Abstract

MESQUITA, Mirta; AVALOS, Sonia; GODOY, Laura  and  ALVAREZ, Elizabeth. Predictive Value of the Complete Blood Count (CBC) in Neonatal Sepsis. Pediatr. (Asunción) [online]. 2011, vol.38, n.1, pp.23-30. ISSN 1683-9803.

ABSTRACT Introduction: Diagnosis of neonatal sepsis remains a challenge. There is no single laboratory test to confirm diagnosis. In recent years the usefulness of various markers has been investigated, but always in conjunction with blood count values (white count, neutrophils etc.). Objective: The purpose of this study is to analyze the complete blood count, a diagnostic tool available in almost all public hospitals, and assess its usefulness in neonatal sepsis. Methodology: A study of diagnostic tests performed at the Niños of Acosta Nu General Pediatric Hospital between January 1, 2008 and December 31, 2009. We analyzed the medical records of newborns clinically diagnosed with sepsis and confirmed by two positive blood cultures. The CBC taken was analyzed as an auxiliary method of diagnosis. Results: 209 newborns were admitted through the hospital’s emergency or pediatric multidisciplinary intensive care units. A study group was formed of the 30 patients with positive blood cultures and a control group from 40 patients with no suspected infection. In the sepsis group leucopenia was shown in 13% (4/30) vs. 10% (4/40) in the control group while leukocytosis was seen in 23% (7/30) of the sepsis group vs. 35% (14/40) in the control group. For leukopenia sensitivity was 13%, specificity 90%, positive predictive value (PPV) 50%, and negative predictive value (NPV) 58%. The positive probability ratio was 3 and negative probability ratio 0.96. In the sepsis group, only 13% (4/30) presented neutropenia vs. 12.5% (5/40) in the control group; neutrophilia (>17500) was observed in 41% (13/30) of the sepsis group and 40% (16/40) of the control group, while in regard to the cutoff points, for neutropenia sensitivity was 13%, specificity 87%, PPV 44%, NPV 57%, positive probability ratio 1, and negative probability ratio 1, while for neutrophilia sensitivity was43%, specificity 80%, PPV 44%, NPV 58% positive probability ratio 2, and negative probability ratio 0.71. Platelets were also analyzed, with thrombocytopenia (<150,000) found in 20% (11/30) in the sepsis group vs.7.5% (3/40) in the control group. The average platelet counts in the sepsis group was 200,000±139,897 while in the control group it was 340,825±15.0417 (p= 0.0002). For thrombocytopenia sensitivity was 36%, specificity 92%, PPV 78%, NPV 66% positive probability ratio 4.5 and negative probability ratio 0.6. Analysis of true positives and false positives using the ROC curve showed an area under the curve for leukocytes of 0.50 (CI 95% 0.36-0.65), for neutrophils 0.57 (CI 95% 0.43-0.71), and for platelets of 0.76 (CI95%0.64-0.88). Conclusion: The CBC alone is not very useful for the diagnosis of neonatal sepsis. Of the blood count values studied, the presence of thrombocytopenia was seen to be a useful tool for diagnosis of neonatal sepsis.

Keywords : Neonatal sepsis; diagnosis; blood counts; thrombocytopenia.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )