Early-onset sepsis (EOS) remains a serious and often fatal illness among infants born preterm, particularly among newborn infants of the lowest gestational age. Currently, most preterm infants with very low birth weight are treated empirically with antibiotics for risk of EOS, often for prolonged periods, in the absence of a culture-confirmed infection. Retrospective studies have revealed that antibiotic exposures after birth are associated with multiple subsequent poor outcomes among preterm infants, making the risk/benefit balance of these antibiotic treatments uncertain. Gestational age is the strongest single predictor of EOS, and the majority of preterm births occur in the setting of other factors associated with risk of EOS, making it difficult to apply risk stratification strategies to preterm infants. Laboratory tests alone have a poor predictive value in preterm EOS. Delivery characteristics of extremely preterm infants present an opportunity to identify those with a lower risk of EOS and may inform decisions to initiate or extend antibiotic therapies. Our purpose for this clinical report is to provide a summary of the current epidemiology of preterm neonatal sepsis and provide guidance for the development of evidence-based approaches to sepsis risk assessment among preterm newborn infants.
Antibiotics are administered shortly after birth to nearly all preterm infants with very low birth weight (VLBW) (birth weight <1500 g) because of the risk of early-onset sepsis (EOS).1,–4 Physicians are often reluctant to discontinue antibiotics once initiated for many reasons, including the relatively high risk of EOS among preterm infants and the relatively high rate of mortality attributable to infection. Particularly among infants with VLBW, neonatal clinicians must determine which infants are most likely to have EOS when nearly all have some degree of respiratory or systemic instability. Poor predictive performance of common laboratory tests and concerns regarding the unreliability of blood cultures add to the difficulty in discriminating at-risk infants. Because gestational age is the strongest predictor of EOS and approximately two-thirds of preterm births are associated with preterm labor, premature rupture of membranes (PROM), or clinical chorioamnionitis,5 risk stratification strategies cannot be applied to preterm newborn infants in the same manner as for term neonates.
Pathogenesis and Current Epidemiology of Preterm Neonatal EOS
Preterm EOS is defined as a blood or cerebrospinal fluid (CSF) culture obtained within 72 hours after birth that is growing a pathogenic bacterial species. This microbiological definition stands in contrast to the functional definitions of sepsis used in pediatric and adult patients, for whom the definition is used to specify a series of time-sensitive interventions. The current overall incidence of EOS in the United States is approximately 0.8 cases per 1000 live births.6 A disproportionate number of cases occur among infants born preterm in a manner that is inversely proportional to gestational age at birth. The incidence of EOS is approximately 0.5 cases per 1000 infants born at ≥37 weeks’ gestation, compared with approximately 1 case per 1000 infants born at 34 to 36 weeks’ gestation, 6 cases per 1000 infants born at <34 weeks’ gestation, 20 cases per 1000 infants born at <29 weeks’ gestation, and 32 cases per 1000 infants born at 22 to 24 weeks’ gestation.6,–10 The incidence of EOS has declined among term infants over the past 25 years, a change attributed to the implementation of evidence-based intrapartum antimicrobial therapy. The impact of such therapies on preterm infants is less clear. Authors of the most recent studies report an EOS incidence among infants with VLBW ranging from 9 to 11 cases per 1000 infants with VLBW, whereas studies from the early 1990s revealed rates of 19 to 32 per 1000 infants.10,11 Improvements among VLBW incidence may be limited to those born at older gestational ages. No significant change over time was observed in a study of 34 636 infants born from 1993 to 2012 at 22 to 28 weeks’ gestation, with the reported incidence ranging from 20.5 to 24.4 per 1000 infants.8 Morbidity and mortality from EOS remain substantial: 95% of preterm infants with EOS require neonatal intensive care for respiratory distress and/or blood pressure support, and 75% of deaths from EOS occur among infants with VLBW.6,10 The mortality rate among those with EOS is an order of magnitude higher among preterm compared with term infants, whether measured by gestational age (1.6% at ≥37 weeks, 30% at 25–28 weeks, and approximately 50% at 22–24 weeks)7,8,10 or birth weight (3.5% among those born at ≥1500 g vs 35% for those born at <1500 g).6
