Reconstruction of knees with combined cruciate deficiencies: a biomechanical study

BACKGROUND: Clinical results of dual cruciate-ligament reconstructions are often poor, with a failure to restore normal anterior-posterior laxity. This could be the result of improper graft tensioning at the time of surgery and stretch-out of one or both grafts from excessive tissue forces. The purpose of this study was to measure anterior-posterior laxities and graft forces in knees before and after reconstructions of both cruciate ligaments performed with a specific graft-tensioning protocol. METHODS: Eleven fresh-frozen cadaveric knee specimens underwent anterior-posterior laxity testing and installation of load cells to record forces in the native cruciate ligaments as the knees were passively extended from 120 degrees to -5 degrees with no applied tibial force, with 100 N of applied anterior and posterior tibial force, and with 5 N-m of applied internal and external tibial torque. Both cruciate ligaments were reconstructed with a bone-patellar tendon-bone allograft. Only isolated cruciate deficiencies were studied. We determined the nominal levels of anterior and posterior cruciate graft tension that restored anterior-posterior laxities to within 2 mm of those of the intact knee and restored anterior cruciate graft forces to within 20 N of those of the native anterior cruciate ligament during passive knee extension. Both grafts were tensioned at 30 degrees of knee flexion, with the posterior cruciate ligament tensioned first. Measurements of anterior-posterior knee laxity and graft forces were repeated with both grafts at their nominal tension levels and with one graft fixed at its nominal tension level and the opposing graft tensioned to 40 N above its nominal level. RESULTS: The anterior and posterior cruciate graft tensions were found to be interrelated; applying tension to one graft changed the tension of the other (fixed) graft and displaced the tibia relative to the femur. The posterior cruciate graft had to be tensioned first to consistently achieve the nominal combination of mean graft forces at 30 degrees of flexion. At these levels, mean forces in the anterior cruciate graft were restored to those of the intact anterior cruciate ligament under nearly all test conditions. However, the mean posterior cruciate graft forces were significantly higher than the intact posterior cruciate ligament forces at full extension under all test conditions. Anterior-posterior laxity was restored between 0 degrees and 90 degrees of flexion with both grafts at their nominal force levels. Overtensioning of the anterior cruciate graft by 40 N significantly increased its mean force levels during passive knee extension between 110 degrees and -5 degrees of flexion, but it did not significantly change anterior-posterior laxity between 0 degrees and 90 degrees of flexion. In contrast, overtensioning of the posterior cruciate graft by 40 N significantly increased posterior cruciate graft forces during passive knee extension at flexion angles of <5 degrees and >95 degrees and significantly decreased anterior-posterior laxities at all flexion angles except full extension. CONCLUSIONS: It was not possible to find levels of graft tension that restored anterior-posterior laxities at all flexion positions and restored forces in both grafts to those of their native cruciate counterparts during passive motion. Our graft-tensioning protocol represented a compromise between these competing objectives. This protocol aimed to restore anterior-posterior laxities and anterior cruciate graft forces to normal levels. The major shortcoming of this tensioning protocol was the dramatically higher posterior cruciate graft forces produced near full extension under all test conditions.     

Movement of the posterior cruciate ligament during knee flexion--MRI analysis.

The movement of the posterior cruciate ligament (PCL) during flexion of the living knee is unknown. The purpose of the present study was to analyze the movement of the PCL using magnetic resonance imaging (MRI). The posterior cruciate ligaments in 20 normal knees were visualized using MRI from extension to deep flexion. Sagittal inclination relative to the longitudinal axis of the tibia was measured and analyzed with reference to the patellar tendon (PT) and the anterior cruciate ligament (ACL). Although the PCL was slack in extension, it straightened with anterior inclination (24.1+/-5.1 degrees ) at 90 degrees flexion. At active maximum flexion (129.2+/-8.1 degrees ), the ligament was almost parallel (3.9+/-7.4 degrees inclination) to the longitudinal axis of the tibia. At passive maximum flexion (158.8+/-5.8 degrees ), the inclination was reversed anteroposteriorly, measuring -23.0+/-6.7 degrees . The PCL and PT moved in a corresponding manner within 20 degrees of discrepancy. The results of this in vivo study of the PCL have clinical relevance to conservative therapy for PCL knee injuries. The results of this study could also be useful in PCL reconstruction surgery to determine the optimum graft position to allow maximum postoperative motion.     

Reconstructive interventions of the posterior cruciate ligament--experimental studies of isometric aspects. Part II: Studies of the posterior cruciate ligament replacement model]

Isometric positioning of the posterior cruciate ligament (PCL) graft is important for successful reconstruction of the PCL-deficient knee. This study documents the relationship between graft placement and changes in intra-articular graft length during passive range of motion of the knee. In eight cadaveric knees the PCL was identified and cut. The specimens were mounted in a stabilizing rig. PCL reconstruction was performed using a 9-mm-thick synthetic cord that was passed through tunnels 10 mm in diameter. Three different femoral graft placement sites were evaluated: (1) in four specimens the tunnel was located around the femoral isometric point, (2) in two specimens the tunnel was positioned over the guide wire 5 mm anterior to the femoral isometric point, (3) in two specimens the tunnel was positioned over the guide wire 5 mm posterior to the isometric femoral point. In all knees only one tibial tunnel was created around the isometric tibial point. The location of the isometric points was described in part I of the study. The proximal end of the cord was fixed to the lateral aspect of the femur. Distally the cord was attached to a measuring unit. The knees were flexed from 0 degree to 110 degrees, and the changes in the graft distance between the femoral attachment sites were measured in 10 degrees steps. Over the entire range of motion measured the femoral tunnels positioned around the isometric point produced femorotibial distance changes of within 2 mm. The anteriorly placed tunnels produced considerable increases in femorotibial distance with knee flexion, e.g. about 8 mm at 110 degrees of flexion.(ABSTRACT TRUNCATED AT 250 WORDS)     

A comparison of crushed ice and continuous flow cold therapy

Crushed ice was compared to continuous flow cold therapy for control of postoperative pain after arthroscopic patellar tendon autograft anterior cruciate ligament (ACL) reconstruction. With all other variables held constant, cold was administered by either continuous flow (group 1) or crushed ice (group 2). The cold therapy was constant for 3 days, then as needed in days 4 through 7. Data were collected by investigator evaluations and patient diaries. Pain was assessed by visual analog scale (VAS) and categorical pain scale (Likert). Eighty-seven patients were included (52 continuous flow and 35 crushed ice). Continuous passive motion averaged 54 hours for group 1 and 43 hours for group 2 (P<.05). Knee motion at one week averaged more (5 degrees/88 degrees) for group 1 (flexion range: 48 degrees-155 degrees) than for group 2 (6 degrees/77 degrees) (flexion range: 25 degrees-125 degrees) (P=.03). Likert pain scores for group 2 patients were always statistically greater than group 1 patients from the first hour through postoperative day 6 (P<.01). The average VAS pain was always greater for group 2 and statistically greater for postoperative day 1 (P<.01). Hydrocodone bitartrate with acetaminophen use in group 2 was greater than in group 1 for postoperative days 1 (P<.001) and 2 (P=.035). The respective cold modality VAS measured performance was 9.1 for group 1 and 7.8 for group 2 (P<.01). During postoperative days 4 through 6, group 1 patients applied their cold modality for 47.9 hours but group 2 patients applied their cold modality for 29.5 hours (P<.01). Compared to crushed ice, continuous flow cold therapy lowered VAS and Likert pain scores more, reduced hydrocodone bitartrate with acetaminophen use, was used more often, increased continuous passive motion, increased 1-week knee flexion, and was given significantly higher performance ratings by patients. Continuous flow cold is superior to crushed ice for outpatient ACL reconstruction pain and should not be considered an equivalent modality.     

Clinical application of biomechanic and functional anatomical findings of the knee joint]

