Acknowledgment

Drs. S. Michael Marcy, Carol Baker and Debra L. Palazzi contributed to this chapter in earlier editions. The authors are indebted to these scholars for their roles in the preparation of this chapter.

Bacterial sepsis in the neonate is a clinical syndrome characterized by systemic signs of infection and accompanied by bacteremia in the first month of life. Meningitis in the neonate usually is a sequela of bacteremia and is discussed in this chapter because meningitis and sepsis typically share a common cause and pathogenesis. Infections of the bones, joints, and soft tissues and of the respiratory, genitourinary, and gastrointestinal tracts can be accompanied by bacteremia, but the cause, clinical features, diagnosis, and management of these infections are sufficiently different to warrant separate discussions. Bloodstream and central nervous system (CNS) infections caused by group B streptococci (GBS), Staphylococcus aureus, and coagulase-negative staphylococci (CoNS), Neisseria gonorrhoeae , Listeria monocytogenes , Salmonella spp., and Mycobacterium tuberculosis are described in detail in other individual chapters. Chapter 2 describes the features of neonatal sepsis and meningitis in developing regions.

The two patterns of disease, early-onset and late-onset, have been associated with systemic bacterial infections during the first month of life ( Table 6-1 ). Early-onset disease typically presents as a fulminant, systemic illness during the first 24 hours of life (median age of onset approximately 6 hours), with the large majority of other cases presenting on the second day of life. Infants with early-onset disease can have a history of one or more obstetric complications, including premature or prolonged rupture of maternal membranes, preterm onset of labor, chorioamnionitis, and peripartum maternal fever, and many of the infants are premature or of low birth weight (LBW). Bacteria responsible for early-onset disease are acquired hours before delivery from the birth canal during delivery after overt or occult rupture of membranes. The mortality rate varies from 3% to as high as 50% in some series, especially with gram-negative pathogens. Late-onset disease has been variably defined for epidemiologic purposes as occurring after 72 hours to 6 days (e.g., GBS) of life. Very-late-onset infection caused by GBS (disease in infants older than 3 months) is discussed in Chapter 12 . Term infants with late-onset infections can have a history of obstetric complications, but these are less characteristic than in early-onset sepsis or meningitis. Bacteria responsible for late-onset sepsis (LOS) and meningitis include those acquired from the maternal genital tract and organisms acquired after birth from human contacts or, infrequently, from contaminated hospital equipment or materials, where prolonged intensive care is needed for a neonate. The mortality rate usually is lower than that for early-onset sepsis but can vary between 2% and 40%, with the latter figure typically for very-low-birth-weight (VLBW) infants with gram-negative sepsis. Because different microorganisms are responsible for disease according to age at onset, the choice of antimicrobial agents also differs. Some organisms, such as Escherichia coli, groups A and B streptococci, and L. monocytogenes, can be responsible for early- and late-onset infections, whereas others, such as S. aureus, CoNS, and Pseudomonas aeruginosa , rarely cause early-onset and typically are associated with late-onset disease. The survival of VLBW infants with prolonged stays in the neonatal intensive care unit (NICU) has been accompanied by increased risk for nosocomial or hospital-associated infections and for very-late-onset disease (see Chapter 35 ).

Table 6-1
Characteristics of Early-Onset and Late-Onset Neonatal Sepsis
Characteristic Early Onset Late Onset
Time of onset (days) 0-6 7-90
Complications of pregnancy or delivery + ±
Source of organism Mother’s genital tract Mother’s genital tract; postnatal environment
Usual clinical presentation Fulminant Slowly progressive or fulminant
Multisystem Focal
Pneumonia frequent Meningitis frequent
Mortality rate (%) 3-50 2-40

Many studies define early-onset sepsis as that which occurs in the first 72 hours of life; others in the first 5 or 6 days of life.

Very small premature infants may have late-onset sepsis beyond 90 days of life.

Higher mortality rates in earlier studies.

Bacteriology

The changing pattern of organisms responsible for neonatal sepsis is well illustrated in a series of reports by pediatricians at the Yale–New Haven Hospital covering the period 1928 to 2003 ( Table 6-2 ). Before development of the sulfonamides, gram-positive cocci, including S. aureus and β-hemolytic streptococci, caused most cases of neonatal sepsis. With the introduction of antimicrobial agents, gram-negative enteric bacilli, particularly E. coli , became the predominant cause of serious infection in the newborn. Reports for the periods of 1966 to 1978 and 1979 to 1988 document the rise to importance of GBS and E. coli as agents of neonatal sepsis. In the most recent analysis from 1989 to 2003, CoNS species, predominantly Staphylococcus epidermidis , emerged as the single most commonly identified agent of neonatal sepsis, with GBS, E. coli , Enterococcus faecalis , S. aureus , and Klebsiella spp. also occurring at substantial frequency. The latest reports also document the problem of sepsis in very premature and LBW infants who have survived with the aid of sophisticated life-support equipment and advances in neonatal intensive care; it is these infants for whom CoNS are particularly threatening. Emerging data from the same center indicate that intrapartum antibiotic prophylaxis protocols, although reducing the overall incidence of early-onset sepsis, may be influencing a higher proportion of septicemia attributable to ampicillin-resistant E. coli .

Table 6-2
Bacteria Causing Neonatal Sepsis at Yale–New Haven Hospital, 1928-2003
No. of Cases
Organism 1928-1932 1933-1943 1944-1957 1958-1965 1966-1978 § 1979-1988 щ 1999-2003
β-Hemolytic streptococci 15 18 11 8 86 83 155
Group A 16 5 0 0 0 0
Group B 2 4 1 76 64 86
Group D (Enterococcus) 0 1 7 9 19 65
Viridans streptococci 11 10
Staphylococcus aureus 11 4 8 2 12 14 70
Staphylococcus epidermidis 36 248
Streptococcus pneumoniae 2 5 3 2 2 2 0
Haemophilus spp. 1 9 9 5
Escherichia coli 10 11 23 33 76 46 106
Pseudomonas aeruginosa 1 0 13 11 5 6 33
Klebsiella and Enterobacter spp. 0 0 0 8 28 25 97
Others 0 6 4 9 21 38 54
Total no. of cases 39 44 62 73 239 270 784
Mortality rate for years 87% 90% 67% 45% 26% 16% 3%

Data from Dunham EC: Septicemia in the newborn, Am J Dis Child 45:229, 1933.

Data from Nyhan WL, Fousek MD: Septicemia of the newborn, Pediatrics 22:268, 1958.

Data from Gluck L, Wood HF, Fousek MD: Septicemia of the newborn, Pediatr Clin North Am 13:1131, 1966.

§ Data from Freedman RM, Ingram DL, Cross I, et al: A half century of neonatal sepsis at Yale, Am J Dis Child 35:140, 1981.

щ Data from Gladstone IM, Ehrenkranz RA, Edberg SC, Baltimore RS: A ten-year review of neonatal sepsis and comparison with the previous fifty-year experience, Pediatr Infect Dis J 9:819, 1990.

Data from Bizzarro MJ, Raskind C, Baltimore RS, Gallagher PG: Seventy-five years of neonatal sepsis at Yale: 1928-2003, Pediatrics 116:595, 2005.

The etiologic pattern of microbial infection observed at Yale Medical Center also has been reported in studies of neonatal sepsis carried out at other centers during the same intervals. Studies indicate that GBS and gram-negative enteric bacilli, predominantly E. coli, were the most frequent pathogens for sepsis, but other organisms were prominent in some centers. S. aureus was an important cause of sepsis in the mid-1980s in Finland and East Africa and a more recently significant pathogen in Connecticut and southern Israel. S. epidermidis was responsible for 53% of cases in Liverpool, CoNS account for 35% to 48% of all LOS in VLBW infants across the United States and in Israel, and Klebsiella and Enterobacter spp. were the most common bacterial pathogens in Tel Aviv. Sepsis and focal infections in neonates in developing countries are further discussed in Chapter 2 .

The Yale data also provide information about the microorganisms responsible for early- and late-onset bacterial sepsis ( Table 6-3 ). GBS were responsible for most early-onset disease. CoNS, S. aureus, E. coli, Enterococcus spp., and Klebsiella spp. were the major pathogens of late-onset disease; a wide variety of gram-positive cocci and gram-negative bacilli are documented as causes of bacterial sepsis in the infant after age 30 days.

Table 6-3
Microbiology of Neonatal Sepsis at Yale–New Haven Hospital, 1989-2003
Data from Bizzaro MJ, Raskind C, Baltimore RS, Gallagher PG: Seventy-five years of neonatal sepsis at Yale: 1928-2003, Pediatrics 116:595, 2005.
No. of Isolates
Age When Cultured (days)
Microorganism 0-4 5-30 >30 Transported Infants Total
Staphylococcus aureus 8 18 20 24 70
Coagulase-negative staphylococci 6 119 42 81 248
Group B streptococci 53 12 7 14 86
Enterococcus spp. 5 21 23 33 82
Viridans streptococci 0 3 3 4 10
Stomatococcus spp. 0 0 0 1 1
Bacillus spp. 1 0 1 0 2
Listeria monocytogenes 1 0 0 0 1
Escherichia coli 25 27 12 41 106
Klebsiella pneumoniae 0 20 9 18 47
Klebsiella oxytoca 0 7 8 4 19
Enterobacter aerogenes 0 1 3 4 8
Enterobacter agglomerans 0 3 1 0 4
Enterobacter cloacae 0 7 5 7 19
Serratia marcescens 0 6 10 7 23
Pseudomonas aeruginosa 2 14 4 13 33
Acinetobacter spp. 1 0 2 1 4
Proteus mirabilis 0 1 1 1 3
Citrobacter freundii 1 0 0 1 2
Haemophilus influenzae 5 0 0 0 5
Bacteriodes spp. 0 0 1 2 3
Yersinia enterocolitica 0 1 0 2 3
Other gram-negative rods 0 3 0 1 4
Candida and other fungi/yeast 3 41 16 18 78
Total 112 304 169 277 862

The mortality rates for neonatal sepsis over time are documented in the Yale Medical Center reports. In the preantibiotic era, neonatal sepsis usually was fatal. Even with the introduction of penicillins and aminoglycosides in the reports from 1944 to 1965, death resulted from sepsis in most infants. Concurrent with the introduction of NICUs and technologic support for cardiorespiratory and metabolic functions beginning in the early 1970s, the mortality rate was reduced to 16%. By 1989 to 2003, mortality from neonatal sepsis in this academic medical center was a rare event, occurring in only 3% of cases. A decline in the incidence of early-onset sepsis, commonly associated with more virulent pathogens, coupled with an increase in late and “late-late”–onset sepsis from CoNS and other commensal species (which together now account for nearly half of all cases), has contributed to the improved survival figures, along with continued advances in care and monitoring of the critically ill infant.

From 2005 to 2008, 658 cases of neonatal early-onset sepsis were reported to the Centers for Disease Control and Prevention (CDC) Active Bacterial Core surveillance (ABCs) sites in four states (California, Connecticut, Georgia, Minnesota), for an incidence of approximately 77 cases per 1000 live births (95% confidence interval [CI], 0.72 to 0.84) associated with a 10.9% mortality rate ( Table 6-4 ). The five most commonly reported pathogens were GBS (37.8%), E. coli (24.2%), viridans streptococci (17.9%), S. aureus (4.0%), and Haemophilus influenzae (4.0%). E. coli infections had the highest case fatality rate at 24.5%. Black preterm infants had the highest disease incidence (5.14 cases/1000 live births) and case fatality ratio (24.4%), whereas nonblack term infants had the lowest incidence (0.40 cases/1000 live births) and case fatality ratio (1.6%). E. coli was the most common infection (1.18 cases/1000 live births) with the highest case fatality ratio (32.1%) among preterm infants, whereas GBS were the leading pathogens among term infants (0.22 cases/1000 live births), with no reported deaths.

