Validity of the Mellits and Cheek method for body-fat estimation in relation to menstrual cycle status in athletes and non-athletes below 22 per cent fat.

The present experiment tested the hypothesis that 22% body fat based on the Mellits and Cheek (1970) regression equation is a minimum or 'trigger' level of fatness required to maintain regularity of menstrual cycles. The validity of the regression equation was cross-validated against densitometric estimates of body fat in 30 athletes (track, gymnastic, crew) and 30 non-athletes who were below 22% fat. For the total sample, the validity correlation was r = 0.47 between body fat (density) and body fat (regression equation); it was r = 0.75 for athletes and r = 0.47 for non-athletes. For all subjects, 60% had regular menses, 27% were oligomenorrheic and 13% were amenorrheic. The lowest body fat (density) was 12.8% for an athlete with a regular cycle and 12.2% for an oligomenorrheic athlete. The results show that: the Mellits and Cheek (1970) height-weight based regression equation to determine fatness is not valid for women below 22% fat, and 13% body fat (not 22%), should be considered an upper-bound estimate of a minimum level of fatness associated with normal ovulatory cycles.