Effects of Anterior Cruciate Ligament Deficiency on Tibiofemoral Cartilage Thickness and Strains in Response to Hopping

Abstract

Introduction:

Changes in knee kinematics after ACL injury may alter loading of the cartilage and thus affect its homeostasis, potentially leading to the development of post-traumatic osteoarthritis (OA). However, there are limited in vivo data to characterize local changes in cartilage thickness and strain in response to dynamic activity in patients with ACL deficiency.

Hypothesis/Purpose:

To compare in vivo tibiofemoral cartilage thickness and cartilage strain resulting from dynamic activity between ACL-deficient and intact contralateral knees. We hypothesized that ACL-deficient knees would show localized reductions in cartilage thickness and elevated cartilage strains.

Study Design:

Controlled laboratory study.

Methods:

Magnetic resonance images were obtained before and after single legged hopping on injured and uninjured knees in 8 patients with unilateral ACL rupture. Three-dimensional models of the bones and articular surfaces were created from both the pre- and post-activity scans. The pre- and post-activity models were registered to each other, and cartilage strain (defined as the normalized difference in cartilage thickness pre- and post-activity), was calculated in regions across the tibial plateau, femoral condyles, and femoral cartilage adjacent to the medial intercondylar notch. These measurements were compared between ACL-deficient and intact knees. Differences in cartilage thickness and strain between knees were tested using multiple ANOVA models with alpha set at p<0.05.

Results:

Compressive strain in the intercondylar notch was elevated in the ACL-deficient knee relative to the uninjured knee. Furthermore, cartilage in the intercondylar notch and adjacent medial tibia was significantly thinner prior to activity in the ACL-deficient knee compared to the intact knee. In these two regions, thinning was significantly influenced by time since injury, with patients with more chronic ACL deficiency (>1 year since injury) experiencing greater thinning.

Conclusion:

In patients with ACL deficiency, the medial femoral condyle adjacent to the intercondylar notch in the ACL deficient knee exhibited elevated cartilage strain and loss of cartilage thickness, particularly with longer time from injury. We hypothesize that these changes may be related to post-traumatic OA development.

INTRODUCTION

Anterior cruciate ligament (ACL) rupture is a common knee injury⁸ that often results in altered joint kinematics.^{2, 7, 12, 13, 18, 19, 28} It is believed that these altered kinematics lead to abnormal cartilage loading, which may subsequently affect cartilage homeostasis^{21, 22} and play a role in the development of post-traumatic osteoarthritis (OA).^{2, 3, 6, 20} More than 50% of ACL-deficient patients exhibit radiographic signs of OA 10–15 years following injury.^{31, 47} Furthermore, the onset of post-traumatic OA in individuals with an ACL injury occurs 15–20 years earlier than in those who have never had an ACL injury.³⁵ However, there are limited data exploring how altered knee kinematics affect the in vivo cartilage mechanical environment under dynamic loading conditions, and how these changes may lead to the development of OA.

Magnetic resonance imaging (MRI) can be used to measure cartilage thickness and compressive strain resulting from physiologic demands on the knee. This can be accomplished by comparing cartilage thickness distributions before and after an activity that loads the cartilage. Due to the biphasic nature of cartilage, water that is exuded during loading re-enters the cartilage matrix in a time dependent manner upon removal of the mechanical load.^{14, 32} Thus, the changes in cartilage thickness between pre- and post-activity MR imaging represent the cartilage deformation induced by loading.^{15, 16, 48} Previously, this principle has been utilized to obtain localized measurements of cartilage strain in response to single-legged hopping in healthy, uninjured individuals.⁴⁰

While previous studies have characterized in vivo cartilage strains in healthy knees,^{9, 29, 40, 48} there are limited data on in vivo tibiofemoral cartilage strains in ACL-deficient knees,^{26, 46} particularly in response to dynamic activities of daily living. Therefore, the purpose of this study was to utilize MR imaging to determine how ACL deficiency affects localized cartilage thickness and strain resulting from single-legged hopping. We hypothesized that increased cartilage strains and decreased cartilage thickness and would be observed in ACL-deficient knees as compared to the intact knees. Specifically, we expected to observe changes within the medial femoral cartilage adjacent to the intercondylar notch of ACL-deficient knees, an area previously shown to be susceptible to cartilage damage and degeneration in patients with ACL injuries.^{17, 33}

