Reprinted from CLINICAL ORTHOPAEDICS, October, 1992Vol. 283© J. 3. Lippincott Co. Printed in U.S.A.

Tibial Tunnel Placement in Anterior CruciateLigament Reconstructions and Graft Impingement

MAJ. STEPHEN M. HOWELL, M.D., U.S.A.F.R., M.C., AND JAMES A. CLARK, M.D.

Fifty-six anterior cruciate ligament (ACL) reconstructions had a magnetic resonance scan of the ACL graft six months after operation. The impingement-free grafts (n = 26) had a low mag-netic resonance signal from origin to insertion. Impinged grafts (n = 30) had an increased mag-netic resonance signal confined to the distal two thirds of the graft. The location of the tibial tunnel (TT) was determined from a lateral roentgeno-gram. Positioning the center of the TT 12-23 mm from the anterior edge of the tibia consistently pro-duced graft impingement and flexion contractures. Roof impingement was avoided and hyperexten-sion was regained when the TT was centered more posteriorly within a 6-mm impingement-free zone (22-28 mm from the anterior edge of the tibia). Sta-bility and knee extension were significantly better when the center of the TT was 2-3 mm posterior to the center of the normal ACL insertion.

Roof impingement is detrimental to the clini-cal course of an anterior cruciate ligament (ACL) reconstruction.4,9 Roof impingement occurs when knee extension is limited by premature impact-ing of the graft against the intercondylar roof.4,11

This differs from a normal knee, in which termi-nal extension is limited by tension in the poste-rior capsule. The amount of roof impingement is determined by the sagittal location of the tibial tunnel. Anterior tibial tunnels have the great-est potential for developing roof impingement.1,11,17

Clinically, roof impingement interferes with rehabilitation. Patients may experience persistent knee effusions, resistant flexion contractures, and pain.12 Second-look arthroscopy and a delayed roofplasty can elimi-nate these symptoms and facilitate rehabilitation of an ACL reconstruction subjected to roof impingement.4

Roof impingement can be analyzed by the magnetic resonance (MR) appearance of the graft. Impinged grafts have a regionalized signal increase in the distal two thirds of the ACL graft.9 These MR signal increases peak at three months and do not improve during the first three years of implantation.9.10

Many studies have confirmed that the sagittal location of the TT can not be controlled by lim-iting tibiofemoral separation distances to 2 mm or less.1,7,8,15,17 A matching femoral region can be found to control graft length changes within 2 mm for any tibial attachment site that is located within the confines of the ACL insertion.7 These biomechanical observations indicate that roof impingement may coexist in reconstructions with acceptable graft excursion profiles because the sagittal location of the TT can vary within the anterior to posterior limits of the normal ACL insertion.

From the Department of Orthopedics (SGHT) and Depart-ment of Radiology (SGHR), David Grant Medical Center, Travis AFB, California.

Presented at the AAOS annual meeting, New Orleans, Feb. 8-13, 1990.

The views expressed herein are those of the authors and do not reflect the official policy or position of the United States Department of Defense or the United States Government.

Reprint requests to Stephen M. Howell, M.D., 7401 Timber-lake, Ste. 103, Sacramento, CA 95823.

Received: October 3, 1990.

The goal of an ACL reconstruction is to select a location for the TT that avoids graft weakening and rehabilitation difficulties from roof impinge-ment, and yet simultaneously retains acceptable graft excursion profiles (2 mm or less) and knee stability. The authors hypothesize that there may be two zones within the ACL insertion: An ante-rior zone that results in roof impingement, and a relatively posterior zone that is free of roof impingement. The purpose of this study was to determine the size and location of the zone for centering the TT that allows the graft to be placed impingement-free and with an ac-ceptable graft-excursion profile.

MATERIALS AND METHODSThe study consisted of 56 patients who had an

arthroscopically assisted ACL reconstruction. Each reconstruction had 2 mm or less excursion, which was determined by a suture attached to a tensiometer (Acufex Microsurgical, Norwood, Massachusetts) that was passed through provisional pilot holes (2.4 mm in diameter) in the femur and tibia. Changes in suture excursion were measured at 30˚ intervals from maxi-mum passive extension to 90˚-105˚ flexion. All knees had a limited sagittal MR scan of the graft at six months after implantation and a lateral roentgeno-gram of the knee in full extension at one year. The patients were divided into two groups, depending on the appearance of the graft on the MR scan.

