In asymptomatic community-based adults, we have demonstrated that increased anthropometric height measures are associated with increased tibial bone area, while knee height as a percentage of body height is associated with a reduced risk of medial tibial cartilage defects. We have also found that knee height is associated with knee cartilage volume. The associations between increased anthropometric measures and increased bone area may simply reflect inherently larger bony structures. However the beneficial associations demonstrated with cartilage morphology suggest that an increased knee height may confer a beneficial biomechanical environment to the chondrocyte of asymptomatic adults.
In the only previous study to have examined knee height, radiographic knee OA was associated with increased knee height among Beijing residents aged 60 years or older [1
]. According to the Kellgren-Lawrence grading system, the diagnosis of radiographic OA is heavily reliant on the presence of osteophytes to classify disease. The role of osteophytes in disease pathogenesis remains unclear, but much of the previous study's association between knee height and radiographic OA [1
] may have been mediated by the presence of osteophytes. To further explore this concept, we found that knee height tended to be significantly associated with an increased risk for the presence of MRI osteophytes (OR 1.2; 95% CI 1.0 - 1.5; p
= 0.06). MRI has been shown to be more sensitive than the joint radiograph at determining the presence of osteophytes [12
]. Furthermore, we demonstrated that increased knee height is associated with increased tibial bone size, which supports knee height being inherently linked to bone morphology and may not necessarily be detrimental to other structures, such as cartilage. This leads to the quandary of determining whether increased knee height is to the detriment of the knee joint. In this study, we substantiate Hunter et al. results by demonstrating a possible deleterious association, particularly with the osteophyte [1
], a key criterion for the diagnosis of radiographic OA.
Nonetheless, despite the deleterious association with the knee osteophyte, we have demonstrated that an increased knee height is associated with increased knee cartilage volume, although this relationship did not persist when knee height as a percentage of body height was examined. This indicates that although knee height is associated with cartilage volume, the relationship is attenuated when the overall stature of an individual is accounted for. In contrast, knee height as a percentage of body height (but not isolated knee height) was associated with a reduced risk of medial tibial, but not lateral cartilage defects. Why isolated knee height is associated with cartilage volume, and knee height as a percentage of body height is associated with reduced medial cartilage defects, but not vice versa, is unclear. However, the direction of these results is consistent and infer that increased knee height (either isolated, or relative to total body height) is associated with beneficial cartilaginous properties (increased cartilage volume and reduced cartilage defects) at the knee, suggesting a protective biological effect. Moreover, the medial compartment specific association between knee height as a percentage of body height and cartilage defects further substantiates an underlying biomechanical mechanism, since knee joint loads are predominantly directed medially [15
]. Cartilage defects are surface lesions which, independent of cartilage volume, predict cartilage loss and pain in both people with and without knee OA [16
]. Why increased joint load, which is speculated to result from an increase in knee height, either alone or as a percentage of body height, benefits the cartilage of asymptomatic people is unclear. It may be that healthy articular cartilage relies upon a certain degree of mechanical stimulation. In childhood, cartilage accrual is greater in physically active children [22
], while in adults, forced immobility results in a rapid decline in knee cartilage [23
]. Mechanical stimulation may therefore be imperative in maintaining cartilage health, although mechanocellular mechanisms may be easily perturbed when disease processes are activated by other means (e.g. in the setting of obesity).
There are a number of other factors that may also account for the differences between the results of our MRI study and the previous radiographic study. In contrast to our population which was Caucasian, the previous study examined Beijing residents. The Chinese have been demonstrated to have more valgus alignment of the distal femur than Caucasians [25
]. Biomechanically, changes in varus-valgus alignment of the lower-limb can ameliorate the external knee adductor moment, which is the major determinant of joint load distribution at the knee [15
]. In contrast to the previous study, we have adjusted for knee alignment. Moreover, we have examined a cohort of people without established knee OA. The disease status of a joint health may be a key determinant of how the chondrocyte responds to external loads. Similarly, articular structures may respond differently to the same stimuli over the life-span. Whereas the previous study of Beijing residents examined people aged 60 or over [1
], we have examined a younger population with only 2.9% being aged 60 or over. In addition, we also examined knee height as a proportion of total body height, which may explain further differences in the findings.
This study has several limitations. Although we excluded people with a diagnosed arthropathy, we have not adjusted for the possibility of radiographic OA, despite adjusting for osteophytes. Since we did not have knee radiographs we determined the presence of osteophytes from MRI. This has previously been shown to be a more sensitive method for determining the presence of osteophytes than radiography [26
]. Moreover, our main findings related to cartilage volume and defects were independent of bone area and cartilage volume respectively, both of which are strongly associated with radiographic knee OA [27
]. Directly assessing cartilage volume is more sensitive than radiography for detecting early OA, since more than 10% of cartilage is already lost before any radiographic OA is detected [27
]. Moreover, since our relatively small (n = 89) cohort was predominantly female (82%), recruited in part, from weight-loss clinics (mean BMI 32 kg m-2
), we were limited in our ability to perform subgroup analyses based on gender. Additionally, we could not meaningfully analyse other subgroups, such as obesity classes, which impacts the generalisability of our findings. Nonetheless, we have limited the confounding effect of gender and weight, by adjusting for these in multivariate analyses. Finally, as this is a cross-sectional study, longitudinal data is needed to confirm the protective effect of an increased knee height on cartilage in those without knee OA.