METHODS

Following Institutional Review Board approval, eight male patients (mean age: 31 years, range: 21 to 47 years; mean body mass index (BMI): 25.6 kg/m²; range: 21.7 to 34.7 kg/m²) with a complete unilateral ACL injury were recruited into this study. Complete unilateral ACL tears were confirmed by clinical examination and MR imaging. Subjects reported dates of injury ranging from 24 days to 13 years and 11 months prior to testing. One patient could not recall the exact time of injury but reported the year of injury, meaning that this individual’s injury occurred between 14 to 25 months prior to testing. Patients were classified into two groups based on time from injury. Four patients were categorized as short time from injury (time from injury less than 2 months, ranging from 24 days to 2 months) and four patients as long time from injury (time from injury greater than 1 year, ranging from 1 year to 13 years and 11 months). None of the patients had undergone an ACL reconstruction at the time of this study, or had a history of injury or surgery to the intact contralateral knee. Subjects had no history of other injury or surgery in their ACL deficient knee. Each subject completed the International Knee Documentation Committee’s (IKDC) questionnaire.^{10, 27}

The protocol used in this study has been described in earlier studies.^{34, 40} To reduce the effects of diurnal cartilage loading,^{9, 48} all study procedures began early in the morning. Patients were asked to refrain from strenuous activity the night before and the morning of their participation. Upon arrival to the MR scanner, patients laid supine for 45 minutes in an adjacent room with their knees relaxed and extended to allow their cartilage to equilibrate.¹⁴ To avoid preloading of the cartilage, patients were transported to the MR scanner in a wheelchair. Pre-activity MR images of both knees were obtained using a 3T scanner (Trio Tim, Siemens) with an 8-channel knee coil. Patients rested in a supine position with their knees relaxed and extended during imaging. Sagittal plane images (resolution: 0.3×0.3×1 mm) were generated using a double-echo steady state (DESS) sequence (flip angle: 25°, repetition time: 17 ms, echo time: 6 ms).^{41, 42, 45} The duration of each MR scan was approximately 9 minutes.

Following pre-activity imaging, subjects were transported via wheelchair to the hall adjacent to the MR scanner. Each subject performed 60 single-legged hops, each 0.6 m in length, while a researcher walked alongside the subject. During hopping, the contralateral knee remained in a flexed position. Patients were instructed not to touch the ground or bear any weight on the contralateral leg. If necessary, the subject could briefly touch the wall or the researcher to maintain balance, but not for support. After completing the hopping activity, patients were quickly seated in a wheelchair and immediately transported into the MR suite to undergo a post-activity MR scan. The average time from completion of the hopping activity to the start of the second DESS scan was 3 minutes and 48 seconds. After completing the hopping task and the scan, the subject repeated the identical hopping and scanning sequence using the contralateral leg. Half of the subjects had their ACL deficient knee tested first, and half had their intact knee tested first.

Prior to analysis, DESS scans were reviewed by a musculoskeletal radiologist with over 30 years of experience to confirm knee status. No definitive meniscus tears were identified in either knee in these subjects. While occasional fissures were found on the medial and lateral compartments of the tibia, no focal partial or full thickness chondral defects were detected.

The DESS images were imported into solid modeling software (Rhinoceros, Robert McNeel and Associates, Seattle, Washington) and the outer margins of the bone and cartilage surfaces of the femur and tibia were manually segmented in each image (Figure 1a). These curves were stacked to create wireframe models (Figure 1b), which were then converted into three-dimensional surface mesh models (Geomagic Studio; Morrisville, NC) (Figure 1c). Previous validation and repeatability studies have demonstrated that this technique is able to detect differences in cartilage thickness to within 1%.^{9, 46} Furthermore, investigators were blinded as to whether the knee was ACL deficient or not.

Figure 1.

MR images were used to generate three-dimensional models of the femur, tibia, and articular cartilage. A) Bony and articular surfaces were manually segmented on each sagittal 3T MR slice, B) stacked to form a wireframe model, and C) converted into three dimensional surface mesh models.

Cartilage thickness was calculated as the distance between each cartilage surface mesh node and the closest mesh node on the corresponding bony surface. In this manner, cartilage thickness maps were generated for each knee (Figure 2). The bony surfaces of the pre-activity and post-activity models were then aligned using an iterative closest point technique (Geomagic Studio), thereby allowing for comparisons of pre- and post-activity cartilage thickness within corresponding regions. Cartilage thickness was then averaged within circular regions distributed evenly across the cartilage surfaces, each with a diameter of 2.5 mm (Figure 3). Cartilage thickness was obtained from 18 regions across the tibial surface, 9 on each plateau, and from 36 points on the femoral surface, 18 in each compartment. Additionally, cartilage thickness was sampled on a region of cartilage on the medial femoral condyle adjacent to the intercondylar notch, a region previously shown to be susceptible to early cartilage degeneration following ACL injury.^{17, 33} Strain within each sampling region was defined as the normalized change in thickness pre- versus post-activity. Overall compartmental strains were defined as the average of the strains of the sampled regions across each compartment.²⁹

Figure 2.

Thickness maps of pre-activity and post-activity tibial and femoral cartilage were generated for each subject. Thicker cartilage is red, and thinner cartilage is blue.

Figure 3.

A series of points was generated for measuring cartilage thickness. Registration of each subject’s ACL deficient and intact knees before and after exercise allowed for site-specific comparisons between models. Nine points in each compartment of the tibial plateau, 18 points on each femoral condyle, and 3 points on the medial femoral condyle adjacent to the femoral notch were placed to create tibial and femoral strain grids.

Descriptive statistics were used to characterize cartilage thickness in each region before and after hopping. A two-way repeated measures analysis of variance ANOVA with knee state (intact versus deficient) and location (medial femur, lateral femur, medial tibia, lateral tibia, and medial femur adjacent to intercondylar notch) was used to detect differences in strain, where time from injury was a categorical factor (long versus short, defined as greater than one year or less than 8 weeks from injury). Three-way repeated measures ANOVA with knee state (intact versus deficient), location (medial tibia, lateral tibia, medial femur, lateral femur, medial femoral condyle adjacent to the intercondylar notch), and exercise (before versus after exercise) was used to detect differences in cartilage thickness, where time from injury was a categorical factor (long versus short). Finally, side-to-side change in thickness, expressed as a percentage (ACL-deficient – intact normalized to intact thickness), was compared across regions using a repeated measures ANOVA with time from injury as a categorical factor. Fisher’s Least Significant Differences post-hoc test was used to detect differences between means as appropriate. Significance level (alpha) was set at p<0.05 for all analyses. All statistical analyses were performed using Statistica (StatSoft, Inc.; Tulsa, OK).

RESULTS

Cartilage thickness decreased after exercise (corresponding to compressive strain, Figure 4) in all regions except for the femoral condyle adjacent to the intercondylar notch in the intact knee. Strain in the ACL-deficient knee in the medial femoral condyle adjacent to the intercondylar notch (6%) was significantly increased compared the intact knee. No statistically significant differences in strain were observed between intact and deficient knees in the medial tibia, lateral tibia, medial femur, or lateral femur, with strains in these regions ranging from 2% to 4% in both knees.

Figure 4.

Regional compressive strain (mean % ± 95% confidence intervals) in intact and ACL deficient knees. Positive strain is indicative of compressive strain. Activity-related compressive strains were detected in both knees in the medial and lateral compartments of the tibia and femur. No statistically significant differences in compressive between the intact and deficient knees were observed in these regions. No statistically significant compressive strain was detected in the femoral condyle adjacent to the intercondylar notch in the intact knee, and this region experienced significantly more compressive strain in the ACL deficient knee as compared to the intact knee. Error bars represent the 95% confidence interval. * denotes significance p < 0.05.

Prior to hopping, statistically significant decreases in cartilage thickness were detected between ACL-deficient and intact knees in the medial tibial plateau (Figure 5) and medial femoral condyle adjacent to the intercondylar notch. No statistically significant differences between cartilage thickness in the intact and ACL-deficient knees were observed in the lateral tibial plateau, medial femoral condyle, or lateral femoral condyle.

Figure 5.

Mean pre-activity regional cartilage thickness (± 95% confidence intervals) in intact and ACL deficient knees. Cartilage in the ACL deficient knee was significantly thinner than in the intact knee in the medial tibial plateau and in the medial femoral notch adjacent to the tibial spine. * denotes significance p < 0.05.

For the medial femoral intercondylar notch and adjacent medial tibial plateau, there was a statistically significant influence of time from injury on the side-to-side difference in thickness (Figure 6), with subjects with a longer time from injury experiencing greater cartilage loss (9%) compared to those with a shorter time from injury (<1%). No significant differences in thickness changes between the ACL deficient and intact knees were detected between the intercondylar notch and medial tibial plateau cartilage.

Figure 6.

Mean percent differences (± 95% confidence intervals) in cartilage thickness in patients with short (n = 4, < 2 months) and long (n = 4, > 1 year) times from injury. In the medial femoral condyle adjacent to the notch and in the medial tibial plateau, there was a statistically significant influence of time from injury on the side-to-side difference in cartilage thickness. Subjects with a long time from injury showed more thinning in their ACL deficient knees relative to their contralateral knees compared to those patients with a short time from injury. * denotes significance p < 0.05.

