Pathogenesis of Intrauterine Infection

Intrauterine infection is a result of clinical or subclinical maternal infection with a variety of agents (cytomegalovirus [CMV], Treponema pallidum, Toxoplasma gondii, rubella virus, varicella virus, parvovirus B19) and hematogenous transplacental transmission to the fetus. Transplacental infection may occur at any time during gestation, and signs and symptoms may be present at birth or may be delayed for months or years (Fig. 103-1). Infection may result in early spontaneous abortion, congenital malformation, intrauterine growth restriction, premature birth, stillbirth, acute or delayed disease in the neonatal period, or asymptomatic persistent infection with sequelae later in life. In some cases, no apparent effects are seen in the newborn infant.

Figure 103-1 Pathogenesis of hematogenous transplacental infections.

(From Klein JO, Remington JS: Current concepts of infections of the fetus and newborn infant. In Remington JS, Klein JO, editors: Infectious diseases of the fetus and newborn infant, ed 5, Philadelphia, 2002, WB Saunders.)

The timing of infection during gestation affects the outcome. First-trimester infection may alter embryogenesis, with resulting congenital malformations (congenital rubella) (Chapter 239). Third-trimester infection often results in active infection at the time of delivery (toxoplasmosis, syphilis) (Chapters 282 and 210). Infections that occur late in gestation may lead to a delay in clinical manifestations until some time after birth (syphilis).

Maternal infection is a necessary prerequisite for transplacental infection. For some etiologic agents (rubella), maternal immunity is effective and antibody is protective for the fetus. For other agents (CMV), maternal antibody may ameliorate the outcome of infection or may have no effect (Chapter 247). Even without maternal antibody, transplacental transmission of infection to a fetus is variable because the placenta may function as an effective barrier.

Pathogenesis of Ascending Bacterial Infection

In most cases, the fetus or neonate is not exposed to potentially pathogenic bacteria until the membranes rupture and the infant passes through the birth canal and/or enters the extrauterine environment. The human birth canal is colonized with aerobic and anaerobic organisms that may result in ascending amniotic infection and/or colonization of the neonate at birth. Vertical transmission of bacterial agents that infect the amniotic fluid and/or vaginal canal may occur in utero or, more commonly, during labor and/or delivery (Fig. 103-2). Chorioamnionitis results from microbial invasion of amniotic fluid, often as a result of prolonged rupture of the chorioamniotic membrane. Amniotic infection may also occur with apparently intact membranes or with a relatively brief duration of membrane rupture. The term chorioamnionitis refers to the clinical syndrome of intrauterine infection, which includes maternal fever, with or without local or systemic signs of chorioamnionitis (uterine tenderness, foul-smelling vaginal discharge/amniotic fluid, maternal leukocytosis, maternal and/or fetal tachycardia). Chorioamnionitis may also be asymptomatic, diagnosed only by amniotic fluid analysis or pathologic examination of the placenta. The rate of histologic chorioamnionitis is inversely related to gestational age at birth (Fig. 103-3) and directly related to duration of membrane rupture. Rupture of membranes for longer than 24 hr was once considered prolonged because microscopic evidence of inflammation of the membranes is uniformly present when the duration of rupture exceeds 24 hr. At 18 hr of membrane rupture, however, the incidence of early-onset disease with group B streptococcus (GBS) increases significantly; 18 hr is the appropriate cutoff for increased risk of neonatal infection.

Figure 103-2 Pathways of ascending or intrapartum infection.

Figure 103-3 Histologic chorioamnionitis in liveborn preterm babies by gestational age (n = 3,928 babies).

(From Lahra MM, Jeffery HE: A fetal response to chorioamnionitis is associated with early survival after preterm birth, Am J Obstet Gynecol 190:147–151, 2004.)

Bacterial colonization does not always result in disease. Factors influencing which colonized infant will experience disease are not well understood but include prematurity, underlying illness, invasive procedures, inoculum size, virulence of the infecting organism, genetic predisposition, the innate immune system, host response, and transplacental maternal antibodies (Fig. 103-4). Aspiration or ingestion of bacteria in amniotic fluid may lead to congenital pneumonia or systemic infection, with manifestations becoming apparent before delivery (fetal distress, tachycardia), at delivery (failure to breathe, respiratory distress, shock), or after a latent period of a few hours (respiratory distress, shock). Aspiration or ingestion of bacteria during the birth process may lead to infection after an interval of 1-2 days.

Figure 103-4 Factors influencing the balance between health and disease in neonates exposed to a potential pathogen. ROM, rupture of membranes.

(Adapted from Baker CJ: Group B streptococcal infections, Clin Perinatol 24:59–70, 1997.)

Resuscitation at birth, particularly if it involves endotracheal intubation, insertion of an umbilical vessel catheter, or both, is associated with an increased risk of bacterial infection. Explanations include the presence of infection at the time of birth or acquisition of infection during the invasive procedures associated with resuscitation.

Pathogenesis of Late-Onset Postnatal Infections

After birth, neonates are exposed to infectious agents in the nursery or in the community. Postnatal infections may be transmitted by direct contact with hospital personnel, the mother, or other family members; from breast milk (HIV, CMV); or from inanimate sources such as contaminated equipment. The most common source of postnatal infections in hospitalized newborns is hand contamination of health care personnel.

Most cases of meningitis result from hematogenous dissemination. Less often, meningitis results from contiguous spread as a result of contamination of open neural tube defects, congenital sinus tracts, or penetrating wounds from fetal scalp sampling or internal fetal electrocardiographic monitors. Abscess formation, ventriculitis, septic infarcts, hydrocephalus, and subdural effusions are complications of meningitis that occur more often in newborn infants than in older children.

Immunoglobulins

Immunoglobulin (Ig) G is actively transported across the placenta, with concentrations in a full-term infant comparable to or higher than those in the mother. The specificity of IgG antibody in cord blood depends on the mother’s previous antigenic exposure and immunologic response. In premature infants, cord IgG levels are directly proportional to gestational age. Studies of type-specific IgG antibodies to GBS have shown that the ratio of cord to maternal serum concentrations is 1.0 at term, 0.5 at 32 wk of gestation, and 0.3 at 28 wk. Levels of maternally derived IgG fall rapidly after birth. Infants with birthweights <1,500 g become significantly hypogammaglobulinemic, with mean plasma IgG concentrations in the range of 200-300 mg/dL in the 1st wk of life. Other classes of immunoglobulins are not transferred across the placenta, although a fetus can synthesize IgA and IgM in response to intrauterine infection.

The presence of passively transferred specific IgG antibody in adequate concentration provides neonates protection against infections to which protection is mediated by that antibody (tetanus, encapsulated bacteria such as GBS). Specific bactericidal and opsonic antibodies against enteric gram-negative bacteria are predominantly in the IgM class. Newborn infants usually lack antibody-mediated protection against Escherichia coli and other Enterobacteriaceae.

Complement

The complement system mediates bactericidal activity against certain organisms such as E. coli and functions as an opsonin with antibody in the phagocytosis of bacteria such as GBS. No transplacental passage of complement from the maternal circulation takes place. A fetus begins to synthesize complement components as early as the 1st trimester. Full-term newborn infants have slightly diminished classical pathway complement activity and moderately diminished alternative pathway activity. Considerable variability, however, is seen in both the concentration and activity of complement components. Premature infants have lower levels of complement components and less complement activity than full-term newborns do. These deficiencies contribute to diminished complement-derived chemotactic activity and to a lesser ability to opsonize certain organisms in the absence of antibody. Opsonization of Staphylococcus aureus is normal in neonatal sera, but various degrees of impairment have been noted with GBS and E. coli.

Neutrophils

Quantitative and qualitative deficiencies of the phagocyte system contribute to the newborn’s susceptibility to infection. Neutrophil migration (chemotaxis) is abnormal at birth in both term and preterm infants. Neonatal neutrophils have decreases in adhesion, aggregation, and deformability, all of which may delay the response to infection. Abnormal expression of cell membrane adhesion molecules (the β₂ integrins and selectins) and abnormalities in the neonatal neutrophil cytoskeleton contribute to abnormal chemotaxis. With adequate opsonization, phagocytosis and killing by neutrophils are comparable in newborn infants and adults. In the presence of infectious or noninfectious stress (respiratory distress syndrome), however, the ability of neonatal neutrophils to phagocytose gram-negative (but not gram-positive) bacteria is decreased. Impairment of the oxidative respiratory burst of neonatal neutrophils is a factor in the increased risk of sepsis, especially in preterm infants.

The number of circulating neutrophils is elevated after birth in both term and preterm infants, with a peak at 12 hr that returns to normal by 22 hr. Band neutrophils constitute less than 15% in normal newborns and may increase in newborns with infection and other stress responses, such as asphyxia.

Neutropenia, which is frequently observed in preterm infants and infants with intrauterine growth restriction, increases the risk for sepsis. The neutrophil storage pool in newborn infants is 20-30% of that in adults and is more likely to be depleted in the face of infection. Mortality is increased when sepsis is associated with severe sepsis-induced neutropenia and bone marrow depletion. Granulocyte colony–stimulating factor (G-CSF) and granulocyte-macrophage CSF (GM-CSF) are cytokines that play important roles in the proliferation, differentiation, functional activation, and survival of phagocytes. These cytokines stimulate myeloid progenitor cells, increase the bone marrow neutrophil storage pool, induce peripheral blood neutrophilia, and influence neutrophil function, including enhancement of bactericidal activity. Although these myeloid colony-stimulating factors influence neutrophil number and function, their clinical utility in the treatment and/or prevention of neonatal sepsis remains undetermined.

