Dr. Hinterwimmer and Mr. Krammer contributed equally to this work.
Cartilage atrophy in the knees of patients after seven weeks of partial load bearing
Article first published online: 5 AUG 2004
Copyright © 2004 by the American College of Rheumatology
Arthritis & Rheumatism
Volume 50, Issue 8, pages 2516–2520, August 2004
How to Cite
Hinterwimmer, S., Krammer, M., Krötz, M., Glaser, C., Baumgart, R., Reiser, M. and Eckstein, F. (2004), Cartilage atrophy in the knees of patients after seven weeks of partial load bearing. Arthritis & Rheumatism, 50: 2516–2520. doi: 10.1002/art.20378
- Issue published online: 5 AUG 2004
- Article first published online: 5 AUG 2004
- Manuscript Accepted: 1 APR 2004
- Manuscript Received: 5 DEC 2003
- Deutsche Forschungsgemeinschaft
It is currently unknown whether human cartilage properties change during short periods of partial load bearing. We used a post–ankle fracture model to explore whether changes in cartilage morphology occur in the knee under conditions of partial load bearing.
The knees of 20 patients with Weber type B and type C fractures were examined using magnetic resonance imaging. The first scan was obtained shortly (mean ± SD 3.2 ± 3.0 days) after the injury, and a second scan was obtained 7 weeks later (mean ± SD 50.7 ± 5.5 days). The morphology (mean and maximum thickness, volume, and surface area) of the patellar, tibial, and femoral cartilage was determined from coronal and axial magnetic resonance images (fat-suppressed gradient-echo).
Between week 0 and week 7, the cross-sectional area of the quadriceps muscle was reduced by 11% (P< 0.001). Changes in the mean (±SD) cartilage thickness ranged from −2.9 ± 3.2% in the patella to −6.6 ± 4.9% in the medial tibia. No significant change in cartilage morphology of the contralateral knee was observed.
Results of this study demonstrate that in a post–ankle fracture model of partial load bearing, cartilage morphology in all knee compartments is subject to significant change. Changes in the femorotibial joint exceeded those in the patella, whereas no change was observed in the contralateral knee. These findings raise the question of whether cartilage is mechanically less competent and particularly vulnerable after states of partial or complete immobilization.
Quantitative magnetic resonance imaging (MRI) is a novel, but established, technique for studying cartilage morphology under physiologic and pathophysiologic conditions (1–5). Recent studies have shown that cartilage thinning occurs during aging (4), in osteoarthritis (OA) (2, 6, 7), after partial meniscectomy (8), and after spinal cord injury (3, 9). Results from animal models have shown that cartilage morphology, composition, and mechanical properties are subject to change during immobilization (10). However, it is as yet unknown whether changes occur in human cartilage during short periods of immobilization or partial load bearing. Such information is important in the postoperative management of OA and in the context of space travel (i.e., when astronauts return to normal gravity conditions after long-term space flight). The relevant question here is whether or not cartilage properties change under conditions of partial load bearing, and, supposing that such changes do occur, whether cartilage is vulnerable to injury when normal load-bearing conditions are reestablished.
In this study, we made use of the fact that patients who have undergone operative treatments of ankle fractures are subject to a 7-week period of partial load bearing. We explored the extent to which the cartilage morphology of the ipsilateral knee changed during the postoperative period in relation to the loss of cross-sectional area of the quadriceps muscle.
PATIENTS AND METHODS
We examined the knee joints of 16 male and 4 female patients with traumatic Weber type B and type C ankle joint fracture (mean ± SD age 34 ± 13 years [range 16–61 years]). Only partial load bearing of the injured limb after surgery is permitted, and we thus explored the effect of partial load bearing on healthy knee cartilage over short periods of time. Patients with a history of knee pain, knee trauma, or knee surgery prior to ankle joint fracture were excluded from the study.
The first MRI measurement of the knee (the same side as the ankle fracture) was obtained as soon as possible after the injury (mean ± SD 3.2 ± 3.0 days). The second measurement was acquired approximately 7 weeks postinjury (mean ± SD 50.7 ± 5.5 days). After the first MRI measurement of the knee, patients underwent surgery of the ankle, during which an internal fixation plate was implanted. During the following 7 weeks, the patients were allowed to walk on crutches, with only partial load bearing (contact with the sole of the foot). A physiotherapist trained patients how not to exceed a 20-kg load on the injured leg. In addition, the patients participated in a standardized therapy program, which included passive range of motion exercise and aqua therapy. This exercise program did not differ between patients. All participants received oral and written information regarding the nature of the examination and the study and signed a statement of informed consent that had been approved by the local ethics committee.
MRI and image analysis.
The injured knee of each patient was examined using a 1T MRI scanner (Magnetom Harmony; Siemens, Erlangen, Germany) and a circular polarized transmit–receive extremity coil. Imaging was performed with a validated (11, 12) spoiled gradient-echo sequence with fat suppression (prepulse) (fast low-angle shot: repetition time [TR] 49 msec, time to echo [TE] 10.3 msec, flip angle [FA] 30°, in-plane resolution 0.156 × 0.156 mm, section thickness 1.5 mm, matrix 1,024 × 1,024 pixels, field of view [FOV] 160 × 160 mm). Patellar cartilage was imaged using an axial scan orientation (acquisition time 14 minutes), and the femorotibial joint was imaged using a coronal section orientation (acquisition time 16 minutes). In 10 of the 20 patients, axial and coronal scans of the contralateral knee were also obtained at weeks 0 and 7. In the other 10 patients, additional axial images of the mid thigh were acquired at baseline and followup, using a fat-suppressed gradient-echo sequence (TR 200 msec, TE 20 msec, FA 90°, resolution 0.4 × 0.4 × 4.0 mm, matrix 512 × 512 pixels, FOV 204 × 204 mm, acquisition time 2 minutes). This was done in order to determine the change in cross-sectional area of the quadriceps muscle between weeks 0 and 7, because this can serve as an objective measure of the magnitude of partial load bearing. The measurement location was defined as the midpoint of the distance between the tip of the greater trochanter and the lateral knee joint space.
All image data sets were rated by an experienced musculoskeletal radiologist (M. Krötz) prior to quantitative cartilage analysis. The images were then segmented by one individual (M. Krammer) using a semiautomated algorithm, as described previously (1–5, 9, 13). The baseline and followup images were segmented during the same session, one after the other, but the investigator (M. Krammer) was blinded to the order of image acquisition (baseline or followup). Blinding of the imaging data (identification of baseline and followup examination) was performed by another investigator (SH). Cartilage volume, mean and maximal cartilage thickness, and size of the cartilaginous joint surface area were computed using proprietary software (1–5, 9). The outline of the quadriceps muscle was segmented manually and its cross-sectional area computed using the same software.
Student's paired t-test was used to assess the level of statistical significance for differences between baseline and followup measurements. The same test was used to compare the rate of change between different cartilage compartments. Because 10 pairwise tests were performed for comparing compartments, the required significance level was set at P < 0.01, to maintain a global significance level of P < 0.1. The correlation of the rate of change in the quadriceps muscle cross-sectional area with the rate of change in various cartilage plates (n = 10) and the correlation of the rate of change between different cartilage plates (n = 20) were assessed using linear regression analysis.
The cross-sectional area of the quadriceps muscle (n = 10) was 11% lower at week 7 than at week 0 (P < 0.001) in the partially loaded limb. Changes in mean cartilage thickness and cartilage volume between week 0 and week 7 (n = 20) were highly significant in all knee compartments (Table 1). The rate of change (the difference in mean cartilage thickness) varied between compartments: −2.9 ± 3.2% in the patella, −6.6 ± 4.9% in the medial tibia, −4.9 ± 4.3% in the lateral tibia, −4.7 ± 4.0% in the medial femoral condyle, and −6.3 ± 2.3% in the lateral femoral condyle (Figure 1). Changes in the femorotibial compartment (specifically, the medial tibia and lateral femur) significantly exceeded changes in the patella. Changes within the femorotibial compartment did not differ significantly from each other. No cartilage lesions were observed in any of the subjects at baseline or followup.
