There is evidence that knee cartilage volume reduces over time for both healthy people and for those with diseases, such as osteoarthritis.1 2 Despite obesity being the most important modifiable osteoarthritis risk factor for all compartments within the knee, including the patellofemoral joint,3^–6 there is a paucity of studies examining whether obesity is associated with the rate of cartilage volume loss for people with or without established osteoarthritis.

Studies that have examined the obesity–osteoarthritis relationship have predominantly used either the body mass index (BMI) or weight (kg) as a surrogate measure of obesity.4 7^–11 Neither of these measures of obesity account for body composition and cannot directly discriminate adipose mass.12 It remains unknown whether the obesity–osteoarthritis relationship is mediated by adipose mass, the distribution of adipose mass such as central adiposity, or a combination of both. We recently demonstrated in a study of healthy adults that women, but not men, with greater body mass, particularly fat mass at baseline, had an associated reduction in their patella cartilage volume 10 years later.13 Whether measures of obesity, such as BMI and weight, or measures of adiposity such as fat mass and percentage of fat mass, are determinants of the rate at which patella cartilage volume is lost has not yet been examined.

Our aim is to determine using a longitudinal study whether surrogate measures of obesity and direct measures of adiposity are associated with the rate of patella cartilage volume loss for healthy adults with no clinical knee osteoarthritis.

METHODS

Subjects

The study was conducted within the Melbourne Collaborative Cohort Study (MCCS), a prospective cohort study of 41 528 community-based residents of Melbourne, Australia, aged 40–69 years at recruitment (1990–4). The aim of the MCCS was to examine the role of lifestyle and genetic factors in the risk of cancer and chronic diseases from middle age and beyond.14 We invited subjects who attended the first year of round 3 follow-up of the MCCS, which commenced in 2003, provided they were aged between 50 and 79 years and met none of the following exclusion criteria: a clinical diagnosis of knee osteoarthritis as defined by the 1986 American College of Rheumatology clinical criteria;15 knee pain lasting for more than 24 h in the past 5 years; a previous knee injury requiring non-weight-bearing treatment for more than 24 h or surgery (including arthroscopy); or a history of any arthritis diagnosed by a medical practitioner. A further exclusion criterion was any contraindication to magnetic resonance imaging (MRI). We also had data pertaining to participation in physical activity between 1990 and 1994. This was derived from questionnaire data at the induction of the MCCS, whereby a subject was asked whether they exercised vigorously (leading to dyspnoea or diaphoresis) for a period of at least 20 minutes at least once a week. We used quota sampling whereby we stopped recruitment when we reached our target sample of approximately 300 subjects.

The study was approved by the Cancer Council Victoria Human Research Ethics Committee and the Standing Committee on Ethics in Research Involving Humans of Monash University. All participants gave written informed consent.

Anthropometric measures

In 2003–4, height and weight were measured using standardised written protocols.16 BMI (weight/height, kg/m²) was calculated from these data, which were again collected at follow-up (2006–7).

Bioelectrical impedance analyses were performed in 2003–4 and again in 2006–7 with a single frequency (50 kHz) electric current produced by a BIA-101A RJL system analyser (RJL Systems, Detroit, Michigan, USA). Resistance and reactance were measured with subjects in the supine position. Non-adipose mass, hereafter termed fat-free mass, was estimated as 9.1536 + (0.4273 × height²/resistance) + (0.1926 × weight) + (0.0667 × reactance) for men and 7.7435 + (0.4542 × height²/ resistance) + (0.119 × weight) + (0.0455 × reactance) for women.17 Adipose mass, hereafter termed fat mass (weight − fat-free mass), and percentage of body fat (fat mass/weight) were subsequently calculated.

MRI and the measurement of patella cartilage volume

In 2003–4 and again in 2006–7, each subject had MRI performed on their dominant knee (defined as the lower limb the subject used to step off from when initiating walking). Imaging was performed in the morning to negate any potential influence that diurnal variation may have on articular structures of the knee. Immediately before MRI, each subject performed similar physical activity, with normal walking preceding the scan, followed by rested non-weight bearing. Knees were imaged in the sagittal plane on a 1.5-T whole-body magnetic resonance unit (Phillips Medical Systems, Eindhoven, The Netherlands) using a commercial transmit–receive extremity coil. The following sequence parameters were used: a T1-weighted fat-suppressed three-dimensional gradient recall acquisition in the steady state; flip angle 55°; repetition time 58 ms; echo time 12 ms; field of view 16 cm; 60 partitions; 512 × 512 matrix; one acquisition time 11 minutes 56 s. Sagittal images were obtained at a partition thickness of 1.5 mm and an in-plate resolution of 0.31 × 0.31 mm (512 × 512 pixels).

