Immaturity of the newborn immune system and defects in structural defenses make the neonate, especially the neonate born prematurely, uniquely susceptible to infection. The skin, for example, normally provides a mechanical barrier between the host and the environment. In infants born before 32 weeks of gestation, the stratum corneum is poorly developed, and the skin is fragile, very permeable, and easily traumatized by routine procedures such as cleansing or removal of adhesive tape. Injured skin provides a portal of entry for infectious agents. Similar defects are seen in the alimentary tract, where low levels of mucosal immunoglobulin A, high gastric pH, and short gastric emptying times increase the susceptibility of the newborn to gastrointestinal infections.

Immune function in the newborn has been extensively reviewed elsewhere (25). Term and preterm newborns have functioning B cells, but there is little antibody synthesis in utero. Postnatally, the B cells make antibodies to protein antigens but respond poorly to polysaccharide antigens, including the bacterial capsular polysaccharides of group B streptococcus (GBS) and Haemophilus influenzae. In the first weeks of life, the newborn depends on passively transferred maternal antibody and the repertoire of antibodies received depends on maternal exposure. Because placental transfer of antibody occurs in the third trimester, infants born at <34 weeks have low levels of immunoglobulin G antibody.

The newborn has a high total T-lymphocytic count, but phenotypic surface markers differ from those in the older child. Cytotoxic T-cell activity is decreased, as is T-cell helper function. The T-cell-dependent antigen-specific response is delayed, and there is limited production of several cytokines. These maturational defects in T cell function enhance the susceptibility of the newborn to intracellular pathogens such as Listeria, Toxoplasma, and Salmonella. Natural killer cell activity, important in control of herpes group viral infections, is also decreased.

Reduced numbers and activity of alveolar macrophages in the lungs of term and preterm infants increase the risk for pulmonary infection. The newborn has a decreased granulocyte storage pool and defective neutrophil and monocyte chemotaxis. Neutrophil phagocytosis and antimicrobial activity are largely intact but may be decreased when bacterial density is high or when opsonins are limited. Although production of complement proteins begins early in gestation, mature activity of the complement system may be delayed until 6 to 10 months of age.

The newborn may develop infection as a result of exposure to maternal flora during labor or delivery, or postpartum from maternal, hospital, or community sources. Postnatally, the hands of healthcare workers are the most common vehicles for transmission of potential pathogens in neonatal units (26,27). Nursery outbreaks of Staphylococcus aureus, enterococcus, a variety of gram-negative bacilli, and viruses have been attributed to hand transmission (1,28, 29, 30, 31, 32, 33, 34 and 35). In one study, gram-negative bacilli were found on the hands of 75% of NICU personnel (36). Usually, hands are transiently contaminated, and hand washing removes the microorganisms and interrupts transmission (37). A few outbreaks have been linked to microbial contamination of hand washing agents (38,39).

Occasionally, personnel who are persistent carriers of potential pathogens such as S. aureus or group A Streptococcus (GAS) have been implicated in nursery outbreaks (40,41). Artificial fingernails may lead to increased hand carriage of gram-negative microorganisms, and healthcare workers with artificial nails have been linked to transmission of Pseudomonas aeruginosa (29,30) and Klebsiella pneumoniae (42).

Patient care equipment may also serve as a vehicle for transmission. Multiple outbreaks have been associated with contaminated respiratory care equipment including ventilator circuits (43,44) laryngoscopes (45), balloons used for manual ventilation (46) and suction devices (47, 48 and 49). Inadequate disinfection of rectal thermometers contributed to nursery outbreaks of Salmonella eimsbuettel (50) and Enterobacter cloace (51).

Infusion of contaminated intravenous fluids, including total parental nutrition solutions and lipid emulsions, may result in bacteremia or meningitis (52, 53, 54, 55, 56, 57, 58, 59 and 60). Exposure to contaminated topical preparations and medications, including contaminated eyewash (61), umbilical cord wash (62), and glycerin (63) may also result in invasive infections. Use of contaminated ultrasound gel resulted in an outbreak of pyoderma in hospitalized neonates (64), while bathing practices have been linked in clusters of listeriosis (65) and Stenotrophomonas infections (66). Blood transfusions may be a source of viruses such as hepatitis A virus, hepatitis B virus (HBV), and hepatitis C virus (HCV) (67, 68 and 69). Before current screening practices, blood products were also a source for perinatal cytomegalovirus (CMV) transmission (70,71). Neonatal transfusionacquired malaria has been rarely reported (72,73).

