Neonatal bacterial sepsis: Clinical features and diagnosis in neonates born at less than 35 weeks gestation

INTRODUCTION — 

Bacterial sepsis remains a major cause of mortality and morbidity in preterm and very low birth weight (VLBW) neonates.

The epidemiology, clinical features, and diagnosis of bacterial sepsis in neonates who are born at <35 weeks gestation will be reviewed here. The management, prevention, and outcome of bacterial sepsis in neonates who are born at <35 weeks gestation are discussed separately. (See "Neonatal bacterial sepsis: Treatment, prevention, and outcome in neonates born at less than 35 weeks gestation".)

Other related topics include:

●Sepsis in neonates ≥35 weeks gestation – (See "Neonatal bacterial sepsis: Clinical features and diagnosis in neonates born at or after 35 weeks gestation" and "Neonatal bacterial sepsis: Treatment, prevention, and outcome in neonates born at or after 35 weeks gestation".)

●Initial management of well-appearing newborns at risk for early-onset sepsis – (See "Approach to risk assessment and initial management of newborns with risk factors for early-onset sepsis".)

●Prevention and management of group B streptococcal (GBS) infection in neonates and young infants – (See "Prevention of early-onset group B streptococcal disease in neonates" and "Group B streptococcal (GBS) infection in neonates and young infants".)

●Outpatient evaluation and management of febrile neonates – (See "The febrile neonate (28 days of age or younger): Initial management" and "The febrile neonate (28 days of age or younger): Outpatient evaluation".)

●Evaluation and management of ill-appearing young infants – (See "Approach to the ill-appearing infant (younger than 90 days of age)".)

TERMINOLOGY — 

The following terms will be used throughout this topic:

●Preterm neonates – In general, preterm neonates are defined as those who are born at <34 weeks gestation. Late preterm (also called near-term) neonates are those who are born between 34 and 36 completed weeks of gestation. Gestational age can be derived from a calculator (calculator 1). (See "Preterm birth: Definitions of prematurity, epidemiology, and risk factors for infant mortality" and "Late preterm infants".)

When discussing sepsis in preterm neonates in this topic, we refer to neonates who are born at <35 weeks gestation. This is consistent with the age threshold used to guide the approach endorsed in the American Academy of Pediatrics (AAP) guidelines on neonatal sepsis [1].

Sepsis in late preterm and term neonates ≥35 weeks gestation is discussed elsewhere. (See "Neonatal bacterial sepsis: Clinical features and diagnosis in neonates born at or after 35 weeks gestation".)

●Very low birth weight (VLBW) neonates – This includes neonates with birth weights <1500 g.

●Neonatal sepsis – In neonates (infants 28 days of age or younger), sepsis is defined as isolation of a pathogenic bacterium from a blood culture. This is in contrast to the definition of sepsis in other populations, which refers to a dysregulated host response and organ dysfunction in the setting of infection but does not require bacteremia.

Sepsis is classified according to the neonate's age at the onset of symptoms:

•Early-onset sepsis (EOS) – This is defined as sepsis that occurs within the first 72 hours after birth, although some experts extend the definition to infections occurring within the first seven days after birth [1,2].

•Late-onset sepsis (LOS) – This is defined as sepsis that occurs at ≥72 hours after birth. Similar to early-onset sepsis, there is variability in its definition, ranging from onset at ≥72 hours of life to ≥7 days of age [2].

Preterm neonates with sepsis typically present with symptoms during their birth hospitalization. The approach to the evaluation and initial management of febrile neonates in the outpatient setting is discussed separately. (See "The febrile neonate (28 days of age or younger): Outpatient evaluation" and "The febrile neonate (28 days of age or younger): Initial management".)

●Healthcare-associated infections – These are defined as infections (eg, sepsis) acquired in the hospital while receiving treatment for other conditions [3].

INCIDENCE — 

The risk of sepsis increases with decreasing gestational age and birth weight [4-8]. The incidence of early-onset sepsis (EOS) in preterm and very low birthweight (VLBW) neonates has remained stable since the 1990s; however, the incidence of late-onset sepsis (LOS) has decreased since the early 2000s among preterm neonates in each gestational age category [7].

In the United States, estimated rates of culture-proven sepsis among preterm and VLBW neonates are as follows:

●EOS [7,9,10]:

•According to gestational age:

-Gestational age ≤25 weeks – 3.1 percent

-Gestational age 26 to 29 weeks – 1.2 percent

-Gestational age >29 weeks – 0.5 percent

•According to birth weight (grams):

-Birth weight <400 to 500 g – 2 percent

-Birth weight 501 to 750 g – 2.5 percent

-Birth weight 751 to 1000 g – 1.7 percent

-Birth weight 1001 to 1250 g – 1 percent

-Birth weight 1251 to 1500 g – 0.7 percent

●LOS [7,8]:

•According to gestational age:

-Gestational age ≤23 weeks – 32 percent

-Gestational age 24 to 25 weeks – 22 percent

-Gestational age 26 to 27 weeks – 11 percent

-Gestational age 28 to 29 weeks – 5 percent

-Gestational age >29 weeks – 2 percent

•According to birth weight (grams):

-Birth weight <400 to 500 g – 43 percent

-Birth weight 501 to 750 g – 43 percent

-Birth weight 751 to 1000 g – 28 percent

-Birth weight 1001 to 1250 g – 15 percent

-Birth weight 1251 to 1500 g – 7 percent

These incidence rates are based on data from two prospective registries in the United States (the National Institute of Child Health and Human Development Neonatal Research Network and the Vermont Oxford Network). Registries of preterm neonates born in other developed countries (including Canada, England, Germany, Israel, Australia, and New Zealand) have yielded similar estimates [5,6,11-14].

MICROBIOLOGY AND PATHOGENESIS

Early-onset sepsis — Early-onset sepsis (EOS) is usually due to vertical transmission; thus, pathogens from the maternal vaginal and fecal flora are the most common causes of infection.

