Longitudinal analysis of cartilage atrophy in the knees of patients with spinal cord injury




A previous cross-sectional study indicated that the morphology of patellar and tibial cartilage is subject to change after spinal cord injury (SCI). The aim of this study was to perform a longitudinal analysis of cartilage atrophy in all knee compartments, including the femoral condyles, in SCI patients over 12 months.


The right knees of 9 patients with complete, traumatic SCI were examined shortly after the injury (mean ± SD 9 ± 4 weeks) and at 6 and 12 months postinjury. Three-dimensional morphology of the patellar, tibial, and femoral cartilage (mean and maximum thickness, volume, and surface area) was determined from coronal and transversal magnetic resonance images (fat-suppressed gradient-echo sequences) using validated postprocessing techniques.


The mean thickness of knee joint cartilage decreased significantly during the first 6 months after injury (range 5–7%; P < 0.05). The mean change at 12 months was 9% in the patella, 11% in the medial tibia, 11% in the medial femoral condyle, 13% in the lateral tibia, and 10% in the lateral femoral condyle (P < 0.05 for all compartments).


This is the first report of a longitudinal analysis of cartilage atrophy in patients with SCI. These data show that human cartilage atrophies in the absence of normal joint loading and movement after SCI, with a rate of change that is higher than that observed in osteoarthritis (OA). A potential clinical implication is that cartilage thinning after SCI may affect the stress distribution in the joint and render it vulnerable to OA. Future studies should focus on whether specific exercise protocols and rehabilitation programs can prevent cartilage thinning.

Integrity of the morphologic and mechanical properties of articular cartilage is a prerequisite for appropriate function of diarthrodial joints. As suggested by various animal studies, a certain amount of mechanical loading and movement of the joint is required in order to maintain normal cartilage morphology, biochemical composition, and biomechanical properties (1). However, the results of these animal studies have been highly variable and inconsistent. Jurvelin et al (2), for example, observed a 9% decrease in cartilage thickness in the canine knee after 11 weeks of rigid immobilization. Haapala et al (3) reported a large decrease (∼20%) in cartilage of the medial femur, but no changes in cartilage of the lateral tibia and lateral femur (4), of canine knee joints. In contrast, Leroux et al (5) found that a shorter period of nonrigid immobilization (4 weeks) did not change the knee cartilage thickness of the dogs in their study.

Spinal cord injury (SCI) causes unloading and restricted movement of the lower limb joints for substantial periods of time. As a consequence, patients with SCI encounter secondary complications, such as decreases in muscle mass, cardiovascular fitness, and bone density (6, 7). Radiographic evidence of narrowing of the hip joint space by more than 50% in 25 of 200 patients with flaccid paralysis was identified in one study (8). Another study showed overgrowth of the epiphyses, periarticular osteoporosis, and joint space narrowing in patients with neuromuscular disorders (9). It was recently shown that 3-dimensional magnetic resonance imaging (MRI), combined with state-of-the-art postprocessing, can provide accurate and highly reproducible data on cartilage morphology in vivo in health and disease states (10, 11). In a recent cross-sectional study, we found significant differences in patellar and tibial cartilage thickness of SCI patients compared with age-matched healthy volunteers (12). The objective of the present study was to longitudinally assess the magnitude and rate of morphologic changes of the cartilage in all compartments of the knee, including the femoral condyles, during unloading and immobilization after SCI.


Study subjects.

We examined the right knees of 9 men with traumatic and complete SCI (mean ± SD age 47 ± 18 years, age range 17–65 years). The first measurement was obtained as soon as possible after the injury (mean ± SD 9 ± 4 weeks). The knees were evaluated again at 6 and 12 months after injury. Seven patients were paraplegic (range of lesion height T4–T12) and 2 were quadriplegic (range of lesion height C4–C6). Patients with a history of pain, trauma, or surgery of the knee prior to the SCI were excluded from the study. One patient developed clear signs of osteoarthritis during the study, as noted on the MR images, and was therefore excluded from the statistical analysis.

During the 12 months of study, all patients received a standardized program of physical therapy at the Swiss Paraplegic Center, including passive range of motion exercises, hydrotherapy, and cardiovascular training (arm crank ergometer training). The physical therapy program did not differ considerably from patient to patient. All participants signed a statement of informed consent, as approved by the Ethics Committee of the canton of Lucerne, Switzerland, after receiving verbal and written information about the study.

Imaging and image processing.