Patella cartilage and bone volumes were determined by image processing on an independent workstation using the Osiris software (University of Geneva). Contours were drawn around the patella on images 1.5 mm apart on sagittal views. The coefficients of variation were 2.1% for patella cartilage volume and 2.2% for patella bone volume.18 The reproducibility of patella cartilage volume measurements when the same subject was imaged a second time, approximately 1 week later, yielded a coefficient of variation of 2.6%. Assessors were blinded to time (ie, 2003/4 or 2006/7 data) and obesity and adiposity data.

Annual change of patella cartilage volume was determined from the equation:

Statistical analyses

With 271 subjects in the study, we had 80% power to demonstrate a correlation as low as 0.2 (alpha error 0.05, two-sided significance) between measures of baseline obesity and change in patella cartilage volume. Outcome was the annual change in patella cartilage volume between 2003–4 and 2006–7. Annual change in patella cartilage volume was assessed for normality before linear regression analyses. Multiple linear regression models were constructed to examine the relationship between measures of obesity (at baseline and change over the study period) and the change in patella cartilage volume, adjusting for potential confounders. In terms of measures of obesity as predictors of patella cartilage volume, we analysed baseline parameters and the change in these parameters (follow-up values − baseline values) separately. Given the significant gender differences in body composition (see table 1), all analyses were performed for the total population (adjusting for gender) and repeated separately for men and women. To determine whether the difference between regression coefficients in men and women was statistically different, the difference in the coefficients was divided by the estimated standard error of the difference and the result compared with a standard normal distribution table. p Values of less than 0.05 were considered to be statistically significant. All analyses were performed using the SPSS statistical package (standard version 15.0).

RESULTS

A total of 297 subjects entered this study at baseline (2003/4) for MRI analyses, of which 233 had body composition analyses performed (fat mass and percentage fat mass). Of the 297 original subjects at baseline, 271 (62% female) completed follow-up in which MRI and body composition analyses were performed (table 1). Twenty-six subjects, nine men and 17 women, were lost to follow-up because of death (n  =  3), poor health (n  =  4), withdrawal of consent (n  =  10), or had developed a contraindication to MRI (pacemaker) (n  =  4) and were unable to be contacted (n  =  5). Some 37.3% (n  =  101) of subjects completing this study had a history of participating in physical activity causing dyspnoea or diaphoresis. The mean time between baseline and follow-up MRI scans was 2.3 years (SD 0.4). The subjects lost to follow-up had greater fat mass (p = 0.03) and tended to have a higher percentage of fat mass (p = 0.07) and larger BMI (p = 0.06) than those who completed the study. Between baseline and follow-up, subjects tended to have a significant increase in their percentage and absolute measures of fat mass (p = 0.01 and p = 0.07, respectively), but did not have a significant change in their weight or BMI (p = 0.82 and p = 0.85, respectively). Patella cartilage volume was lost at a rate of 1.8% per annum (95% CI 1.4% to 2.1%) and there was a significant reduction in patella cartilage volume between baseline and follow-up (p<0.001).

All baseline measures of obesity (BMI, weight) and adiposity (fat mass and percentage fat mass) were associated with a higher annual rate of patella cartilage volume loss before (all p⩽0.06) and after adjustment for age, gender and baseline patella cartilage and bone volume (all p⩽0.04; table 2). When analysing men and women separately, the association between obesity and adiposity measures and the rate of patella cartilage volume loss were significant for women only (table 2). There were, however, no significant differences between the multivariate regression coefficients between men and women when analysing the association between baseline measures of obesity and change in patella cartilage volume (data not shown).

There were no significant associations between change in any of the measures of obesity and adiposity and the annual rate of patella cartilage volume loss before or after adjustment for potential confounders for the total population, or when men and women were analysed separately (data not shown). Moreover, there were no significant differences between the multivariate regression coefficients between men and women when analysing the association between change in measures of obesity and adiposity and change in patella cartilage volume (data not shown).

DISCUSSION

Baseline obesity (BMI and weight) and adiposity (fat mass and percentage of body fat) were associated with an increased annual rate of patella cartilage volume loss over 2 years in healthy adults with no clinical knee osteoarthritis.

No previous study has examined whether measures of obesity or adiposity are associated with the rate of loss of patella cartilage volume. A previous cross-sectional study demonstrated that BMI was negatively associated with patella cartilage thickness and volume in people without knee osteoarthritis,19 and we recently demonstrated that increased obesity and adiposity measures were associated with reduced patella cartilage volume 10 years later in healthy women, but not men.13 Several studies have previously documented a female predisposition towards a reduction in cartilage volume that is associated with greater weight.20 21 In this study, we found that the magnitude and direction of the associations with weight and obesity were similar for men and women, but results remained significant for women only. These gender differences may be explained, at least partly, by the reduced sample size of men in this cohort. However, it may also be that this effect is weaker in men given that women demonstrated significantly greater adipose mass, which may ultimately be a contributing factor towards the female predisposition to knee osteoarthritis.

