Knee cartilage of spinal cord–injured patients displays progressive thinning in the absence of normal joint loading and movement

Authors


Abstract

Objective

Alterations in the morphologic, biochemical, and mechanical properties of cartilage occur after unloading and immobilization in animals. However, the findings have been inconsistent and it is unclear whether such changes also take place in humans. This study tested the hypothesis that progressive thinning of knee joint cartilage is observed after spinal cord injury.

Methods

In this in vivo study, knee cartilage was assessed in patients with complete, traumatic spinal cord injury at 6 (n = 9), 12 (n = 11), and 24 months (n = 6) after injury. Morphologic parameters of the knee cartilage (mean and maximum thickness as well as surface area) were computed from magnetic resonance imaging (MRI) data, and results were compared with those in young, healthy volunteers (n = 9).

Results

After 6 months of injury, the mean articular-cartilage thickness was significantly less in the patella and medial tibia (decrease of 10% and 16%, respectively; P < 0.05), but not in the lateral tibia (decrease of 10%), compared with the MRI findings in healthy volunteers. After 12 and 24 months of injury, the differences amounted to a reduction of 21% and 23%, respectively, in the patella, 24% and 25%, respectively, in the medial tibia, and 16% and 19%, respectively, in the lateral tibia. The changes were significant in all 3 surfaces of the spinal cord–injured joint cartilage (P < 0.05–0.01).

Conclusion

Our data show, for the first time, that progressive thinning (atrophy) of human cartilage occurs in the absence of normal joint loading and movement. This may have important implications for patient management, in particular for spinal cord–injured patients and patients who are immobilized after surgery.

Loading and movement of the joint are of major importance for the maintenance of the morphologic and functional integrity of articular cartilage. Animal studies have demonstrated that joint immobilization and stress deprivation lead to functional adaptation of articular cartilage, and these changes involve the morphologic, biochemical, and biomechanical characteristics of the cartilage matrix (1–13).

Jurvelin et al (3) observed an average decrease in cartilage thickness of 9% in the canine knee after 11 weeks of rigid immobilization, and Haapala et al (10) observed a decrease of ∼20% in the medial femur of dogs. In contrast, Leroux et al (13) found no significant changes in the thickness of canine tibial cartilage after 4 weeks of nonrigid immobilization. Other authors have reported altered proteoglycan synthesis and content, fibrofatty proliferation at the articular surface, and softening of the cartilage during joint immobilization in animals (1–13). However, due to a lack of accurate noninvasive imaging methods, there has so far been no report on the morphologic changes of cartilage in humans following immobilization. This knowledge is important for the anticipation of cartilage changes in patients who are immobilized after surgical procedures or accidents, or after spinal cord injury.

In patients who experience a traumatic injury to the spinal cord, the joints in the lower extremity are unloaded and restricted in movement for long periods of time. These patients consequently encounter secondary complications, such as a decrease in muscle mass, cardiovascular fitness, and bone density (14, 15). Although there have been suggestions that the joints of these patients exhibit joint effusion (16–18), stiffness (17), fibrofatty connective tissue growth (18, 19), and heterotopic ossification (20), only 2 prior studies have attempted to assess the influence of paralysis on the articular cartilage in the lower extremity joints (21, 22). In these studies, a narrowing of the hip joint space by >50% was evident on radiographs from 25 of 200 patients with flaccid paralysis (21). Richardson et al (22) reported an apparent overgrowth of the epiphyses, periarticular osteoporosis, and joint space narrowing in patients with neuromuscular disorders. However, radiography cannot directly visualize cartilage, nor can it discriminate between different cartilage plates in the knee. In particular, accurate and reproducible measurements require a semiflexed, weight-bearing configuration of the knee (23), which cannot be achieved in these patients.

Recently, however, 3-dimensional (3D) magnetic resonance imaging (MRI), combined with state-of-the-art postprocessing, has been shown to provide accurate and highly reproducible data on human cartilage morphology in vivo (24–34). The objective of this study was, therefore, to assess, for the first time, morphologic alterations in human articular cartilage during long-term unloading and immobilization in patients with spinal cord injury. To determine the time course of potential change, we examined patients at various time periods after injury in a cross-sectional study.

