Infections Acquired in the Nursery: Epidemiology and Control

Neonates, especially premature neonates, requiring intensive care support constitute a highly vulnerable population at extreme risk for nosocomial or health care–associated infections. It has been estimated that as many as 6% to 22% of infants who survive 48 or more hours in a high-risk nursery or neonatal intensive care unit (NICU) acquire a nosocomial infection.^{1, 2} Although nosocomial infections have long been recognized in NICUs, only recently have data on rates been documented in the literature. As technology and treatments have advanced to significantly diminish mortality and morbidity among critically ill neonates, especially infants of very low birth weight (less than 1500g), this vulnerability has only increased, as a result of both more profound immune system immaturity and more frequent use of invasive interventions that bypass skin and mucous membrane barriers.¹

Nosocomial infections in neonates carry high attendant morbidity and mortality and health care costs. Prevention and control of these infections, although highly desirable, present a formidable challenge to health care professionals. Because control over birth weight—the most significant predictor of nosocomial infection risk—is limited, proper NICU customs, environment, and procedures (e.g., hand hygiene, antimicrobial usage, catheter-related practices, skin and cord care, visitation policies, unit design, and staffing) can reduce the risk for infection in the NICU. Understanding the epidemiology of nosocomial infections in neonates and methods for their prevention and control is critical to minimizing poor outcomes. This chapter describes the epidemiology, etiology, and clinical characteristics of neonatal nosocomial infections as well as the methods required for effective infection prevention and control.

SPECIAL ISSUES FOR NEONATES

It is well recognized that the immune system of the newborn infant, especially the premature infant, is functionally inferior to that of older infants, children, and adults (see Chapter 4). The lineages of the cells that will develop into the immune system are present at the beginning of the second trimester. The major components of the neonatal immune system, including T cells, neutrophils, monocytes, and the complement pathways, are functionally impaired, however, when compared with those in older infants and adults. For example, neonatal neutrophils show decreased chemotaxis, diminished adherence to the endothelium, and impaired phagocytosis^{3, 4}; neonatal complement levels and opsonic capacity also are reduced, particularly in the premature neonate.^{4, 5} In addition, neonatal T cell lymphokine production, cytotoxicity, delayed-type hypersensitivity, and help for B cell differentiation all are inferior when measured against those in adults.⁶ Antigenic naiveté may account for many of these differences; however, inherent immaturity also appears to account for certain inequities. For example, neonatal T cells are delayed in their ability to generate antigen-specific memory function after HSV infection, even in comparison with naive adult T cells.⁷

Passively acquired maternal immunoglobulin G (IgG) is the sole source of neonatal IgG. Soon after birth, maternal IgG levels begin to fall; weeks later, production of immunoglobulins by the neonate commences. Neonatal IgG levels reach about 60% of adult levels by 1 year of age.⁶ Unfortunately, because much of the maternal IgG is not transferred to the infant until the last 8 to10 weeks’ gestation, premature infants start with significantly lower levels of serum IgG than in their term counterparts, which persist throughout most of the first months of life.

Other issues specific to the premature neonate also affect the functional immune system. For instance, the immature gastrointestinal tract (lack of acidity worsened by use of histamine H₂ blockers and continuous feedings) and easily damaged skin constitute open potential portals of entry for pathogens or commensals. In addition, like other intensive care unit populations, the NICU population frequently experiences extrinsic breeches of the immune system through use of intravascular catheters as well as other invasive equipment and procedures used to care for critically ill patients.

It is generally accepted that colonization with “normal flora” prevents, to some degree, colonization by pathogenic organisms. The neonate begins life essentially sterile. In the healthy term neonate, colonization occurs within the first few days of life. The organisms involved by site are α-hemolytic streptococci in the upper respiratory tract, Staphylococcus epidermidis and other coagulase-negative staphylococci (CoNS) on the skin, and gram-negative bacilli and anaerobes in the gastrointestinal tract. This process of colonization with normal flora is disrupted in infants cared for in an NICU in part because of exposure to the NICU environment, the hands of health care workers (HCWs), antimicrobial agents, and invasive procedures. As a result, the microflora of infants in the NICU can be markedly different from that of healthy term infants.^{8, 9} Multiple antimicrobial agent–resistant CoNS, Klebsiella, Enterobacter, and Citrobacter species colonize the skin and the respiratory and gastrointestinal tracts of a high proportion of NICU neonates by the second week of hospitalization.^{10, 11, 12, 13} In addition, neonates in the NICU become colonized not only with Candida albicans but also with non-albicans Candida species and Malassezia.^{14, 15, 16, 17}

Because colonization of the neonate with pathogenic organisms is a prelude to invasive infection from the same pathogens,⁹ measures to prevent such colonization need to be considered. First, as a result of abnormal colonization, infants in the NICU themselves serve as an important reservoir of potential pathogens. Second, contamination of the hands of HCWs during routine patient care has been well documented.¹⁸ Thus, careful attention to hand hygiene before and after contact with patients and their environment, as well as decontamination of potential fomites, are crucial measures in preventing spread of colonization and infection.

EPIDEMIOLOGY

Incidence

Nosocomial infections in healthy term infants are uncommon unless other conditions require that they be cared for in the NICU for several days to weeks. On the other hand, these other conditions are frequent in neonates of very low birth weight (less than 1500g), who require prolonged NICU care. Understanding the epidemiology of nosocomial infections in NICUs can be challenging, because reported rates vary dramatically by institution. This variation probably results from use of nonstandard definitions of nosocomial infection and from differences in patient populations, such as mean gestational age, birth weight, and severity of underlying illness, which significantly affect the incidence of nosocomial infection.¹⁹

The National Nosocomial Infections Surveillance (NNIS) system is a national surveillance system of the Centers for Disease Control and Prevention (CDC) that uses standardized surveillance protocols and the involvement of multiple medical centers to provide benchmark data for the epidemiology of nosocomial NICU infections. Using standardized definitions, NNIS reported in 1996 that 13,179 nosocomial infections occurred between 1986 and 1994 in 10,296 neonates in 99 NICUs.²⁰ In this study, rates of intravascular catheter– associated bloodstream infection, the most frequent nosocomial infection, ranged from fewer than 5 infections per 1000 umbilical or central catheter days in infants with a birth weight greater than 1500g to almost 15 infections per 1000 catheter days in the lowest-birth-weight group (less than 1000g).

Another national, multicenter surveillance study, the Pediatric Prevention Network’s (PPN) Point Prevalence Survey, was undertaken in 1999 to determine the point prevalence of nosocomial infections in NICUs and to define risk factors associated with development of these infections.²¹ This study included 827 infants from 29 NICUs. Of the 827 infants, 94 (11.4%) had an active nosocomial infection on the day of the survey. Bacteremia accounted for 53% of infections; lower respiratory tract infections, ear-nose-throat infections, and urinary tract infections accounted for 13%, 9%, and 9%, respectively (Table 35–1).

Table 35-1.

Distribution of NICU-Acquired Infections by Birth Weight and Site

Birth Weight (g)	No. of Patients		No. of Infections
	Total Surveyed	With Infections (%)	Total	Bacteremia (%)	Respiratory Infections (%)	ENT Infections (%)	UTIs (%)	Other Infections
<500	13	1/13 (7.7)	1	1/1 (100)	0	0	0	0
501–1000	246	43/246 (17.5)	58	31/58 (53.4)	9/58 (15.5)	5/58 (8.6)	4/58 (7.0)	9/58 (15.5)
1001–1500	147	21/147 (14.3)	26	15/26 (57.7)	2/26 (7.7)	2/26 (7.7)	1/26 (3.8)	6/26 (23.1)
1501–2000	74	2/74 (2.7)	2	2/2 (100)	0	0	0	0
2001–2500	74	5/74 (6.8)	5	2/5 (40)	1/5 (20)	1/5 (20)	1/5 (20)	0
>2500	239	16/239 (6.7)	17	7/17 (41.2)	2/17 (11.8)	2/17 (11.8)	3/17 (17.6)	3/17 (17.6)
Unknown	34	6/347 (1.7)	7	3/7 (42.8)	1/7 (14.3)	0	1/7 (14.3)	2/7 (28.6)
Total	827	94/827 (11.4)	116	61/116 (52.6)	15/116 (12.9)	10/116 (8.6)	10/116 (8.6)	20/116 (17.2)
ENT, ear, nose, or throat; NICU, neonatal intensive care unit; UTIs, urinary tract infections.

Data from Sohn AH, Garrett DO, Sinkowitz-Cochran RL, et al. Prevalence of nosocomial infections in neonatal intensive care unit patients: results from the first national point-prevalence survey. J Pediatr 139:821–827, 2001.

In contrast with the NICU setting, the frequency of nosocomial infection in well-baby nurseries has been estimated to be between 0.3% and 1.7%.^{22, 23, 24} In general, non–life-threatening infections such as conjunctivitis account for a majority of infections in the well-baby population. The remainder of this chapter focuses almost entirely on nosocomial infections in and control measures for the NICU setting.

Maternally Acquired Infections

For purposes of surveillance and tracking, all infections occurring in hospitalized newborns could be considered nosocomial. Infections that are manifested in the first few days of life, however, usually are caused by pathogens transmitted vertically from the maternal genital tract. Unfortunately, no precise time point perfectly distinguishes maternally acquired neonatal infections from those transmitted within the NICU. NNIS has attempted to address this issue by stratifying infections according to whether they are likely to be maternally acquired.²⁰ In 89% of neonates who had an infection thought to be maternally acquired, onset occurred within 48 hours of birth. Use of a cutoff period of 48 hours or less to designate maternally acquired infections allowed 15.3% of bacteremias and 14.5% of pneumonias to be considered as originating from a maternal source. Maternally acquired bloodstream infections were more likely to be caused by group B streptococci, other streptococci, and Escherichia coli, whereas those not maternally acquired usually were caused by coagulase-negative staphylococci.²⁰

Nonmaternal Routes of Transmission

In general, nonmaternal routes of transmission of microorganisms to neonates are divided into three categories: contact (from either direct or indirect contact from an infected person or a contaminated source), droplet (from large respiratory droplets that fall out of the air at a maximum distance of 3 feet), and airborne (from droplet nuclei, which can remain suspended in air for long periods and as a result travel longer distances). Specific microorganisms can be spread by more than one mechanism; in most instances, however, a single mode of spread predominates. The CDC has developed a system of precautions for the control of nosocomial infections that is based on these modes of transmission.²⁵

Contact transmission of bacteria, viruses, and fungi on the hands of HCWs is arguably the most important yet seemingly preventable means of transmission of nosocomial infection. Spread of infection by this means can occur either by transmission of the HCW’s own colonizing or infecting pathogens or, more often, by transmission of pathogens from one patient to another. That the hands of HCWs become contaminated even in touching intact skin of patients has been well demonstrated.¹⁸ Poor compliance with hand hygiene is another means by which the hands of HCWs can spread organisms from one patient to another.^{26, 27} Furthermore, hands of HCWs have been implicated in multiple outbreaks with a variety of different organisms; through experimental studies, a causal link between hand hygiene and nosocomial infection has been established.²⁸

Contact transmission by means of fomites also can occur and has been described as a potential mechanism of spread of pathogens in multiple NICU outbreaks. As described later in this chapter, implicated items have included linens, medical devices, soap dispensers, and breast pumps, to name a few. These observations highlight the need for careful attention to disinfecting items shared between infants.

Spread through large respiratory droplets is an important mode of transmission for pertussis and infections due to Neisseria meningitidis, group A streptococci, and certain respiratory viruses, whereas airborne transmission by means of droplet nuclei is relevant for measles, varicella, and pulmonary tuberculosis. For large droplet or droplet nuclei transmission, usually an ill adult, either an HCW or a parent, is the source of infection in an NICU setting. In general, these organisms are rare sources of outbreaks.

Infusates, medications, and feeding powders or solutions can be intrinsically or extrinsically contaminated and have been reported as the source of outbreaks due to a variety of different pathogens. It is important when possible to mix infusates in a controlled environment (usually the pharmacy), to avoid multiuse sources of medication, and to use bottled or sterilized feeding solutions when breast milk is not available.

Of course, nosocomial infection also can arise from endogenous sources within the neonate. The “abnormal flora” of the neonate residing in the NICU, however, is determined at least in part by the NICU environment and HCWs’ hands. With use of molecular techniques, even organisms typically considered to originate solely from normal flora (e.g., CoNS) have been shown to have clonal spread in the hospital setting, suggesting transmission by means of the hands of HCWs.^{29, 30}

Risk Factors for Nosocomial Infection

As discussed earlier, infants in NICUs have intrinsic factors predisposing them to infection, such as an immature immune system and compromised skin or mucous membrane barriers. In addition, multiple extrinsic factors play important roles in the development of infection, such as presence of indwelling catheters, performance of invasive procedures, and administration of certain medications, such as steroids and antimicrobial agents.

Birth weight is one of the strongest predictors of risk for nosocomial infection. For instance, NNIS data demonstrate that compared with larger infants, low-birth-weight infants are at higher risk of developing bloodstream infections and ventilator-associated pneumonia, even after correction for central intravascular catheter and ventilator use.²⁰ Similarly, in the PPN’s Point Prevalence Survey, infants weighing 1500g or less at birth were 2.69 (95% confidence interval [CI] 1.75% to 4.14%; P < .001) times more likely to have an infection than those weighing more than 1500g.²¹ The relationship between birth weight and nosocomial infection is complicated by multiple other factors that accompany low birth weight and also increase risk for nosocomial infection. Low birth weight, however, has been shown to be an independent predictor for nosocomial infection, after adjustment for use of vascular catheters, parenteral alimentation, and mechanical ventilation.³¹ It is likely that birth weight also is a surrogate marker for other unmeasured factors, such as immune system immaturity.

Central venous catheters (CVCs) increase the risk for development of nosocomial bloodstream infections. In a study by Chien and colleagues 19,507 infants admitted to 17 NICUs in Canada, nosocomial bloodstream infections were found to occur at a rate of 3.1 to 7.2 infections per 1000 catheter days, depending on the type of catheter, versus 2.9 infections per 1000 noncatheter days. Other studies have demonstrated that the association between CVCs and bloodstream infection is independent of birth weight.²¹ Mechanisms for CVC-related nosocomial bloodstream infections probably involve colonization of the catheter by means of the catheter hub, colonization of the skin at the insertion site,³³ or hematogenous spread of pathogens from distant sites of infection or colonization. Bloodstream infections also can result from contaminated intravenous fluids, which have the potential for intrinsic or, especially with use of multiuse vials, extrinsic contamination.

Factors related to the management of CVCs influence the risk of infection. Disconnection of the CVC and the frequency of blood sampling through the catheter increase the frequency of catheter-related infections.³⁴ By contrast, administration of a solution with heparin and exit-site antisepsis decreased infection. Lower frequency of CVC tubing changes (every 72 hours versus every 24 hours) was associated with increased catheter contamination, suggesting a potential for increased risk of infection.³⁵ CVC management techniques, including use of antiseptic-impregnated dressings, antimicrobial-coated catheters, and avoidance of scheduled replacement of CVCs, are discussed in the most recent CDC recommendations, summarized in “Guidelines for the Prevention of Intravascular Catheter–Related Infection,” published in 2002 and prepared by the Hospital Infection Control Practice Advisory Committee.³⁶

It has been suggested that use of peripherally inserted central catheters (PICCs) may be associated with a lower rate of infection than for other CVCs. Studies based in NICUs have yielded conflicting results. In a study by Chien and colleagues,³² the relative risk of bloodstream infection, after adjustment for differences in infant characteristics and admission illness severity, was 2.5 per 1000 catheter days for umbilical venous catheters, 4.6 for PICCs, and 4.3 for Broviac catheters, compared with no catheter (P < .05). Another study also documented similar rates of infection for Broviac catheters and for PICCs.³⁷ By contrast, a higher rate of infection with Broviac catheters than with PICCs was suggested by Brodie and co-workers.³⁸ Further study of different CVCs in NICU infants is needed to delineate infection risks for individual catheter types.

Parenteral alimentation and intralipids have been shown to increase risk of bloodstream infection in premature infants even after adjustment for other covariables such as birth weight and CVC use.³⁸ Etiologic agents often associated are CoNS, Candida species, and Malassezia species. The pathogenesis of this association remains unclear. Potential hypotheses are many. Intralipids, for example, could have a direct effect on the immune system, perhaps through inhibition of interleukin-2.³⁹ Alternatively, as with any intravenous fluids, parenteral alimentation has the potential for intrinsic and extrinsic contamination, and intralipids especially may serve as a growth medium for certain bacteria and fungi. Finally, total parenteral alimentation and intralipids delay the normal development of gastrointestinal mucosa because of lack of enteral feeding, encouraging translocation of pathogens across the gastrointestinal mucosa.

It is well accepted that mechanical ventilation is an important risk factor for nosocomial lower respiratory tract infection. A large multicenter study of 8263 neonates found that mechanical ventilation was a risk factor for bloodstream infection as well, even after adjustment for a number of covariables such as birth weight, parenteral nutrition, and umbilical catheterization.⁴⁰ Clinically obvious respiratory infection appeared to precede some but not all cases of bloodstream infection associated with mechanical ventilation. The study authors suggested that the increased risk of mechanical ventilation could be attributed to colonization of humidified air, as well as to physical trauma from the endotracheal tube and its suctioning.

A number of medications critical to the survival of infants in the NICU increase risk of infection. Broad-spectrum antimicrobial agents, especially with prolonged use, are important in the development of colonization with pathogenic microorganisms.⁹ The widespread use of broad-spectrum antimicrobial agents has been associated with increased colonization with resistant organisms in many settings, including NICUs.¹⁹ In addition to colonization, antimicrobial agents also have been shown to increase risk of infection with resistant bacteria⁴¹ and with fungal pathogens.⁴² Other medications also appear to play a role in nosocomial infection. For instance, infants who receive corticosteroids after delivery are at approximately 1.3 to 1.6 times higher risk for nosocomial bacteremia in the subsequent 2 to 6 weeks than that observed for infants who do not receive this intervention.^{43, 44} In addition, colonization and infection with bacterial and fungal pathogens have been shown to increase with the use of H₂ blockers.^{14, 31}

Measures of illness severity have been developed, in part, in an effort to account for variations in birth weight–adjusted mortality scores between NICUs. The Score for Neonatal Acute Physiology (SNAP) was developed and validated by Richardson and associates,⁴⁵ and the Clinical Risk Index for Babies (CRIB) was developed by the International Neonatal Network.⁴⁶ These scores are highly predictive of neonatal mortality even within narrow birth weight strata and are predictive of nosocomial infection. Thus, in investigating potential risk factors for nosocomial infection, it is important to consider adjusting for illness severity using such measures, in addition to adjusting for other potential confounders.

Other risk factors related to infection include poor hand hygiene and environmental issues, such as understaffing and overcrowding.^{47, 48} These and related issues are discussed later in this chapter under “Prevention and Control.”

CLINICAL MANIFESTATIONS

Nosocomial infections can affect any body site or organ system and manifest in a multitude of different ways. NNIS and PPN data demonstrated that bloodstream infections are the most common manifestation of nosocomial infection and account for 32% to 53% of infections (Table 35–2; see also Table 35–1).^{20, 21} Respiratory infections and eye, ear, nose, or throat infections are second and third in frequency, whereas gastrointestinal infections, urinary tract infections, surgical site infections, meningitis, cellulitis, omphalitis, septic arthritis, and osteomyelitis are reported less frequently.^{20, 21}

Table 35-2.

Most Common Pathogens Causing Nosocomial Infection in NICU Patients: Distribution by Site

Pathogen	Bloodstream	No. of Infections (%)		Pneumonia	Surgical Site
		EENT	GI
Coagulase-negative staphylococci (CoNS)	3833 (51.0)	787 (29.3)	102 (9.6)	434 (16.5)	119 (19.2)
Staphylococcus aureus	561 (7.5)	413 (15.4)		440 (16.7)	138 (22.3)
Group B streptococci	597 (7.9)			150 (5.7)
Enterococcus	467 (6.2)	92 (3.4)		120 (4.6)	55 (8.9)
Candida species	518 (6.9)
Escherichia coli	326 (4.3)	163 (6.1)	147 (13.9)	152 (5.8)	74 (12.0)
Other streptococcal species	205 (2.7)	199 (7.4)		86 (3.3)
Enterobacter species	219 (2.9)	120 (4.5)	58 (5.5)	215 (8.2)	47 (7.6)
Klebsiella pneumoniae	188 (2.5)	76 (2.8)	104 (9.8)	152 (5.8)	39 (6.3)
Pseudomonas aeruginosa		178 (6.6)		308 (11.7)
Haemophilus influenzae		72 (2.7)		38 (1.4)
Viruses		136 (5.1)	317 (30.0a)
Gram-positive anaerobes			99 (9.4)
Other enteric bacilli			8 (0.8)
Miscellaneous organisms	607 (8.1)	449 (26.7)	223 (21.0)	570 (21.7)	147 (23.7)
Total	7521 (100)	2685 (100)	1058 (100)	2665 (100)	619 (100)
EENT, eye, ear, nose, or throat; GI, gastrointestinal; NICU, neonatal intensive care unit.

^aRotavirus constitutes 96.4% of viruses isolated from gastrointestinal infections.

Data from Gaynes RP, Edwards JR, Jarvis WR, et al. Nosocomial infections among neonates in high-risk nurseries in the United States. National Nosocomial Infections Surveillance system. Pediatrics 98:357–361, 1996.

Bloodstream infections are the most common and one of the most potentially serious nosocomial infections that occur in NICU patients. Factors discussed earlier, including birth weight, intravascular catheters, mechanical ventilation, use of parenteral alimentation, and steroids, all have been shown to be associated with an increased risk of bloodstream infection. The most common pathogen associated with nosocomial bloodstream infections is CoNS (see Table 35–2). Staphylococcus aureus, Enterococcus, Candida species, E. coli, Enterobacter species, Klebsiella pneumoniae, and Pseudomonas aeruginosa also play important roles and are associated with higher morbidity and mortality rates than those associated with CoNS.^{49, 50} In one study, the frequency of fulminant sepsis (fatal within 48 hours) was estimated to be 56% (95% CI 38% to 72%) when the bloodstream infection was caused by Pseudomonas species, whereas it was only 1% (95% CI 0% to 4%) when infection was caused by CoNS.⁵⁰ The difficulty of assigning an etiologic role to CoNS on the basis of one blood culture that could be contaminated probably accounts for some distortion of the incidence and mortality data related to this organism, and this problem is discussed in detail in Chapter 6.

Pneumonia accounts for 12% to 18% of NICU nosocomial infections²⁰ and has been associated with prolonged hospital stay and increased mortality.⁵¹ Organisms most commonly associated with nosocomial pneumonia include CoNS, S. aureus, and P. aeruginosa (see Table 35–2). Mechanical ventilation and birth weight are important risk factors for nosocomial respiratory infections.⁵¹ Diagnosis of nosocomial respiratory infections requires correlation of microbiologic results with clinical findings and can be challenging in low-birth-weight infants because of the mostly nonspecific associated signs of illness and often misleading results of radiologic studies.⁵²

Eye, ear, nose, and throat infections account for approximately 8% to 21% of infections, depending on birth weight.²⁰ Common etiologic organisms include CoNS and S. aureus, although gram-negative organisms, such as E. coli, P. aeruginosa, and K. pneumoniae, also can be isolated from these sites (see Table 35–2). Conjunctivitis appears to be the most common of these infections, accounting for 54% to 76%, depending on birth weight.²⁰ Risk factors for neonatal conjunctivitis identified in a study from Nigeria included vaginal delivery, asphyxia, and prolonged rupture of membranes.⁵³

In the NNIS review, gastrointestinal infections were estimated to account for 5% to 11% of nosocomial infections, depending on birth weight. Necrotizing enterocolitis (NEC) was the most common presentation.²⁰ NEC carries high morbidity and mortality rates. A review of 17 NEC epidemics estimated that surgery was required for a mean of 16% (range, 0% to 67%) of infants, and death occurred in a mean of 6% (range, 0% to 88%).⁵⁴ In controlled studies, identified risk factors for NEC have included young chronologic age, low gestational age, low birth weight, and young age at first feeding.⁵⁴ Implication of specific pathogens is complex, requiring careful selection of an appropriate control population and attention to how and where specimens are collected. Pathogens associated with NEC outbreaks have included Pseudomonas species, Salmonella species, E. coli, K. pneumoniae, Enterobacter cloacae, S. epidermidis, Clostridium species, coronavirus, and rotavirus.^{54, 55} The importance of infection control methods such as strict attention to hand hygiene and cohorting patients in the NICU is suggested by the observation that their implementation has been followed by resolution of the outbreak.⁵⁵

ETIOLOGIC AGENTS

A detailed discussion of the cause of nosocomial sepsis and meningitis is found in the chapter on bacterial sepsis (Chapter 6) and chapters describing specific etiologic agents.

Gram-Positive Bacteria

S. aureus is a colonizing agent in neonates and has been a cause of nosocomial infection and outbreaks in well-baby nurseries and NICUs. Methicillin-resistant S. aureus (MRSA) has become a serious nosocomial pathogen, and outbreaks have been reported in many areas of hospitals, including nurseries.^{56, 57, 58} In addition to the usual manifestations of neonatal nosocomial infection (conjunctivitis, bloodstream infections, and pneumonia), nosocomial S. aureus infections can manifest as skin infections,⁵⁹ bone and joint infections,⁶⁰ parotitis,⁶¹ staphylococcal scalded skin syndrome,^{62, 63} toxic shock syndrome,⁵⁶ and disseminated sepsis.

The role of the hands of HCWs in transmitting and spreading pathogenic organisms among infants was demonstrated with S. aureus in the 1960s.^{64, 65} Currently, in a majority of instances, S. aureus transmission is thought to occur by direct contact. Thus, it is not surprising that understaffing and overcrowding have been associated with S. aureus outbreaks in NICUs.^{57, 66} The potential for airborne transmission, however, has been suggested by the occurrence of “cloud babies,” described by Eichenwald and colleagues⁶⁷ in 1960. “Cloud” HCWs also have been described; in such cases, the point source of an outbreak was determined to be a colonized HCW with a viral respiratory infection.^{59, 68} In one of these studies, dispersion of S. aureus from the implicated HCW was found to be much higher after experimental infection with rhinovirus.⁶⁸ More recently, molecular techniques not only have defined outbreaks⁵⁷ but also have demonstrated that transmission to infants probably occurs from colonized HCWs,⁶² and sometimes from colonized parents.⁶⁹

Nasal mupirocin ointment has been used to control outbreaks of both methicillin-susceptible S. aureus and MRSA.^{62, 70} The pharynx, rather than the anterior nares, however, may be a more common site of colonization in neonates and infants,⁷¹ and eradication of the causative organisms with nasal mupirocin may be more difficult in this site.⁷²

Coagulase-Negative Staphylococci

Since the early 1980s, CoNS has been the most common cause of nosocomial infection in the NICU.⁴⁸ Recent NNIS and PPN surveillance estimate that 32% of total pathogens and 48% to 51% of bloodstream infections are caused by these organisms.^{20, 21} A 10-year, prospective, multicenter Australian study corroborated the fact that CoNS accounted for most infections and demonstrated that 57% of all late-onset infections during the study period were due to these organisms.⁷³ This study and a U.S. study of 302 very low birth weight infants from two NICUs reported that the highest risk for CoNS infection was in the most premature infants.^{73, 74} In one study, infection usually manifested between 7 and 14 days of life and was accompanied by a mortality rate of 0.3%.⁷³ Although associated with relatively low mortality rates, bacteremia due to CoNS has been correlated, by means of multivariate analysis, with prolonged NICU stay and increased hospital charges, even after adjustment for birth weight and severity of illness on admission.⁷⁴

Many experts consider infection due to CoNS to be inevitable in neonates in the NICU. Molecular techniques suggest that infections due to S. epidermidis can result from clonal dissemination.^{29, 30} In one study, four clones accounted for 43 of 81 study strains (53%).²⁹ This finding suggests that a portion of CoNS infections may be preventable by strict adherence to infection control practices. The fact that a hand hygiene campaign was associated with increased hand hygiene compliance and a lower rate of CoNS-positive cultures supports this contention.⁷⁵

Enterococcus has been shown to account for 10% of total nosocomial infections in neonates, 6% to 15% of bloodstream infections, 0% to 5% of cases of pneumonia, 17% of urinary tract infections, and 9% of surgical site infections.^{20, 21} Sepsis and meningitis are common manifestations of enterococcal infection during NICU outbreaks^{75, 76}; however, polymicrobial bacteremia and NEC frequently accompany enterococcal sepsis.⁷⁷ Identified risk factors for enterococcal sepsis, after adjustment for birth weight, include use of a nonumbilical CVC, prolonged presence of a CVC, and bowel resection.⁷⁷ Because Enterococcus colonizes the gastrointestinal tract and can survive for long periods of time on inanimate surfaces, the patient’s environment may become contaminated and, along with the infant, serve as a reservoir for ongoing spread of the organism.

