BACKGROUND: The immature platelet fraction (IPF) reflects the degree of reticulated platelets. We evaluated performances of IPF as a biomarker for the discrimination of septic patients from non-septic patients and sepsis severity. METHODS: Total 312 patients admitted between March and July 2013 were enrolled and samples were obtained at admission. Lactate (LA), procalcitonin (PCT), C-reactive protein (CRP), immature granulocyte fraction (IG), immature reticulocyte fraction (IRF), and IPF were analyzed as sepsis biomarkers and their performances were compared. RESULTS: The performance of IPF (area under the curve [AUC]=0.868) in the discrimination of septic patients from non-septic patients was comparable to PCT/CRP/LA/IG (AUC=0.923/0.940/0.781/0.812, P=0.233/0.106/0.186/0.353, respectively), and was significantly better than the IRF (AUC=0.658, P=0.007). Sensitivity (89.8%, 95% confidence interval [CI] 84.9-99.8%) and accuracy (83.2%, 95% CI 78.8-90.0%) of IPF were the best among all biomarkers. The performance of IPF in discriminating septic patients from non-septic patients with local infection showed similar results. However, the IPF could not efficiently discriminate sepsis severity (AUC=0.599), similar to other biomarkers (AUC=0.519-0.752). CONCLUSIONS: The IPF possessed high sensitivity/accuracy in discriminating septic patients from non-septic patients, regardless of local infection status. However, the IPF did not efficiently discriminate sepsis severity. The clinical relevance of IPF as a sepsis biomarker is, therefore, limited to sensitive and accurate discrimination of septic patients from non-septic patients, not discrimination of sepsis severity.