Low energy availability (EA) suppresses many physiological processes, including ovarian function in female athletes. Low EA could also predispose athletes to develop a state of overreaching. This study compared the changes in ad libitum energy intake (EI), exercise energy expenditure (ExEE), and EA among runners completing a training overload (TO) phase. We tested the hypothesis that runners becoming overreached would show decreased EA, suppressed ovarian function and plasma leptin, compared to well-adapted (WA) runners. After 1 menstrual cycle (baseline), 16 eumenorrheic runners performed 4 weeks of TO followed by a 2-week recovery (131±3% and 63±6% of baseline running volume respectively). Seven-day ExEE, EI, running performance (RUNPERF) and plasma [leptin] were assessed for each phase. Salivary [estradiol] was measured daily. Urinary [luteinizing hormone] tests confirmed ovulation. Nine runners adapted positively to TO (WA,ΔRUNPERF: +4±2%); seven were non-functionally overreached (NFOR, ΔRUNPERF -9±2%) as RUNPERF remained suppressed after the recovery period. WA increased EI during TO, maintaining their baseline EA despite a large increase in ExEE (ΔEA=+1.9±1.3 kcal.kgFFM-1.d-1, P=0.17). By contrast, NFOR showed no change in EI, leading to decreased EA (ΔEA=-5.6±2.1 kcal.kgFFM-1.d-1, P=0.04). [Leptin]b, mid-cycle and luteal [estradiol]s decreased in NFOR only. Contrasting with WA, NFOR failed to maintain baseline EA during TO, resulting in poor performance outcomes and suppressed ovarian function.NCT02224976. NOVELTY BULLETS: -Runners adapting positively to training overload (TO) increased ad libitum energy intake, maintaining baseline EA and ovarian function through TO. -By contrast, NFOR runners failed to increase energy intake, showing suppressed EA and ovarian function during TO.