Fig 1.
Study design and patient selection.
Of 1100 enrolled infants, 12 had confirmed early onset sepsis (cEOS: positive blood culture with bacterial neonatal sepsis pathogen within 72 HOL and received antibiotic treatment ≥5 days). Presumed sepsis (PS) subjects were treated for early infection with antibiotics within first 72 HOL, had no positive sterile site culture during NICU stay, and had ≥2 abnormal lab criteria (↓ANC, ↑ I:T ratio, ↑CRP). Late onset sepsis (LOS) subjects had positive blood culture after 72 HOL treated with antibiotics. Control patients had no sterile site culture or antibiotic course >4 days throughout hospitalization. Patients with ≥2 infection episodes were excluded. PS, LOS and control patients were frequency matched to cEOS patients within target gestational age and birth weight ranges (±3 wks and ±400 gms). Infants were randomly selected from the group fitting GA/BW criteria.
Table 1.
Case definition, demographics and placental histopathology.
Fig 2.
Acute phase reactant levels in sepsis groups.
Box and whisker plots displaying distribution of 9 APR biomarkers (SAA, CRP, Hp, ferritin, SAP, PCT, tissue plasminogen activator, fibrinogen, α-2-macroglobulin) in each sepsis category. *Indicates significant difference in APR values between cEOS and control groups. **Indicates significant difference in APR values between cEOS and PS groups.
Table 2.
Cord blood acute phase reactants: comparison of cEOS to other sepsis groups and controls.
Fig 3.
Serum amyloid A, C-reactive protein, and haptoglobin levels in sepsis groups by AI.
Box and whisker plots displaying distribution of 3 APR biomarkers (SAA, CRP, Hp) in each sepsis category with and without fetal AI [bottom row] and any placental AI (maternal and/or fetal) [top row].
Fig 4.
Receiver operating characteristic curves.
ROC curves for serum amyloid A, C-reactive protein, and haptoglobin. The above acute phase reactants had areas under the curve of >95% for prediction of EOS.
Table 3.
Receiver operating characteristic analysis: AUC and biomarker cut-offs.
Table 4.
Linear regression model for sepsis effect on acute phase reactants.