Monocyte-Macrophage System

The monocyte-macrophage system consists of circulating monocytes and tissue macrophages, particularly in the liver, spleen, and lung. Activated macrophages are involved in antigen presentation, phagocytosis, and immune modulation. The number of circulating monocytes in neonatal blood is normal, but the mass or function of macrophages in the reticuloendothelial system is diminished, particularly in preterm infants. In both term and preterm infants, chemotaxis of monocytes is impaired; this impairment affects the inflammatory response in tissues and the results of delayed hypersensitivity skin tests. Monocytes from neonates ingest and kill microorganisms as well as monocytes from adults.

Natural Killer Cells

Natural killer (NK) cells are a subgroup of lymphocytes that are cytolytic against cells infected with viruses. NK cells also lyse cells coated with antibody in a process called antibody-dependent cell-mediated cytotoxicity (ADCC). NK cells appear early in gestation and are present in cord blood in numbers equivalent to those in adults; neonatal NK cells have decreased cytotoxic activity and ADCC in comparison with adult cells. The diminished cytotoxicity against herpes simplex virus (HSV)–infected cells may predispose to disseminated HSV infection in newborns (Chapter 249).

Cytokines/Inflammatory Mediators

The patient’s response to infection and clinical outcome involves a balance between pro-inflammatory and anti-inflammatory cytokines. Several adverse neonatal outcomes, including brain injury, necrotizing enterocolitis (NEC), and bronchopulmonary dysplasia, may be mediated by the cytokine response to infection in the mother, fetus, or newborn. The mediators that have been studied in newborns include tumor necrosis factor-α (TNFα), interleukin 1 (IL-1), IL-4, IL-6, IL-8, IL-10, IL-12, platelet-activating factor, and the leukotrienes. The release of various inflammatory mediators in response to infection offers the potential opportunity to facilitate an early laboratory diagnosis of infection. Potential surrogate markers for bacterial sepsis, pneumonia, and NEC include TNFα, IL-6, and IL-8.

Innate immunity involves nonspecific cellular and humoral responses to an infectious agent without previous exposure. Recognition of pathogens is initiated by soluble components in plasma (including mannose-binding lectin) and by recognition of receptors on monocytes and other cells. Toll-like receptors play an important role in pathogen recognition. Genetic polymorphisms (mutations) of various proteins involved in the immune response may increase the risk and severity of neonatal infections. The neutrophil is another important cellular component of innate immunity. Neutrophil granules contain many enzymes; one protein, bactericidal/permeability-increasing protein (BPI), binds to the endotoxin in the cell wall of gram-negative bacteria. This protein facilitates opsonization and prevents the inflammatory response to endotoxin. BPI activity may be decreased in neonates.

Prematurity

The most important neonatal factor predisposing to infection is prematurity or LBW. Preterm LBW infants have a 3- to 10-fold higher incidence of infection than full-term normal birthweight infants. Possible explanations are as follows: (1) maternal genital tract infection is considered to be an important cause of preterm labor, with an increased risk of vertical transmission to the newborn (Figs. 103-5 and 103-6); (2) the frequency of intra-amniotic infection is inversely related to gestational age (see Fig. 103-3); (3) premature infants have documented immune dysfunction; and (4) premature infants often require prolonged intravenous access, endotracheal intubation, or other invasive procedures that provide a portal of entry or impair barrier and clearance mechanisms.

Figure 103-5 Potential pathways from choriodecidual bacterial colonization to preterm delivery.

(From Goldenberg RL, Hauth JA, Andrews WW: Intrauterine infection and preterm delivery, N Engl J Med 342:1500–1507, 2000. Copyright 2000, Massachusetts Medical Society.)

Figure 103-6 Potential sites of bacterial infection within the uterus.

(From Goldenberg RL, Hauth JA, Andrews WW: Intrauterine infection and preterm delivery, N Engl J Med 342:1500–1507, 2000. Copyright 2000, Massachusetts Medical Society.)

