OBJECTIVE: To evaluate the association between weight variability and disease incidence in women. DESIGN: Prospective cohort study, following women from 1986 through 1992. METHODS: A population-based sample of 33,834 women aged 55-69 y, free of cancer and heart disease, completed a mail-based survey that included self-reported body weights at ages 18, 30, 40, 50 y, and currently. Weight variability was defined as (1) the root mean square error around the slope of weight on age (RMSE); and (2) categorical measures of weight change. Outcome measures were incidence of myocardial infarction (MI); stroke; diabetes; breast, endometrial, lung, or other cancer; total and hip fractures. RESULTS: Adjusted relative risks of MI, stroke, diabetes, and hip fracture increased with increasing weight variability. The age and body mass index-adjusted relative risks (RR) for highest vs lowest quartile of RMSE were: MI: 2.03; stroke: 1.61; diabetes: 1.42; breast cancer: 0.85; endomentrial cancer: 0.88; lung cancer: 1.70; other cancer: 0.93; total fractures: 1.15; hip fractures: 1.45. The strongest associations between weight change categories and disease were for diabetes (RR compared to small gain/stable weight: large cycle, 1.72; small cycle, 1.55; large gain, 1.80; weight loss, 1.91; other pattern, 1.55). Large weight cycles were associated with higher risk of MI (RR = 1.89) and stroke (RR = 1.71). CONCLUSIONS: These findings are consistent with previous studies and suggest that weight variability is associated with higher risk of developing chronic diseases.
Bibliographical noteFunding Information:
This research was supported by the National Cancer Institute Grant R01 CA39742 and by a cooperative agreement from the Associations of Schools of Public Health and the Centers for Disease Control and Prevention. The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute.
- Cardiovascular disease
- Weight cycling
- Weight loss
- Weight variability