*1.1.2. Potential confounders in the manual assessment of IG by microscopic view of a manual smear*

Anyhow, the detection of IGs by microscopic count necessitates experienced laboratory staff; furthermore morphological classification of IGs are subject to a considerable reader bias and interpretative errors; especially in neonates where leukocytosis occurs frequently in the first days of life this method seems to provide only limited reproducibility (Chiesa et al., 2004; Schelonka et al., 1995). Contrariwise in performing a standard 100-cell manual differential small numbers of IGs are often underestimated as they can often be overlooked in samples of leukopenic patients. Another study highlighted the wide range of inter- and intraobserver variance in microscopic band cell identification: A smear of a blood sample from a septic patient was prepared, stained and a PowerPoint presentation was made twice of 100 random cells and sent to 157 different hospital laboratories in the Netherlands for a leukocyte differential. In the rst survey neutrophils were differentiated in segmented and band neutrophils whereas in the second survey no discrimination was made between segmented and band neutrophils. Albeit the morphologic characteristics of a band cell are well defined, this study showed an enormous intervariability of enumeration of band cells so that the authors recommended to cease quantitative reporting of counted band cells especially in regard to other diagnostic tools like C-reactive protein (CRP), procalcitonin, and cytokines (van der Meer et al., 2006). Hence, several authors consider the manual count as inappropriate as a reference method for detection of IGs (Fernandes & Hamaguchi, 2007; Senthilnayagam et al., 2012).

62 Neonatal Bacterial Infection

should be taken into account.

*manual smear* 

and death from sepsis (Christensen et al., 1981).

between 25% and 50% in non-infected infants (Polin, 2012).

higher the degree of elevated IT-ratio was, the higher was the risk of bone marrow depletion

In a retrospective multicenter cohort analysis including 166092 neonates with suspected EOS admitted to 293 NICUs in the United States low WBC counts, low absolute neutrophil counts, and high IT-ratios were associated with increasing odds of infection. Elevated ITratios of >0.2, >0.25, and >0.5 had low sensitivities (54.6%, 47.9%, 21.9%, respectively), but were associated with relatively high specificities (73.7%, 81.7%, 95.7%, respectively) and negative predictive values (NPV) (99.2%, 99.2%, 99.0%, respectively), whereas positive predictive values (PPV) were low (2.5%, 3.2%, 6.0%, respectively). The authors concluded that due to the low sensitivity these CBC-derived indices do not represent reliable diagnostic markers to rule out EOS in neonates (Hornik et al., 2012). The very high negative predictive accuracy of more than 90% is in contrast to high rates of elevated IT-ratios

In a large historical cohort study comprising more than 3100 neonates, patients were evaluated for EOS. In this study a normal WBC was defined as an IT-ratio of less than 0.2 and a total WBC between 6000 and 30000/µL. Two serial normal WBCs with normal ITratios performed 8 to 12 hours apart and a negative blood culture at 24 hours were predictive of healthy newborns in the evaluation of EOS in the rst 24 hours after birth and showed a negative predictive value of 100%. The sensitivity of 2 normal WBCs and a negative blood culture at 24 hours was 100%, as was NPV. The specificity was 51%, and the PPV was 8.8% (Murphy & Weiner, 2012). These results suggest that combinations of parameters and repeated performance of diagnostic tests are likely to increase accuracy.

In a review article Cornbleet reported a wide range of sensitivity and specificity for the ITratio and predicted a possible replacement by the measurement of newly created markers for infection such as inflammatory factors, adhesion molecules, cytokines, neutrophil surface antigens, and bacterial DNA (Cornbleet, 2002). Recent advances in basic science of predicting and diagnosing neonatal sepsis are developing towards more and more sophisticated approaches like the determination of proteomic inflammatory biomarkers in amniotic fluid (Buhimschi et al., 2009). Regarding these new techniques, the diagnostic value of traditional laboratory methods has to be critically analysed. However, comparing these new methods for the detection of neonatal sepsis with the measurement of WBCs including the assessment of the IG count (IGC) as well as the IT-ratio, the additional sample volumes, delayed availability of results, and considerably higher labour and laboratory costs

*1.1.2. Potential confounders in the manual assessment of IG by microscopic view of a* 

Anyhow, the detection of IGs by microscopic count necessitates experienced laboratory staff; furthermore morphological classification of IGs are subject to a considerable reader bias and interpretative errors; especially in neonates where leukocytosis occurs frequently in the first days of life this method seems to provide only limited reproducibility (Chiesa et al.,
