Including both waist and hip circumference in regression models revealed the full strength of their association with morbidity and mortality across the reviewed studies.
Systematic Review
Does including both waist and hip circumference in risk models improve the prediction of cardiovascular diseases, diabetes, and mortality?
Including both waist and hip circumference as separate terms in risk prediction models may improve the estimation of obesity-related risk for cardiovascular disease and mortality by accounting for the protective effect of hip circumference.
Both a larger waist and narrow hips are associated with heightened risk of diabetes, cardiovascular diseases and premature mortality. We review the risk of these outcomes for levels of waist and hip circumferences when terms for both anthropometric measures were included in regression models. MEDLINE and EMBASE were searched (last updated July 2012) for studies reporting the association with the outcomes mentioned earlier for both waist and hip circumferences (unadjusted and with both terms included in the model). Ten studies reported the association between hip circumference and death and/or disease outcomes both unadjusted and adjusted for waist circumference. Five studies reported the risk associated with waist circumference both unadjusted and adjusted for hip circumference. With the exception of one study of venous thromboembolism, the full strength of the association between either waist circumference or hip circumference with morbidity and/or mortality was only apparent when terms for both anthropometric measures were included in regression models. Without accounting for the protective effect of hip circumference, the effect of obesity on risk of death and disease may be seriously underestimated. Considered together (but not as a ratio measure), waist and hip circumference may improve risk prediction models for cardiovascular disease and other outcomes.
Cameron et al. (Wed,) conducted a systematic review in Cardiovascular diseases, diabetes, and mortality. Inclusion of both waist and hip circumference in risk models vs. Unadjusted models or single-measure models was evaluated on Association with morbidity and/or mortality. Including both waist and hip circumference in regression models revealed the full strength of their association with morbidity and mortality across the reviewed studies.