Proprioception in the posterior cruciate ligament deficient knee

Knee Surg, Sports Traumatol, Arthrosc (1999) 7 : 310–317 KNEE © Springer-Verlag 1999 M. R. Safran A. A. Allen S. M. Lephart P. A. Borsa F. H. Fu C. D. Harner Received: 5 September 1998 Accepted: 25 May 1999 M. R. Safran (Y) Department of Orthopaedic Surgery, University of California at Irvine, Kaiser Permanente, Orange County, 3010 West Orange Ave., Anaheim, CA 92804, USA e-mail: [email protected] Tel.: +1-714-2362851 Fax: +1-714-2362886 A. A. Allen Hospital for Special Surgery, New York, New York, USA S. M. Lephart · F. H. Fu · C. D. Harner Neuromuscular Research Laboratory, University of Pittsburgh, 140 Trees Hall, Pittsburgh, PA 15261, USA P. A. Borsa Oregon State University, Corvallis, Oregon, USA Proprioception in the posterior cruciate ligament deficient knee Abstract This study was undertaken to evaluate knee proprioception in patients with isolated unilateral posterior cruciate ligament (PCL) injuries. Eighteen subjects with isolated PCL tears were studied 1–234 months after injury. The threshold to detect passive motion (TTDPM) was used to evaluate kinesthesia and the ability to passively reproduce passive positioning (RPP) to test joint position sense. Two starting positions were tested in all knees: 45 ° (middle range) and 110 ° (end range) to evaluate knee proprioception when the PCL is under different amounts of tension. TTDPM and RPP were tested as the knee moved into flexion and extension from both starting positions. A statistically significant reduction in TTDPM was identified in PCL-in- Introduction The treatment of isolated posterior cruciate ligament (PCL) injuries is controversial. This is partly because the natural history of the isolated PCL injured knee is unclear. It has been noted that many patients function well, even in high level sports, with a torn PCL. This is in direct contradistinction with the ACL injured knee. The reason for this difference is not clear, as the PCL is larger and stronger than the ACL. For years knee surgeons have postulated that the sensory loss associated with ACL injury may affect the results of ACL repair and reconstruction [3, 7, 41]. DuToit [19], Insall et al. [34], and others [42, 47, 53] have all ad- jured knees tested from the 45 ° starting position, moving into flexion and extension. RPP was statistically better in the PCL-deficient knee as tested from 110 ° moving into flexion and extension. No difference was identified in the TTDPM starting at 110 ° or in RPP with the presented angle at 45 ° moving into flexion or extension. These subtle but statistically significant findings suggest that proprioceptive mechanoreceptors may play a clinical role in PCL-intact and PCL-deficient patients. Further, it appears that kinesthesia and joint position sense may function through different mechanisms. Key words Posterior cruciate ligament · Proprioception · Ligament · Knee vocated certain reconstructive techniques due in part to increased afferent preservation. A proprioceptive deficit has been demonstrated following ACL disruption [3, 4, 7, 13]; however, no literature exists reporting proprioception in the PCL deficient knee. Unlike combined ligament injuries involving the PCL, there is much more debate about the natural history and treatment of the “isolated” PCL-deficient knee [9, 15, 16, 18, 22, 39, 51, 63]. It is well known that in the anterior cruciate ligament (ACL) deficient knee instability and reinjury can lead to arthritis over time [8, 23, 31, 46, 48, 54, 57, 61]. Further, authors suggest that function in the ACL-deficient and ACL-reconstructed knee is more reliably predicted by proprioceptive ability than physical examination or knee test scores [4, 7, 13, 33]. Several stud- 311 ies have shown that proprioceptive deficits that exist in ACL-deficient knees can be partially restored by surgical reconstruction [3, 4, 12]. Reduced proprioception has been implicated in the development of knee arthritis [5, 59]. Many authors note that patients with “isolated” PCL deficiency initially function well while progression to degenerative arthritis over time is less well defined [9, 18, 51]. To date no studies have been performed to determine the potential proprioceptive deficits in the PCL deficient knee. With this information in mind, it was the purpose of this study to characterize the proprioceptive changes in the isolated PCL-deficient knee. Materials and methods Eighteen subjects met the criteria for inclusion: isolated PCL injury without contralateral knee injury, surgery or other concomitant knee injury. The 13 men and 5 women averaged 32 years of age (range 19–51). These patients averaged 29 months from time of injury to proprioception testing (range 1–234). There were 10 right knees injured and 8 left knees. The mechanism of injury was hyperextension in 7, flexion/hyperflexion in 8, and unknown in 3. Many of the patients presented to Orthopaedic Sports Medicine Service after initial injury and followed with conservative management. Most of these patients were asked to return for follow-up and testing even though they remained asymptomatic at latest evaluation. All subjects underwent a complete history and thorough knee examination by one of the authors, an orthopedic surgeon specializing in sports injuries. The bilateral knee examination included assessment of knee range of motion, lower extremity alignment, presence of effusion, patellar irritability, patellar motion, anteroposterior and varus-valgus stability, and meniscal signs. All but three individuals (including the one who underwent PCL reconstruction 1 month after the injury) underwent physical therapy stressing quadriceps rehabilitation. In the normal knee the medial tibial plateau normally is 10 mm anterior to the medial femoral condyle with the knee in 90° flexion. Posterior drawer testing for PCL insufficiency is then graded as follows: in grade I injury there is asymmetry side to side, but the medial tibial plateau remains anterior to the medial femoral condyle; in grade II injury the posterior drawer at 90° pushes the tibial plateau to the level of the medial femoral condyle; in grade III injury the medial tibial plateau can be pushed posterior to the medial femoral condyle. The posterior sag is graded similarly: in grade I the medial tibial plateau sits further posteriorly than the “normal” contralateral knee but still anterior to the medial femoral condyle with the knee flexed approximately 90°; in grade II the tibial plateau is at the level of the medial femoral condyle; in grade III the medial tibial plateau rests posterior to the medial femoral condyle. By definition an isolated PCL tear precludes inclusion of patients with injury to the anterior cruciate or collateral ligament in this study. Furthermore, posterior lateral rotatory instability, “dial test,” at 30 ° must be symmetric. It is important to note for this study that we tested patients with truly isolated PCL tears. In the past many so-called isolated grade III PCL injuries have been included in studies as isolated injuries, although concomitant capsular and posterolateral corner involvement likely existed. In this study these patients were excluded. A standard radiographic knee series, including flexion weightbearing posteroanterior view, lateral and sunrise views, and a magnetic resonance imaging scan were obtained on all subjects. KT-1000 instrumented knee testing was also performed to assess the degree of laxity of both knees for comparison. A proprioceptive testing device (PTD) was used to measure kinesthesia as the threshold to detection of passive movement Fig. 1 PTD utilized in this study. The PTD rotates the knee into flexion and extension through the axis of the joint. A rotational transducer interfaced with a digital microprocessor counter provided the angular displacement values directly. The subject is prone. A pneumatic compression boot is placed on each foot to reduce cutaneous input. One pneumatic boot was attached to the moving bar of the PTD, the other to a stationary bar. The subject is blindfolded to eliminate visual cues. For TTDPM testing, headphones with white noise are used to eliminate auditory cues. The subject shown schematically is holding an on-off switch as utilized for RPP testing to passively reproduce the presented angle (TTDPM) and joint position sense by the ability to passively reproduce joint positioning (RPP; Fig. 1). This device has been used previously to assess proprioceptive awareness and is painless to the subjects [44]. The PTD rotates the knee into flexion and extension through the axis of the joint. A rotational transducer interfaced with a digital microprocessor counter provided the angular displacement values directly. A pneumatic compression boot was placed on each foot to reduce cutaneous input. One pneumatic boot was attached to the moving bar of the PTD and the other to a stationary bar. The potential for input from the anterior thigh cutaneous receptors exists since the anterior thigh rests on the stable (not moving) frame of the testing apparatus. The input would likely need to be through the stretching of the distal thigh skin as the knee is moved, although this is felt to be negligible since the distal one-half to one-third of the thigh was not touching the apparatus. The knee was tested moving into flexion and extension from two starting positions. These positions were 45° and 110° of knee flexion. These starting positions were chosen as the middle range of knee motion (45 °) where the capsule, ACL, and PCL are relatively relaxed, and thus the poorest proprioceptive sensory results should be noted. Thus no difference should exist between the injured and uninjured knees. Knee flexion of 110° was chosen since tension on the PCL should play a role in proprioception, and thus some difference in afferent input might be identified. Two familiarity trial tests were performed prior to the subjects being blindfolded and having a headset placed over the ears to negate visual and auditory cues. Testing was performed in a single session with test order of injured and uninjured knee, starting position, and direction of movement being randomized and counterbalanced. The PTD tester was blinded as to the normal and PCL injured knee. Instrument reliability was established previously as intraclass correlations were calculated using a fixed model and ranged from in value from 0.87 to 0.92. Threshold to detection of passive motion TTDPM assessment was started with the motor and shaft of the PTD disengaged. Subjects were blindfolded and had earphones placed over their ears. The subject gave a thumb-up signal to indicate readiness to perform the test. At a random point during the subsequent 20 s knee movement was engaged by the tester. The subject disengaged the PTD by pressing a hand held switch upon perception of sensation of movement at the knee. The PTD rotated the knee at a constant angular velocity of 0.5°/s. This slow speed was chosen