●Route of infection – Early-onset bacterial infections are usually due to [1]:

•Vertical transmission by ascending contaminated amniotic fluid

•Vertical transmission during vaginal delivery from bacteria colonizing or infecting the mother's lower genital tract

●Causative pathogens – Escherichia coli and group B streptococcus (GBS) are the most common causes of neonatal EOS (table 1) [5,6,9].

In very preterm neonates, E. coli is more common as an EOS pathogen compared with GBS [5,6,9,15]. This was illustrated in a study from the National Institute of Child Health and Human Development (NICHD) registry, in which the incidences of EOS due to E. coli and GBS in very low birth weight (VLBW) neonates were 8.1 and 1.6 per 1000 live births, respectively [9]. For larger preterm neonates with birth weights between 1500 and 2500 g, the incidences were 0.7 and 0.3 per 1000 live births, respectively [9].

While coagulase-negative staphylococci (CoNS) is a common pathogen in late-onset sepsis (LOS), the importance of CoNS as a pathogen in EOS in preterm neonates is unclear. This may be due to differences in how EOS due to CoNS is defined in studies. For example, in the NICHD study, CoNS was only considered a cause of EOS if it was isolated in ≥2 separate cultures and the neonate received ≥5 days of antibiotic therapy; with this definition, it accounted for <1 percent of EOS cases [9]. However, in a study from the Canadian Neonatal Network (CNN), any positive culture was considered to be a case of EOS, and CoNS accounted for 11 percent of cases [5].

Determination of the relevance of CoNS in blood cultures in the setting of suspected preterm sepsis is discussed elsewhere. (See 'CoNS isolated' below.)  

Late-onset sepsis — Late-onset sepsis (LOS) can be acquired by vertical transmission and other sources; thus, it is caused by a wider spectrum of bacterial pathogens than EOS.

●Route of infection – Late-onset infections can be acquired by the two following mechanisms:

•Horizontal transmission from direct contact with care providers or environmental sources (ie, healthcare-associated infections).

•Maternal vertical transmission, which results in initial neonatal colonization that later evolves into infection.

●Causative pathogens – In preterm neonates, the most common pathogens associated with LOS are CoNS, Staphylococcus aureus, and gram-negative bacteria (E. coli, Klebsiella, Enterobacter) (table 1).

In a prospective observational study that included 10,501 very preterm neonates with LOS who were born between 2018 and 2020, the relative frequencies of different bacterial pathogens were [8]:

•CoNS (29 percent)

•S. aureus (23 percent)

•E. coli (12 percent)

•Klebsiella (8 percent)

•Enterococcus (5 percent)

•GBS (5 percent)

•Enterobacter (4 percent)

•Pseudomonas (3 percent)

•Serratia (2 percent)

Similar findings were noted in a report based on data from the national neonatal infection surveillance system in Germany [16].

●Other infections – Viral and fungal infections can present similarly to bacterial LOS. (See 'Differential diagnosis' below.)

RISK FACTORS

Risk factors associated with prematurity — Preterm neonates are inherently at increased risk for developing sepsis compared with term neonates due to [3,14,17-21]:

●Immunocompromised status – Preterm neonates have low levels of circulating maternal immunoglobulin G (IgG) because of the decreased opportunity for transplacental transfer that occurs during the third trimester of pregnancy. Even in the presence of adequate IgG concentrations, opsonization and complement functions are reduced in preterm neonates.

●Immature epithelial mucosal barrier – The epithelial barrier in preterm neonates is immature. Thus, the skin and mucosa in these neonates are thin, break down readily, and provide minimal protection.

Risk factors associated with interventions — Preterm neonates are also at increased risk for healthcare-associated (ie, nosocomial) infections due to the use of therapeutic interventions. These include [3,14,17-20,22-25]:

●Invasive devices – Disruption of the intact skin or mucosa due to invasive procedures is more likely to occur in preterm neonates [3,22] and increases the risk of infection, particularly of late-onset sepsis [14,17-19]. This includes the use of invasive devices such as central venous catheters (CVCs), arterial catheters, urinary catheters, and endotracheal tubes. Use of these devices increases with decreasing gestational age, thereby compounding the risk of infection in the most preterm neonates [19]. The longer these devices remain in situ, the higher the risk of infection [17]. In addition, CVCs can increase the risk of coagulase-negative staphylococci (CoNS) infection [23]. (See 'CoNS isolated' below.)  

The type of CVC (ie, peripherally inserted central catheter [PICC] versus umbilical venous catheter [UVC]) does not appear to impact the risk of infection. In a retrospective matched cohort study of 540 preterm neonates (<30 weeks gestation), the rates of catheter-associated bloodstream infection were similar among neonates who received a PICC on the first day after birth (day 1), those who received a UVC on day 1, and those who received a UVC on day 1 that was then changed for a PICC after four days or more (9.3, 7.8, and 8.2 per 1000 catheter days, respectively) [24].

●Parenteral nutrition – Parenteral nutrition administered through a CVC is associated with an increased risk of sepsis [3,14,22]. The administration of lipids may also be an independent risk factor for bacterial and fungal sepsis [3].

●Gastric acid blockade – Use of histamine-2 receptor blockers or proton pump inhibitors is associated with an increased risk of sepsis [20].

●Broad-spectrum antibiotics – Frequent exposure to broad-spectrum antibiotics is associated with an increased risk of fungal and multidrug resistant bacterial infections [3,14].  

Maternal risk factors for intraamniotic infection — Preterm neonates are at increased risk of sepsis due to intraamniotic infection (IAI). Maternal factors that are associated with IAI include:

●Clinical chorioamnionitis.

●Spontaneous preterm labor.

●Preterm birth due to cervical insufficiency or premature rupture of membranes (PROM). IAI may be a cause or a complication of cervical dilatation and PROM [1].