Hospitalized neonates are at risk for food-borne infection. Powdered formula is not sterile, and feeding of reconstituted formula has been associated with gram-negative
bacteremia and meningitis (74). Expressed breast milk may be contaminated during collection (75, 76 and 77) and both breast milk (78, 79 and 80) and formula (81,82) may be contaminated during storage and handling. Feeding practices may also contribute to infection. Pathogens may be acquired during insertion or handling of nasogastric tubes used for feeding (35,83,84). Feedings administered though nasogastric tubes by continuous infusion remain at room temperature for several hours, creating the potential for microbes to proliferate in the reservoir or tubing during infusion (77).

Infected personnel and visitors may introduce pathogens into the nursery, especially during community outbreaks of viral infections (85). Both symptomatic healthcare providers (86) and visitors have been the source of healthcare-associated pertussis in the NICU.

Infection rates in the NICU increase with overcrowding and understaffing. In one report, a 16-fold increase in outbreaks of S. aureus infection was observed when the infant-tonurse ratio exceeded 7 and there was a sevenfold increase when the nursery was crowded (87). Increasing rates of endemic methicillin-resistant S. aureus (MRSA) were also linked to overcrowding and understaffing, with eradication of MRSA when these conditions improved (88). An outbreak of E. cloacae infection was associated with understaffing and overcrowding in another report (89).

Any procedure that disrupts the normal barriers to infection may predispose the newborn to infection. Scalp electrodes, for example, provide a portal of entry for maternal genital microorganisms. Although infectious complications occur in <1% of infants and most are benign abscesses, occasionally severe cellulitis, bacteremia, osteomyelitis, and disseminated herpes simplex virus (HSV) occur (90,91). Osteomyelitis has resulted from infected toe and heel punctures and femoral venipunctures. Surgicalsite infections (SSIs), as well as device-associated infections including catheter-associated bacteremia, bladder catheter-associated urinary tract infection, and ventilatorassociated pneumonia (VAP), are discussed in detail below.

Pustules, cellulitis, subcutaneous abscesses, lymphadenitis, and infections at sites of percutaneous punctures are most often due to S. aureus, or less commonly to streptococci and gram-negative bacilli and other microorganisms (92). Microbes causing infections at scalp monitor sites are more diverse and include maternal genital microorganisms such as HSV (90,91).

Omphalitis is uncommon, occurring in 0.5% of term and 2% of preterm infants in one report (93). The presentation varies from mild erythema or serous drainage to purulent discharge, cellulitis, and acute necrotizing fasciitis of the abdominal wall. S. aureus is most often isolated, but GAS, coagulase-negative staphylococci (CONS), enterococci, gram-negative rods, and anaerobes may also be involved (94). A mortality rate of 7% was reported in a series, with fatalities associated with rapidly progressing cellulitis or necrotizing fasciitis (95).

Mastitis, characterized by redness, swelling, or induration of the breast, is seen occasionally in term infants in the first 3 weeks of life (96). S. aureus is the most common pathogen, although disease due to gram-negative microorganisms such as Escherichia coli and Proteus mirabilis is also reported (97).

Circumcision is the most common surgical procedure performed in the newborn, although NHSN considers circumcision infections with skin and soft-tissue infections rather than with SSIs. Reported infection rates are low, at 0.06% to 0.4% (98,99). Most are simple skin infections, but more serious complications, including necrotizing fasciitis, have occasionally been reported (100). An outbreak of neonatal pustulosis due to a community-associated MRSA strain in one newborn nursery was attributed in part to circumcision practices (101).

Healthcare-associated conjunctivitis is common in NICUs (102,103). The conjunctivae of neonates may become colonized with nasopharyngeal and skin flora during routine care. The immature lacrimal system of preterm neonates facilitates pooling of bacteria and other debris on the surface of the eye, leading to conjunctival infection. The most common bacterial pathogens associated with healthcare-associated conjunctivitis include CONS, S. aureus, Klebsiella species, Pseudomonas aeruoginosa, Serratia marscescens, H. influenzae, and E. coli (102). Viral conjunctivitis also occurs. Risk factors for conjunctivitis include low birth weight and the need for respiratory support, including nasal CPAP and mechanical ventilation. P. aeruginosa conjunctivitis in particular has been associated with contaminated resuscitation equipment (47), and infection in intubated patients has been related to endotracheal tube colonization and eye contamination during suctioning (104). Opthalmalogic exam for retinopathy of prematurity has been associated with the development of conjunctivitis, including outbreaks of adenoviral conjunctivitis (105,106).