|Compartment||Maximum thickness, mm||Mean thickness, mm||Volume, ml||Surface area, cm2|
|0 weeks||5.72 ± 1.17||2.76 ± 0.52||3.95 ± 1.10||14.93 ± 2.54|
|7 weeks||5.69 ± 1.18||2.68 ± 0.49||3.78 ± 1.01||14.70 ± 2.33|
|% change||−0.4 ± 6.4||−2.9 ± 3.2†||−3.8 ± 4.1†||−1.4 ± 2.1‡|
|0 weeks||3.75 ± 0.52||1.79 ± 0.25||2.47 ± 0.61||12.39 ± 1.84|
|7 weeks||3.52 ± 0.52||1.67 ± 0.24||2.27 ± 0.59||12.20 ± 1.88|
|% change||−6.1 ± 7.1‡||−6.6 ± 4.9†||−8.3 ± 5.6†||−1.5 ± 2.7|
|0 weeks||4.82 ± 0.80||2.33 ± 0.37||2.70 ± 0.81||11.17 ± 2.46|
|7 weeks||4.58 ± 0.92||2.22 ± 0.41||2.54 ± 0.77||10.99 ± 2.25|
|% change||−5.3 ± 6.6‡||−4.9 ± 4.3†||−5.6 ± 3.2†||−1.2 ± 3.1|
|0 weeks||3.56 ± 0.70||1.96 ± 0.38||1.22 ± 0.34||5.69 ± 1.21|
|7 weeks||3.44 ± 0.65||1.87 ± 0.38||1.16 ± 0.34||5.61 ± 1.21|
|% change||−3.0 ± 7.4||−4.7 ± 4.0†||−5.0 ± 3.6†||−1.3 ± 1.0‡|
|0 weeks||2.80 ± 0.43||1.67 ± 0.22||1.16 ± 0.32||6.01 ± 1.38|
|7 weeks||2.65 ± 0.40||1.56 ± 0.21||1.08 ± 0.30||6.05 ± 1.39|
|% change||−4.6 ± 10.0||−6.3 ± 2.3†||−6.2 ± 3.4†||−0.7 ± 1.3|
Values for the mean and maximum cartilage thickness, cartilage volume, and joint surface area before and 7 weeks after partial load bearing are shown in Table 1. No significant correlation was observed between the magnitude of loss of quadriceps muscle cross-sectional area and the reduction in mean cartilage thickness in any of the 5 cartilage plates (r < 0.5; n = 10). Also, no significant correlation of cartilage loss was found between the 5 cartilage plates (r < 0.4; n = 20), except for the medial femoral condyle versus the lateral femoral condyle (r = 0.53, P < 0.05). In contralateral knees (n = 10), no significant change in cartilage morphology was observed in any of the knee joint cartilage plates (Figure 1 and Table 2).
|Compartment||Maximum thickness, mm||Mean thickness, mm||Volume, ml||Surface area, cm2|
|0 weeks||5.74 ± 1.04||2.78 ± 0.47||3.87 ± 0.87||14.67 ± 2.04|
|7 weeks||5.77 ± 0.99||2.75 ± 0.44||3.82 ± 0.82||14.62 ± 2.00|
|% change||+0.8 ± 4.5||−0.8 ± 3.4||−0.9 ± 4.2||−0.3 ± 1.4|
|0 weeks||3.59 ± 0.66||1.74 ± 0.32||2.45 ± 0.74||11.72 ± 2.30|
|7 weeks||3.64 ± 0.71||1.75 ± 0.27||2.43 ± 0.69||11.93 ± 2.24|
|% change||+1.3 ± 5.6||+1.3 ± 8.7||−0.3 ± 9.3||+2.0 ± 2.1|
|0 weeks||4.80 ± 0.59||2.41 ± 0.32||2.90 ± 0.86||10.93 ± 2.66|
|7 weeks||4.75 ± 0.61||2.39 ± 0.28||2.84 ± 0.83||10.93 ± 2.51|
|% change||−1.0 ± 4.8||−0.8 ± 3.6||−1.9 ± 4.0||+0.4 ± 3.0|
|0 weeks||3.49 ± 0.53||2.06 ± 0.31||1.24 ± 0.29||5.34 ± 0.80|
|7 weeks||3.45 ± 0.63||2.03 ± 0.35||1.22 ± 0.30||5.38 ± 0.88|
|% change||−1.6 ± 6.0||−2.1 ± 3.8||−2.0 ± 5.6||+0.6 ± 2.5|
|0 weeks||2.88 ± 0.43||1.72 ± 0.22||1.09 ± 0.27||5.34 ± 1.03|
|7 weeks||2.88 ± 0.51||1.74 ± 0.22||1.08 ± 0.28||5.31 ± 1.02|
|% change||−0.2 ± 6.8||+1.5 ± 3.1||−0.50 ± 5.1||−0.5 ± 2.1|
In this study, we used a postoperative model of partial load bearing to explore whether changes in cartilage morphology occur during a short observation period. The model was validated by determining changes in the muscle cross-sectional area; these amounted to 11% on average for the quadriceps muscle. Our results show that, under these conditions, cartilage thinning occurs in all compartments of the knee, although no cartilage lesions were observed. The reduction in cartilage thickness ranged from −2.9% in the patella to −6.6% in the medial tibia. As expected, no changes in cartilage morphology were observed in the contralateral limb, in which normal (or slightly increased) load bearing was maintained over the observation period.
