To evaluate the effects of age, physical activity, and body mass index (BMI) on patella cartilage volume and defects and bone volume in middle-aged women without knee pain.
To evaluate the effects of age, physical activity, and body mass index (BMI) on patella cartilage volume and defects and bone volume in middle-aged women without knee pain.
Magnetic resonance imaging was performed in 176 healthy women, ages 40–67 years, without knee pain to measure patella cartilage and bone volume and patella cartilage defects. The effects of age, physical activity, BMI, smoking, and alcohol were analyzed to determine whether associations existed between these variables and patella cartilage and bone volume and cartilage defects.
Patella cartilage volume decreased with age (P = 0.01) and BMI (P = 0.05) after adjusting for age and patella bone volume. Patella bone volume was positively associated with body height in both the univariate and multivariate models. Cartilage defects in the patellofemoral compartment were present in 36.4% of the study population. Age, weight, and BMI were positively associated with the presence of cartilage defects in the multivariate analysis.
This study demonstrated that although age is positively associated with both patella bone volume and cartilage defects, it is inversely associated with patella cartilage volume in healthy individuals. Moreover, BMI is inversely associated with both patella cartilage volume and patella bone volume in middle-aged women without knee osteoarthritis. Longitudinal studies will be required to determine whether avoiding a high BMI will reduce the risk of developing patellofemoral osteoarthritis.
Osteoarthritis (OA) is a common disease that increases in prevalence with age and is more common in women than in men after midlife (1). It is a joint disease involving both cartilage and bone. With increasing disease severity, articular cartilage is lost and bony changes such as subchondral stiffness and osteophytes (2) become apparent. Although there is a growing body of evidence characterizing the features of joint morphology in the presence of OA, less is known about the factors associated with healthy joint structure.
The knee joint consists of 3 compartments: medial tibiofemoral, lateral tibiofemoral, and patellofemoral. Tibiofemoral disease is predominant, especially in the medial compartment. However, patellofemoral OA is common, with or without the combination of tibiofemoral disease, and has been implicated as the cause of significant symptoms and disability (3, 4). A study by McAlindon et al (5) demonstrated that medial joint and patellofemoral joint OA were significantly associated with disability compared with controls without knee pain. However, higher Health Assessment Questionnaire (HAQ) scores were more common in subjects with patellofemoral joint OA (higher HAQ scores meant more disability) compared with those with medial joint disease. Despite this finding, little is known about the healthy patellofemoral compartment prior to disease onset.
Factors influencing the pathogenesis of OA have been identified through epidemiologic and small group studies that have examined individuals with established joint disease. However, whether factors such as body mass index (BMI) and age affect the healthy joint at the level of the articular cartilage and bone has not been extensively studied, particularly in the patellofemoral joint.
Cartilage loss and bony changes such as subchondral sclerosis and osteophytes are significant radiologic features of OA. There is now increasing interest in examining the interrelated structural features (cartilage volume, cartilage defects and bone) at the knee. Patella cartilage volume has been shown to correlate with radiologic grade of patellofemoral OA (6). We have previously shown that in persons with OA, patella cartilage volume is lost at a rate between 3.7% and 5.3% per year, with no association between change in patella cartilage volume and change in either medial or lateral tibial cartilage volume (7). However, little work has been done to examine factors affecting patella cartilage in healthy individuals. Only 1 recent study has examined this in the patellofemoral joint and demonstrated that patella cartilage was reduced with increasing BMI and age (8). Likewise, there is a paucity of data examining the factors associated with patella bone volume.
Cartilage defects have stimulated increasing interest in recent years. They are commonly found not only in patients with knee OA (9, 10) and those with knee pain requiring arthroscopy (11), but also in healthy individuals without knee pain or radiographic OA (11–14). Cartilage defects have been shown to be significantly associated with disease severity in OA (15), and they predict knee cartilage loss (13) and knee replacement (15). In this cross-sectional study of healthy middle-aged women without symptomatic knee OA, we examined the association between factors such as age, weight, and body height, as well as lifestyle factors such as obesity and physical activity that may affect patella cartilage and bone volume.
Eligible participants were part of a previous cross-sectional study examining androgens in women and had been recruited from a database established from the electoral roll in the southern Australian state of Victoria between April 2002 and August 2003 (16). Of the 1,423 participants ages 18–75 years who participated in the original study, 355 women who were in the desired age group (40–67 years), had not had a hysterectomy, and had agreed to be recontacted about participation in further research studies were eligible. Women were excluded if they had experienced significant knee pain or a knee injury in the last 5 years that necessitated treatment by a doctor or physiotherapist or required rest for more than 1 day; had a contraindication to having a magnetic resonance imaging (MRI) scan, such as the presence of a pacemaker, metal sutures, iron filings in the eye, or claustrophobia; or were unlikely to be available to complete the full 2-year study. The study was approved by the Southern Health Human Research and Ethics Committee and the Monash University Human Research and Ethics Committee, Clayton, Victoria, Australia. All participants gave written informed consent.
Participants were asked a series of questions to establish their menopause status. When menopausal status was uncertain, a decision tree was used to determine status. Physical activity was assessed using 3 questions about walking, activity at work, and sport to create an overall physical activity score (17). At the time of the original cross-sectional study (2002–2003), each participant completed a general health questionnaire that included current medication, smoking status, alcohol consumption, and occupation. Smoking status was assessed by asking, “Are you a current, past, or never smoker?” If current, “How many cigarettes per day?” If ever, “How long since stopped?” Alcohol consumption was assessed by asking participants how many standard alcoholic drinks they drink each day, week, and month. The number of alcoholic drinks per year was then calculated and entered into the main data set. BMI was calculated as weight (kg)/height (m2).
Participation in strenuous, light, and competitive exercise was ascertained using specific questions about these activities (18). Participation in strenuous exercise was assessed by asking, “In the last 14 days did you spend at least 20 minutes doing strenuous exercise, e.g., bicycling, brisk walking, jogging, or aerobics, that was severe enough to raise your pulse rate or cause you to breathe faster?” Participation in light exercise was ascertained by asking, “In the last 14 days have you spent at least 20 minutes doing light exercise, e.g., walking, light housework, slow bicycling, etc. that was not severe enough to cause a pulse rate rise or breathing increase?”
Each woman underwent MRI of her dominant knee between October 2003 and August 2004. The dominant knee was defined as the limb the woman stepped off from to initiate gait. Knees were imaged in the sagittal plane on a 1.5T whole-body magnetic resonance unit (Phillips, Eindhoven, Holland) using a commercial transmit-receive extremity coil, and patella and bone cartilage volume were determined by MRI processing on an independent work station using the Osiris software (The University Hospital of Geneva, Switzerland) as previously described (19, 20).
The same trained observer read each MRI (FSH). The intraobserver coefficients of variation for the patella cartilage volume and patella bone volume measures were 2.1% and 2.2%, respectively. Fifty participants were used to calculate the intraobserver coefficient of variation.
Cartilage defects were graded on the MR images with a modification of a previous classification system (21–23) at the patellofemoral site as follows: grade 0 indicated normal cartilage, grade 1 indicated focal blistering and intracartilaginous low-signal intensity area with an intact surface and base, grade 2 indicated irregularities on the surface or base and loss of thickness <50%, grade 3 indicated deep ulceration with loss of thickness >50%, and grade 4 indicated full-thickness chondral wear with exposure of subchondral bone. We found that cartilage surface in some images was still regular but cartilage adjacent to subchondral bone became irregular, so we included these changes in the classification system. Such observations were mainly labeled as grade 1, unless considerable loss of thickness was also observed. A cartilage defect also had to be present in at least 2 consecutive slices. The cartilage was considered to be normal if the band of intermediate signal intensity had a uniform thickness. The cartilage defects were re-graded 1 month later by the same observer, and the average scores of cartilage defects at the patellar compartment (range 0–4) were used in the study. A prevalent cartilage defect was defined as a cartilage defect score ≥2 at any site within the patellofemoral compartment. Intraobserver and interobserver reliability for the patellar compartment (expressed as intraclass correlation coefficient) were 0.94 and 0.93, respectively, for the total score (24).
The number of women eligible for recruitment into this study was limited by including only those who participated in the previous study, were in our desired age range, had not had a hysterectomy, and had agreed to be contacted about participation in future studies. The final number recruited for this study was 176, which gave us a power of 80% to show a correlation as low as 0.2 between the various risk factors and patella cartilage volume (alpha error 0.05, 2-sided significance), thus explaining up to 4% of the variance of cartilage and bone volume.
Outcome variables (patella cartilage volume, patella cartilage defects, and patella bone volume) were initially assessed for normality before being regressed against other variables. Relationships between variables were explored in univariate and then multiple linear regression models for continuous outcomes and logistic regression for dichotomous outcomes. P values less than 0.05 were considered to be statistically significant. All analyses were performed using the SPSS statistical package, version 12.0.1 (SPSS, Chicago, IL).
A total of 176 women underwent MRI scanning. The demographic features of study participants are presented in Table 1. The mean ± SD age of the women was 52.3 ± 6.7 years. Almost all women (n = 173) had participated in light exercise within the last 2 weeks. Smaller proportions had participated in strenuous activity or competitive sporting activity (Table 1). Nine women reported receiving an oral contraceptive pill whereas 21 women reported using some form of hormone replacement therapy. A total of 35 women were classified as using some sort of hormone treatment (oral contraceptive pill, hormone replacement therapy, and/or other hormone preparations).
|Age, years (n = 176)||52.3 ± 6.7 (40–67)|
|Body height, cm (n = 176)||164 ± 0.06 (1.46–1.82)|
|Weight, kg (n = 176)||72.7 ± 14.1 (46–127)|
|BMI, weight (kg)/height (m2) (n = 176)||27.1 ± 5.5 (16.1–51.5)|
|Patella cartilage volume at baseline, ml||2.5 ± 0.6 (1.13–4.16)|
|Patella bone volume, mm3||19.3 ± 2.8 (13.5–28.8)|
|Patella cartilage defects, no. (%)|
|Grade ≥2||64 (36.4)|
|Grade ≤1||112 (63.6)|
|Grade = 2||40 (22.7)|
|Grade = 3||14 (8)|
|Grade = 4||10 (5.7)|
|Medial tibial cartilage volume, ml||1.50 ± 0.28 (0.86–2.34)|
|Lateral tibial cartilage volume, ml||1.77 ± 0.35 (1.12–3.14)|
|Postmenopausal, no. (%)†||97 (57)|
|Smoking status, no. (%) yes (n = 176)||21 (11.9)|
|Alcohol (at least 1 drink per week), no. (%)||138 (78.9)|
|Light physical activity, no. (%) yes (n = 175)||173 (98.3)|
|Strenuous physical activity, no. (%) yes (n = 175)||138 (78.4)|
In univariate analyses, body height and patella bone volume were positively associated with patella cartilage volume, whereas age and BMI were negatively associated with patella cartilage volume (Table 2). After adjusting for potential confounders including age, body height and weight, and patella bone volume, the positive association between body height and patella cartilage volume did not remain significant. The multivariate model that included age, bone volume, and body height and weight accounted for 31.8% of the variance in patella cartilage volume. When BMI was substituted for body height and weight in the above model including age and patella bone volume, these variables accounted for 31.6% of the variation in patella cartilage volume. Because the 2 models showed no appreciable difference in variation, it was decided that the model that included age, body height, weight, and patella bone volume would be used to assess the association between joint morphology and lifestyle factors in this study. No association was found between smoking status, level of physical activity, or menopausal status and patella cartilage in either the univariate or multivariate analyses. There was a trend for patella cartilage volume to decrease after menopause (P = 0.09). Smoking status remained nonsignificant when examined as packet-years of smoking (P = 0.81). There was a strong correlation between patella and tibial cartilage (R2 = 0.44, P < 0.001), which was independent of age, body height, and weight.
