To investigate the association of being overweight or obese with the presence, extent, and severity of lumbar disc degeneration on magnetic resonance imaging (MRI) in adults.
Wiley Online Library is migrating to a new platform powered by Atypon, the leading provider of scholarly publishing platforms. The new Wiley Online Library will be migrated over the weekend of February 24 & 25 and will be live on February 26, 2018. For more information, please visit our migration page: http://www.wileyactual.com/WOLMigration/Remove maintenance message
To investigate the association of being overweight or obese with the presence, extent, and severity of lumbar disc degeneration on magnetic resonance imaging (MRI) in adults.
A population-based cross-sectional study of 2,599 southern Chinese volunteers was conducted. Subjects underwent radiographic and clinical assessment, and weight and height were measured. Sagittal T2-weighted MRIs of the lumbar spine were obtained. The presence, extent, and severity of disc degeneration and additional radiographic and clinical parameters were assessed. Asian-modified body mass index (BMI) (kg/m2) categories were used. Odds ratios (ORs) and 95% confidence intervals (95% CIs) were calculated.
The study included 1,040 men and 1,559 women (mean age 41.9 years). Disc degeneration was noted in 1,890 subjects (72.7%). BMI was significantly higher in subjects with disc degeneration (mean 23.3 kg/m2) than in subjects without degeneration (mean 21.7 kg/m2) (P < 0.001). A significant increase in the number of degenerated levels (P < 0.001), global severity of disc degeneration (P < 0.001), and end-stage disc degeneration with disc space narrowing (P < 0.001) was noted with elevated BMI, in particular in overweight and obese subjects. In the adjusted multivariate logistic regression model, there was a positive linear trend (r2 = 0.99) between BMI and the overall presence of disc degeneration in overweight (OR 1.30 [95% CI 1.03–1.62]) and obese (OR 1.79 [95% CI 1.17–2.74]) subjects. End-stage disc degeneration with disc space narrowing was significantly more pronounced in obese subjects (adjusted OR 1.72 [95% CI 1.23–2.41] [reference normal weight]).
Our findings, in one of the largest studies to systematically assess lumbar disc degeneration on MRI, indicated a significant association between the presence, extent, and global severity of disc degeneration with weight in overweight and obese adults.
Low back pain is a common debilitating condition worldwide with severe socioeconomic and health care consequences (1, 2). Low back pain can cause functional impairment, diminished quality of life, loss of working ability, potential psychological distress, and increased health care costs (1–3). Although various factors have been implicated in low back pain, intervertebral disc degeneration that is evident radiologically or on advanced imaging (i.e., magnetic resonance imaging [MRI]) is a known cause of low back pain (4–13). In a recent systematic review of MRI studies assessing disc degeneration and low back pain, Chou et al (5) found that disc degeneration is a significant risk factor for the development of such pain. In fact, previous studies have also shown that an increase in the global severity of disc degeneration or the presence of end-stage disc degeneration with disc space narrowing is significantly associated with an increased risk of having low back pain (6, 10, 14).
For many years, a strong emphasis has been placed on identifying risk factors associated with disc degeneration in the lumbar spine in adults. Age, abnormal physical loading (15), and environmental (16, 17), hormonal (18), systemic (19–21), and genetic influences (16, 22–25) have been suggested to contribute to the development of disc degeneration. Some studies have suggested that elevated body mass index (BMI), in particular being overweight or obese, may be related to disc degeneration (17, 26, 27), but such an association remains largely a matter of speculation to date, with several studies providing strong evidence against such a link (6, 28, 29). These inconsistencies can be attributed primarily to the lack of large epidemiologic studies and proper study design, patient-based studies, insufficient statistical analyses, the mode of radiographic/imaging assessment used in defining the phenotype of disc degeneration, and/or conjecture arising from limited radiographic interpretation of additional spinal findings (e.g., Schmorl's nodes) that may contribute to the degenerative process (6, 17, 26, 28–30). Furthermore, the associations between being overweight or obese and the extent of disc degeneration (i.e., the number of lumbar levels with disc degeneration) and the severity of disc degeneration of the lumbar spine remain unknown because previous studies have failed to quantitatively assess such parameters on advanced imaging.