Magnetic resonance scans were performed using a 0.35-T superconducting magnet with a dedicated quadrature detection knee coil (Diasonics, San Fran-cisco, California) during the early portion of the study. Part of the way through the study, the MR system was upgraded to a 1.5-T superconducting magnet (Phil-ips, Amsterdam, Holland). The 1.5-T images were superior in clarity and signal-to-noise ratio, but were otherwise comparable. Imaging was confined to ten 3-mm-thick sagittal sections (0.625 mm2 pixels) cen-tered about the intercondylar region of the knee. The knee was externally rotated 10- 15˚ to optimally align the graft in the sagittal plane. Image acquisition was performed with the standard spin-echo technique using a 1500-millisecond repetition time (TR) and 50-millisecond echo time (TE) (0.35 T) or 1200-mil-lisecond TR and 40-millisecond TE (1.5 T). Encoding and reconstruction was performed with the standard two-dimensional Fourier transformation technique, using 256 phase encoding steps and two excitations (15 minute acquisition time).3,5,9,10,16

The radiologist analyzed each MR scan without knowing the clinical outcome. All patients in whom the entire graft was black, homogeneous, and low in signal (similar to the MR appearance of the poste-rior cruciate ligament) were placed into the impinge-ment-free group (Fig. 1).9 There were 26 patients without evidence of impingement on MR scanning: 23 men and three women (average age, 25 ± 5.5 years). Twenty knees were reconstructed with a dou-ble-looped semitendinosus and gracilis graft, and six had a central-third patellar tendon bone graft. Four knees had an acute and 22 had a chronic reconstruc-tion. The MR scans of the remaining patients had a regionalized increase in the MR signal intensity of the graft. This signal increase was confined to the distal two thirds of the graft (Fig. 2).9,10 These 30 knees in 30 patients (25 men, five women, average age of 25 ± 4.7 years) constitute the impingement group. Twenty-one knees were reconstructed with a double-looped semi-tendinosus and gracilis graft, and nine had a central-third patellar tendon bone graft. Eleven knees were reconstructed acutely. All patellar tendon grafts were oriented with the width (9-10 mm dimension) of the graft in the sagittal plane.

A lateral roentgenogram with the knee in maxi-mal extension was obtained at one year. The roent-genogram was analyzed to evaluate the position of the TT with respect to the tibial plateau and the intercon-dylar roof (Figs. 1 B and 2B). Five lines were con-structed, and four measurements were obtained (Fig. 3). The first line was drawn parallel to the intercon-dylar shelf (Blumensaat’s line), extending across the tibial articular surface. The second line was drawn to overlay the sclerotic, subchondral margin of the concave medial tibial plateau. The third and fourth lines, outlining the anterior and posterior limits of the TT, were drawn overlying the sclerotic tunnel edges. Lastly, a line down the center of the TT was constructed. Four measurements were made: (1) The sagittal depth of the medial tibial plateau (SAG TIB DEP) was measured between the anterior bone-carti-lage junction and the posterior cortex of the medial plateau, identified by the tubercle where the poste-rior cruciate inserts. (2) The distance from the ante-rior edge of the tibia to the center of the tibial tunnel (CTT) was determined. (3) The distance between the anterior wall of the TT and the tibial intersection of the slope of the intercondylar roof was measured (X). (4) The width of the TT orifice was measured (W). The following ratios were constructed to normalize the results from different-sized knees: CTT/(SAG TIB DEP), X/W X 100 (Percent of the width of the TT

188 Howell and ClarkClinical Orthopaedics

and Related Research

orifice either anterior (+) or posterior (-) to the slope of the intercondylar roof = % TT Overlap) (Figs. 1B and 2B).9

The tibial zone that was impingement free was calculated by multiplying the range of CTT/(SAG TIB DEP) by the average sagittal depth of the tibia (60 mm). Similarly, the tibial zone that resulted in roof impingement was determined for the grafts with MR signal changes that indicated impingement.

