With improved obstetrical management and evidence-based use of intrapartum antimicrobial therapy, early-onset neonatal sepsis is becoming less frequent. However, early-onset sepsis remains one of the most common causes of neonatal morbidity and mortality in the preterm population. The identification of neonates at risk for early-onset sepsis is frequently based on a constellation of perinatal risk factors that are neither sensitive nor specific. Furthermore, diagnostic tests for neonatal sepsis have a poor positive predictive accuracy. As a result, clinicians often treat well-appearing infants for extended periods of time, even when bacterial cultures are negative. The optimal treatment of infants with suspected early-onset sepsis is broad-spectrum antimicrobial agents (ampicillin and an aminoglycoside). Once a pathogen is identified, antimicrobial therapy should be narrowed (unless synergism is needed). Recent data suggest an association between prolonged empirical treatment of preterm infants (≥5 days) with broad-spectrum antibiotics and higher risks of late onset sepsis, necrotizing enterocolitis, and mortality. To reduce these risks, antimicrobial therapy should be discontinued at 48 hours in clinical situations in which the probability of sepsis is low. The purpose of this clinical report is to provide a practical and, when possible, evidence-based approach to the management of infants with suspected or proven early-onset sepsis.
Introduction
“Suspected sepsis” is one of the most common diagnoses made in the NICU.1 However, the signs of sepsis are nonspecific, and inflammatory syndromes of noninfectious origin mimic those of neonatal sepsis. Most infants with suspected sepsis recover with supportive care (with or without initiation of antimicrobial therapy). The challenges for clinicians are threefold: (1) identifying neonates with a high likelihood of sepsis promptly and initiating antimicrobial therapy; (2) distinguishing “high-risk” healthy-appearing infants or infants with clinical signs who do not require treatment; and (3) discontinuing antimicrobial therapy once sepsis is deemed unlikely. The purpose of this clinical report is to provide a practical and, when possible, evidence-based approach to the diagnosis and management of early-onset sepsis, defined by the National Institute of Child Health and Human Development and Vermont Oxford Networks as sepsis with onset at ≤3 days of age.
Pathogenesis and Epidemiology of Early-Onset Sepsis
Before birth, the fetus optimally is maintained in a sterile environment. Organisms causing early-onset sepsis ascend from the birth canal either when the amniotic membranes rupture or leak before or during the course of labor, resulting in intra-amniotic infection.2 Commonly referred to as “chorioamnionitis,” intra-amniotic infection indicates infection of the amniotic fluid, membranes, placenta, and/or decidua.
Group B streptococci (GBS) can also enter the amniotic fluid through occult tears. Chorioamnionitis is a major risk factor for neonatal sepsis. Sepsis can begin in utero when the fetus inhales or swallows infected amniotic fluid. The neonate can also develop sepsis in the hours or days after birth when colonized skin or mucosal surfaces are compromised. The essential criterion for the clinical diagnosis of chorioamnionitis is maternal fever. Other criteria are relatively insensitive. When defining intra-amniotic infection (chorioamnionitis) for clinical research studies, the diagnosis is typically based on the presence of maternal fever of greater than 38°C (100.4°F) and at least two of the following criteria: maternal leukocytosis (greater than 15 000 cells/mm3), maternal tachycardia (greater than 100 beats/minute), fetal tachycardia (greater than 160 beats/minute), uterine tenderness, and/or foul odor of the amniotic fluid. These thresholds are associated with higher rates of neonatal and maternal morbidity.
Nonetheless, the diagnosis of chorioamnionitis must be considered even when maternal fever is the sole abnormal finding. Although fever is common in women who receive epidural anesthesia (15%–20%), histologic evidence of acute chorioamnionitis is very common in women who become febrile after an epidural (70.6%).3 Furthermore, most of these women with histologic chorioamnionitis do not have a positive placental culture.3 The incidence of clinical chorioamnionitis varies inversely with gestational age. In the National Institute of Child Health and Human Development Neonatal Research Network, 14% to 28% of women delivering preterm infants at 22 through 28 weeks’ gestation exhibited signs compatible with chorioamnionitis.4 The major risk factors for chorioamnionitis include low parity, spontaneous labor, longer length of labor and membrane rupture, multiple digital vaginal examinations (especially with ruptured membranes), meconium-stained amniotic fluid, internal fetal or uterine monitoring, and presence of genital tract microorganisms (eg, Mycoplasma hominis).5
At term gestation, less than 1% of women with intact membranes will have organisms cultured from amniotic fluid.6 The rate can be higher if the integrity of the amniotic cavity is compromised by procedures before birth (eg, placement of a cerclage or amniocentesis).6 In women with preterm labor and intact membranes, the rate of microbial invasion of the amniotic cavity is 32%, and if there is preterm premature rupture of membranes (PPROM), the rate may be as high as 75%.7 Many of the pathogens recovered from amniotic fluid in women with preterm labor or PPROM (eg, Ureaplasma species or Mycoplasma species) do not cause early-onset sepsis.8,–10 However, both Ureaplasma and Mycoplasma organisms can be recovered from the bloodstream of infants whose birth weight is less than 1500 g.11 When a pathogen (eg, GBS) is recovered from amniotic fluid, the attack rate of neonatal sepsis can be as high as 20%.12 Infants born to women with PPROM who are colonized with GBS have an estimated attack rate of 33% to 50% when intrapartum prophylaxis is not given.13
The major risk factors for early-onset neonatal sepsis are preterm birth, maternal colonization with GBS, rupture of membranes >18 hours, and maternal signs or symptoms of intra-amniotic infection.14,–16 Other variables include ethnicity (ie, black women are at higher risk of being colonized with GBS), low socioeconomic status, male sex, and low Apgar scores. Preterm birth/low birth weight is the risk factor most closely associated with early-onset sepsis.17 Infant birth weight is inversely related to risk of early-onset sepsis. The increased risk of early-onset sepsis in preterm infants is also related to complications of labor and delivery and immaturity of innate and adaptive immunity.18
Diagnostic Testing for Sepsis
The clinical diagnosis of sepsis in the neonate is difficult, because many of the signs of sepsis are nonspecific and are observed with other noninfectious conditions. Although a normal physical examination is evidence that sepsis is not present,19,20 bacteremia can occur in the absence of clinical signs.21 Available diagnostic testing is not helpful in deciding which neonate requires empirical antimicrobial therapy but can assist with the decision to discontinue treatment.22
Blood Culture
A single blood culture in a sufficient volume is required for all neonates with suspected sepsis. Data suggest that 1.0 mL of blood should be the minimum volume drawn for culture when a single pediatric blood culture bottle is used. Dividing the specimen in half and inoculating aerobic and anaerobic bottles is likely to decrease the sensitivity. Although 0.5 mL of blood has previously been considered acceptable, in vitro data from Schelonka et al demonstrated that 0.5 mL would not reliably detect low-level bacteremia (4 colony-forming units [CFU]/mL or less).23 Furthermore, up to 25% of infants with sepsis have low colony count bacteremia (≤4 CFU/mL), and two-thirds of infants younger than 2 months of age have colony counts <10 CFU/mL.24,25 Neal et al demonstrated that more than half of blood specimens inoculated into the aerobic bottle were less than 0.5 mL.26 A study by Connell et al indicated that blood cultures with an adequate volume were twice as likely to yield a positive result.27 A blood culture obtained through an umbilical artery catheter shortly after placement for other clinical indications is an acceptable alternative to a culture drawn from a peripheral vein.28 The risk of recovering a contaminant is greater with a blood culture drawn from an umbilical vein.29 There are, however, data to suggest that a blood culture drawn from the umbilical vein at the time of delivery using a doubly clamped and adequately prepared segment of the cord is a reliable alternative to a culture obtained peripherally.30
Urine Culture
A urine culture should not be part of the sepsis workup in an infant with suspected early-onset sepsis.31 Unlike urinary tract infections in older infants (which are usually ascending infections), urinary tract infections in newborn infants are attributable to seeding of the kidney during an episode of bacteremia.
Gastric Aspirates
The fetus swallows 500 to 1000 mL of amniotic fluid each day. Therefore, if there are white blood cells present in amniotic fluid, they will be present in gastric aspirate specimens at birth. However, these cells represent the maternal response to inflammation and have a poor correlation with neonatal sepsis.32 Gram stains of gastric aspirates to identify bacteria are of limited value and are not routinely recommended.33
Body Surface Cultures
Tracheal Aspirates
Lumbar Puncture
The decision to perform a lumbar puncture in a neonate with suspected early-onset sepsis remains controversial. In the high-risk, healthy-appearing infant, data suggest that the likelihood of meningitis is extremely low.38 In the infant with clinical signs that are thought to be attributable to a noninfectious condition, such as respiratory distress syndrome, the likelihood of meningitis is also low.39 However, in bacteremic infants, the incidence of meningitis may be as high as 23%.40,41 Blood culture alone cannot be used to decide who needs a lumbar puncture, because blood cultures can be negative in up to 38% of infants with meningitis.42,43 The lumbar puncture should be performed in any infant with a positive blood culture, infants whose clinical course or laboratory data strongly suggest bacterial sepsis, and infants who initially worsen with antimicrobial therapy. For any infant who is critically ill and likely to have cardiovascular or respiratory compromise from the procedure, the lumbar puncture can be deferred until the infant is more stable.
Cerebrospinal fluid (CSF) values indicative of neonatal meningitis are controversial. In studies that have excluded infants with “traumatic taps” (or nonbacterial illnesses), the mean number of white blood cells in uninfected preterm or term infants was consistently <10 cells/mm3.44,–50 Cell counts 2 standard deviations from the mean were generally less than 20 cells/mm3.46 In a study by Garges et al, the median number of white blood cells in infants who were born at greater than 34 weeks’ gestation and had bacterial meningitis was 477/mm3.43 In contrast, the median number of white blood cells in infants who were born at less than 34 weeks’ gestation and had meningitis was 110/mm3.51 Infants with meningitis attributable to Gram-negative pathogens typically have higher CSF white blood cell counts than do infants with meningitis attributable to Gram-positive pathogens.52 Adjusting the CSF white blood cell count for the number of red blood cells does not improve the diagnostic utility (loss of sensitivity with marginal gain in specificity).53 In addition, the number of bands in a CSF specimen does not predict meningitis.54 With a delay in analysis (>2 hours), white blood cell counts and glucose concentrations decrease significantly.55
Protein concentrations in uninfected, term newborn infants are <100 mg/dL.44,–50 Preterm infants have CSF protein concentrations that vary inversely with gestational age. In the normoglycemic newborn infant, glucose concentrations in CSF are similar to those in older infants and children (70%–80% of a simultaneously obtained blood specimen). A low glucose concentration is the CSF variable with the greatest specificity for the diagnosis of meningitis.43,51 Protein concentrations are higher and glucose concentrations are lower in term than in preterm infants with meningitis. However, meningitis occurs in infants with normal CSF values, and some of these infants have high bacterial inocula.43,51
Peripheral White Blood Cell Count and Differential Count
Total white blood cell counts have little value in the diagnosis of early-onset sepsis and have a poor positive predictive accuracy.56,57 Many investigators have analyzed subcomponents of the white blood cell count (neutrophil indices)—absolute neutrophil count, absolute band count, and immature to total neutrophil (I/T)ratio—to identify infected infants. Like most diagnostic tests for neonatal sepsis, neutrophil indices have proven most useful for excluding infants without infection rather than identifying infected neonates. Neutropenia may be a better marker for neonatal sepsis and has better specificity than an elevated neutrophil count, because few conditions besides sepsis (maternal pregnancy-induced hypertension, asphyxia, and hemolytic disease) depress the neutrophil count of neonates.58 The definitions for neutropenia vary with gestational age,58,–61 type of delivery (infants born by cesarean delivery without labor have lower counts than infants delivered vaginally),61 site of sampling (neutrophil counts are lower in samples from arterial blood),62 and altitude (infants born at elevated altitudes have higher total neutrophil counts).63 In late preterm and term infants, the definition for neutropenia most commonly used is that suggested by Manroe et al (<1800/mm3 at birth and <7800/mm3 at 12–14 hours of age).58 Schmutz et al reinvestigated these reference ranges using modern cell-counting instrumentation in 30 254 infants born at 23 to 42 weeks’ gestation.61 Infants with diagnoses known to affect neutrophil counts (eg, those born to women with pregnancy-induced hypertension or those with early-onset sepsis) were excluded. In this study, the lower limits of normal for neutrophil values at birth were 3500/mm3 in infants born at >36 weeks’ gestation, 1000/mm3 in infants born at 28 through 36 weeks’ gestation, and 500/mm3 in infants born at <28 weeks’ gestation. Peak values occurred at 6 to 8 hours after birth; the lower limits of normal at that time were 7500/mm3, 3500/mm3, and 1500/mm3 for infants born at >36 weeks’ gestation, 28 to 36 weeks’ gestation, and <28 weeks’ gestation, respectively.61 It is noteworthy that the study by Schmutz et al was performed at 4800 feet above sea level, whereas that of Manroe et al was performed at 500 feet above sea level.
The absolute immature neutrophil count follows a similar pattern to the absolute neutrophil count and peaks at approximately 12 hours of life. The number of immature neutrophils increases from a maximal value of 1100 cells/mm3 at birth to 1500 cells/mm3 at 12 hours of age.58 Absolute immature counts have a poor sensitivity and positive predictive accuracy for early-onset sepsis.22 Furthermore, if exhaustion of bone marrow reserves occurs, the number of immature forms will remain depressed.64
The I/T ratio has the best sensitivity of any of the neutrophil indices. However, with manual counts, there are wide interreader differences in band neutrophil identification.65 The I/T ratio is <0.22 in 96% of healthy preterm infants born at <32 weeks’ gestational age.66 Unlike the absolute neutrophil count and the absolute band count, maximum normal values for the I/T ratio occur at birth (0.16) and decline with increasing postnatal age to a minimum value of 0.12.58 In healthy term infants, the 90th percentile for the I/T ratio is 0.27.59 A single determination of the I/T ratio has a poor positive predictive accuracy (approximately 25%) but a very high negative predictive accuracy (99%).66 The I/T ratio may be elevated in 25% to 50% of uninfected infants.67
Exhaustion of bone marrow reserves will result in low band counts and lead to falsely low ratios. The timing of the white blood cell count is critical.68 Counts obtained 6 to 12 hours after birth are more likely to be abnormal than are counts obtained at birth, because alterations in the numbers (and ratios) of mature and immature neutrophils require an established inflammatory response. Therefore, once the decision is made to start antimicrobial therapy soon after birth, it is worth waiting 6 to 12 hours before ordering a white blood cell count and differential count.68,69
Platelet Counts
Despite the frequency of low platelet counts in infected infants, they are a nonspecific, insensitive, and late indicator of sepsis.70,71 Moreover, platelet counts are not useful to follow clinical response to antimicrobial agents, because they often remain depressed for days to weeks after a sepsis episode.