PATIENTS AND METHODS

The right knees of 26 male patients with traumatic and complete spinal cord injury (mean ± SD age 37 ± 13.7 years, range 19–58 years) were investigated. At the time of the study, 9 individuals (mean ± SD age 45.5 ± 17 years) were 6 months postinjury, 11 individuals (mean ± SD age 32.4 ± 9.6 years) were 12 months postinjury, and 6 individuals (mean ± SD age 31.8 ± 7.5 years) were 24 months postinjury. In the 6-month group, 7 patients were paraplegic (range L1–Th5) and 2 were tetraplegic (C2, C5). Seven patients in the 12-month group were paraplegic (range L1–Th4) and 4 were tetraplegic (range C5–C7). In the 24-month group, 1 patient was paraplegic (Th6) and 5 were tetraplegic (range C5–C7). Patients with a history of knee pain, trauma, or surgery were excluded from the study. After receiving oral and written information, all participants signed a statement of informed consent, which was approved by the Ethics Committee of Lucerne.

The patients were examined with a 1.5T scanner (Magnetom Symphony; Siemens, Erlangen, Germany) and a circular polarized transmit-receive extremity coil. A previously validated fat-suppressed gradient-echo sequence (fast low-angle shot, repetition time 53 msec, echo time 10.3 msec, flip angle 30°) (25, 29) was used to acquire 1 transverse data set of the patellar cartilage and 1 coronal data set of the tibial cartilage (Figure 1). Images were obtained at an in-plane resolution of 0.31 mm × 0.31 mm and a slice thickness of 1.5 mm (matrix 512 × 512 pixels, field of view 160 cm). These orientations were chosen because they involve smaller precision errors than does a sagittal protocol. In the patella, a precision error of only 1.0% (root mean square average coefficient of variation) has been reported for cartilage-thickness measurements in healthy volunteers (30), and precision errors of 2.6% and 2.5% for the medial and lateral tibia, respectively, have been reported (31).

Figure 1.

Transverse magnetic resonance image (top) and coronal image (bottom) of the knee joint at 6 months after spinal cord injury (fat-suppressed fast low-angle shot sequence, repetition time 53 msec, echo time 10.3 msec, flip angle 30°, with a resolution of 0.31 mm × 0.31 mm × 1.5 mm, matrix 512 × 512 pixels, field of view 160 cm).

Segmentation was performed by a single investigator (BV) on a graphics computer (Octane Duo; Silicon Graphics, Mountain View, CA) on a section-by-section basis, using a B-spline Snake algorithm (27). Patellar, medial, and lateral tibial cartilage plates were reconstructed 3-dimensionally, and the volume was determined by numerical integration of the segmented voxels. The size of the joint surface area and the cartilage–bone interface area were computed after triangulation of these surfaces (34), and the mean and maximal cartilage-thickness values were calculated independent of the original section orientation by applying a 3D Euclidean distance transformation algorithm (28). The cartilage-thickness values in the patients were compared with those of a control group, consisting of 9 healthy male volunteers (mean ± SD age 28 ± 7 years, range 25–36 years) with no history of knee pain, trauma, or surgery and no abnormalities of the cartilage in the MR images. The data from the control group were segmented by the same observer (BV), and were consistent with the data previously reported for 49 young, healthy men (mean ± SD age 25.8 ± 3.2 years, range 20–30 years) (33) (Figure 2).

Figure 2.

Mean cartilage thickness of the patella, medial tibia, and lateral tibia of 49 young, healthy men (database), 9 healthy control subjects, and spinal cord–injured subjects 6 months (n = 9), 12 months (n = 11), and 24 months (n = 6) after injury. Bars show the mean ± SD. = P < 0.05 and ∗∗= P < 0.01 compared with controls, by Mann-Whitney U test.

The statistical significance of the differences between the patients and the control group was first assessed by a Kruskal-Wallis test. In a next step, each patient group (6 months, 12 months, and 24 months postinjury) was compared with the control group using a Mann-Whitney U test.