At one year, the difference in passive knee ex-tension between the reconstructed and normal knee, instrumented laxity testing during a manual maximum drawer test using the KT- 1000 (MED-metric, San Diego, California), and the presence or absence of a pivot shift were determined.

Comparisons between the impingement-free and impingement groups were made by using the unpaired Student’s t-test and the chi-square test. Linear regres-sion was used to compare the exten-sion deficit to the

center of the tibial tunnel. Significance was found to be less than 0.05.

RESULTS

The appearance of the graft on MR scan and the relationship of the tibial tunnel to the inter-condylar roof on lateral roentgenogram are shown for a typical impingement-free knee (Fig. 1). The MR signal of the graft in the impingement-free knees was black, homogeneous and low in signal throughout the entire intraarticular course. Impingement-free grafts were similar in appear-ance to the MR signal of the posterior cruciate ligament and patellar tendon (Fig. 1A). A space remained between the graft and intercondylar roof with the knee in full extension. The graft had a straight pathway within the intercondylar notch and tibial tunnel. The lateral roentgeno-

FIGS. 1A AND 1B. Impingement was defined by the MR appearance of the graft. Impingement-free grafts were identified by the MR signal being black, homoge-neous, and low in signal from origin to inser-tion (double- looped semitendinosus/gracilis graft) (A) The width of the TT was entirely posterior to the slope of the intercondylar roof when viewed on a lat-eral roentgenogram exposed with the knee in maximum extension (B) The grafts were located anatomically, entirely within the pathway of the normal ACL.

Tibial Tunnel Placement 189Number 283October, 1992

gram demonstrated that the articular opening of the TT was entirely posterior to the tibial inter-section of the slope of the intercondylar roof in the impingement-free knees (average %TT Over-lap for the group = –11 k 14%) (Fig. 1B).

The appearance of the graft on MR scan and the relationship of the TT to the intercondylar roof on lateral roentgenogram are shown for a knee subjected to impingement (Fig. 2). The MR signal of the impinged graft demonstrates a regional-ized increase in MR signal intensity along the distal two thirds of the intraarticular pathway of the graft (Fig. 2A). There was no space between

the graft and intercondylar roof with the knee in full extension. These grafts were angulated poste-riorly by the intercondylar roof impinging on the anterior fibers of the graft. The lateral roentgeno-gram demonstrated that a portion of the articular opening of the TT was anterior to the tibial inter-section of the slope of the intercondylar roof (% TT Overlap = 60 ± 32%) (Fig. 2B). More than one half of the width of the TT was anterior to the slope of the intercondylar roof in 63% of the im-pinged knees. The percentage of the width of the TT that was anterior or posterior to the slope of the intercondylar roof was significantly different

FIGS. 2A AND 2B. Grafts placed anteriorly suffered from roof impingement (A). The anterior fibers of the graft were located nonanatomically because they coursed anterior to the anterior-most fibers of the normal ACL. The intraarticular pathway of the grafts were deformed posteriorly by the impaction of the intercondylar roof on the anterior fibers of the graft (open arrow). These grafts had a regionalized signal increase in their MR appearance that was confined to the distal two thirds of the graft (double-looped semitendinosus/gracilis graft). A lateral roentgenogram demonstrated that 80% of the TT was located anterior to the tibial intersection of the slope of the intercondylar roof (B).

190 Howell and ClarkClinical Orthopaedics

and Related Research

between the two groups (p < 0.0001).The location of the center of the TT mea-

sured from the lateral roentgenograms was sig-nificantly different for the impingement-free and impinged knees (P < 0.0001). The center of the impingement-free TTs ranged from 37% to 47% from the anterior edge of the tibia (mean, 42 ± 3%). The impingement-free zone for centering the TT was calculated to be 22-28 mm from the anterior edge of the tibia (normalized to a 60-mm-

deep tibial plateau) (Fig. 4).The center of the TT in the impinged knees

ranged from 18% to 38% from the anterior edge of the tibia (mean, 30 ± 4%). Roof im-pingement occurred when the center of the TT was located 11-23 mm from the anterior edge of the tibia (normalized to a 60-mm-deep tibial plateau (Fig. 4).

