Background Although overweight and obesity are associated with cardiovascular disease (CVD), it is unclear which clinical measure of overweight and obesity is the strongest predictor of CVD, and it is unclear whether the various measures of overweight and obesity are indeed independent predictors of CVD. Methods This study was a prospective population-based study of 2493 Danish men and women, age 41–72 years, without major CVD at baseline. At baseline, body mass index, waist circumference, hip circumference, waist-to-hip ratio (WHR), and traditional and new risk factors were recorded. Results Over a median follow-up of 12.6 years, the incidence of a combined CV event (CV death, nonfatal ischemic heart disease, and nonfatal stroke) amounted to 328 cases. Of the various measures of overweight and obesity, in Cox-proportional hazard models, adjusted for age, only WHR was significantly associated with incident CVD with a hazard ratio (95% confidence interval) in women of 2.22 (1.31–3.77; P = 0.0032) in the highest compared with the lowest quartile, and a hazard ratio in men of 1.73 (1.12–2.66; P = 0.014) in the highest compared with the lowest quartile. However, when adjustments were made for the presence of diabetes, dyslipidemia, hypertension, hyperinsulinemia, and inflammation, WHR was no longer significantly (P > 0.41) associated with incident CVD. Conclusion In this study, of the various measures of overweight and obesity, WHR was the only significant predictor of incident CVD, and the relationship between WHR and risk of CVD was mediated by well-known risk factors of CVD.