Table 6-4
Invasive Early-Onset Neonatal Sepsis Cases and Deaths, Centers for Disease Control Active Bacterial Core Surveillance, 2005-2008
Data from Weston EJ, Pondo T, Lewis MM, et al: The burden of invasive early-onset neonatal sepsis in the United States, 2005–2008, Pediatr Infect Dis J 30:937, 2011.
Total Black Preterm Black Term Nonblack Preterm Nonblack Term
Cases (Rate) Deaths (CFR) Cases (Rate) Deaths (CFR) Cases (Rate) Deaths (CFR) Cases (Rate) Deaths (CFR) Cases (Rate) Deaths (CFR)
Total 658 (0.77) 72 (10.9) 131 (5.14) 32 (24.4) 120 (0.89) 2 (1.7) 158 (2.27) 34 (21.5) 249 (0.040) 4 (1.6)
GBS 249 (0.29) 17 (6.8) 40 (1.57) 9 (22.5) 75 (0.55) 0 (0) 38 (0.55) 8 (21.1) 96 (0.15) 0 (0)
Escherichia coli 159 (0.19) 39 (24.5) 46 (1.81) 17 (37.0) 16 (0.12) 1 (6.3) 66 (0.95) 19 (28.8) 31 (0.05) 2 (6.5)
E. coli (Amp-R) 81 (0.09) 16 (19.8) 31 (1.22) 9 (29.0) 3 (0.02) 0 (0) 31 (0.45) 6 (19.4) 16 (0.03) 1 (6.3)
Viridans streptococci 118 (0.14) 3 (2.5) 16 (0.63) 2 (12.5) 16 (0.12) 0 (0) 18 (0.26) 0 (0) 68 (0.11) 1 (1.5)
Staphylococcus aureus 26 (0.03) 2 (7.7) 2 (0.08) 1 (50.0) 5 (0.04) 1 (20.0) 1 (0.01) 0 (0) 18 (0.03) 0 (0)
Haemophilus influenzae 26 (0.03) 4 (15.4) 10 (0.39) 1 (10.0) 0 (0) 0 (0) 12 (0.17) 3 (25.0) 4 (0.006) 0 (0)
Other pathogens 80 (0.09) 7 (8.9) 17 (0.67) 2 (11.8) 8 (0.06) 0 (0) 23 (0.33) 4 (17.4) 32 (0.05) 1 (3.1)
Amp - R , Ampicillin-resistant; CFR , case-fatality ratio; GBS , group B streptococci.
Preterm, <37 weeks of gestation; term, ≥37 weeks of gestation
Rate is per 1000 live births.
Other pathogen categories include: Enterococcus spp. ( n = 21), Listeria monocytogenes ( n = 9), Streptococcus pneumoniae ( n = 8), Citrobacter koseri ( n = 7), Klebsiella pneumoniae ( n = 6), group A Streptococcus ( n = 3), Streptococcus bovis ( n = 3), Bacteroides fragilis ( n = 2), group G Streptococcus ( n = 2).

Occurring in infants the first 72 hours of life:

The incidence of neonatal sepsis showed a strong inverse correlation to birth weight in the latest Yale cohort: greater than 2000 g (0.2%), 1500 to 1999 g (2.5%), 1000 to 1499 g (9.4%), 750 to 999 g (14.8%), and less than 750 g (34.8%). Survival of VLBW infants (<1500 g) has been accompanied by an increased risk for invasive, nosocomial, or health care–associated bacterial infection as a cause of morbidity and mortality. The danger of sepsis is documented in a multicenter trial that enrolled 2416 VLBW infants in a study of the efficacy of intravenous immunoglobulin in preventing nosocomial infections. Sixteen percent of the VLBW infants developed septicemia at a median age of 17 days, with an overall mortality rate of 21% and a hospital stay that averaged 98 days; infants without sepsis had an overall mortality rate of 9% and 58-day average length of stay. Stoll and colleagues reported recent patterns of pathogens causing early-onset sepsis in VLBW infants (400-1500 g) in the centers participating in the National Institute of Child Health and Human Development (NICHD) Neonatal Research Network ( Table 6-5 ). Compared with earlier cohorts, a marked reduction in GBS infections (from 5.9-2.08/1000 live births) and an increase in E. coli infections (3.2-5.09/1000 live births) were noted, although the overall incidence of neonatal sepsis in this population did not change.

Table 6-5
Characteristics and Mortality Rate of 389 U.S. Infants With Early-Onset Sepsis
Data from Stoll BJ, Hansen NI, Sánchez PJ, et al: Early onset neonatal sepsis: the burden of group B streptococcal and E. coli disease continues, Pediatrics 127:817, 2011.
All Patients ( N = 389) Preterm (22-36 wk) with GBS or E. coli Term (37+ wk) with GBS or E. coli Overall with GBS or E. coli
Infant Characteristics, n (%) GBS
( n = 43)
E. coli
( n = 87)
GBS
( n = 117)
E. coli
( n = 20)
GBS
( n = 160)
E. coli
( n = 107)
Birth Weight (g)
401-1500 142 (37%) 27 (63) 66 (76) 0 (0) 0 (0) 27 (17) 66 (62)
1501-2500 51 (13%) 11 (26%) 20 (23%) 3 (3) 0 (0) 14 (9) 20 (19)
2501+ 196 (50%) 5 (12) 1 (1) 114 (97) 20 (100) 119 (74) 21 (20)
Infant Gender
Male 205 (53) 22 (51) 47 (54) 57 (49) 15 (75) 79 (49) 62 (58)
Female 184 (47) 21 (49) 40 (46) 60 (51) 5 (25) 81 (51) 45 (42)
Delivery
Vaginal 181 (47) 18 (43) 29 (33) 63 (54) 9 (47) 81 (51) 37 (35)
Cesarean section 204 (53) 25 (57) 58 (67) 54 (46) 10 (53) 76 (49) 68 (65)
ROM > 18 hr PTD 127 (33) 19 (44) 54 (62) 17 (15) 4 (21) 36 (23) 58 (55)
SROM < 37 wk of gestation 157 (40) 30 (70) 75 (86) NA NA NA NA
Symptoms < 72 hr PTD
Maternal temperature > 38.0° C 102 (26) 3 (7) 27 (32) 42 (36) 6 (30) 45 (28) 33 (31)
Uterine or abdominal tenderness 56 (15) 6 (14) 26 (31) 4 (3) 0 (0) 10 (6) 26 (25)
Maternal tachycardia (>100 bpm) 115 (30) 14 (33) 36 (42) 30 (26) 7 (35) 44 (28) 43 (41)
Fetal tachycardia (>160 bpm) 120 (31) 11 (26) 34 (40) 40 (34) 7 (35) 51 (32) 41 (39)
Placental Pathology Performed 248 (65) 33 (77) 72 (85) 49 (43) 13 (65) 82 (52) 85 (81)
Histologic Chorioamnionitis 190 (77) 30 (91) 63 (88) 32 (65) 8 (62) 62 (76) 71 (84)
Infant Mortality n (%)
All Deaths 61 (16) 13 (30) 33 (38) 2 (2) 2 (10) 15 (9) 35 (33)
Time of Death
0-3 days 35 (57) 7 (54) 21 (64) 1 (50) 1 (50) 8 (53) 22 (63)
4-7 days 12 (20) 1 (8) 7 (21) 1 (50) 1 (50) 2 (13) 8 (23)
8-14 days 3 (5) 0 (0) 2 (6) 0 (0) 0 (0) 0 (0) 2 (6)
>14 days 11 (18) 5 (38) 3 (9) 0 (0) 0 (0) 5 (33) 3 (9)
bpm, Beats per minute; E . coli , Escherichia coli ; GBS , group B streptococci; NA , not available; PTD , prior to delivery; ROM , rupture of membranes; SROM , spontaneous rupture of membranes.

Organisms responsible for bacterial meningitis in the newborn are listed in Table 6-6 , which summarizes data collected from 1932 to 1997 at neonatal centers in the United States, The Netherlands, Great Britain, and Israel. Gram-negative enteric bacilli and GBS currently are responsible for most cases. Organisms that cause acute bacterial meningitis in older children and adults— Streptococcus pneumoniae, Neisseria meningitidis, and type b and nontypeable Haemophilus influenzae —are relatively infrequent causes of meningitis in the neonate. A nationwide survey of causative agents of neonatal meningitis in Sweden between 1976 and 1983 indicated a shift from bacterial to viral or unidentified microorganisms, with lower attributable mortality rates.

Table 6-6
Bacteria Associated With Neonatal Meningitis in Selected Studies
No. of Cases of Association
Organism Boston, 1932-1957, 77 Cases Los Angeles, 1963-1968, 125 Cases Houston, 1967-1972, 51 Cases Multihospital Survey, 1971-1973, 131 Cases The Netherlands, 1976-1982, 280 Cases Great Britain, 1985-1987, 329 Cases Dallas, 1969-1989, 257 Cases Israel, 1986-1994, 32 Cases 31 Great Britain, 1996-1997, 144 Cases
β-Hemolytic streptococci (group not stated) 9 12
β-Hemolytic streptococci
Group A 1 2
Group B 18 41 68 113 134 6 69
Group D 2 4 1
Staphylococcus epidermidis or coagulase-negative Staphylococcus 5 3 9 2 2
Staphylococcus aureus 12 1 3 1 7 4
Streptococcus pneumoniae 7 4 3 2 6 21 18 8
Listeria monocytogenes 6 5 7 12 21 7
Escherichia coli 25 44 16 50 132 2 42 4 26
Pseudomonas aeruginosa 4 1 2 2 4 3 1
Klebsiella and Enterobacte r spp. 3 13 3 19 8 10 4
Proteus spp. 2 5 4 5 8 3 2
Haemophilus spp. 2 2 3 2 12 1
Neisseria meningitidis 1 1 3 14 6
Salmonella spp. 2 4 3 3 2 4 1
Miscellaneous 12 28 1 7 15 32 46 23

Survey of 16 newborn nurseries participating in neonatal meningitis study of intrathecal gentamicin under the direction of Dr. George McCracken, Jr.

Authors report an additional nine cases of gram-positive and six cases of gram-negative meningitis with organisms not otherwise specified.

Authors report 16 cases related to enteric bacteria, including Escherichia coli , Proteus spp., and Klebsiella-Enterobacter group.

Group B Streptococci

Group B β-hemolytic streptococci were implicated in human disease shortly after the precipitin-grouping technique was described. For the past 3 decades, GBS has been the most common pathogen causing invasive disease in neonates throughout the United States and western Europe (see Chapter 12 ).

Streptococcus agalactiae, the species designation of GBS, has a characteristic colonial morphology on suitable solid media. The organism produces a mucoid colony with a narrow zone of β-hemolysis on sheep blood–agar media. The GBS can be differentiated immunochemically on the basis of their type-specific polysaccharides. Ten capsular types—Ia, Ib, II, III, IV, V, VI, VII, VIII, and IX—have been characterized, and most invasive human isolates can be classified as one of these types, with serotypes Ia, III, and V the most prevalent in many recent epidemiologic surveys.

GBS have been isolated from various sites and body fluids, including throat, skin, wounds, exudates, stool, urine, cervix, vagina, blood, joint, pleural or peritoneal fluids, and cerebrospinal fluid (CSF). The organisms frequently are found in the lower gastrointestinal and genital tracts of adult women and men and in the lower gastrointestinal and upper respiratory tracts of newborns. Patterns of early-, late-, and very-late-onset disease have been associated with GBS (see Table 6-1 ). Early-onset disease presents as a multisystem illness, with rapid onset typically during the first day or two of life and is frequently characterized by severe respiratory distress. The pathogenesis is presumed to be similar to that of other forms of early-onset sepsis of neonates. The mortality rate is estimated at 8% but was previously as high as 50% in the 1970s.

Clinical manifestations of late-onset neonatal sepsis are more insidious than those of early-onset disease, and meningitis is frequently a part of the clinical picture. However, some infants with meningitis have a fulminant onset with rapid progression to centrally mediated apnea. Many of the infants are products of a normal pregnancy and delivery and have no problems in the nursery. It is uncertain whether GBS infection was acquired at the time of birth and carried until disease developed, was acquired after delivery from the mother or other household contacts, or was acquired from other infants or personnel in the nursery. In late-onset infection, a majority of strains belong to serotype III. The mortality rate, estimated at 3%, is lower than that for early-onset disease. With increasing survival of extremely-low-birth-weight (ELBW) (<1000 g) infants, very-late-onset disease (>89 days) has been described in the past decade.

In addition to sepsis and meningitis, other manifestations of neonatal disease caused by GBS include pneumonia, empyema, facial cellulitis, ethmoiditis, orbital cellulitis, conjunctivitis, necrotizing fasciitis, osteomyelitis, suppurative arthritis, and impetigo. Bacteremia without systemic or focal signs of sepsis can occur. GBS infection in pregnant women can result in peripartum infections, including septic abortion, chorioamnionitis, peripartum bacteremia, septic pelvic thrombophlebitis, meningitis, and toxic shock syndrome.

Group A Streptococci

Streptococcal puerperal sepsis has been recognized as a cause of morbidity and mortality among parturient women since the 16th century. Neonatal group A streptococcal (GAS) infection now is reported infrequently but can occur rarely in epidemic form in nurseries. The reemergence of virulent GAS infections during the last 4 decades, including invasive disease and toxic shock syndrome, has been reflected in more case reports of severe disease in the pregnant woman and the newborn.

GAS disease in the mother can affect the fetus or newborn in two clinical patterns. Maternal streptococcal bacteremia during pregnancy can lead to in utero infection resulting in fetal loss or stillbirth, or alternatively, acquisition of GAS from the maternal genital tract can cause early-onset neonatal sepsis similar to early-onset GBS disease. In the first form of disease, previously healthy pregnant women with influenza-like signs and symptoms have been reported. This presentation rapidly progressed to disseminated intravascular coagulopathy and shock, with high mortality and risk to the fetus or newborn.

The features of 38 cases of neonatal invasive GAS infection from the literature were recently catalogued. Overall mortality rate in neonatal invasive GAS infection was significantly high, at 31%. Most of these infants presented with early-onset infection (62%), with many occurring in the first 48 hours of life. A specific focus of GAS infection was documented in three quarters of cases; 42% of neonates had pneumonia, sometimes complicated by empyema, and 17% had a toxic-shock–like syndrome presentation. Among the cases of early-onset GAS infection, puerperal sepsis or toxic shock–like syndrome in the mother during the peripartum period was an associated factor in 62% of cases. In late-onset cases of neonatal GAS infection reviewed in this series, soft tissue infections, meningitis, and pneumonia were among the reported clinical manifestations. An earlier review by Greenberg and colleagues on 15 cases of GAS neonatal infection yielded similar statistics on clinical presentations and mortality.

In addition to sepsis, meningitis, and toxin-mediated disease in the neonate, focal infections, including cellulitis, omphalitis, pneumonia, empyema, osteomyelitis, and parotitis, have been reported. Because all GAS are susceptible to β-lactam antibiotics, the current strategy for prevention or treatment of infections caused by GBS also could apply to infections caused by GAS.

Streptococcus Pneumoniae

Although pneumococcal infections in the neonate are unusual occurrences, they are associated with substantial morbidity and mortality. Malhotra and colleagues recently reported four infants with invasive neonatal pneumococcal infections that developed within the first 24 hours of life, with all four having clinical and radiologic features of pneumonia and a pattern of disease rather indistinguishable from typical severe early-onset GBS sepsis. One of the infants was a 33-week premature infant, and one of the mothers had chorioamnionitis before delivery. All four infants survived, with varying levels of supportive care, including extracorporeal membrane oxygenation in a child who also developed meningitis. Two infants were expected to suffer significant long-term sequelae. In another report, fatal pneumococcal bacteremia in a mother 4-weeks postpartum and the same disease and outcome in her healthy term infant who died at 6 weeks of age suggested an absence of protective antibody in mother and child.

Hoffman and colleagues, from the United States Multicenter Pneumococcal Surveillance Group, identified 20 cases of neonatal S. pneumoniae sepsis or meningitis in a review of 4428 episodes of pneumococcal infection at eight children’s hospitals from 1993 to 2001. Ninety percent of the infants were born at term, with a mean age at the onset of infection of 18.1 days. Only two of the mothers had clinically apparent infections at the time of delivery. Eight neonates had meningitis and 12 had bacteremia; 4 of the bacteremic neonates also had pneumonia. The most common infecting pneumococcal serotypes were 19 (32%), 9 (18%), and 18 (11%). Penicillin and ceftriaxone nonsusceptibility were observed in 21.4% and 3.6% of isolates, respectively. Three deaths (15%) occurred, all within 36 hours of presentation. A case report of peripartum transmission of penicillin-resistant S. pneumoniae underlines concern that the increasing use of peripartum ampicillin to prevent GBS disease in the neonate may result in an increase in neonatal infections caused by β-lactam–resistant organisms. A case of purulent pneumococcal pericarditis in a neonate has recently been reported.

Other Streptococci

Human isolates of group C and G streptococci form large β-hemolytic colonies that closely resemble those of GAS and share many virulence genes, including those encoding surface M proteins and the cytotoxin streptolysin S. Group C streptococci have been associated with puerperal sepsis, but neonatal sepsis or meningitis related to these organism is rare. Likewise, group G streptococci are an infrequent cause of neonatal sepsis and pneumonia. Maternal intrapartum transmission was the likely source for most cases, and concurrent endometritis and bacteremia in the mother and sepsis in the neonate have been reported. Recently, a case of neonatal toxic streptococcal shock syndrome attributed to maternal transmission of a group C streptococcus was reported.

Viridans streptococci are a heterogeneous group of α-hemolytic and nonhemolytic streptococci that are constituents of the normal flora of the respiratory and gastrointestinal tracts of infants, children, and adults. There are several classification schemes for these streptococci, and they may bear different designations in the literature. Streptococcus bovis is capable of causing neonatal sepsis and meningitis that is clinically similar to sepsis caused by GBS. Rare cases of fulminant neonatal sepsis or meningitis caused by Streptococcus mitis , Streptococcus gallolyticus , and Streptococcus alactolyticus have been reported.

Viridans streptococci accounted for 23% of isolates from cultures of blood and CSF obtained from neonates at the Jefferson Davis Hospital, Houston; only GBS were more common (28%) as a cause of neonatal sepsis. In this series, most infants had early-onset infection with clinical features similar to those of sepsis caused by other pathogens, but 22.6% had no signs of infection. One infant had meningitis. The case-fatality rate was 8.8%. Sepsis related to viridans streptococci also has been reported from Finland, Liverpool, Indianapolis, and Montreal. Among ventilated neonates in a NICU in Ankara, Turkey, the most prominent bacteria in bronchioalveolar lavage cultures were multidrug-resistant viridans streptococci (66%), and these were also one of the most common bloodstream isolates (29%) in the same population. It is clear from these studies that isolation of viridans streptococci from the blood culture of a neonate suspected to have sepsis cannot be considered a contaminant, as is the case in many other patient populations.

Enterococcus Species

Members of the genus Enterococcus ( E. faecalis and E. faecium ) were formerly classified as group D streptococci; but in the mid-1980s, genomic DNA sequence analysis revealed that taxonomic distinction was appropriate, and a unique genus was established. Enterococci are differentiated from nonenterococci by their ability to grow in 6.5% sodium chloride broth and to withstand heating at 60° C for 30 minutes.

Most cases of enterococcal sepsis in the neonate are caused by E. faecalis, with a smaller number caused by E. faecium ; In the 4 years beginning in 1974, 30 neonates with enterococcal sepsis occurred among 30,059 deliveries at Parkland Memorial Hospital in Dallas. During this period, enterococci were second only to GBS (99 cases) and were more common than E. coli (27 cases) as a cause of neonatal sepsis. The clinical presentation in most cases was similar to that of early-onset sepsis of any cause. Among infants with respiratory distress as a prominent sign of infection, the chest radiographs were similar to those demonstrating the hyaline membrane–appearing pattern of GBS infection. Enterococcal bacteremia during the 10 years beginning January 1977 was reported in 56 neonates from the Jefferson Davis Hospital in Houston, Texas. Patients were segregated among three clinical syndromes: early-onset disease was a mild illness with respiratory distress or diarrhea; late-onset infection often was severe with apnea, bradycardia, shock, and increased requirement for oxygen and mechanical ventilation; and many cases were nosocomial. A large series of 100 cases of enterococcal bacteremia in neonates over a 20-year period at New York Hospital–Cornell Medical Center was evaluated by McNeeley and colleagues. Common characteristics were the presence of a central venous catheter (77%) or a diagnosis of necrotizing enterocolitis (NEC; 33%).

In general, Enterococcus spp. are resistant to cephalosporins, are only moderately susceptible to penicillin G and ampicillin, and require the synergistic activity of penicillin, at high dosage, and an aminoglycoside for maximal bactericidal action; nonenterococcal strains are susceptible to penicillin G, ampicillin, and most cephalosporins. Vancomycin-resistant (VRE) Enterococcus has been reported from NICUs, causing illnesses clinically indistinguishable from vancomycin-sensitive strains, yet raises concerns about the efficacy of antimicrobial agents currently approved for use in neonates. Use of high doses of ampicillin is one option, but other drugs, including daptomycin and the oxazolidinone linezolid, may be required depending on the susceptibility pattern (see Chapter 37 ).

Staphylococcus Aureus and Coagulase-Negative Staphylococci

S. aureus and CoNS, especially S. epidermidis , colonize skin and mucosa. Isolation of S. aureus from tissue, blood, or other body fluids usually is clearly associated with disease. Most episodes of sepsis caused by S. aureus are hospital acquired, and mortality can be high (23% among 216 Swedish neonates with S. aureus bacteremia during 1967 to 1984), with LBW as the most important risk factor. Recently, reports of pneumonia and other severe nosocomial infection in neonates caused by community-acquired methicillin-resistant S. aureus (CA-MRSA) strains, including the epidemic USA300 clone, have been documented. Molecular epidemiologic techniques have established direct transmission of CA-MRSA between postpartum women and among NICU patients.

CoNS include more than 30 different species. S. epidermidis is the dominant species of CoNS responsible for neonatal sepsis, but other species, including Streptococcus capitis, Streptococcus hemolyticus and Streptococcus hominis, have been identified as causes of sepsis in the newborn. A well-documented increased incidence of CoNS sepsis has accompanied the increased survival of VLBW and ELBW infants with developmentally immature immune systems and prolonged stay in NICUs. The CoNS infections have been associated with the introduction of invasive procedures for maintenance and monitoring of the infants, in particular long-term vascular access devices. Levels of serum complement and transplacental anti-CoNS immunoglobulin G (IgG) are inversely correlated with gestational age, and this relative deficiency in preterm infants contributes to their suboptimal opsonization and impaired bacterial killing of CoNS. Because CoNS are present on the skin, isolation of these organisms from a single culture of blood can represent skin contamination but also can indicate bloodstream invasion. Collection of two cultures of blood at separate sites can assist in differentiating skin or blood-culture–bottle contamination from bloodstream invasion in the infant with suspected late-onset sepsis, and adoption of a standard two blood-culture practice can reduce the number of neonates diagnosed with CoNS and exposed to intravenous antibiotic therapy. The significance of a positive blood culture yielding CoNS is discussed in “Microbiologic Techniques.”

Many episodes of sepsis caused by CoNS are associated with the use of vascular catheters. S. epidermidis and other CoNS species can adhere to and grow on surfaces of synthetic polymers used in the manufacture of catheters. Strains obtained from infected ventricular shunts or intravenous catheters produce a mucoid substance (i.e., slime or glycocalyx) that stimulates adherence of microcolonies to various surfaces in the environment and on epithelial surfaces, ultimately leading to establishment of a biofilm. In addition to this adhesin function, the slime may protect staphylococci against antibiotics and host defense mechanisms, such as macrophage phagocytosis, predisposing to persistent infection. Parenteral nutrition with a lipid emulsion administered through a venous catheter having organisms adherent to the polymer provides nutrients for growth of the bacteria, leading to invasion of the bloodstream when the organisms reach an inoculum of sufficient size.