Bloodstream infections (BSIs) are the most common HAI in hospitalized neonates, and most are related to central venous catheters (CVCs) (13,14,22). While there is significant variability in infection rates among NICUs, higher rates of device-associated bloodstream infection are consistently reported among the smallest, most premature infants (24). The pooled mean rates of umbilical catheterassociated BSIs in Level II/III NICUs that reported data to NHSN from 2006 to 2008 ranged from 1/1,000 catheter days in infants with birth weights of >2,500 g to 5.7/1,000 catheter/days in infants with birth weights ≤750 g. Central line-associated BSIs (CLABSIs) for the same time period ranged from 1.2/1,000 catheter days for the largest infants to 4.9/1,000 catheter days in infants in the smallest birthweight category. Even higher rates of catheter-associated BSIs are reported by NICUs in resource-limited settings (107). NHSN does not further stratify CLABSIs by type of catheter, but some data suggest lower rates of infection with percutaneously inserted central catheters (PICCs) compared with tunneled catheters (108), while a multicenter study showed no difference (109).

Coagulase-negative staphylococci, S. aureus, enterococci and Candida species are the most common causes of CLABSIs (110). Among neonates with PICCs, 75% of CLABSIs are caused by CONS (111), although differentiating true CONS bacteremia from culture contamination is challenging. Increasingly, S. aureus CLABSIs are due to MRSA. From 1998 to 2008, the rates of MRSA CLABSIs reported to NHSN rose 49% (112).

In additional to low birth weight and gestational age, mechanical ventilation (113,114) and total parenteral nutrition (114) are risk factors for the development of CLABSIs. In one study involving 3,470 catheter days, sampling of blood through the central line and disconnection of the catheter increased the risk of CLABSIs (115). The importance of catheter location on the incidence of infection remains controversial. One study identified a higher rate of catheter-related sepsis in VLBW infants with femoral PICCs compared to those with nonfemoral PICCs (22.5% vs. 12 or 10.9 vs. 6.8 episodes/1,000 catheter days) (116). Another study found no difference in infection rates when PICCs placed in an upper extremity were compared to those placed in a lower extremity, although the pathogens associated with infections did vary by site (117). Coagulasenegative staphylococcal infections were more common in infants with upper extremity catheters, while more gramnegative infections were diagnosed in children with lower extremity catheters.

Some (118) but not all studies have demonstrated that the risk of infection increases with the duration of catheter use. Umbilical catheterization for more than 5 days is recognized as an independent risk factor for sepsis (113). In one study, catheter duration of 10 to 21 days was associated with a 40-fold risk of gram-negative BSIs in VLBW infants, while the risk was 90-fold with catheter duration of >21 days (119). In a retrospective cohort study in a single NICU, the incidence rate of CLABSIs increased by 14% by day during the first 18 days after PICC insertion. The trend reversed from days 19 to 35 after insertion, but after day 36, the incidence rate of CLABSIs increased by 33% per day.

Healthcare-associated BSIs can also occur in the absence of a central venous catheter. When all causes of late-onset sepsis are considered, the pathogens are similar to those observed with CLABSIs (19,114). Although grampositive infections still account for the majority of infections, the proportion of BSIs caused by gram-negative microorganisms is higher. In patients with short gut syndrome, enteric gram-negative bacilli and yeast predominate, probably because of translocation of bacteria from the gastrointestinal tract and subsequent seeding of the catheter (120). Nasal cannula continuous positive pressure airway use and H₂ blocker/proton pump inhibitor therapy are also risk factors for gram-negative BSIs in VLBW infants (119).

The burden of healthcare-associated meningitis in NICU infants is difficult to quantify, as not all infants with symptoms of infection undergo lumbar puncture. In a study of 9,641 VLBW infants born at NICHD Neonatal Research Network Centers, only 1.4% infants who survived at least 3 days developed meningitis, representing 5% of infants who underwent lumbar puncture (121). Notably, only half of the infants who had a blood culture obtained because of suspected sepsis also had a lumbar puncture performed, suggesting that some cases of meningitis could have been missed. One-third of infants with meningitis had negative blood cultures.