Nosocomial Infections

Nosocomial (hospital-acquired) infections are responsible for significant morbidity and late mortality in hospitalized newborns. Many experts define nosocomial infections in newborns as infections occurring after 3 days of life that are not directly acquired from the mother’s genital tract. For the purposes of surveillance in the acute care setting, the CDC National Healthcare Safety Network (NHSN) defines health care–associated infections in newborns as those that result from passage through the birth canal as well as infections that occur from exogenous sources such as health care personnel, visitors, and equipment/devices in the health care environment. This surveillance definition includes any infection occurring after admission to the neonatal intensive care unit (NICU) that was not transplacentally acquired. Rates of nosocomial infection in healthy term infants who are either rooming-in with their mothers or staying in the well baby nursery are low (<1%). The majority of nosocomial infections occur in preterm or term infants who require intensive care. Risk factors for nosocomial infection in these infants include prematurity, LBW, invasive procedures, indwelling vascular catheters, parenteral nutrition with lipid emulsions, endotracheal tubes, ventricular shunts, alterations in the skin and/or mucous membrane barriers, frequent use of broad-spectrum antibiotics, and prolonged hospital stay. The most frequent nosocomial infections are bloodstream infections associated with an intravascular catheter and pneumonia, especially ventilator-associated pneumonia. Nonetheless, nosocomial sepsis may occur in the absence of a catheter or ventilator. In addition, infants receiving intensive care in the NICU are at risk to acquire community or hospital associated infections during seasonal epidemics (rotavirus, RSV, influenza).

Almost one quarter of VLBW infants (<1,500 g BW) experience nosocomial infections. Rates of infections increase with decreasing gestational age and birthweight. The NICHD Neonatal Research Network has reported rates of 43% for infants 401-750 g; 28% for those 751-1,000 g; 15% for those 1,001-1,250 g; and 7% for those 1,251-1,500 g. The CDC NHSN monitors device-associated nosocomial infection rates. Rates are inversely related to birthweight, and in level III NICUs, they range from 3.7 infections per 1,000 central line days for infants < 750 g to 2.0 infections per 1,000 central line days for those weighing > 2,500 g. The widespread differences in practice regarding the inclusion of lumbar puncture (LP) in the diagnostic evaluation of an infant with suspected sepsis make it more difficult to determine rates of late-onset meningitis.

Various bacterial and fungal agents colonize hospitalized infants, health care workers, and visitors. Pathogenic agents can be transmitted by direct contact or indirectly via contaminated equipment, intravenous fluids, medications, blood products, or enteral feedings. Colonization of the infant’s skin, umbilicus, and respiratory or gastrointestinal tract with pathogenic agents often precedes the development of infection. Antibiotic use interferes with colonization by normal flora, thereby facilitating colonization with more virulent pathogens.

Coagulase-negative staphylococci are the most frequent neonatal nosocomial pathogens. In a cohort of 6,215 VLBW infants in the NICHD Neonatal Research Network, gram-positive agents were associated with 70%, gram-negative with 18%, and fungi with 12% of cases of late-onset sepsis (Table 103-6). Coagulase-negative staphylococcus, the single most common organism, was isolated in 48% of these infections. The emergence of nosocomial bacterial pathogens resistant to multiple antibiotics is a growing concern. Among NICU patients, methicillin-resistant S. aureus, vancomycin-resistant enterococci, and multidrug-resistant gram-negative pathogens are particularly alarming. Organisms responsible for all categories of neonatal sepsis and meningitis may change with time (Table 103-7).

Table 103-6 DISTRIBUTION OF PATHOGENS ASSOCIATED WITH THE 1ST EPISODE OF LATE-ONSET SEPSIS IN VERY LOW BIRTHWEIGHT INFANTS*

* National Institute of Child Health and Human Development Neonatal Research Network, September 1, 1998, through August 31, 2000.

^† Patients with dual infections and patients with presumed coagulase-negative staphylococci (CONS) contaminants excluded. According to the definitions in text, 276 (44%) CONS were definite infections and 353 (56%) were possible infections.

From Stoll BJ, Hansen N, Fanaroff AA, et al: Late-onset sepsis in very low birthweight neonates: the experience of the NICHD Neonatal Research Network, Pediatrics 110:285–291, 2002.

Viral organisms may also cause nosocomial infection in the NICU; they include RSV, varicella, influenza, rotavirus, and enteroviruses. For viral as well as bacterial infections, nursery outbreaks may occur in addition to individual cases. Hospital infection control policies are essential to prevent and/or contain nursery infection outbreaks.

The mean age at onset of the 1st episode of late-onset nosocomial sepsis is 2-3 wk, independent of the infecting pathogen. Nosocomial infections increase the risk of adverse outcomes, including prolonged hospitalization and mortality.

Active surveillance for nosocomial infection is essential in monitoring overall rates of infection, rates of infection with specific pathogens, and antibiotic susceptibility patterns and in identifying clusters of cases or true infectious outbreaks. Surveillance is based on the ongoing review of nursery infections and data from the microbiology laboratory; routine surveillance to detect colonization is not indicated. Culture results should indicate the bacterial isolate and the antimicrobial sensitivity pattern. Assessment of other microbial markers (biotype, serotype, DNA fingerprint) is helpful in epidemics. During epidemics, investigation of possible reservoirs of infection, modes of transmission, and risk factors is necessary. Identification of colonized infants and nursery personnel is also helpful.

