Sepsis is a frequent and serious problem in neonatal intensive care units, particularly in very low birth weight (VLBW) infants. Over 20% of all VLBW infants experience one or more episodes of late-onset sepsis. Unfortunately, diagnosis by blood culture is difficult, time consuming, and lacks sensitivity. Delaying treatment until symptoms arise can be life threatening. On the other hand, recurrent treatment with antibiotics in non- infants holds potential risks to these infants by furthering the development of resistant organisms. A number of research groups are performing detailed studies of individual biomarkers that have shown altered expression in septic infants and may be relevant to the development of the condition. However, the standard method for measuring plasma or serum levels of cytokines, chemokines or other biomarkers is to measure them one at a time using Enzyme-Linked Immunosorbent Assay (ELISA). Clearly, no single molecular marker, or small group of markers, will be able to accurately classify individuals at highest risk. In addition, one-at-a-time assessment of each putative biomarker incurs considerable time, cost and sample volume. The ability to systematically identify protein profiles, predict risk of clinical events, evaluate therapeutic response, and define underlying mechanisms is thereby limited severely. The recent development of bead-based multiplex immunoassays provides an efficient approach for performing a rapid assessment of large numbers of protein antigens. Rules- Based Medicine (RBM) has extended this approach to perform Multi-Analyte Profiles (MAP) of blood proteins using very small sample volumes (10-20 ?L) with a dynamic range of fg/mL to mg/mL. This technology is well suited for screening large numbers of markers in parallel to identify protein profiles associated with late-onset sepsis. During Phase I, RBM, and the University of New Mexico Health Sciences Center (UNMHSC), propose to characterize the progression of protein profiles in the blood of VLBW infants with late-onset sepsis vs. normal VLBW infants over a seven day period, and identify biomarker patterns associated with late-onset sepsis that will significantly improve clinical diagnosis. In addition, a new source for obtaining a diagnostic [blood] sample will be evaluated. During Phase II, a prospective validation of the MAP identified for late-onset sepsis in VLBW infants during Phase I efforts will be performed. The sensitivity, specificity, and positive and negative predictive values for each analyte, as well as, the MAP of biomarkers for predicting late-onset sepsis will be determined. In addition, a proposed normal physiological range of MAP analytes for VLBW infants based on gestational age, birth weight and postnatal age will be developed. The identification of novel biomarker patterns of infants in the early stages of late-onset sepsis, as well as infants at high risk for developing late-onset sepsis, will allow for improved management of the condition by shortening the time course of antibiotic administration, aiding the determination of treatment effectiveness, and will provide a framework for developing and evaluating new treatments.