Neonatal sepsis (NS) is a major cause of morbidity and mortality in both preterm and term infants; early-onset NS (EONS) occurs in newborns within the first 72 h of life (Eichberger and Resch, 2022; Schleier et al., 2023). Transplacental infection or an ascending infection from the cervix may be caused by organisms that colonize the mother’s genitourinary tract and pass through the colonized birth canal during delivery (Abdel-Hakim et al., 2019). The microorganisms most commonly associated with EONS include Group B Streptococcus (GBS), Escherichia coli, Coagulase-negative Staphylococcus, Haemophilus infuenzae, and Listeria monocytogenes (Abdel-Hakim et al., 2019; Kosmeri et al., 2024).

The diagnosis of EONS may be the greatest and most difficult challenge for a neonatologist (Shahkar et al., 2011). With EONS, the newborn may present a history of feeding intolerance, irritability, excessive sleepiness, or “just not looking right” (Ershad et al., 2019; Stein et al., 2023). Vital sign derangements include both hypothermia and fever; there may be tachycardia or bradycardia, signs of poor perfusion, including cool and pale extremities, and a rapid thready pulse (Ershad et al., 2019). Respiratory symptoms and signs are common in EONS, including grunting, nasal flaring, use of accessory muscles of respiration, cyanosis, and episodes of apnea (Ershad et al., 2019). Neurological symptoms and signs include lethargy, seizures, irregular respiration, high pitched cry, hypotonia, hypoactive deep tendon reflexes, and abnormal primitive reflexes (Ershad et al., 2019). Gastrointestinal signs include decreased feeding, vomiting, diarrhea, jaundice, abdominal distension, and hepatosplenomegaly (Ershad et al., 2019). Skin findings include petechiae, impetigo, cellulitis, and abscess (Ershad et al., 2019). Underlying metabolic acidosis secondary to poor perfusion can manifest as tachypnea and respiratory distress in the absence of respiratory tract infection (Ershad et al., 2019).

The signs and symptoms of EONS are similar to those with non-infectious inflammation. This brings more difficulties in the clinical diagnosis, when the infection source cannot be determined (Xing et al., 2023). The current guidelines contain three general approaches toward identifying newborns with an increased risk of EONS: a categorical risk factor assessment, a multivariate risk assessment (EONS calculator), and a risk assessment primarily based on newborn clinical conditions (Nusman et al., 2023). There are many complications associated with EONS, including septic shock, multiple organ failure, and death (Attia Hussein Mahmoud et al., 2023).

Many biomarkers have been studied to determine their utility in the diagnosis of EONS; the biomarkers are divided into acute-phase proteins, cell surface antigens, cytokines and chemokines, and soluble adhesion molecules (Gude et al., 2022). C-reactive protein (CRP), procalcitonin, serum amyloid A, and hepcidin are acute-phase proteins that are used to diagnose EONS; interleukin-1 (IL-1), IL-6, IL-8, tumor necrosis factor (TNF), and soluble TNF receptor (sTNFR) are cytokines and chemokines implicated in the diagnosis of EONS (Gude et al., 2022).

Cytokines are messengers with low molecular weight that are created by macrophages and lymphocytes in response to antigenic stimulations or products of inflammation (Boskabadi and Zakerihamidi, 2018; Gude et al., 2022). IL-6 is an inflammatory cytokine, produced immediately after induction of inflammation, peaking 6 h afterward; IL-6 measurement results can be obtained within a few hours after umbilical cord blood (UCB) sampling, so clinicians can make immediate treatment decisions (Fadilah et al., 2022). IL-8 is a pro-inflammatory cytokine and is predominantly produced by monocytes, macrophages, and endothelial cells. IL-8 regulates the migration and activation of leukocytes; IL-8 level can be evaluated promptly within 1–3 h of infection, and its half-life is less than 4 h (Zhou et al., 2015).

Unfortunately, the gold-standard diagnostic technique of EONS, the blood culture, requires at least 24–72 h to deliver a reliable result, and the prenatal use of antibiotics further decreases its accuracy (Qiu et al., 2018; Liu et al., 2020). Despite its non-specific clinical symptoms, EONS is highly fatal and can lead to serious long-term complications (Liu et al., 2020; Goyal et al., 2024).