The most frequent pathogens isolated in VLBW infants with healthcare-associated meningitis include CONS (29%), Candida species (14%), and enterococci (13%). In the NICHD study, gram-negative microorganisms, including E. coli, Klebsiella, and Serratia species, accounted for 19% of episodes. Multiple NICU outbreaks of gram-negative meningitis have been reported (56,122,123,124, 125, 126 and 127), some of which have been linked to feeding practices. A cluster of Elizabethkingia meningosepticum (formerly Flavobacterium meningosepticum) was linked to contamination of the formula preparation area and bottle stoppers (128). Cases of Chronobacter sakazakii (formerly Enterobacter sakazakii) have been linked to contaminated powdered infant formula (129).

Risk factors for gram-negative meningitis include underlying urinary tract anomalies and hydrocephalus (130). Meningitis may also occur as a complication of ventricular drain and shunt placement (131). In a study of shunt placement for hydrocephalus in newborns weighing <2,000 g, the shunt infection rate was 25% after primary placement and 36% after revision (132). Common causes of shunt infection include CONS, S. aureus, and gram-negative bacilli (133). Young age, prematurity, and the presence of intraventricular hemorrhage, a known complication of prematurity, are associated with an increased risk of shunt infection. In one study, prematurity increased the risk for shunt infection nearly fivefold. Proposed reasons for this increased infection risk may include a high density of CONS on the skin of preterm infants and colonization with more virulent CONS strains. Strategies for the prevention of ventriculoperitoneal (VP) shunt infections are similar to those proposed for the prevention of other SSIs. The use of antibiotic-impregnated shunt systems reduced infections in one small case series of premature infants and deserves further study (133).

Early-onset pneumonia is usually related to intrapartum exposure and is most often due to GBS (134). Maternally transmitted Ureaplasma urealyticum infection occasionally causes pneumonia or respiratory distress in VLBW, premature infants (135).

Most cases of late-onset pneumonia are device-related (134). VAP is common in preterm infants, many of whom require prolonged ventilatory support. In a prospective study, nearly 30% of infants born at <28 weeks of gestation developed VAP; infection was significantly associated with mortality (136). Not surprisingly, the rates of VAP reported through NHSN are consistently highest in infants of <1,000 g birth weight and are similar to those reported in pediatric intensive care units and adult medical intensive care units (24). Pooled mean rates of VAP for level II/III NICUs reporting data to NHSN from 2006 to 2008 ranged from 0.6/1,000 ventilator days in infants of >2,500 g birth weight to 2.7/1,000 ventilator days in infants of ≤750 g birth weight.

The true burden of VAP may not be known, because the diagnosis is difficult in NICU infants. While NHSN surveillance definitions for VAP include criteria for infants <1 year of
age, these definitions have not been validated in premature neonates. Comorbid conditions such as bronchopulmonary dysplasia may mimic the clinical and radiographic features of VAP (137). Ultimately, many infants are treated empirically for presumed pulmonary infection.

Gram-negative microorganisms are thought to cause at least 30% of VAP episodes, but a specific microbiologic diagnosis is often difficult unless there is secondary bacteremia (14,22). Endotracheal cultures are rarely useful in the diagnosis of VAP because the respiratory tract of the intubated newborn rapidly becomes colonized with mixed gram-positive flora by the second week of mechanical ventilation, and gram-negative microorganisms after the fourth week (138,139). The presence of purulence in a specimen suctioned from the endotracheal tube of a mechanically ventilated neonate has a poor positive predictive value for respiratory tract infection, including pneumonia and tracheitis (139). Risk factors for VAP in NICU patients include duration of mechanical ventilation, reintubation, treatment with opiates, and endotracheal suctioning (140). Other modes of assisted ventilation, including nasal continuous positive airway pressure, have been associated with the development of healthcare-associated pneumonia, albeit at lower rates than those associated with mechanical ventilation (141).

Gastrointestinal infections account for ˜1% of infections in NICUs. Microorganisms that cause community-associated outbreaks of gastrointestinal infection can also cause sporadic infections or clusters of infection in the NICU. Rotavirus is most commonly identified (142), although outbreaks of norovirus and adenovirus have also been reported.

