Commonly accepted low energy availability indices failed to identify female soccer players experiencing low energy availability

Background:
Commonly proposed indices of low energy availability fail to identify at risk competitive female football players during pre-season training.

Aim:
This study aimed to determine energy availability (EA) and within-day energy balance (WDEB) in women soccer players during pre-season by assessing energy intake, energy expenditure (EE), physique and systemic glucose availability. We also explored eating disorder risk and athlete food choice. We hypothesised that indicators of low energy availability (LEA) risk would correlate with calculated EA and WDEB variables and that food choice determinants would differ according to EA status.

Methods:
Eleven National Premier League women soccer players (mean ± SD: age 20.2 ± 4.1 years, body mass 63.6 ± 6.8 kg, height 170.9 ± 7.6 cm) participated in this observational cross-sectional study over 3 weeks. Assessment of resting metabolic rate (RMR) and physique traits including bone mineral density (BMD) was conducted during Weeks 1 and 3. During Week 2, dietary intake, EE and continuous monitor-derived glucose was measured for 5 days. EA was calculated daily and WDEB calculated hourly with deficits/surpluses carried continuously throughout the 5 days. Questionnaires were administered throughout the 3 weeks to assess food choice determinants, plus eating disorder and LEA risk, including the Athlete Food Choice Questionnaire, the Eating Disorders Screen for Athletes (EDSA) and the Low Energy Availability in Females Questionnaire (LEAF-Q). RMR ratio, BMD, LEAF-Q and EDSA scores were used as indicators of LEA risk.

Results:
Daily energy and carbohydrate intakes were 42.8 ± 1.8 kcal/kg FFM/d and 3.5 ± 0.2 g/kg BM/d; and EA averaged 30.7 ± 7.5 kcals/kg FFM/d. Approximately, one third (36%) of athletes were at risk of an eating disorder, while approximately half (45%) were at risk of LEA via LEAF-Q, compared with approximately one third (36%) of athletes identified with measured LEA (<30 kcal/kg FFM/d). No athlete achieved optimal EA (>45 kcal/kg FFM/d) and no indicator of LEA risk was associated with calculated EA or WDEB. However, overnight glycaemic variability was positively correlated with EA (r = 0.722, p = 0.012). Food choice determinants did not differ by EA status.

Conclusions:
No athlete met optimal EA on training days, and there was a lack of association between calculated EA and commonly used indicators of LEA risk, including RMR ratio, LEAF-Q and BMD. Reduced variability in overnight glucose measures were associated with calculated EA warranting further investigation as emerging variables in the identification of LEA.