The numbers of overweight and obese individuals are of global concern, and the prevalence continues to rise in many populations (31, 32). Overall, 23.2% and 9.8% of the world's adult population are overweight and obese, respectively (32), and it is expected that such rates will increase exponentially by the year 2030 if left unabated. In fact, based on epidemiologic trends seen in the past (e.g., bubonic plague, pneumonia and infections, cardiovascular disease and cancer, etc), society has entered the obesity phase (31). In the US, it has been estimated that 1 of 3 children are obese, and the likelihood that excess body weight will remain and even lead to more severe obesity in adulthood is high (33). Several European countries (e.g., the UK, Germany, and Croatia) have reported that >60% of their population is overweight. Although it was previously believed that the numbers of overweight and obese individuals did not represent a public health concern among Asian countries because of low prevalence, China, the world's most populous nation and second largest economy, has seen an increase in the numbers of overweight and obese residents due to its financial affluence, the rise of the fast food culture, and the adoption of more Westernized lifestyles, which has affected both urban and rural communities (34). Studies have shown that the prevalence of being overweight or obese in Chinese children and adolescents has tripled from 1991 to 2006 and continues to rise (34).
Although body weight has been shown to be associated with the development of cardiovascular disease, diabetes, and malignancies among other conditions, its effects on disc degeneration, as mentioned above, have remained elusive. Since being overweight, and particularly, obese, have been associated with low back pain (35) and since disc degeneration on MRI is a factor related to low back pain (8, 10, 11), it would appear reasonable to hypothesize that elevated BMI may be instrumental in the development of disc degeneration.
Since MRI is widely regarded as the gold standard of imaging modalities to assess disc degeneration, the Hong Kong Degenerative Disc Disease Cohort Study, composed of southern Chinese subjects, was initiated to address the phenotype of disc degeneration and other spinal findings on MRI (4, 11, 23, 25, 36, 37). To date, this initiative has become the largest population-based study in the world to systematically address radiographic, clinical, and genetic elements related to the spine. Based on findings in patients in this cohort, the authors have recently noted that being overweight or obese is associated with disc degeneration in adolescents (11). The present study addressed the role of BMI, in particular a BMI that indicates being overweight or obese, in the presence, extent, and severity of lumbar disc degeneration in Chinese adults.
Following institutional review board approval, a population-based cohort study was initiated to assess the phenotype of disc degeneration (4, 11, 23, 25, 36–39). Southern Chinese subjects from an overall population of 7.2 million were recruited between 2001 and 2009 by open invitation through newspaper advertisements, e-mails, and posters throughout the university system and communities. The invitation to participate did not discriminate with regard to social or economic demographics. Participants were not recruited based on the presence or absence of low back pain. However, in an effort to assess for selection bias for low back pain in our sample, microsatellite marker frequencies in 10 markers from different chromosomes in 96 individuals were randomly selected. This information was then compared to the same set of microsatellite marker frequencies in 96 university students (23). The outcome of this comparison demonstrated that the sample was representative of the general population. The study sample and recruitment procedures have been described in further detail previously (4, 11, 23, 25, 36–39). A cross-sectional design was used to assess 2,599 individuals who were ≥21 years of age (range 21–63 years). The participants did not have infections, inflammation, neoplasms, deformities, or previous surgery of the lumbar spine. None of the participants were related.
Each participant underwent T2-weighted MRI of the lumbar spine. The technical parameters of the imaging protocol have been described previously (4). The presence and severity of disc degeneration was assessed using the criteria of Schneiderman et al (40) (Table 1), in which higher scores indicate increased severity. A total disc degeneration score was calculated for each individual to assess “global” disc degeneration severity in the overall lumbar spine (4, 23). This total disc degeneration severity score was referred to as the degenerative disc disease score (4, 23). Disc degeneration of the lumbar spine was considered to be present if any level exhibited a degenerative disc disease score of ≥1. The presence of and number of levels with end-stage disc degeneration with disc space narrowing (i.e., a degenerative disc disease score of 3 at any lumbar level) was noted. The extent of disc degeneration was defined as the number of lumbar segments affected and multilevel involvement (i.e., ≥2 levels). The presence and severity of disc herniation (i.e., bulge and extrusion), Schmorl's nodes (i.e., focal morphologic abnormality or defect of the vertebral end plate), high intensity zones of the disc, and vertebral marrow changes (i.e., high signal intensity of the vertebral body) were noted. Two physicians (JK and KMCC) who specialized in the diagnosis and management of spine-related disorders, and who were trained in the assessment of degenerative spine findings on MRI, independently assessed the radiographs in a blinded manner.