Acute-Phase Reactants
A wide variety of acute-phase reactants have been evaluated in neonates with suspected bacterial sepsis. However, only C-reactive protein (CRP) and procalcitonin concentrations have been investigated in sufficiently large studies.72,73 CRP concentration increases within 6 to 8 hours of an infectious episode in neonates and peaks at 24 hours.74,75 The sensitivity of a CRP determination is low at birth, because it requires an inflammatory response (with release of interleukin-6) to increase CRP concentrations.76 The sensitivity improves dramatically if the first determination is made 6 to 12 hours after birth. Benitz et al have demonstrated that excluding a value at birth, 2 normal CRP determinations (8–24 hours after birth and 24 hours later) have a negative predictive accuracy of 99.7% and a negative likelihood ratio of 0.15 for proven neonatal sepsis.76 If CRP determinations remain persistently normal, it is strong evidence that bacterial sepsis is unlikely, and antimicrobial agents can be safely discontinued. Data are insufficient to recommend following sequential CRP concentrations to determine the duration of antimicrobial therapy in an infant with an elevated value (≥1.0 mg/dL).
Procalcitonin concentrations increase within 2 hours of an infectious episode, peak at 12 hours, and normalize within 2 to 3 days in healthy adult volunteers.77 A physiologic increase in procalcitonin concentration occurs within the first 24 hours of birth, and an increase in serum concentrations can occur with noninfectious conditions (eg, respiratory distress syndrome).78 Procalcitonin concentration has a modestly better sensitivity than does CRP concentration but is less specific.73 Chiesa and colleagues have published normal values for procalcitonin concentrations in term and preterm infants.79 There is evidence from studies conducted in adult populations, the majority of which focused on patients with sepsis in the ICU, that significant reductions in use of antimicrobial agents can be achieved in patients whose treatment is guided by procalcitonin concentration.80
Sepsis Screening Panels
Hematologic scoring systems using multiple laboratory values (eg, white blood cell count, differential count, and platelet count) have been recommended as useful diagnostic aids. No matter what combination of tests is used, the positive predictive accuracy of scoring systems is poor unless the score is very high. Rodwell et al described a scoring system in which a score of 1 was assigned to 1 of 7 findings, including abnormalities of leukocyte count, total neutrophil count, increased immature polymorphonuclear leukocyte (PMN) count, increased I/T ratio, immature to mature PMN ratio >0.3, platelet count ≤150 000/mm3, and pronounced degenerative changes (ie, toxic granulations) in PMNs.81 In this study, two-thirds of preterm infants and 90% of term infants with a hematologic score ≥3 did not have proven sepsis.81 Furthermore, scores obtained in the first several hours after birth have been shown to have poorer sensitivity and negative predictive value than scores obtained at 24 hours of age.67 Sepsis screening panels commonly include neutrophil indices and acute-phase reactants (usually CRP concentration). The positive predictive value of the sepsis screen in neonates is poor (<30%); however, the negative predictive accuracy has been high (>99%) in small clinical studies.22 Sepsis screening tests might be of value in deciding which “high-risk” healthy-appearing neonates do not need antimicrobial agents or whether therapy can be safely discontinued.
Treatment of Infants With Suspected Early-Onset Sepsis
In the United States, the most common pathogens responsible for early-onset neonatal sepsis are GBS and Escherichia coli.17 A combination of ampicillin and an aminoglycoside (usually gentamicin) is generally used as initial therapy, and this combination of antimicrobial agents also has synergistic activity against GBS and Listeria monocytogenes.82,83 Third-generation cephalosporins (eg, cefotaxime) represent a reasonable alternative to an aminoglycoside. However, several studies have reported rapid development of resistance when cefotaxime has been used routinely for the treatment of early-onset neonatal sepsis,84 and extensive/prolonged use of third-generation cephalosporins is a risk factor for invasive candidiasis.85 Because of its excellent CSF penetration, empirical or therapeutic use of cefotaxime should be restricted for use in infants with meningitis attributable to Gram-negative organisms.86 Ceftriaxone is contraindicated in neonates because it is highly protein bound and may displace bilirubin, leading to a risk of kernicterus. Bacteremia without an identifiable focus of infection is generally treated for 10 days.87 Uncomplicated meningitis attributable to GBS is treated for a minimum of 14 days.88 Other focal infections secondary to GBS (eg, cerebritis, osteomyelitis, endocarditis) are treated for longer durations.88 Gram-negative meningitis is treated for minimum of 21 days or 14 days after obtaining a negative culture, whichever is longer.88 Treatment of Gram-negative meningitis should include cefotaxime and an aminoglycoside until the results of susceptibility testing are known.87,88
The duration of antimicrobial therapy in infants with negative blood cultures is controversial. Many women receive antimicrobial agents during labor as prophylaxis to prevent early-onset GBS infections or for management of suspected intra-amniontic infection or PPROM. In those instances, postnatal blood cultures may be sterile (false negative). When considering the duration of therapy in infants with negative blood cultures, the decision should include consideration of the clinical course as well as the risks associated with longer courses of antimicrobial agents. In a retrospective study by Cordero and Ayers, the average duration of treatment in 695 infants (<1000 g) with negative blood cultures was 5 ± 3 days.89 Cotten et al have suggested an association with prolonged administration of antimicrobial agents (>5 days) in infants with suspected early-onset sepsis (and negative blood cultures) with death and necrotizing enterocolitis.90 Two recent papers also support this association.91,92
Prevention Strategies for Early-Onset Sepsis
The only intervention proven to decrease the incidence of early-onset neonatal sepsis is maternal treatment with intrapartum intravenous antimicrobial agents for the prevention of GBS infections.93 Adequate prophylaxis is defined as penicillin (the preferred agent), ampicillin, or cefazolin given for ≥4 hours before delivery. Erythromycin is no longer recommended for prophylaxis because of high resistance rates. In parturients who have a nonserious penicillin allergy, cefazolin is the drug of choice. For parturients with a history of serious penicillin allergy (anaphylaxis, angioedema, respiratory compromise, or urticaria), clindamycin is an acceptable alternative agent, but only if the woman’s rectovaginal GBS screening isolate has been tested and documented to be susceptible. If the clindamycin susceptibility is unknown or the GBS isolate is resistant to clindamycin, vancomycin is an alternative agent for prophylaxis. However, neither clindamycin nor vancomycin has been evaluated for efficacy in preventing early-onset GBS sepsis in neonates. Intrapartum antimicrobial agents are indicated for the following situations93:
Positive antenatal cultures or molecular test at admission for GBS (except for women who have a cesarean delivery without labor or membrane rupture)
Unknown maternal colonization status with gestation <37 weeks, rupture of membranes >18 hours, or temperature >100.4°F (>38°C)
GBS bacteriuria during the current pregnancy
Previous infant with invasive GBS disease
Management guidelines for the newborn infant have been published93 and are available online (http://www.cdc.gov/groupbstrep/guidelines/index.html).
