|Home | About | Journals | Submit | Contact Us | Français|
This study investigated the internal fluid pressure of human cadaver meniscal root attachments. A pressure micro-sensor was implanted inside each attachment site. Tibiofemoral joints were compressed to 2x body weight at various flexion angles and pressure recorded for 20 minutes. The anterior cruciate ligament (ACL) was then transected and joints retested. Lastly, a longitudinal incision of the lateral posterior horn was made and the joint retested. Ramp pressure was defined as the pressure when 2x body weight was reached, and equilibrium pressure was recorded at the end of the hold period. The medial posterior attachment was subjected to greater ramp pressure than the medial anterior (p=0.002) and greater equilibrium pressure than all other root attachment sites (p<0.001). Flexion angle had a significant effect on pressure as full extension was greatest at ramp (p=0.040). Transection of the ACL decreased ramp pressure in the lateral posterior attachment (p=0.025) and increased equilibrium pressure (p=0.031) in the medial posterior attachment. The results suggest that repair strategies should be developed which reconstruct the medial posterior attachments to be sufficient to withstand large pressures. Furthermore, since meniscal pressure is highest at full extension, this fact should be considered when prescribing rehabilitation following repair of an attachment.
Knee menisci are semi-lunar, fibrocartilaginous structures that transduce applied compressive loads to circumferential hoop stresses which are attenuated at the tibial plateau via the meniscal root attachments1. These specialized interfaces, located at the lateral anterior (LA), lateral posterior (LP), medial anterior (MA), and medial posterior (MP) horns, are crucial to maintaining mechanical functionality of menisci, thereby preventing osteoarthritis (OA)1–6. Similar to other insertion tissue, the ligament-like root attachments link the meniscus main body to the subchondral bone utilizing a blend of collagen fibers and proteoglycans which assuage complex stress concentrations due to proposed tensile, compressive, and shear loading7. Mechanical evaluation of the root attachments elucidates distinct differences in their material properties8, with the anterior possessing a significantly greater linear elastic modulus than the posterior along the main fiber axis9. This is a presumed adaptation to the distinct mechanical environment that each root attachment site must endure. Magnetic resonance imaging (MRI) of meniscal translation and articular surface contact supports this hypothesis as the anterior root attachment sites demonstrate greater mobility and the posterior compartment exhibiting direct contact during deep knee flexion10. Histological investigation reveals significant differences in the amount of fibrocartilage at the root attachment sites, a constituent directly linked to the amount of compression applied to insertion tissues11,12. Additionally, the proteoglycans present at the meniscus to bone interface aid in resisting the applied stresses by retaining interstitial fluid12. Seitz et al. have recursively determined the forces acting on the anterior root attachment sites during loading revealing very low loads for physiological trajectories13. The posterior root attachment sites, however, are more clinically relevant as they are believed to endure higher forces and are known to be less mobile10,14. Specifically, medial posterior horn tears and root avulsion are most often reported in the literature and are believed to be of greater functional significance15. Thus, it is imperative to determine the mechanical environment of all root attachment sites and how they work in conjunction with the meniscus to support load bearing activities. These data will provide insight into the physical demands that need to be considered when performing meniscal repair and proposing rehabilitation techniques.
In addition to ascertaining the native mechanics, changes in joint biomechanics due to injury, including anterior cruciate ligament (ACL) and meniscal tearing, are of particular interest. Risk of OA is significantly increased following traumatic injury and is due in part to changes in joint mechanics16,17. Specifically, ACL rupture results in posterior shifts in the medial compartment for articular cartilage tibiofemoral contact points, increased contact area, and decreased pressure17–19. Conversely, meniscal lesions have resulted in a significant decrease in tibiofemoral contact area and increase in peak pressure20.