|Normal||No signal changes|
|1||Slight decrease in signal intensity of the nucleus pulposus|
|2||Hypointense nucleus pulposus with normal disc height|
|3||Hypointense nucleus pulposus with disc space narrowing|
Age, sex, physical activity level (exercising <2 days per week versus ≥2 days per week), history of lumbar injury (i.e., an event that led to low back pain), history of smoking, occupation, weight (kg), and height (meters) were obtained for all participants. Physical workload was determined by occupation and characterized as being light, medium, moderate, or heavy and scored for intensity on a scale of 1–4 (41). Objective weight (digital recording) and height data were obtained and recorded at a single institute in a consistent manner for each participant. Subjects removed shoes, excess bulky clothing (such as a jacket), and objects of weight (such as a purse, wallet, or metallic objects) before weight and height measurements were obtained. Height and weight measurements were assessed to the nearest hundredth decimal place and were noted at the time the subject underwent MRI. BMI (kg/m2) categories modified for Asian populations based on World Health Organization guidelines were used (42). Individuals with a BMI of <18.5 kg/m2 were classified as underweight, those with a BMI of 18.5–23.0 kg/m2 as normal, those with a BMI of 23.0–27.5 kg/m2 as overweight, and those with a BMI of >27.5 kg/m2 as obese. This classification scheme was used instead of the standard scheme used in the US or Europe (i.e., <18.5 kg/m2 as underweight, 18.5–25 kg/m2 as normal, 25–30 kg/m2 as overweight, and >30 kg/m2 as obese) to account for ethnic variations in body fat distribution and disease risk, which differ between Asians and Caucasians.
SPSS software, version 17, was used for the statistical analyses. Kappa analysis was performed to assess interrater reliability of the radiographic assessment of disc degeneration and other spinal findings. Kappa statistics >0.90 were considered excellent, 0.80–0.90 were considered good, 0.60–0.80 were considered fair, and <0.60 were considered poor (43). Correlation analysis was performed for continuous variables, and r values of 0.80–1.00 indicated high correlation, 0.60–0.80 indicated a marked degree of correlation, 0.40–0.60 indicated a moderate degree of correlation, and <0.40 indicated a low degree of correlation to no correlation (44). Univariate analysis of various variables in relation to BMI was conducted. Tukey's post hoc pairwise correlation analysis was performed to assess the association of BMI categories with degenerative disc disease score and the number of levels with disc space narrowing. Multivariate logistic regression analysis with adjustments for various covariates was performed. Potential covariates were considered for inclusion in the model if P < 0.20 in univariate analysis. Model selection was performed, interpreted, and adjusted for significant covariates (P < 0.05) in relation to the dependent variable of disc degeneration. Interaction effects were considered. The Hosmer-Lemeshow goodness-of-fit test and Nagelkerke's R2 were used to assess the robustness of the multivariate logistic regression model selected. The odds ratio (OR) was determined as a measure of the strength of association between BMI categories and disc degeneration. With regard to trend association, the best-fit curve was selected based on r2 values. P values less than 0.05 were considered significant, and the corresponding 95% confidence intervals (95% CIs) were determined.
The study included 1,040 men and 1,559 women, with a mean ± SD age of 41.9 ± 9.4 years (range 21–63 years). Excellent interrater reliability was noted (κ >90%). Disc degeneration was present in 1,890 (72.7%) of the subjects. Disc degeneration was seen at L1–L2 in 9.2% of the subjects, at L2–L3 in 16.2% of the subjects, at L3–L4 in 28.9% of the subjects, at L4–L5 in 49.0% of the subjects, and at L5–S1 in 52.0% of the subjects. Men had a significantly higher prevalence of disc degeneration than women (75.6% in men versus 70.8% in women) (P = 0.008). As expected, the prevalence of disc degeneration was found to increase with older age (P < 0.001).
