In this study, we measured P-SEP values in cord blood of term and preterm infants with risk factors for EOS, describing a significant correlation between cord presepsin and clinical early-onset sepsis.

Previously, Seliem and Sultan investigated whether presepsin levels in umbilical cord blood can be used as a predictor of EOS in preterm labor with premature rupture of membranes (PROM). They included only preterm infants born between 24 and 36 weeks of gestation, finding a higher P-SEP in cases with EOS (2231 pg/ml) than in those without (275 pg/ml) [10]. We confirm this trend, with higher P-SEP values in infants with clinical sepsis (909 pg/ml) rather than in infants without (467 pg/ml), but we included not only preterm neonates but also those term-born.

Another difference between our study and their one was the method to measure presepsin blood levels: they stored centrifuged samples (at 1000 × g for 15 min) at -20 °C until analysis and then measured P-SEP values using enzyme-linked immunosorbent assay (ELISA; Abbexa Ltd., Cambridge, UK). Instead, a step forward in the plasmatic measure of P-SEP is represented by a novel, highly sensitive and fully automated method, based on the chemiluminescence (CLEIA) method, providing results in 17 min in six samples simultaneously by a Point-of-Care Testing (POCT) instrument [8, 15]. In our study, we immediately analyzed cord blood samples using this new method although ELISA. The feasibility of this point-of-care method in measuring P-SEP values in EOS has been tested in different NICUs [9, 16,17,18,19,20]: this new method could be considered in the panoply of EOS risk assessment strategies used within neonatal units.

Despite a gradual improvement in trends, neonatal sepsis continues to be a major cause of morbidity and mortality, especially in the VLBW (Very Low Birth Weight) population, with an incidence of 1–4 / 1000 live births [21]. Thus, the early diagnosis of neonatal sepsis is essential due to the rapid evolution of the clinical situation. An early, sensitive and specific laboratory test would be helpful to guide clinicians in deciding whether or not to start antibiotics, avoiding unnecessary treatment, considering that more than 75% of VLBW infants receive prophylactic antibiotics because of the presence of EOS risk factors [22]. The use of antibiotics is not free from drug-related risks, but also risks related to the need for venous access, as well as the discussed risk of necrotizing enterocolitis [23]. Furthermore, exposure to antibiotics via maternal intrapartum antibiotic administration and neonatal administration after birth both alter the composition of the newborn's microbiome through infancy [22].

Currently, the gold standard for the diagnosis of sepsis is blood culture, although EOS is culture-confirmed in only about 1% of VLBW newborns: this rate was 20 times higher than that found in neonates born with birth weights higher than 2500 g according to Stoll’s findings [24]. Furthermore, response times require at least 48–72 h, and the percentage of false negatives related, in most cases, to intrapartum maternal antibiotic therapy, or the cases of false positives associated with contamination of the sample at the time of collection, is not negligible. The consequence of all this is that newborns with risk factors or in the presence of clinical suspicion are all treated with antibiotic therapy without any distinction.

Presepsin has now been widely tested as a marker of sepsis in term and preterm infants [8]. The availability of reference ranges of P-SEP values in the blood of term and preterm neonates has led to greater use of this marker in clinical routine. Pugni et al. found a median value of 603.5 pg/mL in the blood of uninfected term infants, whereas a median value of 620 pg/mL in that of uninfected preterm infants. The reference ranges of presepsin they determined were much higher than those seen in healthy adults, in whom a cut-off value of 600 ng/L has been reported for the discrimination of bacterial sepsis with a sensitivity and specificity of 87.8% and 81.4%, respectively [25]. Among the eventual causes, there is the activation of the innate immune system after birth, which occurs as a result of the shift from a usually sterile intrauterine environment to a world rich in foreign antigens. Furthermore, the newborn’s skin and stomach are rapidly colonized with microbial flora after birth, providing continual stimulation to the innate immune system [26].

Comparing cord P-SEP values of our infants to reference ranges of blood P-SEP values, we found that infants with clinical EOS had at birth higher values (greater than 75th centile), while infants who did not develop clinical sepsis had initially lower values (lower than 50th centile).

An interesting study has recently shown how high P-SEP values early correlate with sepsis’ onset, thus making it possible to identify newborns at higher risk, intervening before the presentation of clinical symptoms, and reducing the doses of antibiotics in lower-risk infants [20].

In recent years, evidence has emerged about using cord blood as a possible diagnostic tool for early sepsis. The measurement of acute phase proteins and cytokines such as CRP, PCT, interleukin-6 (IL-6), interleukin-8 (IL-8), tumor necrosis factor α (TNFα), and interleukin-1β (IL-1β) let us assess the fetal inflammatory response in utero [17], even if from currently available studies none of these markers would be able to confirm or exclude the diagnosis of EOS in the newborn [27].

Seliem and Sultan’s findings about cord presepsin in preterm infants who late developed EOS compared to their healthy peers were interesting: however, beyond the inclusion of only preterm infants, they did not report a cut-off of cord P-SEP to consider [10].

Conversely, our study included infants of all gestational ages with coded risk factors, as per CDC guidelines [11].

Moreover, both infants with positive blood cultures and infants with suspected sepsis were included in the definition of EOS, also in this case, in the presence of coded clinical signs, according to well-defined criteria. This choice derives from the small percentages of positive blood cultures in the neonatal population and from the fact that negative blood cultures do not allow excluding EOS in the presence of a compatible clinical picture. According to this, the incidence of clinical sepsis in our population was 8.6% (8/93 cases), while only in 3/93 cases sepsis was culture-proven (3.2%). This is a major limitation of our study because a discrete proportion of neonates might be ill because of conditions different from sepsis, leading to an overestimation of presepsin accuracy. Therefore, considering the low incidence of culture-proven early-onset sepsis in real life, beyond risk factors, a multicentric study about the presepsin values in cord blood including only infants with culture-proven EOS should be conducted.

The use of cord blood for the dosage of presepsin made it possible to avoid invasive procedures on the newborn (venipuncture, withdrawal from the heel). Furthermore, the point-of-care reading method allows for almost immediate results and seems to be feasible to use in neonatology units. The heterogeneity of the sample examined, including both premature and full-term infants, did not compromise the accuracy of the test studied, thus suggesting a possible transversal use. We identified an ideal cut-off point of 579 pg/ml (579 pg/ml for term infants and 544 pg/ml for preterm infants if considered separately) as an accurate screening method in neonates with risk factors for EOS, possibly avoiding administrating antibiotic prophylaxis in those with low cord P-SEP values.

The positive predictive value of presepsin in cord blood remained quite low with any cut-off, with the eventual risk of many false positive results. Similarly, correlation coefficients (Kendall’s tau), despite significant differences, were not high. Indeed, our findings are limited by the small sample size due to the single-center design and the low number of true positives. We could not assess its capability to discriminate between septic and non-septic patients among clinically ill newborns, because the low number of septic patients and the overlapping clinical picture with other disorders. Therefore, further multicentric studies are needed to confirm our findings. However, the identified value of presepsin in cord blood does not differ significantly from the 50th percentile for presepsin in neonatal blood previously reported in the literature [14].