Disease in newborn infants caused by S. aureus and CoNS is discussed in detail in Chapter 14 .

Listeria Monocytogenes

The potential of L. monocytogenes to contaminate food products and the resultant danger to immunocompromised patients and pregnant women was reconfirmed in a 2002 outbreak involving 46 patients in eight states. This outbreak resulted in seven deaths of adults and miscarriages or stillbirths in three pregnant women. Listeria can be found in unprocessed animal products, including milk, meat, poultry, cheese, ice cream, and processed meats, and on fresh fruits and vegetables. The organism possesses several virulence factors that allow it to infect the fetal placental unit, survive and replicate within human cells, and achieve cell-to-cell spread. Although most people exposed to L. monocytogenes do not develop illness, pregnant women can suffer fetal loss, and the neonate can develop early- or late-onset sepsis and meningitis. Neonatal disease caused by Listeria is discussed in detail in Chapter 13 .

Escherichia Coli

Escherichia coli is second only to GBS as the most common cause of both early- and late-onset neonatal sepsis and meningitis. Coliform organisms are prevalent in the maternal birth canal, and most infants are colonized in their lower gastrointestinal or respiratory tracts during or just before delivery. The antigenic structure of E. coli is complex; members of this species account for more than 145 different somatic (O) antigens, approximately 50 flagellar (H) antigens, and 80 different capsular (K) antigens. Although there is a wide genetic diversity of human commensal isolates of E. coli, strains causing neonatal pathology are derived from a limited number of clones. One of these, the O18:K1:H7 clone, is distributed globally; meanwhile, others such as O83:K1 and O45:K1 are restricted to a smaller subset of countries. The presence of a 134-kDa plasmid encoding iron aquisition systems and other putative virulence genes is characteristic of several of these clones, and loss of the plasmid reduces the virulence more than 100-fold in a neonatal rat model of E. coli meningitis. In a recent analysis comparing E. coli with other agents of early-onset neonatal sepsis, infants with E. coli sepsis ( n = 19) were more likely to be premature, of VLBW (<1500 g), and to have been associated with the intrapartum characteristics of fever, premature or prolonged rupture of membranes, antibiotic use, and presentation in the first 24 hours of life. Fifteen of the 19 E. coli isolates in this study (79%) were ampicillin resistant, and three (16%) were gentamicin resistant; antepartum or intrapartum antibiotic exposure was associated with ampicillin-resistant E. coli sepsis.

The K1 capsular antigen present in certain strains of E. coli is uniquely associated with neonatal meningitis. The K1 antigen is polysialic acid that is immunochemically identical to the capsular antigen of group B N. meningitidis. McCracken and coworkers found K1 strains in the blood or CSF of most (65/77) neonates with meningitis related to E. coli. These strains also were cultured from the blood of some infants (14/36) and adults (43/301) with sepsis but without meningitis. The K1 capsular antigen was present in 88% of 132 strains from neonates with E. coli meningitis reported from The Netherlands. Infants with meningitis caused by K1 strains had significantly higher mortality and morbidity rates than did infants with meningitis caused by non-K1 E. coli strains. The K1 strains have been present in the birth canal of mothers and subsequently in cultures from their newborns, indicating that these newborn infants acquired the organisms vertically from their mothers. However, high rates of carriage of K1 strains by nursery personnel indicate that postnatal acquisition of the K1 strains in the nursery also may occur.

The pathogenesis of E. coli K1 infection is hypothesized to begin with bacterial penetration of the gastrointestinal epithelium to enter the circulation, and efficient transcytosis of gastrointestinal epithelial cell monolayers by the pathogen has been demonstrated in tissue culture. Next, the organisms can establish high-grade bacteremia in the immune-susceptible neonate through the complement resistance properties of its O lipopolysaccharide and K1 capsule–mediated impairment of opsonophagocytic killing. Finally, the pathogen possesses a series of surface protein determinants (OmpA, IbeA-C, CNF1, etc.) that mediate binding to and invasion of brain endothelial cells, as demonstrated in human tissue culture experiments and the neonatal rat model of meningitis.

Klebsiella Species

Klebsiella is a genus of Enterobacteriaceae that has emerged as a significant nosocomial pathogen in neonates. The four recognized species include Klebsiella pneumoniae, Klebsiella oxytoca, Klebsiella terrigena, and Klebsiella planticola. K. pneumoniae, the most common human pathogen, and K. oxytoca cause neonatal infections of the bloodstream, urinary tract, CNS, lung, skin, and soft tissues. Previously thought to be a nonpathogenic organism inhabiting soil and water, K. planticola has been implicated as a cause of neonatal sepsis.

In a 4-year retrospective study from Israel, Klebsiella spp. caused 31% of late-onset neonatal sepsis. Greenberg and colleagues performed an 8-year prospective study of neonatal sepsis and meningitis at Soroka University Medical Center in Israel during 1986 to 1994; 49 (20%) of 250 cases were caused by K. pneumoniae, with a mortality rate of 29%. Klebsiella was also the most common single agent in recent reviews of sepsis in Jamaican and Indian neonates. Risk factors for infection included preterm, VLBW, prolonged rupture of membranes (>24 hours), and cesarean section or instrument delivery. Klebsiella spp. appear to be among the more common causes of liver abscess complicating bacteremia in the neonate.

The reservoirs for transmission of Klebsiella infections include the hands of health care workers and the gastrointestinal tracts of hospitalized infants. Multidrug resistance, in the form of extended-spectrum β-lactamase production, of Klebsiella strains causing neonatal infections and nursery outbreaks has become a substantial problem in some nurseries and is associated with increased morbidity and mortality. Enhanced infection-control measures and changes in use of routine broad-spectrum antibiotics can reduce the frequency of these serious infections.

Enterobacter and Cronobacter Species

Among the Enterobacter spp., Enterobacter cloacae, Enterobacter sakazakii, and Enterobacter hormaechei have caused sepsis and a severe form of necrotizing meningitis in neonates. In 2008, the taxonomy of E. sakazakii was revised, resulting in identification of five species belonging to a new genus, Cronobacter.

Enterobacter septicemia was the most common nosocomial infection in neonates at the Ondokuz Mayis University Hospital in Samsun, Turkey, from 1988 to 1992. Willis and Robinson reviewed 17 cases of neonatal meningitis caused by E. sakazakii; cerebral abscess or cyst formation developed in 77% of the infants, and 50% of the infants died. Bonadio and colleagues reviewed 30 cases of E. cloacae bacteremia in children, including 10 infants younger than 2 months. Of importance was the high frequency of multidrug resistance among isolates from patients in the NICUs that was attributed to routine extended-spectrum cephalosporin use. In a recent review of Enterobacter sepsis in 28 neonates from Taiwan, thrombocytopenia (66%) and increased band-form neutrophils (41%) were common laboratory features, with a reported clinical outcome of 11% mortality, 14% meningitis, and 7% brain abscess.

In addition to the gastrointestinal tracts of hospitalized infants and hands of health care personnel, sources and modes of transmission of Enterobacter infections in the neonate include contaminated infant formula, contaminated total parenteral nutrition fluid, bladder catheterization devices, and contaminated saline. Effective infection-control measures require reinforcement of procedures, including proper hand hygiene, aseptic technique, isolation protocols, and disinfection of environmental surfaces.

Cronobacter spp. are an emerging group of opportunistic gram-negative pathogens that typically affect LBW neonates, causing life-threatening meningitis, sepsis, and NEC. An outbreak of C. sakazakii in a French NICU in 1994 involved 17 cases, including 7 neonates with NEC, 1 case of sepsis, and 1 case of meningitis; 8 infants were colonized but asymptomatic; there were three deaths. Four separable pulse types of C. sakazakii were identified, but the deaths were attributable to only one. C. sakazakii infection in vulnerable infants has often been linked to the consumption of contaminated powdered infant formula.

Citrobacter Species

Organisms of the genus Citrobacter are gram-negative bacilli that are occasional inhabitants of the gastrointestinal tract and are responsible for disease in neonates and debilitated or immunocompromised patients. The genus has undergone frequent changes in nomenclature, making it difficult to relate the types identified in reports of newborn disease over the years. For example, in 1990, Citrobacter koseri replaced Citrobacter diversus.

Citrobacter spp. are responsible for sporadic and epidemic clusters of neonatal sepsis and meningitis, and C. koseri is uniquely associated with brain abscesses. Neonatal disease can occur as early- or late-onset presentations. Brain abscesses caused by C. koseri have been reported in a pair of twins. Outbreaks of C. koseri in NICUs resulting in sepsis and meningitis, septic arthritis, and skin and soft tissue infections were reviewed by Doran. Other focal infections in neonates caused by Citrobacter spp. include bone, pulmonary, and urinary tract infections.

From 1960 to 1980, 74 cases of meningitis caused by Citrobacter spp. were reported to the CDC of the U.S. Public Health Service. In 1999, Doran reviewed an additional 56 cases of neonatal meningitis caused by Citrobacter spp. Combining results from the two studies, brain abscess developed in 73 (76%) of 96 patients for whom information was available. The pathogenesis of brain abscess caused by C. koseri is uncertain; cerebral vasculitis with infarction and bacterial invasion of necrotic tissues is one possible explanation. Studies in the neonatal rat model suggest that the ability of C. koseri to survive phagolysosome fusion and persist intracellularly within macrophages could contribute to the establishment of chronic CNS infection and brain abscess. Such persistence of C. koseri in the CNS is well illustrated by a case report of recovery of the organism from the CSF during a surgical procedure 4 years after treatment of neonatal meningitis. The mortality rate for meningitis caused by Citrobacter spp. was about 30%; most of the infants who survived had some degree of mental retardation. A review of 110 survivors of Citrobacter meningitis revealed only 20 infants who were believed to have structurally intact brains and development that was age appropriate.

Citrobacter spp. usually are resistant to ampicillin and variably susceptible to aminoglycosides. Serial neuroimaging is critical for the diagnosis of cerebral abscess in infants with Citrobacter meningitis. Surgical drainage has been used in some cases with variable success. Choosing antimicrobial agents with the most advantageous susceptibility pattern and selected surgical drainage appears to be the most promising approach to therapy. High neutrophil and CNS penetration and favorable toxicity profiles suggest ciprofloxacin and meropenem as potential antibiotic treatment options for systemic infection or meningitis caused by C. koseri. Plasmid profiles, biotypes, serotypes, and chromosomal restriction endonuclease digests are useful as epidemiologic markers for the study of isolates of C. koseri. Morris and colleagues used these markers to investigate an outbreak of six cases of neonatal meningitis caused by C. koseri in three Baltimore hospitals between 1983 and 1985. Identification of a specific outer membrane protein associated with strains isolated from CSF but uncommon elsewhere can provide a marker for virulent strains of C. koseri according to some investigators.

Serratia Marcescens

Like other members of Enterobacteriaceae, Serratia marcescens increasingly is associated with hospital-acquired infections among infants in the NICU. Late-onset sepsis has occurred in infants infected from health care equipment, the hands of heath care workers, milk bottles, aqueous solutions such as theophylline, hand hygiene washes, and lipid parenteral feeds. The gastrointestinal tracts of hospitalized infants provide a reservoir for transmission and infection. Investigation of an outbreak of multidrug-resistant S. marcescens in NICU identified exposure to inhalational therapy as an independent risk factor for acquisition. Recently, three consecutive outbreaks caused by genetically unrelated S. marcescens clones occurred in a NICU over a 3-year period, with administration of total parenteral nutrition the only statistically significant risk factor identified by logistic regression.