The pathogenesis of preterm EOS is complex. EOS primarily begins in utero and was originally described as the “amniotic infection syndrome.”12,13 Among term infants, EOS pathogenesis most commonly develops during labor and involves ascending colonization and infection of the uterine compartment with maternal gastrointestinal and genitourinary flora, with subsequent colonization and invasive infection of the fetus and/or fetal aspiration of infected amniotic fluid. This intrapartum sequence may be responsible for EOS that develops after PROM or during preterm labor that is induced for maternal indications. However, the pathogenesis of preterm EOS likely begins before the onset of labor in many cases of preterm labor and/or PROM. Intraamniotic infection (IAI) may cause stillbirth in the second and third trimesters.14 In approximately 25% of cases, IAI is the cause of preterm labor and PROM, particularly when these occur at the lowest gestational ages; evidence suggests that microbial-induced maternal inflammation can initiate parturition and elicit fetal inflammatory responses.5,15,–18 Organisms isolated from the intrauterine compartment of women with preterm labor, PROM, or both are primarily vaginal in origin and include low-virulence species, such as Ureaplasma, as well as anaerobic species and well-recognized neonatal pathogens, such as Escherichia coli and group B Streptococcus (GBS).16,–18 The isolation of maternal oral flora and, more rarely, Listeria monocytogenes, suggests a transplacental pathway for some IAIs.16,18,–20 Inflammation inciting parturition may not, however, always be attributable to IAI. Inflammation resulting from immune-mediated rejection of the fetal or placental compartment (from maternal extrauterine infection), as well as that incited by reproductive or nonreproductive microbiota, may all contribute to the pathogenesis of preterm labor and PROM, complicating the interpretation of placental pathology.15,20
Risk Factors for Preterm EOS
Multiple clinical risk factors have been used to assess the risk of EOS among infants born at ≤34 6/7 weeks’ gestation. Univariate analyses of risk factors for EOS among preterm infants have been used to identify gestational age, birth weight, PROM and prolonged rupture of membranes (ROM), preterm onset of labor, maternal age and race, maternal intrapartum fever, mode of delivery, and administration of intrapartum antibiotics to be associated with risk of EOS; however, the independent contribution of any specific factor other than gestational age has been difficult to quantify. For example, among term infants, there is a linear relationship between the duration of ROM and the risk of EOS.9 In contrast, the relationship between PROM and the risk of EOS is not simply described by its occurrence or duration but modified by gestational age as well as by the additional presence of clinical chorioamnionitis and the administration of latency and intrapartum antibiotics.17,21,–24 These observations are likely related to uncertainty regarding the role of intrauterine infection and cervical structural defects in the pathogenesis of spontaneous PROM.24,25
The clinical diagnosis of chorioamnionitis has been used as a primary risk factor for identifying infants at risk for EOS. Most preterm infants with EOS are born to women with this clinical diagnosis.4,26,–29 The American College of Obstetricians and Gynecologists (ACOG) recently advocated for using the term “intraamniotic infection” rather than chorioamnionitis (which is primarily a histologic diagnosis) and published guidance for its diagnosis and management.30 A confirmed diagnosis of IAI is made by a positive result on an amniotic fluid Gram-stain, culture, or placental histopathology. Suspected IAI is diagnosed by maternal intrapartum fever (either a single documented maternal intrapartum temperature of ≥39.0°C or a temperature of 38.0–38.9°C that persists for >30 minutes) and 1 or more of the following: (1) maternal leukocytosis, (2) purulent cervical drainage, and (3) fetal tachycardia. The ACOG recommends that intrapartum antibiotics be administered whenever IAI is diagnosed or suspected and when otherwise unexplained maternal fever occurs during labor. Chorioamnionitis or IAI is strongly associated with EOS in preterm infants, with a number needed to treat of only 6 to 40 infants per case of confirmed EOS.4,26,–29 Conversely, the absence of clinical and histologic chorioamnionitis may be used to identify a group of preterm infants who are at a lower risk for EOS. In a study of 15 318 infants born at 22 to 28 weeks’ gestation, those born by cesarean delivery with membrane rupture at delivery and without clinical chorioamnionitis were significantly less likely to have EOS or die before 12 hours of age.4 The number needed to treat for infants born in these circumstances was approximately 200; with the additional absence of histologic chorioamnionitis, the number needed to treat is approximately 380.4 Another study of 109 cases of EOS occurring among 5313 infants with VLBW over a 25-year period revealed that 97% of cases occurred in infants born with some combination of PROM, preterm labor, or concern for IAI.29 In that report, 2 cases of listeriosis occurred in the context of unexplained fetal distress in otherwise uncomplicated pregnancies.
