Body mass index as a predictor of percent fat in university students and athletes.
J. Ode, J. Knous , M. Reeves, J. Pivarnik, FACSM.
Departments of Kinesiology and Epidemiology
Michigan State University, East Lansing, MI 48824BACKGROUND: In a recent research letter published by the Journal of the American Medical Association, it was reported that a quarter of players within the National Football League (NFL) were classified as class II obese via body mass index (BMI) measures. It was stated the elevated BMI of these athletes could result in future health consequences.
PURPOSE: The purpose of the current research project is to determine the accuracy of BMI as a measure of excessive body fatness in elite college athletes and to compare this with a reference group of college students.
METHODS: A total of 226 (149 male, 77 female) Big Ten athletes and 213 college students (78 male, 135 female) from a large Midwestern University participated in the study. Three male groups: 1) Non-lineman athletes, 2) Linemen, and 3) Students and two female groups: 1) Athletes, and 2) Students were created. Body composition (i.e. %fat) was determined via air plethysmography. A cutpoint of percent body fat of ≥20% for males, and ≥30% for females were used to define the status if overfat. Height was measured via a stadiometer, weight was measured via the BodPod calibrated scale and BMI was calculated. A cutpoint of BMI of 25 was used to define overweight. Using % body fat as the gold standard, the sensitivity, specificity, and predictive values of BMI were calculated. Receiver operator characteristics analysis was used to derive optimal BMI cut points for body fatness.
RESULTS: For the male athlete group, sensitivity of BMI ≥ 25 was high (1.0), but the specificity was low (.27). Sensitivity was also high in the linemen group (1.0). However, it was not possible to estimate specificity because only one individual was less than 20% fat. The sensitivity and specificity of BMI for the student group was (.83) and (.60), respectively. ROC analysis showed the optimal BMI cut points for the athlete, linemen, and student groups were 27.9, 34.1, and 26.5, respectively. For the female athlete group the sensitivity was .86 and the specificity was .78. For the female student group the sensitivity was .43 and the specificity was .94. The optimal BMI cut point for female athletes and students was 26.7 and 22.7, respectively.
CONCLUSIONS: The operating characteristics of the BMI classification system for overweight (i.e. ≥25 vs. <25) varies widely between different populations of college students. There is a high specificity in both male athletes and linemen, which results in few cases of overfat individuals being classified as non-obese. However, the specificity is very low indicating that a large proportion of under fat individuals are classified as overweight. When compared to the male athletes, the male students have a lower sensitivity, but a higher specificity. However, the optimal BMI cut points for both male athletes and students suggest the current recommendations are too low. The sensitivity was higher and specificity lower in female college athletes when compared to students. The low sensitivity for students suggests a large proportion of over-fat individuals being classified as normal weight. The optimal BMI cut point suggest the current recommendations for overweight are too low in female athletes and are elevated in the general female student population.
