When to perform lumbar puncture in infants at risk for meningitis in the neonatal intensive care unit

Introduction

The neonatal period is the most common time in life for the presentation of bacterial meningitis, with an estimated incidence of approximately 0.3 per 1000 live births in developed countries ^, and 0.8 per 1000 life births in underdeveloped countries. However, this figure is likely an underestimate, as 30% to 70% of infants who undergo sepsis evaluation do not have a lumbar puncture (LP) performed. ^, In developed countries, mortality from meningitis ranges between 10% and 15% for term infants and reaches 25% in premature infants. In underdeveloped countries, mortality ranges from 40% to 58%. Although associated mortality has decreased over time, long-term morbidity remains high. ^, Identification of meningitis in infants by clinical examination can be difficult, as early signs are often subtle and nonspecific. Examination of the cerebrospinal fluid (CSF) is essential for diagnosis, identification of pathogens, and appropriate choice of therapy.

The role of the LP as part of the diagnostic evaluation of neonatal sepsis, especially in premature infants in the first 72 hours after birth, remains controversial. The incidence of meningitis among asymptomatic infants with antepartum risk factors for infection alone and premature infants with respiratory distress but no other signs of sepsis is extremely low. However, 20% to 30% of infants with culture-proven early-onset sepsis (EOS) have concomitant bacterial meningitis. ^, Furthermore, about 15% to 38% of infants with confirmed meningitis have negative blood cultures, ^{, ,} suggesting that negative blood cultures cannot exclude meningitis in a sick infant. Therefore, selective evaluation of infants with culture-proven bacteremia results in missed cases of meningitis. Clinicians must additionally account for the variability of the presenting signs of meningitis based on birth weight and gestational age. Initial signs of meningitis in the neonate may include temperature instability, lethargy, apnea, and bradycardia, which are nonspecific and often subtle; more classic meningitic signs such as fever, irritability, seizures, and bulging fontanelle are often seen later in the course of meningitis and in infants weighing >2500 g.

In infants under evaluation for early- and late-onset sepsis, an LP is not performed in approximately 30% to 70% of infants, respectively. ^, The most common reasons for deferring an LP include the low incidence of neonatal meningitis, especially in the first 72 hours after birth in infants at risk for sepsis ^, ; the low yield of the procedure ^{, , ,} ; the risk of complications ; and the fact that very-low-birth-weight (VLBW) infants (<1500 g birth weight) often have respiratory and cardiovascular compromise with the procedure. However, there are advantages to obtaining an LP promptly. Diagnosing meningitis has implications in management and prognosis, as the duration of therapy has to be increased, and antimicrobial agents with higher degrees of central nervous system (CNS) penetration should be used. ^, Initiation of empiric antimicrobial therapy before LP may result in CSF sterilization, leading to underdiagnosis or in unnecessarily prolonged treatment when the possibility of meningitis cannot be excluded. A delay or failure to diagnose meningitis is associated with inappropriate choice in spectrum and duration of antibiotic therapy, partially treated meningitis, increased mortality, and neurologic sequelae.

Increased susceptibility to meningitis in premature infants

Bacteria can enter the CNS by (1) “receptor-mediated” transcellular movement through meningeal endothelial cells, (2) disruption of intracellular junctions of the cerebral microvasculature, or (3) transport across the blood–brain barrier within leukocytes. ^, Infants, especially premature infants, have a number of host defense impairments that increase their vulnerability to serious bacterial, fungal, and viral infections of the CNS, and they lack protective maternal antibodies, which do not cross the placenta until after 32 weeks gestational age. ^, Studies in fetal and neonatal animals have demonstrated immaturity and increased permeability of the blood–brain barrier, which is reflected in the elevated CSF protein content of the premature infant. Premature infants are also more likely to have foreign devices such as endotracheal tubes, arterial and venous catheters, and intraventricular devices, placing them at increased risk of invasive infections.

Accumulating evidence links intrauterine and postnatal neonatal infections with adverse neurodevelopmental outcomes in premature infants. Exposure of the immature brain, and particularly the white matter, to inflammatory mediators causing cytotoxic injury is associated with increased risk for abnormal cognitive and motor functioning. VLBW infants with meningitis are more likely to have major neurologic disability (45% vs. 11%) and subnormal (<70) Mental Developmental Index results (38% vs. 14%) compared with uninfected infants.

