Answers are grounded in this paper and cite their sources.
Neonatal sepsis is a life-threatening condition with a high mortality rate, despite significant advances in perinatology and neonatology. The clinical signs of sepsis in newborns are nonspecific and can mimic a wide range of non-infectious pathologies, making early diagnosis challenging. The aim of this review was to analyze current data on laboratory markers of neonatal sepsis. A search was conducted in the PubMed database using the keywords neonatal sepsis, biomarker, prognosis. The search yielded 66 publications, an additional 9 publications were retrieved from the eLibrary database using the keywords биомаркеры (biomarkers), сепсис (sepsis), прогноз (prognosis). The search period was 5 years (2020–2025), in accordance with PRISMA 2020 guidelines. A review of the current literature confirms that no single laboratory marker, whether classical (C-reactive protein, procalcitonin, interleukin-6) or novel (presepsin, CD64, serum amyloid A, sTREM-1), is ideal due to limited specificity, the influence of non-infectious factors, and variability with gestational age. The most promising approach to diagnosing neonatal sepsis is a combined approach that includes simultaneous measurement of several biomarkers alongside clinical data. Artificial intelligence and machine learning technologies, which can analyze multidimensional data, predict sepsis at a preclinical stage, and create personalized diagnostic algorithms, show particular promise. Further research should therefore focus on developing integrated diagnostic systems to improve the accuracy, speed, and efficacy of sepsis detection in newborns.
No takes yet. Share an insight, caveat, or question.
Golomidov et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: