Procalcitonin is not sufficiently reliable to be the sole marker of neonatal sepsis of nosocomial origin

Background: It has recently been suggested that serum procalcitonin (PCT) is of value in the diagnosis of neonatal sepsis, with varying results. The aim of this prospective multicenter study was to assess the usefulness of PCT as a marker of neonatal sepsis of nosocomial origin. Methods: One hundred infants aged between 4 and 28 days of life admitted to the Neonatology Services of 13 acute-care teaching hospitals in Spain over 1-year with clinical suspicion of neonatal sepsis of nosocomial origin were included in the study. Serum PCT concentrations were determined by a specific immunoluminometric assay. The reliability of PCT for the diagnosis of nosocomial neonatal sepsis at the time of suspicion of infection and at 12–24 h and 36–48 h after the onset of symptoms was calculated by receiver-operating characteristics (ROC) curves. The Youden's index (sensitivity + specificity 1) was used for determination of optimal cutoff values of the diagnostic tests in the different postnatal periods. Sensitivity, specificity, and the likelihood ratio of a positive and negative result with the 95% confidence interval (CI) were calculated. Results: The diagnosis of nosocomial sepsis was confirmed in 61 neonates. Serum PCT concentrations were significantly higher at initial suspicion and at 12–24 h and 36–48 h after the onset of symptoms in neonates with confirmed sepsis than in neonates with clinically suspected but not confirmed sepsis. Optimal PCT thresholds according to ROC curves were 0.59 ng/mL at the time of suspicion of sepsis (sensitivity 81.4%, specificity 80.6%); 1.34 ng/mL within 12–24 h of birth (sensitivity 73.7%, specificity 80.6%), and 0.69 ng/mL within 36–48 h of birth (sensitivity 86.5%, specificity 72.7%). Conclusion: Serum PCT concentrations showed a moderate diagnostic reliability for the detection of nosocomial neonatal sepsis from the time of suspicion of infection. PCT is not sufficiently reliable to be the sole marker of sepsis, but would be useful as part of a full sepsis evaluation. Published: 18 May 2006 BMC Pediatrics 2006, 6:16 doi:10.1186/1471-2431-6-16 Received: 26 January 2006 Accepted: 18 May 2006 This article is available from: http://www.biomedcentral.com/1471-2431/6/16 © 2006 Sastre et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Page 1 of 7 (page number not for citation purposes) BMC Pediatrics 2006, 6:16 http://www.biomedcentral.com/1471-2431/6/16 Background Infections of nosocomial origin are one of the most serious problems in modern neonatal units. Nurseryacquired infections are associated with increased mortality rates, prolonged duration of hospitalization in survivors, and high patient care expenditures. Because very low birth weight (VLBW) infants in neonatal intensive care units are at high risk for nursery-acquired infections, the frequency of these infections has increased in recent decades as a result of increased survival of immature neonates [1]. In a previous study of 30,993 admissions to neonatal units of 27 acute-care hospitals in Spain, the nosocomial sepsis rate was 2.1% with an incidence density of 0.89 per 1000 patient days. Sepsis rate was 15.6% among VLBW infants and 1.16% among those weighing ≥ 1500 g [2]. The prevention and control of these infections are thus major challenges for neonatal intensive care units (NICUs). Rapid diagnosis, however, is problematic because the earliest signs of nosocomial infection may be minimal and are similar to those of various noninfectious conditions. Bacterial cultures are time-consuming and other laboratory tests are either not available for routine use or lack sensitivity or specificity. In this situation, neonates with risk factors for infection or clinical suspicion of infection are empirically treated with antibiotics. On the other hand, the presence of multiresistant nursery flora complicates the choice of antimicrobials. To avoid unnecessary treatment of noninfected neonates, an early, sensitive and specific laboratory test would be helpful to guide clinicians in neonatal units in deciding whether or not to start antibiotics. Over the past few decades, several markers of neonatal infection especially leukocyte indexes and acute-phase reactants, some of which are commonly used in clinical practice, have been studied. However, there is less information on the value of these markers for the diagnosis of sepsis of nosocomial origin compared with neonatal sepsis of vertical transmission, and, to date, no single laboratory test has provided rapid and reliable identification of early infected neonates. This inability has led to search for new diagnostic markers [3,4]. It has been recently reported that procalcitonin (PCT), the prohormone of calcitonin, increases markedly in septic conditions [5] and has appeared to be a good predictor of infection severity. Furthermore, the finding that PCT is released into the circulation within 3 h after endotoxin injection, plateaus at 6 h, and remains elevated for 24 h, makes PCT a promising new agent for early and sensitive identification of severe infection both in adults and children with promising results [6]. The results of recent studies suggested the usefulness of PCT for early diagnosis of early-onset [7-19] and late-onset neonatal sepsis [19-23]. This objective of this prospective multicenter study was to assess the diagnostic usefulness of PCT as a single marker of neonatal sepsis of nosocomial origin. Methods Since 1995 the neonatal services of 28 acute-care teaching hospitals distributed across 10 Autonomous Communities in Spain ("Grupo de Hospitales Castrillo") have been involved in an ongoing prospective surveillance project to assess the incidence and characteristics of nosocomial infections in the neonatal period [2,24]. The neonatal services of 13 hospitals participated in the present study. Between January 2000 and January 2001 all consecutive neonates aged between 4 and 28 days of life with clinical suspicion of sepsis of nosocomial origin were prospectively included in the study if blood samples were available for timed PCT measurement according to three postnatal periods: at the time of appearance of the first clinical manifestations and within 12–24 h and 36–48 h after the onset of symptoms, and complete neonatal and outcome data were collected to classify the infants into two distinct populations: confirmed sepsis and not confirmed sepsis. The study was approved by the Ethics Committees of the participating hospitals and the parents gave their informed consent. Sepsis of nosocomial origin was suspected in the presence of at least three clinical signs and one risk factor for the nosocomial origin of the infectious process (as shown in Table 1), and laboratory signs consistent with infection (abnormal hematologic values and/or C-reactive protein > 1.2 mg/dL). Diagnostic of confirmed nosocomial sepsis was established when blood culture was positive. If the pathogens isolated in blood culture were traditional pathogens of vertical transmission (Streptococcus agalactiae, Escherichia coli) and there was a positive maternal vaginal culture with the same pathogen, the episode was considered of vertical transmission and was excluded from the study. Two successive positive blood cultures from peripheral percutaneous specimens were required for the diagnosis of coagulase-negative staphylococci (CoNS) infection [2]. Only one peripheral blood culture was considered valid when both bottles of the blood culture set grew isolates of CoNS organisms with identical antibiotic susceptibility patterns. In patients with a central line, criteria for CoNS infection included recovery of the same CoNS with identical antibiotic susceptibility patterns from the peripheral blood culture and culture of the catheter tip using the semi-quantitative method of Maki et al. [25]. Because quantitative cultures without removal of the catheter was not a routine microbiologic method in all participating hospitals, the increased bacterial colony counts from blood drawn through the suspected infected device comPage 2 of 7 (page number not for citation purposes) BMC Pediatrics 2006, 6:16 http://www.biomedcentral.com/1471-2431/6/16 pared to blood from a peripheral venipuncture was not required for diagnosis. For all blood cultures it was recommended to submit an amount of ≥ 1 mL to the laboratory.