Outbreaks of bacterial enteritis are occasionally reported. Salmonella outbreaks in neonatal units may be prolonged, in part due to the prolonged incubation period of this microorganism, a high proportion of asymptomatic infants, and prolonged shedding by carriers (143,144). Widespread contamination of the environment and equipment may also facilitate ongoing transmission (145). The newborn with Salmonella gastroenteritis is at risk for bacteremia and focal infections are often seen. Symptomatic Shigella infection in the newborn is rare and transmission to other newborns is unusual; transmission to nursery personnel has been reported (146). Campylobacter jejuni is an uncommon newborn pathogen and is usually of maternal origin, but healthcare-associated outbreaks have been described (147,148).

Necrotizing enterocolitis (NEC) occurs in 1% to 8% of all NICU admissions (149, 150 and 151). It is a disease of prematurity, with preterm infants comprising 90% or more of all cases (92). Rates of NEC are highest in infants born at 30 to 32 weeks of gestation and disease is most common in the second week of life (152,153). The etiology of this disorder is not well understood but is thought to involve vascular compromise, bacterial invasion, and release of inflammatory mediators in the setting of a substrate such as enteral feedings.

While NEC is not strictly an infectious process, it is included in the CDC definitions of HAIs. Outbreaks in nurseries have been temporally associated with the isolation of a number of different microbes, including Klebsiella species; Clostridia species; E. coli species; Serratia species; Pseudomonas species; Staphylococcus epidermidis, Salmonella species, toxigenic E. coli, and S. aureus (154, 155 and 156). Because these microorganisms are commonly found in patients without NEC, proving a causal association has been difficult and association of specific microorganisms with clusters of NEC may reflect patterns of intestinal colonization rather than actual outbreaks of infection. The role of viruses in the pathogenesis of NEC is not well understood. Adenovirus (157), norovirus (158), coronavirus (159), and rotavirus (160,161) have been recovered in infants with NEC but causality remains uncertain.

Prematurity is the principal risk factor for NEC, along with several factors related to infant feeding. The risk of NEC is 6 to 20 times higher in formula-fed infants than in breast milk-fed infants. While breast milk is generally protective, consumption of dry powdered human milk fortifier has been identified as a risk factor for NEC (162,163). One study of >15,000 infants that utilized administrative data identified exposure to antenatal steroids as a risk factor for NEC independent of birth weight (151).

A number of strategies have been evaluated for the prevention of NEC. Enteral probiotic supplementation reduces the risk of severe NEC and death in preterm infants with birth weights of at least 1,000 g (164). In one study, oral immunoglobulin containing immunoglobulin A and immunoglobulin G was protective when fed to low-birth-weight infants for whom breast milk was not available (165). Prophylactic oral vancomycin protected VLBW infants against NEC and may be indicated in specific situations, but routine use may increase the risk of colonization with resistant microorganisms (166). Outbreaks of NEC, even those not temporally associated with isolation of a particular microbe, have been aborted with typical infection control interventions such as cohorting, environmental cleaning, and enhanced hand hygiene (167).

Some studies suggest that healthcare-associated urinary tract infections (HA-UTIs) are uncommon in NICU patients. In a national point prevalence survey of NICU HAIs in the United States, urinary tract infection occurred in only 10/827 patients (14). Most infections occurred in infants with birth weights of ≤1,000 g or >2,500 g. Similar rates were reported from a multicenter trial in Europe: only 0.3% of infants developed HA-UTIs (13). Single center studies report rates of HA-UTIs ranging from 11.6 to 19/1,000 admissions (168, 169 and 170). Higher rates are reported in infants undergoing surgery (170) and in VLBW infants (171). In a retrospective review of NICU infants weighing ≤1,500 g at birth, HA-UTIs developed in 8.1%.

The most common bacterial pathogens causing HAUTIs in NICU patients include E. coli, Enterobacter, and Klebsiella species. In one 6-year retrospective study, Candida species were the most common cause of HA-UTIs (169). Compared to bacterial UTIs, candidal UTIs occurred in less mature and younger (34 vs. 97 days) infants, and were more likely to be associated with bloodstream infection.