Of the study subjects, 187 (7.2%) were underweight, 1,245 (47.9%) were of normal weight, 937 (36.1%) were overweight, and 230 (8.8%) were obese. The mean ± SD BMI was 22.9 kg/m2 (range 14.2–35.8 kg/m2). Although women had a higher correlation between increasing age and BMI values than men (r = 0.334, P < 0.001 and r = 0.243, P < 0.001, respectively), the correlation between age and BMI in the entire cohort was poor (r = 0.288, P < 0.001) (Figure 1). BMI was significantly higher in subjects with disc degeneration (mean ± SD 23.3 ± 3.2 kg/m2 [range 14.2–35.8 kg/m2]) than in subjects without disc degeneration (mean ± SD 21.7 ± 3.1 kg/m2 [range 15.1–35.6 kg/m2]) (P < 0.001). The mean ± SD number of levels with disc degeneration was 0.8 ± 1.1 (range 0–5) in underweight subjects, 1.4 ± 1.3 (range 0–5) in subjects with normal weight, 1.8 ± 1.4 (range 0–5) in overweight subjects, and 2.0 ± 1.3 (range 0–5) in obese subjects (P < 0.001). Multilevel disc degeneration was noted in 24.6%, 41.4%, 54.9%, and 65.2% of underweight, normal weight, overweight, and obese subjects, respectively. Individuals who were overweight and those who were obese had an increased likelihood of multilevel disc degeneration involvement (P < 0.001).
The overall mean ± SD degenerative disc disease score was 2.8 ± 9.4 (range 0–14). Increasing age was significantly but poorly correlated with degenerative disc disease score among all subjects (r = 0.403, P < 0.001) and was similar in men (r = 0.395, P < 0.001) and women (r = 0.410, P < 0.001) (Figure 1). The mean ± SD degenerative disc disease scores for underweight, normal, overweight, and obese subjects were 1.5 ± 2.1 (range 0–9), 2.5 ± 2.5 (range 0–14), 3.3 ± 2.8 (range 0–14), and 3.8 ± 2.8 (range 0–13), respectively (P < 0.001), indicating an increase in the severity of disc degeneration with increased BMI (Figure 2). Pairwise comparisons assessing degenerative disc disease score showed a significant difference between underweight and normal subjects (P < 0.001) and between normal and overweight subjects (P < 0.001), but not between overweight and obese subjects (P = 0.092) (Figure 2).
End-stage disc degeneration and disc space narrowing (Schneiderman grade 3) was noted in 619 (23.8%) of the subjects. The mean ± SD number of levels with end-stage disc degeneration and disc space narrowing (Schneiderman grade 3) of the lumbar spine was 0.2 ± 0.5 (range 0–2) in underweight subjects, 0.3 ± 0.6 (range 0–4) in normal subjects, 0.4 ± 0.7 (range 0–4) in overweight subjects, and 0.5 ± 0.9 (range 0–4) in obese subjects (P < 0.001) (Figure 3). Pairwise comparisons of the number of levels with end-stage disc degeneration and disc space narrowing (Schneiderman grade 3) showed no significant difference between underweight and normal subjects (P = 0.322), but did show a significant difference between normal and overweight subjects (P < 0.001) and between overweight and obese subjects (P = 0.012) (Figure 3).
In the multivariate logistic regression model adjusted for age, workload, presence or absence of Schmorl's nodes, previous lumbar injury, and vertebral marrow changes, there was a positive linear trend association (r2 = 0.99, y = 0.39x + 0.2) between BMI and the presence of disc degeneration (OR 1 for normal weight [reference], OR 0.59 [95% CI 0.41–0.84] for underweight, OR 1.30 [95% CI 1.03–1.62] for overweight, and OR 1.79 [95% CI 1.17–2.74] for obese) (Figure 4). In a multivariate logistic regression model adjusted for the same covariates, similar findings were noted for the presence of multilevel disc degeneration. In a multivariate logistic regression model adjusted for age, sex, presence or absence of Schmorl's nodes, previous lumbar injury, vertebral marrow changes, and workload, a quadratic trend association (r2 = 0.99; y = 0.13x2−0.40x+1.28) was noted between BMI and the presence of end-stage disc degeneration with disc space narrowing (Schneiderman grade 3) (OR 1 for normal weight [reference], OR 0.97 [95% CI 0.61–1.53] for underweight, OR 1.21 [95% CI 0.96–1.51] for overweight, and OR 1.72 [95% CI 1.23–2.41] for obese) (Figure 5).
Since disc degeneration is related to low back pain and since being overweight or obese is also associated with low back pain, it is reasonable to assume that disc degeneration and low back pain may have increased body weight in common as a risk factor. However, to date, the relationship between body weight and disc degeneration has been a subject of controversy. In a study of 270 elderly Japanese subjects, Hangai et al (17) showed that high BMI values were a risk factor for developing disc degeneration evident on MRI. In their study of 129 middle-aged Finnish men, Liuke et al (26) showed that obesity was associated with the development of disc degeneration.
Conversely, in the Rotterdam Study, which is a population-based study of Dutch subjects, a cross-sectional analysis of 2,819 individuals ≥55 years old who underwent radiography did not show any association between elevated BMI and disc space narrowing (6). In the Chingford Study, a prospective longitudinal assessment of 1,003 elderly women from the UK, a potential trend between elevated BMI values and the development of disc space narrowing was observed on plain radiographs; however, the association was not statistically significant (29). In a cross-sectional analysis of 187 North American subjects from the Framingham Study who underwent computed tomography (CT) to assess spinal degenerative changes, Kalichman et al (30) found a higher prevalence of facet joint osteoarthritis but not disc space narrowing in obese individuals.
More recently, based on their assessment of 44 pairs of male monozygotic twins in the Finnish Twin Cohort who had a difference in body weight of ≥8 kg, Videman et al (28) concluded that being overweight or obese was not associated with disc degeneration, based on quantitative signal variation on MRI. In fact, those authors concluded that greater body mass is “not harmful to the discs” and that it may “delay” disc degeneration. However, that study was limited by its small sample size and by the use of the criteria of an 8 kg difference in weight, which may have created subgroups that were not comparable.
Our study, which included 2,599 southern Chinese subjects ranging in age from 21 to 63 years, demonstrated a significant association between elevated BMI values and the presence of disc degeneration of the lumbar spine on MRI. Furthermore, this is the first study to show that elevated BMI values, in particular BMI values indicating being overweight or obese, are significantly associated with degeneration at a greater number of lumbar levels, the global severity of disc degeneration, and end-stage disc degeneration with disc space narrowing. Notably, we demonstrated a significantly positive linear trend association between BMI and disc degeneration of the lumbar spine, where individuals who were overweight and those who were obese were 1.3 and 1.8 times more likely, respectively, to have disc degeneration than individuals of normal weight. Furthermore, our study showed a significant increase in the likelihood of end-stage disc degeneration with disc space narrowing in obese subjects. End-stage disc degeneration with disc space narrowing (Schneiderman grade 3) is often coupled with degenerative changes throughout the vertebral motion segment (e.g., canal and neuroforaminal stenosis, ligamentous thickening) and altered lumbar kinematics that increase the risk of low back pain (6). This severity of disc degeneration may help explain the increased prevalence of prolonged and chronic low back pain in overweight and obese individuals discussed in a recent systematic review by Shiri et al (35).
Although previous studies did not show a significant association between elevated BMI and disc degeneration, this is likely due to the lack of the use of advanced imaging (i.e., MRI) to assess the subtle disc changes that may otherwise be missed on plain radiographs or CT and the manner in which degenerative changes were assessed throughout the lumbar spine (i.e., assessing disc degeneration at each lumbar level or in specific regions). Therefore, it is very plausible that if the method of assessment of disc degeneration was consistent throughout studies, the effects of being overweight or obese on degenerative changes would be found to be more pronounced. Furthermore, our study entailed a large number of subjects, who were systematically assessed for additional radiographic findings (e.g., Schmorl's nodes) that may have an impact on the phenotype of disc degeneration (37). Using such a large sample further facilitated controlling for various confounding variables associated with disc degeneration, providing the ability to arrive at robust conclusions in the assessment of the relationship of being overweight or obese to disc degeneration and its various dimensions.
Being overweight and, in particular, obese are influential factors related to disc degeneration, but the exact mechanism of this association remains unknown. For numerous years, it has been postulated that being overweight or obese contributes to the compressive loading of the disc, leading to disc degeneration. However, recent studies have shown a linear association between hand osteoarthritis and atherosclerosis in elderly women (45). Such findings are independent of increased loading effects that could be caused by body weight and perhaps may contradict the notion that being overweight or obese may contribute to arthritic changes due to altered biomechanics (e.g., of the knee or hip).
Some authors contend that body weight may be a surrogate for physical loading and that musculoskeletal structures adapt to such forces (46, 47). However, in our study, we showed that being underweight was associated with a significantly lower likelihood of having disc degeneration. These findings suggest that physical loading on the disc in the form of elevated BMI may biomechanically affect the disc or that it may be synonymous with an alternative mechanism that is not yet well understood. For example, such a phenomenon may be mediated by an inflammatory response of secondary mediators secreted by adipocytes, such as adipocytokines (i.e., adiponectin, leptin, and resistin), macrophage-derived factors (i.e., interleukin-1β [IL-1β]), or proinflammatory cytokines and chemokines (i.e., C-reactive protein, tumor necrosis factor α, and IL-6) (48). Another potential mechanism may be vascular insufficiency to the vertebrae and subsequently to the disc, brought on by atherosclerosis or high serum lipid levels that can affect nutrient and metabolite transport into the disc (17, 19). Other possible mechanisms could involve a metabolic disorder (49) or gene–environment interaction effects (50). Nonetheless, our study has further broadened the understanding of factors related to lumbar disc degeneration in adults.
Although our study provides further insight into disc degeneration, there are certain potential limitations to consider. First, this study was conducted in a southern Chinese population, and our findings may not be generalizable to other populations. However, similar genetic factors found to be associated with degenerative changes of the spine in the Chinese cohort examined in this study have been replicated in Japanese and Finnish populations. Moreover, focusing on southern Chinese subjects minimizes genetic variation that may confound the assessment of disc degeneration, which can occur in mixed population studies. Alternatively, variations exist between populations that may affect the impact that being overweight or obese has on disc degeneration; thus, factors related to disc degeneration may be population- or ethnicity-dependent and should be investigated further.
A second limitation of this study was the cross-sectional design, which presents a challenge in assessing the causal pathway leading to disc degeneration. Therefore, we are hesitant to conclude that being overweight or obese leads to the development of disc degeneration without more prospective analyses. However, due to the nature of the degenerative process, it would not seem plausible that being overweight or obese is entirely a result of disc degeneration. Nonetheless, a prospective, long-term analysis of our study cohort is being conducted to further validate the findings of this study and to determine the effects of being overweight or obese on the progression of disc degeneration severity.
Disc degeneration is a complex multifactorial process characterized by structural and chemical changes in the disc. Future clinical or basic science studies assessing risk factors for disc degeneration should be cognizant of elevated BMI values, in particular BMI values that indicate being overweight or obese, and their effects on disease. In addition, future research efforts should be directed toward gaining a deeper understanding of the mechanism(s) that may contribute to disc degeneration and low back pain in overweight and obese individuals, with the goal of developing preventative or novel therapeutic interventions.
This study, the largest population-based study ever conducted to assess the phenotype of disc degeneration on MRI, is, to our knowledge, the first to show that BMI values are significantly higher in individuals with disc degeneration than in individuals with nondegenerated discs. Elevated BMI values, in particular BMI values indicating being overweight or obese, were associated with a greater extent and increased global severity of disc degeneration. Moreover, end-stage disc degeneration with disc space narrowing was significantly more pronounced in obese subjects. Furthermore, since this study was population based, it is of tremendous public health importance. Since there is abundant evidence in the literature demonstrating the strong association of disc degeneration on MRI with low back pain (2, 4, 5, 7–13), and since the present study clearly illustrates that being overweight or obese is a strong determinant related to disc degeneration of the lumbar spine, the public should be well informed that weight control is as important for preventing low back pain as it is for other conditions, such as heart disease and diabetes. Prevention and treatment of being overweight or obese must be a public health priority. If successful, such outcomes may lead to the prevention or minimization of the extent and severity of disc degeneration, which in turn may also decrease the risk of developing low back pain and the subsequent need for medical management, resulting in a more productive and healthier society.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Drs. Samartzis and Cheung had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Samartzis, Karppinen, Chan, Cheung.
Acquisition of data. Karppinen, Cheung.
Analysis and interpretation of data. Samartzis, Chan, Luk, Cheung.
The authors wish to thank Ms Yu Pei and Professor Kathryn Cheah of the Department of Biochemistry and Ms Florence P. S. Mok of the Department of Orthopaedics and Traumatology at the University of Hong Kong, Hong Kong, China for assistance with this study.