The duration of ruptured membranes is commonly considered when assessing the neonate's risk of infection; however, it is unclear whether prolonged premature rupture of membranes (ie, 18 or more hours) is an independent risk factor for sepsis in preterm neonates in the absence of intraamniotic infection [26,27].

●Maternal group B streptococcus (GBS) colonization – Maternal GBS bacteriuria during the current pregnancy, prior delivery of a neonate with GBS disease, and maternal colonization are some risk factors for early-onset GBS sepsis.

Maternal GBS screening and intrapartum antibiotic prophylaxis (IAP) reduce but do not eliminate the risk of GBS infection in the neonate. Risk factors for and the prevention of GBS infection are discussed separately. (See "Group B streptococcal (GBS) infection in neonates and young infants", section on 'Risk factors' and "Prevention of early-onset group B streptococcal disease in neonates".)

These risk factors are particularly important to consider in the evaluation of EOS. (See 'Early-onset (<72 hours)' below.)

CLINICAL MANIFESTATIONS

Spectrum of findings — In preterm neonates, the spectrum of sepsis symptoms ranges from nonspecific subtle findings (eg, mild increase in the frequency of apnea) to fulminant septic shock.

The underlying etiologic agent may influence the clinical presentation:

●Gram-negative sepsis can be associated with a fulminant course of severe sepsis and/or septic shock with a high risk of mortality [28].

●Sepsis due to coagulase-negative staphylococci (CoNS) infection tends to be more subtle. This was illustrated in a retrospective review of preterm neonates with culture-proven CoNS bacteremia in which most neonates on the first day of sepsis had new-onset apnea or bradycardia or new or increased respiratory support needs [29]. In this study, the most frequent presenting signs that prompted evaluation included hypoxia, apnea, bradycardia, lethargy, and increased gastric residuals [29].

General signs and symptoms — Nonspecific signs frequently observed in preterm neonates with sepsis include [9,30,31]:

●Respiratory distress that ranges from mild tachypnea to respiratory failure

●Increased respiratory support requirement

●Lethargy or hypotonia

●Change in frequency or severity of apnea (as compared with baseline)

●Feeding intolerance

●Temperature instability

●Increase in heart rate

Severe sepsis and septic shock — Sepsis is considered severe when it is associated with cardiovascular or pulmonary compromise or multiorgan failure (ie, dysfunction in two or more organ systems).

The clinical findings of neonatal shock include (see "Neonatal shock: Etiology, clinical manifestations, and evaluation", section on 'Distributive shock'):

●Cool extremities, acrocyanosis, and pallor

●Changes in heart rate (initially tachycardia and, in the later stages, which may be terminal, bradycardia)

●Neurologic changes such as lethargy, irritability, and nonresponsiveness

●Hypotension or evidence of poor perfusion

●Oliguria

Septic shock results in inadequate tissue perfusion and lactic acidosis. Shock in this setting is likely multifactorial, resulting from reduced systemic vascular resistance, reduced circulating intravascular volume due to capillary leak and third spacing, and, potentially, myocardial dysfunction. (See "Neonatal shock: Etiology, clinical manifestations, and evaluation", section on 'Distributive shock'.)

In the preterm neonate, evidence suggests that systemic inflammation triggered by sepsis can contribute to the development of other neonatal morbidities such as necrotizing enterocolitis, bronchopulmonary dysplasia, and intraventricular hemorrhage [32-36]. (See "Neonatal necrotizing enterocolitis: Pathology and pathogenesis" and "Germinal matrix and intraventricular hemorrhage (GMH-IVH) in the newborn: Risk factors, clinical features, screening, and diagnosis".)

APPROACH TO EVALUATION AND EMPIRIC MANAGEMENT

Early-onset (<72 hours)

Assessing the risk for early-onset sepsis — The first step in the evaluation of preterm neonates for early-onset sepsis (EOS) is to categorize the risk (high versus low) based on the circumstances leading to preterm birth (algorithm 1) [37-42]:

●Preterm birth due to circumstances associated with IAI – Neonates born preterm due to circumstances associated with maternal intraamniotic infection (IAI) are at the highest risk for EOS (see 'Approach to neonates at increased risk' below). This includes birth due to:

•Spontaneous preterm labor

•Cervical insufficiency

•Premature rupture of membranes (PROM)

•Unexplained nonreassuring fetal status

•Diagnosed intraamniotic infection

●Delivery for maternal or fetal factors without above features – Preterm birth may also be due to maternal or fetal factors (eg, maternal preeclampsia or other medical illness, fetal growth restriction, placental insufficiency). Although these circumstances are generally less concerning as risk factors for EOS, potential issues during labor may result in risk for maternal IAI and therefore increase the risk for infection in the neonate. Thus, risk assessment in this situation depends on whether labor was induced prior to delivery:

•Labor prior to delivery – For neonates born via vaginal or cesarean delivery after preterm labor, the risk for EOS depends on whether there were concerning features for IAI.

If IAI was suspected (eg, maternal fever or indication for intrapartum antibiotic prophylaxis (IAP) without receipt of adequate IAP prior to delivery), we consider these neonates to be at high risk for EOS (see 'Approach to neonates at increased risk' below). Indications for IAP are discussed in detail elsewhere. (See "Prevention of early-onset group B streptococcal disease in neonates", section on 'Candidates for intrapartum antibiotic prophylaxis'.)

Otherwise, these neonates are considered to be at lower risk for EOS. (See 'Approach to neonates at low risk' below.)

•No labor prior to delivery – Neonates born via cesarean delivery to mothers who did not labor prior to delivery and in whom rupture of membranes occurred at the time of delivery are at lowest risk for EOS. (See 'Approach to neonates at low risk' below.)

Risk stratification by delivery characteristics is supported by retrospective studies in preterm and very low birth weight neonates that have found a low incidence of sepsis in neonates born by cesarean delivery in the absence of labor or other risk factors for IAI [39-42]. As an example, in one study of over 14,000 extremely preterm (gestational age <28 weeks) neonates who survived >12 hours following delivery, EOS incidence was 0.5 percent (29 out of 5640) in those deemed low risk based on the above categorization compared with 2.5 percent (209 out of 8422) in those at higher risk (adjusted relative risk 0.24 [95% CI 0.16-0.36]) [40]). Low-risk neonates also had a lower combined risk of EOS or death within 12 hours of delivery. In addition, the implementation of EOS risk assessment decreased antibiotic initiation in low-risk neonates without increased risk of infection or death [41]. Decreasing antibiotic use is important since early and frequent use of broad-spectrum antibiotics have been associated with adverse outcomes in preterm neonates such as increased risk of mortality, fungal and multidrug resistant bacterial infections, necrotizing enterocolitis, and bronchopulmonary dysplasia [1,3,14,43].

The clinical condition of the newborn following initial delivery room stabilization may increase concern for EOS. However, cardiorespiratory instability and other clinical findings are common in preterm neonates during the transitional period after delivery and, given the low overall incidence of EOS, may be more likely due to other causes. Distinguishing between these findings and the clinical signs of EOS can be challenging in this population. Relying on clinical features alone might lead to an overestimation of the presumptive diagnosis of EOS and antibiotic overuse. (See 'Clinical manifestations' above and 'Differential diagnosis' below.)

Risk stratification schemas for EOS that have been validated and are routinely used in term neonates (eg, the EOS calculator) are not appropriate for assessing risk in preterm neonates, particularly for those who are born at <34 weeks gestation. The risk factors for EOS upon which these tools are based vary for preterm neonates. (See "Neonatal bacterial sepsis: Clinical features and diagnosis in neonates born at or after 35 weeks gestation", section on 'Early-onset sepsis calculator'.)

Approach to neonates at increased risk — Neonates at increased risk for EOS based on risk assessment (see 'Assessing the risk for early-onset sepsis' above) require laboratory evaluation for sepsis including blood and cerebrospinal fluid (CSF) cultures (unless lumbar puncture [LP] would compromise the neonate's condition) and initiation of empiric parenteral antibiotics pending culture results. Empiric antibiotics are initiated due to the risk of death and severe morbidity in neonatal sepsis. Empiric antibiotic selection is discussed in more detail elsewhere. (See "Neonatal bacterial sepsis: Treatment, prevention, and outcome in neonates born at less than 35 weeks gestation", section on 'Antibiotic therapy'.)

Blood cultures are the cornerstone of early-onset sepsis evaluation in preterm neonates (see 'Blood cultures' below). In addition, preterm neonates at increased risk for sepsis are also at risk for meningitis. Clinical signs of early meningitis can be subtle; thus, obtaining CSF studies in these high-risk neonates is important to confirm the presence or absence of infectious meningitis. However, in cases where it is not feasible to perform an LP because the neonate is too unstable to tolerate the procedure, LP should be deferred until the neonate is stable. LP should not delay empiric antibiotic initiation. (See 'Cerebrospinal fluid studies' below.)

Preterm neonates undergoing evaluation for EOS may also warrant investigation (eg, cell counts [CBC], viral testing, imaging, metabolic screening) for alternative diagnoses (eg, perinatally acquired viral infections such as cytomegalovirus [CMV], pneumonia, metabolic conditions), as discussed below. (See 'Other studies in selected cases' below and 'Testing for alternative diagnoses' below.)

We do not routinely perform urine or surface site cultures in the evaluation of suspected EOS. Isolated urinary tract infection is unlikely in neonates within the first 72 hours of age. If urine culture is performed, a positive result in this setting is a reflection of high-grade bacteremia rather than an isolated urinary tract infection. (See "Urinary tract infections in neonates".)

Approach to neonates at low risk — For neonates at low risk for EOS, we do not routinely check blood cultures and we do not start empiric antibiotics. Although some centers may obtain blood cultures in such patients without starting empiric antibiotics, this is not our practice. If EOS risk is high enough to obtain a blood culture, it is generally high enough to start empiric antibiotics. Support for this approach comes from data indicating a very low incidence of EOS in neonates categorized as low risk. (See 'Assessing the risk for early-onset sepsis' above.)

For such neonates, if there is significant clinical instability that does not improve after initial delivery room stabilization or there are signs of severe sepsis (eg, shock), it is generally appropriate to perform laboratory evaluation for sepsis and administer empiric parenteral antibiotics, as for neonates at increased risk. (See 'Approach to neonates at increased risk' above.)

Late-onset (≥72 hours) — Because the signs and symptoms of late-onset sepsis (LOS) are subtle and nonspecific, neonates ≥72 hours of age who have clinical features consistent with severe sepsis or shock, or who deviate significantly from the usual pattern of activity or feeding, require laboratory evaluation for sepsis and treatment with empiric parenteral antibiotics pending culture results. (See "Neonatal bacterial sepsis: Treatment, prevention, and outcome in neonates born at less than 35 weeks gestation", section on 'Empiric antibiotic therapy'.)

In these neonates, we obtain blood cultures, CSF studies, and a urine culture (via catheterization or bladder tap) prior to initiating antibiotics. In cases where it is not feasible to perform an LP because the neonate is too unstable to tolerate the procedure, LP should be deferred until the neonate is stable. LP should not delay empiric antibiotic initiation. (See 'Diagnostic tests' below.)

