The study, published in BMC Pediatrics by Martin Jouza, Petr Jabandziev, and colleagues at University Hospital Brno and Masaryk University, focused on microRNAs, or miRNAs. These are short, non-coding RNA fragments, typically around 19 to 24 nucleotides long, that do not carry instructions for building proteins. Instead, they act as fine-tuners of gene expression, binding to messenger RNA transcripts and dampening their translation. Crucially for diagnostic medicine, miRNAs are remarkably stable in body fluids, protected from degradation by their small size and by packaging inside membrane-bound vesicles or association with carrier proteins. That stability makes them attractive candidates for liquid biopsy-style biomarkers, molecules measurable from a simple blood draw rather than invasive tissue sampling.

The rationale for looking at miRNAs in sepsis is rooted in immunology. Sepsis is fundamentally a dysregulated inflammatory response to infection, and miRNAs are deeply embedded in the control of that response. Individual miRNAs regulate networks of cytokine signaling, neutrophil function, and macrophage polarization. When infection tips the immune system into overdrive, the circulating profile of these molecules shifts measurably. Previous smaller studies had hinted that specific miRNAs change in septic newborns, but independent confirmation has been scarce, and the field has repeatedly seen promising biomarker candidates fail when tested in new cohorts. The Brno team set out to provide exactly that independent test.

The researchers assembled a cohort of 64 newborns with sepsis and 21 control infants without signs of systemic infection. Within the sepsis group, 20 patients were classified as having early-onset sepsis, typically acquired around the time of birth, and 41 had late-onset sepsis, acquired later during hospitalization. Blood cultures were positive in 45 percent of cases, with Staphylococcus epidermidis, Escherichia coli, and Staphylococcus aureus standing out as the most common pathogens. From each patient, peripheral blood was collected and plasma was separated, from which total RNA enriched for small RNA species was extracted.

On the technical side, the team selected eight candidate miRNAs based on a systematic search of the existing literature: hsa-miR-29a-3p, hsa-miR-96-5p, hsa-miR-185-5p, hsa-miR-16-5p, hsa-miR-15a-5p, hsa-miR-132-3p, hsa-miR-223-3p, and hsa-miR-26a-5p. Levels of each molecule were quantified using quantitative polymerase chain reaction, the workhorse technique of nucleic acid measurement, which amplifies specific RNA sequences and allows their relative abundance to be calculated. To control for technical variation between samples, the researchers normalized the results to the average expression of all measured miRNAs, a common strategy when no single endogenous reference is reliable across inflammatory states. Statistical analysis was performed in GraphPad Prism 8.

The results were clear-cut. Of the eight miRNAs examined, three were significantly dysregulated in septic newborns compared with controls: miR-15a, miR-132, and miR-223. MiR-15a showed the most pronounced effect, with a p-value of 0.0009, indicating a difference between groups that would arise by chance less than one time in a thousand. To translate these molecular findings into something clinically usable, the team built a combined diagnostic model using logistic regression, a statistical method that weighs multiple inputs to estimate the probability that a patient belongs to one group or another. Incorporating the three most significantly dysregulated miRNAs, the model achieved an area under the curve, or AUC, of 0.78. In diagnostic terms, an AUC of 1.0 represents perfect discrimination and 0.5 represents a coin flip, so 0.78 indicates meaningful, though not definitive, diagnostic power.

Several secondary findings add depth to the picture. No significant differences in miRNA expression emerged between early-onset and late-onset sepsis, suggesting that the molecular signature reflects the septic process itself rather than the timing of infection. More intriguingly, the researchers tracked miRNA levels longitudinally. MiR-15a showed significant dysregulation during the acute phase of the disease and returned to baseline levels by day seven, behaving like a dynamic readout of the inflammatory storm rather than a static marker. When the team stratified patients by nSOFA score, a neonatal adaptation of the Sequential Organ Failure Assessment used to gauge sepsis severity, they found divergent miRNA normalization trajectories between severity subgroups. MiR-15a and miR-223 were significantly dysregulated in patients with higher nSOFA scores, hinting that these molecules could flag which infants are heading toward severe disease before the clinical picture fully declares itself.

The researchers also examined how the miRNA signals relate to C-reactive protein, or CRP, the standard inflammatory marker measured in virtually every suspected sepsis workup. CRP was systematically assessed in all patients, and correlations between CRP and miR-15a, miR-132, and miR-223 were analyzed. CRP has well-known limitations in newborns: it rises slowly after the onset of infection, can be elevated for benign reasons related to birth trauma, and lacks specificity. A biomarker panel that captures complementary information, or that responds faster than CRP, would fill a genuine clinical gap. The study’s design, pairing molecular measurements with routine CRP data, positions the miRNAs as potential partners rather than immediate replacements for existing tests.

The authors conclude that miRNAs could serve as semi-invasive biomarkers for neonatal sepsis, and they emphasize the prognostic dimension of their findings. Because miR-15a and miR-223 tracked with nSOFA-based severity, they may function as predictive biomarkers for severe disease progression, not merely diagnostic ones. In practical terms, a neonatologist armed with such a test could, in principle, stratify patients at the bedside, intensify monitoring or escalate antibiotics for infants whose molecular profile signals high risk, and personalize treatment intensity in a condition where both undertreatment and overtreatment carry serious costs. The study is also notable as the first independent evaluation of these particular miRNA findings in neonatal sepsis, addressing a reproducibility problem that has long dogged the biomarker literature.

Caution is still warranted before such a test reaches the nursery. The cohort, while well characterized, is modest in size, and the AUC of 0.78, though respectable, falls short of the near-perfect discrimination clinicians would want for a stand-alone diagnostic. Normalization strategies for circulating miRNAs remain debated, and standardization across laboratories will be essential before any clinical deployment. Still, the convergence of diagnostic signal, longitudinal dynamics, and severity correlation in a single validated panel marks real progress. If larger multicenter trials confirm these results, the humble microRNA, a molecule once dismissed as genetic noise, could become a routine part of the arsenal against one of newborn medicine’s deadliest adversaries.

Subject of Research: Circulating microRNA biomarkers for diagnosing and predicting severity of neonatal sepsis

Article Title: Dysregulated microRNA expression in neonatal sepsis: validation study of potential novel biomarkers

Article References: Jouza, M., Bohosova, J., Sonawane, S., Naar, O., Slaba, K., Stanikova, A., Slaby, O., & Jabandziev, P. (2026). Dysregulated microRNA expression in neonatal sepsis: validation study of potential novel biomarkers. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07703-8

Image Credits: AI Generated

DOI: 10.1186/s12887-026-07703-8

Keywords: microRNA, neonatal sepsis, biomarkers, miR-15a, miR-223, nSOFA, logistic regression, quantitative PCR, C-reactive protein, neonatology, non-coding RNA, liquid biopsy