Study participants
The Melbourne Collaborative Cohort Study (MCCS) is a well-established cohort study that recruited 41,514 participants (17,045 men, 99.3% aged 40-69 years) during 1990-1994[9]. The study protocol was approved by the Cancer Council Victoria’s Human Research Ethics Committee[9]. For the current study, 2,590 (6.2%) were excluded because they had either: died or left Australia or reported having an arthroplasty prior to 1 January 2001; or their first recorded procedure was a revision arthroplasty[9], leaving 38,924 participants available for analysis.
Demographic data, anthropometric measurements and classification of obesity categories
At baseline, demographic and lifestyle data, including date of birth, sex, country of birth, education, smoking and physical activity, were collected using standard questionnaires[9]. WC, height and weight were measured using standard procedures[9]. Obesity was defined by BMI ≥30 kg/m2 or WC ≥102 cm for men and ≥88 cm for women[10]. Obesity status was classified based on combination of BMI-and WC-defined obesity: (i) not obese (non-obese BMI and non-obese WC, BMIN/WCN); (ii) WC-defined obesity only (non-obese BMI and obese WC, BMIN/WCO); (iii) BMI-defined obesity only (non-obese WC and obese BMI, BMIO/WCN); (iv) BMI- and WC-defined obesity (BMIO/WCO), where N = non-obese and O = obese.
Incidence of total knee arthroplasty for OA
The Australian Orthopaedic Association National Joint Replacement Registry (AOA NJRR) collects information on prostheses, patient demographics, type and reason for arthroplasty, with an almost complete data relating to arthroplasty (>99%) in Australia[11]. Linking the MCCS records to the AOA NJRR identified those who had a primary total knee arthroplasty performed between 1 January 2001 and 31 December 2013. Knee OA was defined as the first primary total knee arthroplasty with a contemporaneous diagnosis of OA, as recorded in the AOA NJRR. If one person had multiple arthroplasties, the first recorded procedure was considered the event. The linkage study was approved by the Human Research Ethics Committee of Cancer Council Victoria and Monash University.
Statistical analysis
Cox proportional hazard regression models were used to estimate the hazard ratio (HR) and 95% confidence interval (CI) for the incidence of total knee arthroplasty due to OA associated with obesity categories, with age as the time scale. Follow-up for arthroplasty (calculation of person-time) began 1 January 2001 and ended at the date of first knee arthroplasty for OA or date of censoring. Participants were censored at either the date of first knee arthroplasty for indications other than OA, the date of death, or end of follow-up, whichever came first. All analyses were adjusted for sex, education, smoking status, physical activity and country of birth. As men and women demonstrate different obesity category distributions and have different risks of knee arthroplasty, stratified analyses by gender were performed. Tests based on Cox regression methods showed no evidence that proportional hazard assumptions were violated for any analysis. All statistical analyses were performed using Stata 15.0 (StataCorp LP., College Station, TX, USA).