|Univariate regression coefficient||P||Multivariate regression coefficient||P||R2 for model|
|Age, years||−14.0 (−26.2, −1.7)||0.03||−14.2 (−24.8, −3.5)||0.01||0.318†|
|Bone volume, cm3||0.10 (0.08, 0.13)||< 0.001||0.10 (0.07, 0.13)||< 0.001||0.318|
|Body height, cm||3.1 (1.9, 4.2)||< 0.001||0.9 (−0.4, 2.1)||0.19||0.318|
|Weight, kg||−2.7 (−8.5, 3.1)||0.37||−4.3 (−9.2, 0.7)||0.09||0.318|
|BMI, kg/m2||−20.8 (−35.6, −6.0)||0.01||−12.6 (−25.3, 0.1)||0.05||0.316‡|
|Smoking status, yes (n = 176)||40.0 (−49.8, 129.8)||0.38||100.7 (−112.0, 313.3)||0.69||0.321†|
|Alcohol||135.1 (−66.7, 336.7)||0.21||−31.3 (−205.6, 143.0)||0.72||0.301†|
|Strenuous activity, yes/no||26.2 (−174.0, 226.4)||0.8||−52.3 (−224.1, 120.1)||0.55||0.319†|
|Light exercise, yes/no||−46.3 (−683.1, 590.1)||0.9||−86.2 (−622.2, 450.0)||0.75||0.318†|
|Menopause, yes/no||−93.1 (−260.1, 74.1)||0.3||190.1 (−29.0, 409.1)||0.09||0.330†|
Sixty-four (36.4%) women had a defect (grade 2 or higher) in their patella cartilage (Table 1). Both univariate and multivariate analyses demonstrated a significant association between age, cartilage volume, BMI, and weight and presence of defects in the patella cartilage (Table 3). The multivariate model that included age, cartilage and bone volume, body height, and weight accounted for 45.8% of the variance in patella cartilage defects. After adjusting for the above covariates, there was a trend towards statistical significance between strenuous activity and the presence of cartilage defects in the patellar compartment (odds ratio [OR] 2.3; 95% confidence interval [95% CI] 0.9, 6.0; P = 0.08). The magnitude of the effect of strenuous activity tended to increase as the definition of defects was changed to a higher grade, suggesting a dose effect (data not shown). In univariate analysis, postmenopausal state was positively associated with cartilage defects in the patella cartilage (OR 3.7; 95% CI 1.9, 7.3; P < 0.001), but this relationship did not persist after adjusting for age, body height, weight, cartilage volume, and bone volume (P = 0.36) (Table 3).
|Univariate regression||P||Multiple regression||P||R2 for model|
|Age, years||1.10 (1.07, 1.20)||< 0.001||1.14 (1.07, 1.21)||< 0.001||0.458†|
|Patella cartilage volume, cm3||0.998 (0.997, 0.999)||< 0.001||0.998 (0.997, 0.999)||< 0.001||0.458†|
|Bone volume, cm3||1.000 (1.000, 1.000)||0.08||1.000 (1.000, 1.000)||0.42||0.458†|
|Body height, cm||0.96 (0.92, 1.0)||0.13||1.01 (0.95, 1.08)||0.71||0.458†|
|Weight, kg||1.04 (1.01, 1.06)||0.002||1.04 (1.01, 1.07)||0.006||0.458†|
|BMI, kg/m2||1.12 (1.1, 1.2)||< 0.001||1.09 (1.02, 1.17)||0.01||0.446‡|
|Smoking status, yes/no||0.5 (0.2, 1.5)||0.20||0.6 (0.16, 2.22)||0.45||0.462†|
|Alcohol||0.9 (0.4, 2.1)||0.83||0.4 (0.1, 1.2)||0.09||0.471†|
|Light exercise, yes/no||0.3 (0.03, 3.1)||0.30||0.32 (0.03, 4.05)||0.38||0.463†|
|Strenuous activity, yes/no||1.3 (0.6, 2.8)||0.49||2.3 (0.9, 6.0)||0.08||0.474†|
|Menopause, yes/no||3.7 (1.9, 7.3)||< 0.001||1.8 (0.5, 6.2)||0.36||0.564†|
In univariate analyses, body height was positively associated with patella bone volume (Table 4). After adjusting for potential confounders including age, body height, and weight, the association between age and patella bone volume became significant (P = 0.04). Body height also remained significant in the multivariate model that included age, body height, and weight. The multivariate model that included age, body height, and weight accounted for 27.5% of the variance in patella bone volume. When BMI was substituted for body height and weight in the above model including age, these variables accounted for only 2.1% of the variation in patella bone volume. No association was found between smoking status, level of physical activity, or menopausal status and patella bone volume in either the univariate or multivariate analyses.
