,had instability that necessitates revision procedures.Some of these individuals already have early degen-erative joint disease and are looking for a less invasivetype of procedure. Nonirradiated patellar tendon allo-graft performed endoscopically is a minimally inva-sive surgical technique. Bone-tendon-bone allograftallows graft customization with regard to graft width,potential primary bone graft, and customization ofbone plugs. Additionally, if there is a potential graftconstruct mismatch, this allows creation of a longerbone plug for a “free bone block” modification fixa-tion (Fig 18).12 As the bone-tendon-bone technique isthe most commonly used primary ACL surgical tech-nique, this revision patellar tendon allograft provides afamiliar technique to the surgeon and interferencescrew fixation provides for solid graft fixation.

Another preoperative consideration involves previ-ously created skin incisions. The surgeon should as-sess skin mobility and adherent tissue may requireconsideration of a skin expander. One may use formerincisions to preclude skin bridges that may result inwound dehiscence. It may be necessary to customizethe incision to accommodate previous incisions.

TECHNICAL CONSIDERATIONS

The failed 2-incision surgery can be approachedendoscopically. The majority of failed endoscopictechniques can be approached endoscopically, or afailed endoscopic technique can be approached with a2-incision femoral technique. The majority of ourpatients have been revised endoscopically. The basicfundamentals of revision ACL surgery can be directlyderived from primary ACL reconstruction tech-niques.29-31

At the time of surgery, before the induction ofanesthesia, one must verify that an allograft is avail-able. A careful examination is performed under anes-thesia to exclude the possibility of a missed patholax-


ity that has been associated with ACL failures.22,32 Atourniquet is placed proximal on the affected extrem-ity, which is then placed in a well-padded GYN legholder with the hip and knee flexed to protect thecommon peroneal and femoral nerve. The waist andfoot of the table are flexed. One should be able to flexthe knee 110°. Preoperative antibiotics are adminis-tered, cephalosporin 1 g, unless there is a penicillinallergy, then 600 mg of Cleocin is used.

Standard arthroscopic portals are created. A system-atic diagnostic arthroscopy is performed to verify ar-ticular surfaces, chondral pathology, and meniscal sta-tus. Appropriate meniscal work is performed. Theknee should be placed in a “figure-4” position to makecertain that the patient does not have an abnormal

“gap sign” which would be suggestive of a postero-lateral laxity (F. R. Noyes, personal communications)

The residual ACL tibial footprint tissue is debridedso that one can determine the appropriate strategy forentrance placement. The creation of a long tibial tun-nel should be attempted and an accessory inferomedialor a transpatellar portal may be necessary to optimizetunnel orientation and positioning. One should avoidbeing trapped by the standard inferomedial portal asthe soft tissues in this situation may result in creationof a more sagittally oriented tibial tunnel. In general,if an entrance site is created midway between the tibialtubercle and the posteromedial corner of the tibia anappropriate medial to lateral orientation is created.One can actually overlay a pin or instrument over theflexed knee to ascertain this orientation. Once thetibial aiming device is provisionally positioned, anincision is created to approach the medial metaphysealregion of the tibia. If the screw or hardware is visibleor easily identified, it is in general removed unless thetunnel is so abnormally positioned that one can easilybypass the hardware. If there is an overlap, it may benecessary to preliminarily remove the screw and thenreinsert this at the time of graft fixation, thus, creatinga stacked screw fixation (Figs 4 and 19). If the screwis not readily identifiable, then it is advisable to useC-arm fluoroscopy or fluoroscan to identify its loca-tion so that inadvertent cortical destruction does notoccur while attempting to remove hardware. Once thehardware is removed, the tibial aiming device may berepositioned. A variable angle rather than fixed angletibial aimer is used and in revision situations, a 55°angle is selected on the aiming device for orientation.An inferomedial accessory portal is generally usedalthough sometimes a transpatellar portal is used fortibial aimer placement. Intra-articularly, the posterioredge of the anterior horn lateral meniscus in the coro-nal plane is used as a relative reference point wherewe would like our pin to exit within the former ACLinsertion region. A “point to elbow” aimer is used; thestylette of the aimer is moved back 3 to 4 mm toachieve this goal. As the pin is drilled, it frequentlywill enter a portion of the previously drilled tibialtunnel. Therefore, the pin often enters within the intra-articular tunnel entrance region. There may be pinmicromotion; it should be secured at the time offemoral reaming. The pin may be stabilized with aKocher clamp or it may be tapped up into the femoralroof. Once the tibial tunnel is reamed, the intra-artic-ular edges are smoothed with a chamfer reamer andhand rasp. The arthroscope is routinely placed retro-grade up the tibial tunnel to assess the continuity of