Antibiotic Stewardship in Preterm EOS Management
Currently, most premature infants with VLBW are treated empirically with antibiotics for risk of EOS, often for prolonged periods, even in the absence of a culture-confirmed infection. Prolonged empirical antibiotics are administered to approximately 35% to 50% of infants with a low gestational age, with significant center-specific variation.1,–4 Antibiotic drugs are administered for many reasons, including the relatively high incidence of EOS among preterm infants, the relatively high rate of mortality attributable to infection, and the frequency of clinical instability after birth. Empirical antibiotics administered to very preterm infants in the first days after birth have been associated with an increased risk of subsequent poor outcomes.1,4,31,–33 One multicenter study of 4039 infants born from 1998 to 2001 with a birth weight of <1000 g revealed that those infants who died or had a diagnosis of necrotizing enterocolitis (NEC) before hospital discharge were significantly more likely to have received prolonged empirical antibiotic therapy in the first week after birth.1 The authors of the study estimated that the risk of NEC increased by 7% for each additional day of antibiotics administered in the absence of culture-confirmed EOS. Authors of a single-center study of infants with VLBW estimated that the risk of NEC increased by 20% for each additional day of antibiotics administered in the absence of a culture-confirmed infection.31 Authors of another study of 11 669 infants with VLBW assessed the overall rate of antibiotic use and found that higher rates during the first week after birth or during the entire hospitalization were both associated with increased mortality, even when adjusted for multiple predictors of neonatal morbidity and mortality.33 One concern in each of these studies is that some infants categorized as uninfected may in fact have suffered from EOS. Yet, even among 5640 infants born at 22 to 28 weeks’ gestation at a lower risk for EOS, those who received prolonged empirical antibiotic therapy during the first week after birth had higher rates of death and bronchopulmonary dysplasia.4 Several explanations are possible for all of these findings, including simply that physicians administer the most antibiotics to the sickest infants. Other potential mechanisms include the role of antibiotics in promoting dysbiosis of the gut, skin, and respiratory tract, affecting the interactions between colonizing flora in maintaining health and promoting immunity; it is also possible that antibiotics and dysbiosis function as modulators of vascular development.34,35 Although the full relationship between early neonatal antibiotic exposures and subsequent childhood health remains to be defined, current evidence suggests that such exposures do affect preterm infants. Physicians should consider the risk/benefit balance of initiating antibiotic therapy for risk of EOS as well as for continuing empirical antibiotic therapy in the absence of a culture-confirmed infection.
Risk Categorization for Preterm Infants
Perhaps the greatest contributor to the nearly universal practice of empirical antibiotic administration to preterm infants is the uncertainty in EOS risk assessment. Because gestational age is the strongest predictor of EOS, and two-thirds of preterm births are associated with preterm labor, PROM, or clinical concern for intrauterine infection,5 risk stratification strategies cannot be applied to preterm infants in the same manner as for term neonates. In particular, the Neonatal Early-Onset Sepsis Risk Calculator does not apply to infants born before 34 0/7 weeks’ gestation.36 The objective of EOS risk assessment among preterm infants is, therefore, to determine which infants are at the lowest risk for infection and who, despite clinical instability, may be spared administration of empirical antibiotics. The circumstances of preterm birth may provide the best current approach to EOS management for preterm infants.
Preterm Infants at Lower Risk for EOS
Criteria for preterm infants to be considered at a lower risk for EOS include the following: (1) obstetric indications for preterm birth (such as maternal preeclampsia or other noninfectious medical illness or placental insufficiency), (2) birth by cesarean delivery, and (3) absence of labor, attempts to induce labor, or any ROM before delivery. Acceptable initial approaches to these infants might include (1) no laboratory evaluation and no empirical antibiotic therapy, or (2) a blood culture and clinical monitoring. For infants who do not improve after initial stabilization and/or those who have severe systemic instability, the administration of empirical antibiotics may be reasonable but is not mandatory.
Infants in this category who are born by vaginal or cesarean delivery after efforts to induce labor and/or ROM before delivery are subject to factors associated with the pathogenesis of EOS during delivery. If any concern for infection arises during the process of delivery, the infant should be managed as recommended below for preterm infants at a higher risk for EOS. Otherwise, an acceptable approach to these infants is to obtain a blood culture and to initiate antibiotic therapy for infants with respiratory and/or cardiovascular instability after birth.
Preterm Infants at Higher Risk for EOS
Infants born preterm because of cervical incompetence, preterm labor, PROM, chorioamnionitis or IAI, and/or acute and otherwise unexplained onset of nonreassuring fetal status are at the highest risk for EOS. In these cases, IAI may be the cause of preterm birth or a secondary complication of PROM and cervical dilatation. IAI may also be the cause of unexplained fetal distress. The most reasonable approach to these infants is to perform a blood culture and start empirical antibiotic treatment. Obtaining CSF for culture before the administration of antibiotics should be considered if the infant will tolerate the procedure and if it will not delay the initiation of antibiotic therapy.