Clinical Challenges
Challenge 1: Identifying Neonates With Clinical Signs of Sepsis With a “High Likelihood” of Early-Onset Sepsis Who Require Antimicrobial Agents Soon After Birth
Most infants with early-onset sepsis exhibit abnormal signs in the first 24 hours of life. Approximately 1% of infants will appear healthy at birth and then develop signs of infection after a variable time period.21 Every critically ill infant should be evaluated and receive empirical broad-spectrum antimicrobial therapy after cultures, even when there are no obvious risk factors for sepsis. The greatest difficulty faced by clinicians is distinguishing neonates with early signs of sepsis from neonates with noninfectious conditions with relatively mild findings (eg, tachypnea with or without an oxygen requirement). In this situation, data are insufficient to guide management. In more mature neonates without risk factors for infection who clinically improve over the first 6 hours of life (eg, need for oxygen is decreasing and respiratory distress is resolving), it is reasonable to withhold antimicrobial therapy and monitor the neonates closely. The 6-hour window should not be considered absolute; however, most infants without infection demonstrate some improvement over that time period. Any worsening of the infant’s condition should prompt starting antimicrobial agents after cultures have been obtained.
Challenge 2: Identifying Healthy-Appearing Neonates With a “High Likelihood” of Early-Onset Sepsis Who Require Antimicrobial Agents Soon After Birth
This category includes infants with 1 of the risk factors for sepsis noted previously (colonization with GBS, prolonged rupture of membranes >18 hours, or maternal chorioamnionitis). GBS is not a risk factor if the mother has received adequate intrapartum therapy (penicillin, ampicillin, or cefazolin for at least 4 hours before delivery) or has a cesarean delivery with intact membranes in the absence of labor.93 The risk of infection in the newborn infant varies considerably with the risk factor present. The greatest risk of early-onset sepsis occurs in infants born to women with chorioamnionitis who are also colonized with GBS and did not receive intrapartum antimicrobial agents. Early-onset sepsis does occur in infants who appear healthy at birth.21 Therefore, some clinicians use diagnostic tests with a high negative predictive accuracy as reassurance that infection is not present (allowing them to withhold antimicrobial agents). The decision of whether to treat a high-risk infant depends on the risk factors present, the frequency of observations, and gestational age. The threshold for initiating antimicrobial treatment generally decreases with increasing numbers of risk factors for infection and greater degrees of prematurity. Suggested algorithms for management of healthy-appearing, high-risk infants are shown in Figs 1, 2, and 3. Screening blood cultures have not been shown to be of value.21
Evaluation of asymptomatic infants <37 weeks’ gestation with risk factors for sepsis. aThe diagnosis of chorioamnionitis is problematic and has important implications for the management of the newborn infant. Therefore, pediatric providers are encouraged to speak with their obstetrical colleagues whenever the diagnosis is made. bLumbar puncture is indicated in any infant with a positive blood culture or in whom sepsis is highly suspected on the basis of clinical signs, response to treatment, and laboratory results. IAP, intrapartum antimicrobial prophylaxis; WBC, white blood cell; Diff, differential white blood cell count.
Evaluation of asymptomatic infants <37 weeks’ gestation with risk factors for sepsis. aThe diagnosis of chorioamnionitis is problematic and has important implications for the management of the newborn infant. Therefore, pediatric providers are encouraged to speak with their obstetrical colleagues whenever the diagnosis is made. bLumbar puncture is indicated in any infant with a positive blood culture or in whom sepsis is highly suspected on the basis of clinical signs, response to treatment, and laboratory results. IAP, intrapartum antimicrobial prophylaxis; WBC, white blood cell; Diff, differential white blood cell count.
Evaluation of asymptomatic infants ≥37 weeks’ gestation with risk factors for sepsis. aThe diagnosis of chorioamnionitis is problematic and has important implications for the management of the newborn infant. Therefore, pediatric providers are encouraged to speak with their obstetrical colleagues whenever the diagnosis is made. bLumbar puncture is indicated in any infant with a positive blood culture or in whom sepsis is highly suspected on the basis of clinical signs, response to treatment, and laboratory results. WBC, white blood cell; Diff, differential white blood cell count.
Evaluation of asymptomatic infants ≥37 weeks’ gestation with risk factors for sepsis. aThe diagnosis of chorioamnionitis is problematic and has important implications for the management of the newborn infant. Therefore, pediatric providers are encouraged to speak with their obstetrical colleagues whenever the diagnosis is made. bLumbar puncture is indicated in any infant with a positive blood culture or in whom sepsis is highly suspected on the basis of clinical signs, response to treatment, and laboratory results. WBC, white blood cell; Diff, differential white blood cell count.
Evaluation of asymptomatic infants ≥37 weeks’ gestation with risk factors for sepsis (no chorioamnionitis). aInadequate treatment: Defined as the use of an antibiotic other than penicillin, ampicillin, or cefazolin or if the duration of antibiotics before delivery was <4 h. bDischarge at 24 h is acceptable if other discharge criteria have been met, access to medical care is readily accessible, and a person who is able to comply fully with instructions for home observation will be present. If any of these conditions is not met, the infant should be observed in the hospital for at least 48 h and until discharge criteria are achieved. IAP, intrapartum antimicrobial prophylaxis; WBC, white blood cell; Diff, differential white blood cell count.
Evaluation of asymptomatic infants ≥37 weeks’ gestation with risk factors for sepsis (no chorioamnionitis). aInadequate treatment: Defined as the use of an antibiotic other than penicillin, ampicillin, or cefazolin or if the duration of antibiotics before delivery was <4 h. bDischarge at 24 h is acceptable if other discharge criteria have been met, access to medical care is readily accessible, and a person who is able to comply fully with instructions for home observation will be present. If any of these conditions is not met, the infant should be observed in the hospital for at least 48 h and until discharge criteria are achieved. IAP, intrapartum antimicrobial prophylaxis; WBC, white blood cell; Diff, differential white blood cell count.
Conclusions
The diagnosis and management of neonates with suspected early-onset sepsis are based on scientific principles modified by the “art and experience” of the practitioner. The following are well-established concepts related to neonatal sepsis:
Neonatal sepsis is a major cause of morbidity and mortality.
Diagnostic tests for early-onset sepsis (other than blood or CSF cultures) are useful for identifying infants with a low probability of sepsis but not at identifying infants likely to be infected.
One milliliter of blood drawn before initiating antimicrobial therapy is needed to adequately detect bacteremia if a pediatric blood culture bottle is used.
Cultures of superficial body sites, gastric aspirates, and urine are of no value in the diagnosis of early-onset sepsis.
Lumbar puncture is not needed in all infants with suspected sepsis (especially those who appear healthy) but should be performed for infants with signs of sepsis who can safely undergo the procedure, for infants with a positive blood culture, for infants likely to be bacteremic (on the basis of laboratory data), and infants who do not respond to antimicrobial therapy in the expected manner.
The optimal treatment of infants with suspected early-onset sepsis is broad-spectrum antimicrobial agents (ampicillin and an aminoglycoside). Once the pathogen is identified, antimicrobial therapy should be narrowed (unless synergism is needed).
Antimicrobial therapy should be discontinued at 48 hours in clinical situations in which the probability of sepsis is low.
