Jane D. Siegel
The newborn nursery includes healthy, full-term infants who weigh ≥2,000 grams (g) at birth in the normal newborn areas as well as high-risk infants who weigh <2,000 g at birth and are term infants but have complex medical problems in the neonatal intensive care unit (NICU) or high-risk nursery (HRN) area. These infants have an increased risk of acquiring infection because all components of their host defense system are deficient compared with those of older infants or adults, and the severity of these deficiencies is increased as gestational age decreases [1]. Survival of prematurely born infants has improved in recent years as a result of more advanced high-risk obstetrical care and neonatal supportive care including the use of surfactant replacement for treatment of hyaline membrane disease, mechanical ventilation including conventional and high-frequency ventilation, extracorporeal membrane oxygenation (ECMO) and continuous hemofiltration to support cardiopulmonary and renal function, noninvasive ventilation (e.g., continuous positive airway pressure [CPAP]) and cardiac intervention techniques, improved surgical techniques, and screening and chemoprophylaxis for early onset group B streptococcal (GBS) diseases. Therefore, increasing numbers of infants of very low birth weight (VLBW) (500–1,000 g, ≥24 weeks gestation) are surviving but require prolonged length of NICU stay and are at increased risk for infection. The National Nosocomial Infections Surveillance (NNIS) system of the Center for Disease Control and Prevention (CDC) reported that the group of VLBW infants (≤1,000 g) has the highest rates of central line including umbilical catheter-associated bloodstream infections (CLA-BSIs) (mean 9.1; range 1.6–16.1) CLA-BSIs per 1,000 central venous catheter (CVC)-days compared with all other intensive care units (ICUs) reporting in the NNIS system [2,3,4,5] (Table 25-1). The rates of infection caused by bacteria and Candida sp. vary considerably among NICUs in the United States [6,7] and Canada [8] even after adjustment for known risk factors that could be practice related. Identification of practices in the best performing units and of widespread implementation has been a successful strategy of the Vermont Oxford Collaborative to reduce infection and other complication rates associated with NICU care (www.vtoxford.org/home.aspx) [9,10,11]. Device-associated infection rates have decreased over time in NNIS HRNs as in other NNIS ICUs, most likely a reflection of the beneficial effect of consistent implementation of recommended practices, measurement of rates, and feedback to the primary caregivers.
In the absence of in utero infection, the neonate is first exposed to microorganisms during passage through the birth canal. Subsequently, the normal skin and mucous membrane microflora are derived from environmental sources. Healthy term infants usually have a short hospital stay of <72 hours and in the managed care era often barely 24 hours. Therefore, infections are acquired infrequently in the normal newborn nursery (<1% of all admissions) and could not become manifest until after discharge, making
P.418
surveillance particularly challenging. In contrast, VLBW infants could remain in the NICU for several weeks to months with continued exposure to many devices and invasive procedures, antibiotic-resistant hospital flora, and antimicrobial agents that further influence the composition of the microflora. Thus, meaningful analysis of healthcare-associated infection (HAI) rates in the newborn nursery must use consistent definitions and risk stratification to account for the heterogeneity of its population. The HRN is the focus of HAI surveillance and prevention because of the associated increase in morbidity and mortality.
|
TABLE 25-1 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Definitions
Time of Onset
Although the term nosocomial infection has been replaced by healthcare-associated infection in most populations studied by healthcare epidemiologists, nosocomial is an appropriate term for most NICU infections because neonates are generally not readmitted to nurseries after discharge. Occasionally, infants are transferred to other institutions for specialized surgical procedures and then return to the referring nursery for prolonged care, and HAI would be a more appropriate term for surgical site infections (SSIs) in such instances.
Various investigators define early onset disease as positive cultures of a normally sterile body fluid obtained within the first 3, 7, or 10 days of life. For the study of HAIs, the most appropriate time interval is 3 days. Infections that appear at <48 hours of age are considered to be maternally acquired. Approximately 15% of BSIs and pneumonias in the HRN are maternally acquired [4]. Outbreaks of early onset infections rarely are reported and have remained either unexplained [12] or have been associated with fetal scalp electrode placement during labor [13], contaminated resuscitation equipment in the delivery room [14,15], and contaminated materials used within the first few hours of life (e.g., hydrocolloid dressings manufactured in large sheets precut by healthcare workers [HCWs] and used to secure umbilical catheters or endotracheal tubes contaminated with A. baumanni) [16].
For the purposes of HAI tracking, positive cultures obtained >3 days of life are considered late onset disease. Because clinical manifestations of infection are often delayed and determining whether an infection was acquired from the mother or from transmission within the nursery is difficult, NNIS reports all infections except those that are transmitted transplacentally as HAIs [4]. Differentiation of early and late onset disease is most useful when designing prophylaxis regimens. It is recommended that bacterial infections other than urinary tract infection (UTI) occurring within the first month after discharge from the
P.419
nursery be reported to the infant's nursery to facilitate prompt identification of an outbreak (e.g., skin infections associated with Staphylococcus aureus and streptococcus group A or B, omphalitis, and bacterial diarrhea, especially Salmonella species) [14,17,18,19].
