Frisch and colleagues and others have used the Mellits and Cheek method to estimate body fatness in women athletes, and a value (22%) obtained from this calculation has been interpreted as a critical fatness for the maintenance and restoration of regular menstrual cycles. In comparison with hydrostatic weighing, the Mellits and Cheek method was shown to be inaccurate in predicting either mean or individual percent body fat in women runners and body builders. No fatness threshold for the maintenance of regular menses was found. There was no difference in mean body fat between eumenorrheic (n = 19) and amenorrheic (n = 13) athletes, all of whom were less than 22% body fat.
Loucks, Anne B.; Horvath, Steven M.; and Freedson, Patty S.
"Menstrual Status and Validation of Body Fat Prediction in Athletes,"
2, Article 15.
Available at: https://digitalcommons.wayne.edu/humbiol/vol56/iss2/15