FIGURE 18. In the event of a graft construct mismatch, one maycompensate by using a free bone block modification. The distalbone plug is sharply removed and No. 5 Ticron suture is run upproximally and distally. This effectively creates a pseudoquadtendon graft. The graft is then secured on the femoral side. Thebone plug is then placed along the anterior aspect of the graft andthen, while this is secured, interference screw fixation is per-formed. (Do not reinforce this with additional fixation.) The con-cept has also been mechanically assessed in the laboratory. (Re-printed with permission from Bach BR Jr, Mazzocca A, Fox JA.Revision anterior cruciate ligament surgery. In: Grana WA, ed.Orthopaedic knowledge online. Rosemont, IL: American Academyof Orthopaedic Surgeons, 2003. Available at www.aaos.org/oko.Accessed May 15, 2003.7)

24 B. R. BACH, JR.

the tibial osseous tube. Additionally, this allows forbetter visualization of any adherent soft tissue whichmay need to be debrided.

At this point, attention is directed towards the fem-oral socket. The femoral interference screw must beidentified and soft tissue and overlapping bone care-fully removed. It is critical to remove bone from theperiphery of the screw so that it is not stripped whilebeing removed. A spinal needle is placed percutane-ously or through the inferomedial portal to identify theangle that the knee has to be flexed to optimize screwremoval. The hyperflex Nitenol wire is positioned intothe cannulated screw and a screwdriver is inserted forscrew removal. Situations have been encounteredwhere the screw is so vertically oriented that we have

actually had to approach removal through an acces-sory inferolateral portal. In a vertically oriented screwposition, one may be able to potentially bypass thescrew by further inserting the screw. This is anotheroption that may maintain a medial buttress for thefemoral socket. The soft tissue must be carefully re-moved, exposing the lateral wall and notchplasty mayneed to be performed. The over-the-top position mustbe carefully identified and confirmation of this posi-tion achieved with probe palpation. A 7-mm femoraloffset aimer is placed generally retrograde through thetibial tunnel and keyed off the over-the-top positionwith the knee flexed at approximately 80°. If the tibialtunnel placement results in a tendency for the aimingdevice to slide more superiorly toward the 11:30 or 12o’clock position, then it is advisable to actually placethe offset aimer through the accessory inferomedialportal, hyperflex the knee and drill the femoral tunnelin this manner. This is a technique modification thathas been advocated by O’Donnell.33 Once the offsetaimer is positioned, it is drilled with a 3/32-inchSteinmann pin. This may be more difficult to drill thanin index ACL surgeries as the pin frequently may bedrilled through the cortex of the previous graft boneblock. The pin is drilled to a depth of 1.5 inches andover-reamed with a cannulated 10-mm endoscopicacorn reamer. An endoscopic footprint is created toconfirm that there has been no “blow out” of theposterior cortical wall. Reaming is generally com-pleted to a depth of 35 mm. Once again, because of thepreviously placed bone plug, this reaming may have adifferent sensation than the tactile sensation of ream-ing an index femoral tunnel. Once the femoral socketis reamed, the femoral socket must be carefully in-spected. This can be achieved by rotating the arthro-scope to visualize up the femoral tunnel, placing thearthroscope in the inferomedial portal to directly viewthis socket, or sliding the scope retrograde via thetibial tunnel with the knee flexed up into the socket.As this is done the arthroscope is rotated to confirmthat there has been no intraosseous “blow out” andconfirm that a thin posterior cortical rim remains.