In addition, we obtain cultures from any other potential foci of infection (eg, purulent eye drainage, skin lesions or pustules, tracheal aspirates in mechanically ventilated neonates, central venous or other catheters, pleural or peritoneal fluid, bone or joint aspiration), cell counts (CBC), and other tests (eg, viral testing, imaging, metabolic screening) to evaluate for alternative diagnoses (eg, perinatally or postnatally acquired viral infections such as cytomegalovirus, herpes simplex virus, respiratory syncytial virus, pneumonia, metabolic conditions), as discussed below. LOS can be due to heterogeneous causes, including viral or other nonbacterial infections and interventions such as the use of invasive devices. (See 'Other studies in selected cases' below and 'Testing for alternative diagnoses' below.)

Novel monitoring techniques using algorithms to detect pathologic heart rate variability, respiratory instability, and/or decreased spontaneous movement in the neonate have been suggested as tools to help predict LOS in preterm neonates [44]. However, they lack diagnostic specificity and may result in an increased number of false positive results (eg, isolation of contaminant species). Further study is needed before this strategy is routinely used in clinical practice.

DIAGNOSTIC TESTS — 

Appropriate cultures and other microbiology studies usually distinguish neonatal bacterial sepsis from other types of infections or noninfectious conditions. (See 'Differential diagnosis' below.)

Ancillary studies may also be used in selected cases.

Blood cultures — The definitive diagnosis of neonatal sepsis is established with blood cultures, which can be obtained by venous or arterial puncture or by sampling from a newly inserted arterial or venous catheter. Properly obtained blood cultures are highly sensitive and specific for diagnosing neonatal sepsis. The sensitivity of blood cultures is dependent upon several factors:

●Timing – Blood cultures should be drawn before starting antibiotics. The available evidence suggests that exposure to intrapartum antibiotics does not reduce the sensitivity of blood cultures to detect early-onset sepsis (EOS) in the neonate [45,46].

●Volume of blood – A minimum volume of 1 mL of blood is required for blood cultures [1]. The sensitivity of blood culture to detect bacteremia depends upon using an adequate volume of blood to inoculate the culture bottle [47-49].

●Number of cultures – In neonates, obtaining a single blood culture is usually sufficient. Some centers obtain both aerobic and anaerobic cultures for all initial EOS and late-onset sepsis (LOS) evaluations. Anaerobic blood cultures may lead to earlier identification and can also identify pathogens not recovered by aerobic culture [50]. Obtaining more than one blood culture is also helpful in assessing for catheter associated infections and in the interpretation of blood culture results, particularly if coagulase-negative staphylococci (CoNS) is isolated. However, obtaining an adequate volume of blood for cultures can be difficult and only a single culture may be possible. (See 'CoNS isolated' below.)

Cerebrospinal fluid studies — If a lumbar puncture (LP) is performed, cerebrospinal fluid (CSF) should be sent for cell count with differential, protein, glucose, Gram stain, and culture. Molecular testing, if available, may assist in the diagnosis but is often unnecessary, as it does not provide information on antimicrobial susceptibility and culture is usually sufficient to identify the organism. Even when CSF cultures are not positive (eg, if empiric antibiotics were initiated prior to LP), CSF findings that suggest meningitis include a pleocytosis (CSF white blood cell [WBC] count >15/microL) with neutrophilic predominance, elevated CSF protein, and low CSF glucose. The interpretation of CSF parameters in neonates is discussed in greater detail separately. (See "Bacterial meningitis in the neonate: Clinical features and diagnosis", section on 'Interpretation of CSF parameters'.)

Other studies in selected cases

●CBC – Cell counts (CBC) can be used in combination with other tests, risk factors, and clinical assessment to help differentiate sepsis from other causes of illness. Both the absolute neutrophil count and the ratio of immature to total neutrophil counts (I/T ratio) have been used as markers for neonatal sepsis. However, as is true for term and late preterm neonates, these tests are not useful in isolation to accurately predict neonatal sepsis. (See "Neonatal bacterial sepsis: Clinical features and diagnosis in neonates born at or after 35 weeks gestation", section on 'Complete blood count'.)  

This was illustrated in a multicenter study of 166,092 both term and preterm neonates (mean gestational age of 34.6 weeks) who were suspected to have EOS [51]. Although the probability of a positive blood culture within the first three days of age increased with a low white blood cell count (<5000/microL), absolute neutropenia (<1000 neutrophils/microL), and an elevated I/T ratio, these indices had poor sensitivities and were insufficient to accurately diagnose neonatal sepsis.

In another analysis of the same cohort of neonates, LOS was associated with both low and high white blood cell counts (<1000 and >50,000/microL, respectively), high absolute neutrophil count (>17,670/microL), elevated I/T ratio (0.2 or higher), and low platelet count (<50,000/microL) [52]. However, the sensitivity of these indices was inadequate to reliably make the diagnosis of LOS.

●Limited role for molecular diagnostics – Molecular techniques (eg, polymerase chain reaction [PCR]) have the potential to offer a reliable and more timely method for diagnosis than bacterial cultures as they have a more rapid turnaround time and require smaller sample volumes [53,54]. While molecular methods are increasingly used to assist in the diagnosis of central nervous system (CNS) infections in neonates and children and are in development to aid in the detection of bacteremia, their role in the routine evaluation of suspected neonatal sepsis (ie, to detect bacteremia) is limited by cost and diagnostic accuracy, particularly as blood cultures are sufficiently sensitive to detect bacterial sepsis.

The role of PCR testing in neonatal meningitis is discussed separately. (See "Bacterial meningitis in the neonate: Clinical features and diagnosis", section on 'Molecular methods'.)

●Limited role for inflammatory markers – We do not use C-reactive protein (CRP) routinely in the diagnosis of sepsis. CRP is an acute phase reactant synthesized in the liver that increases in inflammatory conditions, including sepsis. Although an elevated CRP (>1 mg/dL) is 90 percent sensitive in detecting neonatal sepsis, it has poor specificity (eg, is elevated in other noninfectious inflammatory conditions such as maternal fever or fetal distress) and is thus a poor predictor for neonatal sepsis, particularly EOS [55,56]. However, CRP levels do not seem to be affected by gestational age [57] and may be useful as part of a comprehensive evaluation for LOS [58-62]. (See "Neonatal bacterial sepsis: Clinical features and diagnosis in neonates born at or after 35 weeks gestation", section on 'Other inflammatory markers'.)