In order to study the functional anatomy of the knee joint, careful anatomical dissections were conducted on over 130 fresh-frozen cadaveric knee specimens. We found no evidence to support the two-bundle and three-bundle theories of cruciate ligament fiber patterns. The longest fibers in the anterior cruciate ligaments (ACL) measured 37 mm, and the longest in the posterior cruciate ligament (PCL), 41 mm. Cruciate ligament insertions follow a transition line on tibia and femur. Usually not all the fibers of the cruciate ligaments are taut at the same time. They are progressively recruited according to the biomechanical demands placed on them. Fibre recruitment in the ACL is from knee flexion to extension and in the PCL from extension to flexion. The concept of fiber recruitment was recently evaluated mathematically. As a working hypothesis, the knee joint can be looked upon as a biological realization of the crossed four-bar linkage, even in three dimensions. In vitro measurements have shown that correct graft placement in cruciate reconstructions is critical for knee biomechanics. Incorrect placement of grafts may lead to decreased range of motion and/or increased laxity. Distance changes of 3 mm between femoral origin and tibial insertion of a graft may lead to a 400% increase of graft preload and will thus easily reach published pull-out forces for some of the graft fixation methods (button = 248 N). Precise drill guides and isometers may be helpful in any operative technique (open, arthroscopic). Using the IKDC evaluation form and the KT-1000 arthrometer, our studies on 25 patients demonstrated a direct correlation between intraoperative graft tracking and the clinical outcome 2 years after operation. Biomechanical studies to investigate in vivo strain patterns of the anterior cruciate ligament and in vitro strain patterns of isometrically placed cruciate graft reconstructions showed that they did not reach critical fixation failure or graft rupture loads. The highest values were measured in Lachman and anterior drawer testing. In the 20-kg Lachman test the maximum load was 96 +/- 6 N. This was twice the maximum loads of any of the quadriceps exercises (40-50 N). With an isometrically placed graft, full active range of motion postoperatively and exercises with isolated or combined contractions of quadriceps and hamstring muscles will result in small graft loads. Postoperative immobilization with plaster and/or protection with braces may therefore become obsolete.     

One- and two-strand posterior cruciate ligament reconstructions: cyclic fatigue testing.

This study examined how one- and two-strand posterior cruciate ligament (PCL) reconstructions resist the return of posterior translation during repetitive knee cycling. The femoral attachment of the one-strand graft and the anterior strand of the two-strand (AD2) grafts were located within the anterior one-third of the femoral PCL footprint. The second strand was placed within the middle third of the femoral footprint in one of three locations: middle-distal (MD), middle-middle (MM), or middle-proximal (MP). During repetitive knee cycling from 5 degrees to 120 degrees flexion with a 100 N posterior force, the intact knee had less than 1mm of residual posterior translation after 2048 flexion-extension cycles. Under similar cyclic conditions, the AD2-MM reconstruction achieved the most cycles before failure; however, this two-strand configuration failed in less than 700 cycles. The other reconstructions, either one strand or two strand, failed in less than 350 cycles. The surface failure location for 19 of 25 graft strands was within the femoral one-third of the strand. We concluded that one- and two-strand reconstructions under moderate loading and a range of motion from 5 degrees to 120 degrees flexion have an unacceptably high cyclic failure rate suggesting modifications of the allowable postoperative knee flexion and loading.     

Second-look arthroscopy with removal of bioabsorbable tacks

Eleven years after tearing her anterior cruciate ligament (ACL) (not reconstructed), a 36-year-old dancer reinjured her knee and required arthroscopic ACL reconstruction. At arthroscopy, the medial meniscus had a bucket-handle tear that was repaired by using three bioabsorbable tacks. The ACL was then repaired in the usual manner. Because of persistent posterior knee pain throughout her rehabilitation, we performed "second-look" arthroscopy 14 weeks after reconstruction. The meniscus had healed and was stable; however, tack motion was evident and the tacks were easily removed. Inspection of the tacks showed that the barbs had been resorbed. The patient recovered uneventfully, and pain-free flexion 28 days after surgery was 0 degrees-136 degrees. We believe this to be the first reported case demonstrating the early stages of tack degradation in meniscal repair.     

Graft options for ACL reconstruction

Once the decision for anterior cruciate ligament (ACL) reconstruction has been made, the orthopedist has to decide which graft substitute will best restore normal anterior knee stability. This article will present an overview of graft options as well as fixation options most commonly used in ACL reconstructions.     

Intraoperative measurement of knee kinematics in reconstruction of the anterior cruciate ligament

Our objectives were to establish the envelope of passive movement and to demonstrate the kinematic behaviour of the knee during standard clinical tests before and after reconstruction of the anterior cruciate ligament (ACL). An electromagnetic device was used to measure movement of the joint during surgery. Reconstruction of the ACL significantly reduced the overall envelope of tibial rotation (10 degrees to 90 degrees flexion), moved this envelope into external rotation from 0 degrees to 20 degrees flexion, and reduced the anterior position of the tibial plateau (5 degrees to 30 degrees flexion) (p < 0.05 for all). During the pivot-shift test in early flexion there was progressive anterior tibial subluxation with internal rotation. These subluxations reversed suddenly around a mean position of 36 +/- 9 degrees of flexion of the knee and consisted of an external tibial rotation of 13 +/- 8 degrees combined with a posterior tibial translation of 12 +/- 8 mm. This abnormal movement was abolished after reconstruction of the ACL.     