In a review of neonatal bacteremia and meningitis caused by S. marcescens by Campbell and colleagues, 11 (29%) of 38 infants had meningitis as a complication of their bacteremia. Mean gestational age and birth weight were 28 weeks and 1099 g, respectively. All patients required mechanical ventilation, 90% had central venous catheters in situ, 90% had received prior antibiotics, 50% had a prior intraventricular hemorrhage, 40% had a hemodynamically significant patent ductus arteriosis treated medically or surgically, and 20% had NEC with perforation. All patients were treated for a minimum of 21 days with combination antimicrobial therapy that included a third-generation cephalosporin or an ureidopenicillin and an aminoglycoside, typically gentamicin. Three of 10 patients died. Four of the 7 survivors developed severe hydrocephalus requiring ventriculoperitoneal shunt placement and had poor neurologic outcome. Poor neurologic outcome also was documented in a report of S. marcescens brain abscess, resulting in multicystic encephalomalacia and severe developmental retardation. Combination therapy with high-dose amikacin and meropenem was associated with clinical improvement in a case of S. marcescens brain abscess in a 34-week premature neonate.

Pseudomonas Aeruginosa

Pseudomonas aeruginosa usually is a cause of late-onset disease in infants who are presumably infected from their endogenous flora or from equipment, from aqueous solutions, or occasionally from the hands of health care workers. An outbreak of P. aeruginosa sepsis in a French NICU was associated with contamination of a milk bank pasteurizer. Stevens and colleagues reported nine infants with Pseudomonas sepsis, four of whom presented in the first 72 hours of life. In three of these infants, the initial signs were those of respiratory distress, and chest radiographs were consistent with hyaline membrane disease. Noma (i.e., gangrenous lesions of the nose, lips, and mouth) in a neonate has been associated with bacteremia caused by P. aeruginosa.

A retrospective review of sepsis in infants admitted over the 10-year period from 1988 through 1997 to the NICU at Children’s Hospital of the King’s Daughters in Norfolk, Virginia, identified 825 cases of LOS. Infants with Pseudomonas sepsis had the highest frequency of clinically fulminant onset (56%), and 20 (56%) of the 36 (56%) infants with Pseudomonas sepsis died within 48 hours of blood-culture collection.

P. aeruginosa conjunctivitis in the neonate is a danger because it is rapidly destructive to the tissues of the eye and because it may lead to sepsis and meningitis. Shah and Gallagher reviewed the course of 18 infants at Yale–New Haven Hospital NICU who had P. aeruginosa isolated from cultures of the conjunctiva during the 10 years beginning in 1986. Five infants developed bacteremia, including 3 with meningitis, and 2 infants died. A cluster of four fatal cases of P. aeruginosa pneumonia and bacteremia among neonates was traced by genotypic fingerprinting to their shared exposure to a health care worker experiencing intermittent otitis externa. A case of fatal multidrug-resistant pseudomonal sepsis with ecthyma gangrenosum was recently reported in a premature neonate, shortly after the infant was discharged after a prolonged course of empirical antibiotic therapy secondary to maternal chorioamnionitis.

Salmonella Species

Non-Typhi Salmonella infection is an uncommon cause of sepsis and meningitis in neonates, but a significant proportion of cases of Salmonella meningitis occur in young infants. The CDC observed that approximately one third of 290 Salmonella isolates from CSF reported during 1968 to 1979 were from patients younger than 3 months, and more than one half were from infants younger than 1 year. A 21-year review of gram-negative enteric meningitis in Dallas, beginning in 1969, identified Salmonella as the cause in 4 of 72 cases. Investigators from Turkey reported 7 cases of neonatal meningitis caused by Salmonella during 1995 to 2001. Two of the five survivors developed communicating hydrocephalus, and one had a subdural empyema. Cases of neonatal meningitis caused by Salmonella enterica ser. Ancona, in which the pathogen was isolated simultaneously from the newborn’s CSF, parental fecal samples, and the mother’s breast milk ; S. enterica ser. Arizona meningitis in a 13-day-old girl ; and septicemia caused by S. Paratyphi B were recently reported.

Reed and Klugman reviewed 10 cases of neonatal typhoid that occurred in a rural African hospital. Six of the infants had early-onset sepsis with acquisition of the organism from the maternal genital tract, and 4 had late-onset infection with acquisition from a carrier or an environmental source. Two neonates developed meningitis, and 3 died. Recurrent S. entertidis meningitis in a neonate after a 3-week course of ceftriaxone and ciprofloxacin was recently described.

Neisseria Meningitidis

Although Neisseria meningitidis is a leading cause of bacterial sepsis and meningitis among children and adolescents, it rarely is associated with invasive infection in neonates. N. meningitidis may colonize the female genital tract and has been associated with pelvic inflammatory disease. The infant can be infected at delivery by organisms present in the maternal genital tract, or intrauterine infection can result during maternal meningococcemia. Meningococcal sepsis is rare in the neonate, but more than 50 cases (including 13 from the preantibiotic era) have been described. Early- and late-onset forms of meningococcal sepsis in neonates have been reported. Purpura similar to that of meningococcemia in older children has been observed in a 15-day-old and a 25-day-old infant.

Shepard and colleagues from the CDC reported 22 neonates with invasive meningococcal disease from a 10-year active, population-based surveillance of 10 states with diverse populations and more than 31 million persons. The average annual incidence was 9 cases per 100,000 people (vs. 973.8/100,000 for GBS). Sixteen patients had meningitis, and 6 of these also had meningococcemia. Six patients had early-onset disease. The overall mortality rate was 14%. Ten isolates were serogroup B, 4 were serogroup C, 3 were serogroup Y, 1 was nongroupable, and 4 were unavailable for analysis. Cases of meningococcal meningitis in infants successfully treated with no evidence of neurologic sequelae have been described.

Haemophilus Influenzae

Because of the introduction of H. influenzae type b conjugate vaccines in 1988, there has been a substantial decrease in the incidence in H. influenzae type b disease in infants and children in the United States and many other countries. Given the estimated proportion of individuals that are completely immunized, the decrease in H. influenzae type b invasive disease has exceeded expectations. The reduction in H. influenzae carriage associated with vaccination and the consequent decreased transmission from immunized children to unimmunized infants and children likely explains this effect.

Despite increased reporting of invasive infections caused by nontypeable H. influenzae in adults and older children, such infections in neonates remain uncommon. Five clinical syndromes have been associated with neonatal disease caused by H. influenzae : sepsis or respiratory distress syndrome, pneumonia, meningitis, soft tissue or joint infection, and otitis media or mastoiditis. The overall mortality rate was 5.5% for 45 cases reviewed by Friesen and Cho ; the mortality rate was 90% for 20 infants with a gestation lasting less than 30 weeks. Clinical and epidemiologic characteristics were similar to those of neonatal disease caused by GBS, including early- (within 24 hours of birth) and late-onset presentations, signs simulating respiratory distress syndrome, and a high mortality rate. Autopsy of infants with bacteremia related to nontypeable H. influenzae and signs of respiratory distress syndrome revealed hyaline membranes with gram-negative coccobacilli within the membranes, similar to findings of hyaline membranes caused by GBS. Examination of placentas from mothers of infants with sepsis caused by nontypeable H. influenzae revealed acute chorioamnionitis and acute villitis in some. H. influenzae also has been responsible for maternal disease, including bacteremia, chorioamnionitis, acute or chronic salpingitis, and tubo-ovarian abscess. Recently, a cluster of 8 cases of early-onset infections over 53 months caused by β-lactamase negative, nontypeable H. influenzae was reported from a NICU in Israel. In this series, a presentation resembling pneumonia, rather than classic respiratory distress syndrome, characterized the infants’ respiratory problems.

Neonatal sepsis caused by Haemophilus parainfluenzae and Haemophilus aphrophilus have also been reported.

Anaerobic Bacteria

Improvements in techniques for isolation and identification of the various genera and species of anaerobic bacteria have provided a better understanding of the anaerobic flora of humans and their role in disease. With the exception of Clostridium tetani and Clostridium botulinum, all of the anaerobic bacteria belong to the normal flora of humans. Anaerobes are present on the skin, in the mouth, in the intestines, and in the genital tract. They account for the greatest proportion of the bacteria of the stool. All are present in the intestines and have been isolated from the external genitalia or vagina of pregnant and nonpregnant women. Newborns are colonized with these organisms during or just before delivery. A literature review by Brook in 1990, on neonatal bacteremia caused by anaerobic bacteria, included 179 cases, with a mortality rate of 26%. Bacteroides and Clostridium spp. were the most common isolates. Predisposing factors for infection included premature rupture of membranes, preterm delivery, and NEC.

Anaerobic bacteria have been isolated from the blood of newborns with sepsis, from various organs at autopsy, from an infant with an adrenal abscess, from an infant with an infected cephalhematoma, and from infants with necrotizing fasciitis of the scalp associated with placement of a scalp electrode. Feder reviewed meningitis caused by Bacteroides fragilis; seven of nine reported cases occurred in neonates.

The incidence of neonatal sepsis caused by anaerobic bacteria remains uncertain, but recent data are available from some surveys that suggest the incidence is low (<5%). Noel and colleagues identified 29 episodes of anaerobic bacteremia in neonates in the intensive care unit (ICU) at New York Hospital during 18 years. Chow and coworkers analyzed 59 cases of neonatal sepsis associated with anaerobic pathogens and classified them into four groups: transient bacteremia after premature rupture of membranes and maternal amnionitis, sepsis after postoperative complications, fulminant septicemia (in the case of clostridial infections), and intrauterine death associated with septic abortion. The mortality rate associated with neonatal anaerobic sepsis reported in the literature ranges from 4% to 38%.

Serious infections of the bloodstream or CNS of neonates caused by Bacillus cereus have been reported, and in certain cases have proven intractable and refractory to antibiotic therapy. Magnetic resonance imaging of B. cereus meningoencephalitis reveals a pattern of hemorrhage and early cavitation accompanied by selective white matter destruction. An outbreak of B. cereus infections in a NICU was traced to contamination of balloons used in mechanical ventilation. Bacteroides fragilis has been identified as a cause of pneumonia, sepsis, or meningitis in the immediate newborn period.

Infections caused by Clostridium spp. can be localized, as in the case of omphalitis, cellulitis, and necrotizing fasciitis, or can manifest as sepsis or meningitis. Disease in neonates has been related to Clostridium perfringens, Clostridium septicum, Clostridium sordellii, Clostridium butyricum, Clostridium tertium, and Clostridium paraputrificum. The presenting signs usually are similar to those of other forms of bacterial sepsis. Chaney reported a case of bacteremia caused by C. perfringens in mother and child in which the neonate had classic features of adult clostridial sepsis, including active hemolysis, hyperbilirubinemia, and hemoglobinuria. Motz and colleagues reviewed five cases of clostridial meningitis caused by C. butyricum and C. perfringens . Clostridial sepsis is accompanied by a high mortality rate.

Neonatal Tetanus

Neonatal tetanus is caused by the gram-positive anaerobic spore-forming bacillus, C. tetani. The organism is present in soil and can be present in human and animal feces. Infection usually occurs after contamination of the umbilical stump. Maternal and neonatal tetanus are important causes of mortality in developing countries, claiming an estimated 180,000 lives annually. In the United States, tetanus in the newborn is exceedingly rare. Since 1984, only three cases of neonatal tetanus have been reported. The most recent case, reported from Montana in 1998, was an infant born to an unimmunized mother; the parents used a C. tetani –contaminated clay powder to accelerate drying of the umbilical cord. The use of this product had been promoted on an Internet site on “cord care” for use by midwives.