The mean IKDC score across all patients was 70 with a minimum score of 38 and a maximum score of 100. Mean IKDC scores were significantly different between subjects with a long time from injury (mean IKDC = 89) versus short time from injury (mean IKDC = 49). IKDC scores were not correlated with compressive strain in any region.

DISCUSSION

Maintenance of normal cartilage loading is critical to its homeostasis, and the disruption of normal cartilage mechanical function may lead to the development of post-traumatic OA.^{3, 21, 22} Alterations in knee kinematics resulting from ACL deficiency may lead to changes in cartilage mechanics and subsequently altered cartilage metabolism. Thus, in the present study, cartilage strain resulting from a hopping activity and cartilage thickness were compared between ACL-deficient and intact contralateral knees. While we observed a statistically significant strain response to hopping in the medial and lateral compartments of the tibia and femur in both knees, we did not identify statistically significant differences in strain between the ACL-deficient and intact knees in any of these regions. The only region for which the cartilage strain response was significantly elevated in the ACL-deficient knee relative to the intact knee was in the medial intercondylar notch of the femur. Notably, this study also identified a statistically significant reduction in cartilage thickness between the ACL-deficient and intact knees in the medial intercondylar notch, as well as in the adjacent medial tibial plateau. This effect was most pronounced in those with more chronic ACL deficiency. Taken together, these findings suggest that altered local cartilage mechanics, as indicated by increased strain, may be related to local cartilage thinning in ACL-deficient knees.

The increased cartilage strain observed in the intercondylar notch of ACL-deficient knees relative to the intact knees may be the due to abnormal kinematics resulting from ACL deficiency. To this point, several studies have described kinematic changes resulting from ACL deficiency during both dynamic and static activities.^{4, 7, 12, 28, 30, 50} Specifically, in a study utilizing biplanar fluoroscopy, ACL-deficient knees were shown to have increased anterior tibial translation during gait as compared to uninjured knees.⁷ Using a similar methodology, increased anterior and medial translation of the tibia as well as internal rotation of the tibia relative to the femur in ACL-deficient knees was demonstrated during a step-up activity.²⁸ These increased translations were primarily identified when the knee was positioned at low flexion angles. Additionally, studies have shown increased medial tibial translation throughout the range of motion, as well as increased anterior translation and internal rotation near full extension, during quasi-static lunge with ACL deficiency.¹² Finally, a recent study indicated a medial subluxation of the tibia on standing radiographs after ACL deficiency.⁴³ Taken together, these kinematic studies suggest that when hopping on an ACL-deficient knee, the tibia may be excessively translated anteriorly and medially relative to the femur. Due to the proximity of the medial intercondylar notch to the tibial spine, medial translation of the tibia during knee flexion may result in elevated medial intercondylar notch compressive strain. Future studies should relate measurements of cartilage function to kinematic measurements to further clarify this mechanism.

There is evidence to suggest that potentially compensatory changes in landing kinematics can be measured during a single-legged forward hop in the contralateral knee after ACL reconstruction.²⁵ However, in this study a similar magnitude of strain response was measured in the ACL-deficient knees as compared to the intact knees in the lateral and medial plateaus of the tibia, as well as both femoral condyles. The magnitudes of the compressive strain (ranging from 2% to 5%) in the medial and lateral tibia reported here are congruent with a recent study⁴⁰ which investigated the compressive strain response to the same hopping activity in normal healthy patients and reported tibial strains of about 5%. Similarly, femoral compressive strains measured in the intact contralateral knee of ACL-deficient patients in the present study were also comparable to the femoral strains previously measured in the intact knees of uninjured patients. The similarity in the strain response measured in the intact contralateral knees as compared to normal controls suggest that the intact contralateral knees of ACL-deficient patients may be an appropriate control for quantifying the effect of ACL deficiency on cartilage mechanics.