Lead Author
Richard A. Polin, MD
Committee on Fetus and Newborn, 2011–2012
Lu-Ann Papile, MD, Chairperson
Jill E. Baley, MD
William Benitz, MD
Waldemar A. Carlo, MD
James Cummings, MD
Praveen Kumar, MD
Richard A. Polin, MD
Rosemarie C. Tan, MD, PhD
Kasper S. Wang, MD
Kristi L. Watterberg, MD
FORMER COMMITTEE MEMBER
Vinod K. Bhutani, MD
Liaisons
CAPT Wanda Denise Barfield, MD, MPH – Centers for Disease Control and Prevention
George Macones, MD – American College of Obstetricians and Gynecologists
Ann L. Jefferies, MD – Canadian Paediatric Society
Rosalie O. Mainous, PhD, RNC, NNP – National Association of Neonatal Nurses
Tonse N. K. Raju, MD, DCH – National Institutes of Health
FORMER LIAISON
William Barth, Jr, MD – American College of Obstetricians and Gynecologists
Staff
Jim Couto, MA
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.
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.
Comments
Re:Empirical Antibiotic Therapy for Suspected Early-Onset Bacterial Sepsis
We have received four letters to the editor[1-4] regarding the recently published clinical report from the Committee on Fetus and Newborn (COFN) titled “Management of Neonates With Suspected or Proven Early-Onset Sepsis,”[5] which raise a number of important issues. Like the authors of these letters, we are concerned about unnecessary evaluations for sepsis and treatment of asymptomatic infants for an inappropriate length of time. The dangers of prolonged antibiotic therapy in the preterm population are well described.[6,7,8] Neither the Centers for Disease Control and Prevention (CDC) guideline on group B streptococcal disease (GBS)[9] nor the COFN clinical report on early-onset sepsis addresses the duration of antibiotic therapy. Furthermore, although there are data to support which groups of high-risk infants should receive empirical antibiotics, there are only observational data regarding the duration of antibiotic therapy in late-preterm and term infants[10,11] and no data in the very low birth weight population.
The risk of sepsis in asymptomatic infants with risk factors for sepsis is low but significant (0.5-1.0%).[10,12] The major area of uncertainty is the number of days an asymptomatic infant born to a woman treated for chorioamnionitis (or other risk factors) should receive antibiotics when the infant’s laboratory studies are abnormal. We would not treat a well-appearing, asymptomatic term infant with a negative blood culture longer than 48 to 72 hours, whose mother was treated for chorioamnionitis, even when the infant’s laboratory data are abnormal. Our rationale is as follows. The postnatal blood culture is difficult to interpret when antibiotics have been administered before delivery. Adjunct tests for neonatal sepsis generally have a high negative predictive value (predicting absence of sepsis) and mediocre positive predictive value for identifying infected infants (40% or less).[13] Experts might debate if a 40% risk of sepsis is high; however, by 48 to 72 hours of life, a normal physical examination in an otherwise well infant should exclude the possibility of early-onset sepsis.[9] When laboratory studies in the infant are normal, we would stop antibiotics by 48 hours. It is worth noting that in the study by Jackson et al,[10] only 13% of asymptomatic infants (of greater than or equal to 35 weeks’ gestation) had an elevated I/T ratio at age 12 hours (using the reference ranges of Manroe). This is the neutrophil index with the highest sensitivity for early-onset neonatal sepsis.
In a preterm infant born to a woman with chorioamnionitis, prolonged rupture of membranes (PROM) greater than 18 hours, or inadequately treated maternal colonization with GBS, we suggest treatment for 72 hours (and not 48-72 hours) given the higher risk of sepsis in the preterm population and the greater difficulty with clinical assessments. The guidance of the COFN to treat preterm infants born to women with PROM greater than 18 hours or women colonized with GBS who received inadequate (or no) intrapartum antibiotic prophylaxis (IAP) differs from the CDC GBS guidelines, which recommended a limited evaluation but no treatment. COFN made that recommendation because of the significantly higher risk of infection in the preterm population when those risk factors for sepsis are also present.[12]
We agree with the Cotten et al that treatment for 48 hours might be an acceptable alternative in the late preterm infant or term infant. However, that needs to be determined by the treating physician. As stated in the CDC GBS guideline, “This revised algorithm is not an exclusive approach to management; variation that incorporates individual circumstances or institutional preferences may be appropriate.” The COFN agrees with that philosophy.
It is also worth emphasizing that the algorithms apply to only a small segment of the neonatal population. In the CDC Active Bacterial Core surveillance system, the incidence of chorioamnionitis was 3.1% and the incidence of PROM greater than 18 hours was 7.2%.[14] Furthermore, as noted above, only a minority of infants will have an elevated I/T ratio by age 12 hours. Therefore, our guidance to treat those infants 48 to 72 hours will only apply to a small subset of the term neonatal population. There are also some preterm infants (mostly late preterm infants) who are completely asymptomatic at birth and who remain asymptomatic. Our algorithm applies to those infants. With increasing degrees of prematurity, most preterm infants will be symptomatic. In the study by Cordero and Ayers, the average duration of antimicrobial therapy in extremely low birth weight infants with negative blood cultures was 5 plus/minus 3 days.[15] We hope this guidance will reduce the duration of antimicrobial therapy in that population.
Dr. Puopolo raises a few other issues we would like to address. The lack of specificity in the diagnosis of chorioamnionitis is a concern for all practitioners. Until the American College of Obstetrics and Gynecology arrives at a new definition, we would follow the recommendation in the 2010 CDC GBS guidelines, “Consultation with obstetric providers to assess whether chorioamnionitis was suspected is important to determine neonatal management.” As noted by Dr. Puopolo, the COFN suggests a white blood count and differential count at 6 to 12 hours of life in term infants with risk factors for sepsis other than chorioamnionitis (i.e, infants born to women who are GBS positive who received inadequate intrapartum antibiotic prophylaxis and infants born following PROM greater than 18 hours). The CDC GBS guideline recommends a limited evaluation (blood culture and white blood count and differential count) and observation if there has been inadequate IAP and PROM greater than 18 hours. In an ideal world where every infant can be observed closely for changes in clinical activity or vital signs, no additional testing should be needed. If close observation of an “at-risk” term infant can be guaranteed with frequent vital signs (every 2 to 4 hours) for 24 hours, that remains an acceptable alternative to testing. However, in many nurseries, observations occur less frequently and may not identify abnormal clinical signs at the earliest possible time point. A single white blood count and differential count (without a blood culture) is not an unreasonable way to be reassured that the infant is very unlikely to be infected. An abnormal result would prompt a blood culture in the COFN algorithm (and perhaps closer observation), but no treatment. Dr. Puopolo notes her recently published paper, which provides a formula to estimate the probability of sepsis in infants’ greater than or equal to 34 weeks gestation.[16] We like this approach, and if validated in other studies, it is likely to decrease the number of sepsis evaluations in that population. Lastly, Sise et al suggest an additional biomarker for the diagnosis of sepsis (urinary neutrophil gelatinase associated lipocalin [NGAL])[17]. However, it needs further evaluation in infants with early-onset sepsis.
Finally, we have chosen to publish this report as guidance for all practitioners who care for ill infants. The COFN wanted to make sure the guidance applies to practitioners with varying levels of expertise who care for infants in a wide variety of settings. Revised algorithms are being submitted to Pediatrics as an erratum to the clinical report.
1. Cotten M, Benjamin DK, Clark R et al. Empirical antibiotic therapy for suspected early-onset sepsis. Pediatrics
2. Puopolo K. Letter to the editor. Pediatrics
3. Sise ME, Parravicini E and Barasch J,Urinary Neutrophil Gelatinase Associated Lipocalin (NGAL) identifies neonates with high probability of sepsis. Pediatrics
4. Sukumar M. Letter to the editor. Pediatrics
5. Polin RA and the Committee on Fetus and Newborn. Management of neonates with suspected or proven early-onset bacterial sepsis. Pediatrics. 2012;129(5):1006–1015.
6. Cotten CM, Taylor S, Stoll B, Goldberg RN, Hansen NI, Sanchez PJ, et al. Prolonged duration of initial empirical antibiotic treatment is associated with increased rates of necrotizing enterocolitis and death for extremely low birth weight infants. Pediatrics 2009;123(1): 58-66.
7. Alexander VN, Northrup V, Bizzarro MJ. Antibiotic Exposure in the Newborn Intensive Care Unit and the Risk of Necrotizing Enterocolitis. J Pediatr 2011;1593(3):392-397.
8. Kuppala VS, Meinzen-Der J, Morrow AL, Schibler KR. Prolonged Initial Empirical Antibiotic Treatment is Associated with Adverse Outcomes in Premature Infants. J Pediatr 2011;159(5):720-725.
9. Centers for Disease Control and Prevention. Prevention of Perinatal Group B Streptococcal Disease: Revised Guidelines from CDC 2010. MMWR 2010;59 RR-10, 1-32.
10. Escobar GJ, Li DK, Armstrong MA et al. Neonatal sepsis workups in infants ≥ 2000 grams at birth: A population based study. Pediatrics 2000;106(2 pt 1);256-63
11. Jackson, GL, Engle WD, Sendelbach DM et al. Are complete blood counts useful in the evaluation of asymptomatic neonates exposed to suspected chorioamnionitis? Pediatrics. 2004;113, 1173-1180.
12. Ottolini MC, Lundgren K, Mirkinson LJ, Cason S and Ottolini MG. Utility of complete blood count and blood culture screening to diagnose neonatal sepsis in the asymptomatic at risk newborn. Pediatr. Infect. Did. J. 2003;22(5):430-434.
13. Gerdes JS. Clinicopathologic approach to the diagnosis of neonatal sepsis. Clin Perinatol. 1991;18(2) 361-381.
14. Van-Dyke M, Phares CR, Lynfield R, Thomas AR et al. Evaluation of universal antenatal screening for group B streptococcus. N Engl J. Med. 2009;360:2626-2636.
15. Cordero L, Ayers LW. Duration of empiric antibiotics for suspected early-onset sepsis in extremely low birth weight infants. Infect. Control Hosp. Epidemiol. 2003;34 (9): 662-666.
16. Puopolo KM, Draper D, Wi S, Newman TB, Zupancic J, Lieberman E, et al. Estimating the probability of neonatal early-onset infection on the basis of maternal risk factors. Pediatrics 2011;128(5):e1155-e1163.
17. Parravicini E, Nemerofsky SL, Michelson KA, et al. Urinary neutrophil gelatinase-associated lipocalin is a promising biomarker for late onset culture-positive sepsis in very low birth weight infants. Pediatr Res 2010;67:636-40.
Richard A. Polin, MD, on behalf of the Committee on Fetus and Newborn, 2011-2012
1. Lu-Ann Papille MD Chairperson
2. Jill E. Baley MD
3. William Benitz MD
4. Waldemar Carlo MD
5. James Cummings MD
6. Eric C Eichenwald MD
7. Praveen Kumar MD
8. Richard A Polin MD
9. Kristy Watterberg MD
10. Rose Tan MD
11. Kasper S Wang MD
12. Vinod K Bhutani MD
Conflict of Interest:
None declared
Need Clarification on " abnormal labs"
The clinical report on this topic was fascinating to read.However I have the following observations /concerns A. The committee recommends treating with prolonged course of broad spectrum antibiotics in well infants who have " abnormal labs". It is unclear what is considered as abnormal lab result. Is it just one abnormal I/T ratio or it is abnormal CBC + elevated CRP? The poor positive predictive value of either of these tests is well documented in published studies. B. The committee acknowledges the fact prolonged antibiotic course in the neonates can be harmful. Hence it is important to know how long the antibiotics ( 7 or 10 days) should be continued in culture negative babies with initial abnormal and/or CRP Sincerely
Mike Sukumar MD
Conflict of Interest:
None declared
Response to the American Academy of Pediatrics, Committee on the Fetus and Newborn statement, "Management of Neonates with Suspected or Proven Early-Onset Bacterial Sepsis"
To the Editor: I am writing in response to the American Academy of Pediatrics Committee on the Fetus and Newborn (COFN) statement, "Management of Neonates with Suspected or Proven Early-Onset Bacterial Sepsis," published in the May 2012 edition of Pediatrics (1).
I am glad that Dr. Polin and COFN have given emphasis to the current issues of concern in the evaluation of both preterm and term infants for early-onset sepsis (EOS). However, several of the recommendations continue traditional approaches to this problem, without acknowledging recent data on this subject that both raises concern about prolonged antibiotic administration and offers alternative approaches to risk assessment. My specific concerns are as follows:
* The statement continues to refer to the risk presented by maternal chorioamnionitis without offering a standard definition of this clinical diagnosis. From a practical standpoint, many obstetricians and neonatologists use this term almost interchangeably with intrapartum maternal fever. Our obstetrical colleagues often express reluctance to definitively make - or rule out - a diagnosis that they know will influence neonatal care. Both the COFN statement and the Centers for Disease Control and Prevention revised guidelines for the prevention of perinatal group B Streptococcal (GBS) disease (2) can be difficult to implement given the lack of precision in this diagnosis.
* Figure 2 proposes an algorithm for the management of asymptomatic term infants with exposure to chorioamnionitis. This algorithm recommends extending antibiotic therapy (for an undefined period) if the infant remains well, with a negative blood culture, if screening laboratory data is abnormal and intrapartum antibiotics were administered to the mother. This recommendation continues the belief that the administration of intrapartum antibiotics simply renders the blood culture unreliable, rather than that these antibiotics provide protection to the infant. What if the mother did not receive intrapartum antibiotics? Can the infant then have antibiotics stopped and be sent home after 48 hours, despite the abnormal laboratory values? I recognize that clinicians fear discharging an infant who will later fall ill at home. Gabriel Escobar's work on neonatal sepsis evaluations among infants with birth weight > 2000 g demonstrated that this is a relatively rare event. This study included follow-up after discharge through the first week of life among 2785 infants evaluated for EOS. Among these infants, 4/2785 (0.14%) were later readmitted and diagnosed with viral or bacterial infection in the first week of life (3).
* Figure 1 proposes an algorithm for the management of asymptomatic preterm infants with any risk factor for infection. This algorithm also recommends extending antibiotic therapy (for an undefined period) if the infant remains well, with a negative blood culture, if screening laboratory data is abnormal and intrapartum antibiotics were administered to the mother. The issue of evaluating preterm infants is more complex than term infants, and I would argue that "preterm infants < 37 weeks gestation" is too broad a category, given the very different risk among infants with very-low birth weight, compared to late preterm infants. In fact, the consequences of prolonged antibiotic treatment for culture- negative sepsis may be highest among the smallest infants. At least three recent papers find negative effects on survival and the incidence of NEC after prolonged antibiotic treatment of very-low birth weight in the first week of life (4-6). Underlying the COFN recommendations is the belief that blood cultures are unreliable indicators of infection if obtained after maternal intrapartum antibiotic treatment. Preterm infants are not generally discharged from the hospital in the first week of life. Given the potential dangers of antibiotic treatment to the smallest of these infants, I would suggest that these infants present an opportunity to study the overall risks and benefits of a policy of treating only culture- positive infection.
