The determinants of change in tibial cartilage volume in osteoarthritic knees




The rate of change in osteoarthritic (OA) tibial articular cartilage and the factors that influence it are not known. We examined a cohort of subjects with OA to determine the change in articular knee cartilage volume over the course of 2 years and to identify factors which might influence such change and its rate.


One hundred twenty-three subjects with OA underwent baseline knee radiography and magnetic resonance imaging (MRI) on their symptomatic knee. They were followed up 2 years later with a repeat MRI of the same knee. Knee cartilage volume was measured at baseline and at followup. Risk factors assessed at baseline were tested for their association with change in knee cartilage volume over time.


Mean ± SD total tibial articular cartilage decreased by 5.3 ± 5.2% (95% confidence interval [95% CI] 4.4%, 6.2%) per year. The annual percentages of loss of medial and lateral tibial cartilage were 4.7 ± 6.5% (95% CI 3.6%, 5.9%) and 5.3 ± 7.2% (95% CI 4.1%, 6.6%), respectively. Initial cartilage volume was the most significant determinant of loss of tibial cartilage in all compartments, while age was a significant determinant of lateral tibial cartilage loss, when possible confounders were accounted for.


In OA, tibial cartilage volume is lost at a rate of ∼5% per year. The main factor affecting cartilage loss is initial cartilage volume. Our results suggest that cartilage loss may be more rapid early in disease. Further study is required to determine whether the rate of cartilage loss in OA is steady or phasic, and to identify factors amenable to intervention to reduce cartilage loss.

Osteoarthritis (OA) is a major cause of work-related and long-term disability in people age >50 years (1). Despite this, factors influencing progression are poorly understood. This, in part, relates to the difficulty of measuring disease progression in OA.

Minimum joint space width, as an approximation of articular cartilage, has been recommended as the best measure of disease progression in OA, being more sensitive to change than global scoring of the joint (2, 3). Despite standardization of radiographic technique, the rates of progression of loss of joint space width in OA vary widely (4, 5). This variability may be, in part, due to the difficulty of obtaining reproducible radiographs, which is dependent on many factors, including consistent anatomic position of the joint, beam alignment, and distance between the joint and film (6, 7). The radiographic joint space consists of articular cartilage, meniscal cartilage, and possibly other structures, including effusions (8). Mild-to-moderate joint space loss has also been shown to be effected by meniscal extrusion, and not by loss of articular cartilage. Thus, although it is the currently recommended method, the use of change in minimum joint space width as a surrogate marker of articular cartilage and disease progression in OA may not provide entirely accurate results.

Magnetic resonance imaging (MRI) visualizes all components of the joint simultaneously. We and others have validated this as a method for measuring articular cartilage volume accurately and reproducibly in healthy individuals, those with OA, and children (9–14). Because this method measures cartilage in 3 dimensions, the results are less likely to be influenced by positioning. This is important in longitudinal studies. Cartilage volume has been shown to correlate with radiographic grade of OA (15). It is possible that articular cartilage volume may be a useful measure for disease progression in OA.

We examined a cohort of subjects with predominantly mild-to-moderate symptomatic OA over the course of 2 years. We sought to determine the change in articular knee cartilage volume during that time and to identify factors which might influence such change and its rate.


Subjects with early knee OA were recruited by advertising through local newspapers and the Victorian branch of the Arthritis Foundation of Australia, as well as in collaboration with general practitioners, rheumatologists, and orthopedic surgeons. The study was approved by the ethics committee of the Alfred and Caulfield Hospitals in Melbourne, Australia. All subjects gave informed consent.

One hundred thirty-two subjects entered the study. Inclusion criteria were age >40 years and symptomatic (score of >20% on at least 1 pain dimension of the Western Ontario and McMaster Universities Osteoarthritis Index [WOMAC; see ref. 16] and presence of osteophytes) knee OA (according to the clinical and radiographic criteria of the American College of Rheumatology [17]). Subjects were excluded if any other form of arthritis was present, if there were any contraindications to MRI (e.g., pacemaker, cerebral aneurysm clip, cochlear implant, presence of shrapnel in strategic locations, metal in the eye, or claustrophobia), or if they were unable to walk 50 feet without the use of assistive devices, had hemiparesis of either lower limb, or were planning to undergo total knee replacement.