Determining physiological root attachment mechanics is complicated due to tight spatial constraints when considering transducer fixation, particularly in the posterior compartments13. Recent development of a novel, custom-made (Luna Innovations, Lunainc.com) minimally invasive, pressure micro-sensor (250µm in diameter) facilitates direct observation of fluid pressure within biological tissues, with implications for the surrounding stress environment21–23. These sensors are implanted using a 21-ga syringe needle and remain rooted to the tissue via a small barb on the outer surface. After insertion the needle is then retracted from the tissue. The latest iteration of these sensors exhibits relative immunity to temperature, electromagnetic-interference, and corrosive environments. Their dynamic operating range is in excess of 250mmHg, sampled at 960Hz, with accuracy of 2% of full-scale output and repeatability and hysteresis better than 1%24–26.
Based on clinical and laboratory evidence it is hypothesized that 1) the fluid pressure within the posterior root attachment sites would be highest during loading, 2) fluid pressure will decrease in the anterior and increase in posterior sites with greater flexion angles and 3) ACL rupture and disruption of meniscal integrity will alter fluid pressure within the root attachment sites.
Six human cadaver knees (ages: 41–61, average age 55) were acquired from the Mayo Clinic donor program with institutional review board approval. Knees selected had no history of joint surgery nor showed any apparent signs of articular cartilage thinning based on radiographic examination by an orthopaedic surgeon. Specimens were dissected such that the patellar tendon, patella, and a portion of the quadriceps as well as the structural ligaments and most of the synovial sac were left intact. A threaded rod was bonded inside the medullary cavity of the femur and the tibia was potted using fiber strand compound (Bondo-Glass, 3M, St. Paul, MN). The joints were then mounted in a uni-axial material tester (Model 8872, Instron Corporation, Canton, MA) with custom made fixtures that fix the position of the tibia while allowing free rotation of the femur about all three axes (Figure 1). Load was applied to the tibiofemoral joint through the use of a custom fixture that was attached to the threaded rod and had a flared shoulder that contacted the cut surface of the femur. The pressure microsensors were then inserted into the meniscal root attachments using a 21.5-ga needle. Sensor location was determined by visually approximating the middle of the root attachment in the transverse plane and then inserting the needle immediately adjacent to the meniscal horn (Figure 2). The sensors were inserted by making small incisions in the proximal portion of the synovial sac (near the femoral condyles and above the menisci) in order to gain physical and visual access to the joint space. There was no obvious loss of synovial fluid during this process nor during loading of the joint. The soft tissues occluding any of the insertion sites were temporarily pushed aside using forceps in order to perform sensor placement. The sensors monitored the relative change in pressure during loading.
Sensor calibration is described in detail by Cottler et al. 200925. In brief, the sensor is placed in a sealed water filled pressure chamber that is in series with pressurized air pressure and a digitally controlled hydraulic pressure regulation system (Ruska Model # 7250, GE Sensing, Houston, TX). This system then steps through the repeated cycles of pressure from 0–250 mmHg in 25 mmHg increments. The resultant sensor output is then calibrated by fitting the known pressure-time history using a fourth order polynomial function. Sensor performance is then validated using a custom-built pressure chamber mounted to an electromechanical material test stand (ElectroForce 3200, Bose Corp, Eden Prairie, MN). The sensor is then subjected to a 0.02 Hz sawtooth wave that increases in a linear fashion at 8.8 mmHg/s from ambient to 250 mmHg and then decreased back to ambient. This cycle is repeated five times. The pressure chamber was validated using a NIST traceable 500 mmHg pressure sensor (Sensotec FPG gage pressure transducer, Sensotec, Colombus, OH). Results from this technique have shown that the sensors exhibit accuracy of 1.79±0.90, repeatability of 0.48±0.29, and hysteresis of 0.60±0.18 (reported as percentage of full-scale output).
Knees were tested at five flexion angles (0 °, 15 °, *15 °, 30 °, 45 °). These flexion angles were selected because they represent the range of knee motion during the weight-bearing phases of gait 27. Three of the angles positioned the knee with the simulated ankle behind the hip (15°, 30°, 45°), one directly in line (0 – full extension), and one with the ankle in front of the hip (*15°) (Figure 1). An electric actuator (Model EC2, Danaher Motion, Radford, VA) applied a pre-load of 45 N to the quadriceps complex of each knee. A variable joint compressive load was applied to each knee based on the body weight of the donor. Knee joints were consistently compressed to two times body weight using the servohydraulic material test stand. The load was ramped up over 100 sec and then held constant for 20 minutes while pressure was recorded continuously throughout the entire loading history. This time frame was chosen from pilot testing which showed less than a 1% change in pressure over time. After loading, the joint was unloaded for 20 minutes, and then tested under load at another flexion angle. Upon completion of all joint angle positions for the intact joint, the anterior cruciate ligament was transected (ACLT) and the test cycle was repeated. Lastly, a longitudinal incision was cut into the lateral posterior meniscus and the test sequence was again repeated (ACLT+cut). This type of lesion was performed due to the significantly higher rate of combined ACL-lateral meniscus injury in alpine skiers 28,29. The ACL transection and longitudinal incision were achieved using a scalpel blade. The ACL was transected in the middle of the ligament. The meniscal incision was performed by placing the knee at full extension then guiding the flat side of the blade along the femoral condyle and then straight down into the meniscus main body. This placed a lesion in the red-white zone of the meniscus. Knees were wrapped in saline soaked gauze to keep tissues moist during testing. Flexion angle was randomized to eliminate any potential test sequence effects.
In order to compare loading angles and conditions, specific points of interest were identified from the pressure-time histories: pressure at the start of holding cycle (“ramp”), after 20 minutes of constant load (“equilibrium”). Data were analyzed using a three-factor ANOVA with repeated measures (root attachment site, physical condition, and flexion angle). Post-hoc analysis using Tukey’s method was used to identify explicit differences, p < 0.05 was considered statistically significant. Datasets were checked for Gaussian distribution using a Shapiro-Wilk test for normality. All statistics were performed using commercial statistical software (Minitab Inc., State College, PA). The sample size chosen was based on statistical power analysis of initial pilot data (n = 3). The study by Richards et al. Orthopedics 2008 was used for determining the expected change in pressure due to flexion angle where they reported changes up to 100 mmHg (13.3 kPa) for a change in 15° of flexion30. Statistical power was set at β = 0.75.
The fiber optic micro-sensors recorded changes in pressure over time in meniscal root attachment tissue (Figure 3). Pressure-time series data demonstrated consistent, though uniquely different, patterns for the root attachment sites. Medial anterior and posterior root attachment sites exhibited little change or an increase in fluid pressure over time with one exception at 30°. The LA root attachment exhibited little change or decreased with time (Figure 4). The LP root attachment consistently decreased in fluid pressure over time.
Examining fluid pressure for each root attachment site the MP was significantly greater than the MA at ramp pressure (p = 0.002). At equilibrium, the MP root attachment had significantly greater pressure than all other root attachment sites (p < 0.001). When considering joint flexion angle, the fluid ramp pressure at 0° of flexion was significantly greater than the *15°, 30°, and 45° angles (p = 0.040). This effect was reduced at equilibrium as zero is only greater than 45° at this time point (p = 0.028, Figure 5).
Ramp and equilibrium pressures were affected considerably following ACLT, specifically at 0° flexion. Post-hoc analysis revealed a significant difference between the healthy and the ACLT and ACLT+cut in the LP root attachment for ramp pressure (p = 0.025) and MP root attachment for equilibrium (p = 0.031) (Figure 5).
This work reports, for the first time, direct measurement of the relative change in internal pressure in human meniscal root attachments due to loading. Overall, the root attachment sites endured the greatest pressure at full extension. Specifically, the medial posterior root attachment was subjected to the highest amount of pressure during loading. Lastly, transection of the ACL resulted in an increase in pressure in the medial posterior and decrease in the lateral anterior. Given the large pressures experienced by the posterior root attachments, it is not surprising that the medial posterior root attachment is the most likely to fail and care should be taken to ensure that repair strategies are developed to ensure reconstruction of these root attachments is sufficient to withstand the large pressures15,31. Furthermore, since overall pressure is highest at full extension, this fact may be an important consideration when prescribing rehabilitation following repair of a root attachment32.