The emergence of vancomycin-resistant enterococci (VRE) is a concern in all hospital settings, and VRE have been the cause of at least one outbreak in the NICU setting.⁷⁸ In the neonate, resistant strains appear to cause clinical syndromes indistinguishable from those due to susceptible enterococci.⁷⁷ The conditions promoting VRE infection, such as severe underlying disease and use of broad-spectrum antimicrobial agents, especially vancomycin, can be difficult to alter in many NICU settings. Guidelines for the prevention and control of VRE infection have been published; these focus on infection control tools such as rapid identification of a VRE-colonized or VRE-infected patient, cohorting, isolation, and barrier precautions.⁸⁰

Historically, before the recognized importance of hand hygiene and the availability of antimicrobial agents, group A streptococci (GAS) were a major cause of puerperal sepsis and fatal neonatal sepsis. Although less common now, GAS continue to be a cause of well-baby and NICU outbreaks.^{81, 82, 83, 84} GAS-associated clinical manifestations include severe sepsis and soft tissue infections. One report described a high frequency of “indolent omphalitis”; in this outbreak, the umbilical stump appeared to be an important site of GAS colonization and an ongoing reservoir of the organism.⁸¹ Routine cord care included daily alcohol application. After multiple attempts, the outbreak finally was interrupted after a 15-day interval during which bacitracin ointment was applied to the umbilical stump in all infants, and affected infants received intramuscular penicillin. Molecular techniques have enhanced the ability to define outbreaks, and use of these techniques has suggested that transmission can occur between mother and infant, between HCW and infant, and between infants— probably indirectly on the hands of HCWs.^{82, 83} In one recurring outbreak, inadequate laundry practices appeared to be a contributing factor.⁸⁵

NNIS data have shown that group B streptococci (GBS) infections account for less than 2% of non–maternally acquired nosocomial bloodstream and pneumonia infections.²⁰ A number of studies from the 1970s and 1980s demonstrated nosocomial colonization of infants born to GBS-negative women.^{86, 87, 88, 89, 90} These studies suggested a rate of transmission to babies born to seronegative mothers as high as 12% to 27%.^{87, 88} A recent case-control study evaluating risk factors for late-onset GBS infection demonstrated that premature birth was a strong predictor.⁹¹ In that study, 50% of the infants with late-onset GBS infection were born at less than 37 weeks of gestation (compared with 15% of controls), and only 38% of the mothers of these infants were colonized with GBS, suggesting possible nosocomial transmission of GBS during the NICU stay.

The hands of HCWs are assumed to account for the transmission of most cases of nosocomial GBS infection. Breast milk also has been implicated as a potential mode of acquisition, however. In one report, GBS probably was transmitted from breast milk to one set of premature triplets between days 12 and 63 of life.⁹² Two maternal vaginal swabs taken before delivery did not grow GBS, but repeated cultures of the mother’s breast milk yielded a pure growth of GBS (greater than 10⁵ colony-forming units [CFU]/mL) despite no evidence of mastitis. In this report, antimicrobial therapy administered to the mother appeared to eradicate the organism.

Gram-Negative Bacteria

The Enterobacteriaciae family has long been recognized as an important cause of nosocomial infection. Neonatal infection can be manifested as sepsis, pneumonia, urinary tract infections, and soft tissue infections; morbidity and mortality rates frequently are high.⁹³ Enterobacter species, K. pneumoniae, E. coli, and Serratia marcescens are the members of the family Enterobacteriaciae most commonly encountered in the NICU.

Enterobacter species have been estimated to account for 3% of bloodstream infections, 8% of cases of pneumonia, and 8% of surgical site infections in the NICU setting (see Table 35–2). Outbreaks due to Enterobacter species in NICUs have been associated with thermometers,⁹⁴ a multidose vial of dextrose,⁹⁵ intravenous fluids,⁹⁶ and powdered formula,⁹⁷ as well as with understaffing, overcrowding, and poor hand hygiene practices.⁹⁸ In one outbreak in which contaminated saline was linked to the initial cases, subsequent ongoing transmission was documented, presumably by means of the hands of HCWs and the environment.⁹⁹ In that study, early gestational age, low birth weight, exposure to personnel with contaminated hands, and E. cloacae colonization of the stool were associated with E. cloacae bacteremia, whereas use of CVCs and mechanical ventilation was not.

K. pneumoniae has been estimated to account for a similar proportion of infections in the NICU setting to that identified for Enterobacter species. Investigations in outbreaks involving Klebsiella species have implicated contaminated breast milk,¹⁰⁰ infusion therapy practices,¹⁰¹ intravenous dextrose,¹⁰² cockroaches, ¹³⁴ disinfectant,¹⁰⁴ incubator humidifiers,¹⁰⁵ thermometers, oxygen saturation probes,¹⁰⁶ and ultrasonography coupling gel.¹⁰⁸ In a surveillance study of 383 NICU infants in Brazil, 50% became colonized with Klebsiella.¹³ In this study, colonization was associated with use of a cephalosporin and aminoglycoside combination therapy, as well as with longer duration of the NICU stay.

E. coli has been estimated to cause 4% of bloodstream, 14% of gastrointestinal, and 12% of surgical site infections. E. coli also has been responsible for outbreaks of pyelonephritis,¹⁰⁸ gastroenteritis,^{109, 110} and NEC.

S. marcescens is an opportunistic pathogen that survives in relatively harsh environments. Disease due to S. marcescens often is manifested as meningitis, bacteremia, and pneumonia.¹¹¹ S. marcescens infections have a high potential for morbidity and mortality.^{112, 113} S. marcescens outbreaks have been associated with, but not limited to, contaminated soap,¹¹⁴ multiuse bottles of theophylline,¹¹⁵ formula,¹¹⁵ enteral feeding additives,¹¹⁶ breast pumps,^{117, 118} and transducers from internal monitors.¹¹⁶ Although point source environmental contamination is important in Serratia outbreaks, in many of these outbreaks and in reports in which no point source was identified,¹¹⁹ patient-to-patient spread of the organism by means of the hands of HCWs appeared to be an important mechanism of spread.¹¹³

Extended-spectrum β-lactamases (ESBLs) are plasmid-mediated resistance factors produced by members of the Enterobacteriaceae family. ESBLs inactivate third-generation cephalosporins and aztreonam. They most commonly occur in K. pneumoniae and E. coli but have increasingly been found in other gram-negative bacilli. Colonization with ESBL-producing organisms has been associated with administration of certain antimicrobials and longer duration of hospitalization, whereas infection has been associated with prior colonization and use of CVCs.¹³ That the ESBL-containing plasmids can be transmitted to other Enterobacteriaceae organisms has been demonstrated in NICU outbreaks in which the implicated plasmid spread from Klebsiella species to E. coli, E. cloacae, and Citrobacter freundii.^{120, 121} The gastrointestinal tract in neonates and the hands of HCWs serve as reservoirs for members of the Enterobacteriaceae family. Thus, in general, measures aimed at controlling spread of organisms in this family have focused on attention on hand hygiene, cohorting of patient and staff, and observation of isolation precautions.^{121, 122}

P. aeruginosa, an opportunistic pathogen that persists in relatively harsh environments, frequently has been associated with nosocomial infections and outbreaks in the NICU setting. Nosocomial P. aeruginosa infections vary in their clinical presentation, but the most common manifestations are respiratory, ear, nose, or throat and bloodstream infections.²¹ From the PPN data it has been estimated that P. aeruginosa species account for 6.8% of total pathogens, 5% of bloodstream infections, and 15% of respiratory infections.²¹ P. aeruginosa infections, particularly bloodstream infections, have been associated with a very high mortality rate.¹²³

Feeding intolerance, prolonged parenteral alimentation, and long-term intravenous antimicrobial therapy have been identified as risk factors for Pseudomonas infection.¹²³ Outbreaks due to P. aeruginosa have been linked with contaminated hand lotion,¹²⁴ respiratory therapy solution, ¹²⁵ a water bath used to thaw fresh-frozen plasma,¹²⁶ a blood gas analyzer,¹²⁷ and bathing sources. In one case, neonatal Pseudomonas sepsis and meningitis were shown by pulsed-field gel electrophoresis to be associated with shower tubing from a tub used by the infant’s mother during labor.¹²⁸ Of importance, HCWs and their contaminated hands also have been linked with Pseudomonas infections in the NICU setting. In a study of a New York outbreak, recovery of Pseudomonas from the hands of HCWs was associated with older age and history of use of artificial nails.¹²⁹ This and other studies suggest that the risk of transmission of Pseudomonas to patients is higher among HCWs with onychmycosis or those who wear long artificial or long natural nails.^{129, 130} As a result of these and other findings, the CDC revised its 2002 hand hygiene recommendations to include a recommendation against the presence of HCWs with artificial fingernails in intensive care units.¹³¹

Bordetella pertussis is a rare cause of nosocomial infection in neonates. When B. pertussis infection occurs, parents and HCWs typically are the source. A parent was the source of an outbreak involving three neonates and one nurse in a special care nursery in Australia.¹³² In 1999 in Knoxville, Tennessee, an outbreak involving six neonates probably was due to transmission of infection by an HCW.¹³³ As a result of the Tennessee outbreak, 166 infants received erythromycin prophylaxis. Subsequently, an increase in infantile hypertrophic pyloric stenosis was noted by local pediatric surgeons. Results of a CDC investigation suggested a causal role of erythromycin in the cases of hypertrophic pyloric stenosis.^{133, 134} Erythromycin remains the recommended agent of choice for prophylaxis after pertussis exposure, but parents should be informed of the risk and signs of hypertrophic pyloric stenosis, and cases associated with erythromycin use should be reported to MedWatch.¹³⁵

Other Bacterial Pathogens

Newborn infants are particularly prone to infection and disease following exposure to Mycobacterium tuberculosis. A cluster of multidrug-resistant M. tuberculosis infections was noted in three infants born during a 2-week period in one New York hospital.¹³⁶ Investigation implicated an HCW who visited the nursery several times during that period. Pulmonary and extrapulmonary disease occurred in three infants 4 to15 months after exposure, highlighting the vulnerability of the newborn population.¹³⁶ Tuberculosis screening of HCWs, ultraviolet lighting, and a high number of air exchanges appear to be effective methods in preventing nosocomial tuberculosis infection.¹³⁷ The CDC’s “Guidelines for Preventing the Transmission of Mycobacterium tuberculosis in Health-Care Settings” emphasizes (1) use of engineering controls and personal protective equipment, (2) risk assessments for the development of institutional tuberculosis control plans, (3) early identification and management of individuals with tuberculosis infection and disease, (4) tuberculosis screening programs for HCWs, (5) HCW education and training, and (6) evaluation of tuberculosis control programs.¹³⁸

Fungi

Candida species are an increasingly important cause of nosocomial infection in NICU patients and have been estimated to account for 6.9% of bloodstream infections and 42% of urinary tract infections.^{20, 139} Prospective studies have estimated colonization rates with Candida to be 12% to 54% in low-birth-weight neonates,^{17, 140, 141, 142} and colonization has been associated with subsequent invasive disease.¹⁴¹ The mortality rate can be high in invasive candidiasis. In one study of 34 patients with fungemia due to Candida species, a case-fatality rate of 41% was reported.¹⁴³ Risk factors for fungal infections in neonates are similar to risk factors for bacterial infections; low birth weight and gestational age are important predictors. In addition, a prospective, multicenter study of 2157 infants found that use of a third-generation cephalosporin, presence of a CVC, intravenously administered lipids, and H₂ blocker therapy were associated with Candida colonization after adjusting for length of stay, birth weight of 1000g or less, and gestational age less than 32 weeks.¹⁴ Candida parapsilosis appears to be the most frequent species associated with nosocomial Candida infection in NICU infants. Both cross-contamination and maternal reservoirs are sources of nosocomial Candida albicans infection, as demonstrated in studies using molecular typing methods.^{144, 145, 146}

Malassezia species, lipophilic yeasts, frequently colonize NICU patients. In one French study, 30 of 54 preterm neonates (56%) became colonized with Malassezia furfur.¹⁴⁷ Malassezia pachydermatis, a zoonotic organism present on the skin and in the ear canals of healthy dogs and cats, also has been associated with nosocomial outbreaks in the NICU setting.^{147, 148} In one report, the outbreak appeared to be linked to colonization of HCWs’ pet dogs.¹⁴⁷

Pichia anomala, or Hansenula anomala, a yeast found in soil and pigeon droppings, and on plants and fruits, also can colonize the human throat and gastrointestinal tract. In general, it is an unusual cause of nosocomial infection in neonates, but it was the cause of two reported outbreaks in this setting.^{149, 150} In both reports, carriage on the hands of HCWs appeared to be a factor.