Weight was measured to the nearest 0.1 kg (shoes and bulky clothing removed) using a single pair of electronic scales. Height was measured to the nearest 0.1 cm (shoes removed) using a stadiometer. Body mass index (BMI; weight [kg]/height [m2]) was calculated. Function and pain were assessed by the Short Form 36 Health Survey (18) and by the WOMAC.

At baseline, each subject had a weight-bearing anteroposterior tibiofemoral radiograph taken of the symptomatic knee in full extension. Where both knees had OA and were symptomatic, the knee with the less severe radiographic OA was used. Radiographs were independently scored by 2 trained observers who used a published atlas to classify disease in the tibiofemoral joint. The radiographic features of tibiofemoral OA were graded in each compartment on a 4-point scale (0–3) for individual features of osteophytes and joint space narrowing (19). In the case of disagreement between observers, the films were reviewed with a third independent observer. Intraobserver reproducibility for agreement on features of OA was 0.93 for osteophytes (grades 0 and 1 versus grades 2 and 3) and 0.93 for joint space narrowing (grades 0 and 1 versus grades 2 and 3). Interobserver reproducibility was 0.86 for osteophytes and 0.85 for joint space narrowing (by kappa statistic) (20).

Each subject had an MRI performed on his or her symptomatic knee at baseline and ∼2 years later. Knee cartilage volume was determined by means of image processing on an independent work station using the software program Osiris, as previously described (10, 20). Knees were imaged in the sagittal plane on the same 1.5T whole-body MR unit (Signa Advantage HiSpeed; General Electric Medical Systems, Milwaukee, WI) using a commercial receive-only extremity coil. The following sequence and parameters were used: a T1-weighted, fat-suppressed, 3-dimensional gradient recall acquisition in the steady state; flip angle 55 degrees; repetition time 58 msec; echo time 12 msec; field of view 16 cm; 60 partitions; 512 × 192 matrix; one acquisition time 11 minutes, 56 seconds. Sagittal images were obtained at a partition thickness of 1.5 mm and an in-plane resolution of 0.31 × 0.83 mm (512 × 192 pixels).

Two trained observers read each MRI. The scans were measured by 2 observers independently. Each subject's baseline and followup MRI scans were scored unpaired and blinded to subject identification and timing of MRI. The same 2 observers measured cartilage volume on each scan once. Their results were compared. If the results were within ±20%, an average of the results was used. If they were outside this range, the measurements were repeated until the independent measures were within ±20%, and the averages were used. Repeat measurements were made blinded to the results of the comparison of the previous results. The coefficients of variation (CVs) for the measurement of total, medial, and lateral cartilage volumes were 2.6%, 3.4%, and 2.0%, respectively (20).

Medial and lateral tibial plateau areas were determined by creating an isotropic volume from the input images which were reformatted in the axial plane. Areas were directly measured from these images. CVs for the medial and lateral tibial plateau areas were 2.3% and 2.4%, respectively (20). The average of the areas assessed at baseline and at 2-year followup was used.

Descriptive statistics for characteristics of the subjects were tabulated. Independent t-tests were used for comparison of means. The chi-square test was used to compare nominal characteristics between the groups. Change in cartilage volume (followup cartilage volume subtracted from initial cartilage volume) and percentage change (1 − [followup cartilage volume/initial cartilage volume], expressed as a percentage) over the period of time was divided by the time between MRI scans to obtain an annual rate of change. Principal outcome measures in analyses were the annual percentage of cartilage loss from baseline and the volume of cartilage lost annually. Multiple linear regression techniques were used to explore the possible factors affecting the rate of change in cartilage volume, including age, sex, height, weight, BMI, WOMAC scores (pain, stiffness, function), initial cartilage volume, bone size, and radiographic features (osteophytes, joint space narrowing in the studied compartment and in the opposite compartment). Results were also explored using quartiles of significant factors identified. All analyses were performed using the SPSS statistical package (version 10.0.5; SPSS, Chicago, IL).