The Kaiser Permanente (KP) group in Northern California applied the Bayesian approach to create a multivariate model of predicting infant-specific EONS risk—the sepsis risk calculator (SRC), derived and validated from a case-control study of blood culture-proven EONS (Puopolo et al., 2011; Escobar et al., 2014; Kuzniewicz et al., 2017; Goel et al., 2020). The SRC derives the “sepsis risk score” (SRS) at birth by modifying the population risk with intrapartum risk factors of the specific mother–infant pair; it then incorporates the infant’s clinical status to compute the final SRS (Goel et al., 2020). Although the SRS computation algorithm is evidence based, the clinical management recommendations (based on the SRS) are a consensus opinion of the KP clinicians (Goel et al., 2020). For example, infants with SRS <1 are assigned normal care, SRS 1–3 receive enhanced observation ± blood culture, and symptomatic or SRS >3 are treated with antibiotics (Goel et al., 2020).

No currently available test is able to provide perfect diagnostic accuracy, and false-negative as well as false-positive results may occur (Cobo et al., 2013; van Leeuwen et al., 2024). Thus, empirical antibiotic therapy remains the current practice in all neonates with suspicion of EONS (Cobo et al., 2013). However, empirical treatment also increases exposure to adverse drug effects and nosocomial complications and poses a high risk of developing resistant strains (Cobo et al., 2013).

Considering that the majority of studies have left a door open for searching new diagnostic tools that will help the clinicians in diagnosing EONS, this review aims to analyze if the cytokines determined from the UCB of newborns may help in the rapid and accurate diagnosis of EONS in newborns originating from pregnancies with maternal–fetal infectious risk.

We searched the PubMed, Scopus, and Web of Science databases for articles indexed from 1987 through 3 September 2024 that assessed the relationship between interleukins and EONS. We searched for these articles without restrictions on publication date. The search terms used were “interleukin” AND “early” AND “onset” AND “neonatal” AND “sepsis.” We aimed to include human-based studies conducted on neonatal populations, literature data published in English language, and research articles that analyzed a correlation between levels of cytokines determined from UCB and EONS. Exclusion criteria consisted of non-English language literature data, studies that did not meet our article objective, articles without available abstracts, duplicates, case reports, editorials, letters to the editor, review articles and meta-analyses, as well as experimental, animal studies.

The article selection process first consisted of exclusion of duplicates and then of the records written in non-English language or without available abstracts, a task which was performed by authors RMA and SMO. Each of the three authors of the article examined the title and abstracts of the identified reports to exclude irrelevant articles for the reviewer’s objectives. SMO and RMA accessed the full-length text of the manuscripts and checked for compliance to the inclusion criteria. Eventual disagreements between authors had been thoroughly debated and discussed by all authors. The inclusion of each individual record belonging to this review was established upon mutual agreement.

The following information was extracted from articles belonging to the final selection pool: author name, year of publication, type of study, population and study group assignment, and main findings of the article related to IL levels detected from UCB and the association between these ILs and EONS.

The search retrieved a total number of 1,052 records. After exclusion of 296 duplicate articles, 5 non-English language articles and 9 articles without an available abstract, a number of 742 articles were screened. Then, 670 articles that were not in line with the reviewer’s objectives were excluded (637—because they did not study the level of cytokines in EONS and 33—because they did not evaluate the level of cytokines in UCB). A number of 72 relevant articles were assessed, and after the exclusion of case reports, editorials, letters to the editor, review articles, meta-analyses, and experimental studies, 18 admissible articles were systematically selected. The article selection process is detailed in Figure 1 and was performed in accordance with the PRISMA 2020 statement (Page et al., 2021).

A summary of the main findings of the studies that complied with our election criteria is provided in Table 1 .

EONS resulting from bacterial bloodstream infections remains a serious clinical concern and is currently defined as a “life-threatening organ dysfunction caused by a dysregulated immune response that occurs as the result of an infection” (Shankar-Hari et al., 2016; Greenfield et al., 2021). The dysregulated immune response is obvious through the initial hyperinflammatory response driven by proinflammatory cytokines and chemokines (Shankar-Hari et al., 2016; Greenfield et al., 2021). Campos et al. argued that routine cytokine assays entail increased cost, and they have not reliably established reference ranges, as cytokine levels vary widely from person to person and are highly influenced by external factors (Campos et al., 2010).

Cytokines may have great potential for diagnosis of EONS and have been studied for many years. The promising determination of cytokines from UCB in cases of maternal–fetal infectious risk remains a challenge because of the variety of the study populations enrolled, which could have led to such different results between studies. IL-6 seems to be the cytokine that was mostly studied and also IL-6 may be a good predictor of EONS. This statement was sustained by the studies of González-Andrade et al. and Basu et al. that described UCB IL-6 concentrations as a highly sensitive and specific early diagnostic parameter of EONS (Basu et al., 2012; González-Andrade et al., 2019). However, the study population of González-Andrade et al. was represented by full-term newborns, while that of Basu et al. was represented by preterm, LBW neonates (Basu et al., 2012; González-Andrade et al., 2019). Available conventional sepsis screening markers, including hematologic indices and CRP, were not reliable early markers for EONS (Basu et al., 2012).