Efforts have not been successful in identifying tests that can accurately and rapidly predict neonatal sepsis while awaiting blood culture results. In addition to neutrophil counts and CRP, procalcitonin is a widely studied test. Other tests that have been evaluated include interleukin-6, interleukin-8, and tumor necrosis factor-alpha levels; however, these studies are primarily used in the research setting and are not routinely available in hospital laboratories [62,63]. (See "Neonatal bacterial sepsis: Clinical features and diagnosis in neonates born at or after 35 weeks gestation", section on 'Laboratory tests'.)

Testing for alternative diagnoses — Additional studies are often obtained to differentiate sepsis from conditions with similar presentations. These include (table 2 and table 3) (see 'Differential diagnosis' below):

●Imaging of the chest, abdomen, or head to assess for other infectious, respiratory, and nervous system conditions (eg, pneumonia, respiratory distress syndrome, necrotizing enterocolitis, intraventricular hemorrhage, etc). (See "Neonatal pneumonia" and "Respiratory distress syndrome (RDS) in preterm neonates: Clinical features and diagnosis" and "Neonatal necrotizing enterocolitis: Clinical features and diagnosis" and "Germinal matrix and intraventricular hemorrhage (GMH-IVH) in the newborn: Risk factors, clinical features, screening, and diagnosis".)

●Testing for viral infection, including for respiratory syncytial virus, SARS-CoV-2, influenza, rhinovirus, enterovirus, cytomegalovirus, herpes simplex virus). (See "Respiratory syncytial virus infection: Clinical features and diagnosis in infants and children" and "COVID-19: Clinical manifestations and diagnosis in children", section on 'Approach to diagnosis' and "Congenital cytomegalovirus (cCMV) infection: Clinical features and diagnosis" and "Neonatal herpes simplex virus (HSV) infection: Clinical features and diagnosis".)

●Serum glucose to assess for hypoglycemia. (See "Pathogenesis, screening, and diagnosis of neonatal hypoglycemia", section on 'Clinical presentation'.)

●Metabolic screening to differentiate from inborn errors of metabolism. (See "Inborn errors of metabolism: Epidemiology, pathogenesis, and clinical features" and "Metabolic emergencies in suspected inborn errors of metabolism: Presentation, evaluation, and management".)

●Pulse oximetry screening and echocardiography to differentiate from critical congenital heart disease. (See "Evaluation of suspected critical congenital heart disease (CHD) in the newborn".)

ESTABLISHING THE DIAGNOSIS

Blood culture positive

Non-CoNS pathogen isolated — In preterm neonates whose blood culture is positive for a non-coagulase-negative staphylococci (CoNS) pathogen, the diagnosis of sepsis is established. For neonates with culture-proven sepsis, parenteral antibiotic therapy is tailored based on the isolated pathogen and its antimicrobial susceptibility pattern.

If blood cultures are positive, cerebrospinal fluid (CSF) studies should be obtained if not already performed. (See 'Diagnostic tests' above and "Neonatal bacterial sepsis: Treatment, prevention, and outcome in neonates born at less than 35 weeks gestation", section on 'Organism-specific therapy'.)

CoNS isolated — While CoNS is commonly isolated in culture in preterm neonates, it may or may not reflect true infection. In some cases, the positive culture may be due to contamination or colonization of a catheter hub (eg, when blood cultures are drawn from indwelling catheters). Obtaining two blood cultures can help to distinguish between true CoNS infection versus contaminant. We make the diagnosis of bacterial sepsis when CoNS are isolated from two or more blood cultures. In such cases, true bacteremia is more likely than contamination of the specimen, which may be reflected by a single positive blood culture [23].

Because of the difficulty in obtaining an adequate blood volume for culture in preterm neonates, the result from only a single positive blood culture may be available. In this setting, it is challenging to decide whether or not isolation of CoNS reflects true infection [3]. Repeating blood cultures (including collecting cultures from newly placed indwelling catheters) may be useful in determining true or ongoing infection. A single positive culture for CoNS with a negative repeat culture in the absence of indwelling catheters is likely a contaminant [23]. However, further management, including obtaining repeat blood cultures and the decision to continue antibiotic therapy, should also be based on clinical judgment and other risk factors for CoNS infection (eg, timing of blood culture collection with respect to length of antibiotic therapy, the susceptibility of isolated CoNS to empiric antibiotics, clinical response to empiric therapy, and clinical suspicion for infection based on gestational age and other risk factors [23]).  

The challenge of distinguishing true CoNS neonatal infection was illustrated in a retrospective study from a single tertiary center that identified 134 preterm neonates with 149 episodes of positive CoNS culture from blood or a usual sterile site [23]. Of these episodes, 27 percent were considered proven infection (defined as two positive cultures and ≥2 clinical signs of sepsis [temperature instability, cardiorespiratory or gastrointestinal disturbances, lethargy or irritability]), 37 percent were considered probable infection (based upon blinded review of clinical and laboratory data by two pediatric infectious disease specialists), and 36 percent were felt to reflect contamination. In this cohort, risk factors for proven and probable CoNS infection were similar to those for late-onset sepsis (LOS) due to other causes and included birth weight <2000 g, gestational age <34 weeks, and presence of a central venous catheter.  

Blood culture negative

Clinically improved or well appearing — Preterm neonates with mild and/or transient symptoms who have negative cultures at 36 to 48 hours are unlikely to have sepsis. Empiric antibiotic therapy should be discontinued after 36 to 48 hours in these neonates. Discontinuing unnecessary antibiotic therapy is important to reduce the risk of antibiotic resistance and other adverse outcomes [1].

Clinically unwell — Preterm neonates with ongoing concerning symptoms despite negative cultures should be evaluated for other possible explanations of their clinical findings while continuing empiric antibiotics (table 2). This includes evaluation for site-specific bacterial infections (eg, pneumonia, cellulitis, meningitis), nonbacterial infections, and noninfectious causes (eg, patent ductus arteriosus, respiratory distress syndrome). Viral infections (eg, enterovirus, parechovirus, postnatally acquired cytomegalovirus [CMV]) are a particularly common nonbacterial cause of suspected sepsis in neonates. (See 'Testing for alternative diagnoses' above and 'Differential diagnosis' below.)  

If a focal bacterial infection is identified, empiric antibiotic therapy should be tailored to treat the specific cause. If a nonbacterial cause is identified, antibiotics should be discontinued. In addition to routinely evaluating for causes of illness, clinicians should engage in efforts to limit the use of antibiotics in preterm neonates with negative cultures by avoiding sampling errors. (See 'Blood cultures' above.)  

Only a minority of neonates who undergo sepsis evaluation are ultimately diagnosed with proven sepsis. "Culture-negative sepsis" is a term that has been used to describe neonates who have sterile blood cultures yet have a clinical course that is concerning for sepsis. These neonates are often presumed to have sepsis and are treated with a full course of antibiotic therapy. However, this practice results in overdiagnosis of neonatal sepsis and overuse of antibiotics in preterm neonates [64-66]. This contributes to rising rates of drug-resistant bacteria in neonatal care units and it may have other adverse effects (eg, increased risk of necrotizing enterocolitis) [21,43]. (See "Neonatal necrotizing enterocolitis: Pathology and pathogenesis", section on 'Microbial dysbiosis'.)

Antibiotic resistance and the approach to the evaluation of the ill-appearing young infant is discussed in more detail elsewhere. (See "Neonatal bacterial sepsis: Treatment, prevention, and outcome in neonates born at less than 35 weeks gestation", section on 'Antibiotic therapy' and "Approach to the ill-appearing infant (younger than 90 days of age)".)

DIFFERENTIAL DIAGNOSIS — 

The differential diagnosis for neonatal sepsis in preterm neonates is broad and is similar to that for term or near-term newborns (table 2). It includes infectious and noninfectious etiologies:

●Focal bacterial infections – Focal bacterial infections can occur with or without associated bacteremia. Examples include:

•Urinary tract infection (see "Urinary tract infections in neonates")

•Pneumonia (see "Neonatal pneumonia")

•Meningitis (see "Bacterial meningitis in the neonate: Clinical features and diagnosis")

•Cellulitis (see "Skin and soft tissue infections in neonates: Evaluation and management")

•Necrotizing enterocolitis (see "Neonatal necrotizing enterocolitis: Clinical features and diagnosis")

●Nonbacterial infections – Nonbacterial infections that may mimic bacterial sepsis in preterm neonates include:

•Fungal infection – Candidiasis (see "Candida infections in neonates: Epidemiology, clinical manifestations, and diagnosis" and "Unusual fungal infections in the neonate")

•Viral infections – Enteroviruses, herpes simplex virus, cytomegalovirus, influenza viruses, and respiratory syncytial virus (see "Neonatal herpes simplex virus (HSV) infection: Clinical features and diagnosis" and "Enterovirus and parechovirus infections: Clinical features, laboratory diagnosis, treatment, and prevention", section on 'Neonates' and "Overview of cytomegalovirus (CMV) infections in children" and "Respiratory syncytial virus infection: Clinical features and diagnosis in infants and children")

•Spirochetal infections – Syphilis (see "Congenital syphilis: Clinical manifestations, evaluation, and diagnosis")

•Parasitic infections – Congenital malaria and toxoplasmosis (see "Congenital toxoplasmosis: Clinical features and diagnosis", section on 'Clinical features' and "Malaria in pregnancy: Epidemiology, clinical manifestations, diagnosis, and outcome", section on 'Vertical transmission')

●Noninfectious etiologies – Noninfectious causes of cardiorespiratory instability in preterm neonates include:

•Respiratory distress syndrome and other noninfectious causes of neonatal respiratory distress (table 3) (see "Respiratory distress syndrome (RDS) in preterm neonates: Clinical features and diagnosis" and "Meconium aspiration syndrome: Pathophysiology, clinical manifestations, and diagnosis")

•Metabolic conditions (eg, hypoglycemia, inborn errors of metabolism) (see "Pathogenesis, screening, and diagnosis of neonatal hypoglycemia" and "Inborn errors of metabolism: Epidemiology, pathogenesis, and clinical features" and "Metabolic emergencies in suspected inborn errors of metabolism: Presentation, evaluation, and management")

•Patent ductus arteriosus, if large in size (see "Patent ductus arteriosus (PDA) in preterm infants: Clinical features and diagnosis")

•Critical congenital heart disease (see "Evaluation of suspected critical congenital heart disease (CHD) in the newborn")

•Central nervous system insult (see "Germinal matrix and intraventricular hemorrhage (GMH-IVH) in the newborn: Risk factors, clinical features, screening, and diagnosis" and "Neonatal encephalopathy: Clinical features and diagnosis")

•Congenital anomalies (eg, diaphragmatic hernia, tracheoesophageal fistula) (see "Congenital diaphragmatic hernia (CDH) in the neonate: Clinical features and diagnosis" and "Congenital anomalies of the intrathoracic airways and tracheoesophageal fistula")

•Side effects of certain medications (eg, opiates, prostaglandins, immunizations) (see "Cyanotic congenital heart disease (CHD) in the newborn: Causes, evaluation, and initial management", section on 'Prostaglandin E1' and "Management and prevention of pain in neonates", section on 'Adverse effects')

•Withdrawal from in utero substance exposure (ie, neonatal abstinence syndrome) (see "Neonatal abstinence syndrome (NAS): Clinical features and diagnosis")

SOCIETY GUIDELINE LINKS — 

Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Sepsis in neonates" and "Society guideline links: Group B streptococcal infection in pregnant people and neonates".)