One hundred twenty-three subjects (93%) completed followup (Table 1). Nine subjects were lost to followup: 2 moved overseas or interstate, 3 were too busy to continue in the study, 2 had knee surgery, 1 died of complications related to diabetes mellitus and chronic obstructive airways disease, and 1 subject was too ill to continue (due to multiple sclerosis). Of those subjects completing the study, most had mild or moderate OA, with only 23 (19%) having severe (grade 3) medial tibiofemoral osteophytes and/or joint space narrowing and only 16 (13%) having severe lateral tibiofemoral osteophytes and/or joint space narrowing. The cartilage volumes and tibial plateau areas were significantly greater in the men than in the women (Table 1).

Table 1. Characteristics of the study population*
 Total (n = 123)Men (n = 52)Women (n = 71)P
  • *

    Except where indicated otherwise, values are the mean ± SD. BMI = body mass index; WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index.

  • Men vs. women, by Student's t-test or chi-square test, as appropriate.

  • Values are the number of subjects in each grade. One radiograph was lost after the study commenced.

  • §

    By chi-square test.

Age, years63.1 ± 10.663.7 ± 10.662.6 ± 10.70.57
Height, cm167.3 ± 9.1175.1 ± 5.9161.6 ± 6.3<0.001
Weight, kg80.2 ± 14.983.9 ± 13.377.5 ± 15.40.018
BMI, weight (kg)/height (m2)28.7 ± 5.127.3 ± 3.429.7 ± 5.90.008
Time between scans, years1.94 ± 0.221.94 ± 0.191.94 ± 0.250.97
WOMAC score (pain, stiffness, function)427.1 ± 225.5395.1 ± 220.6450.2 ± 227.80.19
Kellgren-Lawrence grade    
Tibial plateau area, μm2    
 Medial2,072 ± 3922,387 ± 3341,841 ± 242<0.001
 Lateral1,366 ± 2541,496 ± 2351,270 ± 222<0.001
Cartilage volume at baseline, μm3    
 Medial1,731 ± 4741,989 ± 4991,542 ± 352<0.001
 Lateral1,922 ± 5762,277 ± 5681,662 ± 424<0.001

The mean ± SD amount of “total” tibial cartilage (medial + lateral tibial cartilage) lost per year ([cartilage at commencement of study − cartilage at end of study]/time between scans) was 201.1 ± 191 μm3 (Table 2). When this was calculated as a percentage of the initial baseline cartilage ([initial cartilage volume – followup cartilage volume]/initial cartilage volume per year), this represented a mean ± SD annual rate of loss of “total” tibial cartilage of 5.3 ± 5.2% (95% confidence interval [95% CI] 4.4%, 6.2%) (Figure 1). The mean ± SD amounts of medial and lateral tibial cartilage lost per year were 88.4 ± 115 μm3 and 112.7 ± 127 μm3, respectively (Table 2). Medial tibial cartilage was lost at an annual rate of 4.7 ± 6.5% (95% CI 3.6%, 5.9%) of initial cartilage, and lateral tibial cartilage was lost at an annual rate of 5.3 ± 7.2% (95% CI 4.1%, 6.6%) of initial cartilage.

Table 2. Annual change in tibial cartilage*
 TotalMenWomenDifference (95% CI)PAdjusted difference (95% CI)§P
  • *

    Except where indicated otherwise, values are the mean ± SD. 95% CI = 95% confidence interval.

  • Between men and women.

  • Men vs. women.

  • §

    Between men and women, results adjusted for differences in bone size and initial cartilage volume.

  • Men vs. women, adjusted difference.

  • #

    Medial + lateral tibial cartilage.

Change, %       
 Total#5.3 ± 5.25.6 ± 4.75.1 ± 5.60.55 (−1.28, 2.38)0.560.41 (−2.32, 3.14)0.77
 Medial4.7 ± 6.55.1 ± 6.54.5 ± 6.50.62 (−1.73, 2.96)0.610.40 (−2.87, 3.66)0.81
 Lateral5.3 ± 7.25.7 ± 5.65.1 ± 8.20.63 (−1.84, 3.10)0.61−1.15 (−4.49, 2.19)0.50
Amount lost, μm3       
 Total#201.1 ± 191239 ± 198173 ± 18265.9 (−3.3, 135.2)0.065.78 (−88.96, 100.51)0.90
 Medial88.4 ± 115104.6 ± 12876.6 ± 10228.1 (−14.8, 70.9)0.18−4.26 (−58.3, 49.7)0.88
 Lateral112.7 ± 127134.5 ± 12996.7 ± 12537.9 (−8.1, 83.8)0.10−20.9 (−75.74, 33.89)0.45
Figure 1.

Individual annual percentage rate of total tibial cartilage loss. The y-axis represents the number of subjects with a given percentage of tibial cartilage loss. The mean ± SD annual rate of loss was 5.3 ± 5.19% (n = 123 subjects).

The annual percentage and volume loss of medial and lateral tibial cartilages were significantly correlated (r = 0.27, P = 0.003 and r = 0.24, P = 0.007, respectively, by Pearson's correlation coefficients). There was no significant sex difference in the annual percentage change of cartilage (Table 2). Although the amount of cartilage lost by men each year was greater than that lost by women, after adjustment for bone size and initial cartilage volume, this difference disappeared (Table 2).

Univariate analysis showed that initial cartilage volume was the only significant predictor of cartilage loss and annual percentage loss in both tibial cartilages (Table 3 and Figures 2 and 3). The significance of initial cartilage volume increased after adjustment for age, sex, BMI, and bone size (corresponding tibial plateau area). However, for total tibial volume, a possible effect of increasing age and BMI was also observed (P = 0.08 and P = 0.09, respectively) (Table 3). Apart from initial cartilage volume, the factors affecting annual percentage loss of medial and lateral tibial cartilages differed (Table 3). The annual percentage loss of lateral tibial cartilage was significantly affected by initial cartilage volume (P = 0.002) and age (P = 0.016). The only significant factor affecting annual percentage medial tibial cartilage loss was initial cartilage volume (P = 0.01). The factors affecting amount of cartilage lost were the same as those affecting percentage change in all compartments. Change in weight during the period of the study, pain, grade of osteophytes, and grade of joint space narrowing had no significant effect on change in medial or lateral tibial cartilage volume.

Table 3. Factors affecting annual percentage cartilage loss*
Cartilage, factorRegression coefficient95% CIP
Univariate analysisMultivariate analysis
  • *

    95% CI = 95% confidence interval; BMI = body mass index.

  • Factors were defined as follows: age = percentage change per 1-year increase in age; sex = men compared with women; height = percentage change per cm increase in height; weight = percentage change per kg increase in weight; BMI = percentage change per unit increase in BMI; initial cartilage volume = percentage change per 1-μm3 increase in initial cartilage volume; bone size = percentage change per 1-μm2 increase in bone area.

  • Includes age, sex, BMI, initial cartilage volume, and bone size in regression equation.

 Age0.0050.04−0.07, 0.150.49
 Sex−0.05−0.64−4.02, 2.730.71
 BMI0.0850.11−0.12, 0.350.35
 Initial cartilage volume0.210.0040.00009, 0.0070.01
 Bone size−0.01−0.003−0.007, 0.0010.19
 Age0.120.150.03, 0.270.02
 Sex−0.041.01−2.43, 4.450.56
 BMI0.100.18−0.08, 0.430.17
 Initial cartilage volume0.250.0040.002, 0.0070.002
 Bone size−0.08−0.003−0.009, 0.0020.28
 Age0.080.08−0.01, 0.170.08
 Sex−0.05−0.74−3.55, 2.080.61
 BMI0.130.16−0.03, 0.350.09
 Initial cartilage volume0.150.0010.000001, 0.0030.05
 Bone size−0.04−0.002−0.004, 0.00050.12
Figure 2.

Mean annual percentage medial tibial cartilage loss by quartile of initial cartilage volume (quartile 1 = lowest cartilage volume). P = 0.07 for the difference between groups. 95% CI = 95% confidence interval.

Figure 3.

Mean annual percentage lateral tibial cartilage loss by quartile of initial cartilage volume (quartile 1 = lowest cartilage volume). P = 0.04 for the difference between groups. 95% CI = 95% confidence interval.


In our cohort of 123 symptomatic subjects with mild-to-moderate OA followed up for 2 years, we found that the rate of tibial cartilage loss was ∼5% per year. The main determinant of cartilage loss was the subjects' initial cartilage volume, with age also shown to be significant in the lateral compartment. Apart from initial cartilage volume, the factors affecting change in the medial and lateral tibial cartilages differed, with age being a significant determinant of lateral cartilage loss, but not of medial cartilage change. There was a significant correlation between the percentage loss of medial and lateral tibial cartilage.

Little is known about longitudinal change in cartilage volume. Until recently, the radiographic approximation of articular cartilage (joint space width) has been used to follow progression of OA. Studies of the rate of change in OA have provided widely varying results, from 0.06 mm/year to 0.6 mm/year, a 10-fold difference (21). How this relates to change in a 3-dimensional structure, such as the cartilage volume we measured, is unknown. More recently, MRI has been used to measure articular cartilage volume (9, 10). Investigators in two studies of small numbers of subjects (16 and 11 subjects with OA) reported, in abstract form only, that articular cartilage volume is lost at a rate of ∼6% per year, which is similar to our findings (22, 23). However, it has also been suggested that cartilage loss in OA may not be linear, which is in accordance with our finding of the importance of initial cartilage volume (24–26).

In this report, we have presented the average (mean ± SD) change in cartilage volume of the cohort. When dealing with the average change within a group, Student's paired t-test, with 95% CIs, is the appropriate tool. However, for percentage change in an individual, the minimum detectable difference (at a 5% level of significance) is 2.8 multiplied by the CV for a single volume measurement (27). In our study, this is ±7.28% per year. Hence, only 2 subjects had a statistically significant increase in cartilage volume during the course of the study. Both of these had grade 2 joint space narrowing. It may be that in these subjects a true increase in cartilage volume occurred, either by an increase in the amount of cartilage or by swelling in existing cartilage (28, 29). It is possible that there is an increase in joint cartilage in a subgroup of subjects with OA. However, it is also possible that our CIs are underestimated and that these subjects did not have a real gain in cartilage (27). Longitudinal studies of longer duration with larger numbers of subjects will be needed to determine this.

We showed only a moderate-to-weak correlation between cartilage loss in the medial and that in the lateral compartment. This is not surprising, since it is likely that although one compartment is mainly affected in OA, pathologic changes may affect the rest of the knee. Our findings support those of a previous study, based on radiographic measurement of progression of knee OA, that also suggested that medial and lateral tibiofemoral OA occur asymmetrically (30). Thus, although initial cartilage volume was the strongest determinant of cartilage loss in the medial and lateral compartments, the strength of this association was less when total tibial cartilage was examined. This is probably because it consists of the combination of the medial and the lateral cartilage, where in most cases one compartment was more significantly affected than the other. When the results of the two compartments were combined, the impact of initial cartilage volume on each compartment's cartilage loss was reduced by that of the other compartment.

In our study, age appeared to be an important determinant of the rate of loss of tibial cartilage in the lateral compartment. This is consistent with what is known about factors affecting the progression of radiographic OA (31). Although others have suggested that BMI is associated with a higher incidence of progression of joint space narrowing (24, 31), we did not demonstrate a significant effect of BMI on the rate of loss of tibial cartilage. We found no effect of sex on loss of tibial cartilage. In studies of radiographic progression of knee OA, the effect of sex has been inconsistent (24, 32, 33). The only factor, reported in abstract form only, which has been identified to increase cartilage loss (as assessed by MRI) is a meniscal tear (34). Since our study included only a small number of patients with previous meniscal injuries, we could not examine this. In this study, we showed some differences in the effects of age and BMI on medial and lateral tibial cartilage loss. Most epidemiologic studies now examine these two compartments as separate entities. However, apart from the effect of varus and valgus deformity on progression, we are not aware of any previous studies on the effect of different risk factors on disease progression in these two compartments (30).