* Figure 3 advocates EOS evaluation of asymptomatic infants born at term, on the basis of inadequate GBS prophylaxis or with ROM equal to or greater than 18 hours - in the absence of any indication of chorioamnionitis. These recommendations contradict those contained in the CDC 2010 GBS guidelines, and thus present neonatal clinicians with conflicting expert opinions, making local care decisions even more difficult.
Overall, the algorithms contained in this statement continue the practice of giving consideration to EOS risk factors in isolation, without consideration of the relative contribution of each risk factor, or the effect of combinations of these factors. Recently, Gabriel Escobar, myself and colleagues published a large case-control study revisiting intrapartum risk factors for EOS in the era of GBS prophylaxis (7). We used objective data only, to develop a multivariate model for prediction of EOS among term and late-preterm infants. We validated this model and demonstrated that it can perform better than algorithms based on cut-off values. The model only uses objective data; perhaps most important, we used "highest maternal intrapartum temperature" rather than the clinical diagnosis of chorioamnionitis. The model also accounts for use of all types and time frames of intrapartum antibiotics. This model can be incorporated with an electronic medical record, but a publically-accessible risk calculator is now available at http://www.dor.kaiser.org/external/DORExternal/research/InfectionProbabilityCalculator.aspx. This model is not meant to be used in isolation - infant clinical status and laboratory values must also be considered - but we hope it can form the basis of a more effective means of identifying term and late preterm infants who should be subject to EOS evaluations.
Neonatal EOS evaluation remains a subject of significant controversy, and a dangerous one at that -among term infants it is a fairly low incidence, but very high consequence condition. The COFN statement has given attention to important issues, and hopefully will motivate neonatal clinicians to continue to seek safe and effective approaches to EOS risk.
References:
1. Polin RA and the Committee on the Fetus and Newborn. Management of Neonates with Suspected or Proven Early-Onset Bacterial Sepsis. Pediatrics 2012;129(5):1006-1015. 2. Verani JR, McGee L, Schrag SJ; Division of Bacterial Diseases, National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention (CDC). Prevention of perinatal group B streptococcal disease--revised guidelines from CDC, 2010. MMWR Recomm Rep 2010;59:1-36. 3. Escobar GJ, Li DK, Armstrong MA, Gardner MN, Folck BF, Verdi JE, Xiong B, Bergen R. Neonatal sepsis workups in infants >/=2000 grams at birth: A population-based study. Pediatrics 2000;106(2):256-263. 4. Cotten CM, Taylor S, Stoll B, Goldberg RN, Hansen NI, Sanchez PJ, et al. Prolonged duration of initial empirical antibiotic treatment is associated with increased rates of necrotizing enterocolitis and death for extremely low birth weight infants. Pediatrics 2009;123(1): 58-66. 5. Alexander VN, Northrup V, Bizzarro MJ. Antibiotic Exposure in the Newborn Intensive Care Unit and the Risk of Necrotizing Enterocolitis. J Pediatr 2011;1593():392-397. 6. Kuppala VS, Meinzen-Der J, Morrow AL, Schibler KR. Prolonged Initial Empirical Antibiotic Treatment is Associated with Adverse Outcomes in Premature Infants. J Pediatr 2011;159(5):720-725 7. Puopolo KM, Draper D, Wi S, Newman TB, Zupancic J, Lieberman E, et al. Estimating the probability of neonatal early-onset infection on the basis of maternal risk factors. Pediatrics 2011;128(5):e1155-e1163.
Conflict of Interest:
None declared
Urinary Neutrophil Gelatinase Associated Lipocalin (NGAL) identifies neonates with high probability of sepsis
The diagnosis of "suspected sepsis" in neonates in the intensive care unit is challenging given the non-specific signs of sepsis and the poor diagnostic performance of currently used laboratory markers and the unfortunate delay in bacterial culture data.(1,2) In this excellent review, Polin reports that while sepsis screening panels and scoring systems that include multiple laboratory values may help exclude neonatal sepsis, their positive predictive value is very poor, less than 30%.(3,4) While the negative predictive value is of great importance, this review identifies the importance of identifying a marker of "high likelihood" of early-onset sepsis in neonates who require antimicrobial agents soon after birth - in short, a marker with a useful positive predictive value.(4) Neutrophil gelatinase associated lipocalin (NGAL) is a member of the lipocalin superfamily expressed by neutrophils and kidney tubular epithelia in response to ischemia, hypoxia, sepsis and drug toxicity. NGAL acts as an iron scavenger preventing bacterial growth and hence is a critical component of the defense against infection.(5) NGAL is a robust marker of acute kidney injury in adults(6) and children, and we have identified it as a marker of sepsis in adults and in very low birth weight (VLBW) infants.(7) We studied NGAL as an early biomarker of late-onset blood culture positive sepsis in VLBW infants and found 75% sensitivity, 84% specificity, 67% positive predictive value, and 89% negative predictive value, when compared to VLBW infants without sepsis. The sensitivity and positive predictive value of this single urinary biomarker far exceeded even panels of biomarkers used to detect sepsis.(4) Although the test characteristics may be different in early-onset compared to late- onset bacterial sepsis and in full-term compared to VLBW infants, we believe NGAL may offer not only a robust negative predictive value, but also improved positive predictive values and hence may be able to identify infants at high risk of being infected. NGAL deserves further investigation in neonatal populations.
1. Rozycki HJ, Stahl GE, Baumgart S. Impaired sensitivity of a single early leukocyte count in screening for neonatal sepsis. Pediatr Infect Dis J 1987;6:440-2. 2. Manzoni P, Mostert M, Galletto P, et al. Is thrombocytopenia suggestive of organism-specific response in neonatal sepsis? Pediatr Int 2009;51:206- 10. 3. Gerdes JS. Clinicopathologic approach to the diagnosis of neonatal sepsis. Clin Perinatol 1991;18:361-81. 4. Polin RA. Management of Neonates With Suspected or Proven Early-Onset Bacterial Sepsis. Pediatrics 2012;129:1006-15. 5. Flo TH, Smith KD, Sato S, et al. Lipocalin 2 mediates an innate immune response to bacterial infection by sequestrating iron. Nature 2004;432:917 -21. 6. Sise ME, Forster C, Maarouf O, Schmidt-Ott K, Nickolas TL, Barasch J. Urinary Neutrophil Gelatinase Associated Lipocalin (uNGAL) identifies serious bacterial sepsis in patients with systemic inflammatory response syndrome. World Congress of Nephrology Vancouver, Canada 2011. 7. Parravicini E, Nemerofsky SL, Michelson KA, et al. Urinary neutrophil gelatinase-associated lipocalin is a promising biomarker for late onset culture-positive sepsis in very low birth weight infants. Pediatr Res 2010;67:636-40.
Conflict of Interest:
None declared
Empirical Antibiotic Therapy for Suspected Early-Onset Bacterial Sepsis
We are writing in response to the recently published article, "Management of Neonates with Suspected or Proven Early-Onset Bacterial Sepsis" by Richard Polin and the American Academy of Pediatrics (AAP) Committee on the Fetus and Newborn (COFN) (1).
We agree with Dr. Polin and the committee that empirical antimicrobial initiation and cessation are based on scientific principles modified by art and experience. The review is thorough and well referenced. However, we are concerned that the algorithms provided in this report may lead to increased use and duration of empirical antibiotics.
