Sepsis of vertical transmission in the newborn infant has been a nearly constant concern in Neonatology, but development of nosocomial infections from acquisition of bacterial and other microbial pathogens in the nursery has become a challenging complication as a result of increased survival of VLBW infants or premature infants affected with severe diseases. However, there are relatively few studies of the value of serum PCT for the diagnosis of nosocomial neonatal sepsis, and none of them is a multicenter one. Monneret et al.  assessed daily variations in serum PCT in comparison with C-reactive protein (CRP) in 94 control and infected newborn infants, with 14 infected infants and 17 controls between the 4 and 28 days of life. Although in this study, procalcitonin seems to be an interesting marker of neonatal sepsis (early PCT peak compared with CRP), data to estimate diagnostic reliability are lacking. Chiesa and associates [9, 21] reported the results of a study that included 23 cases of nosocomial infection and 92 asymptomatic controls between 3 and 30 days of life, in which serum PCT concentration discriminated all cases of sepsis with a 100% sensitivity and specificity. All infected infants had PCT values of ≥ 2 ng/mL and all controls ≤ 1 ng/mL. Enguix et al.  evaluated serum PCT as a diagnostic marker of bacterial sepsis in newborns aged 3–30 days admitted to the NICU (20 neonates with sepsis, 26 neonates without sepsis), and found a sensitivity of 98.6% and specificity of 88.9% for an optimum diagnostic cutoff value of ≥ 8.05 ng/mL, without statistically significant differences in comparison with CRP and serum amyloid. Kawczynski et al.  evaluated PCT and CRP in 48 newborn infants who suffered from nosocomial sepsis, reporting sensitivity of 89.6% for PCT and 66.7% for RCP at the onset of sepsis, that improved to 100% and 89.6% respectively 24 hours later. Unfortunately, only septic newborns where included, so it was no possible to calculate specificity. More recently, Vazzalwar et al  assessed PCT for the diagnosis of late-onset sepsis in 67 very low birth weight infants. At a PCT cutoff value of 1.0 ng/mL sensitivity was 97% and specificity 80%, while with CRP sensitivity was 72% and specificity 93%.
In the present study, the diagnostic reliability of serum PCT was modest, with sensitivities and specificities slightly higher than 80% at the time of suspicion of nosocomial infection. It should be noted that in the studies of Enguix et al.  and Chiesa et al. , a control group formed by asymptomatic infants without evidence of infection or with a diagnosis on admission easily differentiable from nosocomial sepsis was included, which may overestimate the reliability of a diagnostic test . We studied a homogeneous group of neonates with clinically suspected nosocomial sepsis in which pre-established criteria for the definitive diagnosis of sepsis were applied, an approach that closely resembles the clinical scenario where the diagnostic test is intended to be used. On the other hand, because of the requirement of strict microbiologic criteria for the final diagnosis of sepsis, it may be possible that some cases of true bloodstream infection could not have been confirmed due to false negative blood cultures.
The use of PCT for the diagnosis of sepsis of vertical transmission is influenced by the physiologic peak of this marker during the first 48 h of life [9, 27], but this phenomenon is obviated because nosocomial sepsis develops on a later time. It has been reported that elevations of PCT in neonates infected with CoNS are lower than serum PCT increases in sepsis caused by other pathogens . In our study, serum PCT concentrations were significantly lower in the group of 22 patients with infections caused by CoNS than in the remaining 39 patients with infections of other etiologies. This is a relevant finding given the high frequency of infections caused by CoNS in the NICU.
Diagnostic reliability of serum PCT was not compared with CRP or other infection markers, such as leukocyte count, given that these laboratory tests were not standardized and were performed according to techniques of each participating hospital, and they were used to define sepsis. Although this may be considered a limitation of the present findings, the study was to assess the diagnostic usefulness of PCT as a single marker of neonatal sepsis of nosocomial origin. In addition, a comparative study would have required a bigger sample size in order to detect differences between infection markers.
Studies about diagnostic tests should be carried out in a patient sample that includes an appropriate spectrum of patients to whom the diagnostic test will be applied in clinical practice, and compared with a sufficiently reliable "gold standard" . Unfortunately, an ideal "gold standard" for the definition of neonatal sepsis is not available and, for this reason, different definitions according to clinical, microbiologic, or laboratory criteria are used. Recently, the international pediatric sepsis consensus conference  modified the adult systemic inflammatory response syndrome (SIRS) criteria for children, incorporating pediatric physiologic variables for the subcategories of newborn, neonate, infant, child, and adolescent. However, premature infants were not included in the newborns group. Therefore, from the NICUs perspective, reference values of heart rate, respiratory rate, leukocyte count, and systolic blood pressure included in the definition of SIRS should be conveniently updated, as well as evidences (clinical, microbiologic, laboratory, radiologic) required to consider infection as the cause of SIRS. On the other hand, little guideline or consensus exists in literature for the differentiation between neonatal sepsis of vertical transmission and nosocomial sepsis, usually defined as early-onset and late-onset sepsis. The chronologic criteria, however, is inadequate for cases of early-onset nosocomial sepsis or late-onset bloodstream infection of vertical transmission that have a different profile of causative organisms and distinct therapeutic approach. An international consensus definition of these relevant aspects of pediatric sepsis will facilitate the performance of clinical studies in neonates with sepsis and the application of the results obtained in the daily practice.