Infections acquired by newborns after discharge from the nursery are usually community acquired. They have the same epidemiologic considerations as other community-acquired infections in infants and children, except for protection provided by maternal antibody.

Bacterial Sepsis

Neonates with bacterial sepsis may have either nonspecific signs and symptoms or focal signs of infection (Tables 103-9 and 103-10), including temperature instability, hypotension, poor perfusion with pallor and mottled skin, metabolic acidosis, tachycardia or bradycardia, apnea, respiratory distress, grunting, cyanosis, irritability, lethargy, seizures, feeding intolerance, abdominal distention, jaundice, petechiae, purpura, and bleeding. International criteria for bacterial sepsis are listed in Table 103-11. The initial manifestation may involve only limited symptomatology and only 1 system, such as apnea alone or tachypnea with retractions or tachycardia, or it may be an acute catastrophic manifestation with multiorgan dysfunction. Infants should be reevaluated over time to determine whether the symptoms have progressed from mild to severe. Later complications of sepsis include respiratory failure, pulmonary hypertension, cardiac failure, shock, renal failure, liver dysfunction, cerebral edema or thrombosis, adrenal hemorrhage and/or insufficiency, bone marrow dysfunction (neutropenia, thrombocytopenia, anemia), and disseminated intravascular coagulopathy (DIC).

Table 103-11 CLINICAL CRITERIA FOR THE DIAGNOSIS OF SEPSIS

IMCI, Integrated Management of Childhood Illness; WHO, World Health Organization.

* Any of the signs listed implies high suspicion for serious bacterial infection.

^† Each symptom or sign is associated with a score. The score indicates the probability of disease.

From Vergnano S, Sharland M, Kazembe P, et al: Neonatal sepsis: an international perspective, Arch Dis Child Fetal Neonatal Ed 90:F220–F224, 2005.

A variety of noninfectious conditions can occur together with neonatal infection or can make the diagnosis of infection more difficult. Respiratory distress syndrome (RDS) secondary to surfactant deficiency can coexist with bacterial pneumonia. Because bacterial sepsis can be rapidly progressive, the physician must be alert to the signs and symptoms of possible infection and must initiate diagnostic evaluation and empirical therapy in a timely manner. The differential diagnosis of many of the signs and symptoms that suggest infection is extensive; these noninfectious disorders must also be considered (Table 103-12).

Systemic Inflammatory Response Syndrome

The clinical manifestations of infection depend on the virulence of the infecting organism and the body’s inflammatory response. The term systemic inflammatory response syndrome (SIRS) is most frequently used to describe this unique process of infection and the subsequent systemic response (Chapter 64). In addition to infection, SIRS may result from trauma, hemorrhagic shock, other causes of ischemia, NEC, and pancreatitis.

Patients with SIRS have a spectrum of clinical symptoms that represent progressive stages of the pathologic process. In adults, SIRS is defined by the presence of 2 or more of the following: (1) fever or hypothermia, (2) tachycardia, (3) tachypnea, and (4) abnormal white blood cell (WBC) count or an increase in immature forms. In neonates and pediatric patients, SIRS manifests as temperature instability, respiratory dysfunction (altered gas exchange, hypoxemia, acute respiratory distress syndrome [ARDS]), cardiac dysfunction (tachycardia, delayed capillary refill, hypotension), and perfusion abnormalities (oliguria, metabolic acidosis) (Table 103-13). Increased vascular permeability results in capillary leak into peripheral tissues and the lungs, with resultant pulmonary and peripheral edema. DIC results in the more severely affected cases. The cascade of escalating tissue injury may lead to multisystem organ failure and death.

Suspected Intrauterine Infection

The acronym TORCH refers to toxoplasmosis, other agents (syphilis, varicella, parvovirus B19, many more), rubella, CMV, and HSV. Although the term may be helpful in remembering some of the etiologic agents of intrauterine infection, the TORCH battery of serologic tests has a poor diagnostic yield, and specific diagnostic studies should be selected for each etiologic agent under consideration. CMV and HSV require culture or polymerase chain reaction (PCR) methods; toxoplasmosis is diagnosed by serologic tests and PCR, whereas syphilis and rubella are diagnosed by serologic methods (Table 103-15).

In most cases of suspected fetal infection, concern is not raised until the pregnant woman has been ill for several weeks or, in retrospect, after delivery. At this time, the maternal immune response to the suspected pathogen may no longer reflect an acute infection; that is, the specific IgM response is no longer detectable and the IgG response has already reached a plateau. Many of the pathogen-specific IgM serologic assays require considerable skill to perform and tend to be less reliable than the more common IgG assays. As a result, IgM assay results can be either falsely negative or falsely positive.