Role of LP in the evaluation of early-onset sepsis

The Centers for Disease Control and Prevention defines EOS as blood or CSF culture-proven infection within the first week after birth. In hospitalized premature infants, EOS has also been defined as culture-proven infections occurring within the first 3 days after birth. ^, The incidence of EOS in the United States is estimated to be 0.8 to 1 per 1000 live births, with the highest incidence among the most premature infants. ^, Due to widespread implementation of intrapartum antibiotic prophylaxis, the incidence of EOS has declined among term infants; however, whether or not there is a similar decrease in the incidence for premature infants remains unclear. Group B Streptococcus (GBS) remains the most frequent isolated pathogen associated with EOS in the United States and other developed countries among term infants, and Escherichia coli has emerged as the most common cause in VLBW infants. ^{, ,} Among infants with early-onset GBS disease, the most commonly identified syndromes are bacteremia without focus (83%), pneumonia (9%), and meningitis (7%). Collection of CSF via LP is needed to rule out meningitis; however, its role in infants with suspected EOS remains controversial, and clinical practice varies greatly by center.

The yield of CSF cultures in asymptomatic infants who undergo evaluation due to perinatal risk factors for infection is extremely low in the first few days after birth. ^{, , ,} In a retrospective review of 3423 asymptomatic full-term infants evaluated because of maternal risk factors for infection in the first 7 days after birth, no cases of meningitis were observed. A prospective study of 712 asymptomatic infants who underwent LP during the first week after birth for suspected sepsis found nine positive CSF cultures (13%); however, only one infant had concomitant bacteremia and a clinical course consistent with meningitis, and the remaining cases were considered contaminants. Consistent with these findings, a review of 506 infants at risk for infection or with suspected sepsis found no cases of meningitis among 263 infants who underwent LP within the first 72 hours after birth. However, the incidence of positive CSF cultures increases with advancing postnatal age; after 7 days of age, the incidence of meningitis in infants evaluated for sepsis may be as high as 10%. Conversely, a retrospective review of 169,849 infants born in U.S. Army hospitals during a 5-year period identified 43 cases of meningitis in the first 72 hours after birth. Of these 43 cases, five occurred in premature infants with respiratory distress syndrome (RDS), eight were born at term with no specific CNS symptoms and negative blood cultures, and three had positive blood cultures but were asymptomatic. Those findings were worrisome because it suggested that meningitis can occur in infants with negative blood cultures who appeared completely well; however, the retrospective nature of that case series makes that observation uncertain. Taking the available evidence into account, it is likely safe to omit routine LPs in well-appearing infants evaluated because of maternal risk factors ( Table 7.1 ).

Similarly, the incidence of meningitis in premature infants with RDS is low. ^, In a retrospective review of 238 infants born between 23 and 40 weeks’ gestation admitted with RDS and evaluated for suspected sepsis within the first 24 hours after birth, no cases of meningitis were found among 203 CSF cultures collected. The authors suggested reserving LPs for infants with RDS with positive blood cultures or infants with other concomitant signs after 24 hours of age, such as hypothermia or hyperthermia, poor feeding, or specific CNS signs. Supporting a selective approach in premature infants with RDS, only four cases of proven early-onset meningitis were found among 1495 infants at 27 to 36 weeks’ gestation who were admitted with respiratory distress and evaluated for sepsis with an LP within the first 24 hours after birth. Another retrospective case review of 1390 premature infants who were ≤34 weeks’ gestation and evaluated for sepsis identified 32 infants with bacteremia in the first 6 hours after birth but no missed cases of meningitis (confirmed by autopsy in infants who died). Therefore, deferring or omitting LP from EOS evaluation in premature infants with RDS when the clinician believes that the infant’s signs correspond to a noninfectious condition is a reasonable approach (see Table 7.1 ).

Current recommendations by the American Academy of Pediatrics (AAP) Committee on Fetus and Newborn suggest performing an LP as part of the evaluation of EOS in (1) infants with bacteremia, (2) infants with signs of infection when clinically stable, (3) infants with suspected bacteremia based on laboratory findings, and (4) infants with clinical deterioration despite initial antibiotic treatment.