PURPOSE: Altitude training is popular among athletes to augment oxygen delivery capabilities to tissues and to improve physical performance. Hypoxia inducible factor-1 (HIF-1) controls the expression of several genes' encoding involved in physiological responses towards reduced oxygen availability, in particular by increasing serum erythropoietin (EPO). It may be involved in the individual variability for erythropoietic markers and/or sea-level performance of athletes using altitude during their training. Therefore, we investigated whether, before training, evolutions of hif-1alpha and ahif (HIF-1alpha natural antisense) transcript amounts and HIF-1alpha protein quantities in leukocytes measured during an acute hypoxia normobaric test (3 h at 3000 m at rest) could allow to predict poor and good responders for hematological markers after a "living high-training low" protocol. METHODS: Eighteen elite swimmers were divided into two groups that followed a 13-d training program: "living low-training low" (1200 m) (LL) or "living high (2500-3000 m)-training low (1200 m)" (LH). RESULTS: During the initial hypoxia test, a strong interindividual variability in the amounts of HIF-1alpha mRNA, aHIF transcript, and HIF-1alpha protein was observed in athlete leukocytes (after vs before): -82%/+396%, -100%/+229%, and -100%/+633%, respectively. After the test, serum erythropoietin concentration was increased (11.2 +/- 0.8 vs 9.8 +/- 0.8 IU.L(-1); +18%, P = 0.01). After the training protocol, total red cell volume (+7.6%, P = 0.04) and circulating hemoglobin amount (48.8 +/- 2.8 vs 45.5 +/- 3.0 mmol; i.e., +7.9%, P = 0.02) were significantly augmented in LH. CONCLUSION: We conclude that hif-1alpha gene expression quantification in leukocytes after a 3-h hypoxia test performed before training does not predict poor and good responder athletes to "living high-training low" model.