Universal screening for maternal group B streptococcal (GBS) colonization and use of intrapartum antibiotic prophylaxis have resulted in substantial reduction in early-onset GBS among newborns. Despite this progress, GBS remains the leading cause of early-onset neonatal sepsis in the United States, with continued burden of disease (2). In 2010, the Centers for Disease Control and Prevention (CDC) issued revised Guidelines for the Prevention of Perinatal Group B Streptococcal Disease (3). The algorithm included in the 2010 CDC guidelines identifies only two categories of infants for empirical treatment: those with signs of sepsis and those with maternal chorioamnionitis. Well-appearing infants <37 weeks gestation with inadequate intrapartum antibiotic prophylaxis (IAP) and infants >37 weeks with rupture of membranes >18 hours are to receive limited evaluation and observation for at least 48 hours. As the consequences of missing early-onset sepsis can be grave, it is difficult to fault clinicians for initiating empirical antibiotics in preterm infants with risk factors. This may be the reason for the alternative strategy depicted in Figure 1 of the clinical report by Polin et al., which suggests treating all well-appearing infants <37 weeks gestation with inadequate IAP with empirical antibiotics (1).
Although Polin and the COFN state that "Antimicrobial therapy should be discontinued at 48 hours in clinical situations in which the probability of sepsis is low," Figures 1 and 2 in the report may lead clinicians to prolong empirical antibiotic courses. For example, Figure 1 indicates that if cultures are negative and infants appear well but laboratory data are abnormal, antibiotics are to be continued. Duration of empirical therapy is not discussed, and the only laboratory tests listed on the algorithms are from the first 12 postnatal hours. We are concerned that only using laboratory results obtained in the first 6-12 postnatal hours to determine duration of empirical therapy does not provide adequate positive or negative predictive value in most situations and will lead to increases in duration of antimicrobial courses, despite sterile cultures. For example, would an abnormal test be defined as one elevated C-reactive protein (CRP), or one elevated band count, or one elevated white blood cell count? In a report of over 800 well-appearing near-term and term infants exposed to suspected chorioamnionitis, 99% of the infants had at least one abnormal finding on the complete blood cell count, while only 0.5% had a positive blood culture (4). Thus, following the algorithm outlined in Figure 1, as many as 99% of infants with suspected chorioamnionitis might be candidates for empirical antibiotic courses longer than 48 hours, despite sterile cultures. Guidance on definitions and timing of tests such as CRP after the first 12 hours that can provide reassurance regarding the very low likelihood of sepsis would be helpful for clinicians (5).
The COFN suggestions could be strengthened with recent guidance from the Centers for Disease Control and Prevention to "Take an Antibiotic Time Out" when culture results are available at 24-48 hours: "Stop and reassess therapy. Antibiotics are generally started before a patient's full clinical picture is known. Now that additional information is available, including microbiology, radiographic, and clinical information, clinicians should ask themselves if the antibiotic is still warranted or, more importantly, is this antibiotic still effective against this organism? It is the time to reevaluate why the therapy was started in the first place and to gather all of the evidence on whether there should be changes in the course of therapy or the antibiotics should be stopped altogether if an infection no longer appears likely" (6).
Unfortunately no antibiotic is without risk. Every time we administer antibiotics, especially prolonged antibiotics, we expose each infant to increased risk of subsequent infection with resistant organisms, invasive candidiasis, necrotizing enterocolitis, late-onset sepsis, and death (7,8). The risk for each infant of iatrogenic harm from prolonged antibiotic therapy is admittedly small (<1%). However, the risk of culture-proven early-onset bacteremia is 0.1%, and the risk of culture- negative early-onset bacteremia (false negative blood culture) is also likely <1%.2 Thus, continued antibiotics as outlined in the algorithms proposed by the COFN may well harm more infants than they will protect.
The proposed algorithms have potential public health implications as well. When we administer surfactant or caffeine to infant A, there is no potential risk to infant B located in the same nursery. Antibiotics are different. Every time we provide prolonged antibiotics to one infant, we expose every infant in the nursery to a small increased risk of resistant infection (9).
We agree that clinicians face difficult choices in initiation and duration of empirical antibiotics for early-onset sepsis. Too little use may result in preventable neonatal deaths from infection; too much use, especially prolonged use, may lead to increased mortality and morbidity (7,8). We agree that clinicians should have a low threshold for starting antibiotics in high-risk infants, but timely discontinuation of antibiotics in the face of negative blood cultures is necessary to reduce the risks associated with prolonged use of empirical antibiotics.
References
1. Polin RA.; the Committee on Fetus and Newborn. Management of neonates with suspected or proven early-onset bacterial sepsis. Pediatrics. 2012;129(5):1006-1015.
2. Stoll BJ, Hansen NI, S?nchez PJ, Faix RG, Poindexter BB, Van Meurs KP, Bizzarro MJ, Goldberg RN, Frantz ID 3rd, Hale EC, Shankaran S, Kennedy K, Carlo WA, Watterberg KL, Bell EF, Walsh MC, Schibler K, Laptook AR, Shane AL, Schrag SJ, Das A, Higgins RD; Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network. Early onset neonatal sepsis: the burden of group B Streptococcal and E. coli disease continues. Pediatrics. 2011;127(5):817-826. Erratum in: Pediatrics. 2011;128(2):390. Baker CJ, Byington CL, Polin RA. Policy statement: recommendations for the prevention of perinatal group B streptococcal (GBS) disease. Pediatrics. 2011;128(3):611-616.
3. Verani JR, McGee L, Schrag SJ; Division of Bacterial Diseases, National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention (CDC).Prevention of perinatal group B streptococcal disease--revised guidelines from CDC, 2010. MMWR Recomm Rep. 2010;59(RR-10):1-36.
4. Jackson GL, Engle WD, Sendelbach DM, Vedro DA, Josey S, Vinson J, Bryant C, Hahn G, Rosenfeld CR. Are complete blood cell counts useful in the evaluation of asymptomatic neonates exposed to suspected chorioamnionitis? Pediatrics. 2004;113(5):1173-1180.
5. Benitz WE, Han MY, Madan A, Ramachandra P. Serial serum C-reactive protein levels in the diagnosis of neonatal infection. Pediatrics. 1998;102(4):E41.
6. Centers for Disease Control and Prevention. 12-Step Program to Prevent Antimicrobial Resistance in Health Care Settings. Available at: http://www.cdc.gov/drugresistance/healthcare/default.html. Accessed May 15, 2012.
7. Cotten CM, Taylor S, Stoll B, Goldberg RN, Hansen NI, S?nchez PJ, Ambalavanan N, Benjamin DK Jr; NICHD Neonatal Research Network. Prolonged duration of initial empirical antibiotic treatment is associated with increased rates of necrotizing enterocolitis and death for extremely low birth weight infants. Pediatrics. 2009;123(1):58-66.
8. Kuppala VS, Meinzen-Derr J, Morrow AL, Schibler KR. Prolonged initial empirical antibiotic treatment is associated with adverse outcomes in premature infants. J Pediatr. 2011;159(5):720-725.
9. Patel SJ, Saiman L. Antibiotic resistance in neonatal intensive care unit pathogens: mechanisms, clinical impact, and prevention including antibiotic stewardship. Clin Perinatol. 2010;37(3):547-563.
Conflict of Interest:
None declared