Laboratory Testing
Blood Culture
In the absence of validated, clinically available molecular diagnostic tests, a blood culture remains the diagnostic standard for EOS. Newborn surface cultures and gastric aspirate analysis cannot be used to diagnose EOS, and a urine culture is not indicated in sepsis evaluations performed at <72 hours of age. Modern blood culture systems use optimized enriched culture media with antimicrobial neutralization properties, continuous-read detection systems, and specialized pediatric culture bottles. Although concerns have been raised regarding incomplete detection of low-level bacteremia and the effects of intrapartum antibiotic administration,27,37 these systems reliably detect bacteremia at a level of 1 to 10 colony-forming units if a minimum of 1 mL of blood is inoculated; authors of several studies report no effect of intrapartum antibiotics on time to positivity.38,–42 Culture media containing antimicrobial neutralization elements efficiently neutralize β-lactam antibiotics and gentamicin.39 A median blood culture time to positivity <24 hours is reported among VLBW infants when using contemporary blood culture techniques.29,43,–46 Pediatric blood culture bottles generally require a minimum of 1 mL of blood for optimal recovery of organisms.47,48 The use of 2 separate bottles may provide the opportunity to determine if commensal species are true infections by comparing growth in the two.49,50 Use of 1 aerobic and 1 anaerobic culture bottle may optimize organism recovery. Most neonatal pathogens, including GBS, E coli, coagulase-negative Staphylococcus, and Staphylococcus aureus, will grow in anaerobic conditions. One study revealed that with routine use of both pediatric aerobic and adult anaerobic blood cultures, strict anaerobic species (primarily Bacteroides fragilis) were isolated in 16% of EOS cases in preterm infants with VLBW.29 An anaerobic blood culture is routinely performed among adult patients at risk for infection and can be used for neonatal blood cultures. Individual centers may benefit from collaborative discussion with the laboratory where cultures are processed to optimize local processes.
CSF Culture
The incidence of meningitis is higher among preterm infants (approximately 0.7 cases per 1000 live births at 22–28 weeks’ gestation)4 compared with the incidence in the overall birth population (approximately 0.02–0.04 cases per 1000 live births).6,10 In the study of differential EOS risk among very preterm infants, meningitis did not occur at all among lower-risk preterm infants.4 The true incidence of meningitis among preterm infants may be underestimated because of the common practice of performing a lumbar puncture after the initiation of empirical antibiotic therapy. Although most preterm infants with culture-confirmed early-onset meningitis grow the same organism from blood cultures, the concordance is not 100%, and CSF cell count parameters may not always identify meningitis.51 If a CSF culture has not been obtained before the initiation of empirical antibiotics, physicians should balance the physiologic stability of the infant, the risk of EOS, and the potential harms associated with prolonged antibiotic therapy when making the decision to perform a lumbar puncture in preterm infants who are critically ill.
White Blood Cell Count
The white blood cell (WBC) count, differential (immature-to-total neutrophil ratio), and absolute neutrophil count are commonly used to assess risk of EOS. Multiple clinical factors can affect the WBC count and differential, including gestational age at birth, sex, and mode of delivery.52,–55 Fetal bone marrow depression attributable to maternal preeclampsia or placental insufficiency, as well as prolonged exposure to inflammatory signals (such as PROM), frequently result in abnormal values in the absence of infection. Most studies in which the performance characteristics of the complete blood cell (CBC) count in predicting infection is addressed have been focused on term infants. In 1 large multicenter study, the authors assessed the relationship between the WBC count and culture-confirmed EOS and analyzed data separately for infants born at <34 weeks’ gestation.56 They found that all components of the CBC count lacked sensitivity for predicting EOS. The highest likelihood ratios (LRs) for EOS were associated with extreme values. A positive LR of >3 (ie, a likelihood of infection at least 3 times higher than the entire group of infants born at <34 weeks’ gestation) was associated with a WBC count of <1000 cells per μL, an absolute neutrophil count of <1000, and an immature-to-total neutrophil ratio of >0.25. A total WBC count of >50 000 cells per μL (LR, 2.3) and a platelet count of <50 000 (LR, 2.2) had a modest relationship to EOS.