Invasive mold infections are a rare cause of nosocomial infection in neonates, but when they occur, they are associated with high mortality rate. Aspergillus infections may manifest as pulmonary, central nervous system, gastrointestinal, or disseminated disease. A cutaneous presentation, with or without subsequent dissemination, appears to be the most common presentation for hospitalized premature infants without underlying immune deficiency.^{151, 152} Often, skin maceration is the presumed portal of entry. In a series of four patients who died of disseminated Aspergillus infection that started cutaneously, a contaminated device used to collect urine from the male infants was implicated.¹⁵² Similarly, contaminated wooden tongue depressors, used as splints for intravenous and arterial cannulation sites, were associated with cutaneous infection due to Rhizopus microsporus in four premature infants.¹⁵³ In addition to preterm birth, use of broad-spectrum antimicrobial agents, steroid therapy, and hyperglycemia are thought to be risk factors for mold infection.

Even zoophilic dermatophytes have been described as a source of nosocomial infection in neonates. In one report, five neonatal cases in one unit were traced to an infected nurse and her cat.¹⁵⁴ Prolonged therapy for both the nurse and her cat was necessary to clear their infections.

Viruses

Rotavirus

Although many pathogens can cause nosocomial gastroenteritis, rotavirus is responsible for 95% or more of viral infections in high-risk nurseries, including the NICU.²⁰ In one longitudinal study, rotavirus infection developed during hospitalization in 95 of 194 neonates (49%).¹⁵⁵ In this study, rotavirus was manifested as frequent and watery stools in term infants and as abdominal distention and bloody, mucoid stools in the preterm neonates.

A high titer of virus is excreted in stool of infected persons, and the organism is viable on hands and in the environment for relatively prolonged periods of time.^{156, 157} Attention to hand hygiene and disinfection of potential fomites are crucial in preventing spread of infection. This concept is illustrated by the results of one study in which rotavirus infection was associated with ungloved nasogastric tube feeding.¹⁵⁷

Respiratory Viruses

Respiratory viruses including influenza A virus, parainfluenza virus, coronavirus, respiratory syncytial virus, and adenovirus have been reported to cause nosocomial infections in NICU patients.^{158, 159, 160, 161} Associated clinical findings include rhinorrhea, tachypnea, retractions, nasal flaring, rales, and wheezing, but illness also can be manifested as apnea, sepsis-like illness, and gastrointestinal symptoms.^{161, 162} Identified risk factors for acquisition vary from study to study but have included low birth weight, low gestational age, twin pregnancy, mechanical ventilation, and high CRIB score.^{159, 160, 161, 162} Contact and droplet transmission are the most common modes of spread of infection, again highlighting the importance of scrupulous hand hygiene in delivery of patient care for this population.

Enteroviruses

Numerous nursery and NICU outbreaks of enteroviral infection have been reported.^{163, 164} In the neonate with enteroviral infection, clinical manifestations can range from mild gastroenteritis to a severe and fulminant sepsis-like syndrome or meningitis/encephalitis. The latter presentation can be associated with a high mortality rate.¹⁶⁴ In index cases, the patient may have acquired disease vertically, with subsequent horizontal spread leading to outbreaks^{164, 165}; with other viral pathogens, virus can be shed into the stool for prolonged periods, enabling patient-to-patient transmission by the hands of HCWs when hand hygiene procedures are improperly performed.

Cytomegalovirus

Congenitally acquired cytomegalovirus (CMV) infection is a cause of morbidity and occasionally death, whereas postnatally acquired CMV infection follows a benign course in virtually all healthy term infants. Postnatal CMV infection, however, can cause considerable morbidity and death in premature infants.¹⁶⁶ Hepatitis, neutropenia, thrombocytopenia, sepsis-like syndrome, pneumonitis, and development of chronic lung disease each have been associated with postnatal acquisition of CMV in premature infants.^{167, 168} With the routine use of CMV-seronegative blood products in these neonates, a majority of postnatal CMV infections appear to be acquired through breast milk.¹⁶⁹ It has been estimated that transmission by this mode occurs in approximately 37% of breast-fed infants of mothers with CMV detected in breast milk.¹⁷⁰ In one study, approximately 50% of these infants had clinical features of infection, and 12% presented with a sepsis-like syndrome. Nosocomial person-to-person transmission has been documented,^{171, 172} but the extent to which this occurs is controversial.¹⁷³ At present, no proven, highly effective method is available for removing CMV from breast milk without destroying its beneficial components. Some data, however, suggest that freezing the breast milk before use may decrease the CMV titer, thereby limiting subsequent transmission.¹⁷⁴

Herpes Simplex Virus

In a majority of cases, neonatal herpes simplex virus (HSV) infection is acquired vertically from the mother. Nursery transmission of HSV infection is rare but has been described.^{175, 176, 177} In each of these cases, HSV-1 was involved. In one infant, the source of virus was thought to be a patient’s father, who had active herpes labialis.¹⁷⁵ Subsequent spread of virus from this first infant to a second infant was thought to have occurred by means of the hands of an HCW. In another report, the source of HSV for the index case, an infant who died of respiratory distress in whom evidence of HSV infection was found at postmortem examination of the brain, was unknown.¹⁷⁶ The hands of HCWs were implicated in the spread of HSV to three subsequent cases, however. In another report, direct spread from an HCW was thought to be responsible for transmission of HSV to three infants over a period of approximately 3 years.¹⁷⁷ Studies of adults with herpes labialis suggest a high frequency of recovery of virus from the mouth and the hands (78% and 67%, respectively).¹⁷⁸ In this same study, HSV was shown to survive for 2 to 4 hours on skin, cloth, and plastic. Implementing contact precautions for infants with HSV and instructing HCWs with active herpes labialis regarding control measures, such as covering the lesion, not touching the lesion, and using strict hand hygiene, are reasonable means to prevent nosocomial transmission of HSV. If there are concerns that an HCW would be unable to comply with control measures or if the HCW has a herpetic whitlow, such persons should be restricted from patient contact.

Varicella-Zoster Virus

Nosocomial transmission of varicella in the NICU setting, although unusual, has been described.¹⁷⁹ Large-scale outbreaks in nurseries and NICUs are rare, most probably because of the high rate of varicella-zoster virus (VZV) immunity in HCWs and pregnant women. Premature infants born at less than 28 weeks of gestation are unlikely to have received protective levels of VZV IgG from their mothers, so their potential risk is significant if an exposure occurs. Transmission is most likely to occur from an adult with early, unrecognized symptoms of varicella. In such instances, the potential risk for VZV-seronegative exposed infants and HCWs is substantial, especially if the patient in the index case is an HCW.¹⁸⁰ For this reason, it is recommended that HCWs be screened for prior varicella infection by history, with subsequent immunization as indicated.

Hepatitis A

Hepatitis A is a rare cause of nosocomial infection in NICUs, but a number of outbreaks in this setting have been reported.^{181, 182, 183} In most instances, disease in neonates is clinically silent. Neonatal cases often are detected only through recognition of the symptomatic secondary adult cases. In one report, disease was acquired by patients in the index cases through blood transfusion from a donor with acute hepatitis A.¹⁸³ Of note, the virus subsequently spread to another 11 infants, 22 nurses, and 8 other HCWs. Overall, hepatitis A affected 20% of the patients and 24% of the nurses. Lapses in infection control practices and the prolonged shedding of the virus in infants stool probably contributed to the rapid spread and high attack rate documented in the outbreak. Outbreaks such as this one are unlikely because of current blood product practices to eliminate transmissible agents from donor blood.

PREVENTION AND CONTROL

An effective infection control program that focuses on reducing risk on a prospective basis can decrease the incidence of nosocomial infections.^{184, 185} The principal function of such a program is to protect the infant and the HCW from risk of hospital-acquired infection in a manner that is cost-effective. Activities crucial to achieving and maintaining this goal include collection and management of critical data relating to surveillance for nosocomial infection, and direct intervention to interrupt the transmission of infectious diseases.²²

Surveillance

Reducing the incidence of nosocomial infection for neonates must begin with surveillance for these events. Surveillance has been defined as “a comprehensive method of measuring outcomes and related processes of care, analyzing the data, and providing information to members of the health care team to assist in improving those outcomes.”¹⁸⁶ Essential elements of a surveillance program include the following:

• •Defining the population and data elements as concisely as possible
• •Collecting relevant data using systematic methods
• •Consolidating and tabulating data to facilitate evaluation
• •Analyzing and interpreting data
• •Reporting data to those who can bring about change¹⁸⁷

Surveillance systems necessarily vary, depending on the population; accordingly, a written plan, based on sound epidemiologic principles,¹⁸⁷ should be in place to track rates of infection over time. Because new risks can emerge, such as new interventional technology or drugs, changing patient demographics, and new pathogens and resistance patterns, the plan should be reviewed and updated frequently.¹⁸⁸ The Joint Commission on Accreditation of Healthcare Organizations (JCAHO) recommends that hospitals have a written infection control plan that includes a description of prioritized risks; a statement of the goals of the infection control program; a description of the hospital’s strategies to minimize, reduce, or eliminate the prioritized risks; and a description of how the strategies will be evaluated. The JCAHO further recommends that hospitals identify risks for transmission and acquisition of infectious agents (Table 35–3) and formally review this analysis annually and whenever significant changes occur in any of the risk factors.¹⁸⁹

Table 35-3.

Factors Influencing Transmission and Acquisition of Infectious Agents within a Hospital

Geographic location and community environment of the hospital

Characteristics of the population served

Care, treatment, and services provided

Actions after analysis of the hospital’s infection prevention and control data

Data from Joint Commission on Accreditation of Healthcare Organizations. Surveillance, Prevention and Control of Infection, 2005 Pre-publication edition. Oak Brook Terrace, Ill, Joint Commission on Accreditation of Healthcare Organizations, 2003, pp 1–11. Available at http://www.jcaho.org/accredited+organizations/patient+safety/infection+control/05_ic_std_hap.pdf.

The definitions provided by the CDC’s NNIS system are the most comprehensive and widely used set of case definitions for the outcomes of nosocomial infections.¹⁸⁹ The system provides high-risk nursery–specific data collection methods as well as denominator data and allows external benchmarking of infection rates for this population.^{190, 191} The NNIS system defines a nosocomial infection as a localized or systemic process that results from adverse reaction to the presence of an infectious agent(s) or its toxin(s) and that was not present or incubating at the time of admission to the hospital. NNIS also recognizes as special situations, and defines as nosocomial, some infections in neonates that result from passage through the birth canal but do not become clinically apparent until several or more days after birth. It does not, however, consider infections that are known or proved to have been acquired transplacentally to be nosocomial.¹⁹¹

Distinction between maternal and hospital sources of infection is important, although difficult at times, because control measures designed to prevent acquisition from hospital sources will be ineffective in preventing perinatal acquisition of pathogens.¹⁹² Surveillance for infections in healthy newborns also is challenging because of the typically short length of stay. Infections can develop after discharge, and these are more difficult for infection control practitioners (ICPs) to capture. Methods for postdischarge surveillance have been developed, but because most neonatal infections that occur following discharge are noninvasive,¹⁹³ such surveillance has not been widely implemented, because of concerns about the cost-effectiveness of these labor-intensive processes.

The ultimate goal of surveillance is to achieve outcome objectives (e.g., decreases in infection rates, morbidity, mortality, or cost).¹⁸⁷ Baseline infection rates for an inpatient unit must be established so that the endemic rate of infection can be understood and addressed. In the NICU, concurrent surveillance (initiated while the infant is in the hospital) should be conducted by persons trained to collect and interpret clinical information. Typically, such persons are ICPs working closely with HCWs and using various data sources (Table 35–4).

Table 35-4.

Sources of Surveillance Data

Admission records

Patient records

Closed medical records

Kardex

Temperature and vital signs records

Microbiology reports

Antimicrobial susceptibility reports

Verbal and written reports

Radiographic reports

Interviews with caregivers

Interviews with and observation of patient

Postdischarge reports

Autopsy reports

Data from Lee TB, Baker OG, Lee JT, et al. Recommended practices for surveillance. Association for Professionals in Infection Control and Epidemiology, Inc. Surveillance Initiative Working Group. Am J Infect Control 26:277–288, 1998.

Using NNIS or other accepted definitions, the ICP should collect data regarding cases of nosocomial infection in the NICU population as well as population-specific denominator data. Denominators must be carefully chosen to represent the population at risk. Attempts to stratify risk should take into account both underlying infant-specific risks and those resulting from therapeutic or diagnostic interventions.¹⁸⁶ Risk stratification techniques that attempt to control for distribution of risk have included severity of illness score, intensity of care required, and birth weight.⁷⁴ Because the risk for developing nosocomial infection is greater for lower-birth-weight infants,⁴⁸ the NNIS system breaks down data collection and analysis into birth weight categories (Tables 35–5 and 35–6).^{191, 194} The use of invasive devices, however, also is an important factor to consider. The appropriate denominator for an infection related to the use of a medical device, such as a CVC-related primary bloodstream infection, according to NNIS, would be total device days for the population during the surveillance period.

Table 35-5.

National Nosocomial Infections Surveillance NICU Birth Weight Categories

≤1000 g

1001–1500 g

1501–2500 g

>2500 g

NICU, neonatal intensive care unit.

Data from Centers for Disease Control and Prevention, Division of Health Care Quality Promotion. National Nosocomial Infections Surveillance (NNIS) system report, data summary from January 1992 to June 2003, issued August 2003. Am J Infect Control 30:481–498, 2003.

Table 35-6.

Distribution of Device-Associated Infection Rates by Birth Weight Category^a

Birth Weight (g)	Pooled Mean
	Umbilical and CR-BSIb	VAPc
≤1000	10.6	3.3
1001–1500	6.4	2.5
1501–2500	4.1	2.1
>2500	3.7	1.4
NICU, neonatal intensive care unit; VAP, ventilator-associated pneumonia.

