Clinical Approach to the Neonate With Suspected Infection

Early-Onset Sepsis in Late Preterm and Term Infants

Because most babies are born at or near term, the majority of neonatal sepsis cases occur in this population, despite the current low incidence of disease (0.17–0.90 cases per 1000 live births ). Although this reflects a substantial reduction in the attack rate for early-onset group B streptococcal (GBS) sepsis since adoption of intrapartum prophylaxis strategies, GBS remains the leading cause of EOS among late preterm and term infants (see Fig. 90.1A ). Although ampicillin resistance is becoming more prevalent (38%) among E. coli isolates from term infants with EOS, most remain susceptible to gentamicin, and resistance to both is unusual (<8%); ampicillin and gentamicin remain appropriate for initial antibiotic treatment. Risk of EOS is increased with earlier gestational age at birth, higher maximum intrapartum maternal temperature, prolonged rupture of membranes, and evidence of maternal GBS colonization (positive antepartum screening culture, GBS bacteriuria during current pregnancy, or a prior infant with invasive GBS disease).

Recognition of these associations enabled development of strategies for prevention of EOS (GBS, in particular). The expectation that these risk indicators might also be useful for postnatal identification of infants at high risk for EOS has not been borne out, however, likely because of effectiveness of the preventive measures based on those findings. For example, nearly three-quarters of infants with GBS EOS now are born to women with negative antepartum screening cultures. As the incidence of EOS declined, without a concomitant decline (likely an increase) in empiric postpartum treatment, concerns about high rates of unnecessary treatment arose. Among well-appearing infants ≥35 weeks’ gestation, consensus guidelines based on this approach resulted in treatment of nearly 200 infants for each one with confirmed bacteremia. Recommended treatment of all infants born to women with chorioamnionitis is addressed in greater detail in Chapter 91 . The inefficiency of these strategies led to reappraisals of their underlying rationale, and reports of alternative approaches began to appear.

This changing landscape was addressed in a Clinical Report from the American Academy of Pediatrics Committee on Fetus and Newborn in 2018, which endorsed three options for managing ascertainment of EOS ( Fig. 90.2 ). The first is the categorical paradigm previously recommended by Centers for Disease Control and Prevention (CDC) and American Academy of Pediatrics (AAP) guidelines, in which fulfillment of any single criterion leads to recommendation of diagnostic evaluation and empiric treatment. This legacy option allows that approach to remain within the standard of care but recognizes that it is fraught with inefficiency, uncertain efficacy in an era of widespread intrapartum antibiotic prophylaxis, and substantial challenges of implementation. Similar categorical guidelines, incorporating both single findings that mandate treatment and others for which two or more are required to achieve the treatment threshold, are in use in the UK.

The second option is use of individualized quantitative risk estimates that integrate multiple risk indicators for each newborn infant, as exemplified by the Neonatal Early-Onset Sepsis Calculator offered by the Kaiser Permanente Division of Research (see Fig. 90.2B ). In this paradigm, traditional intrapartum indicators of EOS risk (gestational age, maternal temperature, duration of ruptured membranes, GBS colonization status, and nature and duration of antibiotic prophylaxis) form the basis for an initial estimate of absolute risk, which is then adjusted based on findings from physical examination. Although these absolute risk estimates are inaccurate (partly by design) and have not been externally validated, they provide an index for risk stratification and decision making. Adoption of this paradigm into clinical practice has been associated with substantial reductions in diagnostic evaluations and empiric antibiotic use. Within the Northern California Kaiser system, for example, the proportion of babies subjected to diagnostic testing decreased from 14.5% to 4.9% and antibiotic administration in the first 24 hours after birth decreased from 5% to 2.6%.

The third option relies on recognition of clinical signs of illness in newborn infants (see Fig. 90.2C ). Building on the observation that there were no cases of EOS in 46,000 asymptomatic term babies across three birth cohorts, ^{, ,} clinical investigators in Italy, the US, and Norway have reported experience with reliance on serial examinations, reserving empiric treatment for infants who show signs of illness. These experiences differ in specific protocols for observation and in criteria for selection of infants for close observation, ranging from selective use in infants with intrapartum risk factors (Italy ) to universal application in all newborn infants (US ). Typically, screening consists of focused examination (general well-being, skin color and perfusion, retractions, grunting) by nursing staff at 2- to 4-hour intervals for the first 12–48 hours after birth, with more comprehensive evaluation by a pediatrician or neonatologist if signs develop. The optimal constellation of examination elements remains to be determined, and criteria for initiation of diagnostic testing and empiric treatment are not clearly delineated. Typical criteria are provided in Table 90.1 . The required duration of observation also is uncertain, but almost all infants with EOS who develop clinical signs do so by 36 hours of age ( Fig. 90.3 ). In two reports of such strategies, there were substantial reductions in both diagnostic testing (from 11.6% to 1.6% of infants and from 15.3% to 6.3% of infants, respectively) and empiric treatment (from 12.8% to 0.6% and 11.1% to 4.1%, respectively). ^, Antibiotic exposure among term infants (≥37 weeks at birth) was similarly reduced in a hospital in Norway, from an already relatively low rate of 2.9% to 1.3%.

In addition to these measured effects, these less aggressive intervention strategies reduce separation of babies from their mothers, facilitate breastfeeding and parent-infant bonding, and reduce parental anxiety. These changes were achieved without evident increase in delayed treatment of infants with bacteremia, mortality attributable to infection, or readmission to the hospital for unrecognized infection.

At each delivery facility, strategy selection will be determined by available capabilities and resources, informed by relative costs and benefits of each approach ( Table 90.2 ). Because no stratagem ascertains all infants who are or will become infected at the initial evaluation immediately after birth, birthing facilities must have an organized process for recognizing the large proportion of infants (50%–60%) with EOS who fall ill thereafter. Notably, these infants were successfully identified in both the risk calculator and serial examination strategies, highlighting both the necessity for ongoing surveillance and operational convergence of these approaches.