We did not observe any significant associations between change in obesity or adiposity measures and change in patella cartilage volume over the same time period. This result, however, does not necessarily support the notion that a gain in obesity or adiposity is not detrimental to the health of articular patella cartilage. Although the duration of this study was not long enough to observe a significant increase in obesity measures (weight (p = 0.82) or BMI (p = 0.85)), it was of sufficient duration to observe a trend towards significant increases in adiposity measures (fat mass (p = 0.07) and percentage of fat (p = 0.01)). Whereas a gain in either obesity or adiposity measures does not appear to confer any additional effect on the rate of loss of patella cartilage volume over the same period, independent of baseline measures, increased baseline measures of obesity or adiposity are significant determinants of the rate at which patella cartilage volume is lost. Taken together, these results may be a consequence of the very long-term effect of risk factors such as obesity on cartilage in an otherwise healthy population, whereby the detrimental effect of current gain in obesity measures may be reflected in the longer term.

How baseline obesity and adiposity increase the rate of loss of patella cartilage volume is unclear, but is likely to result from a combination of excess mass and hormonal factors. Although it is recognised that some degree of mechanical stimulation is required to maintain cartilage health,22 increased loading imparted by excessive body mass may expedite cartilage degeneration. Alternatively, dysregulation of lipid homeostasis may be crucial in mediating the deleterious changes that occur at the patella cartilage in the obese. Adipose tissue was previously thought to be a passive store of energy, but is now considered an endocrine organ, releasing various factors, including cytokines such as tumour necrosis factor and IL-1, as well as adipokines, such as leptin, adiponectin and resistin.23 Of these, IL-1 and tumour necrosis factor alpha, have recently been shown to play a key role in cartilage destruction in osteoarthritis.24 Moreover, leptin receptors have been found at articular cartilage,25 and significant levels of leptin were observed in the cartilage and osteophytes of people with osteoarthritis.26 Nevertheless, the potential for these factors to contribute towards accelerated cartilage destruction in health or disease is unclear. Interventions targeting a reduction in measures of obesity may potentially help to reduce the risk of the onset and/or progression of patellofemoral osteoarthritis. Indeed, at the tibiofemoral joint, it had been shown that women who had decreased their BMI by 2 units or more over 10 years decreased the odds of developing knee osteoarthritis by over 50%.4 Whether similar patterns are apparent at the patellofemoral joint requires further investigation.

We examined asymptomatic people and used novel and sensitive measures to assess the rate of change of patella cartilage volume. However, we did not obtain knee radiographs and some subjects may have had asymptomatic, radiographic osteoarthritis. Nevertheless, we have excluded people with clinical knee osteoarthritis based on the American College of Rheumatology criteria, as well as excluding individuals with significant knee injury in the past, pain at baseline, knee surgery or a medical diagnosis of any other type of arthritis. Moreover, the radiographic grade of patellofemoral osteoarthritis is correlated with patella cartilage volume,27 and we have adjusted for baseline patella cartilage volume in the analyses. Nevertheless, an increased annual rate of cartilage loss may have occurred due to other extraneous and unadjusted factors, such as a period of sustained prolonged non-weight bearing following new-onset paraplegia, or forced immobility following ankle fracture, which have been shown to be associated with cartilage volume reduction.28 29 This is unlikely in this study, because participants were healthy subjects with no clinical knee pathology precluding them from weight-bearing tasks and rendering them immobile at the time of MRI. In addition, we demonstrated that measures of obesity and adiposity were significantly associated with the rate of patella cartilage volume loss independent of participation in physical activity leading to dyspnoea or diaphoresis. We also examined change in obesity and adiposity measures and although we demonstrated significantly increased adiposity measures at follow-up compared with baseline, we failed to observe a significant change in obesity measures. It may have been that an average follow-up time of 2.3 years was too short to observe significant changes in weight or BMI. Finally, those people who did not complete the study tended to have greater measures of obesity and adiposity than those who did. It is not surprising that many of the reasons subjects were lost to follow-up were related to poor health status (eg, death, exclusion from MRI due to pacemaker). Had these subjects been able to complete the study, the strength of the associations between baseline measures of obesity and adiposity and the rate of patella cartilage volume loss may have been strengthened.

We have demonstrated that increased baseline measures of obesity (BMI and weight) and adiposity (fat mass and percentage fat mass) are associated with an increased annual rate of patella cartilage volume loss over 2 years in healthy adults with no clinical knee osteoarthritis. Weight-loss interventions that reduce body mass, or specifically target a reduction in fat mass, may help to reduce the rate at which patella cartilage volume is lost and subsequently the risk of patellofemoral osteoarthritis.