In many developing countries, both the incidence and mortality of neonatal tetanus remain startlingly high. Mustafa and colleagues conducted a retrospective neonatal tetanus survey among rural and displaced communities in the East Nile province in the Sudan and observed an incidence of neonatal tetanus of 7.1 cases per 1000 live births, more than double that reported from the stable rural community (3.2/1000). In both communities, coverage with two doses of tetanus toxoid was about 58%. Mortality attributable to neonatal tetanus in Djakarta in 1982 was 6.9 deaths per 1000 live births, and in the island provinces of Indonesia, it was 10.7 deaths per 1000 live births. Among 62 cases of neonatal tetanus in Ethiopia, 90% were born at home and 70% lacked antenatal care. Three quarters of infants in this series died in hospital, and risk factors for fatal outcome included an incubation period of less than 1 week, onset of symptoms at less than 48 hours, tachycardia, and fever. The mortality rate for neonates with tetanus in Lima, Peru was 45% and was not improved with use of intrathecal tetanus antitoxin. However, a meta-analysis of intrathecal therapy in tetanus suggested benefit in adults but not in neonates. A recent systematic review of prognostic factors in neonatal tetanus indicated that LBW and age of onset less than or equal to 5 to 7 days were crucial factors increasing the odds of death.

Application of contaminated materials to the umbilical cord is associated with deep-rooted customs and rituals in developing countries. A case-control study to identify risk factors for neonatal tetanus in rural Pakistan identified application of ghee (i.e., clarified butter from the milk of water buffaloes or cows) to the umbilical wound as the single most important risk factor. Although commercial ghee is available in Pakistan, the ghee used in rural areas is made at home from unpasteurized milk. Oudesluys-Murphy observed that application of some materials, including ghee and a stone wrapped in wet cloth, increased the risk of neonatal tetanus among Yoruba women but that other practices of cord care decreased the incidence, including searing of the cord with heat in China during the Ming dynasty and use of a candle flame to scar the cord in Guatemala.

Neonatal tetanus is a preventable disease; use of hygienic techniques at delivery and a program of tetanus toxoid immunization of children and young adults, particularly of pregnant women, are effective in eliminating this lethal disease. A systematic review of interventions to reduce neonatal tetanus mortality found vaccination of pregnant women with tetanus toxoid to be the key factor; in resource-poor countries such as Pakistan, this single intervention coupled with regular effective antenatal checkups and clean delivery practices effectively reduces neonatal tetanus.

Mixed Infections

Multiple organisms frequently are present in brain, liver, or lung abscesses; lung aspirate after pneumonia; or pleural empyema, but multiple organisms are found infrequently in cultures of blood or CSF. When several species are found, the significance of each is uncertain because it is possible that one or more of the organisms in a mixed culture is a contaminant.

Bacteremia with more than one organism occurs in patients with immunodeficiency, major congenital abnormalities, or contamination of a body fluid with multiple organisms, as is present in peritonitis, typically as a sequela of severe NEC in the VLBW infant. Neonatal meningitis caused by S. pneumoniae and Acinetobacter calcoaceticus and sepsis caused by P. aeruginosa and Yersinia enterocolitica have been reported. Although included in a series of cases of neonatal sepsis by some investigators, mixed cultures are not identified by most. Mixed infections were reported by Tessin and coworkers in 5% of 231 Swedish neonates, by Vesikari and associates in 4% of 377 Finnish infants, and by Bruun and Paerregaard in 7% of 81 Danish neonates. Faix and Kovarik reviewed the records of 385 specimens of blood or CSF submitted to the microbiology laboratories at the University of Michigan Medical Center from September 1971 to June 1986. More than one organism was present in 38 specimens from 385 infants in the NICU; 15 (3.9%) infants had multiple pathogens associated with clinical signs of sepsis or meningitis. The mortality was high (60%). Factors predisposing to mixed infection included prolonged rupture of membranes (>24 hours), total parenteral nutrition, NEC, presence of an intravascular catheter or ventriculostomy, and entities associated with multiple pathogens, including peritonitis, pseudomembranous colitis, and hepatic necrosis. Chow and colleagues reported polymicrobial bacteremia in eight newborns with anaerobic co-isolates or aerobic and anaerobic organisms in combination. An outbreak of polymicrobial bacteremia caused by K. pneumoniae and E. cloacae associated with use of a contaminated lipid emulsion was reported by Jarvis and colleagues.

Mixed infections also can include bacteria and viruses or bacteria and fungi, typically Candida, in the situation of intravascular central catheter or peritoneal infections associated with bowel perforation. Sferra and Pacini reported mixed viral-bacterial meningitis in five patients, including neonates with CSF isolates of enterovirus and GBS in a 10-day-old child and enterovirus and Salmonella in a 12-day-old child.

Uncommon Bacterial Pathogens

A large number of additional bacterial pathogens have been identified as rare or uncommon causes for neonatal sepsis and meningitis. These are listed in Table 6-7 with references, and were reviewed by Giacoia.

Table 6-7
Unusual Pathogens Responsible for Neonatal Sepsis and Meningitis
Organism Reference
Achromobacter spp.
Acinetobacter spp.
Bacillus anthracis
Bacillus cereus
Borrelia (relapsing fever)
Brucella spp.
Burkholderia cepacia
Burkholderia pseudomallei
Campylobacter spp.
Capnocytophaga spp.
Corynebacterium spp.
Edwardsiella tarda
Escherichia hermanii
Chryseobacterium (Flavobacterium) spp.
Gardnerella vaginalis
Helicobacter cinaedi
Lactobacillus spp.
Leptospira spp.
Leuconostoc spp.
Morganella morganii
Mycoplasma hominis
Ochrobactrum anthropi
Pantoea agglomerans
Pasteurella spp.
Plesiomonas spp.
Proteus mirabilis
Pseudomonas pseudomallei
Psychrobacter immobilis
Ralstonia pickettii
Rothia dentocariosa
Shigella sonnei
Staphylococcus capitis
Stomatococcus mucilaginosus
Vibrio cholerae
Yersinia enterocolitica
Yersinia pestis

Epidemiology

Incidence of Sepsis and Meningitis

The reported incidence of neonatal sepsis varies from less than 1 to 8.1 cases per 1000 live births.

References .

The increased use of intrapartum antibiotic prophylaxis for women with GBS colonization, with or without other risk factors associated with neonatal GBS disease, has been associated with a 70% reduction in the incidence of early-onset GBS sepsis to 0.44 per 1000 live births in 1999, a rate comparable to that of LOS (see Chapter 12 ).

The incidence of meningitis usually is a fraction of the number of neonates with early-onset sepsis. During the 8-year period from 1986 to 1994 at the Soroka University Medical Center in southern Israel, Greenberg and colleagues found incidences of neonatal bacterial sepsis and meningitis of 3.2 and 0.5 cases per 1000 live births, respectively. Certain pathogens that cause bloodstream invasion, such as GBS, E. coli, and L. monocytogenes, are more likely to be accompanied by meningeal invasion than others (e.g., S. aureus ). Meningitis is more frequent during the first month of life than in any subsequent period.

Characteristics of Infants who Develop Sepsis

Host susceptibility, socioeconomic factors, obstetric and nursery practices, and the health and nutrition of mothers are important in the pathogenesis of neonatal sepsis and meningitis. Infants who develop sepsis, particularly early-onset disease, usually have a history of one or more risk factors associated with the pregnancy and delivery that significantly increase the risk for neonatal infection. These factors include preterm delivery or LBW, premature rupture of membranes (i.e., rupture before the onset of labor), prolonged time of rupture of membranes, maternal peripartum infection, septic or traumatic delivery, and fetal hypoxia.

Birth Weight

The factor associated most significantly with enhanced risk for bacterial sepsis and meningitis in neonates is LBW (see Table 6-5 ). Infection is the most common cause of death in VLBW infants ; the diagnosis of early-onset sepsis in this population is associated with a threefold increase in mortality. However, with the exception of infection caused by GBS, it is unusual for a term infant to develop early-onset sepsis after an uneventful pregnancy and delivery. In a U.K. study, neonates born weighing less than 2000 g acquired meningitis six times more frequently than did infants weighing greater than 2000 g. The lower the infant’s birth weight, the higher is the incidence of sepsis (see Table 6-5 ). An Israeli study of 5555 VLBW infants documented the increased risk of LOS with decreasing birth weight; LOS occurred in 16.8% of neonates with a birth weight of 1250 to 1500 g, 30.6% of neonates weighing 1000 to 1249 g, 46.4% of those weighing 750 to 999 g, and 53% of those weighing less than 750 g at birth. In NICHD prospective disease surveillance from 2006 to 2009, the incidence of infection per 1000 live births was 0.57 for infants with birth weight greater than 2500 g, 1.38 for infants with birth weight 1500 to 2500 g, and 10.96 for infants of birth weight 400 to 1500 g.

Risk Factors of Infant and Mother

The relative importance of other factors associated with systemic infection in the newborn is more difficult to define. Greenberg and coworkers found that certain conditions were common in their prospective study of 229 infants with sepsis and meningitis: 130 (57%) were premature (<37 weeks of gestation), 64 (28%) were delivered by cesarean section or instrumental delivery, 43 (19%) had an Apgar score of less than 7 at 5 minutes, and 27 (2%) had a prolonged (>24 hours) interval after rupture of maternal membranes. Investigators in Pakistan found that maternal urinary tract infection and maternal fever, vaginal discharge, and vaginal examinations during labor were maternal factors significantly associated with neonatal early-onset sepsis, whereas low Apgar scores at birth and the need for endotracheal intubation were significant neonatal risk factors. Attack rates for early-onset sepsis are affected by birth weight, duration of rupture of membranes, and occurrence of maternal peripartum fever. Uterine or abdominal tenderness and/or maternal or fetal tachycardia are other suggestive signs (see Table 6-5 ).

Maternal fever during labor or after delivery suggests a concurrent infectious event in mother and infant, but noninfectious events may be responsible for maternal fever. Use of epidural analgesia for pain relief during labor is associated with increases in maternal temperature. Intrapartum fever of greater than 38° C (100.4° F) occurred an average of 6 hours after initiation of the epidural anesthesia in 14.5% of women receiving an epidural anesthetic, compared with 1.0% of women not receiving an epidural agent; the rate of fever increased from 7% in women with labors of less than 6 hours to 36% for labors lasting longer than 18 hours. There was no difference in the incidence of neonatal sepsis in the infants born to 1045 women who received epidural analgesia (0.3%), compared with infants born to women who did not have epidural analgesia (0.2%). Fetal core temperature may be elevated during maternal temperature elevation, and increased temperature may be present transiently in the neonate after delivery.

Ethnicity

The Collaborative Perinatal Research Study provides historical information on 38,500 pregnancies ; selected data for white and black women are presented in Tables 6-4 and 6-8 . Black women had a higher rate of premature rupture of membranes lasting more than 24 hours (21.4%), compared with white women (10.8%); black women had a higher rate of puerperal infection (4.1%), compared with white women (3.6%); and more black infants weighed less than 2500 g at birth (13.4%), compared with white infants (7.1%). Recent published data concurs with that observed 30 years ago. The National Center for Health Statistics reports continued disparities between blacks and whites in maternal and infant health indicators. In 2010, significant differences were found between non-Hispanic blacks and the general population in terms of neonatal mortality (11.46 vs. 6.14 deaths/1000 live births), LBW (13.6% vs. 7.7%), and preterm delivery less than 37 weeks’ gestation (17.1% vs. 12.0%). An earlier review of the literature from 1966 to 1994 reported significantly increased rates of severe histologic chorioamnionitis, maternal fever during labor, prolonged rupture of membranes, and early neonatal mortality from sepsis in blacks compared with whites.

Table 6-8
Selected Characteristics of Women, Their Pregnancies, and Newborns
Data from Niswander KR, Gordon M: The women and their pregnancies. The Collaborative Perinatal Study of the National Institute of Neurological Diseases and Stroke. U.S. Department of Health, Education and Welfare Publication No. (NIH) 73-379, Washington, DC, 1972, U.S. Government Printing Office.
Percent with Characteristics
Characteristic White Women Black Women
Premature Rupture of Membranes: Time From Rupture to Onset of Labor (hr)
<8 70.9 56.7
8-23 18.3 21.9
24-48 5.4 11.7
≥49 5.4 9.7
Puerperal Infection 3.6 4.1
Type of Delivery
Vaginal vertex 91.7 92.4
Vaginal breech 3.3 2.6
Cesarean section 4.9 5.0
Birth Weight < 2500 g 7.1 13.4
Neutrophilic Infiltration of
Amnion 9.0 7.9
Chorion 13.1 15.6
Umbilical vein 14.6 7.5

Approximately 18,700 white women and 19,800 black women were evaluated.