to minimize contribution from muscle receptors. Three trials from a starting position of 45° and 110° knee flexion moving into both flexion and extension were performed. The number of degrees the PTD moved the knee by the time the subject disengaged the motor was recorded as the TTDPM. Both the injured and uninjured knees were tested. Mean TTDPM values were calculated for the four test conditions. Reproduction of passive positioning The subjects were blindfolded but permitted to communicate with the PTD tester during RPP testing. As with the TTDPM testing, 45° and 110° knee flexion were used as starting positions (reference angle). After confirmation of the subject’s readiness, the knee was moved passively 10 ° into further flexion or extension (presented angle) by the tester. The angles were presented at variable velocities in order to reduce any time associated cues. The limb was held in the presented angle position for 10 s, and the subject was asked to concentrate on this position. The limb was then returned passively to the reference angle by the examiner. The subject was then instructed to manipulate the on/off switch to reproduce the previously presented angle at an angular velocity of 0.5 °/s. This was recorded and repeated for each of the three trials moving into flexion and extension. The difference between the presented angle and the angle that was repositioned by the subject was calculated as the error of reproduction. The mean of three trials was calculated for the four test conditions. PCL injured and uninjured knee mean differences were analyzed using a paired t test for both TTDPM and RPP testing. Pearson product-moment correlation coefficients were established between all dependent variables. Results Although all of the patients complained of mechanical symptoms (locking, giving way) initially, ten (55%) of the patients noted some instability symptoms at follow-up. No subjects complained of locking of the knee. Thirteen (72%) complained of some pain in the knee at follow-up, the majority having anterior knee pain. Four (22%) subjects had instability with activities of daily living, and six had occasional giving-way with sporting activities at the most recent evaluation. Nine (50%) complained of knee pain or instability going up or down stairs, and six (33%) complained of slight recurrent effusions, primarily with a high level of activity. Nine (50%) were able to return to their previous level of activity, including four collegiate/professional athletes. On physical examination 11 subjects had full flexion and extension of the affected knee. Six subjects had an average 9° loss of flexion (range 5–12 °), and one lacked 4° of knee hyperextension. Three subjects had a small effu- Degrees of Passive motion 312 2.5 2 1.5 1 0.5 0 Injured Uninjured 45 Degrees-Flexion Injured Uninjured 45 Degrees-Extension * P < 0.05, denote significant difference Fig. 2 Kinesthestic sensation as measured by the TTDPM of the PCL-injured and normal contralateral knees at the 45° starting angle, moving into flexion and extension. Kinesthesia was statistically significantly worse in the PCL-injured knee than in the uninjured knee, both moving into flexion and moving into extension sion, two had trace effusion. Ten subjects (56%) had crepitation of the patellofemoral joint. Twelve (67%) had tenderness of the medial facet of the patella, and six (33%) had tenderness of the lateral facet of the patella (four had tenderness of both the medial and lateral facets). Fifteen (83%) patients had a grade II posterior drawer test on physical examination, while the remaining three (17%) had a grade 3 posterior drawer test. Two subjects had minimal laxity of the medial collateral ligament (MCL), while there was no other ligamentous injury identified on physical examination (ACL, MCL, LCL, rotatory instability) of any of the subjects. KT-1000 instrumented testing revealed an average 7.5 mm (range 2.5–12.5 mm) manual maximum side to side difference at 90° and 3.1 mm manual maximum side to side difference at 20° of knee flexion. Plain radiographs showed no degenerative changes in any of the 18 subjects. Magnetic resonance imaging of all 18 subjects revealed an isolated PCL tear with no meniscal or chondral injuries. For TTDPM, starting at 45°, the PCL injured knee averaged 1.5 ± 0.2°, while the uninjured knee averaged 1.2 ± 0.1° (P = 0.051) as the knee was moving into extension (Fig. 2, Table 1). TTDPM at 45 ° moving into flexion averaged 1.9 ± 0.4° for the involved knee and 1.2 ± 0.2° for the uninvolved knee (P = 0.022; Fig. 2, Table 1). At 110° TTDPM values did not differ statistically significantly between PCL-injured and normal knees moving into flexion and into extension (Fig. 3, Table 1). Testing RPP at 110°, the injured knee averaged 2.3 ± 0.4° error from the true test angle and the uninjured knee average 3.1 ± 0.6° error as the test angle was brought into more extension (P = 0.050; Fig. 4, Table 1). RPP at 110° testing into flexion showed an average error of 2.2 ± 0.3° for the involved knee and 3.0 ± 0.4° for the uninvolved 313 Test Injured knee Uninjured knee P TTDPM 45° extension 45° flexion 1.51 ± 0.19° 1.87 ± 0.35° 1.19 ± 0.10° 1.21 ± 0.22° 0.051 0.022 RPP 45° extension 45° flexion 2.43 ± 0.33° 2.97 ± 0.47° 2.75 ± 0.33° 3.42 ± 0.30° 0.224 0.155 TTDPM 110° extension 110° flexion 1.54 ± 0.22° 1.38 ± 0.24° 1.43 ± 0.18° 1.29 ± 0.25° 0.119 0.290 RPP 110° extension 110° flexion 2.28 ± 0.37° 2.15 ± 0.28° 3.11 ± 0.56° 2.96 ± 0.36° 0.050 0.050 