The first objective of this article was to determine the diagnostic accuracy of tumor necrosis factor-alpha, interleukin-6 (IL-6), and interleukin-8 (IL-8) in differentiating infected from noninfected neonates during the first 24 hours of suspected sepsis and to compare them to the currently used laboratory parameters: C-reactive protein (CRP), immature-to-total neutrophil ratio, and leukocyte and platelet count. The secondary objective was to compare the cytokine levels in subpopulations of neonates. Seventy-five premature and 30 term infants were enrolled. Blood samples for the "currently used laboratory tests" and the cytokine levels were obtained at the first suspicion of sepsis ("0-hour") and 18 to 30 hours later ("24-hours"). Patients were classified as septic (48) or nonseptic (57). Thirty-two septic patients had positive blood cultures and 16 showed clinical signs of sepsis. Twenty septic patients had early-onset and 28 had late-onset sepsis. Sensitivity, specificity, and positive and negative predictive values (PPV and NPV) were calculated for each test. Receiver-operating characteristic curves were analyzed to determine the optimal thresholds. A combination of CRP > 10 pg/mL plus IL-6 > 18 pg/mL (sensitivity = 89%, specificity = 73%, PPV = 70%, NPV = 90%) was the best "0-hour" test, and CRP (sensitivity = 78%, specificity = 94%) was the best "24-hours" test. Lower IL-6 at 0-hour (p = 0.018) and IL-8 at 24 hours (p = 0.023) were detected among the patients infected with coagulase-negative staphylococci then with other bacteria. In conclusion, a combination of CRP + IL-6 provided additional diagnostic accuracy for differentiation between septic and nonseptic patients during the first 24 hours of suspected sepsis.