The Pediatric Infectious Disease Journal •  Volume 33, Number 10, October 2014 www.pidj.com | 1027 ORIGINAL STUDIES Background: Bloodstream infections (BSI) remain a leading cause of mor- bidity and mortality among infants admitted to neonatal intensive care units (NICUs). At the time of evaluation for suspected BSI, presenting signs may be nonspecific. We sought to determine the clinical signs and risk factors associated with laboratory-confirmed BSI among infants evaluated for late- onset sepsis in a tertiary NICU. Methods: This prospective cohort study included infants >3 days of age admitted to a level 4 NICU from July 2006 to October 2009 for whom a blood culture was drawn for suspected sepsis. Clinicians documented pre- senting signs at the time of culture. Laboratory-confirmed BSI was defined as per the National Healthcare Safety Network. Multivariate analyses were performed using a logistic regression random effects model. Results: Six-hundred and eighty eligible episodes of suspected BSI were recorded in 409 infants. Enteral contrast within the preceding 48 hours was the most significant risk factor for laboratory-confirmed BSI [Odds Ratio: 9.58 (95% confidence interval: 2.03–45.19)] followed by presence of a cen- tral venous catheter. Apnea and hypotension were the most strongly associ- ated presenting signs. Conclusion: Among infants evaluated in a tertiary NICU, recent exposure to enteral contrast was associated with increased odds of developing BSI. Apnea and hypotension were the most strongly associated clinical signs of infection. Key Words: apnea, bloodstream infection, central venous line, enteral con- trast, hypotension, NICU (Pediatr Infect Dis J 2014;33:1027–1032) D espite increased prevention efforts, bloodstream infections (BSI) remain a leading cause of morbidity and mortality among infants admitted to neonatal intensive care units (NICUs). 1–3 For this vulnerable population, BSI not only increase health care costs and the duration of hospitalization, but also increase the like- lihood of growth impairment, long-term neurodevelopmental dis- ability and death. 4–6 At the time of evaluation for suspected BSI, presenting signs such as apnea, temperature instability, hypogly- cemia and feeding intolerance may be nonspecific. Factors such as gestational age, presence of a central line and history of surgery may increase an infant’s likelihood for infection. Although previ- ous studies have addressed risk factors for BSI among inborn pre- mature infants, infants in tertiary NICUs, who are hospitalized for long periods and commonly require surgical or subspecialty care, have not been studied extensively. 7,8 Identifying factors associated with BSI in this unique population may reveal risk factors that could help guide appropriate treatment. The incidence of BSI in NICUs varies greatly depending on the gestational age, severity of illness, composition of inborn/ outborn infants and definition of BSI. 9–12 We sought to determine which presenting signs and risk factors are associated with labora- tory-confirmed BSI among infants in a tertiary NICU. PATIENTS AND METHODS Enrollment This prospective cohort study included infants >3 days of age hospitalized at the Children’s Hospital of Philadelphia New- born/Infant Intensive Care Unit, a tertiary level IV NICU, between July 2006 and October 2009 for whom a blood culture was drawn for suspected BSI. 13 Target enrollment for this study was 400 sub- jects based on estimates of symptom prevalence among the infected and noninfected infants. The committee for the protection of human subjects approved this study with a waiver of informed consent. At the time of culture, the collecting clinician documented on a checklist the presence or absence of clinical signs: hypother- mia (temperature <36.5°C), fever (temperature >38.0°C), hypogly- cemia (blood glucose <50 mg/dL), hyperglycemia (blood glucose >140 mg/dL) and unexplained metabolic acidosis (the presence of an acidosis unrelated to a known cause, including hypercarbia, hypotension or hypoxemia). Additional signs-lethargy, apnea, brad- ycardia, tachycardia, hypotension, poor perfusion, increased residu- als, abdominal distension, tachypnea or increased work of breath- ing, increased ventilatory requirements and suspicious blood count or C-reactive protein-were also documented at the time of blood culture collection and were based on deviation from normative val- ues for a given patient as per assessment by the collecting clinician. Demographic characteristics, presence of central lines or arterial catheters, preceding exposures (eg, enteral contrast study or line removal within 48 hours before BSI evaluation, histamine-2 recep- tor antagonist (H2-blocker) or proton-pump inhibitor use within the preceding 14 days) and interventions (eg, history of prior surgeries) were abstracted from the medical record by the study investigators. Feeding was defined as having received greater than trophic enteral feeds (ie, >10 mL/kg/d) at the time of the episode. Soft tissue infec- tions included cellulitis and skin abscesses. History of gastrointes- tinal pathology included the prior diagnosis of any of the following conditions: omphalocele, gastroschisis, necrotizing enterocolitis, spontaneous intestinal perforation, intestinal atresia, intestinal mal- rotation, Hirschprung’s disease or other bowel obstruction. Copyright © 2014 by Lippincott Williams & Wilkins ISSN: 0891-3668/14/3310-1027 DOI: 10.1097/INF.0000000000000386 Risk Factors Associated With Laboratory-conirmed  Bloodstream Infections in a Tertiary Neonatal Intensive  Care Unit Michael A. Padula, MD, MBI,*†‡ Maya L. Dewan, MD, MPH,† Samir S. Shah, MD, MSCE,§ Amy M. Padula, PhD, MSc,¶ Lakshmi Srinivasan, MB BS,*†‡ Karin L. McGowan, PhD, MS,‖ Kaitilin R. Mahoney, BA,* and Mary C. Harris, MD*†‡ Accepted for publication April 16, 2014. From the *Division of Neonatology; †Department of Pediatrics, The Children’s Hospital of Philadelphia; ‡Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA; §Division of Infectious Diseases, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH; ¶Division of Neonatol- ogy, Department of Pediatrics, Stanford University School of Medicine, Palo Alto, CA; and ‖Department of Pathology and Laboratory Medicine, Univer- sity of Pennsylvania, Philadelphia, PA. The authors have no funding or conflicts of interest disclose. Address for correspondence: Michael A. Padula, MD, MBI, The Children’s Hos- pital of Philadelphia, Division of Neonatology, 2NW49 Main, 34th St. & Civic Center Blvd., Philadelphia, PA 19104. E-mail: padula@email.chop.edu.