The difference in maximum passive knee

FIG. 3. The sagittal location of the TT and degree of roof impingement were determined from calcula-tions derived from five lines drawn on a lateral roent-genogram taken with the knee in maximal extension. The first line was drawn parallel to the intercondylar shelf (Blumensaat’s line), extending across the tibial articular surface. The second line was drawn to over-lie the sclerotic, subchondral margin of the concave medial tibial plateau. The third (number not visible on anterior- most line) and fourth lines outlined the ante-rior and posterior limits of the tibial tunnel. The center of the tibial tunnel was marked (five).

FIG. 4. The lateral roentgenogram outlines the broad anterior (.4) impingement zone and the nar-rower, more posterior (p), anatomic, impingement- free zone. There is little overlap between the zones. The nonanatomic impingement zone ranges from 18 to 38% from the anterior tibial edge, expressed as a percentage of the sagittal depth of the medial tibial plateau (11-23 mm). The more preferable anatomic, impingement-free zone ranges from 37 to 47% (22-28 mm). Centering the TT in the impingement-free zone places the entire TT posterior to the slope of the inter-condylar roof (dotted line) and avoids roof impinge-ment.

Tibial Tunnel Placement 191Number 283October, 1992

extension between the reconstructed and normal knee was 1° ± 2° in the impingment- free group and 4° ± 3° in the impingement group (p < 0.001) (Fig. 5). Knee extension deficit increased as the center of the TT was placed more anteriorly (r = 0.376; p < 0.0 1).

Knee stability was also affected by the loca-tion of the TT (Table 1) (Fig. 6). The manual maximum translation difference between the sur-gically treated and normal side averaged 0.3 ± 1.1 mm in the impingement-free knees and 1.2 ± 2.1 mm in the knees subjected to impingement (p < 0.05). Only one knee in the impingement-free group developed a positive pivot shift (4%), com-pared with five knees with graft impingement that developed recurrent instability (17%) (not signif-icant).

Four knees had a second arthroscopy after graft insertion. One impinged hamstring graft was arthroscopically visualized and a biopsy was per-formed during removal of painful hardware at six months. The graft was separated into several fas-cicles, covered with a thin synovial envelope, and repopulated with cells. This knee remained clini-

cally stable at three-year follow-up examination. Three patellar-tendon-bone grafts required a delayed arthroscopic roofplasty because of pain-ful flexion contractures at two, three, and nine months, respectively, after graft implantation. Bone removal from the intercondylar roof reversed the pain, effusion, and flexion contrac-ture by three weeks. The two-and three-month grafts showed some minor, anterior surface fray-ing, but remained as one ligament bundle. The nine-month graft was more frayed than the others and separated into several distinct, large fascicles. All four of these grafts remained stable at the 16-36- month follow-up examination.

DISCUSSION

Uncorrected graft impingement had a detri-mental effect on the clinical outcome of the ACL reconstruction. Clinically, impinged knees were observed to have limited knee extension and an increased incidence of instability. A primary goal of ACL reconstructions should be to avoid graft impingement.

This study identified the ideal region for cen-

FIG. 5. Roof impingement has prevented this reconstructed knee from regaining the hyperextension evident in the normal knee at three years postoperatively. The width of the TT was 50% anterior to the slope of the intercondylar roof, and the center of the TT was 32% from the anterior tibial edge. The double-looped hamstring graft in this patient demonstrated an increased signal in the distal two thirds of the graft. The laxity measure-ments were identical in each knee at three years. In this study, many stable knees suffered from roof impinge-ment, with limited extension as the only clinical finding.

193 Howell and ClarkClinical Orthopaedics

and Related Research

tering the TT so that the ACL graft was not subjected to roof impingement, and yet had an acceptable excursion profile (2 mm or less tibial-femoral separation distance during passive knee motion from maximum extension to 90-105° flex-ion). The tibial attachment site of the normal ACL was found to be subdivided into two unequal zones; a small 6-mm-deep posterior impinge-ment-free zone and a larger 12-mm-deep anterior impingement zone (Fig. 4). Roof impingement was avoided by positioning the center of the TT within the posterior impingement-free zone (37-47% from the anterior edge of the tibia).9,11

Centering the graft in the impingement-free zone confined all the graft fibers within the original

pathway of the normal ACL, thereby achieving an anatomic placement for the graft.11