In a study of GBS disease in infants from the Atlanta metropolitan area, black infants had a higher incidence than nonblack infants of early-onset disease; the risk of late-onset disease was 35 times greater in black than in white infants. Thirty percent of early-onset disease and 92% of late-onset disease could be attributed to black race, after controlling for other significant risk factors, such as LBW and maternal age younger than 20 years. The increased incidence of GBS disease in blacks of all ages was observed in a survey by the CDC in selected counties in California, Georgia, and Tennessee and the entire state of Oklahoma. The rate of disease of 13.5 cases per 100,000 blacks was significantly higher than the 4.5 cases per 100,000 whites. In neonates with early-onset infection, 2.7 cases per 1000 live births occurred in blacks and 1.3 cases per 1000 live births occurred in whites. Maternal factors, such as socioeconomic status, nutrition, recently acquired sexually transmitted diseases, or racial differences in maternally acquired protective antibodies, may result in the increased risk of GBS disease among blacks.

Gender

Historical data have suggested that there is a predominance of male neonates affected by sepsis and meningitis but not by in utero infections ( Table 6-9 ). This difference partially may reflect the fact that female infants had lower rates of respiratory distress syndrome (i.e., hyaline membrane disease) than did male infants. Torday and colleagues studied fetal pulmonary maturity by determining lecithin-to-sphingomyelin ratios and concentrations of saturated phosphatidylcholine and cortisol in amniotic fluid of fetuses between 28 and 40 weeks of gestation. Female infants had higher indices of pulmonary maturity than did male infants. These data provide a biochemical basis for the increased risk of respiratory distress syndrome in male infants and the possible role of these factors of pulmonary maturation in the development of pulmonary infection. Later studies failed to confirm a significant increased risk for bacterial sepsis and meningitis among male infants.

Table 6-9
Incidence of Fetal and Neonatal Infections by Sex
Data based on a review of the literature and study of Johns Hopkins Hospital case records, 1930-1963. Washburn TC, Medearis DN J, Childs B: Sex differences in susceptibility to infections, Pediatrics 35:57, 1965.
No. of Infants
Infection Male Female Ratio of Male to Female
Intrauterine Infections
Syphilis 118 134 0.89
Tuberculosis 15 14 1.07
Toxoplasmosis 118 103 1.14
Listeriosis 26 37 0.70
Perinatal Sepsis
Gram-negative organisms 82 34 2.41
Gram-positive organisms 58 31 1.87
Perinatal Meningitis
Gram-negative organisms 126 44 2.87
Gram-positive organisms 45 39 1.15

Geographic Factors

The cause of neonatal sepsis varies from hospital to hospital and from one community to another. These differences probably reflect characteristics of the population served, including unique cultural features and sexual practices, local obstetric and nursery practices, and patterns of antimicrobial agent use. The bacteriology of neonatal sepsis and meningitis in western Europe

References .

and Jamaica is generally similar to that in the United States. In tropical areas, a somewhat different pattern can be observed. In Riyadh, Saudi Arabia, from 1980 through 1984, E. coli, Klebsiella, and Serratia spp. were the dominant causes of neonatal sepsis; group B streptococci were an infrequent cause. However, later data from this geographic location revealed E. coli and CoNS, respectively, were the most common pathogens causing early-onset and LOS.

Every year four million neonatal deaths occur. About one third of the deaths are due to sepsis. The highest numbers of neonatal deaths are in South Central Asian countries and sub-Saharan Africa. The global perspective of neonatal sepsis is discussed in Chapter 2 . The most common isolates responsible for neonatal sepsis vary by country but include a wide spectrum of gram-negative and gram-positive species, the most common of which are E. coli, S. aureus, Pseudomonas, and Klebsiella . Multidrug-resistant strains are an increasing threat to intervention programs. In a recent meta-analysis of 19 neonatal sepsis studies identified from 13 developing countries, Staphylococcus aureus , Klebsiella spp., and Escherichia coli accounted for 55% (39%-70%) of culture-positive sepsis on weighted prevalence.

GBS is the most frequent cause of early- and late-onset sepsis in the United States, but the rates and risk factors for maternal and neonatal GBS colonization and disease vary in different communities. Amin and colleagues in the United Arab Emirates evaluated 563 pregnant women from similar socioeconomic and ethnic backgrounds and reported a GBS colonization rate of 10.1%. In Athens, Greece, maternal and neonatal colonization rates were 6.6% and 2.4%, respectively, with a vertical transmission rate of 22.5%. Middle-class women followed in the private setting were more frequently colonized with GBS than those followed in a public hospital. No association was found between colonization with GBS and maternal age, nationality, marital status, previous obstetric history, cesarean section, infant birth weight, or preterm birth.

Stoll and Schuchat reviewed data on female genital colonization with GBS from 34 reports in the literature and emphasized the importance of appropriate specimen collection and inoculation into selective (antibiotic containing) broth media in the ascertainment of accurate colonization rates. Analysis of data from studies using adequate methods revealed regional GBS colonization rates of 12% in India and Pakistan, 19% in Asian and Pacific countries, 19% in sub-Saharan Africa, 22% in the Middle East and North Africa, and 14% in the Americas. A comparison of studies that did and did not use selective broth media revealed significantly higher GBS colonization rates in the populations where selective broth media was used to assess colonization. Other reasons for varying rates of GBS colonization and disease may include socioeconomic factors or differences in sexual practices, hygiene, or nutrition.

Socioeconomic Factors

The lifestyle pattern of mothers, including cultural practices, housing, nutrition, and level of income, appears to be important in determining infants at risk for infection. The most significant factors enhancing risk for neonatal sepsis are LBW and prematurity, and the incidence of these is inversely related to socioeconomic status. Various criteria for determining socioeconomic status have been used, but no completely satisfactory and reproducible standard is available. Maternal education, resources, and access to health care can affect the risk of neonatal sepsis. A CDC report evaluating the awareness of perinatal GBS infection among women of childbearing age in the United States revealed that women with a high school education or less; women with a household income of less than $25,000; and women reporting black, Asian/Pacific Islander, or other ethnicity had lower awareness of perinatal GBS infections than other women.

Procedures

Most VLBW infants have one or more procedures that place them at risk for infection. Any disruption of the protective capability of the intact skin or mucosa can be associated with infection. In a multicenter study of NICU patients, increased risk of bacteremia was associated with parenteral nutrition, mechanical ventilation, peripherally inserted central catheters, peripheral venous catheters, and umbilical artery catheters.

Nursery Outbreaks or Epidemics

The nursery is a small community of highly susceptible infants where patients have contact with many adults, including parents, physicians, nurses, respiratory therapists, and diagnostic imaging technicians (see Chapter 14 ). Siblings may enter the nursery or mothers’ hospital suites and represent an additional source of infection. In these circumstances, outbreaks or epidemics of respiratory and gastrointestinal illness, most of which is caused by nonbacterial agents, can occur. Spread of microorganisms to the infant occurs by droplets from the respiratory tracts of parents, nursery personnel, or other infants. Organisms can be transferred from infant to infant by the hands of health care workers. Individuals with open or draining lesions are especially hazardous agents of transmission.

Staphylococcal infection and disease are a concern in many nurseries in the United States (see Chapter 14 ). Epidemics or outbreaks associated with contamination of nursery equipment and solutions caused by Proteus spp., Klebsiella spp., S. marcescens, Pseudomonas spp., and Flavobacterium also have been reported. An unusual and unexplained outbreak of early-onset GBS sepsis with an attack rate of 14 per 1000 live births occurred in Kansas City during January through August of 1990.

The availability of molecular techniques to distinguish bacterial strains provides an important epidemiologic tool in the investigation of nursery outbreaks. Previously, methods to determine strain relatedness relied on antibiotic susceptibility patterns, biochemical profiles, and plasmid or phage analysis. More recent techniques permit the discrimination of strains based on bacterial chromosomal polymorphisms. Pulse-field gel electrophoresis, ribotyping, multilocus sequence typing, and polymerase chain reaction (PCR)-based methods are widely used tools to assign strain identity or relatedness.

Antimicrobial agents play a major role in the ecology of the microbial flora in the nursery. Extensive and prolonged use of these drugs eliminates susceptible strains and allows proliferation of resistant subpopulations of neonatal flora. There is selective pressure toward colonization by microorganisms that are resistant to the antimicrobial agents used in the nurseries and, because of cross-resistance patterns, to similar drugs within an antimicrobial class.

A historical example of the selective pressure of a systemic antimicrobial agent is provided by Gezon and coworkers in their use of benzathine penicillin G to control an outbreak of GAS disease. All infants entering the nursery during a 3-week period were treated with a single intramuscular dose of penicillin. Before institution of this policy, most strains of S. aureus in the nursery were susceptible to penicillin G. One week after initiation of the prophylactic regimen and for the next 2 years, almost all strains of S. aureus isolated from newborns in this nursery were resistant to penicillin G.

During a 4-month period in 1997, van der Zwet and colleagues investigated a nosocomial nursery outbreak of gentamicin-resistant K. pneumoniae in which 13 neonates became colonized and 3 became infected. Molecular typing of strains revealed clonal similarity of isolates from 8 neonates. The nursery outbreak was terminated by the substitution of amikacin for gentamicin in neonates when treatment with an aminoglycoside was believed to be warranted. Development of resistance in gram-negative enteric bacilli also has been documented in an Israeli study after widespread use of aminoglycosides.

Extensive or routine use of third-generation cephalosporins in the nursery, especially for all neonates with suspected sepsis, can lead to more rapid emergence of drug-resistant gram-negative enteric bacilli than occurs with the standard regimen of ampicillin and an aminoglycoside. Investigators in Brazil performed a prospective investigation of extended-spectrum β-lactamase (ESBL)-producing K. pneumoniae colonization and infection during the 2-year period from 1997 to 1999 in the NICU. A significant independent risk factor for colonization was receipt of a cephalosporin and an aminoglycoside. Previous colonization was an independent risk factor for infection. In India, Jain and coworkers concluded that indiscriminate use of third-generation cephalosporins was responsible for the selection of ESBL-producing, multiresistant strains in their NICU, where ESBL production was detected in 86.6% of Klebsiella spp., 73.4% of Enterobacter spp., and 63.6% of E. coli strains. Nosocomial infections in the nursery and their epidemiology and management are further discussed in Chapter 35 .

Pathogenesis

The developing fetus is relatively protected from the microbial flora of the mother. However, procedures disturbing the integrity of the uterine contents, such as amniocentesis, cervical cerclage, transcervical chorionic villus sampling, or percutaneous umbilical blood sampling, can permit entry of skin or vaginal organisms into the amniotic sac, causing amnionitis and secondary fetal infection.

Initial colonization of the neonate usually takes place after rupture of the maternal membranes. In most cases, the infant is colonized with the microflora of the birth canal during delivery. However, if delivery is delayed, vaginal bacteria may ascend the birth canal and, in some cases, produce inflammation of the fetal membranes, umbilical cord, and placenta. Fetal infection can then result from aspiration of infected amniotic fluid, leading to stillbirth, premature delivery, or neonatal sepsis. The organisms most commonly isolated from infected amniotic fluid are GBS, E. coli and other enteric bacilli, anaerobic bacteria, and genital mycoplasmas.

There are studies reporting that amniotic fluid inhibits the growth of E. coli and other bacteria because of the presence of lysozyme, transferrin, immune globulins (IgA and IgG but not IgM), zinc and phosphate, and lipid-rich substances. The addition of meconium to amniotic fluid in vitro has resulted in increased growth of E. coli and GBS in some studies. However, in other in vitro studies of the bacteriostatic activity of amniotic fluid, there is not inhibition of the growth of GBS. Further discussion of bacterial inhibition by amniotic fluid is available in Chapter 3 .

