Diagnostic tests for sepsis aim to either detect the infectious agent (such as microbiological cultures) or detect host markers that commonly change in response to an infection (such as C-reactive protein). The latter category of tests has advantages compared to culture-based methods, including a quick turnaround time and in some cases lower requirements for blood samples. They also provide information on the immune response of the host, a critical determinant of clinical outcome. However, they do not always differentiate nonspecific host inflammation from true infection and can inadvertently lead to antibiotic overuse. Multiple noninfectious conditions unique to neonates in the first days after birth can lead to inflammatory marker profiles that mimic those seen among infected infants. Our goal was to review noninfectious conditions and patient characteristics that alter host inflammatory markers commonly used for the diagnosis of early-onset sepsis. Recognizing these conditions can focus the use of biomarkers on patients most likely to benefit while avoiding scenarios that promote false positives. We highlight approaches that may improve biomarker performance and emphasize the need to use patient outcomes, in addition to conventional diagnostic performance analysis, to establish clinical utility.