The different dimensions and sagittal con-tours between the normal ACL and the graft explain why roof impingement may be a common problem. The broad insertion of the normal ACL is nearly twice as wide as the more proximal por-tion of the ACL that originates from the femur.11,14 The distal portion of the ACL flares anteriorly to accommodate the contour of the intercondylar roof when the knee is extended.11,13,14 A tubular or rect-angular graft cannot replicate this anterior flare of the normal ACL insertion.7 Centering the TT in the anterior impingement zone results in roof impingement because some of the anterior

FIG. 6. A scattergram demonstrates that the instrumented laxity difference (using a MMT) becomes greater and more variable as the TT is moved more anterior (18-39%). Eight impinged grafts had 3 or more mm of side-to-side laxity. Only one unimpinged knee had a laxity difference that was greater than 3 mm. Multiple patients with identical data are represented by the solid vertical bar bisected by either a circle (impinged) or square (unimpinged).

Tibial Tunnel Placement 193Number 283October, 1992

graft fibers course the anterior graft fibers course nonanatomically. These anterior graft fibers lie anterior to the normal flare of the ACL fibers. This nonanatomic TT allows the intercondylar roof to prematurely impact against the straight anterior edge of the graft during knee extension.11

Current recommendations for centering the TT have been to place it either eccentrically (5 mm anterior and medial to the center of the ACL insertion) or within the center of the ACL inser-tion.2,6 An MR study of normal knees that mapped the sagittal dimensions of the normal ACL inser-tion showed that the eccentric location was 29% and the center of the tibial attachment site was 38% from the anterior edge of the tibia.11 Both of these TT locations are within the anterior impingement zone, Moving the center of the TT just 2-3 mm posterior to the center of the ACL insertion avoids roof impingement by anatomi-cally aligning all of the graft fibers within the pathway of the normal ACL.9,11 A roofplasty will be required in both acute and chronic reconstruc-tions if the center of the TT is placed eccentri-cally or centrally.11

Roof impingement may have been avoided if contact between the intercondylar roof and the graft had been recognized at operation and corrected. The authors were cognizant of the potential problem of roof impingement; how-ever, awareness was not sufficient to relieve roof

impingement in 30 of the patients. The intra-operative correction of roof impingement was incomplete because the graft-roof relationship was obscured by the trochlea articulating with the tibial plateau during the final 5˚- 10˚ of exten-sion. Contact between the graft and intercondylar roof could not be seen by the surgeon with the knee fully extended.11

The regionalized MR signal increase con-fined to the distal two thirds of the anteriorly placed grafts supports the assertion that roof impingement was the cause of the signal in-crease. The MR scans confirmed that there was direct contact between roof and graft that often caused posterior bowing when the grafts were aligned in anterior TTs. Impingement-free grafts had a space between the intercondylar roof and the graft. Reinjury or improper tensioning of the graft would have been expected to cause a more diffuse MR signal increase extending along the proximal third of the graft and into the TT.9,10

Grafts centered anatomically in the im-pingement-free zone had better outcomes than grafts centered in anterior TTs. Impingement-free knees regained extension equal to the unoperated side and were consistently more stable than the grafts placed more anteriorly in the impingement zone.9,11

Most of the impinged grafts in this study and others, however, continued to provide stability to

TABLE 1. Comparison of Clinical Parameters

Impinged Nonimpinged Parameter (n = 30) (n = 24) Significance

Lachman Test (number of patients with 9 2 p = 0.036 positive test postoperatively)Pivot Shift Test (number of patients with 5 1 ns positive test postoperatively) p = 0.12 Manual Maximum Test (KT- 1000) (difference 1.3 ± 2.0 mm 0.3 ± 1.0 mm p = 0.05 in translation between involved and normal knee)Number of second operations 4 0 p = 0.05Center of TT 30 ± 5% 42 ± 3% p = 0.0001

TT, tibial tunnel.

194 Howell and ClarkClinical Orthopaedics

and Related Research

the reconstructed knee. A stable, functional knee can coexist with an impinged graft. 9,10 Anterior cruciate ligament grafts must have some capacity to adapt and remodel to repetitive roof impinge-ment.