3 2.5 2 1.5 1 0.5 2.5 2 1.5 0 Injured Uninjured 110 Degrees-Flexion Injured Uninjured 110 Degrees-Extension * P < 0.05, denote significant difference Fig. 4 Joint Position Sense as measured by RPP of the PCL-injured and normal contralateral knees at the 110° starting angle, moving into flexion and extension. Joint position sense was statistically significantly better in the PCL injured knee than in the uninjured knee both moving into flexion and moving into extension 4 3,5 Degrees of error TTDPM values represent the average number of degrees (with standard error) before the subject sensed the knee moving. RPP values represent the average number of degrees error (with standard error) of the subjects’ knee angle from the true presented angle Degrees of Passive motion 4 3,5 Degrees of error Table 1 Data for each testing scenario (45 °, 110 ° starting angle, extension test knee moving into extension, flexion test knee moving into flexion) 3 2.5 2 1.5 1 0.5 0 Injured Uninjured 45 Degrees-Flexion 1 Injured Uninjured 45 Degrees-Extension * P < 0.05, denote significant difference 0.5 0 Injured Uninjured 110 Degrees-Flexion Injured Uninjured 110 Degrees-Extension * P < 0.05, denote significant difference Fig. 3 Kinesthestic sensation as measured by the TTDPM of the PCL-injured and normal contralateral knees at the 110° starting angle, moving into flexion and extension. Kinesthesia was not statistically significantly different in the PCL injured knee than in the uninjured knee either moving into flexion or moving into extension knee (P = 0.050; Fig. 4, Table 1). RPP values in 45 ° did not differ significantly between PCL-deficient and normal knee with the test angle in more flexion or extension (Fig. 5, Table 1). A correlation matrix revealed a significant correlation between the time from injury and the ability to passively reproduce a joint angle at 110° flexion moving into extension (r = 0.687). Thus, the longer time from injury, the better was the subject’s RPP. Knee laxity measurements, as quantified by KT-1000 knee ligament arthrometer (MedMetric, San Diego, Calif., USA), were not correlated with proprioceptive measurements. Fig. 5 Joint Position Sense as measured by RPP of the PCL-injured and normal contralateral knees at the 45° starting angle, moving into flexion and extension. Joint position sense was not statistically significantly different between the PCL-injured and the contralateral, normal knee, moving into flexion or moving into extension Discussion Proprioception is considered a specialized variation of the sensory modality of touch and encompasses the sensations of joint movement (kinesthesia) and joint position (joint position sense). Conscious proprioception is essential for proper function in activities of daily living, sports, and occupational tasks. Unconscious proprioception modulates muscle function and initiates reflex stabilization. Much effort has been dedicated to elucidating the mechanical function of knee articular structures and the corresponding mechanical deficits that occur secondary to disruption of these structures. Knee articular structures may also have a significant sensory function which plays a role in dynamic joint stability, acute and chronic injury, pathological wearing, and rehabilitation training. 314 Extrinsic innervation of joints follows Hilton’s law [67], which states that joints are innervated by articular branches of the nerves supplying the muscles that cross that joint. The afferent innervation of joints is based on peripheral receptors located in articular, muscular, and cutaneous structures. Articular receptors include nociceptive free nerve endings and proprioceptive mechanoreceptors. Ruffini endings, Pacinian corpuscles, and Golgi tendon organs are mechanoreceptors that have been histomorphologically identified in the ACL [30, 40–42, 52, 55, 56], PCL [38, 56], meniscus [45, 49, 50, 68], lateral collateral ligament [17] and infrapatellar fat pad [43]. Mechanoreceptors transduce some function of mechanical deformation into a frequency modulated neural signal which is transmitted via cortical and reflex pathways. An increased stimulus of deformation is coded by an increased afferent discharge rate or an increased population of activated receptors. Grigg and Hoffman [27, 29] have correlated mechanoreceptor afferent discharge with strain energy density and have calibrated mechanoreceptors as in vivo load cells in the posterior capsule of the feline knee. Receptors demonstrate different adaptive properties based on their response to a continuous stimulus. Quick adapting (QA) mechanoreceptors, such as the Pacinian corpuscle, decrease their discharge rate to extinction within milliseconds of the onset of a continuous stimulus. Slow-adapting (SA) mechanoreceptors, such as the Ruffini ending and the Golgi tendon organ, continue their discharge in response to a continuous stimulus. QA mechanoreceptors are very sensitive to changes in stimulation and are therefore thought to mediate the sensation of joint motion. Different populations of SA mechanoreceptors are maximally stimulated at specific joint angles, and thus a continuum of SA receptors is thought to mediate the sensation of joint position [10, 32, 36]. In animal models these mechanoreceptors respond to active or passive motion with maximal stimulation occurring at the extremes of knee motion [26–28, 37]. Stimulation of these receptors results in reflex muscle contraction about the joint [6, 20, 35, 60]. The muscle