Other Inflammatory Markers
Other markers of inflammation, including C-reactive protein (CRP), procalcitonin, interleukins (soluble interleukin 2 receptor, interleukin 6, and interleukin 8), tumor necrosis factor α, and CD64 are addressed in multiple studies.57,–60 Both CRP and procalcitonin concentrations increase in newborn infants in response to a variety of inflammatory stimuli, including infection, asphyxia, and pneumothorax. Procalcitonin concentrations also increase naturally over the first 24 to 36 hours after birth.60 Single values of CRP or procalcitonin obtained after birth to assess the risk of EOS are neither sufficiently sensitive nor specific to guide EOS care decisions. Consistently normal values of CRP and procalcitonin over the first 48 hours of age are associated with the absence of EOS, but serial abnormal values alone should not be used to extend antibiotic therapy in the absence of a culture-confirmed infection.
Treatment of Preterm EOS
The microbiology of EOS in the United States is largely unchanged over the past 10 years. Authors of national surveillance studies continue to identify E coli as the most common bacteria isolated in EOS cases that occur among preterm infants, whether defined by a gestational age of <34 weeks or by a birth weight of <1500 g. Overall, E coli is isolated in approximately 50%, and GBS is isolated in approximately 20% of all EOS cases occurring among infants born at <34 weeks’ gestation.6 Fungal organisms are isolated in <1% of cases. Approximately 10% of cases are caused by other Gram-positive organisms (predominantly viridans group streptococci and enterococci), and approximately 20% of cases are caused by other Gram-negative organisms. S aureus (approximately 1%–2%) and L monocytogenes (approximately 1%) are uncommon causes of preterm EOS.4,6,11 If an anaerobic culture is routinely performed, strict anaerobic bacteria are isolated in up to 15% of EOS cases among preterm infants with VLBW, with B fragilis being the predominant anaerobic species isolated.29
Ampicillin and gentamicin are the first choice for empirical therapy for EOS. This combination will be effective against GBS, most other streptococcal and enterococcal species, and L monocytogenes. Although two-thirds of E coli EOS isolates and most other Gram-negative EOS isolates are resistant to ampicillin, the majority remain sensitive to gentamicin.6 Extended-spectrum, β-lactamase-producing organisms are only rarely reported among EOS cases in the United States. Therefore, the routine empirical use of broader-spectrum antibiotics is not warranted and may be harmful.61 Nonetheless, 1% to 2% of E coli cases were resistant to both ampicillin and gentamicin in recent surveillance studies by the Centers for Disease Control and Prevention, and B fragilis is not uniformly sensitive to these medications.6,62 Therefore, among preterm infants who are severely ill and at the highest risk for Gram-negative EOS (such as infants with VLBW born after prolonged PROM and infants exposed to prolonged courses of antepartum antibiotic therapy63,–65), the empirical addition of broader-spectrum antibiotic therapy may be considered until culture results are available. The choice of additional therapy should be guided by local antibiotic resistance data.
When EOS is confirmed by a blood culture, a lumbar puncture should be performed if not previously done. Antibiotic therapy should use the narrowest spectrum of appropriate agents once antimicrobial sensitivities are known. The duration of therapy should be guided by expert references (eg, the American Academy of Pediatrics [AAP] Red Book: Report of the Committee on Infectious Diseases) and informed by the results of a CSF analysis and the achievement of sterile blood and CSF cultures. Consultation with infectious disease specialists should be considered for cases complicated by meningitis or other site-specific infections and for cases with complex antibiotic resistance patterns.
When initial blood culture results are negative, antibiotic therapy should be discontinued by 36 to 48 hours of incubation, unless there is evidence of site-specific infection. Persistent cardiorespiratory instability is common among infants with VLBW and is not alone an indication for prolonged empirical antibiotic administration. Continuing empirical antibiotic administration in response to laboratory test abnormalities alone is rarely justified, particularly among preterm infants born in the setting of maternal obstetric conditions known to affect fetal hematopoiesis.
Prevention Strategies
The only proven preventive strategy for EOS is the appropriate administration of maternal intrapartum antibiotic prophylaxis. The most current recommendations from national organizations, such as the AAP, ACOG, and Centers for Disease Control and Prevention, should be followed for the administration of GBS intrapartum prophylaxis as well as for the administration of intrapartum antibiotic therapy when there is suspected or confirmed IAI. Neonatal practices are focused on the identification and empirical antibiotic treatment of preterm neonates at risk for EOS; these practices cannot prevent EOS. The empirical administration of intramuscular penicillin to all newborn infants to prevent neonatal, GBS-specific EOS is not justified and is not endorsed by the AAP. Neither GBS intrapartum antibiotic prophylaxis nor any neonatal EOS practice will prevent late-onset GBS infection or any other form of late-onset bacterial infection. Preterm infants are particularly susceptible to late-onset GBS infection, with approximately 40% of late-onset GBS cases occurring among infants born at ≤34 6/7 weeks’ gestation.66,67
Summary Points
The epidemiology, microbiology, and pathogenesis of EOS differ substantially between term infants and preterm infants with VLBW.