We did not observe differences in pre-activity cartilage thickness between the ACL-deficient and intact knees in either the medial or lateral compartments of the femur or the lateral compartment of the tibia in this study. However, cartilage was significantly thinner in ACL-deficient knees as compared to ACL-intact knees in the intercondylar notch and the adjacent medial tibial plateau. These findings are in line with prior in vivo imaging studies.^{17, 33} Specifically, previous work has reported that the intercondylar notch is the first area to exhibit radiographic signs of post-traumatic OA in patients with chronic ACL deficiency.¹⁷ Focal cartilage thinning has also been observed in the intercondylar notch of patients with non-anatomic ACL reconstruction an average of 18 months after surgery.^{11, 33} Since non-anatomic placement of the ACL graft has been shown to alter tibiofemoral kinematics in a manner similar to ACL-deficient kinematics,¹ cartilage in this region is likely to be affected by ACL deficiency. These results regarding cartilage thinning in the intercondylar notch are particularly compelling because this is the same region where increased compressive strain following hopping was observed in the present study. It is possible that thinning occurs at earlier time points after ACL injury in this region of high strain, which may propagate to other cartilage regions at later time points, as evidenced by the thinner medial tibial cartilage observed in this study. To this point, we observed increased cartilage thinning in the group of subjects with longer time from injury (defined as >1 year), relative to patients with short time from injury (< 2 months) in both the medial intercondylar notch and the adjacent medial tibial plateau. Future prospective studies that track cartilage thickness changes in acutely ACL-deficient knees and at several later time points may help identify the timeline by which regional changes in morphology and strain changes may occur, and provide insight into the optimal timing of clinical interventions.

The patients in this cohort likely fall into the category of “copers”, or individuals with ACL injury who are able to obtain clinically adequate outcomes without ACL reconstruction.^{36, 37} This is evidenced by the fact that these subjects had the capacity to perform the relatively physically challenging hopping activity successfully on their ACL-deficient knee. However, patients with short time from injury had lower IKDC scores than patients with long time from injury. This finding suggests that those with more acute injuries may still be experiencing injury-related disability, whereas those with more chronic injuries had returned to their pre-injury levels of function. We did not observe a relationship between patient-reported knee function (IKDC score) and cartilage strain in this study. An explanation for this finding is that changes in cartilage mechanics and cartilage thinning may precede patients’ subjective evaluation of knee joint impairment that may be associated with the development post-traumatic OA.

Although the patients in this study were asked to refrain from strenuous activity the night before and morning of the study, and then rested supine for 45 minutes prior to testing, any preloading of the cartilage would result in measurement of lower pre-activity cartilage thickness. This would result in an underestimate of activity-related strain. Furthermore, the time required for patient transport and positioning for the post-activity MR scan may have allowed for some cartilage equilibration, which would also contribute to the underestimation of strain. In this study, the average time between the hopping activity and the post-activity MR scan was minimized to 3 minutes and 48 seconds. Previous work has suggested that patellar cartilage only recovers 50% of its volume in 45 minutes¹⁶, which is a long period of time relative to the delay between the completion of the hopping activity and the post-activity scan. Nonetheless, the strains observed in this study are likely to be underestimates. The time course of cartilage thickness recovery following activity, however, needs to be more thoroughly addressed and should be a focus of future research.

It should also be noted that the cartilage strains calculated in this study are a function of total thickness changes and do not account for the depth-dependent heterogeneity of cartilage mechanical properties.^{5, 49} Further investigation with other imaging modalities (e.g., T1ρ imaging)^{24, 38, 44, 51} may characterize how dynamic activity can affect the relationship between mechanical deformation and changes in composition due to exudation of water with loading.^{23, 39} Such information may further support the measurement of zone-specific strain within cartilage to help determine if increased strain is a result of increases in localized forces or changes in cartilage mechanical properties as a result of ACL deficiency.

In conclusion, this study examines the effects of ACL deficiency on cartilage thickness and acute cartilage strains resulting from dynamic activity. Notably, the present study found decreased cartilage thickness and elevated compressive strains and in the intercondylar notch of the medial femoral condyle in ACL-deficient knees relative to intact contralateral knees. This cartilage location has previously been shown to exhibit signs of cartilage degeneration. An understanding of changes in cartilage morphology and in the strain response to loading may provide new insight into the mechanisms of post-traumatic OA development with ACL deficiency.

Clinical relevance:

This study suggests that altered mechanical loading is related to localized cartilage thinning after ACL injury.

What is known:

Abnormal knee motion is believed to change cartilage loading and predispose the knee to cartilage degeneration. However, there is limited data quantifying the role of altered cartilage biomechanics on cartilage health in vivo.

What this study adds:

There are limited in vivo data describing the effect of ACL deficiency on cartilage function. The present study compared cartilage thickness and strains in response to dynamic activity between ACL-deficient and intact contralateral knees and determined that ACL deficiency may lead to localized regions of increased strain that are associated with reduced cartilage thickness in the tibiofemoral cartilage.