^aNICU component of reported data, January 1995 to June 2003 (VAP data are for January 2002 through June 2003 only).

^bNumber of umbilical and central line–related (CR) bloodstream infections (BSIs) × 1000/number of umbilical and central line days.

^cNumber of VAP cases × 1000/number of ventilator days.

Data from Centers for Disease Control and Prevention, Division of Health Care Quality Promotion. National Nosocomial Infections Surveillance (NNIS) system report, data summary from January 1992 to June 2003, issued August 2003. Am J Infect Control 30:481–498, 2003.

The formula generally used for calculating nosocomial infection rates is (x/y)k, where x equals the number of events (infections) over a specific time period, y equals the population at risk for development of the outcome, and k is a constant and a multiple of 10. Rates can be expressed as a percentage (k = 100), although device-related infections usually are expressed as events per 1000 device days (k = 1000). A value should be selected for k that results in a rate greater than 1.¹⁸⁷

Because use of invasive devices is such a significant risk factor both for bloodstream infection and ventilator-associated pneumonia, assessing NICU practices with device use may be warranted. NNIS provides a benchmark for NICU device utilization broken down into birth weight categories. An NICU device utilization ratio can be calculated using the following formula:

In those units with device utilization ratios above the NNIS 90th percentile, investigation into the practices surrounding use of invasive devices may be warranted.¹⁹⁵ Calculating monthly and annual rates to employ as benchmarks can assist in identification of a potential problem in device-related procedures.

Surveillance data must be arranged and presented in a way that facilitates interpretation, comparison both directed internally and with comparable external benchmarks, and dissemination within the organization. Quality improvement tools (e.g., control and run charts) can be useful for these purposes. Statistical tools should be used to determine the significance of findings, although statistical significance should always be balanced with the evaluation of clinical significance.¹⁸⁷ External benchmarking through interhospital comparison is a valuable tool for improving quality of care^{196, 197} but should be performed only when surveillance methodologies (e.g., case definitions, case finding, data collection methods, intensity of surveillance)¹⁸⁷ can reasonably be assumed to be consistent between facilities.

Few overall infection rates in NICUs are available, but a small study done in 17 children’s hospitals performing NICU nosocomial infection surveillance reported a median nosocomial infection rate of 8.9 infections per 1000 patient days (range, 4.6 to 18.1).¹⁹⁸ NNIS does not provide a benchmark for overall infection rates within NICUs. Instead, NNIS provides birth weight–stratified device-associated infection rates for umbilical and central intravascular line–associated bloodstream infections. The most recent rates for catheter-related bloodstream infections (137 to 143 NICUs reporting) and ventilator-associated pneumonias (78 to 96 NICUs reporting) are summarized in Table 35–6.¹⁹⁴

Once arranged and interpreted, nosocomial infection data must be shared with personnel who can effect change and implement infection control interventions. Written reports summarizing the data and appropriate control charts should be provided to the facility’s infection control committee, unit leaders, and members of the hospital administration on an ongoing basis. The interval between reports is determined by the needs of the institution. In addition to formal written reports, face-to-face reports are appropriate in the event of identification of a serious problem or an outbreak. ICPs can serve as consultants to assist NICU or neonatology service leaders in addressing infection rate increases or outbreak management.

Outbreak (Epidemic) Investigation

Surveillance activities typically identify endemic nosocomial infections (i.e., those infections that represent the usual level of disease within the nursery or NICU).¹⁹⁹ Although the rate can fluctuate over time, in the absence of interventions that successfully reduce risk of infection, the difference rarely is statistically significant. Establishing an NICU’s endemic infection rate and expected variation around that rate allows the ICP to rapidly identify unusual increases in rates that may indicate on outbreak (epidemic) of a particular infection. Using baseline surveillance data along with aggregate data from sources such as the NNIS system allows the ICP to develop meaningful threshold rates for initiating outbreak investigation.¹⁸⁸ Alternatively, HCWs can be the first to sense an increase in infections, which then can be confirmed or refuted by surveillance data.²⁰⁰ Even a single case of infection due to an unusual and potentially dangerous pathogen (e.g., Salmonella) can constitute the index case for a subsequent outbreak and thus merits rapid and comprehensive investigation. Outbreaks may need to be reported to health authorities, depending on local and state requirements as well as the organism involved.

Numerous studies have described nursery and NICU epidemics caused by a variety of pathogens (Table 35–7), and most such epidemics have required the coordinated efforts of ICPs, NICU leadership, staff, and hospital administration for resolution.^{79, 106, 115, 136, 157, 161, 162, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212} Outbreak investigation and intervention should be approached systematically, applying sound epidemiologic principles. In general, the process should include the following^{199, 200}:

• Confirming that an outbreak exists, by comparing the outbreak infection rate with baseline data (or with rates reported in the literature if baseline data are not available), and communicating concerns to stakeholders within the institution (and to those in other agencies if notification of health authorities is necessary)

• Assembling the appropriate personnel to assist in developing a case definition and in planning immediate measures to prevent new cases

• Performing active surveillance using the case definition to search for additional infections and collecting critical data and specimens

• Characterizing cases of infection by person, place, and time, including plotting of an epidemic curve (to facilitate identification of shared risk factors among involved patients, such as invasive devices, proximity to other infected patients or temporal association with infection in such patients, common underlying diagnoses, shared medical or nursing staff, surgery, and medications, including antimicrobial agents)

• Formulating a working hypothesis and testing this hypothesis (if the severity of the problem warrants this level of study, and provided that the institution has and can commit the necessary resources), with use of analytic approaches, including case-control and cohort studies, as appropriate to determine the likely cause of the outbreak

• Instituting and evaluating control measures, which can be implemented anywhere in the foregoing process (more directed measures become possible as more is learned about the outbreak, and efficacy of control measures can be judged on whether the outbreak resolves, as indicated by return of number of cases to endemic levels or by cessation of occurrence of infections)

• Reporting findings to appropriate personnel, including unit staff, hospital administration, and public health authorities (if involved in management of the outbreak), in comprehensive written reports, including summaries of how the outbreak was first recognized, study and analysis methods used, interventions implemented to resolve the epidemic, results, and a discussion of any other important outcomes or surveillance and control measures identified

Table 35-7.

Reported Nursery Outbreaks of Infection

Causative Organism	Source	Reference	Year	Location
Staphylococcus aureus (pyoderma)	Hospital staff	212	2002	Taipei, Taiwan
Staphylococcus aureus	Skin barrier paste (Stomahesive)	206	2000	Leeds, UK
MRSA	Horizontal transmissiona?	207	2001	Washington, DC
Enterococcus faecium (VRE)	Unknown	79	2001	Omaha
Clostridium species	Horizontal transmission?	210	2002	Manitoba, Canada
Serratia marcescens	Horizontal transmission	204	1998	Leipzig, Germany
S. marcescens	Milk bottles	115	2002	Zurich
Enterobacter sakazakii	Powdered milk formula	201	1998	Brussels
Klebsiella pneumoniae, antibiotic-resistant	Environment, breast milk, horizontal transmission	106	2001	London
Acinetobacter species	Air conditioners	202	1996	Bahamas
Pseudomonas aeruginosa	Unknown	203	1999	Maryland
Chryselbacterium meningosepticum	Sink taps	211	1996	London
Salmonella enterica	Horizontal transmission	209	1999	Rio de Janeiro
Tuberculosis, multidrug-resistant	Hospital staff?	136	1998	New York
Adenovirus type 8	Horizontal transmission?	205	1998	Heidelberg, Germany
Parainfluenza virus type 3	Horizontal transmission?	161	1996	Winnipeg, Canada
Influenza A virus	Unknown	162	1999	Barcelona, Spain
Respiratory syncytial virus	Unknown	208	2002	Riyadh, Saudi Arabia
Rotavirus	Environment	157	2002	Holland
MRSA, methicillin-resistant Staphylococcus aureus; VRE, vancomycin-resistant enterococci.

^aHorizontal transmission refers to indirect contact transmission by contaminated equipment or health care workers’ hands.

Interventions used to control and limit outbreaks usually have consisted of isolation and cohorting of infected or colonized infants to prevent transmission of organisms. Transmission-based precautions, a system developed by the CDC, can be used to determine the most effective barrier precautions to use with affected patients. Cohorting, or placing infants infected or colonized with the outbreak organism together in geographically segregated areas and assigning dedicated staff and equipment to their care, also has been used successfully to halt outbreaks in nurseries and NICUs. In extreme cases, closure of an NICU to admissions has been necessary to bring an outbreak under control.^{57, 106, 161}

Every attempt should be made to identify the source of a nursery outbreak, although this is not always possible. Sources implicated in NICU outbreaks have included medications, equipment, and enteral feeding solutions; person-to-person transmission and environmental reservoirs also have been reported.^{102, 106, 157, 201, 202, 213} Efforts to identify the source may include culturing of specimens from HCWs, equipment, and the environment, although careful consideration should be given to the potential benefits before initiating these measures. Culture of samples from the environment and equipment, in view of the vast number of objects that could be contaminated, usually is not helpful in identifying the source of an outbreak unless specific case characteristics or microbiologic data strongly suggest the location. Culture of specimens obtained from HCWs when person-to-person transmission is suspected may be more likely to identify the source of an outbreak, but it must be remembered that an HCW whose culture specimen yields the outbreak organism may have been transiently colonized while working with an affected infant, rather than constituting the source of the infection. Management of culture-positive HCWs (possible furlough, treatment, and return to work criteria) should be planned in advance of widespread culture surveillance and should involve supervisors of affected employees and occupational health services.²⁰⁰

Standard and Transmission-Based Precautions in the Nursery

The most widely accepted guideline for preventing the transmission of infections in hospitals was developed by the CDC. Most recently revised in 1996, the system contains two tiers of precautions. The first and most important, standard precautions, was designed for the management of all hospitalized patients regardless of their diagnosis or presumed infection status. The second, transmission-based precautions, is intended for patients documented or suspected to be infected or colonized with highly transmissible or epidemiologically important pathogens for which additional precautions to interrupt transmission are needed.²⁵

Standard precautions are designed to reduce the risk of transmission of microorganisms from both recognized and unrecognized sources and are to be followed for the care of all patients, including neonates. They apply to blood; all body fluids, secretions, and excretions except sweat; nonintact skin; and mucous membranes. Components of standard precautions include hand hygiene and wearing gloves, gowns, and masks and other forms of eye protection.

Hand hygiene plays a key role for caregivers in the reduction of nosocomial infection for patients^{27, 214} and in prevention of nosocomial or health care–associated infections. Hand hygiene should be performed before and after all patient contacts; before donning sterile gloves to perform an invasive procedure; after contact with blood, body fluids or excretions, mucous membranes, nonintact skin, and wound dressings; in moving from a contaminated body site to a clean body site during patient care (i.e., from changing a diaper to performing mouth care); after contact with inanimate objects in the immediate vicinity of the patient; after removing gloves; and before eating and after using the restroom.¹³¹ When hands are visibly soiled or contaminated with proteinaceous materials, blood, or body fluids, and after using the restroom, hands should be washed with antimicrobial soap and water. Soaps containing 2% to 4% chlorhexidine gluconate or 0.3% triclosan¹⁵⁶ are recommended for hand washing in nurseries.¹⁹² When hands are not visibly soiled, alcohol-based hand rubs, foams, or gels are an important tool for hand hygiene. Compared with washing with soap and water, use of the alcohol-based products is at least as effective against a variety of pathogens and requires less time, and these agents are less damaging to skin. The CDC “Guideline for Hand Hygiene in the Health Care Setting” calls for use of alcohol hand rubs, foams, or gels as the primary method to clean hands, except when hands are visibly soiled.¹³¹ Programs that have been successful in improving hand hygiene and decreasing nosocomial infection have used multidisciplinary teams to develop interventions focusing on use of the alcohol rubs in the setting of institutional commitment and support for the initiative.^{27, 215}

HCWs should wash hands and forearms to the elbows on arrival in the nursery. A 3-minute scrub has been suggested,²¹⁶ but consensus on optimal duration of initial hand hygiene is lacking. At a minimum, the initial wash should be long enough to ensure thorough washing and rinsing of all parts of the hands and forearms. Routine hand washing throughout care delivery should consist of wetting the hands, applying product, rubbing all surfaces of the hands and fingers vigorously for at least 15 seconds, rinsing, and patting dry with disposable towels.¹³¹ Wearing hand jewelry has been associated with increased microbial load on hands. Whether this results in increased transmission of pathogens is not known. Many experts, however, recommend that hand and wrist jewelry not be worn in the nursery.^{217, 218} In addition, the CDC guideline states that staff who have direct contact with infants in NICUs should not wear artificial fingernails or nail extenders.¹³¹ Only natural nails kept less than ¹/₄ inch long should be allowed.

Clean, nonsterile gloves are to be worn whenever contact with blood, body fluids, secretions, excretions, and contaminated items is anticipated. The HCW should change gloves when moving from dirty to clean tasks performed on the same patient, such as after changing a diaper and before suctioning a patient, and whenever they become soiled. Because hands can become contaminated during removal of gloves, and because gloves may have tiny, unnoticeable defects, wearing gloves is not a substitute for hand hygiene. Hand hygiene must be performed immediately after glove removal.²⁵

Personnel in nurseries including the NICU historically have worn cover gowns for all routine patient contact. The practice has not been found to reduce infection or colonization in neonates and is unnecessary.^{219, 220} Instead, CDC guidelines recommend nonsterile, fluid-resistant gowns to be worn as barrier protection when soiling of clothing is anticipated and in performing procedures likely to result in splashing or spraying of body substances.²⁵ Possible examples of such procedures in the NICU are placing an arterial line and irrigating a wound. The Perinatal Guidelines of the American Academy of Pediatrics and the American College of Obstetricians and Gynecologists recommend that a long-sleeved gown be worn over clothing when a neonate is held outside the bassinette by nursery personnel.²²¹

Nonsterile masks, face shields, goggles, and other eye protectors are worn in various combinations to provide barrier protection and should be used during procedures and patient care activities that are likely to generate splashes or sprays of body substances and fluids.²⁵

Standard precautions also require that reusable patient care equipment be cleaned and appropriately reprocessed between patients; that soiled linen be handled carefully to prevent contamination of skin, clothing, or the environment; that sharps (i.e., needles, scalpels) be handled carefully to prevent exposure to blood-borne pathogens; and that mouthpieces and other resuscitation devices be used, rather than mouth-to-mouth methods of resuscitation.²⁵

In addition to standard precautions, which must be used for every patient, the CDC recommends transmission-based precautions when the patient is known or suspected to be infected or colonized with epidemiologically important or highly transmissible organisms. Always used in addition to standard precautions, transmission-based precautions comprise three categories: contact precautions, droplet precautions, and airborne precautions.