The changes observed during partial load bearing are in the same range as the annual cartilage loss observed in patients with OA (6), albeit the period of partial load bearing was only 7 weeks. It is unclear at present whether these changes correspond to a true loss of cartilage or, rather, reflect a state of compositional reorganization of the cartilage matrix. Although experimental MRI methods for evaluating cartilage composition are currently being developed (14), there does not yet exist an established, validated protocol for the reproducible assessment of a defined component of cartilage for an entire cartilage plate. Measurement of cartilage morphology, in contrast, has been shown previously to deliver valid and highly reproducible results under in vivo imaging conditions (1–5, 9, 13, 15, 16).
Limitations of this study include the facts that quantitative analysis was performed by a single investigator and that intraobserver and interobserver variability were evaluated only for a similar protocol in a previous study (13) and not directly within the context of this study. Axial and coronal imaging orientations were chosen because they involve smaller precision errors compared with sagittal scans (13, 15, 16). In the case of the patella, precision errors of only 1.0% have been reported (root mean square coefficient of variation) (15), and precision errors in the tibia and femoral condyles have been shown to range from 2.5% to 3.2% (13).
It is interesting to note that, using this methodology, morphologic changes were observed during a very short observation period (7 weeks). The changes observed during this period of partial load bearing are in the same range as those observed per annum in patients with OA (6). These findings coincide with the recent observation of an annual rate of cartilage thinning of ∼10% in patients with spinal cord injury (9). In contrast to the previous study (9), in the current study changes in the patella were significantly less than those in other joint compartments. This may be attributable to the physiotherapy exercise program that included activities in which the femoropatellar joint compartment was loaded to a greater extent than was the femorotibial joint compartment (e.g., underwater exercise).
This study is the first to show that morphologic alterations of human cartilage occur during short periods of partial load bearing. Future studies will have to clarify whether or not these changes are reversible, what time periods and loading regimens are required to reverse the changes, and whether the cartilage is more vulnerable to mechanical insult after periods of complete or partial immobilization. The latter question is particularly relevant in the context of space travel, because astronauts may experience an increased risk of OA when returning to normal gravity conditions after long-term space flight.
In conclusion, the results of this study using a post–ankle fracture model of partial load bearing show that cartilage morphology is subject to significant change in all knee compartments. Changes in the femorotibial joint exceeded those in the patella, whereas no changes were observed in the contralateral knee. These findings raise the question of whether cartilage is mechanically less competent and particularly vulnerable after states of partial or complete immobilization.
We thank the patients for participating in this study. We thank Hans-Jürgen Pfeifer, MD (Institute of Clinical Radiology, Ludwig-Maximilians-Universitat München, Munich, Germany) and Karl-Hans Englmeier, PhD (Medis Institute, GSF National Research Center, Neuherberg, Germany) for technical advice.