The purpose of this study was to measure the influence of breathing frequency (BF) on heart rate variability (HRV) and specifically on the Low Frequency/High Frequency (LF/HF) ratio in athletes. Fifteen male athletes were subjected to HRV measurements under six randomised breathing conditions: spontaneous breathing frequency (SBF) and five others at controlled breathing frequencies (CBF) (0.20; 0.175; 0.15; 0.125 and 0.10 Hz). The subjects were divided in two groups: the first group included athletes with SBF <0.15 Hz (infSBF) and the second athletes with SBF higher than 0.15 Hz (supSBF). Fatigue and training load were evaluated using a validated questionnaire. There was no difference between the two groups for the fatigue questionnaire and training load. However, the LF/HF ratio during SBF was higher in infSBF than in supSBF (6.82 +/- 4.55 vs. 0.72 +/- 0.52; p<0.001). The SBF and LF/HF ratio were significantly correlated (R=-0.69; p=0.004). For the five CBF, no differences were found between groups; however, LF/HF ratios were very significantly different between sessions at 0.20; 0.175; 0.15 Hz and 0.125; 0.10 Hz. In this study, BF was the main modulator of the LF/HF ratio in both controlled breathing and spontaneous breathing. Although, none of the subjects of the infSBF group were overtrained, during SBF they all presented LF/HF ratios higher than four commonly interpreted as an overtraining syndrome. During each CBF, all athletes presented spectral energy mainly concentrated around their BF. Consequently, spectral energy was located either in LF or in HF band. These results demonstrate that the LF/HF ratio is unreliable for studying athletes presenting SBF close to 0.15 Hz leading to misclassification in fatigue.