The patients were examined with a 1.5T MRI scanner (Magnetom Symphony; Siemens, Erlangen, Germany), and a circular-polarized transmit–receive extremity coil. Transverse and coronal images were obtained using a validated fat-suppressed gradient-echo sequence (fast low-angle shot; repetition time 53 msec, echo time 10.3 msec, flip angle 30°, spatial resolution 0.31 × 0.31 × 1.5 mm, and matrix 512 × 512 pixels, field of view 160 cm) as described previously (12) (Figure 1). These orientations were chosen because they involve smaller precision errors than do sagittal scans (11). In the patella, precision errors of only 1.0% (root mean square coefficient of variation) have been reported with this technique, whereas precision errors in the tibia and femoral condyles range from 2.5% to 3.2%.

Figure 1.

Transverse (left) and coronal (right) magnetic resonance images of the knee joint of a spinal cord–injured patient. Fat-suppressed fast low-angle shot sequence, with a repetition time/echo time of 53/10.3 msec, flip angle of 30°, and spatial resolution of 0.31 × 0.31 × 1.5 mm (matrix of 512 × 512 pixels and 160-cm field of view).

Semiautomatic segmentation was performed by a single investigator (BV) using a graphics computer (Octane Duo; Silicon Graphics, Mountain View, CA) on a section-by-section basis. After manual initialization of a contour around the cartilage, a B-spline snake algorithm was used to attract the contour to the cartilage boundaries. The performance of the algorithm was then checked visually by the same observer (BV) and corrected manually, if required. Finally, the cartilage volume, mean and maximal thickness, and size of the joint and the bone–cartilage interface area were computed as described previously (10, 11).

Statistical analysis.

Wilcoxon's paired samples test was used to assess statistical significance of the differences between baseline and followup measurements. A significance level of 0.05 was chosen unless stated otherwise. The same test was used to compare the rate of change (as a percentage of the initial value between joint surfaces) at 12 months.


The total knee cartilage volume (patella, medial and lateral tibia, and medial and lateral femur) was reduced by 7% at 6 months postinjury. At 12 months postinjury, there was a 10% reduction in the total knee cartilage volume compared with the baseline values.

The magnitude and rate of change in the mean and maximal cartilage thickness for the patella, medial tibia, lateral tibia, medial femoral condyle, and lateral femoral condyle are shown in Table 1 and Figure 2. At 6 months post-SCI, changes in mean cartilage thickness ranged from 5% in the patella to 7% in the medial and lateral tibia. The change was significant in all compartments, but there was no significant difference in the rate of change between the 5 cartilage plates.

Table 1. Morphology of knee joint cartilage in 8 patients with spinal cord injuries, soon after injury and at 6 and 12 months postinjury*
Joint compartment, assessment timeThickness, mmVolume, mlSurface area, cm2
  • *

    Values are the mean ± SD.

  • P < 0.05 versus 0 months, by Wilcoxon's paired samples test.

 0 months5.04 ± 0.702.53 ± 0.324.11 ± 0.7313.43 ± 1.43
 6 months4.76 ± 0.672.40 ± 0.373.87 ± 0.6813.50 ± 1.37
 12 months4.72 ± 0.782.30 ± 0.323.68 ± 0.6413.26 ± 1.45
Medial tibia    
 0 months3.33 ± 0.441.61 ± 0.222.25 ± 0.6811.58 ± 2.67
 6 months3.16 ± 0.461.50 ± 0.252.06 ± 0.5911.40 ± 2.61
 12 months3.27 ± 0.541.44 ± 0.232.06 ± 0.6011.62 ± 2.76
Lateral tibia    
 0 months4.48 ± 0.562.28 ± 0.272.74 ± 0.6310.55 ± 1.67
 6 months4.26 ± 0.582.11 ± 0.252.45 ± 0.7810.39 ± 1.79
 12 months4.35 ± 0.571.98 ± 0.282.40 ± 0.6310.40 ± 1.83
Medial femoral condyle    
 0 months3.14 ± 0.391.80 ± 0.290.99 ± 0.203.98 ± 0.68
 6 months2.97 ± 0.471.68 ± 0.300.94 ± 0.204.06 ± 0.80
 12 months2.81 ± 0.481.59 ± 0.250.87 ± 0.163.94 ± 0.69
Lateral femoral condyle    
 0 months2.70 ± 0.291.60 ± 0.151.04 ± 0.244.81 ± 1.04
 6 months2.55 ± 0.271.50 ± 0.170.96 ± 0.204.71 ± 1.07
 12 months2.46 ± 0.331.41 ± 0.150.91 ± 0.184.66 ± 0.98
Figure 2.

Mean thickness of the cartilage of the patella, medial and lateral tibia, and medial and lateral femoral condyles of 8 patients with spinal cord injury. Measurements were made as soon as possible after the injury and at 6 and 12 months after the injury. Values are percentages of the initial mean cartilage thickness.

At 12 months after SCI, changes in mean cartilage thickness ranged from 9% in the patella to 13% in the lateral tibia. Again, the changes were significant in all compartments (P < 0.05). The relative rate of thinning in the 5 compartments did not differ significantly at 6 months (Figure 2). Comparisons of the relationship between the age of the patients and the magnitude (rate) of change at 12 months in the various cartilage plates yielded no significant associations.

There was no significant difference in the size of the bone–cartilage interface or the articular surface between baseline and either followup point. Thus, the changes in the cartilage volume followed a pattern similar to that of the changes in the mean cartilage thickness (Table 1).


In this study, we used a longitudinal design to analyze the magnitude and rate of cartilage thinning during restricted motion and unloading of the knee in patients with spinal cord injury. We have previously reported significant differences in cartilage thickness in different groups of patients at different time intervals after SCI (12); however, due to the high intersubject variability in morphologic cartilage properties (13, 14), that cross-sectional study did not allow us to reliably determine the magnitude and rate of change in cartilage morphology. Moreover, in the current study, we also explored changes in the femoral condyles. Femoral condyle changes have not been previously studied in humans (12), although animal studies have indicated that changes during immobilization may differ between different cartilage plates and compartments of the knee (2–5).

The mean thickness of knee joint cartilage in patients in the present study decreased by 5–7% after 6 months of immobilization and by 9–13% after 12 months of immobilization. These annual changes exceed those that occur during normal aging (14), which average ∼0.4% in most knee joint cartilage plates. The changes observed after SCI thus exceed changes during normal aging by a factor of >20:1. The changes after SCI also exceed by a factor of ∼2:1 (∼5% per year) the changes reported in patients with Kellgren/Lawrence grades 1–3 osteoarthritis (14). It is interesting to note that restricted movement and loading appear to have more severe effects on the maintenance of the volume and thickness of previously healthy cartilage than has been observed during the progression of advanced osteoarthritis. In this context, it is also of relevance that all SCI patients received a standardized program of physical therapy, including passive range of motion exercise, hydrotherapy, and cardiovascular training. It is possible that atrophic cartilage changes may have even been higher in the absence of these therapeutic efforts.

We found relatively uniform changes in cartilage thickness in all knee compartments, with a trend toward lower rates in the patella and higher rates in the lateral tibia, but no significant differences between cartilage plates. Whether this represents a random finding due to the small sample size or is due to, for example, a more flexed position of the knee after SCI is currently unclear. In any case, this finding is in contrast with the observations by Haapala et al in their studies of the knee joints of dogs. Those investigators found cartilage thickness changes only in the medial tibia and medial femur (3), without changes in the lateral tibia and femur (4), after 11 weeks of rigid immobilization of the knee joint. In those studies, however, cartilage thickness was measured in only a few areas of histologic sections. In contrast, the technique applied here permits one to analyze the cartilage thickness throughout the entire joint surfaces, taking into account out-of-plane deviations of the minimum distance between the articular surface and the bone–cartilage interface (10, 11). The changes after SCI, as observed in this study, also clearly exceed the precision errors associated with the technique we applied, which range from 1% in the patella to ∼3% in the femorotibial cartilage plates.

Given the relatively large change in cartilage morphology during the first year after SCI, it is likely that substantial biochemical, structural, and mechanical changes are involved, but these features are, at present, more difficult to reliably measure in vivo. Whether the morphologic changes continue at the same rate after 1 year remains to be determined. The extent to which the changes can be slowed or stopped by therapeutic interventions, such as mobilization and physical rehabilitation programs, is also not known.

The findings of the current study have some clinical implications. If patients are immobilized for substantial periods of time, thinning of the cartilage may cause an alteration in the congruity of the femorotibial joint surfaces, and thus, cause changes in the stress distribution throughout the joint (15). This may render the joint vulnerable to osteoarthritic degeneration. Therefore, it may be important to initiate therapeutic programs as soon as possible after SCI, so that cartilage atrophy is prevented at the earliest possible time point. Several types of post-SCI remobilization programs are available, such as functional electrostimulated cycling or treadmill training using robotic orthoses. It will therefore be a challenging issue for future studies to objectively evaluate which specific programs are most beneficial in maintaining normal morphologic properties of the knee joint cartilage. Such exercise programs will also be important in the context of postoperative immobilization and long-term space flight, in which equal measures may have to be taken to avoid cartilage atrophy, loss of mechanical competence, and the associated changes in the joints.


The authors thank all of the subjects for participating in this study and acknowledge the contribution of Dr. Phil Jungen (Swiss Paraplegic Centre, Nottwil, Switzerland) in the data collection. We thank Dr. Christian Glaser and Prof. Dr. Maximilian Reiser (Institute of Clinical Radiology, LMU, Munich, Germany), and PD Dr. Karl-Hans Englmeier (Medis Institute, GSF National Research Center, Neuherberg, Germany) for technical advice and for allowing us to use the software for quantitative cartilage analysis.