Most infants who develop HA-UTIs do not have underlying renal anomalies or vesicoureterial reflux. Exposure to
intermittent or indwelling urinary catheters is a well-described risk factor in pediatric patients (168). Although catheters are used less frequently in newborns than in older children and adults, they carry a high risk of infection. This may be because of the use of feeding tubes, which are not well stabilized for bladder drainage rather than the balloon-tipped catheters used in older children (172).

Neonates are at increased risk for SSIs compared to children and adults (173,174). In a study involving 1,094 neonates and 1,433 surgical procedures, 17% of patients developed a wound infection (175). Staphylococci were most commonly isolated. Increased incision length, increased duration of surgery, and contamination of the operative site were all associated with an increased risk of infection but gestational age and birth weight were not. Rates of SSI are likely to vary by patient population, type of surgery, surgical site infection risk, severity of illness, and duration of operation. At present, NHSN does not provide risk-stratified, comparative data for neonates who develop SSIs, but such data are needed.

Guidelines for the prevention of SSI have been published. Although details of SSI prevention in pediatric surgical practice are not addressed in these guidelines, it is noted that SSI prevention measures effective in adult surgical care are indicated in pediatric surgical care (176). Use of prophylactic antibiotics was associated with a lower infection rate after potentially contaminated surgery but not after clean surgery (177). Unfortunately, there are few data on the efficacy of antibiotic prophylaxis for surgical procedures in the neonate (175) and detailed recommendations that address issues unique to this population are lacking. The current wound classification system used for determining the need for antibiotic prophylaxis (clean, clean contaminated, contaminated, and dirty or infected) was created based on adult surgeries, and it is unclear how these guidelines apply to neonatal operations (178). The American Academy of Pediatrics (AAP) recommends systemic antibiotic prophylaxis “when the probability or morbidity of postoperative infection is high and the benefits of preventing wound infection outweigh potential risks for adverse drug reactions and emergence of resistant microorganisms” (179). Specifically, prophylaxis is indicated for some clean wound surgeries and most clean contaminated wound surgeries (Table 52-1). Surgeries that involve contaminated wounds or dirty and infected wounds require treatment rather than prophylaxis. Ampicillin and gentamicin are recommended agents for surgical prophylaxis in all neonates <72 hours of age. For older neonates, prophylaxis is targeted to colonizing microorganisms, healthcare-associated microorganisms, and the operative site. Unfortunately, there is little consensus among pediatric surgeons about how these recommendations apply to specific neonatal surgical procedures, which agents should be used, and what duration of therapy is appropriate (180).

Neonatal osteomyelitis and septic arthritis are uncommon manifestations of HAI, with only one to three cases reported per 1,000 admissions to intensive care nurseries (181). Both are usually hematogenous in origin and, as such, are most often caused by the same agents that cause neonatal sepsis, including GBS, S. aureus, a variety of gramnegative bacilli, and Candida species (182). Osteomyelitis may also occur as a result of direct trauma related to heel puncture (183), femoral venipuncture (184), and fetal monitoring electrodes (185). Most cases of bone and joint infection are sporadic, although outbreaks have been reported (186).

S. aureus causes a wide spectrum of disease in neonates, from superficial skin infections to severe invasive disease including bacteremia, meningitis, osteomyelitis, and SSIs. Colonization is common; from 20% to 90% of newborns acquire the microorganism in the first week of life (187). In NICUs, healthcare worker hands are implicated in transmission. In one study, 20/37 infants handled by a caregiver for 10 minutes through the portholes of incubators acquired that caregiver’s strain of S. aureus (27). Transmission may be reduced by hand washing (26,188). Less commonly, outbreaks have been traced to nasal carriage by a healthcare worker, or the so-called cloud phenomenon (40,41).

Increasingly, staphylococcal infections in NICUs are due to MRSA. According to data reported to NHSN, late-onset MRSA infection in US NICUs increased 300% from 1995 to 2004 (189). In units that conduct active surveillance, up to 40% of infants may be colonized (190). Infants with MRSA colonization have a significantly higher rate of MRSA infection than those without colonization. In one study, the risk of infection in colonized infants was 26% versus 2% in noncolonized infants (p < .001; odds ratio 19.86). Other risk factors for infection include low birth weight and prematurity (190,191).