spindle receptor is a complex fusiform, SA receptor found within skeletal muscle. Via afferents and efferent to intrafusal muscle fibers, the muscle spindle receptor can detect and regulate muscle strain over a large range of extrafusal muscle length. There is considerable debate over the relative contribution of muscle receptors versus joint receptors to proprioception, with traditional views emphasizing muscle receptors [10, 14, 24–26]. Recent work suggests that muscle receptors and joint receptors are probably complementary components of an intricate afferent system in which each receptor modifies the function of the other [6, 21, 27]. Kinesthesia is assessed functionally by measuring TTDPM, and joint position sense by measuring RPP. In patients with unilateral joint involvement the contralateral knee serves as an internal control, and uninjured knees in a normative population serve as external controls. Using these measures in the knee, investigators have found proprioceptive deficits with aging [2, 5, 58], arthrosis [2, 5, 59], and ACL disruption [3, 4, 7, 13]. These processes damage articular structures containing mechanoreceptors and are thus hypothesized to result in partial deafferentation with resultant proprioceptive deficits. Proprioceptive enhancement has been found to occur in ballet dancers [1] and also with the use of an elastic knee sleeve [4, 44], suggesting that training and bracing may have proprioceptive benefits. The use of the TTDPM as a measure of kinesthesia has been established by previous studies. Slow, painless, passive motion was used in this investigation, as this is thought to maximally stimulate slow-adapting joint mechanoreceptors while minimally stimulating muscle receptors [3]. Although we focused primarily on joint receptors in joint injury, muscle receptors are an integral component of a complex afferent system and may also play a role in kinesthetic awareness of slow, passive motion. In addition to reflex pathways, joint mechanoreceptors have been shown to have cortical pathways that account for conscious appreciation of joint movement and position. While much research exists about ACL deficiency, including information about proprioception in the ACL deficient knee [3, 4, 7, 12, 13, 42, 44], data regarding PCL injuries are lacking [11, 62, 64]. We have studied the proprioceptive function of a selective group of human subjects with isolated injuries to the PCL. Few other studies identifying proprioceptive mechanoreceptors within the substance of the PCL have attempted to assess the clinical function of these mechanoreceptors within the PCL. One previous study found better joint position sense in patients undergoing total knee arthroplasty using a PCL retaining prosthesis than with a PCL-sacrificing implant [64]. Joint position sense was measured by the subjects moving a hand held knee model to replicate the perceived amount of passively placed knee flexion. Twenty-five subjects tested underwent a total knee replacement with a PCL retaining prosthesis more than 1 year prior to testing and were compared with nine age-matched controls and 30 patients with PCL-sacrificing knee arthroplasties [64]. However, Tibone et al. [62], in another PCL proprioception related study, reported no electromyographic differences between PCL-deficient (some with posterolateral corner injury) and PCL-reconstructed knees (using medial head of the gastrocnemius) during functional tasks [62]. Both groups had abnormal findings during gait [62]. It may be that those with PCL reconstruction, who were quite symptomatic preoperatively, may have had greater proprioceptive deficits. Thus the discrepancy as compared with our findings may be due to many possible factors such as nonanatomic reconstruction (medial head of gastrocnemius using muscles that may affect proprioceptive input versus our intra-articular reconstruction of the anterolateral bundle of the PCL), large preoperative proprio- 315 ceptive deficits that may only be incompletely restored, or their testing protocol which is functional requiring input from muscle and other fibers as compared with our slow moving, passive model. One published study more similar to ours studied the threshold to detect passive positioning in eight patients with isolated PCL deficient knees [11]. Their eight patients averaged 34 years of age, seven were men, and the average time from injury to testing was 3 years (8 months– 6 years). These patients were tested for TTDPM at 0.5°/s in the sitting position with their knee moved into flexion or extension from 37°. These authors found statistically significantly less ability to detect passive motion in the PCL-injured knee than in the normal, contralateral knee [11]. We also found statistically significant differences in the TTDPM at a similar range (45° moving both into flexion and extension), although we did not find the reduced TTDPM at 110° of knee flexion. This greater degree of extension was not studied by Clark et al. [11], nor was RPP. We studied the proprioceptive function of a selective group of human subjects with isolated injuries to the PCL. The subjects studied are for the most part examples of the clinical best-case scenario. These are subjects who have been treated nonoperatively for isolated grade 2 and 3 posterior laxity of the knee. Subjects with more significant injury are more likely to undergo early knee ligament reconstruction. Therefore individuals who undergo early ligament reconstruction may be expected to