Neonatal antibody titers are often difficult to interpret because (1) IgG is acquired from the mother by transplacental passage and (2) determination of neonatal IgM titers to specific pathogens is technically difficult to perform and is not universally available. IgM titers to specific pathogens have high specificity but only moderate sensitivity; they should not be used to preclude infection. Paired maternal and fetal-neonatal IgG titers showing higher newborn IgG levels or rising IgG titers during infancy may be used to diagnose some congenital infections (syphilis). Total cord blood IgM or IgA (neither is actively transported across the placenta to the fetus) and the presence of IgM–rheumatoid factor in neonatal serum are nonspecific tests for intrauterine infection.

If the likelihood of maternal infection with a known teratogenic agent is high, fetal ultrasound examination is recommended. If the examination demonstrates either a physical abnormality or delayed growth for gestational age, examination of a fetal blood sample may be warranted. Cordocentesis can provide a sufficient sample for both total and pathogen-specific IgM assays, for PCR, or for culture. The total IgM value is important because the normal fetal IgM level is <5 mg/dL. Any elevation in total IgM may indicate an underlying fetal infection. Specific IgM antibody tests are available for CMV, T. pallidum, parvovirus B19, and toxoplasmosis. IgM tests are useful only when the results are strongly positive. A negative pathogen-specific IgM finding does not rule out that pathogen as a cause of fetopathy.

If maternal serologic studies point to a specific pathogen, it is sometimes possible to detect the organism in amniotic fluid or fetal blood (culture, PCR). Amniocentesis can be performed and the fluid sent for analysis. The presence of CMV, Toxoplasma, or parvovirus in amniotic fluid indicates that the fetus is infected and at high risk, but it does not always mean that the fetus will have severe sequelae. In contrast, HSV and varicella-zoster virus (VZV) are rarely isolated from amniotic fluid samples. CMV, Toxoplasma, and parvovirus can also be identified from cordocentesis sampling.

Parvovirus does not grow in the cell cultures commonly available in the virology laboratory. An IgM response is not always detectable in women with primary infection. When fetal parvovirus infection is suspected, testing of fetal blood or amniotic fluid by PCR is recommended in addition to testing for a specific IgM response in the fetus. PCR may also be used for the diagnosis of toxoplasmosis, CMV, HSV, rubella, and syphilis.

Neonatal infections with CMV, Toxoplasma, rubella, HSV, and syphilis present a diagnostic dilemma because (1) their clinical features overlap and may initially be indistinguishable; (2) disease may be inapparent; (3) maternal infection is often asymptomatic; (4) special laboratory studies may be needed; and (5) appropriate treatment of toxoplasmosis, syphilis, cytomegalovirus, and HSV, which may reduce significant long-term morbidity, is predicated on an accurate diagnosis. Common shared features that should suggest the diagnosis of an intrauterine infection include intrauterine growth restriction, hematologic involvement (anemia, neutropenia, thrombocytopenia, petechiae, purpura), ocular signs (chorioretinitis, cataracts, keratoconjunctivitis, glaucoma, microphthalmos), CNS signs (microcephaly, aseptic meningitis, hydrocephaly, intracranial calcifications), other organ system involvement (pneumonia, myocarditis, nephritis, hepatitis with hepatosplenomegaly, jaundice), and nonimmune hydrops. Diagnostic studies in newborns with suspected chronic intrauterine infection should specifically test for each diagnostic consideration. Systemic infections with CMV, HSV, and enteroviruses frequently involve the liver; if these infections are suspected, liver function tests should be performed. Neonatal HSV CNS disease may be confirmed by viral culture or by PCR identification from CSF. Given that approximately 50% of infants infected with HSV do not have CNS disease, cultures of skin, eyes, and mouth should also be performed in all infants with suspected HSV disease.

Suspected Bacterial or Fungal Infections

Bacterial or fungal infection is diagnosed by isolating the etiologic agent from a normally sterile body site (blood, CSF, urine, joint fluid). Obtaining 2 blood culture specimens by venipuncture from different sites avoids confusion caused by skin contamination and increases the likelihood of bacterial detection. Samples should be obtained from an umbilical catheter only at the time of initial insertion. A peripheral venous sample should also be obtained when blood is drawn for culture from central venous catheters. Although blood cultures are usually the basis for a diagnosis of bacterial infection, the bacteremic phase of the illness may be missed by poor timing or inadequate blood sample size. Low-level bacteremia (<10 colony-forming units/mL) has been observed in some infants from birth to 2 mo of age with positive culture results. Automated blood culture systems (BACTEC, Becton Dickinson; BacT/Alert, Organon Teknika), which continuously monitor blood cultures by checking each bottle every few minutes, have led to earlier detection of bacterial growth. After positive signaling in the automated system, the specific pathogen is identified by biochemical tests. PCR technology is emerging for more rapid accurate identification of a number of viral and bacterial agents.