Infants born at ≤34 6/7 weeks’ gestation can be categorized by level of risk for EOS by the circumstances of their preterm birth.
∘ Infants born preterm by cesarean delivery because of maternal noninfectious illness or placental insufficiency in the absence of labor, attempts to induce labor, or ROM before delivery are at a relatively low risk for EOS. Depending on the clinical condition of the neonate, physicians should consider the risk/benefit balance of an EOS evaluation and empirical antibiotic therapy.
∘ Infants born preterm because of maternal cervical incompetence, preterm labor, PROM, clinical concern for IAI, or acute onset of unexplained nonreassuring fetal status are at the highest risk for EOS. Such neonates should undergo EOS evaluation with a blood culture and empirical antibiotic treatment.
∘ Obstetric and neonatal care providers should communicate and document the circumstances of preterm birth to facilitate EOS risk assessment among preterm infants.
Clinical centers should consider the development of locally appropriate written guidelines for preterm EOS risk assessment and clinical management. After guidelines are implemented, ongoing surveillance, designed to identify low-frequency adverse events and affirm efficacy, is recommended.
The diagnosis of EOS is made by a blood or CSF culture. EOS cannot be diagnosed by laboratory tests alone, such as CBC count or CRP levels.
The combination of ampicillin and gentamicin is the most appropriate empirical antibiotic regimen for infants at risk for EOS. Empirical administration of additional broad-spectrum antibiotics may be indicated in preterm infants who are severely ill and at a high risk for EOS, particularly after prolonged antepartum maternal antibiotic treatment.
When blood cultures are sterile, antibiotic therapy should be discontinued by 36 to 48 hours of incubation, unless there is clear evidence of site-specific infection. Persistent cardiorespiratory instability is common among preterm infants with VLBW and is not alone an indication for prolonged empirical antibiotic administration. Laboratory test abnormalities alone rarely justify prolonged empirical antibiotic administration, particularly among preterm infants at a lower risk for EOS.
- AAP
American Academy of Pediatrics
- ACOG
American College of Obstetricians and Gynecologists
- CBC
complete blood cell
- CRP
C-reactive protein
- CSF
cerebrospinal fluid
- EOS
early-onset sepsis
- GBS
group B Streptococcus
- IAI
intraamniotic infection
- LR
likelihood ratio
- NEC
necrotizing enterocolitis
- PROM
premature rupture of membranes
- ROM
rupture of membranes
- VLBW
very low birth weight
- WBC
white blood cell
This document is copyrighted and is property of the American Academy of Pediatrics and its Board of Directors. All authors have filed conflict of interest statements with the American Academy of Pediatrics. Any conflicts have been resolved through a process approved by the Board of Directors. The American Academy of Pediatrics has neither solicited nor accepted any commercial involvement in the development of the content of this publication.
Clinical reports from the American Academy of Pediatrics benefit from expertise and resources of liaisons and internal (AAP) and external reviewers. However, clinical reports from the American Academy of Pediatrics may not reflect the views of the liaisons or the organizations or government agencies that they represent.
The guidance in this report does not indicate an exclusive course of treatment or serve as a standard of medical care. Variations, taking into account individual circumstances, may be appropriate.
All clinical reports from the American Academy of Pediatrics automatically expire 5 years after publication unless reaffirmed, revised, or retired at or before that time.
FUNDING: No external funding.
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2018-2894.
