S100A8/A9 is the first predictive marker for neonatal sepsis

Abstract

Neonatal sepsis is a leading cause of childhood mortality worldwide particularly affecting preterm infants, who are often exposed to empirical antibiotics since sepsis biomarkers lack sensitivity in this patient group and markers predicting the risk of sepsis have not been identified yet.1, 2 In our study, serum S100A8/A9 proved as an independent predictive marker of late-onset neonatal sepsis (LOS) in preterm infants, which for the first time offers the opportunity to change current treatment policies by improving antibiotic stewardship and timely individualized therapeutic intervention. Serum S100A8/A9 (also known as serum calprotectin) is an alarmin respective damage-associated molecular pattern (DAMP) that is rapidly released from myeloid cells upon stress or cell damage, acting then as secondary amplifier of inflammation.3 In adults, the S100A8/A9 serum level is used as one of the most sensitive biomarkers in inflammatory processes.4 Its value as sepsis marker in neonates remains unclear and needs to be tested. On the other hand, we previously identified S100A8/A9 as an essential sepsis-protective regulator of neonatal immunity5, 6 with serum levels being physiologically increased after birth.7 Thereby, significantly higher levels in term compared to preterm infants6 suggest that low S100A8/A9 in neonates is associated with an increased risk of sepsis. However, other parameters potentially influencing neonatal S100A8/A9 serum levels are ill defined, wherefore the value of S100A8/A9 as an independent predictive marker of neonatal sepsis also remains to be demonstrated. A total of 289 preterm infants born below 32 gestational weeks were prospectively enrolled. In a case–control study, of 41 infants serum levels of S100A8/A9, C-reactive protein (CRP) and interleukin 6 (IL-6) were determined at the onset of neonatal sepsis and compared with 50 matched healthy controls (Table S1 and Methods in the Supporting Information). However, S100A8/A9 did not distinguish septic infants from healthy controls due to already elevated S100A8/A9 levels in nondiseased infants (Figure 1A). Consequently, serum S100A8/A9 did not outcompete sepsis marker like CRP and IL-6 (Figure 1B–D) currently used in the clinical routine. Next, we investigated the influence of the following common sepsis risk factors on S100A8/A9 serum levels during the first 2 days of life in a cohort of 198 at birth healthy preterm infants (Table S2 and Methods in the Supporting Information): delivery by cesarean section (CS), birth weight (BW) of <1000 g, BW small for gestational age (SGA) (below the 10th percentile), male sex, insertion of a central venous line (CVL), and the need of invasive mechanical ventilation (IMV)8, 9 as detailed in the Supporting Information. Children born by CS had significantly lower S100A8/A9 levels compared to those born by vaginal delivery (VD) (Figure 2A). Among infants born by CS S100A8/A9 levels were higher after secondary CS compared to primary CS (Figure S1), suggesting dependence of perinatal S100A8/A9 levels on labor-associated stress, which is in line with increased levels in trauma patients or marathon runners.7 In addition to the mode of delivery (MOD), also BW (Figure 2B), SGA status (Figure 2C), and sex (Figure 2D) were important determinants of the increase of S100A8/A9 serum levels at the beginning of life but not the need of CVL or IMV (Figure 2E and F). Subgroup analyses in preterm newborns delivered via CS, with a BW of <1000 g, SGA status, or male sex revealed that despite generally lower values (Figure 2A–D), S100A8/A9 levels at birth within these groups were still significantly lower in children with LOS compared to children that remained without LOS (Figure S2). To test the value of serum S100A8/A9 at birth as independent sepsis risk marker, we then performed a nested model comparison while accounting for all identified confounding factors. This analysis validated that serum levels of S100A8/A9 at birth highly significantly and inversely correlated with the occurrence of sepsis, independent of the MOD, BW, SGA status, and sex (Figure 3A). This is in accordance with the sepsis-protective function of S100A8/A9 in neonates shown previously.6, 7, 10 Calculating the effect sizes of all sepsis risk factors considered in this study and detailed in the Supporting Information revealed the strongest association of LOS with low S100A8/A9 levels compared to IMV, SGA status, low BW, male sex, MOD, and presence of CVL (Figure 3B). S100A8/A9 values within the normal range of adults, that is, below a cut-off level of 300 ng/ml, increased the OR of LOS to 8.3-fold (95%CI 4.15–16.72, p < 0.0001) with a positive predictive value of 0.92, a sensitivity of 0.89 and a specificity of 0.51. Above a S100A8/A9 level of 1000 ng/ml, the LOS risk decreased to 0.2 with a negative predictive value of 0.95 (OR 0.2, 95% CI 0.08–0.48, p = 0.0004). These cut-off values might be valuable for neonatologists enabling them to identify newborn preterm infants at low risk of LOS and to avoid unnecessary empirical antibiotic treatment. In this regard, S100A8/A9 is superior to CRP and IL-6 whose negative predictive power is limited due to the late peak of CRP after sepsis onset2 and the varying baseline levels of IL-6 in neonates.1 In summary, our data establish serum S100A8/A9 on the first 2 days of life as the first biomarker that reliably predicts the sepsis risk in preterm neonates, independent of patient characteristics. Useful S100A8/A9 cut-off values are provided that stratify preterm infants into high or low sepsis risk groups at the earliest perceivable moment after birth. Analysing postnatal S100A8/A9 serum levels is a simple clinical assay that could ensure timely sepsis management reducing mortality rates and aid in avoiding unnecessary antibiotic treatment improving antibiotic stewardship in this vulnerable patient group. We are grateful to all infants and their parents who participated in our study. This work was supported by the Interdisciplinary Centre of Clinical Research at the University of Münster [Vo2/011/19]; the Cluster of Excellence Cells in Motion and the Collaborative Research Centre 1009 [B08/B09]; and the Clinical Research Unit 342 [P3/P5] to TV and JR; by the Appenrodt Foundation; the Deutsche Forschungsgemeinschaft (DFG) [VI 538/6-1]; and the DFG under Germany's Excellence Strategy—EXC 2155–390874280 [RESIST]; and the Volkswagen Foundation [Az 90005] to DV. Sabine Pirr, Louise Dauter, and Dorothee Viemann conceived and designed the study; Sabine Pirr, Louise Dauter, Bettina Bohnhorst, and Dorothee Viemann acquired the data and collected and prepared the samples for analysis; Sabine Pirr, Louise Dauter, Thomas Vogl, Thomas Ulas, Johannes Roth, and Dorothee Viemann performed the analyses and interpreted the data; Sabine Pirr, Louise Dauter, and Dorothee Viemann drafted the manuscript; and all authors critically revised the manuscript. Sabine Pirr and Louise Dauter share first authorship and contributed equally to this study and manuscript. The authors report no potential conflict of interest. De-identified patient data from the results reported in this article will be made available for 5 years after publication upon request to the corresponding author. Researchers will need to provide a proposal and sign a data access agreement. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts