Waist Circumference (WC) and not Body Mass Index (BMI) Predicts All-cause Mortality and Diabetes Mellitus (DM): a PreCIS Database Study
Abstract Number: 1028-P
Authors: HARPREET BAJAJ, DANIELLE BRENNAN, BYRON HOOGWERF, KRUPA DOSHI, SANGEETA KASHYAP, Cleveland, OH
Institutions: Cleveland, OH
Results: There is ongoing uncertainty about whether WC adds value to risk prediction models over BMI.
We compared the associations of WC and BMI with all-cause mortality and prevalence of DM in 4622 consecutive patients with BMI <35 (64% male (M), 88% Caucasian, 56 ± 13 years, BMI 28 ± 4) who presented to our Preventive Cardiology clinic from 1995-2006. This cohort was classified retrospectively into gender specific BMI and WC categories (figure - legend). Mortality was obtained prospectively from the Social Security death index over 5.4 years (median). DM was defined by patient history, DM medications and/or fasting glucose ≥126 mg/dl.
Mortality was greater in high vs. normal WC group in both genders, whereas BMI categories had similar mortality (figure). Within the overweight group, high vs. normal WC predicted mortality (M p<0.001, F p=0.009). In the Cox proportional hazards model (entire cohort), the unadjusted hazard ratio (HR) of death for WC (continuous) was 1.02 (p=0.023) in M and 1.04 (p<0.001) in F. After adjustment for traditional cardiovascular (CV) risk factors, HR was no longer significant in M, but remained associated in F (HR=1.03, p=0.009).
Age-adjusted DM was associated with WC and BMI groups in both genders. However, after adjustment for WC, DM was no longer statistically different among BMI groups. When continuous BMI and WC values were used together in a regression model, WC alone was significantly associated with DM (OR: M 1.05, F 1.06, both p<0.001).
We conclude that a 1 cm difference in WC for men and women is associated with an increase in all-cause mortality of 2% and 4% and in DM prevalence of 5% and 6%, respectively. This excess mortality is explained by traditional CV risk factors in men, but not in women. While BMI categorization does not portend mortality risk in this cohort, the NIH classification - subcategorizing BMI by WC - predicts mortality in the overweight group in particular. Overall, our observations support and add data from a cohort of patients typically seen in clinical practice to the recent consensus statement (NAASO, ASN and ADA) about the clinical utility of WC measurements.