RESULTS

There was a significantly lower mean cartilage thickness in the patella, medial, and lateral tibia of the spinal cord–injured patients compared with the control group of young, healthy volunteers (P < 0.01, by Kruskal-Wallis test) (Figure 2). The maximal cartilage thickness in the medial tibia was also significantly decreased, but maximal changes in the patella and lateral tibia failed to reach statistical significance at the 5% error level.

In the patella, the differences in the mean and maximal cartilage thickness in the patients versus the volunteers amounted to 10% and 8%, respectively, at 6 months, 21% and 15%, respectively, at 12 months, and 23% and 16%, respectively, at 24 months postinjury (Table 1). The differences in mean thickness were significant for all groups of patients (P < 0.01, by Mann-Whitney U test).

Table 1. Cartilage morphology of the patella of healthy, young men (controls) and spinal cord–injured (SCI) subjects 6, 12, and 24 months postinjury*
 Controls (n = 9)SCI subjects
6 months (n = 9)12 months (n = 11)24 months (n = 6)
  • *

    Values are the mean ± SD.

  • P < 0.01 versus control group, by Mann-Whitney U test.

  • P < 0.05 versus control group, by Mann-Whitney U test.

Mean thickness, mm2.79 ± 0.192.48 ± 0.202.17 ± 0.302.12 ± 0.47
Maximal thickness, mm5.55 ± 0.535.13 ± 0.584.70 ± 0.824.69 ± 1.40
Joint surface area, cm213.8 ± 1.814.5 ± 1.712.9 ± 1.913.6 ± 1.8
Volume, ml4.43 ± 0.674.27 ± 0.713.39 ± 0.553.49 ± 0.95

In the medial tibia, the differences were higher than in the patella. Changes in the mean and maximal thickness amounted to 16% for both the mean and maximal thickness at 6 months, 24% and 19%, respectively, at 12 months, and 25% and 14%, respectively, at 24 months postinjury compared with the healthy controls (Table 2). The differences in mean and maximal thickness were significant for all groups of patients (P < 0.05).

Table 2. Cartilage morphology of the medial tibia of healthy, young men (controls) and spinal cord–injured (SCI) subjects 6, 12, and 24 months postinjury*
 Controls (n = 9)SCI subjects
6 months (n = 9)12 months (n = 11)24 months (n = 6)
  • *

    Values are the mean ± SD.

  • P < 0.05 versus control group, by Mann-Whitney U test.

  • P < 0.01 versus control group, by Mann-Whitney U test.

Mean thickness, mm1.65 ± 0.171.38 ± 0.171.26 ± 0.201.24 ± 0.22
Maximal thickness, mm3.77 ± 0.393.16 ± 0.443.05 ± 0.493.22 ± 0.39
Joint surface area, cm212.0 ± 1.712.3 ± 2.011.8 ± 1.21.11 ± 0.8
Volume, ml2.34 ± 0.461.89 ± 0.271.81 ± 0.341.63 ± 0.26

In the lateral tibia, the differences were lower than in the medial tibia. The changes in the mean and maximal thickness amounted to 10% for both the mean and maximal thickness at 6 months, 16% and 16%, respectively, at 12 months, and 19% and 7%, respectively, at 24 months postinjury compared with the healthy controls (Table 3). The mean cartilage thickness was significantly lower at 12 and 24 months (P < 0.05), but not at 6 months. The maximal cartilage thickness only displayed significant differences at 12 months postinjury (P < 0.05).

Table 3. Cartilage morphology of the lateral tibia of healthy, young men (controls) and spinal cord–injured (SCI) subjects 6, 12, and 24 months postinjury*
 Controls (n = 9)SCI subjects
6 months (n = 9)12 months (n = 11)24 months (n = 6)
  • *

    Values are the mean ± SD.

  • P < 0.05 versus control group, by Mann-Whitney U test.

Mean thickness, mm2.08 ± 0.201.86 ± 0.301.73 ± 0.291.67 ± 0.27
Maximal thickness, mm4.56 ± 0.714.12 ± 0.543.85 ± 0.694.24 ± 0.55
Joint surface area, cm211.6 ± 1.711.4 ± 1.410.9 ± 1.111.0 ± 1.1
Volume, ml2.53 ± 0.442.52 ± 0.612.25 ± 0.362.17 ± 0.40

The size of the bone–cartilage interface or articular surface showed no significant differences between groups. The changes in cartilage volume in patients followed a pattern similar to that of the mean cartilage thickness (Tables 1–3.