Infection of the mother at the time of birth, particularly genital infection, can play a significant role in the development of infection in the neonate. Transplacental hematogenous infection during or shortly before delivery (including the period of separation of the placenta) is possible, although it is more likely that the infant is infected just before or during passage through the birth canal. Among reports of concurrent bacteremia in mother and neonate are cases caused by H. influenzae type b, H. parainfluenzae, S. pneumoniae, GAS, N. meningitidis, Citrobacter spp., and Morganella morganii, and concurrent cases of meningitis have been reported as caused by S. pneumoniae, N. meningitidis, and GBS. Many neonates are bacteremic at the time of delivery, which indicates that invasive infection occurred antepartum. Infants with signs of sepsis during the first 24 hours of life also have the highest mortality rate. These data suggest the importance of initiating chemoprophylaxis for women with GBS colonization or other risk factors for invasive disease in the neonate at the time of onset of labor (see Chapter 12 ).

Microorganisms acquired by the newborn infant just before or during birth colonize the skin and mucosal surfaces, including the conjunctivae, nasopharynx, oropharynx, gastrointestinal tract, umbilical cord, and in the female infant, the external genitalia. Normal skin flora of the newborn includes CoNS, diphtheroids, and E. coli. In most cases, the microorganisms proliferate at the initial site of attachment without resulting in illness. On occasion, contiguous areas may be infected by direct extension (e.g., sinusitis and otitis can occasionally occur from upper respiratory tract colonization).

Bacteria can be inoculated into the skin and soft tissue by obstetric forceps, and organisms may infect these tissues if abrasions or congenital defects are present. Scalp abscesses can occur in infants who have electrodes placed during labor for monitoring of heart rate. The incidence of this type of infection in the hands of experienced clinicians, however, is generally quite low (0.1% to 5.2%). A 10-year survey of neonatal enterococcal bacteremia detected 6 of 44 infants with scalp abscesses as the probable source of their bacteremia. The investigators were unable from the data available to deduce whether these abscesses were associated with fetal scalp monitoring, intravenous infusion, or other procedures that resulted in loss of the skin barrier.

Transient bacteremia can accompany procedures that traumatize mucosal membranes, such as endotracheal suctioning. Invasion of the bloodstream also can follow multiplication of organisms in the upper respiratory tract or other foci. Although the source of bacteremia frequently is inapparent, careful inspection can reveal a focus, such as an infected circumcision site or infection of the umbilical stump, in some neonates. Metastatic foci of infection can follow bacteremia and can involve the lungs, kidney, spleen, bones, or CNS.

Most cases of neonatal meningitis result from bacteremia. Fetal meningitis followed by stillbirth or hydrocephalus, presumably because of maternal bacteremia and transplacentally acquired infection, has been described but is exceedingly rare. Although CSF leaks caused by spiral fetal scalp electrodes do occur, no cases of meningitis have been traced to this source. After delivery, the meninges can be invaded directly from an infected skin lesion, with spread through the soft tissues and skull sutures and along thrombosed bridging veins, but in most circumstances, bacteria gain access to the brain through the bloodstream to the choroid plexus during the course of sepsis. Infants with developmental defects, such as a midline dermal sinus or myelomeningocele, are particularly susceptible to invasion of underlying nervous tissue.

Brain abscesses can result from hematogenous spread of microorganisms (i.e., septic emboli) and proliferation in tissue that is devitalized because of anoxia or vasculitis with hemorrhage or infarction. Certain organisms are more likely than others to invade nervous tissue and cause local or widespread necrosis. Most cases of meningitis related to C. koseri (formerly C. diversus ) and E. sakazakii are associated with cyst and abscess formation. Other gram-negative bacilli with potential to cause brain abscesses include Proteus, Citrobacter, Pseudomonas, S. marcescens, and occasionally GBS. Volpe comments that bacteria associated with brain abscesses are those that cause meningitis with severe vasculitis.

Host Factors Predisposing to Neonatal Bacterial Sepsis

Infants with one or more predisposing factors (e.g., LBW, premature rupture of membranes, septic or traumatic delivery, fetal hypoxia, maternal peripartum infection) are at increased risk for sepsis. Microbial factors such as inoculum size and virulence properties of the organism undoubtedly are significant. Immature function of phagocytes and decreased inflammatory and immune effector responses are characteristic of very small infants and can contribute to the unique susceptibility of the fetus and newborn (see Chapter 4 ).

Metabolic factors are likely to be important in increasing risk for sepsis and severity of the disease. Fetal hypoxia and acidosis can impede certain host defense mechanisms or allow localization of organisms in necrotic tissues. Infants with hyperbilirubinemia can suffer impairment of various immune functions, including neutrophil bactericidal activity, antibody response, lymphocyte proliferation, and complement functions (see Chapter 4 ). The indirect hyperbilirubinemia that commonly occurs with breastfeeding jaundice rarely is associated with neonatal sepsis. Late-onset jaundice and direct hyperbilirubinemia can be the result of an infectious process. In one study from Turkey, more than one third of infants with late-onset direct hyperbilirubinemia had culture-proven sepsis, with gram-negative enteric bacteria, including E. coli, the most common etiologic agents. Evidence of diffuse hepatocellular damage and bile stasis have been described in such infected and jaundiced infants.

Hypothermia in newborns, generally defined as a rectal temperature equal to or less than 35° C (95° F), is associated with a significant increase in the incidence of sepsis, meningitis, pneumonia, and other serious bacterial infections. In developing countries, hypothermia is a leading cause of death during the winter. Hypothermia frequently is accompanied by abnormal leukocyte counts, acidosis, and uremia, each of which can interfere with resistance to infection. However, the exact cause of increased morbidity in infants presenting with hypothermia remains poorly understood. In many infants, it is unclear whether hypothermia predisposes to or results from bacterial infection. For example, in a large outbreak of S. marcescens neonatal infections affecting 159 cases in Gaza City, Palestine, hypothermia was the single most common presenting symptom, recorded in 38% of cases.

Infants with galactosemia have increased susceptibility to sepsis caused by gram-negative enteric bacilli, in particular E. coli. Among 8 infants identified with galactosemia by routine newborn screening in Massachusetts, 4 had systemic infection caused by E. coli. Three of these 4 infants died of sepsis and meningitis; the fourth infant, who had a urinary tract infection, survived. A survey of state programs in which newborns are screened for galactosemia revealed that among 32 infants detected, 10 had systemic infection, and 9 died of bacteremia. E. coli was the infecting organism in 9 of the infants. It appears that galactosemic neonates have an unusual predisposition to severe infection with E. coli, and bacterial sepsis is a significant cause of death among these infants. Depressed neutrophil function resulting from elevated serum galactose levels is postulated to be a possible cause of their predisposition to sepsis. The gold standard for diagnosis of classic galactosemia is measurement of galactose-1-phosphate uridyltransferase (GALT) activity in erythrocytes, and the sole therapy is galactose restriction in the diet. Shurin observed that infants became ill when serum galactose levels were high and when glucose levels were likely to be low, and that susceptibility to infection diminished when dietary control was initiated.

Other inherited metabolic diseases have not been associated with a higher incidence of neonatal bacterial infection. A poorly documented increase in the relative frequency of sepsis has been observed among infants with hereditary fructose intolerance. Infants with methylmalonic acidemia and other inborn errors of branched-chain amino acid metabolism manifest neutropenia as a result of bone marrow suppression by accumulated metabolites; however, no increased incidence of infection has been described in this group of infants.

Iron may have an important role in the susceptibility of neonates to infection, but this continues to be controversial. Iron added to serum in vitro enhances the growth of many organisms, including E. coli, Klebsiella spp., Pseudomonas spp., Salmonella spp., L. monocytogenes, and S. aureus. The siderophore IroN is a proven virulence factor for the bacteremic phase of E. coli K1 sepsis and meningitis in the neonatal rat infection model. Iron-binding proteins, lactoferrin and transferrin, are present in serum, saliva, and breast milk. However, the newborn has low levels of these proteins. The iron sequestering capacity of oral bovine lactoferrin supplementation may be one contributing factor to its reported efficacy in prophylaxis of bacterial sepsis in VLBW infants.

Barry and Reeve demonstrated an increased incidence of sepsis in Polynesian infants who were treated with intramuscular iron as prophylaxis for iron-deficiency anemia. The regimen was shown to be effective in preventing anemia of infancy, but an extraordinary increase in bacterial sepsis occurred. The incidence of sepsis in newborns receiving iron was 17 cases per 1000 live births, whereas the incidence of sepsis in infants who did not receive iron was 3 cases per 1000 live births; during a comparable period, the rate of sepsis for European infants was 0.6 cases per 1000 live births. Special features of sepsis in the infants who received iron soon after birth were late onset, paucity of adverse perinatal factors, and predominance of E. coli as the cause of sepsis. During the period studied, E. coli was responsible for 26 of 27 cases of sepsis in iron-treated Polynesian infants and for none of three cases of sepsis in the infants who did not receive iron. Results of this study were similar to the experience reported by Farmer for New Zealand infants given intramuscular iron. The incidence of meningitis caused by E. coli increased fivefold in infants who received iron and decreased when the use of iron was terminated. Conventional iron-supplemented human milk fortifiers, however, appear safe and do not contribute to a higher rate of sepsis in preterm infants.

Infection in Twins

Studies have suggested a higher risk for contracting ascending intrauterine infection in the first than the second born of twins. Comparing delivery methods, no difference was observed in the incidence of neonatal sepsis in twins delivered in the vertex/vertex position when compared with cases requiring uterine manipulation (vertex/breech extraction). However, vaginal delivery of twin A, followed by cesarean delivery of twin B, may be associated with a higher rate of endometritis and neonatal sepsis when compared with cases where both twins are delivered by cesarean section.

A recent large study from the NICHD Neonatal Research Network conducted from 2002 to 2008 identified LOS occurring in 25.0% (3797/15,178) of singleton and 22.6% (1196/5294) of multiple-birth infants with VLBW (401-1500 g). CoNS accounted for 53.2% of episodes in singletons and 49.2% in multiples. A similar concordance of LOS in same-sex and unlike-sex twin pairs suggested that susceptibility to LOS among VLBW infants is not genetically determined. No difference in complications or sepsis risk exists among twin pregnancies conceived spontaneously or through in vitro fertilization.

Edwards and colleagues studied GBS infection in 12 index cases of multiple gestations. Early-onset disease occurred in both twins in one pair and in one twin in five other pairs; late-onset infection occurred in both infants in two pairs and in one twin in four other pairs. Cases of late-onset GBS disease in twin pairs occurred closely in time to one another: 19 and 20 days in one set and at 28 and 32 days of age in the other set. In another case report of late-onset GBS infection in identical twins, twin A suffered fulminant fatal meningitis whereas twin B recovered completely. The GBS isolates proved to be genetically identical; clinical variables associated with the adverse outcome in twin A were longer duration of fever before antibiotics and the development of neutropenia. In twins, the presence of virulent organisms in the environment, especially the maternal genital tract; their absence of specific maternal antibodies; and their similar genetic heritage probably contribute to the risk for invasive infection. It seems logical that twins, particularly if monochorionic, should have high rates of simultaneous early-onset infection, but it is particularly intriguing that some cases of late-onset disease occur in twins almost simultaneously. However, the incidence of infection in preterm twins co-bedding in the nursery did not differ from those cared for in separate beds.

Infections in twins, including disease related to Toxoplasma pallidum, echoviruses 18 and 19, and Toxoplasma gondii, are discussed in Chapter 16, Chapter 25, Chapter 31 , respectively. Examples of neonatal infections in twins include those caused by GAS (case report of streptococcal sepsis in a mother and infant twins), Salmonella spp., C. koseri (brain abscesses in twins, malaria, coccidioidomycosis, cytomegalovirus infection, and rubella ).

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here