Documentation of a positive blood culture result is the first diagnostic criterion that must be met for sepsis (see Table 103-15). It is important to note, however, that some patients with bacterial infection may have negative blood culture results (“clinical infection”), and other approaches to identification of infection are needed. A variety of diagnostic markers of infection are being evaluated. Although the total WBC count and differential counts and the ratio of immature to total neutrophils have limitations in sensitivity and specificity, an immature-to-total neutrophil ratio of ≥0.2 suggests bacterial infection. Neutropenia is more common than neutrophilia in severe neonatal sepsis, but neutropenia also occurs in association with maternal hypertension, preeclampsia, and intrauterine growth restriction. Thrombocytopenia is a nonspecific indicator of infection. Tests to demonstrate an inflammatory response include determinations of C-reactive protein, procalcitonin, haptoglobin, fibrinogen, proteomic markers in amniotic fluid, inflammatory cytokines (including IL-6, IL-8, and TNFα), and cell surface markers. It is unclear which surrogate markers for infection are most helpful.

When the clinical findings suggest an acute infection and the site of infection is unclear, laboratory studies should be performed, including blood cultures, LP, urine examination, and a chest radiograph. Urine should be collected by catheterization or suprapubic aspiration; urine culture for bacteria can be omitted in suspected early-onset infections because hematogenous spread to the urinary tract is rare at this point. Examination of the buffy coat with Gram or methylene blue stain may demonstrate intracellular pathogens. Demonstration of bacteria and inflammatory cells in gram-stained gastric aspirates on the 1st day of life may reflect maternal amnionitis, which is a risk factor for early-onset infection. Stains of endotracheal secretions in infants with early-onset pneumonia may demonstrate intracellular bacteria, and cultures may reveal either pathogens or upper respiratory tract flora. Careful examination of the placenta can be helpful in the diagnosis of both chronic and acute intrauterine infections.

Diagnostic evaluation (including blood culture) is indicated for asymptomatic infants born to mothers with chorioamnionitis. The probability of neonatal infection correlates with the degree of prematurity and bacterial contamination of the amniotic fluid. Some experts recommend presumptive treatment with antibiotics. By contrast, all symptomatic infants should be treated with antibiotics after blood cultures are obtained. There is controversy over whether an LP is necessary for all term infants with suspected early-onset sepsis. If the blood culture result is positive or if the infant becomes symptomatic, LP should definitely be performed. If the mother has been treated with antibiotics for chorioamnionitis, the newborn’s blood culture result may be negative, and the clinician must rely on clinical observation and other laboratory tests (see Fig. 103-9).

The diagnosis of pneumonia in a neonate is usually presumptive; microbiologic proof of infection is generally lacking because lung tissue is not easily cultured. CDC definitions of pneumonia do not target newborns, particularly the high-risk group of VLBW infants on mechanical ventilation. Although some clinicians rely on the results of bacteriologic culture of material obtained from the trachea as “proof” of cause, interpretation of such cultures has many pitfalls. These cultures often reflect upper respiratory tract commensal organisms and may have no etiologic significance. Even cultures performed on material obtained by bronchoalveolar lavage in a neonate are unreliable because the small bronchoscopes used in neonates cannot be protected from contamination as they are introduced into the distal airways. Short of tissue obtained by lung biopsy, the only reliable bacteriologic cultures are those performed on specimens obtained from blood or pleural fluid. Unfortunately, blood culture results are usually negative, and sufficient pleural fluid for culture is rarely present.

Interpretation of fungal cultures is associated with the same problems as that of bacterial cultures. Cultures of respiratory secretions for U. urealyticum and other genital Mycoplasma species are of little value because normal neonates are often colonized with these agents as a result of contamination with secretions from the maternal genital tract. Cultures for respiratory viruses and C. trachomatis may be valuable; these organisms are never indigenous flora, and isolation of them therefore suggests an etiologic role.

Other tests of potential value in evaluating neonates with possible infectious pneumonitis are discussed under diagnosis of infections (Chapter 164). The differential diagnosis of pneumonitis in neonates is broad and includes RDS, meconium aspiration syndrome, persistent pulmonary hypertension, diaphragmatic hernia, transient tachypnea of the newborn, congenital heart disease, and BPD.

