Sensitivity And Specificity Of Resting Metabolic Rate Measures To Predict Exercise Associated Menstrual Disturbances

Abstract

Energy deficiency in exercising women can lead to menstrual disturbances (MD). There is no gold standard to accurately estimate energy deficiency. Ratios of measured to predicted resting metabolic rate (RMR) have been used as a proxy to categorize women as energy deficient. PURPOSE: To evaluate whether measured to predicted RMR ratios are predictive of amenorrhea or other MD. METHODS: We performed a cross-sectional comparison of 223 exercising women (≥2 hrs/wk, age 18-35 years, BMI 16-30 kg/m) who were ovulatory (OV), amenorrheic (AMEN), or subclinical MD (sMD) (including oligomenorrhea, anovulation, and luteal phase defects). Menstrual status was determined using urinary measures of reproductive hormones and menstrual calendars. Body composition was measured with DXA and RMR with the SensorMedics Vmax. Harris-Benedict, Cunningham, and DXA equations were used to calculate predicted HBRMR, CRMR, and DXARMR and to calculate the measured to predicted RMR ratio. ANOVA and Kruskal-Wallis tests determined group differences and logistic regression determined predictors of AMEN or any MD. Calculations of sensitivity, specificity and positive predictive value (PPV) assessed accuracy of predictions. RESULTS: Groups did not differ in lean or fat free mass. AMEN had lower body mass (p<0.01) than sMD, and lower BMI, percent body fat, fat mass (p<0.001) and measured RMR (1172 ± 21 kcal/d) (p<0.05) than OV (1227 ± 20 kcal/d) and sMD (1233.68 ± 17 kcal/d). HBRMR was lower in AMEN (1402± 8 kcal/d) vs sMD (1434 ± 9 kcal/d) (p<0.05). CRMR ratio (0.84 ± 0.01) was lower in AMEN vs OV (0.88 ± 0.01) (p<0.05), but DXARMR ratio (0.90 ± 0.01) was lower in AMEN vs both OV (0.96 ± 0.01) and sMD (0.95 ± 0.01) (p<0.01). Each ratio predicted AMEN (HBRMR: χ =4.822, p<0.05; CRMR: χ =8.708, p<0.01; DXARMR: χ =14.068, p<0.001), but only DXARMR ratio predicted any MD (χ =6.795, p<0.01). DXARMR ratio correctly identified the most women with AMEN (ppv=0.5; sensitivity= 0.49, specificity= 0.74) and with any MD (AMEN+sMD: ppv=0.75; sensitivity= 0.39, specificity= 0.75). CONCLUSIONS: Each ratio may be used to predict AMEN, but only DXARMR significantly predicts MD, regardless of severity. Similarly, DXARMR ratio correctly identified the most subjects. DXARMR ratio can be utilized to correctly identify women with AMEN or MD secondary to energy deficiency.