Before reaming the femoral tunnel, determinationof whether accessory bone grafting is needed is made.If there has been some overlap in the tunnel, the screwis removed and the defect is packed with allograftcancellous graft. The graft is placed through a cleararthroscopic cannula and tapped into position provid-ing a firm osseous construct. There have also beentimes that allograft cortical match sticks are used toprovide additional fixation (Fig 20). At this point, thefemoral socket is re-reamed. Depending on the clini-

FIGURE 19. This right knee anteroposterior view shows the use ofstacked femoral screws to provide additional fixation. (Reprintedwith permission from Bach BR Jr, Mazzocca A, Fox JA. Revisionanterior cruciate ligament surgery. In: Grana WA, ed. Orthopaedicknowledge online. Rosemont, IL: American Academy of Ortho-paedic Surgeons, 2003. Available at www.aaos.org/oko. AccessedMay 15, 2003.7)


cal situation and the status of the tunnels, the allograftmay be fashioned concurrently or may delayed untilthe tunnel status has been determined. A push-inrather than pull-through technique is used for graftplacement. A curved hemostat is inserted through theinferomedial portal and the femoral bone plug isgrasped at the junction of the proximal and middlethird and guided up into the femoral socket. The graftis oriented with the cortex posterior and in the coronalplane. The femoral plug is left slightly proud initiallyto act as a skid for placement of the 14-inch hyperflexNitenol pin, which is inserted through an accessoryinferomedial portal. It is provisionally positioned andthen the knee is further flexed to between 100° and110° and the pin is completely slid up until it reachesthe depths of the femoral socket. This will assure

parallel or near parallel placement of the anteriorfemoral interference screw. At this point, the knee isbrought back to the resting position of approximately80° and the bone plug is further positioned at thefemoral articular entrance. At this point, assessment ismade to determine whether there is a construct mis-match. If the construct match is appropriate, femoralfixation is achieved with a 7 � 25 mm metal cannu-lated interference screw. Bioabsorbable (i.e., radiolu-cent) interference screws have not been used in revi-sion ACL surgery. If the bone is soft, a 9-mm diameterscrew is used. Once the graft is secured fixation isassessed by applying manual tension distally on thegraft. The knee is then multiply cycled to assess grossisometry. In general, as the knee is brought from 90°to complete extension in the terminal 20° of extension,the graft translates into the tibial tunnel by approxi-mately 1 to 2 mm. The graft is multiply cycled andattention is directed towards tibial fixation. The graftis generally rotated 180° towards the lateral wall. Onthe femoral side, the graft is oriented with the cortexposterior and in the coronal plane. On the tibial side,the graft is oriented with the cortex anterior and in thecoronal plane. The tibial graft is secured in completeextension with a 9 � 20 or 9 � 25 mm interferencescrew. If there is slight recession of the graft, a longertibial interference screw is used rather than buryingthe screw intraosseously. If there has been some over-lap with the tibial tunnel entrance sites, a “stacked”screw is used for reinforcement as an additional but-tress fixation. If the bone is soft, augmentation fixationwith a screw and post or staple is performed. Follow-ing tibial fixation, the graft is visualized to assess itsintegrity, the knee multiply cycled, and its fixationreevaluated. A Lachman, anterior drawer, and pivot-shift test is performed on the operating table. A He-movac drain is not used. Bupivicaine is injected intothe surgical wound region and intra-articularly. Amotorized cryotherapy device (Iceman; DJ Orthopae-dics, Carlsbad, CA) is used and a hinged drop-lockbrace (DonJoy T-ROM, DJ Orthopaedics) is used.Surgeries are performed on an outpatient basis. Thepatient is seen the following day for a dressingchange, assessment of motion and evaluation ofwhether a hemarthrosis requires aspiration.

RESULTS

A nonrandomized retrospective review at a 2-yearminimum follow-up evaluation of 32 of 38 patients(84%) who underwent a revision reconstruction withnonirradiated patellar tendon allograft used for a failed

FIGURE 20. This right knee anteroposterior view shows the use ofmatch-stick corticocancellous graft to provide additional medialbuttress at the time of revision surgery. (Reprinted with permissionfrom Bach BR Jr, Mazzocca A, Fox JA. Revision anterior cruciateligament surgery. In: Grana WA, ed. Orthopaedic knowledge on-line. Rosemont, IL: American Academy of Orthopaedic Surgeons,2003. Available at www.aaos.org/oko. Accessed May 15, 2003.7)

26 B. R. BACH, JR.

patellar tendon autograft was recently conducted. Therevision procedures were performed between 1993and 1999, the methodology used was similar to pre-viously published primary ACL reconstruction fol-low-up surveys.1-3 Clinical, radiographic, arthromet-ric, and functional evaluations were performed. TheTegner and Lysholm surveys along with Noyes, Koos,IKDC and SF-12 rating scales were used.34-38

Twenty-four of the patients underwent an endo-scopic revision procedure and 8 patients underwent a2-incision revision procedure. As a generalization,hardware was bypassed unless it was necessary toremove to perform adequate tunnels. In this survey, anidentical rehabilitation (accelerated program) identicalto our index ACL reconstructions was used.2 A cus-tom ACL orthosis was used between 6 and 12 monthspostoperatively for return to sports activities.

Of interest is that only 2 of these 32 patients hadbeen initially reconstructed by the senior surgeon(B.R.B.) from our center. At revision reconstruction,the mean interval to revision surgery was 50 months(range, 9 to 101 months). At follow-up, the meaninterval from revision surgery was 4.8 years (range,2.1 to 12.1 years). At the time of revision the averageage was 28 years (range, 16 to 57 years). There were20 right knees and 12 left knees revised in 14 malesand 18 females. The average patient had 2.8 surgicalprocedures performed before the revision reconstruc-tion.

Varying degrees of chondromalacia was observedin 70% of our study group and in at least 1 of the 3articular compartments. Fifty percent of these patients(n � 16) had previous meniscal surgery involvingeither the lateral and/or medial meniscus. Radiograph-ically, the medial compartment was normal in only 17subjects (53%), the lateral compartment (67%) and thepatellofemoral compartment (68%).

At follow-up, 87% of the patients had a negative orgrade 1 Lachman with a firm endpoint and 89% of thepatients had a negative or grade 1 pivot-shift. How-ever, if the presence of a pivot-shift was used to definea failure, 29% of our patients overall had a demon-strable pivot-shift. Of note, only one individual had agrade 2 pivot-shift phenomenon. Motion was reliablyrecovered with a mean of 135° of flexion. The meanheel height difference in extension was 0.5 cm (range,0 to 4 cm). One individual had thigh girth atrophy ofgreater than 1 cm.

Functional indices showed small mean deficits onsingle-leg hop, timed single-leg hop over 6 meters,and in vertical jump. However, a wide range in stan-dard deviations was noted. In fact, 32% of the patients

tested better on the affected knee than the unaffectedknee for each of these 3 tests although they were notthe same individuals. Moreover, for these 3 functionaltests, 88%, 85%, and 66% of the subjects had less than10% side-to-side differences for these 3 tests, respec-tively.

Preoperative arthrometric evaluations were similarto failed index ACL reconstructions.2,39 The meanmanual maximum translations on the affected kneeaveraged 14 mm (range, 8 to 21 mm), whereas theunaffected knee averaged 6.5 mm (range, 2.5 to 13mm). Preoperatively, 78% of the patients had side-to-side differences of between 3 and 5 mm and 9% haddifferences of �5 mm. At follow-up, the mean man-ual-maximum translation was 9.4 mm (range, 6 to 18mm) for the affected knee. The mean maximum-man-ual side-to-side difference was 1.9 mm (range, �1 to8 mm). Twenty seven patients (84%) had side-to-sidedifferences of �3 mm. Three patients (9%) had dif-ferences ranging between 3 and 5 mm, and 2 individ-uals (6%) had differences of �5 mm. Preoperatively,the differences of affected and unaffected knees werehighly statistically significant, postoperatively side-to-side differences were not significant, and improve-ment from preoperative maximum-manual translationto postoperative manual-maximum translation washighly significant.

Table 1 summarizes the various postoperative ratingscores that we used. Twenty-eight of these 32 patients(87%) indicated that they were either mostly or com-pletely satisfied with their surgical revision. However,only 50% of the patients were completely satisfied.Three patients (10%) were somewhat satisfied and 1patient (3%) was dissatisfied. No postoperative infec-tions occurred in this group. No additional surgeries

TABLE 1. Knee Scoring Scales

Scale Mean Range SD

IKDC 71 23-97 22KOOS Pain 84 36-100 18KOOS Symptom 77 25-100 21KOOS ADL 91 50-100 14Tegner (before any surgery) 8.4 2-10 2.1Tegner (before revision) 5.0 0-10 3.3Tegner (most recent follow-up) 6.3 0-10 2.6Lysholm 75 30-100 22Noyes Sports Function 72 0-100 26Noyes Functional ADLs 30 17-40 30Modified Cincinnati 7.2 2-10 2.2SF-12 Mental 55 27-66 8SF-12 Physical 48 20-59 11Visual Analog Scale 2.9 0-9 2.5


were performed for arthrofibrosis and no hardwarerequired secondary removal.