Tibial tunnel placement in anterior cruciate ligament reconstructions and graft impingement.

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 magnetic resonance signal from origin to insertion. Impinged grafts (n = 30) had an increased magnetic resonance signal confined to the distal two thirds of the graft. The location of the tibial tunnel (TT) was determined from a lateral roentgenogram. Positioning the center of the TT 12-23 mm from the anterior edge of the tibia consistently produced graft impingement and flexion contractures. Roof impingement was avoided and hyperextension 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). Stability 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.     

Cyclops syndrome: loss of extension following intra-articular anterior cruciate ligament reconstruction

Arthrofibrosis is one of the recognized complications following traditional anterior cruciate ligament (ACL) reconstruction. With the advent of arthroscopic assisted ACL reconstructions, the extent of potential arthrofibrosis appeared to be less. However, 13 patients after intra-articular ACL reconstruction using a patella tendon autograft developed a similar symptom complex. In addition to postoperative loss of full extension, there was an audible and palpable clunk with terminal extension. These patients had similar arthroscopic findings of a nodule that formed anterolateral to the tibial tunnel placement of the graft. The arthroscopic appearance of the soft tissue mass with its surface vessels was reminiscent of a "cyclops." After arthroscopy with debridement and manipulation of the knee, extension was improved in all cases. The average range of motion immediately after the procedure was 6.0-130 degrees, compared with 16-103 degrees preoperatively. The range of motion at last follow-up averaged 3.8 degrees of extension and 138 degrees of flexion. All patients had greater than 130 degrees of flexion. There were no complications attributed to the manipulation and arthroscopic lysis of adhesions, and no patient experienced loss of graft integrity or knee stability. The "cyclops" nodule was examined grossly and microscopically and demonstrated peripheral fibrous tissue with a central region of granulation tissue in all specimens. In addition, two specimens were noted to include bony fragments and three specimens contained cartilaginous tissue.     

Isometry testing for anterior cruciate ligament reconstruction revisited

The purpose of this study was to determine the effect, if any, that varying the distal testing position (tibial level) has on isometry data produced with a common anatomic proximal testing position at the native anterior cruciate ligament (ACL) origin. During ACL reconstruction in 25 knees, in vivo isometry measurements were recorded using two different isometry testing methods, which differed in the tibial level of the distal fixation testing point. Method 1 tested distally at a point 13 cm peripheral to the native aCL insertion on a vector in line with the tibial tunnel. Method 2 tested distally at a point central in the native ACL insertion at the level of the intercondylar floor. All tibial tunnels were standardized with similar sagittal tunnel-plateau angles and similar tunnel lengths. The proximal testing point was standardized at a point that was anatomically located at or near the central ACL origin 7 mm anterior to “over the top” in all knees. Using these methods, length changes between the proximal and distal testing points were recorded in each knee with each testing method, with the knee ranged from 70° of flexion to full extension and from 70° to 140° of flexion. From these data, a total excursion from 0° to 140° of flexion was calculated. A nonanatomic distal testing point (Method 1) produced a 6 mm ± 1 mm total excursion, whereas anatomic testing points (Method 2) in the same knees produced a 1 mm ± 1 mm total excursion. From these data, the authors conclude that the tibial level of the distal isometry testing point has a significant effect on the resultant isometry measurement such that anatomic testing points are most isometric. Isometers that produce data between nonanatomic testing points should not be used to position tunnels for ACL reconstruction and should not be used to assume the elongation forces an ACL substitute will see when fixed at different points. Conversely, the clinical relevance of this study is that both anatomic graft position and anatomic graft fixation position are important and, when achieved, should result in minimal graft elongation with early postoperative range of motion, leading to a more stable long-term result.     