Our results seem to differ from those of previous studies of the radiographic progression of OA, which relate the rapidity of progression to the severity of disease. Those studies have focused on the medial tibiofemoral compartment. Indeed, in subjects with OA, one study, which examined the lateral compartment, failed to demonstrate any change in lateral joint space width (35). However, we found similar tibial cartilage losses in both compartments. Radiographic progression has been shown to occur more rapidly in subjects with more severe disease in the hip and knee (35–37). This may appear contradictory to our findings, since joint space narrowing is seen as a surrogate marker for articular cartilage. However, radiographic joint space width includes articular cartilage, the menisci, and other structures (8). Since our results relate to directly measured changes in knee cartilage, it may be that changes in joint space width also relate to changes in these other structures or even to biomechanical changes associated with changes in articular cartilage as OA progresses. In our study, we have measured the changes in a 3-dimensional structure, using measurement in 3 dimensions. In contrast, joint space narrowing is a 1-dimensional assessment of this 3-dimensional structure.

The importance of initial cartilage volume in determining both the amount of cartilage lost and the percentage of cartilage lost suggests that cartilage loss may be more rapid early in disease, when more cartilage is present. A similar trend was observed when we examined the effect of radiographic grade of joint space narrowing, where subjects with increasing grade had reduced loss of tibial cartilage. This suggests that the rate of cartilage loss at the knee is not steady, but may be phasic. This will need to be examined in further studies. However, it may be analogous to the situation described for bone mineral density (BMD) in the femoral neck bone, where, in an age-stratified cohort study of women, the main predictor of femoral neck bone loss in women of different ages was baseline femoral neck BMD (38). It is possible that a floor effect exists when little or no cartilage remains. However, we excluded subjects with end-stage OA (Kellgren-Lawrence grade IV [39]) and are therefore unable to comment on this. Nor can we determine from our data whether there is a minimum amount of cartilage volume that is so small as to be beyond the level of detection for cartilage loss for the method being used.

Measurement of cartilage volume is limited by the contrast between articular cartilage and the adjacent tissues. However, our method has been validated against cadavers and has excellent reproducibility, with CVs of 2–3% (10, 20). It may be expected that as cartilage volume decreases with increasing severity of OA, the delineation of cartilage may be less than that in healthy individuals due to effusion, repair tissue, osteophytes, and the like. However, the accuracy of the method is similar in normal and OA knees (12, 13). Positioning and partial volume averaging may be expected to contribute to error. This method has been shown to be highly reproducible using 2-mm slices through the knee (40). We used 1.5 mm–thick slices to further minimize this. To improve in-plane resolution, we used a matrix of 512 × 192 pixels, resulting in an in-plane resolution of 0.31 × 0.83 mm.

Knee cartilage volume has potential advantages over measurement of cartilage thickness, being derived from multiple cross-sectional slices taken at 1.5-mm intervals across the knee. Although OA is a field change, cartilage damage may be focal. Because the whole cartilage is examined, cartilage volume measurement is less subject to change in positioning in the knee, which is a potential problem in reselecting the same position in longitudinal studies using cartilage thickness (13). Cartilage volume has been shown to be stable throughout the day, while cartilage thins in locations that encounter the greatest force (41). The potential disadvantage of using cartilage volume measurement is that the measurement may be more time consuming.

To our knowledge, this is the largest reported longitudinal study of MRI-measured knee cartilage volume. We recruited subjects with symptomatic OA from a broad base and did not select for any particular subgroup of patients with OA. Thus, it is unlikely that we selected a group of subjects more likely to lose cartilage. Nevertheless, this will need to be confirmed using larger numbers of subjects, followed up for longer periods of time, particularly to determine the role of other potential risk factors, such as current activity level, grade of OA, change in body weight, and pain.

In subjects with knee OA, tibial cartilage volume is lost at a rate of ∼5% per year. The main factor affecting cartilage loss is initial cartilage volume. Lateral tibial cartilage volume loss is affected significantly by age. Our results suggest that cartilage loss may be more rapid early in disease, when more cartilage is present. Further studies are required to determine whether the rate of cartilage loss in OA is steady or phasic, and to identify factors amenable to intervention to reduce cartilage loss.


We would like to acknowledge Andrew Forbes and Rory Wolfe for advice with statistics, Judy Hankin for doing duplicate volume measurements, the MRI Unit at Alfred Hospital for their cooperation, and Kevin Morris for technical support. We would especially like to thank the study participants, who made this study possible.