The results of our survey indicated that, in ourhands, revision ACL reconstruction was not compa-rable to an index ACL reconstructive procedure whenusing a nonirradiated patellar tendon allograft. Thepatients’ subjective satisfaction level was slightly lessthan primary ACL reconstructions performed witheither an autograft or allograft. As previous studieshave shown, between 92% and 95% mostly are com-pletely satisfied subjective ratings in index ACL re-constructions.1-3 Our results are quite similar to thosereported by Noyes et al.,40 who also observed a highincidence of abnormal articular cartilage at the time ofrevision as well as a high degree of previous meniscalsurgeries. In their series, a combination of grafts wereused including irradiated and nonirradiated grafts, lig-ament augmentation devices and extra-articular ili-otibial band procedures.40 Noyes reported an 88%postoperative pivot-shift grades of negative or 1� thatwere identical to our observations. In another study,Noyes and Barber-Westin41 reviewed 54 patients whounderwent a patellar tendon autograft revision. In thissurvey the failure rate was reported as 24%, but 6 ofthe 13 failures had a reharvested patellar tendon au-tograft. In the Pittsburgh series using irradiated freshfrozen allograft, 36% of their cohort had a KT-1000maximum manual side-to-side difference of �5.5mm.42 We do not believe that it is the irradiatedallograft per se that is the contributing factor to thehigher failure rates observed in this patient population,as we are currently retrospectively reviewing the useof nonirradiated patellar tendon allografts for indexprocedures. We have carefully evaluated 50 primaryallograft ACL reconstruction patients at a minimum of2 years postreconstruction and a negative pivot-shifthas been noted in 90% of the patients (n � 48)(Aadalen K, et al. Primary arthroscopy-assisted ACLreconstruction using nonirradiated patellar tendon al-lografts. Presented at the Linvatec Residents-FellowsSports Medicine Course, Usseppa Island, FL, May 15,2003).

Our survey had several potential weaknesses. It wasa nonrandomized retrospective survey and severalquestions arise. These include whether a differentgraft source may have resulted in improved results.Additionally, as we used an aggressive rehabilitationprogram, perhaps a more conservative rehabilitationprogram may have resulted in fewer clinical failures.I do not believe that there were unrecognized ligamen-tous patholaxities in our cohort group, but this hasbeen associated with failed ACL surgical proce-

dures.22,31 No meniscal transplants were performed inthis cohort, and perhaps, if judiciously used, several ofour patients might have benefited from additional sec-ondary restraints.

Nevertheless, our study has strengths. One graftsource was used for a specific type of graft failure, onespecific not-for-profit tissue bank was use, and thepatient population was clearly defined. Additionally,multiple available outcome scores were employed.Arthrometric evaluations were performed by one ex-perienced individual. Our previous index ACL recon-struction surveys served as a comparative bench-mark.1-3

In summary, revision ACL reconstructive surgery isvery challenging. It requires considerable thought pre-operatively to define potential technical errors thatmay have contributed to the failure, creating a preop-erative game plan, and then technically executing achallenging surgical procedure. We feel that, regard-less of graft utilized, that there will be inherently ahigher failure rate with revision ACL procedures formultiple reasons. We continue to use nonirradiatedpatellar tendon allograft for revision ACL surgery. Itis critical to counsel patients preoperatively to haverealistic expectations and to recognize that revisionACL surgery is effectively a salvage procedure.

REFERENCES

1. Bach BR Jr, Jones GT, Sweet FA, Hager CA. Arthroscopy-assisted anterior cruciate ligament reconstruction using patel-lar tendon substitution. Two- to four-year follow-up results.Am J Sports Med 1994;22:758-767.

2. Bach BR Jr, Levy ME, Bojchuk J, Tradonsky S, Bush-JosephCA, Khan NH. Single-incision endoscopic anterior cruciateligament reconstruction using patellar tendon autograft. Min-imum two-year follow-up evaluation. Am J Sports Med 1998;26:30-40.

3. Bach BR Jr, Tradonsky S, Bojchuk J, Levy ME, Bush-JosephCA, Khan NH. Arthroscopically assisted anterior cruciate lig-ament reconstruction using patellar tendon autograft. Five- tonine-year follow-up evaluation. Am J Sports Med 1998;26:20-29.

4. Williams RJ III, Warren RF, Carson EW, Wickiewicz TL.Revision anterior cruciate ligament reconstruction: The Hos-pital for Special Surgery experience. Tech Orthop 1998;13:375-383.

5. Steadman JR, Saterbak AM. Revision anterior cruciate liga-ment reconstruction: Techniques and tips—The Vail experi-ence. Tech Orthop 1998;13:384-390.

6. Clancy WG Jr, Pietropaoli MP. Revision anterior cruciateligament reconstruction using the “anatomic-endoscopic”method. Tech Orthop 1998;13:391-410.

7. Bach BR Jr, Mazzocca A, Fox JA. Revision anterior cruciateligament surgery. In: Grana WA, ed. Orthopaedic knowledgeonline. Rosemont, IL: American Academy of OrthopaedicSurgeons, 2003. Available at www.aaos.org/oko. AccessedMay 15, 2003.

28 B. R. BACH, JR.

8. Brown CH Jr, Carson EW. Revision anterior cruciate ligamentsurgery. Clin Sports Med 1999;18:109-171.

9. Barber FA, McGuire DA, Johnson DH. Should allografts beused for routine anterior cruciate ligament reconstructions?Arthroscopy 2003;19:421-425.

10. Noyes FR, Barber-Westin SD. Revision anterior cruciate lig-ament reconstruction: Report of 11-year experience and resultsin 114 consecutive patients. Instr Course Lect 2001;50:451-461.

11. Uribe JW, Hechtman KS, Zvijac JE, Tjin ATEW. Revisionanterior cruciate ligament surgery: Experience from Miami.Clin Orthop 1996;325:91-99.

12. Novak PJ, Williams JS, Wexler G, Bach BR Jr, Bush-JosephCA. Comparison of screw post fixation and free bone blockinterference fixation for anterior cruciate ligament soft tissuegrafts: Biomechanical considerations. Arthroscopy 1996;12:470-473.

13. Dworsky B, Jewell BF, Bach BR Jr. Interference screw diver-gence in endoscopic anterior cruciate ligament reconstruction.Arthroscopy 1996;12:45-49.

14. Bush-Joseph CA, Bach BR Jr. Migration of femoral interfer-ences screw after anterior cruciate ligament reconstruction.Am J Knee Surg 1998;11:32-34.

15. Hamner DL, Brown CH Jr, Steiner ME, et al. Hamstringtendon grafts for reconstruction of the anterior cruciate liga-ment: Biomechanical evaluation of the use of multiple strandsand tensioning techniques. J Bone Joint Surg Am 1999;81:549-557.

16. Kousa P, Jarvinen TLN, Vihavainen M, Kannus P, Jarvinen M.The fixation strength of six hamstring tendon graft fixationdevices in anterior cruciate ligament reconstruction. Part I:Femoral site. Am J Sports Med 2003;31:174-181.

17. Kousa P, Jarvinen TLN, Vihavainen M, Kannus P, Jarvinen M.The fixation strength of six hamstring tendon graft fixationdevices in anterior cruciate ligament reconstruction. Part I:Tibial site. Am J Sports Med 2003;31:182-188.

18. Brand JC Jr, Pienkowski D, Steenlage E, et al. Interferencescrew fixation strength of a quadrupled hamstring tendon graftis directly related to bone mineral density and insertion torque.Am J Sports Med 2000;28:705-710.

19. Caborn DNM, Coen M, Neef R, et al. Quadrupled semitendi-nosus-gracillis autograft fixation in the femoral tunnel: A com-parison between a metal and a bioabsorbable interferencescrew. Arthroscopy 1998;14:241-245.

20. Giurea M, Zorilla P, Amis AA, et al. Comparative pull-out andcyclic-loading strength tests of anchorage of hamstring tendongrafts in anterior cruciate ligament reconstruction. Am J SportsMed 1999;27:621-625.

21. Magen HE, Howell SM, Hull ML. Structural properties of sixtibial fixation methods for anterior cruciate ligament soft tissuegrafts. Am J Sports Med 1999;27:35-43.

22. O’Brien SJ, Warren RF, Pavlov H, Panariello R, WickiewiczTL. Reconstruction of the chronically insufficient ACL withthe central third of the patellar ligament. J Bone Joint Surg Am1991;73:278-286.

23. Daniel DM, Malcolm LL, Losse G, Stone ML, Sachs R, BurksR. Instrumented measurement of anterior laxity of the knee.J Bone Joint Surg Am 1985;67:720-726.

24. Bach BR Jr, Warren RF, Wickiewicz TL. The pivot-shiftphenomenon: Results and description of a modified clinical

test for anterior cruciate ligament insufficiency. Am J SportsMed 1988;16:571-576.

25. Bush-Joseph C, Bach BR Jr, Bryan J. Posterior cortical viola-tion during endoscopic ACL reconstruction. Am J Knee Surg1995;8:130-133.

26. Bach BR Jr. Special report—Observations on interferencescrew morphologies. Arthroscopy 2000;16:E10. Available atwww.arthroscopyjournal.org.

27. Adams DJ, Curran AR, Koneczny S, Utz M, Steiner ME,Scheller AD. Cyclic creep and failure of paired human patellartendon allografts sterilized by allowash versus allowash pluslow dose gamma irradiation. Proc Orthop Res Soc 2001;739.

28. Fideler BM, Vangsness CT Jr, Lu B, Orlando C, Moore T.Gamma irradiation: Effects on biomechanical properties ofhuman bone-patellar tendon-bone allografts. Am J Sports Med1995;23: 643-6.

29. Hardin GT, Bach BR Jr, Bush-Joseph CA. Endoscopic single-incision anterior cruciate ligament reconstruction using patel-lar tendon autograft. Case Report. Am J Knee Surg 1992;5:144-155.

30. Bach BR Jr. Arthroscopy-assisted patellar tendon substitutionfor anterior cruciate ligament insufficiency. Surgical tech-niques. Am J Knee Surg 1989;2:3-20.

31. Ferrari JD, Bush-Joseph CA, Bach BR Jr. Endoscopic anteriorcruciate ligament reconstruction with patellar tendon au-tograft: Surgical technique. Tech Orthop 1998;13:262-274.

32. Ferrari JD, Bach BR Jr. Posterolateral Instability of the Knee.Sports Med Arthrosc Rev 1999;7:273-288.

33. O’Donnell JB, Scerpella TA. Endoscopic anterior cruciateligament reconstruction: modified technique and radiographicreview. Arthroscopy 1995;11:577-584.

34. Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD.Knee Injury and Osteoarthritis Outcome Score (KOOS)—Development of a self-administered outcome measure. J Or-thop Sports Phys Ther 1998;28:88-96.

35. Anderson AF. Rating scales. In: Fu FH, Harner CD, VinceKG, eds. Knee surgery. Baltimore: Williams & Wilkins, 1994;275-296.

36. Noyes FR. The Noyes Knee Rating System. An assessment ofsubjective, objective, ligamentous, and functional parameters.Cincinnati: Sports Medicine and Research and EducationFoundation, 1990.

37. Tegner Y, Lysholm J. Rating systems in the evaluation of kneeligament injuries. Clin Orthop 1985;198:43-49.

38. Lysholm J, Gillquist J. Evaluation of knee ligament surgeryresults with special emphasis on use of a scoring scale. Am JSports Med 1982;10:150-154.

39. Bach BR Jr, Warren RF, Flynn WM, Kroll M, Wickiewicz TL.Arthrometric evaluation of knees that have a torn anteriorcruciate ligament. J Bone Joint Surg Am 1990;72:1299-1306.

40. Noyes FR, Barber-Westin SD, Roberts CS. Use of allograftsafter failed treatment of rupture of the anterior cruciate liga-ment. J Bone Joint Surg Am 1994;76:1019-1031.

41. Noyes FR, Barber-Westin SD. Revision anterior cruciate sur-gery with use of bone-patellar tendon-bone autogenous grafts.J Bone Joint Surg Am 2001;83:1131-1143.

42. Johnson DL, Swenson TM, Irrgang JJ, Fu FH, Harner CD.Revision anterior cruciate ligament surgery: Experience fromPittsburgh. Clin Orthop 1996;325:100-109.
