BACKGROUND: During early diagnostic studies on Acute Lymphoblastic leukemia (ALL) patients, it is difficult to differentiate the origin of fever. Antibiotics are administered first until defervescence prior to chemotherapy. Thus, we tested the effect of administration of both antibiotics and prednisone in early phase prior to chemotherapy. METHODS: From January 1, 1989 to March 20, 2013, 28 patients diagnosed with ALL at Wonju Severance Christian hospital were selected. They were febrile but had no evidence of infection. Prednisone was given with antibiotics prior to chemotherapy to group 1, and antibiotics only to group 2. Age, sex, height, weight, initial complete blood cell count, blood chemistry, time to recover from febrile state, time to initiate remission induction therapy, period of remission induction therapy and time to complete remission were recorded. RESULTS: Ages and heights showed statistical differences between groups. Initial neutrophil counts, hemoglobin, C-reactive protein level and platelet counts revealed no differences. Prednisone was administered 3 days after the admission in group 1 and 5 days in group 2. Fever subsidence time, times to start remission induction, duration of the treatment, time to complete remission did not suggest significant difference. However, log-rank test demonstrated that the fever subsided earlier in group 1 (P<0.05). CONCLUSION: Early glucocorticoid use in febrile ALL patients without evidence of infection revealed contradictory results on fever subsidence time depending on different statistical methods, and a prospective study with larger sample is required to corroborate the finding.