exhibit more significant proprioceptive differences. Proprioceptive deficits have been identified in the ACL deficient knee [3, 44]. These proprioceptive losses are reduced more significantly at 15° than at 45° flexion in the ACL-deficient knee. This is expected since the ACL has more force at 15° flexion and thus more input in functioning mechanoreceptors. Further, Barrett [4], in a study of ACL-reconstructed patients, found patient satisfaction and function to be correlated with proprioceptive function rather than with clinical examination and knee scores. We have shown that isolated PCL deficiency in the human knee does result in reduced kinesthesia, as tested by the threshold to detect passive positioning and enhanced RPP. There are many potential reasons for these findings that are beyond the scope of this study and are outlined below, although one potential reason is that the proprioceptive mechanoreceptors within the PCL have some clinical function. Proprioceptive deficits in studies of patients with ACL disruption reveal greater proprioceptive deficits, both in magnitude and over a greater range of motion, than the findings presented here for PCL deficiency. It has been argued that proprioception may play a protective role in acute injury through reflex muscular splinting [44]. The protective reflex arc initiated by mechanoreceptors and muscle spindle receptors occurs much more quickly than the reflex arc initiated by nociceptors (70–100 m/s vs. 1 m/s). Thus proprioception may play a more significant role than pain sensation in preventing injury in the acute setting. More importantly, proprioceptive deficits may play a more significant role in the etiology of chronic injuries and reinjury. Initial knee injury results in partial deafferentation and sensory deficits which can predispose to further injury [41]. Proprioceptive deficits may also contribute to the etiology of degenerative joint disease through pathological wearing of a joint with poor sensation. It is unclear whether the proprioceptive deficits that accompany degenerative joint disease are a result of the underlying pathological process or contribute to the etiology of the pathological process. It may be surmised then that the apparent loss of proprioception over a greater range of motion in the ACL deficient knee may help explain why the so-called isolated PCL deficient knee has a relatively more “benign course.” This study does not attempt to explain why TTDPM is reduced only at 45°, or why RPP appears to be better at 110°. There are many possible explanations, including the altered kinematics of the PCL deficient knee, variable coordinated input between the ACL and PCL at varying degrees of flexion (the ACL may have significant resultant force when flexed) [63], and even that the PCL still contributes proprioceptive information as it may heal in a lengthened position. Further still, the effect of physical therapy following injury may enhance proprioceptive function of the remaining mechanoreceptors. Further, it may be that kinesthesia and joint position sense are mediated through different pathways. We hypothesize that the altered proprioceptive input in the PCL-deficient knee may be due to proprioceptive function of the mechanoreceptors within the PCL, and that they may play a role in the clinical function of patients with PCL injuries. Areas of limitations and further study include the relative importance of control design, effect of gender, length of follow-up, and effect of physical therapy. We have chosen to use the contralateral knee as the control since proprioception has been shown to be affected by training, age, injury, and other factors. Some investigators have found lower proprioceptive capabilities in the noninjured knees of ACL-deficient patients than in age- and sexmatched controls. As understanding of proprioception in the human knee, and specifically cruciate ligaments, is in its infancy, the relative strengths and limitations of both methodologies is unclear. The number of patients in our study did not allow statistically or clinically significant evaluation of possible gender differences in proprioception following isolated PCL disruption, nor for a more meaningful understanding of the influence of time on proprioception following isolated PCL tears, although these are two of many issues that needs to be evaluated further in the future. Furthermore, due to the relatively small numbers of patients studied, a correlation of physical therapy with outcome and 316 proprioception could not be made. It does appear evident that a longer period from injury is correlated with enhanced proprioception (RPP only). This may be counterintuitive since the proposed natural history of isolated PCL rupture is to develop degenerative arthritis [18]. It is known that proprioception is reduced in the arthritic knee [2, 5, 58]. However, our investigation had only one patient studied nearly 20 years following PCL injury, and he had no radiographic evidence of arthritic change, as with all of the other patients (part of the exclusionary criteria). Thus the issue of the length of time after injury and the effects on proprioception are unclear. Further investigation with larger numbers of subjects with a greater range of time from injury to testing, and further follow-up of the subjects in this study to evaluate proprioceptive changes in the same individuals over time may help elucidate the effects of time on proprioception. Acknowledgements This work