Body Mass Index versus Percent Body Fat as a Predictor of Mortality in Men Aged 20-49

<h3>Context:</h3> Body mass index (BMI) is commonly used as a risk marker for mortality and cardiometabolic disease, but its predictive ability has come into question. Other body composition measures may have more utility in practice. <h3>Objective:</h3> To examine the predictive ability of accepted thresholds of BMI versus body fat percentage for 15-year mortality risk in young men and further characterize the BMI &gt;=25 and BF% &gt;=28% cohorts by comparing their respective prevalence of markers of cardiometabolic syndrome, diet, and exercise. <h3>Study Design &amp; Analysis:</h3> In this retrospective cohort analysis, Cox proportional hazards regression was computed for 15-year mortality risk. Healthy BMI was defined as 18.5-24.9 and unhealthy BMI as &gt;= 25. After a sensitivity analysis in men aged 20-49, healthy BF% was defined as &lt; 28 and unhealthy BF% as &gt;=28. 15-year mortality risk for the predictors of BMI and BF% was computed. Results were adjusted for age, poverty, and race. <h3>Dataset:</h3> NHANES 1999-2004 linked to the National Death Index which provides population estimates of the US population. <h3>Population Studied:</h3> Men aged 20-49 <h3>Instrument:</h3> Cox proportional hazards for 15-year mortality <h3>Outcome Measures:</h3> 15 year all-cause, heart disease, and/or cancer mortality <h3>Results:</h3> BF% is a strong significant predictor of 15-year, all-cause mortality [unadjusted hazard ratio (HR) 2.16, 95% confidence interval (CI) 1.44-3.25, P&lt;0.001; adjusted HR 1.93, CI 1.27-2.94, P=0.002] and heart disease mortality (unadjusted HR 5.55, 95% CI 1.53-20.14, P=0.009; adjusted HR 5.099, 95% CI 1.32-19.62, P= 0.02). BMI was not significant as a predictor in young adult men using healthy weight BMI versus overweight/obese. Significantly more subjects had a waist circumference &gt;40 inches, hypertension, and hyperlipidemia in the BF% &gt;= 28% group than the BMI &gt;=25 group, but there was no difference in diabetes, calorie intake, or nutrient intake. While cardiovascular fitness was the same between these groups, the BF&gt;=28% group performed significantly more moderate and vigorous activity over the past 30 days. <h3>Conclusions:</h3> BF% is a much better predictor of downstream mortality in young adult men than BMI. This could ultimately lead to a change in how we measure body composition for risk stratification. BF% is a better predictor of certain markers of cardiometabolic syndrome, and these patients may be more likely to increase their exercise if appropriately counseled to improve their health.