膝关节后内侧结构损伤合并单一交叉韧带断裂的早期治疗

目的 探讨膝关节后内侧结构损伤合并单一交叉韧带断裂进行早期手术的疗效.方法 2002年1月至2005年12月共治疗12例后内侧结构损伤合并单一交叉韧带断裂患者,其中10例合并前交叉韧带(ACL)断裂,2例合并后交叉韧带(PCL)断裂.交叉韧带损伤术前Lysholm评分为50～60分(平均56.7分).关节镜下重建交叉韧带,开放修复后内侧结构.8例采用自体半腱肌、股薄重建ACL(transfix术式),2例采用骨.髌腱.骨重建ACL.2例采用一端带骨块的异体跟腱蓖建PCL.后内侧结构损伤修复:8例采用星状钢板螺钉同定,2例采用GⅡ锚钉固定.1例采用自体半肌腱、股薄肌移植重建,1 例采用端对端缝合.结果 12例中除2例随访4个月后失访外,其余10例患者术后获平均12个月(6～18个月)随访.交叉韧带损伤重建后Lysholm评分为74～94分(平均81.2分).后内侧结构修复后10例膝伸屈范围正常,2例伸直受限5.外翻应力试验于O啦时,9例正常,2例弱阳性(+),1例阳性(++).结论 膝后内侧结构损伤合并单一交叉韧带断裂时,早期重建交叉韧带同时一期修复膝后内侧结构可以较好地恢复膝关节稳定性.     

Revision anterior cruciate ligament reconstruction using the lateral third of the ipsilateral patellar tendon after failure of a central-third graft: a preliminary report on 10 patients

Long-term outcomes were reported for 10 (77%) of 13 cases of revision anterior cruciate ligament (ACL) reconstruction using the lateral third of the ipsilateral patellar tendon as a graft. All primary ACL reconstructions were ipsilateral central-third bone-patellar tendon-bone graft procedures. Mean age at follow-up was 30.7 years, and mean time from revision ACL surgery to follow-up was 42.9 months. At follow-up, average KT-1000 difference between knees was 2.4 mm. All patients had a negative pivot shift, extension within 5 degrees of the contralateral knee, and flexion within 15 degrees. Mean bilateral comparison ratios for isokinetic strength and hop testing were: extension, 83.5%; flexion, 96%; and single-leg hop 96.9%. No patella fractures or tendon ruptures had occurred. All patients had returned to their previous work level, and 8 of the 10 patients could participate in at least "moderate" sports activities (e.g., skiing and tennis). The results were comparable to published outcome reports for both primary and revision ACL reconstruction. The lateral third of the ipsilateral patellar tendon is a good graft option for revision ACL reconstruction.     

Factors affecting graft force in surgical reconstruction of the anterior cruciate ligament

The effect of the maximum unloaded graft length (Lo) and femoral fixation hole location on graft force with the knee under anteriorly directed tibial loads was measured in five fresh cadaver knees with a reconstruction of the anterior cruciate ligament (ACL). The reconstruction was performed using a composite graft consisting of the semitendinosus and gracilis tendons augmented with the Kennedy ligament augmentation device (LAD). Buckle transducers were used to measure ligament and graft forces. The total graft force was adjusted to match the intact ACL at 30 degrees flexion using a force-setting method so that a standardized reference configuration could be repeatedly obtained. The graft force was highly sensitive to Lo, typically changing by 50% with a change in Lo of 3 mm. Variation in femoral hole location of 5 mm anterior, posterior, proximal, and distal to the anatomic position produced changes in graft force, particularly at 60 degrees and 90 degrees flexion; however, these changes were not statistically significant. The effect of femoral hole location varied considerably between knees. This variability makes predicting proper hole placement difficult, and suggests the need to adjust each knee at surgery to account for this variable femoral hole position sensitivity.     

Tension in a double loop tendon anterior cruciate graft during a simulated open chain knee extension exercise.

Concerns exist regarding the tension developed in a reconstructed anterior cruciate ligament (ACL) during open chain knee extension exercises used to rehabilitate the knee. Therefore, the primary objective was to measure tension in an ACL graft during a simulated open chain knee extension exercise as a function of ankle weight. A secondary objective was to determine whether the graft tension was reduced with relatively high stiffness fixation. The open chain exercise was simulated in seven cadaveric specimens in which the ACL had been reconstructed with double loop tendon grafts. Graft tension was measured at 15 degrees of flexion as the effective ankle weight was increased from 22.5 to 67.5 and then to 112.5 N for three different fixation stiffnesses (25, 125, and 225 N/mm). The initial tension was set to restore the 225 N anterior limit of motion to that of the intact knee at 30 degrees of flexion. Increasing the ankle weight caused the graft tension to increase significantly (p<0.0001), but the increase with the highest ankle weight was only 62 N on average. Increasing the fixation stiffness caused the graft tension to decrease significantly (p<0.0001) because the initial tension decreased by 107 N as the fixation stiffness increased. Because the graft tension with the highest ankle weight was limited to 112 N on average, high stiffness fixation methods, which are also resistant to lengthening in the region of the fixation, may reduce the risk of graft construct lengthening during open chain knee extension exercises.     