IL-6 concentrations in UCB, and to a lesser degree in maternal blood, that were taken during delivery were very sensitive, reliable, and early markers of EONS. They could provide an accurate indication of whether a neonate will develop EONS, thus, offering the opportunity for prompt diagnostic and aggressive therapeutic intervention (Hatzidaki et al., 2005). In contrast to this study, Mercer et al. stated that initial maternal plasma markers were not associated with adverse neonatal outcomes; maternal plasma markers evaluated at delivery are less closely associated with neonatal complications than UCB markers (Mercer et al., 2012). UCB reflected the intrauterine fetal environment and was associated with subsequent infectious and composite morbidity (Mercer et al., 2012). According to the authors, the combination of levels of multiple markers and clinical findings could further refine the predictive value of these tests, but their sample size was inadequate to evaluate this (Mercer et al., 2012). The higher levels of IL-6 and G-CSF seen in UCB compared to maternal blood suggested strong fetal/placental participation in the inflammatory response related to intrauterine infection and preterm parturition; these elevations reflected a subclinical inflammatory process that was present at the time of initial evaluation of these women (Mercer et al., 2012).

An increase in IL-8 in UCB might indicate fetal inflammation, but no significant increase was seen in regard to EONS in the newborns with PPROM, when compared to the control group (p > 0.05) (Satar et al., 2008).

Cernada et al. speculated that performing analytical assays in UCB could avoid repeated blood extractions in the first hours of postnatal life. The values of IL-6 can be ready in 30–60 min as part of the routine clinical laboratory, so it could be useful to take decisions in the first hours of life depending on these biomarkers (Cernada et al., 2012). A limitation of their study was the fact that the authors enrolled preterm (50% of neonates were preterm) and full-term newborns that had completely different immune responses to EONS (Cernada et al., 2012). Similar to Cernada et al., Lehrnbecher et al. concluded that IL-6 and IL-8 may indicate EONS earlier than CRP (Lehrnbecher et al., 1996).

Excessive inflammation in EONS might be involved in the damage associated with EONS (Chen et al., 2022). Several studies have demonstrated that serum pro-inflammatory TNF-a, IL-1β, IL-6, and IL-8 levels are rapidly and strikingly elevated in EONS (Kurt et al., 2007; Wu et al., 2016; Leal et al., 2019). In contrast to these studies, Perenyi et al. concluded that elevated UCB levels may predict maternal chorioamnionitis, rather than the EONS (Perenyi et al., 1999). IL-6 becomes detectable within 1 h after triggering an infectious stimulus, reaches its plateau in 4 to 6 h, and the IL-6 release may be essentially abolished by 36 h, even while the infection remains unchallenged; as IL-6 blood level drops, acute phase reactants (such as CRP) increase (Perenyi et al., 1999).

Septic neonates had significantly higher UCB levels of IL-6 and IL-8 than non-septic neonates. But determinations of IL-6 in neonatal peripheral blood and from UCB showed comparable results in indicating EONS (Lehrnbecher et al., 1996). The association of IL-6 and IL-8 detected from UCB with EONS was also sustained by El Farargy et al. But in their study, EONS was more common in preterm than in full-term neonates (p = 0.007) (El Farargy et al., 2018).

Weeks et al. speculated that UCB IL-6 levels could assist neonatologists in determining which infants were at greatest risk for infectious morbidity, NEC, and high-grade IVH. The authors hypothesized that UCB IL-6 levels could be useful in directing the use of diagnostic tests and determining which neonates require postnatal therapies, such as antibiotics (Weeks et al., 1997). Neonates with EONS had higher levels of inflammatory and oxidative stress markers in UCB before the onset of symptoms; UCB IL-6, IL-10, and oxidative stress markers were significantly higher in infants with EONS, even in the proven sepsis group (Cancelier et al., 2009).

Recent studies have brought the following new markers to the scene: IL-18 and IL-27. A significant increase in IL-18 in the UCB was observed in the sepsis group, but all the other markers that were investigated did not show significant differences (Janec et al., 2023). IL-27 had a better performance in distinguishing neonates with true infection from neonates without infection, and a combined performance of IL-27 with PCT in the UCB and at 24 h of life showed greater prediction of EONS with p < 0.01 for each parameter/parameter combination (Sultan et al., 2023). IL-27 had an increased risk of EONS with an odds ratio of 9.13 and p < 0.01 (Sultan et al., 2023).

Nakstad et al. stimulated UCB samples from 20 healthy term pregnancies for 2 h with a GBS III isolate from a patient and a commercially available GBS Ia strain; nonstimulated samples served as controls (Nakstad et al., 2016). The authors concluded that IL-6, IL-8, and potentially CD11b could be useful in diagnosing neonatal GBS infection in an early stage (Nakstad et al., 2016). Only in the study of Berner et al. were the levels of UCB IL-8 more elevated after cesarean section (114.9 vs. 23.1 pg/ml, p < 0.001) (Berner et al., 2000).

Neither serum IL-6 nor IL-10 alone was a useful diagnostic index of EONS in equine neonates, but the number of animals involved in the study was too small (15 septic neonatal foals vs. 15 age-matched control foals) for the identification of a concrete value; the serum IL-6:IL-10 ratio was likely to provide a valuable prognosticator for EONS (Burton et al., 2009).

Intrapartum antibiotic exposure can result in a partially treated infant delaying the onset of clinical signs and symptoms of infection and further complicating the expedient definitive diagnosis of EONS (Najati et al., 2009). This information was in opposition to that of Mercer et al. who claimed that evaluated maternal systemic inflammatory markers were marginally affected by antibiotic treatment during conservative management of PROM remote from term. That evaluation of these systemic inflammatory markers at initiation of treatment was not likely to provide adequately predictive information regarding subsequent adverse maternal and neonatal outcomes. While increased levels of the evaluated UCB pro-inflammatory cytokines were associated with subsequent adverse neonatal outcomes, significant overlap in values between newborns with and without subsequent complications limited the predictive and clinical value of these markers for use in clinical practice; the authors concluded that it is plausible that the inflammatory cascade cannot be stopped once initiated (Mercer et al., 2012).

Unfortunately, there is no adequate laboratory test, single or in combination, which has sufficient sensitivity and specificity to allow neonatologists to safely rule out EONS. For that reason, all newborns with clinical signs that are suggestive of EONS should be treated empirically with antibiotics, after peripheral and blood cultures have been taken. Careful observation of apparently good-looking full-term newborns that are born from pregnancies with infectious risk is fundamental during the first 24 h of life. The main limitations of current studies remain the small study samples, the heterogeneous study populations included, and the lack of stratification of the studied population according to GA. GA groups have different risk factors that may be associated with EONS. The results might have differed if the GA-dependent study group division would have been considered as follows: full-term neonates—more than 37 weeks of gestation, preterm neonates with GA less than 24 weeks of gestation, 24–28 weeks of gestation, 28–32 weeks of gestation, 32–34 weeks of gestation, and 34–37 weeks of gestation. Another heterogeneity of the studies was caused by different laboratory techniques used to quantify cytokine levels. Future studies should focus on establishing standardized cut-off values for IL-6 and IL-8 accounting for variations in GA and maternal risk factors.

Practical applications of the findings are huge because the collection of blood from UCB at birth is minimally invasive for the newborn. Cytokines detected from UCB may predict from birth which newborns are at risk of presenting EONS. However, we need precise cut-off values of IL-6 and IL-8 to use them in clinical practice. UCB levels of IL-6 and IL-8 may be cost effective, and this approach of predicting EONS may reduce unnecessary antibiotic use.

It is well known that EONS contributes to morbidity and mortality in the neonatal period, so that early diagnosis and early and targeted treatment may improve its prognosis and outcome. Testing blood from the UCB is less invasive. The cytokines that seem to be more accurate in early diagnosis of EONS in many studies are IL-6 and IL-8. IL-6 and IL-8 levels have been shown to be rapidly released in response to infection often before the onset of clinical symptoms and before routine positive laboratory tests. The incidence of EONS varies in countries around the world. There is a need for further multicenter, large-sized studies, such as prospective cohort studies that will assess the role of IL-6 and IL-8 detected from both maternal blood and UCB in early diagnosis of EONS and in reducing the abuse of antibiotics. The methodology needs to take account of GA stratification. GA-dependent references for each marker are also claimed. Moreover, validated cut-off values of these cytokines may guide clinicians in the careful administration of antibiotics which will overall reduce antimicrobial resistance. The potential integration of cytokine biomarkers with machine learning models is promising and should also be considered in further studies.