Contact precautions involve the use of barriers to prevent transmission of organisms by direct or indirect contact with the patient or contaminated objects in the patient’s immediate environment.²⁵ Sources of indirect contact transmission in nurseries can include patient care equipment such as monitor leads, thermometers, isolettes, breast pumps,¹⁸⁶ toys, and instruments and contaminated hands.²²²

The patient requiring contact precautions should be placed in a private room whenever possible but, after consultation with an infection control practitioner, can be cohorted with a patient infected with the same microorganism but no other infection.²⁵ Many nurseries, however, have few if any isolation rooms. The American Academy of Pediatrics states that infected neonates requiring contact precautions can be safely cared for without an isolation room if staffing is adequate to allow appropriate hand hygiene, a 4- to 6-foot-wide space can be provided between care stations, adequate hand hygiene facilities are available, and staff members are well trained regarding infection transmission modes.²²¹

HCWs should wear clean, nonsterile gloves when entering the room or space of a patient requiring contact precautions and should wear a cover gown when their clothing will have contact with the infant, environmental surfaces, or items in the infant’s area. A cover gown also should be worn when the infant has excretions or secretions that are not well contained, such as diarrhea or wound drainage, which may escape the diaper or dressing. Infant care equipment should be dedicated to the patient if possible so that it is not shared with others.²⁵

Examples of conditions in the neonate that require contact precautions include neonatal mucocutaneous herpes simplex virus infection, respiratory syncytial virus infection, varicella (also see airborne precautions), infection or colonization with a resistant organism such as MRSA or a multiple drug–resistant gram-negative bacillus, and congenital rubella syndrome.

Droplet precautions are intended to reduce the risk of transmission of infectious agents in large-particle droplets from an infected person. Such transmission usually occurs when the infected person generates droplets during coughing, sneezing, or talking, or during procedures such as suctioning. These relatively large droplets travel only short distances and do not remain suspended in the air, but can be deposited on the conjunctiva, nasal mucosa, and/or mouth of persons working within 3 feet of the infected patient.²⁵ Patients requiring droplet precautions should be placed in private rooms (see earlier discussion of isolation rooms in nurseries in the paragraph on contact precautions), and staff should wear masks when working within 3 feet of the patient.²⁵ Examples of conditions in the neonate that necessitate droplet precautions are pertussis and invasive N. meningitidis infection.

Airborne precautions are designed to reduce the risk of airborne transmission of infectious agents.²⁵ Because of their small size, airborne droplet nuclei and dust particles containing infectious agents or spores can be widely spread on air currents or through ventilation systems and inhaled by or deposited on susceptible hosts. Special air-handling systems and ventilation are required to prevent transmission. Patients requiring airborne precautions should be placed in private rooms in negative air-pressure ventilation with 6 to 12 air changes per hour. Air should be externally exhausted or subjected to high-efficiency particulate air (HEPA) filtration if it is recirculated.²²²

Examples of conditions in the neonate for which airborne precautions are required are varicella-zoster virus infections and measles. Susceptible HCWs should not enter the rooms of patients with these viral infections. If assignment cannot be avoided, susceptible staff members should wear masks to deliver care. If immunity has been documented, staff members need not wear masks.²²² Airborne precautions also are required for active pulmonary tuberculosis, and although neonates are rarely contagious, the CDC recommends isolating patients while they are being evaluated.²¹³ A more important consideration is the need to isolate the family of a suspected tuberculosis patient until an evaluation for pulmonary tuberculosis has been completed, because the source of infection frequently is a member of the child’s family.^{223, 224}

Physical Environment

Before the 1990s, well-baby nurseries and many NICUs were constructed as large, brightly lit open wards with rows of incubators surrounded by equipment. Sinks could be provided in such rooms only around the periphery, limiting access to hand hygiene facilities for staff and families. In these NICUs, parents’ time with their infant was severely restricted, and the units were designed for the convenience and function of the HCW.²²⁵ More recently, perinatal care professionals have come to understand that neonates (and especially preterm infants) can benefit from a quiet, soothing atmosphere and protection from unnecessary light, noise, handling, uncomfortable positioning, and sleep disruptions.²²⁶

If infants are kept in a central nursery rather than rooming-in with mothers, at least 30 square feet of floor space should be provided per neonate, and bassinets should be at least 3 feet apart.²¹⁶ Teams designing units delivering higher levels of perinatal care, including NICUs, should plan individual bed areas large enough for families to stay at the bedside for extended periods of time without interfering with the staff’s ability to care for the child. If individual rooms cannot be provided, at least 150 square feet of floor space should be allowed for each neonate in an NICU, incubators or overhead warmers should be separated by at least 6 to 8 feet, and aisles should be at least 8 feet wide.^{216, 227}

A scrub sink with foot, knee, or touchless (electronic sensor) controls should be provided at the entrance to every nursery and should be large and deep enough to control splashing. Sinks in patient care areas should be provided at a minimum ratio of 1 sink for at least every 6 to 8 stations in the well-baby nursery and 1 sink for every 3 or 4 stations in higher-level nurseries, including the NICU.²¹⁶ Every bed position should be within 20 feet of a hand-washing sink and accessible for children and persons in wheelchairs.²²⁷ For NICUs composed of individual rooms, a hand-washing sink should be located in each room near the door to facilitate hand hygiene on entering and leaving the room.

Environmental surfaces should be designed so that they are easy to clean and do not harbor microorganisms. Sink taps and drains, for instance, have been implicated in outbreaks of infection.^{228, 229} Installing sinks with seamless construction may minimize this risk by decreasing areas where water can pool and microorganisms proliferate. Faucet aerators have been implicated in outbreaks of infection and should be avoided in the intensive care unit.²³⁰ Although carpeting can reduce noise levels in a busy NICU, the CDC “Guidelines for Environmental Infection Control in Health-Care Facilities” recommend against use of carpeting in areas where spills are likely, including intensive care units. The guidelines further recommend against upholstered furniture in NICUs.²³¹ If, for reasons of noise reduction and developmentally appropriate care, porous surfaces such as carpeting and cloth upholstery are selected for the NICU, cleaning must be performed carefully. Carpet should be vacuumed regularly with equipment fitted with HEPA filters, and upholstered furniture should be removed from inpatient areas to be cleaned.

Attention also should be paid to air-handling systems. According to the Perinatal Guidelines, minimal standards for inpatient perinatal care areas include six air changes per hour, and a minimum of two changes should consist entirely of outside air. Air delivered to the NICU should be filtered with at least 90% efficiency. In addition, nurseries should include at least one isolation room capable of providing negative pressure vented to the outside, observation windows with blinds for privacy, and the capability for remote monitoring.^{227, 232}

General Housekeeping

Floors and other horizontal surfaces should be cleaned daily by trained personnel using Environmental Protection Agency (EPA)–registered hospital disinfectants/detergents. These products (including phenolics and other chemical surface disinfectants) must be prepared in accordance with manufacturers’ recommendations and used carefully to avoid exposing neonates to these products. Phenolics should not be used on surfaces that come in direct contact with neonates’ skin.²³¹ High-touch areas, such as counter tops, work surfaces, doorknobs, and light switches, may need to be cleaned more frequently because they can be heavily contaminated during the process of delivering care. Hard, nonporous surfaces should be “wet dusted” rather than dry dusted, to avoid dispersing particulates into the air, and then disinfected using standard hospital disinfectants.²³¹ Sinks should be scrubbed daily with a disinfectant detergent. Walls, windows, and curtains should be cleaned regularly to prevent dust accumulation, but daily cleaning is not necessary unless they are visibly soiled.

Bassinets and incubators should be cleaned and disinfected between infants, but care must be taken to rinse cleaning products from surfaces with water before use. Care units should not be cleaned with phenolics or other chemical germicides during an infant’s stay. Instead, infants who remain in the nursery for long periods of time should periodically be moved to freshly cleaned and disinfected units.²³¹

Patient care equipment must be cleaned, disinfected, and, when appropriate, sterilized between patients. Sterilization (required for devices that enter the vascular system, tissue, or sterile body cavities) and higher levels of disinfection (required for equipment that comes in contact with mucous membranes or that has prolonged or intimate contact with the newborn’s skin) must be performed under controlled conditions in the central processing department of the hospital. Examples of patient care equipment that require these levels of processing are endotracheal tubes, resuscitation bags, and face masks.^{216, 232, 233} Low-level disinfection is required for less critical equipment, such as stethoscopes or blood pressure cuffs, and usually can be performed at point of use (e.g., the bedside), although this type of equipment should be dedicated to individual patients whenever possible.

Linens

Requirements for linen handling and management for neonates do not vary appreciably from those for other hospitalized patients. Although soiled linen can contain large numbers of organisms capable of causing infections, transmission to patients appears to be rare. Studies suggesting linen as a source of infection often have failed to confirm it as the source of infection.²³⁴ At least one report, however, has implicated linen in the transmission of group A streptococci.⁸⁵ Investigation of this outbreak revealed that clothing worn by the neonates was being washed in the local hospital “mini laundry,” rather than being processed under the usual laundry contract. Investigation of the dryers revealed extensive contamination with the outbreak organism. This case illustrates the importance of having standard hospital laundry protocols and ensuring that appropriate water and dryer temperatures are maintained. When such protocols are followed, the mechanical actions of washing and rinsing, combined with hot water and/or the addition of chemicals such as chlorine bleach, and a final commercial dryer and/or ironing step significantly reduce bacterial counts.^{235, 236} Few hospitals in the United States use cloth diapers, but regardless of type used, soiled diapers should be carefully bagged in plastic and removed from the unit every 8 hours.²¹⁶

Health Care Workers

HCWs caring for neonates have the potential both to transmit infections to infants and to acquire infections from patients. Educating HCWs about infection control principles is crucial to preventing such transmission. Hospitals should provide education about infection control policies, procedures, and guidelines to staff in all job categories during new employee orientation and on a regular basis throughout employment. The content of this education should include hand hygiene, principles of infection control, the importance of individual responsibility for infection control, and the importance of collaborating with the infection control department in monitoring and investigating potentially harmful infectious exposures and outbreaks.

Transmission of infectious organisms between patients and HCWs has been well documented. Several studies have indicated that a high proportion of HCWs acquire RSV (34% to 56%) when working with infected children, and these workers appear to be important in the spread of the illness within hospitals.^{237, 238, 239} Although 82% of the infected HCWs in one of these RSV studies were asymptomatic, staff should be aware of the importance of self-screening for communicable disease. They should be encouraged to report personal infectious illnesses to supervisors, who in turn should report them to occupational health services and infection control. In general, HCWs with respiratory, cutaneous, mucocutaneous, or gastrointestinal infections should not deliver direct patient care to neonates.²¹⁶ In addition, seronegative staff members exposed to illnesses, such as varicella and measles, should not work during the contagious portion of the incubation period.²³²

Staff members with HSV infection rarely have been implicated in transmission of the virus to infants and thus do not need to be routinely excluded from direct patient care. Those with herpes labialis or cold sores should be instructed to cover the lesions and not to touch their lesions, and to comply with hand hygiene policies. Persons with genital lesions also are unlikely to transmit HSV so long as hand hygiene policies are followed. However, HCWs who are unlikely or unable to comply with the infection control measures and those with herpetic whitlow should not deliver direct patient care to neonates until lesions have healed.²⁴⁰

Acquisition of CMV often is a concern of pregnant HCWs because of the potential effect on the fetus. Approximately 1% of newborn infants in most nurseries and a higher percentage of older children (up to 70% of children 1 to 3 years of age in child-care centers) excrete CMV without clinical manifestations.²⁴¹ The risk of acquiring CMV infection has not been shown to be higher for HCWs than for the general population.^{242, 243} For this reason, pregnant caregivers need not be excluded from the care of neonates suspected to be shedding CMV. They should be advised of the importance of standard precautions.

HCWs in well-baby nurseries and NICUs should be as free from transmissible infectious diseases as possible,²¹⁶ and ensuring that they are immune to vaccine-preventable diseases is an essential part of a personnel health program. The CDC recommends several immunizations for health care personnel (Table 35–8).

Table 35-8.

Immunizing Agents Strongly Recommended for Health Care Workers

Vaccine	Recommendation
Hepatitis B recombinant vaccine	Vaccinate all HCWs at risk of exposure to blood and body fluids.
Influenza vaccine	Vaccinate HCWs annually.
Measles live-virus vaccine	Vaccine should be considered for all HCWs, including those born before 1957, who have no proof of immunity (receipt of 2 doses of live vaccine on or after first birthday, physician-diagnosed measles, or serologic evidence of immunity).
Mumps live-virus vaccine	HCWs believed to be susceptible can be vaccinated; adults born before 1957 can be considered immune.
Rubella live-virus vaccine	HCWs, both male and female, who lack documentation of receipt of vaccine on or after first birthday or serologic evidence of immunity should be vaccinated; adults born before 1957 can be considered immune, except women of childbearing age.
Varicella-zoster live-virus vaccine	HCWs without a reliable history of varicella or serologic evidence of varicella immunity should be vaccinated.
HCW, health care worker.