The diagnosis of meningitis is confirmed by examination of CSF and identification of a bacterium, virus, or fungus by culture, antigen, or the use of PCR. The importance of the LP as part of the diagnostic evaluation of the neonate with suspected sepsis has been the subject of debate; clinical practice varies. For term infants with suspected early-onset sepsis, many clinicians start with a blood culture and a complete blood count, because 70-85% of term neonates with bacterial meningitis have positive blood culture results. Examination and culture of CSF are undertaken in term infants with symptoms and/or bacteremia. Many clinicians defer the LP in severely ill infants with suspected early-onset infection because of the fear of respiratory and/or cardiovascular compromise. In these situations, blood culture should be performed and treatment initiated for presumed meningitis until LP can be safely performed.

Normal, uninfected infants from 0-4 wk of age may have the following elevated CSF findings: protein 84 ± 45 mg/dL, glucose 46 ± 10 mg/dL, and leukocyte count 11 ± 10 mm³ with the 90th percentile being 22. The proportion of polymorphonuclear leukocytes is 2.2 ± 3.8% with the 90th percentile being 6. Elevated CSF protein values and leukocyte counts and hypoglycorrhachia may develop in preterm infants after intraventricular hemorrhage. Many nonpyogenic congenital infections (toxoplasmosis, CMV, HSV, syphilis producing an aseptic meningitis) can also produce alterations in CSF protein value and leukocyte count.

Gram staining of CSF yields a positive result in most patients with bacterial meningitis. The leukocyte count is usually elevated, with a predominance of neutrophils (>70-90%); the number is often >1,000 but may be <100 in infants with neutropenia or early in the disease. Microorganisms are recovered from most patients who have not been pretreated with antibiotics. Bacteria have also been isolated from CSF that did not have an abnormal number of cells (<25) or an abnormal protein level (<200 mg/dL), thus underscoring the importance of performing a culture and Gram stain on all CSF specimens. Contamination of CSF by bacteremia after traumatic LP may occur rarely. Culture-negative meningitis may be seen with antibiotic pretreatment, a brain abscess, or infection with Mycobacterium hominis, U. urealyticum, Bacteroides fragilis, enterovirus, or HSV. Use of PCR has improved the ability to detect viruses in CSF. Head ultrasonography or, more often, CT with contrast enhancement may be helpful in diagnosing ventriculitis and brain abscess.

Prevention of Nosocomial Infection

Principles for the prevention of nosocomial infection include adherence to universal precautions with all patient contact, avoiding nursery crowding and limiting nurse-to-patient ratios, strict compliance with handwashing, meticulous neonatal skin care, minimizing the risk of catheter contamination, decreasing the number of venipunctures and heelsticks, reducing the duration of catheter and mechanical ventilation days, encouraging appropriate advancement of enteral feedings, providing education and feedback to nursery personnel, and ongoing monitoring and surveillance of nosocomial infection rates in the NICU (Table 103-18).

Most nosocomial infections in the NICU are bloodstream infections associated with intravascular catheters. Catheters used in neonates include peripheral intravenous catheters, umbilical catheters, peripherally inserted central catheters, and surgically placed central venous catheters (CVCs). Efforts to reduce catheter-related infections include proper antisepsis of the skin before insertion of the catheter, sterile precautions during catheter insertion, aseptic technique when entering the line, minimizing repeated entry into the line for blood sampling, sterile preparation of fluids to be used with a CVC, and, finally, minimizing the number of catheter days.

The skin is an important mechanical barrier to infection. VLBW infants are born with an ineffective epidermal barrier that results in increases in transepidermal water loss and risk for infection. Efforts to reduce traumatic injury to this immature skin are important, including a reduction in the number of heelsticks.

Handwashing remains the most important and effective means of reducing nosocomial infections. Several expert groups have established guidelines for effective handwashing. Antimicrobial soaps or alcohol-based preparations are recommended. Proper handwashing is essential before entry into the NICU and before each patient contact. Barriers to compliance with handwashing include overcrowding, excessive patient-to-nurse ratios, poorly located sinks and inadequate supplies, lack of easy-to-use alcohol-based products at the bedside, concern about skin irritation, and inadequate knowledge, including the mistaken belief that the use of gloves obviates the need for handwashing. Ongoing education of staff regarding practices that are likely to reduce nosocomial infections and active surveillance of infection rates are important components of nosocomial infection control.

Neonatal immunization, commonly used in the USA to protect the newborn against hepatitis B infection and in other countries to protect against TB, is a promising strategy for a number of postnatal infections and deserves further study. Neonatal immunization is particularly attractive because the infant’s birth hospitalization is the most reliable point of health care contact.

Oral administration of bovine lactoferrin either alone or with a probiotic (Lactobacillus rhamnosus GG) for 30 days in LBW infants has been demonstrated to reduce the incidence of late-onset bacterial and fungal sepsis; probiotics have also reduced the incidence of NEC (Chapter 96.2).