The clinical manifestations of MRSA in neonates are similar to those seen with methicillin-susceptible Staphylococcus
aureus (MSSA). Bloodstream infections and skin and soft tissue infections (including postoperative wound infections) predominate, although in one report, infants with MRSA infections were younger at presentation than infants with MSSA infection (23 vs. 32 days, p < .03) (192).

As with MSSA, transmission of MRSA within NICUs is most often ascribed to transient hand carriage on the unwashed hands of healthcare workers (88,193,194). Overcrowding and understaffing may facilitate the MRSA spread (87,195). Other sources include colonized or infected healthcare workers (196), an infected or colonized family member (197), or contaminated breast milk (198). Rates of maternal vaginal or rectovaginal MRSA colonization in pregnant women range from 0% to 10% (199, 200, 201 and 202). Nevertheless, maternal to child transmission with subsequent early-onset MRSA infection is rare (200,201). Communityacquired MRSA chorioamniotitis with subsequent transmission to a premature neonate who developed MRSA sepsis has been described (203).

MRSA has become endemic in some neonatal units, with ongoing introduction and transmission of multiple, distinct molecular strains over time (204). Nevertheless, several centers have eradicated transmission using a multimodal approach to infection prevention (88,205,206,207). Active surveillance cultures, aggressive implementation of Contact Precautions with monitoring of compliance, cohorting, and decolonization of patients and healthcare workers eradicated MRSA from one 18-bed NICU in a community hospital for 2.5 years (205). Creation of distinct nursing cohorts to care for MRSA-colonized or -infected infants, MRSA-exposed infants, and newly admitted infants contributed to the control of MRSA outbreaks in centers in the United States and France (206,208).

Decolonization of MRSA carriers, including colonized infants, is controversial. The optimal decolonization regimen remains unknown. In two, small observational studies, application of mupirocin to the anterior nares of colonized infants did not reliably eradicate colonization or prevent infection (206,208).

A working group composed of infection prevention experts at Chicago-area hospitals has published consensus recommendations for the control of MRSA in NICUs (209). Key features of the recommendations include ready availability of hand hygiene products with monitoring of hand hygiene compliance, cohorting and Contact Precautions for colonized or infected infants, periodic active surveillance culture of patients’ anterior nares, molecular analysis of MRSA isolates, and communication between regional NICUs to prevent spread between institutions with patient transfers. Notably, screening of healthcare workers to detect MRSA colonization is recommended only to corroborate or refute epidemiologic data that link a particular HCW to transmission.

CONS are the most common cause of HAIs in NICU infants. They cause 35% of late-onset sepsis in neonates (19,210) and are a common cause of CLABSIs (14,111). A syndrome of persistent bacteremia accompanied by thrombocytopenia has been described in neonates, even in the absence of CVCs (211). Endocarditis, abscesses, omphalitis, SSIs, and meningitis occur occasionally, and a mild form of NEC has been described. The role of CONS as a cause of neonatal pneumonia is suggested by the isolation of microorganisms in the blood of infants with pulmonary infiltrates. CONS are also the major cause of ventricular shunt and drain infections (212). The virulence of these common skin commensals in neonates may relate in part to their ability to adhere to catheter surfaces and proliferate to form a multilayered biofilm (213,214). Production of slime may also help the microorganism evade host defenses (***214a).

Low birth weight, low gestational age, and procedures that interfere with skin or mucosal integrity, including placement of intravascular catheters, endotracheal tubes, and feeding tubes, increase the risk for CONS infection (19,215,216). Receipt of intravenous lipids is also a risk factor; lipids may enhance the growth of CONS in colonized catheters (217,218). CONS infections prolong the hospital stay of infected newborns and increase healthcare costs but are not associated with increased mortality (219,220).

Effective measures for the prevention of CONS infection include limiting the use of invasive devices and aseptic technique for insertion and handling of intravascular and other prosthetic devices. Addition of vancomycin to intravenous fluids has been shown to reduce the incidence of CONS bacteremia in neonates but is not recommended for routine use because of potential for inducing vancomycin resistance (221). Transmission of CONS with heteroresistance to vancomycin in NICU has been reported (222).

GBS is a leading cause of neonatal sepsis and meningitis among newborns. Approximately 10% to 30% of pregnant women are colonized with GBS, 50% of these women will transmit the bacteria to their newborns, and 2% of colonized infants will develop disease. Early-onset disease usually presents on the first day of life with septic shock, pneumonia, and severe multiorgan failure. Intrapartum prophylaxis of GBS-colonized women has decreased the incidence of disease in infants in the first 72 hours of life (223).

