![[ Visit Client Website ]](images/banner.gif)
Methods: Data were from the Johnston County Osteoarthritis Project, a longitudinal study of African-Americans and Whites aged 45 years and older in Johnston County, NC from T0 (1990-1998) to T1 (1999- 2003 (n=1,480). Weight change was defined as change from initial weight and was coded as a 5-level variable with categories defined as: ≥5% loss, >3 to <5% loss, ± 3%, >3 to <5% gain, and ≥5% gain. Indicator variables were used to make contrasts between ≥5% loss (weight loss), ± 3% (weight maintenance), and ≥5% (weight gain) with weight gain as the referent. Incident rKOA was defined as Kellgren-Lawrence (K-L) grade of 0 or 1 at T0 and K-L ≥ 2 at T1. Knee-based Weibull proportional hazards models with adjustment for the correlation between knees were used to calculate hazard ratios and 95% confidence intervals for the association between weight change incident rKOA. All models were adjusted for age, race, sex, height, and the mean of weights from T0 and T1.
Results: Of the 1,480 individuals, 63.2% were female and 25.9% were African-American with mean (SD) age and BMI of 59.4 (9.4) years and 28.6 (5.5) kg/m2, respectively. Mean (SD) follow-up time was 5.9 (1.3) years (range 3.6 to 13.2), during which rKOA developed in 415 (14.9%) of 2,788 knees. Compared to those who gained weight (31.4%), those who maintained weight (32.8%) were no less likely to develop incident rKOA [HR=1.02 (95% CI=0.77, 1.35)], but those who lost weight (16.7%) were at reduced risk [0.71 (0.49, 1.01)].
Disclosure: L. M. Abbate, None; J. Stevens, None; T. A. Schwartz, None; L. F. Callahan, None; J. B. Renner, None; C. G. Helmick, None; J. M. Jordan, None.