Methods

A total of 226 college-aged athletes and 213 college-aged nonathletes participated. Three male groups (athletes, football linemen, and nonathletes) and two female groups (athletes and nonathletes) were created. BMI was calculated. Percent fat was determined via BOD POD. BMI> or= 25 kg. m (-2) was used to define overweight. Twenty percent fat for males and 33% fat for females were used to define overfatness. Using% fat as the criterion, sensitivity and specificity of BMI were calculated. Receiver operator characteristic curves determined optimal BMI cut points for% fat.

Results

Sensitivity was high (0.83-1.0) and specificity was low (0.27-0.66) in male athletes, male nonathletes, and female athletes. Sensitivity was high in linemen (1.0). Sensitivity was low (0.56) and specificity was high (0.90) in female nonathletes. Optimal BMI cut points for male athletes, linemen, male nonathletes, female athletes, and female nonathletes were 27.9, 34.1, 26.5, 27.7, and 24.0 kg. m (-2), respectively.

Conclusions

BMI should be used cautiously when classifying fatness in college athletes and nonathletes. Our results support the need for different BMI classifications of overweight in these populations.