References
Lead Authors
Karen M. Puopolo, MD, PhD, FAAP
William E. Benitz, MD, FAAP
Theoklis E. Zaoutis, MD, MSCE, FAAP
Committee on Fetus and Newborn, 2017–2018
James Cummings, MD, Chairperson
Sandra Juul, MD
Ivan Hand, MD
Eric Eichenwald, MD
Brenda Poindexter, MD
Dan L. Stewart, MD
Susan W. Aucott, MD
Karen M. Puopolo, MD, PhD, FAAP
Jay P. Goldsmith, MD
Kristi Watterberg, MD, Immediate Past Chairperson
Liaisons
Kasper S. Wang, MD – American Academy of Pediatrics Section on Surgery
Thierry Lacaze, MD – Canadian Paediatric Society
Joseph Wax, MD – American College of Obstetricians and Gynecologists
Tonse N.K. Raju, MD, DCH – National Institutes of Health
Wanda Barfield, MD, MPH, CAPT USPHS – Centers for Disease Control and Prevention
Erin Keels, MS, APRN, NNP-BC – National Association of Neonatal Nurses
Staff
Jim Couto, MA
Committee on Infectious Diseases, 2017–2018
Carrie L. Byington, MD, FAAP, Chairperson
Yvonne A. Maldonado, MD, FAAP, Vice Chairperson
Ritu Banerjee, MD, PhD, FAAP
Elizabeth D. Barnett, MD, FAAP
James D. Campbell, MD, MS, FAAP
Jeffrey S. Gerber, MD, PhD, FAAP
Ruth Lynfield, MD, FAAP
Flor M. Munoz, MD, MSc, FAAP
Dawn Nolt, MD, MPH, FAAP
Ann-Christine Nyquist, MD, MSPH, FAAP
Sean T. O’Leary, MD, MPH, FAAP
Mobeen H. Rathore, MD, FAAP
Mark H. Sawyer, MD, FAAP
William J. Steinbach, MD, FAAP
Tina Q. Tan, MD, FAAP
Theoklis E. Zaoutis, MD, MSCE, FAAP
ex Officio
David W. Kimberlin, MD, FAAP – Red Book Editor
Michael T. Brady, MD, FAAP – Red Book Associate Editor
Mary Anne Jackson, MD, FAAP – Red Book Associate Editor
Sarah S. Long, MD, FAAP – Red Book Associate Editor
Henry H. Bernstein, DO, MHCM, FAAP – Red Book Online Associate Editor
H. Cody Meissner, MD, FAAP – Visual Red Book Associate Editor
Liaisons
Amanda C. Cohn, MD, FAAP – Centers for Disease Control and Prevention
Jamie Deseda-Tous, MD – Sociedad Latinoamericana de Infectología Pediátrica
Karen M. Farizo, MD – US Food and Drug Administration
Marc Fischer, MD, FAAP – Centers for Disease Control and Prevention
Natasha Halasa, MD, MPH, FAAP – Pediatric Infectious Diseases Society
Nicole Le Saux, MD – Canadian Pediatric Society
Scot Moore, MD, FAAP – Committee on Practice Ambulatory Medicine
Angela K. Shen, ScD, MPH – National Vaccine Program Office
Neil S. Silverman, MD – American College of Obstetricians and Gynecologists
James J. Stevermer, MD, MSPH, FAAFP – American Academy of Family Physicians
Jeffrey R. Starke, MD, FAAP – American Thoracic Society
Kay M. Tomashek, MD, MPH, DTM – National Institutes of Health
Staff
Jennifer M. Frantz, MPH
Competing Interests
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
Comments
RE: RE: Management of Neonates Born at ≤34 6/7 Weeks’ Gestation With Suspected or Proven Early-Onset Bacterial Sepsis
We appreciate the thoughtful commentary by Dr. Burchfield. His concerns are commonly encountered by neonatal clinicians and were considered carefully in the preparation of the revised Clinical Reports.
His first concern addresses the decision to continue or stop antibiotics among infants receiving empiric treatment when blood cultures are sterile after maternal intrapartum antibiotic treatment. Culture media with antibiotic neutralization material are widely used in clinical microbiology laboratories to mitigate the impact of circulating antibiotics. If appropriate blood volumes and cultures methods are used, a sterile blood culture provides strong evidence that the infant is not bacteremic. Dr. Burchfield also raises concern that an infant who is not bacteremic still may have a focal bacterial infection that will recrudesce if antibiotics are stopped prematurely. This theoretical concern cannot be definitively resolved without a controlled trial of antibiotic administration in the absence of bacteremia. Nonetheless, available observational data (cited in the Clinical Reports) suggest that extended antibiotic administration in the absence of a positive blood culture does not result in better outcomes among preterm infants, whether measured by the incidence of death or significant morbidities.
We do not agree that the data of Barton, et al. support his concern. In that autopsy case series of 111 infants with birth weight 300-1000 grams who died in 1990-1993, the authors identified infection as the primary cause of death if white blood cell infiltrates were observed histologically in autopsied organs. “Congenital infection” was diagnosed in 30 cases, among which only 11 infants had blood cultures obtained before death. Pathogenic bacteria were isolated from 9/11 (no data is provided for 2 cultures), demonstrating that neonatal blood cultures confirmed the authors’ histologic diagnosis in at least 82% of the cases.