Our findings correlate well with findings that the medial meniscus body and horn are believed to be subjected to greater forces during loading10,14. Seitz et al. demonstrated that the anterior sites are subjected to very low forces, approximately 10–15N13; accordingly our data also show relatively small changes in fluid pressure for these sites. With the micro-sensor employed here we were able to directly monitor and compare the pressure between all root attachment sites and affirm that while the anterior sites develop relatively low internal pressure the medial posterior site is subjected to significantly greater pressure. Hauch et al. determined the human medial posterior site to possess significantly lower failure properties and linear modulus9. Recently, Johannsen et al. reported that the medial posterior root attachment has a smaller insertion footprint than the lateral posterior31. Coupling these findings with our results corresponds with clinical findings that medial posterior root injuries are most common and represent a major disruption in effective knee biomechanics15.
Unexpectedly, the pressure was often highest at 0° of flexion. This may be due to the load concentrated at the cartilage-cartilage contact area forcing interstitial fluid radially outwards, consolidating in the root attachment site compartments. Alternatively or additionally, at 0° flexion, the root attachments may be loaded the most to keep the joint load transferred through the meniscus and off the underlying articular cartilage33. Large tensile loads in conjunction with Poisson’s effect could in turn reduce cross sectional area and increase fluid pressure. In contrast, Seitz et al. and Stärke et al. reported no difference in anterior forces with changes in flexion angle 13,34. Differences between these studies and the current may be due to patellofemoral contact35.
While care was taken to ensure accurate and translatable results there are several limitations of this study that must be identified. This study aimed at recreating conditions that are physiologically relevant; however, static loading conditions were examined rather than dynamic cycling. Dynamic systems have proven vital in studying changes in joint kinematics; however, few achieve physiologically relevant joint loads. We chose to focus on achieving loads that approached physiological levels to obtain meaningful results within the root attachment sites36,37, particularly considering the low loads for the anterior sites as determined by Seitz et al.13. Additionally, ramp loading speeds were kept relatively slow to prevent dislodging the sensors from the tissue. The relatively stiff/tough collagen fibrils of the meniscal root attachments do not provide fixation as secure as muscle tissue, for which these sensors were originally developed. Also, sensor fixation was performed visually and may not have resulted in perfectly reproduced placement between specimens. This may have influenced the results as previous testing has shown an inhomogeneous distribution in material properties between the insertion mid-substance and the bony root9. However, no significant differences were found when comparing material properties throughout the outer-middle-inner portions within each root attachment site9. While the absolute pressure may have been impacted by the visual placement method, care was taken to ensure consistent placement of sensors so that the conclusions drawn in this study were not influenced by sensor placement. It should be noted that our findings agree with previous results that infer relatively higher in vivo loads of the medial posterior root attachment, when compared to the other root attachment sites, based on imaging and mechanical testing modalities9,10,15. Work continues to refine these sensors for a wide array of laboratory and clinical applications25. Another limitation is the implementation of only a single muscle group, the quadriceps. Clearly, inclusion of additional muscle groups can alter the contact mechanics; however, we opted to include the quadriceps because it aided in joint alignment during setup. Furthermore, inclusion of this muscle group maintained patellofemoral contact which helped to retain interstitial fluid. A similar limitation is that the loading apparatus most likely gave a 50/50 load distribution, whereas the in-vivo load distribution is more likely 60/4038. This difference in loading would have accentuated but not changed the results obtained in this study because the greatest pressure changes were observed in the medial root attachments.
In summary, this study has demonstrated for the first time that the mechanical environment, as indicated by fluid pressure, of the meniscal root attachments during loading are unique between sites and can change significantly with flexion angle and loss of ACL integrity. This information can be used as means to validate whole joint mechanical models, improving understanding of the anatomical role of each root attachment site, and changes within each that occur with injury. Clinically, the individual root attachment sites should be considered distinct organs and may require site-specific repair strategies and rehabilitation prescriptions. Future work should leverage these sensors in observing changes due to different injury models and reparative techniques.
This work was supported by grants from the National Institutes of Health (R01 HD31476, R15 253 ARG051906, and F31 AG039975).
The authors have no conflicts of interest to disclose.