Continuous passive motion after arthroscopically assisted anterior cruciate ligament reconstruction: comparison of short- versus long-term use

The use of continuous passive motion (CPM) following anterior cruciate ligament (ACL) reconstruction has become common. The duration of use of CPM for maximal therapeutic benefit is not known. This study compared 4-day CPM use with 14-day CPM use following arthroscopic ACL reconstruction using a bone-patellar tendon-bone autograft prospectively in 20 patients. The patients were randomly allocated to the CPM 4-day group [6 h daily CPM for 4 days in hospital followed by intermittent passive motion (IPM) at home] or to the CPM 14-day group (6 h daily CPM for 14 days). The objective parameters measured were girth measurements at four lower limb locations for joint swelling and muscle atrophy; range of motion of the knee, measured goniometrically; and KT-1000 arthrometry measurements for joint laxity. The measurements were made prior to surgery, and on days 2, 7, 14, and 42, postoperatively. There were no statistically significant differences (p greater than 0.05) at 42 days postoperatively between groups in all parameters measured with the exception of KT-1000 laxity at 42 days.     

Method for setting total graft force and load sharing in augmented ACL grafts

The objective of this study was to develop a method for obtaining a controllable and reproducible immediate postoperative mechanical state in a knee with an anterior cruciate ligament (ACL) reconstruction. This method, called the force-setting technique, was demonstrated using a composite graft consisting of the middle third of the patellar tendon with bone blocks (PT) and the ligament augmentation device (LAD). The total graft force was set to match the force in the intact ACL at 30 degrees flexion with the knee under the same standardized external load, while at the same time the load sharing between the biologic and augmentation components was controlled. The total graft force was set to match the ACL force three separate times in each knee, with ratios of load sharing set at the following levels: 50% PT-50% LAD, 25% PT-75% LAD, and 75% PT-25% LAD. ACL, PT, LAD, and collateral forces were measured using buckle transducers, and three-dimensional knee motion was measured using an instrumented spatial linkage as 90 N anteriorly directed tibial loads were applied to eight specimens at 0 degree, 30 degrees, 60 degrees, and 90 degrees flexion with an intact ACL, an excised ACL, and the three load-sharing reconstruction states. The total graft force could be consistently set to within an average of 2% of the intact ACL force at 30 degrees flexion, and load sharing between the graft segments could be set to within an average of 5.1% of the desired ratio at 30 degrees.(ABSTRACT TRUNCATED AT 250 WORDS)     

Posterior cruciate ligament removal contributes to abnormal knee motion during posterior stabilized total knee arthroplasty

Abnormal anterior translation of the femur on the tibia has been observed in mid flexion (20–60°) following posterior stabilized total knee arthroplasty. The underlying biomechanical causes of this abnormal motion remain unknown. The purpose of this study was to isolate the effects of posterior cruciate ligament removal on knee motion after total knee arthroplasty. We posed two questions: Does removing the posterior cruciate ligament introduce abnormal anterior femoral translation? Does implanting a posterior stabilized prosthesis change the kinematics from the cruciate deficient case? Using a navigation system, we measured passive knee kinematics of ten male osteoarthritic patients during surgery after initial exposure, after removing the anterior cruciate ligament, after removing the posterior cruciate ligament, and after implanting the prosthesis. Passively flexing and extending the knee, we calculated anterior femoral translation and the flexion angle at which femoral rollback began. Removing the posterior cruciate ligament doubled anterior translation (from 5.1 ± 4.3 mm to 10.4 ± 5.1 mm) and increased the flexion angle at which femoral rollback began (from 31.2 ± 9.6° to 49.3 ± 7.3°). Implanting the prosthesis increased the amount of anterior translation (to 16.1 ± 4.4 mm), and did not change the flexion angle at which femoral rollback began. Abnormal anterior translation was observed in low and mid flexion (0–60°) after removing the posterior cruciate ligament, and normal motion was not restored by the posterior stabilized prosthesis. © 2008 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 26:1494–1499, 2008