Predictive Factors and Its Usefulness in Early Diagnosis of Neonatal Sepsis

Lee, Young Hyuk; Kim, Seong Woo; Kim, Min Hee; Choi, Young Sook

Abstract

PURPOSE
Early diagnosis of neonatal sepsis is very difficult because of no specific clinical and laboratory findings. It also takes at least 48 hours of incubation period to isolate the organism by culture study. So several laboratory tests have been evaluated for their usefulness in rapid detection of the neonatal sepsis. Those are evaluated either singly or in combination with a defined scoring system include leukocyte count with differential count, platelet count, C-reactive protein level, erythrocyte sedimentation rate, haptoglobin level, fibronectin level, leukocyte alkaline phosphatase and so on. But no single test or combination with others has proved superior to the leukocyte count and differential count as a reliable indirect indicator of neonatal bacterial infection. We performed this study to determine the appropriate screening test for early detection of neonatal sepsis.
METHODS
During the period of May 1991 through April 1997, we selected 200 neonates who were admitted to the neonatal intensive care unit of Kon-Kuk University Medical Center Seoul Hospital. All of the cases were retrospectively evaluated and divided two groups; sepsis group-88 neonates who were confirmed by blood cultures, and control group-112 neonates who had no evidence of neonatal bacterial infection.
RESULTS
The results were as follows; 1) The sex ratio of male to female was 1.5:1 in the sepsis group and showed significant difference between two groups (P<0.05). The incidence of neonatal sepsis in prernature infant was higher in sepsis group than control group (P<0.05), and mean body weight was lower in sepsis group (2351.4148.3g) than control group (Z821.8 142.6g) significantly (P<0.05). 2) Predisposing perinatal factors associated with neonatal sepsis were premature rupture of membrane (> or = 24hrs) (14.5%) meconiurn staining (6.8%), asphyxia (Apgar score < or = 6 at 5mins) (5.7%), eclampsia and preeclampsia (4.5%), maternal infection (3.4%) and bleeding (including placenta previa, abruptio placenta) (1.1%) in order of frequency. Among th, only premature rupture of membrane was significant difference between two groups (P<0.05). Others risk factors including umbilical catheterization, endotracheal intubation, ventilatory care, total parenteral nutrition were also signi- ficant difference between two groups (P<0.05). 3) The common presenting symptoms of neonatal sepsis were jaundice (48.9%), poor feeding (45.5%), ternperature instability (43.2%), lethargy (30.7%), irritability, dyspnea, diarrhea, vomiting, tachypnea, and cyanosis in order of frequency. Among the above symptoms, poor feeding, dyspnea and cyanosis were significant difference between two groups (P<0.05). 4) The peripheral blood findings (leukocyte count, platelet count, ESR) showed no significant differences between two groups (P>0.05). The acute phase reactants (APR) score above two (37/88) and positive C-reactive protein (51/88) in the sepsis group were regarded as significantly high compared to the control group. 5) In the cases with APR score above two including positive C-reactive protein and abnormal total leukocyte count, sensitivity was 17.0%, specificity 97.3% positive predictive predictive value 83.3%, and negative predictive value 60.0%.
CONCLUSIONS
The higher frequency of neonatal sepsis was proved in the cases of APR score above two including positive C-reactive protein. In the cases with abnormal total leukocyte count and APR score above two including positive C-resctive protein, the specificity was 97.3% and the positive predictive value was 83.8%. So APR score above two including positive C-reactive protein and abnormal total leukocyte count could be regarded as an useful test method for early detection of neonatal sepsis.