In summary, relying on the tensiometer to help select the location for the TT may result in large variations in the sagittal location of the TT. Impingement will result if the tunnel is placed nonanatomically in the anterior impinge-ment zone, in which case removal of bone from the intercondylar roof will be required to avoid impingement. This roofplasty needs to be per-formed in both acute and chronic reconstructions. Roof impingement, if untreated, will result in a significant loss of extension at one-year and an increased incidence of instability. Placement of the graft anatomically within the more posterior impingement-free zone avoids roof impinge-ment. Knees with impingement-free graft placement are more likely to be stable and regain complete extension.

REFERENCES 1. Bradley, J.. Fitzpatrick, D., Daniel, D., Shertcliff, T., and

O’Connor, J.: Orientation of the cruciate ligament in the sag-ittal plane. J. Bone Joint Surg. 70B:94. 1988.

2. Clancy, W., Nelson, D.. Reider, B., and Narechania, R.: Ante-rior cruciate ligament reconstruction using one-third patellar ligament, augmented by extra-articular tendon transfers. J. Bone Joint Surg. 64A:352, 1982.

3. Davis, P., Crooks, L., Arakawa, M., McRee, R., Kaufman, L., and Margulis, A.: NMR characteristics of normal and abnor-mal rat tissues. In Kaufman, L., Crooks, L., and Margulis, A. (eds.): Nuclear Magnetic Imaging Medicine. Tokyo, Igaku-Shoin, 1981, pp. 71-100.

4. Fullerton, L. R., and ,Andrews, J. R.: Mechanical block to

extension following augmentation of the anterior cruciate lig-ament: A case report. Am. J. Sports Med. 12: 166, 1984.

5. Gallimore, G., and Harms, S.: Knee injuries: High-resolution MR imaging. Radiology 160:457, 1986.

6. Good, L., Odensten, M., and Gillquist, J.: Precision in recon-struction of the anterior cruciate ligament. Acta Orthop. Scand. 58:658, 1987.

7. Hefzy, M., Grood, E., and Noyes, F.: Factors affecting the region of most isometric femoral attachments part II: The anterior cruciate ligament. Am J. Sports Med. 17:208, 1989.

8. Hoogland, T., and Hillen. 3.: lntra-articular reconstruction of the anterior cruciate ligament: An experimental study of length changes in different ligament reconstructions, Clin. Orthop. 185:197, 1984.

9. Howell, S. M., Berns, G. S., and Farley, T. E.: Signal inten-sity measurements of unimpinged and impinged anterior cru-ciate ligament grafts. Radiology, 179:639, 1991.

10. Howell, S. M,, Clark, J. A., and Blasier, R. D.: Serial mag-netic resonance imaging of hamstring ACL autografts during the first year of implantation: A preliminary study. Am. J. Sports Med. 19:42, 199 1.

11. Howell, S. M., Clark, J. A., and Farley, T. E.: A rationale for predicting anterior cruciate graft impingement by the inter-condylar roof: An MRI study. Am. J. Sports Med. 19:276, 1991.

12. Lukianov, A. V., Richmond, J. C., Barrett, G. R., and Gillquist. J.: A multicenter study on the results of anterior cruciate liga-ment reconstruction using a dacron ligament prosthesis in salvage cases. Am. J. Sports Med. 17:380, 1989.

13. Norwood, L. A., and Cross, M. J.: The intercondylar shelf and the anterior cruciate ligament. Am. J. Sports Med. 5:171, 1977.

14. Odensten, M.. and Gillquist, J.: Functional anatomy of the anterior cruciate ligament and a rationale for reconstruction. J. Bone Joint Surg. 67A:257, 1985.

15. Penner, D., Daniel, D., Wood, P., and Mishra, D.: An in vitro study of anterior cruciate ligament graft placement and isom-etry. Am. J. Sports Med. 16:238, 1988.

16. Reicher, M., Hartzman, S., Bassett, L., Mandelbaum, B., Duckwiler, G., and Gold: R.: MR imaging of the knee: Part 1. Traumatic disorders. Radiology 162:547, 1987.

17. Sidles, J., Larson, R., Garbini. J., Downey, D., and Matsen F.: Ligament length relationships in the moving knee, J. Orthop. Res. 6:593, 1988.

Tibial Tunnel Placement 195Number 283October, 1992