Fasted or weight-category athletes manage their training under strict diet conditions that could impair the stress-recovery balance and result in acute or chronic fatigue. However, to date, no validated biomarker are available to quantify this phenomena. The aim of this study was to assess the validity of a specific index combining plasma albumin and weight change to detect nutrition-related risks of fatigue increase and under-performance in athletes experiencing particular nutritional conditions. An athlete's nutrition risk index (ANRI) equation, based on data from lightweight and heavyweight rowers, was developed using relationship between plasma albumin concentrations combined to weight changes with sport performance and overtraining scores and was tested by odds ratio for failure. The accuracy and sensitivity of this former specific equation was subsequently tested on runners observing the Ramadan-fasting as well as on boxers after a short weight-loss period. Independently of training and performance, lightweight rowers presented lower nutritional parameters than heavyweight (albumin: 37.4 ± 2.7 vs 39.9 ± 1.8 g·L−1, P < 0.05; weight state: 94.5 ± 1.8 vs 99.9 ± 0.9%, P < 0.01). In lightweight, ANRI was related with overtraining score (R2 = 0.21, P < 0.01), risks for failure in competition were enhanced when ANRI increased (OR:2.5, P = 0.03). Relationship of ANRI with overtraining score tended to be also significant in runners (R2 = 0.32, P = 0.06) but not in boxers (P = 0.4). Albumin concentrations combined to weight loss appeared relevant to delineate nutrition-related risks of fatigue and/or competitive failure associated with mid-term diets (about 30 days) as observed in rowers and Ramadan-fasted runners. ANRI may benefit to athletes monitoring by delineating effects of their weight loss program.