Data from Bolyard EA, Tablan OC, Williams WW, et al. Guideline for infection control in health care personnel, 1998. Hospital Infection Control Practices Advisory Committee. Infect Control Hosp Epidemiol 19:407–463, 1998.

Staffing levels in a patient care setting also can affect patient outcomes. A number of studies suggest that as patient-to-nurse ratios in intensive care units increase, so do nosocomial infections and mortality rates.^{47, 132, 244} Although optimal staffing ratios have not been established for NICUs and will vary according to characteristics of individual units, one study demonstrated that the incidence of clustered S. aureus infections was 16 times higher after periods when the infant-to-nurse ratio exceeded 7:1. Decreased compliance with hand hygiene during a period of understaffing frequently is cited as contributing to nosocomial infection rate increases.⁹⁸ Further study is necessary to determine best practice surrounding staffing levels in NICUs.

Family-Centered Care: Parents and Visitors to the Newborn Infant

The first NICUs in the late 1960s grouped infants together in large, brightly lit rooms with incubators placed in rows. Parents were allowed very little time with their babies and even less physical contact. In the decades since, it has been recognized that “the parent is the most important caregiver and constant influence in an infant’s life”²²⁵ and that HCWs working in NICUs should encourage parents to become involved in the nonmedical aspects of their child’s care. Principles of family-centered care also include liberal NICU visitation for relatives, siblings, and family friends and the involvement of parents in the development of nursery policies and programs that promote parenting skills.²²⁶

Care must be taken, however, to minimize risk of infection for the neonate. Mothers can transmit infections to neonates both during delivery and post partum, although separation of mother and newborn rarely is indicated. In the absence of certain specific infections, mothers, including those with postpartum fever not attributed to a specific infection, should be allowed to handle their infants if the following conditions are met:

• •They feel well enough.
• •They wash their hands well under supervision.
• •A clean gown is worn.
• •Contact of the neonate with contaminated dressings, linen, clothing, or pads is avoided.²⁴⁵

A mother with a transmissible illness not requiring separation from her infant should be carefully educated about the mode of transmission and precautions necessary to protect her infant. Personal protective equipment, such as cover gowns, gloves, and masks, and hand hygiene facilities should be readily available to her, and she should perform hand hygiene and don a long-sleeved cover gown before handling her infant. If wounds or abscesses are present, drainage should be contained within a dressing. If drainage cannot be completely contained, separation from the infant may be necessary. Care should be taken to prevent the infant from coming in contact with soiled linens, clothing, dressings, or other potentially contaminated items. The mother with active genital HSV lesions need not be separated from her infant if the foregoing precautions are taken. Those with herpes labialis should not kiss or nuzzle their infants until lesions have cleared; lesions should be covered and a surgical mask may be worn until the lesions are crusted and dry, and careful hand hygiene should be stressed.

Mothers with viral respiratory infections should be made aware that many of these illnesses are transmitted by contact with infected secretions as well as by droplet spread, that soiled tissues should be disposed of carefully, and that hand hygiene is critical to transmission prevention. Masks can be worn to reduce the risk of droplet transmission.^{221, 232}

As previously mentioned, although very few infections require separation of mother and infant, women with untreated active pulmonary tuberculosis should be separated from their infants until they no longer are contagious. Mothers with group A streptococcal infections, especially when involving draining wounds, also should be isolated from their infants until they no longer are contagious. Less certain is the necessity of separating mothers with peripartum varicella (onset of infection within 5 days before or 2 days after delivery) from their uninfected infants. The Perinatal Guidelines recommend that such infants remain with their mothers after receiving varicella-zoster immune globulin (VZIG) but caution that infant and mother must be carefully managed in airborne and contact precautions²⁴⁵ to prevent transmission within the nursery. Some experts recommend separating these mothers from their infants until all lesions are dried and crusted.²⁴⁶

Breast-feeding

Numerous studies support the value of human milk for infants (see Chapter 5). Besides providing optimal nutritional content for infants, it has been shown to be associated with a lower incidence of infections and sepsis in the first year of life.^{16, 247} Although contraindications to breast-feeding are few, mothers who have active untreated tuberculosis, human immunodeficiency virus (HIV) infection, breast abscesses (as opposed to simple mastitis that is being treated with antimicrobial therapy), or HSV lesions around the nipples should not breast-feed. Mothers who are hepatitis B surface antigen positive may breast-feed, because ingestion of an infected mother’s milk has not been shown to increase the risk of transmission to her child, but the infant must receive hepatitis B virus immune globulin (HBIG) and vaccine immediately after birth.²⁴⁸ Because systemic disease may develop in preterm infants with low concentrations of transplacentally acquired antibodies to CMV following ingestion of milk of CMV-seropositive mothers, decisions regarding breast-feeding should consider the benefits of human milk versus the risk of CMV transmission. Freezing breast milk has been shown to decrease viral titers but does not eliminate CMV; pasteurization of human milk can inactivate CMV. Either method may be considered in attempts to decrease risk of transmission for breast-feeding NICU neonates.²⁴⁹

Neonates in the NICU frequently are incapable of breast-feeding because of maternal separation, unstable respiratory status, and immaturity of the sucking reflex. For these reasons, mothers of such infants must use a breast pump to collect milk for administration through a feeding tube. Pumping, collection, and storage of breast milk create opportunities for contamination of the milk, and for cross-infection if equipment is shared between mothers. Several studies have demonstrated contamination of breast pumps, contamination of expressed milk that had been frozen and thawed, and higher levels of stool colonization with aerobic bacteria in infants fed precollected breast milk.^{16, 250, 251}

Consensus is lacking on the safe level of microbiologic contamination of breast milk, and most expressed breast milk contains normal skin flora. Although breast milk containing greater than 100CFU/mL of gram-negative bacteria has been reported to cause feeding intolerance and to be associated with suspected sepsis, routine bacterial culturing of expressed breast milk is not recommended.^{249, 250} Instead, efforts to ensure safety of expressed milk should focus on optimal collection, storage, and administration techniques. Cleaning and disinfection of breast pumps should be included in educational material provided to nursing mothers (Table 35–9). In addition, mothers should be instructed to perform hand hygiene and cleanse nipples with cotton and plain water before expressing milk in sterile containers.^{192, 249}

Table 35-9.

Collection and Storage of Expressed Breast Milk

Each mother is supplied with a personal pumping kit.

Nursing staff instruct mothers in techniques of milk expression and appropriate procedures for cleaning breast pump parts:

Wipe all horizontal surfaces at the pumping station with hospital disinfectant before and after pumping.

Wash hands with soapy water before and after pumping.

Wash all parts of the breast pump kit that have been in contact with milk in hot water and dish detergent or in a dishwasher.

Expressed milk is collected in sterile, single-use plastic (polycarbonate or polypropylene) containers.

Breast milk containers are labeled with infant’s name and the date and time of collection.

Administration containers (bottle or syringe) are similarly labeled when breast milk is transferred from collection containers.

All HCWs wear gloves when handling and administering breast milk.

Two persons check the labeled administration container against the infant’s hospital identification band before administering breast milk (may be two HCWs or one HCW and a family member).

HCW, health care worker.

From Infection Control Policy, Children’s Hospital and Regional Medical Center, Seattle, 2003.

Expressed breast milk can be refrigerated for up to 48 hours and can be safely frozen (–20°C ± 2°C [–4°F ± 3.6°F]) for up to 6 months.¹⁹² It can be thawed quickly under warm running water (avoiding contamination with tap water) or gradually in a refrigerator. Exposure to high temperatures, as may be experienced in a microwave, can destroy valuable components of the milk. Thawed breast milk can be stored in the refrigerator for up to 24 hours before it must be discarded. To avoid proliferation of microorganisms, milk administered through a feeding tube by continuous infusion should hang no longer than 4 to 6 hours before replacement of the milk, container, and tubing.²⁴⁵

For mothers who choose not to breast-feed, commercial infant formula is available. Most hospitals now use sterile, ready-to-feed formulas provided by the manufacturer in bottles, with sterile nipples to attach just before use. Nipples are best attached at the bedside just before feeding, and the unit should be used immediately and discarded within 4 hours after the bottle is uncapped.²⁴⁵

Specialty and less commonly used formulas may not be available as a ready-to-feed product, and breast milk supplements do not come in liquid form. After a recent report of a case of fatal Enterobacter sakazakii meningitis in a neonate fed contaminated powdered infant formula,⁹⁷ concerns have risen about the safety of these products. Although powdered formulas are not sterile, preparation and storage practices can decrease the possibility of proliferation of microorganisms after preparation. The CDC, the Food and Drug Administration, and the American Dietetic Association offered updated recommendations on the safe preparation and administration of commercial formula after the recall of the product linked to the E. sakazakii case. These recommendations instruct the care provider as follows:

• •Use alternatives (ready-to-feed or concentrated formulas) to powdered infant formula whenever possible.
• •Prepare formula using aseptic technique in a designated formula preparation room.
• •Refrigerate prepared formula so that a temperature of 2° to 3°C is reached by 4 hours after preparation, and discard any reconstituted formula stored longer than 24 hours.
• •Limit ambient-temperature hang time of continuously infused formula to no longer than 4 hours.
• •Use hygienic handling techniques at feeding time, and avoid open delivery systems.
• •Have written guidelines for managing a manufacturer’s recall of contaminated formula.²⁵²

The FDA also recommended that boiling water be used to prepare powdered formulas, but concerns about this recommendation include potential damage to formula components from the high temperature of the water, a lack of evidence that using this method would kill potential pathogens in the formula, and risk of injury to persons preparing the formula.²⁵²

Co-bedding

The concept of co-bedding, or the bunking of twin infants (or other multiples) in a single isolette or crib, is being explored in NICUs for the potential benefits offered to the babies. Co-bedding as a component of developmentally supportive care is based on the premise that extrauterine adaptation of twin neonates is enhanced by continued physical contact with the other twin.²⁵³ Potential benefits need further study but may include increased bonding, decreased need for temperature support, and easier transition to home. It is certainly possible, however, for one of a set of multiples to be infected while the others are not, and for parents to be implicated as vectors in infection transmission. It also is possible for invasive devices and intravascular catheters to be dislodged by close contact with an active sibling. Therefore, exclusion criteria for co-bedding infants should include clinical findings suggesting infection that could be transmitted to a sibling (e.g., draining wound) and the need for drains and central venous or arterial lines.^{253, 254, 255}

Visitors

The principles of family-centered care encourage liberal visitation policies, both in the well-baby nursery (or rooming-in scenario) and in the NICU. Parents, including fathers, should be allowed unlimited visitation to their newborns, and siblings also should be allowed liberal visitation. Expanding the number of visitors to neonates may, however, increase the risk of disease exposure if education and screening for symptoms of infection are not implemented. Written policies should be in place to guide sibling visits, and parents should be encouraged to share the responsibility of protecting their newborn from contagious illnesses. The Perinatal Guidelines regarding persons who visit newborns are listed in Table 35–10.

Table 35-10.

Guidelines for Sibling Visits to Well-Baby and High-Risk Nurseries

Sibling visits should be encouraged for healthy and ill newborns.

Parents should be interviewed at a site outside the nursery to establish that the siblings are not ill before allowing them to visit.

Children with fever or other symptoms of an acute illness such as upper respiratory infection or gastroenteritis, or those recently exposed to a known communicable disease such as chickenpox, should not be allowed to visit.

Visiting children should visit only their sibling.

Children should be prepared in advance for their visit.

Visitors should be adequately observed and monitored by hospital staff.

Children should carefully wash their hands before patient contact.

Throughout the visit, siblings should be supervised by parents or another responsible adult.

Data from American Academy of Pediatrics and American College of Obstetricians and Gynecologists. Care of the neonate. In Gilstrap LC, Oh W (eds). Guidelines for Perinatal Care, 5th ed. Elk Grove Village, Ill, American Academy of Pediatrics, and Washington DC, American College of Obstetricians and Gynecologists, 2002, pp 331–353.

Adult visitors to neonates, including parents, have been implicated in outbreaks of infections including P. aeruginosa infection, pertussis, and Salmonella infection.^{132, 255, 256} Accordingly, the principles for sibling visitation should be applied to adult visitors as well. They should be screened for symptoms of contagious illness, instructed to perform hand hygiene before entering the NICU and before and after touching the neonate, and should interact only with the family member they came to the hospital to visit. Families of neonates who have lengthy NICU stays may come to know each other well and serve as sources of emotional support to one another. Nevertheless, they should be educated about the potential of transmitting microorganisms and infections between families if standard precautions and physical separation are not maintained, even though they may be sharing an inpatient space.