INFORMATION FOR PATIENTS — 

UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Sepsis in newborn babies (The Basics)")

SUMMARY AND RECOMMENDATIONS

●Incidence and microbiology – Neonatal bacterial sepsis, defined as sepsis caused by bacteremia, is a major cause of mortality and morbidity in preterm (gestational age [GA] <35 weeks) and very low birth weight (VLBW; <1500 g) neonates. Risk increases with decreasing GA and birth weight (BW). Neonatal sepsis is classified by the neonate's age at the time of onset of symptoms:

•Early-onset sepsis (EOS; onset within 72 hours after birth) – EOS occurs in approximately 1 to 2 percent of VLBW neonates. It is due to vertically transmitted bacteria. Escherichia coli and Group B Streptococcus (GBS) are the most common pathogens (table 1). (See 'Incidence' above and 'Early-onset sepsis' above.)

•Late-onset sepsis (LOS; onset at ≥72 hours after birth) – LOS occurs in approximately 10 to 40 percent of VLBW neonates. LOS is due to vertically transmitted bacteria or from environmental horizontal transmission. Coagulase-negative staphylococci (CoNS), Staphylococcus aureus, and gram-negative bacteria (E. coli, Klebsiella,) are the most commonly isolated pathogens (table 1). (See 'Incidence' above and 'Late-onset sepsis' above.)

●Risk factors – Preterm neonates are at higher risk of sepsis compared with term neonates because they have decreased immunocompetence, their epithelial mucosal barrier is immature, and they are more likely to require invasive devices (eg, central venous catheters, mechanical ventilation) and other interventions. Maternal risk factors also contribute to risk for EOS. (See 'Risk factors' above.)

●Clinical manifestations – In preterm neonates, the spectrum of symptoms of neonatal sepsis ranges from nonspecific subtle findings (eg, mild increase in the frequency of apnea) to fulminant septic shock (eg, hypotension or evidence of poor perfusion). Nonspecific signs frequently observed in preterm neonates with sepsis include (see 'Clinical manifestations' above):

•Respiratory distress and/or increase in respiratory support requirement

•Lethargy or hypotonia

•Change in frequency or severity of apnea (as compared with baseline)

•Feeding intolerance

•Temperature instability

•Increase in heart rate

●Evaluation and empiric management

•EOS – The evaluation of preterm neonates for EOS requires categorization of the risk (high versus low) based on the circumstances leading to preterm birth (algorithm 1) (see 'Assessing the risk for early-onset sepsis' above):

-Neonates at increased risk – Neonates born preterm due to circumstances associated with maternal intraamniotic infection (IAI) are at the highest risk for EOS. This includes birth due to premature rupture of membranes (PROM), unexplained nonreassuring fetal status, diagnosis of or other concern for IAI. These neonates warrant laboratory evaluation for sepsis with blood cultures (ideally two bottles) and cultures of other potential foci of infection; we obtain cerebrospinal fluid (CSF) studies in those with signs of meningitis or severe sepsis. (See 'Approach to neonates at increased risk' above.)

In these neonates, we initiate empiric parenteral antibiotics pending culture results, as discussed separately. (See "Neonatal bacterial sepsis: Treatment, prevention, and outcome in neonates born at less than 35 weeks gestation", section on 'Empiric antibiotic therapy'.)

-Neonates at low risk– Preterm neonates born by cesarean delivery without any attempt to induce labor and with ROM at delivery are at the lowest risk for EOS. Neonates born by vaginal delivery for maternal causes without any concern for IAI are also at low risk. In these neonates, we do not routinely check blood cultures and we do not initiate empiric antibiotic therapy unless there is significant clinical instability that does not improve after initial delivery room stabilization or signs of shock. Support for this approach comes from data indicating a very low incidence of EOS in patients categorized as low risk. (See 'Approach to neonates at low risk' above.)

•LOS – Preterm neonates older than 72 hours who have clinical features consistent with severe sepsis or shock, or who deviate significantly from the usual pattern of activity or feeding, warrant laboratory evaluation for LOS with cultures of blood (ideally two bottles), urine, CSF, and other potential foci of infection. (See 'Late-onset (≥72 hours)' above.)

In these neonates, we initiate empiric parenteral antibiotics pending culture results, as discussed separately. (See "Neonatal bacterial sepsis: Treatment, prevention, and outcome in neonates born at less than 35 weeks gestation", section on 'Empiric antibiotic therapy'.)

•Other testing – Preterm neonates undergoing evaluation for sepsis also warrant workup for alternative diagnoses, which generally includes a complete blood count, chest radiographs, and serum glucose. (See 'Diagnostic tests' above.)

●Establishing the diagnosis

•Blood culture positive – If blood cultures are positive for a non-CoNS pathogen, the diagnosis of sepsis is established, and antibiotic therapy is tailored based on the isolated pathogen and its antimicrobial susceptibility pattern. If a coagulase-negative staphylococcus (CoNS) is isolated, we diagnose sepsis if it grows from two blood cultures; otherwise, clinical assessment is necessary to distinguish true infection from contamination. (See 'Blood culture positive' above.)    

•Blood culture negative – If blood cultures are negative at 36 to 48 hours and the neonate has mild and/or transient symptoms, sepsis is unlikely, and empiric antibiotic therapy should be discontinued. If blood cultures are negative but the neonate remains clinically unwell, evaluation for alternative causes (including other infections) is necessary. (See 'Blood culture negative' above.)

●Differential diagnosis – The differential diagnosis of neonatal sepsis includes other systemic infections, inborn errors of metabolism, critical congenital heart disease, and neonatal respiratory distress (table 2 and table 3). (See 'Differential diagnosis' above.)