There were no statistically significant differences in cartilage thickness, joint surface area, or cartilage volume between the reference group used for the comparison with the patients and the larger database of 49 healthy volunteers (Figure 2).

DISCUSSION

In this study, we investigated whether human articular cartilage shows morphologic changes in spinal cord–injured patients. Our data show, for the first time, that progressive thinning of human cartilage occurs in the absence of normal joint loading and movement.

So far, in vivo analysis of human articular cartilage has been hampered, because no validated noninvasive measurement method has been available. Three-dimensional MRI combined with state-of-the-art postprocessing has only recently been shown to provide accurate and highly reproducible data on the cartilage morphology of living patients (24–34). Precision errors for the given imaging sequence and resolution have been shown to range from only 1–3% (30, 31), whereas we observed differences between groups of up to 20% in this study. The technology available therefore allows us to characterize with high reliability the time-dependent changes of human articular cartilage following spinal cord injury.

At 6 months postinjury, the largest changes in mean and maximal thickness were found in the medial tibia. In the lateral tibia, no substantial changes occurred, but a significant trend was also apparent. These results are consistent with the findings reported in some animal studies (3, 10). In dogs, the cartilage thickness of the medial tibia showed more substantial changes than the patellar or lateral tibial cartilage after a rigid immobilization of 11 weeks (3). Haapala et al (10) did not find significant changes in the lateral tibia after 11 weeks of rigid immobilization in dogs, while Jurvelin et al (3) found only a small degree of thinning of ∼4% in the cartilage. In contrast to these results, a recent study by Leroux et al (13) reported no changes in cartilage thickness in the medial tibia after 4 weeks of nonrigid immobilization in dogs. Thickness measurements in these studies were derived from histologic sections at only a few points. In contrast, the technique used here allows us to analyze the cartilage thickness throughout the whole joint, taking into account out-of-plane deviations of the minimal distance between the articular surface and bone–cartilage interface.

There are numerous clinical implications of the current findings. Thinning of the cartilage may render the joint unstable, and these joints may encounter abnormal stresses during passive standing training in spinal cord–injured patients. Adjunct therapies, such as functional electrostimulated cycling, may, however, be able to prevent cartilage thinning following spinal cord injury (18). Other forms of exercise, such as continuous passive motion (35), may be used to prevent cartilage changes during postoperative immobilization. This is of particular importance, since some animal studies have indicated that changes of the cartilage during immobilization are not fully reversible during remobilization (10, 11).

O'Connor made a distinction between unloading alone and unloading combined with nonrigid immobilization (8). Unloading alone caused thinning of the uncalcified cartilage in the medial tibia, but did not change the total thickness of the cartilage. Unloading combined with restricted movement did not cause changes in the uncalcified cartilage thickness of the medial tibia, but the uncalcified cartilage in the medial anterior tibia was decreased by 10%. Further studies will therefore have to identify whether thinning of human cartilage is caused by a decrease in weight-bearing or by a decrease in joint movement. Moreover, it will be necessary to investigate to what extent thinning of human cartilage is associated with biochemical, structural, and mechanical changes, whether changes proceed after 24 months, and to what extent the changes can be reversed during remobilization. Such data will be important to anticipate cartilage changes in patients who are immobilized after surgical procedures. The current techniques will also permit the evaluation of the effectiveness of potential countermeasures that would be designed to preserve the morphologic and functional competence of articular cartilage.

Acknowledgements

The authors thank all of the subjects for participating in this study. We acknowledge the contributions of Drs. Phil Jungen and Hans Hawighorst (Swiss Paraplegic Centre, Nottwil, Switzerland) in data collection. We also thank Drs. Christian Glaser, Maximilian Reiser (Institute of Clinical Radiology, Ludwig Maximillians University), and Karl-Hans Englmeier (National Research Center, Neuherberg, Germany) for providing technical advice and for allowing us to use the software for quantitative cartilage analysis.

Ancillary