was presented at the Specialty Day Meeting of the American Orthopaedic Society for Sports Medicine and the 63rd Annual Meeting of the American Academy of Orthopaedic Surgeons, both in Atlanta, Georgia, in February 1996. The study was performed at the Neuromuscular Research Laboratory, Department of Orthopaedic Surgery, University of Pittsburgh, Pittsburgh, Penn., USA. References 1. Barrack RL, Skinner HB, Brunet ME, Cook SD (1983) Joint kinesthesia in the highly trained knee. J Sports Med Phys Fitness 24 : 18–20 2. Barrack RL, Skinner HB, Cook SD, Haddad RJ Jr (1983) Effect of articular disease and total knee arthroplasty on knee joint-position sense. J Neurophysiol 50 : 684–687 3. Barrack RL, Skinner HB, Buckley SL (1989) Joint proprioception in the anterior cruciate ligament deficient knee. Am J Sports Med 17 : 1–6 4. Barrett DS (1991) Proprioception and function after anterior cruciate reconstruction. J Bone Joint Surg Br 73 : 833–837 5. Barrett DS, Cobb AG, Bentley G (1991) Joint proprioception in normal, osteoarthritic and replaced knee. J Bone Joint Surg Br 73 : 53–56 6. Baxendale RA, Ferrell WR, Wood L (1988) Responses of quadriceps motor units to mechanical stimulation of knee joint receptors in the decerebrate cat. Brain Res 453 : 150–156 7. Beard DJ, Kyberd PJ, Fergusson CM, Dodd CAF (1993) Proprioception after rupture of the anterior cruciate ligament. An objective indication of the need for surgery? J Bone Joint Surg Br 75 : 311–315 8. Chick RR, Jackson RW (1978) Tears of the anterior cruciate ligament in young athletes. J Bone Joint Surg Am 60 : 970–973 9. Clancy WG Jr, Shelbourne KD, Zoellner GB, Keene JS, Reider B, Rosenberg TD (1983) Treatment of knee joint instability secondary to rupture of the posterior cruciate ligament. J Bone Joint Surg Am 65 : 310–322 10. Clark FJ, Burgess PR (1975) Slowly adapting receptors in cat knee joint: can they signal joint angle? J Neurophysiol 38 : 1448–1463 11. Clark P, MacDonald PB, Sutherland K (1996) Analysis of proprioception in the posterior cruciate ligament-deficient knee. Knee Surg Sports Traumatol Arthrosc 4 : 225–227 12. Co FH, Skinner HB, Cannon WD (1991) Proprioception of the knee following ACL reconstruction and its relation to heelstrike impulse. Transactions of the 37th Annual Meeting of the Orthopaedic Research Society, Anaheim, p 603 13. Corrigan JP, Cahsman WF, Brady MP (1992) Proprioception in the cruciate deficient knee. J Bone Joint Surg Br 74 : 247–250 14. Cross MM, McCloskey DI (1973) Position sense following surgical removal of joints in man. Brain 55 : 443–445 15. Cross MJ, Powell JF (1984) Long term follow-up of posterior cruciate ligament rupture. A study of 116 cases. Am J Sports Med 12 : 292–297 16. Dandy DJ, Pusey RJ (1982) The longterm results of unrepaired tears of the posterior cruciate ligament. J Bone Joint Surg Br 64 : 92–94 17. DeAvila GA, O’Connor BL, Visco DM, Sisk TD (1989) The mechanoreceptor innervation of the human fibular collateral ligament. J Anat 162 : 1–7 18. Dejour H, Walch G, Peyrot J, Eberhard P (1988) The natural history of rupture of the posterior cruciate ligament. Fr J Orthop Surg 2 : 112–120 19. DuToit GT (1967) Knee joint cruciate ligament substitution. The Lindemann (Heidelberg) operation. S Afr J Surg 5 : 25–30 20. Ekholm J, Eklund G, Skoglund S (1960) On the reflex effects from the knee joint of the cat. Acta Physiol Scand 50 : 167–174 21. Ferrell WR (1985) The response of slowly adapting mechanoreceptors in the cat knee joint to tetanic contraction of hind limb muscles. Q J Exp Physiol 70 : 337–345 22. Fowler PJ, Messieh SS (1987) Isolated posterior cruciate ligament injuries in athletes. Am J Sports Med 15 : 553– 557 23. Giove TP, Miller SJ III, Kent BE, Sanford TL, Garrick JG (1983) Nonoperative treatment of the torn anterior cruciate ligament. J Bone Joint Surg Am 65 : 184–192 24. Goodwin GM, McCloskey DI, Matthews PB (1972) The persistence of appreciable kinesthesia after paralyzing joint afferents but preserving muscle afferents. Brain Res 37 : 326– 329 25. Goodwin GM, McCloskey DI, Matthews PB (1972) The contribution of muscle afferents to kinesthesia shown by vibration induced illusions of movement and by the effects of paralyzing joint afferents. Brain 95 : 705– 748 26. Grigg P (1975) Mechanical factors influencing response of joint afferent neurons from cat knee. J Neurophysiol 38 : 1473–1484 27. Grigg P (1976) Response of joint afferent neurons in cat medial articular nerve to active and passive movements of the knee. Brain Res 118 : 482–485 28. Grigg P, Hoffman AH (1984) Ruffini mechanoreceptors in isolated joint capsule: responses correlated with strain energy density. Somatosens Res 2 : 149–162 29. Grigg P, Hoffman AH (1989) Calibrating joint capsule mechanoreceptors as in vivo soft tissue load cells. J Biomech 22 : 781–785 30. Halata Z, Haus J (1989) The ultrastructure of sensory nerve endings in human anterior cruciate ligament. Anat Embryol 179 : 415–421 31. Hawkins RJ, Misamore GW, Merritt TR (1986) Follow up of the acute nonoperated isolated anterior cruciate ligament tear. Am J Sports Med 14 : 205– 210 317 32. Heetderks WJ (1978) Principal component analysis of neural population responses of knee joint proprioceptor in cat. Brain Res 156 : 51–65 33. Ihara H, Nakayama A (1986) Dynamic joint control training for knee ligament injuries. Am J Sports Med 14 : 309–315 34. Insall J, Joseph DM, Aglietti P, Campbell RD (1981) Bone-block iliotibialband transfer for anterior cruciate insufficiency. J Bone Joint Surg Am 63 : 560–569 35. Johansson H, Sjölander P, Sojka P (1990) Activity in receptor afferents from the anterior cruciate ligament evokes reflex effects on fusimotor neurones. Neurosci Res 8 : 54–59 36. Johansson H, Sjölander P, Sojka P (1991) Receptors in the knee joint ligaments and their role in the biomechanics of the joint. Crit Rev Biomed Eng 18 : 341–368 37. Johansson H, Sjölander P, Sojka P (1991) A sensory role for the cruciate ligaments. Clin Orthop 268 : 161–178 38. Katonis PG, Assimakokpoulos AP, Agapitos MV, Exarchou EI (1991) Mechanoreceptors in the posterior cruciate ligament. Histologic study on cadaver knees. Acta Orthop Scand 62 : 276–278 39. Keller PM, Shelbourne D, McCarroll JR, Rettig AC (1993) Nonoperatively treated isolated posterior cruciate ligament injuries. Am J Sports Med 21 : 132–136 40. Kennedy JC, Weinberg HW, Wilson AS (1974) The anatomy and function of the anterior cruciate ligament as determined by clinical and morphological studies. J Bone Joint Surg Am 56 : 223–235 41. Kennedy JC, Alexander IJ, Hayes KC (1982) Nerve supply of the human knee and its functional importance. Am J Sports Med 10 : 329–335 42. Krauspe R, Schmidt M, Schaible HG (1992) Sensory innervation of the anterior cruciate ligament. An electrophysiological study of the response of single identified mechanoreceptors in the cat. J Bone Joint Surg Am 74 : 390–397 43. Krenn V, Hofmann S, Engel A (1990) First description of mechanoreceptors in the corpus adiposum infrapatellare of man. Acta Anat 137 : 187–188 View publication stats 44. Lephart SM, Kocher MS, Fu FH, Borsa PA, Harner CD (1992) Proprioception following ACL reconstruction. J Sports Rehabil 1 : 188–196 45. MacKenzie WG, Shim SS, Day B, Leung G (1985) The blood and nerve supply of the knee meniscus in man. Anat Rec 211 : 115A-116A 46. McDaniel WJ, Dameron TB Jr (1980) The untreated anterior cruciate ligament rupture. Clin Orthop 172 : 158– 163 47. Noyes FR, Butler DL, Paulos LE, Grood ES (1983) Intra-articular cruciate reconstruction. I perspectives on graft strength, vascularization, and immediate motion after replacement. Clin Orthop 172 : 71–77 48. Noyes FR, Mooar PA, Matthews DA, Butler DL (1983) The symptomatic anterior cruciate ligament deficient knee. I the long-term functional disability in athletically active individuals. J Bone Joint Surg Am 65 : 154–162 49. O’Connor BL (1976) The histologic structure of dog knee menisci with comment on its possible significance. Am J Anat 147 : 407–417 50. O’Connor BL, McConnaughey JS (1978) The structure and innervation of cat knee menisci and their relation to a ‘sensory hypothesis’ of meniscal function. Am J Anat 153 : 431–442 51. Parolie JM, Bergfeld JA (1986) Longterm results of nonoperative treatment of isolated posterior cruciate ligament injuries in the athlete. Am J Sports Med 14 : 35–38 52. Rivard CH, Yahia LH, Rhalmi S, Newman N (1993) Immunohistochemical demonstration of sensory nerve fibers and endings in human anterior cruciate ligaments. Transactions of the 39th Annual meeting of the Orthopaedic Research Society, San Francisco, CA, p 61 53. Safran MR, Caldwell GL Jr, Fu FH (1994) Proprioception Considerations in Surgery. J Sports Rehab 1 : 105–115 54. Satku K, Kumar VP, Ngoi SS (1986) Anterior cruciate ligament injuries. To counsel or to operate? J Bone Joint Surg Br 68 : 458–461 55. Schultz RA, Miller DC, Kerr CS, Micheli L (1984) Mechanoreceptors in human cruciate ligaments: a histologic study. J Bone Joint Surg Am 66 : 1072– 1076 56. Schutte MJ, Dabezies EJ, Zimny ML, Happel LT (1987) Neural anatomy of the human anterior cruciate ligament. J Bone Joint Surg Am 69 : 243–247 57. Sherman MF, Warren RF, Marshall JL, Savatsky GJ (1988) A clinical and radiographical analysis of 127 anterior cruciate insufficient knees. Clin Orthop 227 : 229–237 58. Skinner HB, Barrack RL, Cook SD (1984) Age-related decline in proprioception. Clin Orthop 184 : 208–211 59. Skinner HB, Barrack RL, Cook SD, Haddad RJ Jr (1984) Joint position sense in total knee arthroplasty. J Orthop Res 1 : 276–283 60. Sojka P, Sjolander P, Johansson H, Djupsjobacka M (1991) Influence from stretch-sensitive receptors in the collateral ligaments of the knee joint on the gamma-muscle spindle systems of flexor and extensor muscles. Neurosci Res 11 : 55–62 61. Sommerlath K, Lysholm J, Gillquist J (1991) The long-term course after treatment of acute anterior cruciate ligament ruptures. Am J Sports Med 19 : 156–162 62. Tibone JE, Antich TJ, Perry J, Moynes D (1988) Functional analysis of untreated and reconstructed posterior cruciate ligament injuries. Am J Sports Med 16 : 217–223 63. Torg JS, Barton TM, Pavlov H, Stine R (1989) Natural history of the posterior cruciate ligament-deficient knee. Clin Orthop 246 : 208–216 64. Warren PJ, Olanlokun TK, Cobb AG, Bentley G (1993) Proprioception after knee arthroplasty. The influence of prosthetic design. Clin Orthop 297 : 182–187 65. Wascher DC, Markolf KL, Shapiro MS, Finerman GAM (1993) Direct in vitro measurement of forces in the cruciate ligaments. I. The effects of multiplane loading in the intact knee. J Bone Joint Surg Am 75 : 377–386 66. Williams WJ (1981) A systems-oriented evaluation of the role of joint receptors and other afferents in position and motion sense. Crit Rev Biomed Eng 7 : 23–77 67. Wyke BD (1967) The neurology of joints. Ann Royal Coll Surg 41 : 25–49 68. Zimny ML, Albright DJ, Dabezies E (1988) Mechanoreceptors in the human medial meniscus. Acta Anat 133 : 35– 40