Skin and Cord Care

Bathing the newborn is standard practice in nurseries, but very little standardization in frequency or cleansing product exists. If not performed carefully, bathing actually can be detrimental to the infant, resulting in hypothermia, increased crying with resulting increases in oxygen consumption, respiratory distress, and instability of vital signs.²⁵⁷ Although the initial bath or cleansing should be delayed until the neonate’s temperature has been stable for several hours, removing blood and drying the skin immediately after delivery may remove potentially infectious microorganisms such as hepatitis B virus, HSV, and HIV, minimizing risk to the neonate from maternal infection.²⁴⁹ When the newborn requires an intramuscular injection in the delivery room, infection sites should be cleansed with alcohol to prevent transmission of organisms that may be present in maternal blood and body fluids.¹⁹⁵ For routine bathing in the first few weeks of life, plain warm water should be used. This is especially important for preterm infants, as well as full-term infants with barrier compromise such as abrasions or dermatitis. If a soap is necessary for heavily soiled areas, a mild pH-neutral product without additives should be used, and duration of soaping should be restricted to less than 5 minutes no more than three times per week.²⁵⁷

Few randomized studies comparing cord care regimens and infection rates have been performed, and consensus has not been reached on best practice regarding care of the umbilical cord stump. A review published in 2003 described care regimens used for more than 2 decades, including combinations of triple dye, chlorhexidine, 70% alcohol, bacitracin, hexachlorophene, povidone-iodine, and “dry care” (soap and water cleansing of soiled periumbilical skin) and found variable impact on colonization of the stump.²⁵⁸ The study authors suggested that dry cord care alone may be insufficient and that chlorhexidine seemed to be a favorable antiseptic choice for cord care because of its activity against gram-positive and gram-negative bacteria. They went on to stress, however, that large, well-designed studies were required before firm conclusions could be drawn. The current Perinatal Guidelines do not recommend a specific regimen but warn that use of alcohol alone is not an effective method of preventing umbilical cord colonization and omphalitis.²⁴⁹ The Perinatal Guidelines further recommend that diapers be folded away from and below the stump and that emollients not be applied to the stump.²⁵⁷

Ocular Prophylaxis

Although blindness resulting from neonatal conjunctivitis is rare in the United States, with a reported incidence of 1.6% or less, the rate among the 80 million infants born annually throughout the world is as high as 23%.²⁵⁹ Chlamydia trachomatis has been the most common etiologic agent in the United States, but other organisms such as Neisseria gonorrhoeae, S. aureus, and E. coli also can cause ophthalmia neonatorum.²⁶⁰ Use of 1% silver nitrate drops, at one time the agent of choice, is no longer recommended because of concerns about associated chemical irritation. Agents thought to be equally efficacious and now recommended include 1% tetracycline and 0.5% erythromycin ophthalmic ointments, administered from sterile single-use tubes or vials.^{156, 245} Povidone-iodine (2.5%) ophthalmic solution also can be used and in one study was shown to be more effective than silver nitrate or erythromycin in the prevention of ophthalmia neonatorum. Bacterial resistance has not been seen with this agent, it causes less toxicity than either silver nitrate or erythromycin, and it is less expensive—a definite consideration in developing countries.²⁵⁹ Whatever the agent selected, it should reach all parts of the conjunctival sac, and the eyes should not be irrigated after application.

Ophthalmic agents will not necessarily prevent ocular or disseminated gonorrhea in infants born to mothers with active infection at time of delivery. These infants should be given parenteral antimicrobial therapy as well as ocular prophylaxis.^{195, 261} Some experts also advise giving infants born to mothers with untreated genital chlamydial infections a course of oral erythromycin beginning on the second or third day of life.²⁶¹

Device-Related Infections

Primary Bloodstream Infections

Primary bloodstream infections (defined by the CDC NNIS System as being due to a pathogen cultured from one or more blood specimens not related to an infection at another site) account for a large proportion of infections in NICU infants,²¹ and most are related to the use of an intravascular catheter.³⁶ Peripheral intravenous catheters (PIVs) are the most frequently used devices for the neonate for intravenous therapy of short duration. When longer access is necessary, nontunneled CVCs such as umbilical catheters and PICCs most commonly are used.¹⁹⁵ The most recent data available from NNIS (August 2003) revealed that the mean umbilical catheter– and CVC-associated bloodstream infection rates for NICUs ranged from 10.6 per 1000 catheter days for infants whose birth weight was less than 1000 g to 3.7 per 1000 catheter days in infants whose birth weight was 2500g or more.¹⁹⁴ The CDC recommends implementing strategies to reduce the incidence of such infections that strike a balance between patient safety and cost-effectiveness.

Few large studies of risks related to intravascular devices have been performed in NICU patients. As a result, intravascular device recommendations for neonates are based on those developed for adults and older pediatric patients (Table 35–11). Several differences in their management should be considered. Although the CDC recommends, in certain circumstances, using antimicrobial- or antiseptic-impregnated CVCs in adults whose catheters are expected to remain in place more than 5 days,³⁶ these catheters are not available in sizes small enough for neonates. Of more importance, studies to evaluate their safety in neonates, especially premature neonates of very low birth weight, have not been performed. In addition, although the CDC recommends changing the insertion site of PIVs at least every 72 to 96 hours in adults, data suggest that leaving PIVs in place in pediatric patients does not increase the risk of complications.²⁶² The 2002 CDC guidelines recommend that PIVs be left in place in children until therapy is completed, unless complications occur.

Table 35-11.

Strategies for Prevention of Catheter-Related Bloodstream Infections in Adult and Pediatric Patients

Conduct surveillance in NICUs to determine catheter-related bloodstream infection rates, monitor trends, and identify infection control lapses.

Investigate events leading to unexpected life-threatening or fatal outcomes.

Select the catheter, insertion technique, and insertion site with the lowest risk for complications for the anticipated type and duration of intravenous therapy.

Use a CVC with the minimal number of ports essential for management of the patient. Designate one port for hyperalimentation if a multilumen catheter is used.

Educate HCWs who insert and maintain catheters, and assess their knowledge and competence periodically.

Use aseptic technique and maximal sterile barriers during insertion of CVCs (cap, mask, sterile gown, sterile gloves, and a large sterile barrier).

Do not routinely replace CVCs, PICCs, or pulmonary artery catheters to prevent catheter-related infections. Do not remove on the basis of fever alone.

In pediatric patients, leave peripheral venous catheters in place until intravenous therapy is completed unless a complication (e.g., phlebitis, infiltration) occurs.

Remove intravascular catheters promptly when no longer essential.

Observe proper hand hygiene procedures either by washing with antiseptic-containing soap and water or use of waterless alcohol-based products before and after working with intravascular lines.

Disinfect skin with an appropriate antiseptic before catheter insertion and during dressing changes. A 2% chlorhexidine-based preparation is preferred.

Do not use topical antibiotic ointment or creams on insertion sites, except when using dialysis catheters.

Use either sterile gauze or sterile, transparent, semipermeable dressing to cover the catheter site. Replace gauze dressings on short-term CVC sites every 2 days and transparent dressings at least weekly, except in pediatric patients, in whom the risk of dislodging the catheter outweighs the benefit of changing the dressing. Change if damp, loosened, or visibly soiled.

Replace dressings on tunneled or implanted CVC sites no more than once per week until the insertion site has healed.

Chlorhexidine sponge dressings are contraindicated in neonates younger than 7 days or those born at a gestational age of less than 26 weeks.

Clean injection ports with 70% alcohol or an iodophor before accessing the system.

Use disposable transducer assemblies with peripheral arterial catheters and pressure monitoring devices. Keep all components of such systems sterile, and do not administer dextrose-containing solutions or parenteral nutrition fluids through them.

CVC, central venous catheter; HCW, health care worker; NICU, neonatal intensive care unit; PICC, peripherally inserted central catheter.

Data from Centers for Disease Control and Prevention. Guidelines for prevention of intravascular catheter-related infections. MMWR Morb Mortal Wkly Rep 51(No. RR-10):32, 2002.

Careful skin antisepsis before insertion of an intravascular catheter is critical to prevention of intravascular device– related bacteremia, although care in the selection of a product for use on neonatal skin is required. Chlorhexidine preparations are recommended by the CDC because these products have been found to be superior to povidone-iodine in reducing the risk for peripheral catheter colonization in neonates. Residues left on the skin by chlorhexidine prolong its half-life, providing improved protection for catheters in neonates that must be left in place for longer periods of time.²⁵⁷

Umbilical veins and arteries are available for CVC insertion only in neonates and are typically used for several days; thereafter, the CVC is replaced with another, nontunneled CVC or PICC if continued central venous access is required. The umbilicus provides a site that can be cannulated easily, allowing for collection of blood specimens and hemodynamic measurements, but after birth, the umbilicus quickly becomes heavily colonized with skin flora and other microorganisms. Colonization and catheter-related bloodstream infection rates for umbilical vein and umbilical artery catheters are similar. Colonization rates for umbilical artery catheters are estimated to be 40% to 55%; the estimated rate for umbilical artery catheter–related bloodstream infection is 5%.³⁶ Colonization rates are from 22% to 59% for umbilical vein catheters; rates for umbilical vein catheter–related bloodstream infections are 3% to 8%.³⁶ A summary of the CDC recommendations for management of umbilical catheters³⁶ is presented in Table 35–12.

Table 35-12.

Summary of CDC Recommendations for Management of Umbilical Catheters

Cleanse umbilical insertion site with an antiseptic before catheter insertion. Avoid tincture of iodine; povidone-iodine can be used.

Add low doses of heparin to fluid infused through umbilical artery catheters.

Remove and do not replace umbilical catheters if signs of catheter-related bloodstream infection, vascular insufficiency, or thrombosis are present.

Remove umbilical catheters as soon as possible when no longer needed or if any sign of vascular insufficiency to the lower extremities is observed.

Umbilical artery catheters should not be left in place for longer than 5 days.

Umbilical venous catheters should be removed as soon as possible when no longer needed but can be used for up to 14 days if managed aseptically.

CDC, Centers for Disease Control and Prevention.

Data from Centers for Disease Control and Prevention. Guidelines for prevention of intravascular catheter-related infections. MMWR Morb Mortal Wkly Rep 51(No RR-10):32, 2002.

Ventilator-Associated Pneumonia

As mentioned earlier, NNIS data indicate that nosocomial pneumonia is the second most common infection type in NICU patients. Risk factors for ventilator-associated pneumonia can be grouped as host-related (prematurity, low birth weight, sedation or use of paralytic agents), device-related (endotracheal intubation, mechanical ventilation, orogastric or nasogastric tube placement) and factors that increase bacterial colonization of the stomach or nasopharynx (broad-spectrum antimicrobial agents, antacids, or H₂ blockers).^{51, 264, 265} Ventilator-associated pneumonia generally refers to bacterial pneumonia that develops in patients who are receiving mechanical ventilation. Aspiration and direct inoculation of bacteria are the primary routes of entry into the lower respiratory tract; the source of these organisms may be the patient’s endogenous flora or transmission from other patients, staff members, or the environment.^{266, 267} Few studies have been performed to assess the effectiveness of prevention strategies in pediatric patients. Strategies to prevent ventilator-associated pneumonia in the NICU patient are therefore based primarily on studies performed in adults (Table 35–13). Hand hygiene remains critical to the prevention of ventilator-associated pneumonia, and HCWs should consistently apply the principles of standard precautions to the care of the ventilated patient, wearing gloves to handle respiratory secretions or objects contaminated by them, and changing gloves and performing hand hygiene between contacts with a contaminated body site and the respiratory tract or a respiratory tract device.

Table 35-13.

Effective Strategies for Prevention of Ventilator-Associated Pneumonia

Removal of nasogastric or endotracheal tube as soon as clinically feasible

Adequate hand hygiene between patients

Semirecumbent positioning of the patient

Avoidance of unnecessary reintubation

Provision of adequate nutritional support

Avoidance of gastric overdistention

Scheduled drainage of condensate from ventilator circuits

Data from Kollef MH. Current concepts: the prevention of ventilator-associated pneumonia. N Engl J Med 340:627–634, 1999.

Because mechanical ventilation is a significant risk factor for the development of nosocomial infection or ventilator-associated pneumonia, weaning from ventilation and removing endotracheal tubes as soon as indication for their use ceases are key infection control strategies. As an alternative to endotracheal intubation, noninvasive nasal continuous positive airway pressure (CPAP) ventilation avoids some of the common risk factors for ventilator-associated pneumonia and has been used successfully for neonates.^{268, 269} Respiratory care equipment that comes in contact with mucous membranes of ventilated patients or that is part of the ventilator circuit should be single use (discarded after one-time use with a single patient) or be subjected to sterilization or high-level disinfection between patients. Wet heat pasteurization (processing at 76°C for 30 minutes) or chemical disinfectants can be used to achieve high-level disinfection of reusable respiratory equipment.²⁶³ Ventilator circuits should be changed no more frequently than every 48 hours, and evidence suggests that extending the length of time between changes to 7 days does not increase the risk of ventilator-associated pneumonia.^{270, 271} Circuits should be monitored for accumulation of condensate and drained periodically, with care taken to avoid allowing the condensate, a potential reservoir for pathogens, to drain toward the patient.^{263, 267} Sterile fluids should be used for nebulization, and sterile water should be used to rinse reusable semicritical equipment and devices such as in-line medication nebulizers.²⁶³

Basic infection control measures, such as hand hygiene and wearing gloves during suctioning and respiratory manipulation, also can reduce the risk of nosocomial pneumonia. Both open, single-use and closed, multiuse suction systems are available. If an open system is used, a sterile single-use catheter should be used each time the patient is suctioned. Closed systems, which do not need to be changed daily and can be used for up to 7 days,²⁷² have the advantage of lower costs and decreased environmental cross-contamination²⁵⁸ but have not been shown to decrease the incidence of nosocomial pneumonia when compared with open systems.^{273, 274}

Although not well studied in pediatric patients, aspiration of oropharyngeal secretions is believed to contribute to development of ventilator-associated pneumonia in adults.²⁷⁵ Placing the mechanically ventilated patient in a semi-recumbent position or elevating the head of the bed in an attempt to minimize aspiration is recommended unless medically contraindicated. Also, placement of enteral feeding tubes should be verified before their use.^{263, 267} To prevent regurgitation and potential aspiration of stomach contents by the sedated patient, overdistention of the stomach should be avoided by regular monitoring of the patient’s intestinal motility, serial measurement of residual gastric volume or abdominal girth, reducing the use of narcotics and anticholinergic agents, and adjusting the rate and volume of enteral feedings.^{263, 267} Oral decontamination, with the intent of decreasing oropharyngeal colonization, has been studied in adults and seems to lower the incidence of ventilator-associated pneumonia (although not duration of ventilation or mortality rate),^{275, 276} but further work is needed to determine whether this is an effective strategy in neonates. In addition, medications such as sucralfate, as opposed to histamine H₂ receptor antagonists and antacids, which raise gastric pH and can potentially result in increased bacterial colonization of the stomach, have been used to prevent development of stress ulcers and have been associated with lower incidence of ventilator-associated pneumonia in adults.²⁷⁷ Two studies suggest, however, that this approach is of no benefit in pediatric patients, but the authors stress that additional studies with larger sample sizes are needed to confirm these findings.^{278, 279}