Finally, Dr. Burchfield suggests that clinical decision-making will be enhanced by obtaining serial CRP values from infants born in the setting of intrapartum antibiotic administration. While it is true that consistently low CRP levels over the first 48 hours imply an ~6-fold reduction in the probability of culture-proven sepsis (1), the positive predictive value of CRP levels up to 6 mg/dL was <5% in the cited study and even levels >6 mg/dL had poor positive predictive values for culture-proven sepsis. Nonetheless, the 3-fold increase in the posterior probability of infection implied by an elevated CRP level may convince some clinicians to extend antibiotic treatments in specific culture-negative cases. The Reports state, “Consistently normal values of CRP and procalcitonin over the first 48 hours of age are associated with the absence of EOS, but serial abnormal values alone should not be used to extend antibiotic therapy in the absence of culture-confirmed infection.” The key word is alone; context is important.
Dr. Burchfield’s commentary highlights the difficulty of uncertainty. The revised Clinical Reports discuss the relative merits of different approaches to risk assessment among term and preterm infants. In the end, no predictive model, clinical algorithm, or laboratory test result will perform perfectly, and we agree with Dr. Burchfield that individual cases will always require some degree of clinical judgement.
References
1. Benitz WE, Han MY, Madan A, Ramachandra P. Serial serum C-reactive protein levels in the diagnosis of neonatal infection. Pediatrics 1998;102:E41.
RE: Management of Neonates Born at ≤34 6/7 Weeks’ Gestation With Suspected or Proven Early-Onset Bacterial Sepsis
The Clinical Report is important and timely in attempting to limit antibiotic use to those patients most likely to benefit from them, but I am concerned that the implications of the statement may leave some patients undertreated.
I am most concerned with discontinuing antibiotics in high-risk, ill patients who may have received antibiotics trans-placentally and whose blood cultures show no growth. The authors attempt to justify this by showing time to positive blood cultures are no different in babies whose mothers did versus did not receive antibiotics. This negates the concept of bacterial killing time after exposure to antibiotics, which for Group B Streptococcus, is approximately 4-6 hours (1). I am concerned about possible partial treatment of true bacterial infections with early termination of antibiotics.
Neonatal clinicians frequently manage patients exposed to maternal antibiotics for hours, and this conceivably will inhibit bacterial growth. According to the report, with a positive blood culture, clinicians should use Red Book guidelines to determine length of therapy (for instance with Group B Streptococcus, 10-14 days depending on the CSF result). In circumstances where antibiotics MAY HAVE inhibited bacterial growth in the blood culture, surely stopping antibiotics at 36-48 hours is not advised. It would be as if a positive blood culture was followed after 24 hours of antibiotic therapy with a repeat culture that is negative, and antibiotics are stopped after a 2-3 day treatment. That negative culture should give the clinician confidence that they chose the correct antibiotic, but not that a sufficient antibiotic course has been delivered.
In addition, clinicians need to remember that blood is only one organ, and other organs may be infected without the blood being inoculated. In the autopsy review of neonatal deaths by Barton et al (2), infection was not diagnosed clinically 61% of the time with deaths in these cases attributed to RDS or immaturity, many of whom were not treated with antibiotics.
Finding no growth in the blood culture at 36-48 hours after multiple hours of in-utero antibiotic exposure may be reassuring when no other markers of infection are present, and this is where the use of inflammatory markers such as CRP can be helpful. Although the clinical report mentions other conditions that may be associated with elevations of the CRP, such as asphyxia and pneumothorax (not referenced), certainly an inflammatory response from infection remains the overwhelming etiology of an elevation. In the face of significant risk for neonatal infection, like PPROM and chorioamnionitis, in a mother who received antibiotics prior to delivery AND the baby has elevation of the non-specific inflammatory markers, it not only seems reasonable but also prudent to consider that the CRP or procalcitonin is elevated due to infection. In addition, as shown by Benitz (3), serial normal CRP values have a 99% negative predictive value of serious infection at 48 hrs and should give confidence to discontinue antibiotics. This approach should meet the goal of limiting antibiotics to those most likely infected while minimizing exposure to those not infected.
1. Swingle HM, Bucciarelli RL, Ayoub EM. Synergy between penicillins and low concentrations of gentamicin in the killing of group B streptococci. J Infect Dis. 1985:152(3):515-20. PMID: 3897398
2. Barton L, Hodgman JE, Pavlova Z. Causes of Death in the Extremely Low Birth Weight Infant. Pediatrics 1999 Feb;103(2):446-51. PMID:9925839
3. Benitz WE . Adjunct laboratory tests in the diagnosis of early-onset neonatal sepsis. Clin Perinatol. 2010;37(2):421–438.PMID:20569816