|Univariate regression coefficient||P||Multivariate regression coefficient||P||R2 for model|
|Age, years||17.7 (−45.0, 80.1)||0.57||57.3 (3.0, 112.1)||0.039||0.275†|
|Body height, cm||216.1 (161.1, 271.2)||< 0.001||227.1 (171.0, 283.1)||< 0.001||0.275†|
|Weight, kg||12.0 (−17.4, 41.4)||0.4||−1.4 (−27.1, 24.0)||0.91||0.275†|
|BMI, kg/m2||−69.3 (−144.3, 5.6)||0.07||−71.1 (146.3, 4.2)||0.06||0.021‡|
|Smoking status, yes (n = 176)||−209.3 (−1,482.1, 1,063.5)||0.7||84.1 (−1,014,1, 1,182.1)||0.88||0.275†|
|Alcohol||1,202.3 (204.6, 2,200.1)||0.02||927.6 (50.3, 1,804.9)||0.04||0.274†|
|Light exercise, yes/no||−393.3 (−3,581.0, 2,794.5)||0.81||−555.4 (−3,314.1, 2,203.2)||0.69||0.275†|
|Strenuous activity, yes/no||415.1 (−586.4, 1,416.1)||0.41||643.1 (−233.1, 1,518.0)||0.15||0.283†|
|Menopause||17.0 (821.4, 854.5)||0.97||−760.1 (−1,885.3, 366.2)||0.18||0.279†|
In this cross-sectional study of 176 healthy women, we found that in relation to age there was a decrease in patella cartilage volume (independent of bone volume), an increase in patella cartilage defects (independent of patella cartilage volume), and an increase in patella bone volume. Increasing BMI was associated with a decrease in patella cartilage volume and an increase in patella cartilage defects, and had a trend towards association with a decrease in patella bone volume. Smoking status had no effect on patella cartilage volume or bone, but there was a trend towards increasing alcohol consumption being associated with an increase in patella bone volume. There was no significant effect of physical activity on patella cartilage or bone, although there was a trend towards an increase in patella cartilage defects with increasing strenuous physical activity. Similarly, there was a trend towards a reduction of patella cartilage volume with menopause.
In this study, age was negatively associated with patella cartilage volume but positively associated with patella cartilage defects. This is consistent with a recent study by Ding et al that demonstrated that age was negatively associated with patella cartilage volume but not tibial cartilage volume, and that patella cartilage defects increased with age (25, 26). An autopsy study previously reported that patella cartilage in women age >50 years showed progressive thinning with increasing age, which was more obvious in women than men (27). Indeed, patella cartilage volume has been shown to be lost at 4.5% per year in persons with established OA (7). The present study suggests that even in the absence of disease, it is likely that patella cartilage volume may also be lost with the passage of time. Whether there is a discrepancy in the rate of cartilage volume reduction between healthy and diseased states is yet to be examined.
This study demonstrated that age is positively associated with patella bone volume. Few studies have examined the association between age and patella bone volume in individuals without OA. One study that examined subjects with features of OA demonstrated that both medial and lateral tibial surface bone area as well as patella bone volume were positively associated with age. However, these associations decreased in magnitude after adjustment for radiographic OA, suggesting that the generalizability of these results is limited to individuals with OA only. Our results are the first to demonstrate that age is positively associated with patella bone volume in persons without OA. Moreover, given that we have also demonstrated that patella cartilage defects are increased and volume is reduced with age, it is possible that the enlarged surface volume of bone combined with the reduced cartilage volume may predispose exposed cartilage to fissuring, which cannot be repaired, initiating or facilitating the process of OA. The rate of change in these bony and cartilaginous changes in both healthy and diseased states is of major interest.
Our results demonstrated that both patella cartilage volume and bone volume are inversely associated with BMI and that cartilage defects increase with increasing BMI. The association between BMI and cartilage defects is consistent with the study by Ding et al (26). To our knowledge, no other study has examined the association between patella bone volume and BMI in the absence of OA. Although our study is the first to examine and report an inverse relationship between patella bone volume and BMI, one other study reported that tibial bone enlargement occurs with increasing BMI. Given that the direction of the documented associations between BMI and bone volume differs between the patella and tibia, it is possible that the effect of BMI on different bony volumes (e.g., patella versus tibial) are mechanistically different. Indeed, the knee adduction moment, which exerts a compressive force in the frontal plane and is a major determinant of medial tibial load, has been previously found to be associated with bony enlargement of the medial tibia (28), and added body mass is likely to amplify such forces. It is unlikely that the knee adductor moment exerts such an influence on the patella bone. This highlights the fact that the forces applied to the tibia are markedly different from those applied to the patella, and this may be important in mediating the different association between BMI and bone volumes. This may also imply that the pathogenic mechanisms for OA of the tibiofemoral and patellofemoral joint may differ, particularly from a biomechanical perspective.
Previously, the only other study that had examined the association between BMI and patella cartilage properties described a negative association between BMI and patella cartilage thickness, but not volume (29). A study that examined the association between BMI and cartilage volume at other sites within the knee joint, such as the tibial cartilage volume, noted an inverse association between BMI and medial and lateral tibial cartilage volume in healthy men (30). Our study is the first to examine and support an inverse association between BMI and patella cartilage volume. Given that both patella and tibial cartilage volume are inversely associated with BMI, it is possible that the effect of added body mass on cartilage volume is mediated by systemic/metabolic factors rather than biomechanical factors only. This is speculative but is suggested by the consistent direction of the association between cartilage volume and BMI at both the patella and tibial sites, despite the considerable differences in the joint biomechanics between the 2 sites (e.g., the tibia is a relatively greater weight-bearing structure compared with the patella).
