Background: Neonatal sepsis represents one of the common diseases in the neonatal intensive care unit. Here we aim to evaluate the differences between a group of preterm newborns with sepsis and a control group in relation to clinical and laboratory variables. In addition, our goal is to establish potential predictors of early-onset sepsis (EOS) and late-onset sepsis (LOS).
Methods: The study included 113 preterm newborns with sepsis (EOS-63.72%/LOS-36.28%). Laboratory deter-minations included full blood count, CRP, biochemical determinations, blood culture.
Results: The most important univariate neonatal predictors were gestational age (p < 0.001), surfactant adminis-tration (p < 0.001), mechanical ventilation (p < 0.001), heart failure (p < 0.001), a history of hypocalcemia (p = 0.037), Apgar score at 1 minute lower than 7 (p = 0.001), birth weight < 1,500 g (p = 0.005), number of hospi-talization days (p = 0.048), and number of weight recovery days < 10 (p < 0.05). The WBC and CRP parameters remained significant univariate predictors of sepsis on day 7 (p = 0.002; OR = 2.01 per 10,000 mm3 increase of WBC, 95% CI: (1.30; 3.09) and p = 0.001; OR = 4.27, 95% CI: (1.85; 9.88), respectively). Logistic regression anal-ysis showed maternal urinary tract infection (OR = 3.05), heart failure (OR = 5.28), the number of hospitalization days (OR = 1.09) and CRP (OR = 3.26) were significant independent risk factors for neonatal sepsis in preterms. The univariate predictors of EOS were gestational age (p = 0.002), birth weight (p = 0.014), 1-minute Apgar score (p = 0.012), maternal urinary tract infection (p = 0.008), surfactant administration (p < 0.001), heart failure (p < 0.001), and CRP level (p < 0.001). Surfactant administration (OR = 6.73) and CRP level (OR = 3.51) represent predictors of EOS in preterms according to the multivariate model. The univariate predictors of LOS were gesta-tional age (p = 0.001), birth weight (p = 0.048), 1-minute Apgar score (p = 0.001), surfactant administration (p < 0.001), hypocalcemia (p = 0.03), heart failure (p = 0.003), CRP level (p < 0.001), mechanical ventilation (p < 0.001), and the number of hospitalization days (p < 0.001). In the multivariate model, the number of hospitali-zation days (OR = 1.11) and heart failure (OR = 5.98) are independent predictors for LOS in preterms.
Conclusions: The study confirms the presence of maternal urinary tract infection, hospitalization days, heart fail-ure, and CRP level as predictors of neonatal sepsis in preterms with differences between EOS and LOS.