In contrast, late-onset disease is seen in infants up to 3 months of age or older and may manifest as bacteremia, meningitis, or focal infection such as osteomyelitis. The risk factors for late-onset disease are not well described, but only 50% of cases are attributable to maternal colonization (224). Infection may result from a colonized healthcare worker or occasionally as a result of cross-transmission from other infants in the nursery. Healthcare-associated transmission has been reported, especially in crowded nurseries with a high rate of maternal colonization (225).

Enterococci, including Enterococcus faecalis and Enterococcus faecium, are common commensals of the gastrointestinal tract and the vagina. Maternally acquired infection can result in sepsis during the first week of life, although symptoms are generally less severe than those observed with GBS (226). Late-onset infection commonly manifests as bacteremia, often in the setting of NEC or other gastrointestinal tract pathology (227). Bacteremia is often polymicrobial. Enterococci may also cause focal infections such as abscess, urinary tract infection, or meningitis.

Not long ago, vancomycin-resistant enterococci (VRE) were rarely encountered in hospitalized neonates (14), but
recent reports suggest that the incidence of VRE, like other MDRO, is increasing (228,229) Clinical cultures identify a minority of patients harboring VRE. In the setting of active surveillance cultures, the prevalence of colonization has ranged from 2% to >40% (228,229,230). Risk factors for VRE acquisition include low birth weight and exposure to antimicrobials (229,230). Clonal spread in one NICU was linked to a contaminated electronic thermometer (228). Routine infection control measures, including Contact Precautions, cohorting, hand hygiene, and environmental cleaning, have been successful in terminating VRE outbreaks (228,229).

GAS infections may be acquired vertically from colonized mothers or from colonized or infected healthcare personnel (231, 232, 233, 234 and 235). Most infants described in reports of nursery outbreaks developed mild, indolent omphalitis (232, 233 and 234) although severe disease, including sepsis and necrotizing cellulitis, also occur (235). Treatment of the umbilical stump with bacitracin (233),triple dye (234), or chlorhexidine (236) may reduce colonization and disease. Identification and treatment of GAS carriers is essential for the control of outbreaks. Administration of Penicillin G has been effective in some outbreaks. In other reports, eradication of colonization required clindamycin treatment (235).

Listeria is usually maternally acquired (237). Maternal infection is food borne, and clusters of infection in newborns usually indicate community outbreaks. Early-onset disease, often associated with maternal symptoms, presents with pneumonia and rash and multisystem disease. Meningitis is the major form of late-onset disease. Control measures include advising pregnant women to avoid unpasteurized milk products and foods epidemiologically associated with an outbreak and diagnosing and treating infection in pregnancy. Nursery transmission is reported but rare (238). Contaminated resuscitation equipment (239) and mineral oil used to bathe infants (65) have been implicated.

Streptococcus pneumoniae is an unusual cause of neonatal sepsis. Early-onset infection may be associated with maternal sepsis and has a poor prognosis (240). Healthcare-associated transmission has been reported (241).

Leuconostoc species may cause sepsis and meningitis in premature infants (242). Case reports describe underlying gastrointestinal tract disease and central venous catheter use in infants diagnosed with this uncommon, vancomycinresistant pathogen.

From 2% to 70% of infants may be asymptomatically colonized with Clostridum difficile, including toxigenic strains (243,244). Rates of colonization decrease with age, falling in the second year of age to 6%. Rates of colonization in children >2 years of age are similar to those in adults (˜3%).

Infants may acquire colonization early in the first week of life (245). Studies examining the risk factors for C. difficile have failed to show a consistent association between acquisition of the organism and the mode of delivery or receipt of formula versus breast milk. However, healthcare acquisition of the microorganism is well described in NICUs and C. difficile contamination of the NICU environment has been demonstrated (246).

Most studies have failed to show an epidemiologic association between colonization and disease in infants <1 year of age, including NICU patients. C. difficile toxin was recovered from the stools of 55% of patients in one NICU, but signs of enteric disease, including NEC, occurred with equal frequency in both toxin-positive and toxin-negative infants (247). Data from neonatal rabbits suggest that the lack of disease in colonized infants may be related to the absence of a receptor for toxin A in immature enterocytes (248).