We found that strenuous physical activity had no effect on patellar cartilage volume. One previous study that compared 7 professional weight lifters with 7 sprinters and 14 untrained volunteers reported that patella cartilage deformation showed a “dose-dependent” response, where more intense loading led to greater deformation (31). This finding suggests that adult patella cartilage may be amendable to training effects in vivo. It may be that the type of activity measured by our questionnaire, which focused on general physical activity, did not capture the type of activity that is most likely to influence patella cartilage or bony properties. For example, as the knee flexes to 15 degrees at initial contact during walking, the patellofemoral joint reaction force is reportedly 50% of the total body weight, while at 60 degrees knee flexion, the retropatellar force may increase to 3.3 times the total body weight (32). Therefore, it is possible that individuals who participate in activities requiring deep knee flexion, such as weightlifting, may have altered patellofemoral joint cartilage and bone. Future studies are required to examine these issues.
The risk factors age and BMI were related to the presence of patella defects in the same way as they are related to patella cartilage volume. These findings suggest that obesity and older age are related to an increased risk of patellofemoral OA, possibly due to the deterioration of cartilage health in the patellofemoral compartment. Strenuous physical activity showed a trend towards statistical significance for patella cartilage defects, but light activity did not. This finding may indicate that laborious physical pressure on the joint may affect the health of the cartilage in that joint.
In this study, we found no effect of smoking status on patella cartilage or bone, but there was a trend towards increasing alcohol consumption being associated with an increase in patella bone volume. This finding is consistent with a finding by Williams et al that moderate alcohol consumption is not harmful to bone health in women and may even be beneficial (33). Whether this then impacts the risk of patellofemoral OA is unclear. However, interestingly, in this study we also found an indication that alcohol consumption may be associated with a reduced risk of cartilage defects (OR 0.4, P = 0.09). Further work is needed in this area.
Our study was limited by the examination of healthy middle-aged women only. As such, we can only generalize these results to healthy, middle-aged women, and whether similar associations are apparent in men and persons with OA is unknown. In this study, we classified level of physical activity based on the exercise the women had done in the last 2 weeks. It has previously been shown that levels of physical activity are stable in women over the course of a year and that recall of physical activity over 12 months is poor (34). For these reasons we used recall of physical activity in the last 2 weeks because this is likely to reflect activity in the medium term in this population. A strength of this study is that we excluded subjects with a significant injury requiring surgery or non–weight-bearing treatment for more than 24 hours. However, we cannot exclude the possibility that some of our findings may relate to minor degrees of trauma. Moreover, in this study we used MRI to assess knee cartilage volume in healthy asymptomatic subjects. MRI allows for early detection of change in cartilage, in contrast to radiologic measures of OA that require loss of at least 13% of cartilage before any radiologic OA is identified (35). Subsequently, MRI is a more sensitive measure of assessing cartilage volume than plain radiologic assessment. To confirm these findings and to establish a cause and effect picture, longitudinal studies are necessary.
This is the first study to demonstrate that age is positively associated with patella bone volume and cartilage defects while inversely associated with patella cartilage volume in middle-aged women without knee OA. Moreover, this is the first study to demonstrate that added body mass is inversely associated with both patella cartilage volume and bone volume and positively associated with patella cartilage defects. The role of these associations in the pathogenesis of patellofemoral OA will require longitudinal examination. Moreover, whether these associations are consistent for other joint sites such as the tibiofemoral joint is unclear, and it is probable that the mechanisms influencing joint morphology at the patellofemoral joint are different from those affecting the tibiofemoral compartments.
Dr. Cicuttini had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Davis, Cicuttini.
Acquisition of data. Hanna, Davis, Cicuttini.
Analysis and interpretation of data. Hanna, Wluka, Teichtahl, Cicuttini.
Manuscript preparation. Hanna, Bell, Davis, Wluka, Teichtahl, Cicuttini.
Statistical analysis. Hanna, Bell, Cicuttini.
We wish to acknowledge the support of Roy Morgan International for their assistance with the recruitment of women to the original cross-sectional study from which women for this study were recruited.