Device-Related Infections
Standardized definitions developed and updated by CDC, are used to track rates of NICU infections associated with CVCs, including umbilical catheters and ventilator-associated pneumonia (VAP) [20]. Although the CDC definition of CLA-BSI includes culture negative, clinical sepsis in young infants, such infections are infrequently reported in NNIS, and it is likely that only laboratory-confirmed episodes will be reported to the National Healthcare Safety Network (NHSN), formerly NNIS. (www.cdc.gov/ncidod/hip/NNIS/members/pneumonia/Final/PneumoCriteriaV1.pdf)
Calculation of Device-Related Infection Rates in the High-Risk Nursery
Studies [3,21,22] have demonstrated the advantages of calculating device-associated HAI rates to control for duration of exposure to the primary risk factors. Device utilization (DU) ratios are useful for interhospital comparisons as long as each hospital has collected the data and the calculated ratios use the same definitions and methods. The DU ratio is the measure of an ICU's invasive practices that constitutes an extrinsic risk factor for HAI. The DU ratio also can serve as a marker for severity of illness or the patients' intrinsic susceptibility to infection. If the DU ratio is > the 90th percentile, a specific hospital is considered a high outlier and further investigation of that specific practice could be warranted. Device-associated infection rates and DU ratios are calculated as follows:
Central line-associated primary bloodstream infection (CLA-BSI) rate:
Ventilator-associated pneumonia (VAP) rate:
Device utilization (DU) ratio:
Severity of Illness Scoring Systems
Measures of illness severity other than birth weight have been applied to the study of neonatal HAI risk since 1993 when several scoring systems were first described [23,24,25,26,27]. The Score for Neonatal Acute Physiology (SNAP) uses the worst recorded values of >24 routinely measured physiologic variables during the first 24 hours of stay. The SNAP-Perinatal Extension (SNAP-PE) adds scoring for birth weight, small for gestational age status, and low Apgar score (<7 at 5 minutes). Using multiple regression analysis, a study of coagulase-negative staphylococcal (CONS) BSIs demonstrated a 53.9% increase in a patient's risk of experiencing at least one nosocomial BSI episode associated with each 5-point increment in the admission day SNAP [24]. Further analysis of the association between admission-day therapies and BSI attack rates among these patients using the Neonatal Therapeutic Intervention Scoring System (NTISS) that bases severity of illness on 62 specific therapeutic interventions demonstrated a significant positive association of phototherapy and blood pressure support with colloid or vasopressor administration. One study that used SNAP and NTISS reported abnormal heart rate characteristics and worsening SNAP scores 24 hours before the clinical suspicion of sepsis [25]. A study of eradication of endemic methicillin-resistant S. aureus (MRSA) infections from the Parkland Health and Hospital System neonatal intensive care unit (NICU) [26] used the number of patient-care hours determined for each infant by the nursing staff according to a workload quantification method called the (GRASP) system [27] to calculate a time-and-intensity-of-care–adjusted HAI incidence density (see Chapters 6 and 8). Use of the sum of all infants' daily patient-care hours as the denominator effectively controlled for the difference in risk factors between the intensive care and intermediate care areas. Further validation is required before these scoring systems can be applied routinely.
A European scoring system, Clinical Risk Index for Babies (CRIB) also has been used but conflicting results concerning its accuracy in predicting outcome have been reported [28]. A model dependent on variables collected before admission to the NICU proposed by the National Institute of Child Health and Human Development (NICHD) neonatal research network has published performance worse than the CRIB and SNAP models. At this time, whether any of these scoring systems offer an advantage over birth weight, Apgar scores, and length of stay for risk stratification and predicting outcome is uncertain [5,23].
Distinguishing True Pathogens From Blood Culture Contaminants
Sepsis caused by pathogens that are common skin contaminants (e.g., CONS) can be associated with low colony counts [29,30,31] and relatively few symptoms. Analysis of sites from which isolates are recovered (e.g., peripheral, specific ports of CVCs and time to positivity) assists the clinician in distinguishing blood culture contaminants from true pathogens. Time to positivity is an indicator of the quantity of bacteria and is an easily implemented replacement for quantitative blood cultures. When there is a >2 hour difference in time to positivity for various sites, the site that turned positive first is more likely to be the source of the bacteremia or fungemia [32,33,34]. The
P.420
following suggestions can optimize the clinician's accuracy in distinguishing true sepsis from contamination (see Chapter 9):
Risk Factors for HAI
The intrinsic and extrinsic factors for HAI have been reviewed [37] and are summarized in Table 25-2.
|
TABLE 25-2 |
||||||||||||||||||||||||||||||
|
Intrinsic Risk Factors
Birth weight and gestational age are the most important risk factors for the HAI development. The decreased function of the immune system in the most premature infants accounts for most of the increased intrinsic risk of infection [1]. There is minimal active transport of maternal IgG antibodies across the placenta until 32 weeks' gestation, neutrophils have defective chemotaxis or phagocytosis, and the classic and alternative complement pathways have decreased activity. Attempts to improve the neonate's immune function have included exchange transfusion [38], white blood cell transfusion [39], administration of intravenous immune globulin (IVIG) therapeutically [40] or prophylactically [41,42], and administration of recombinant human granulocyte colony-stimulating factor (G-CSF) [43]. Although these studies have been instructive, they have not demonstrated efficacy in adequately controlled trials, and no recommendations for routine use of these products have been made. In addition to providing enhanced opsonophagocytic activity, IVIG infusions are associated with a prompt release of neutrophils from the marrow neutrophil storage pool into the peripheral circulation and enhanced chemotaxis of neutrophils to the site of bacterial infection [40]. Antibody replacement could be more effective by using products containing high titers of antibody against the specific infecting agent, but such products are not yet available for routine use [44]. However, a significant protective effect has not been demonstrated in trials of two different preparations of intravenous S. aureus immune globulin [45,46].
Neonates are particularly vulnerable to colonization with virulent and/or antibiotic-resistant bacteria because their mucosal surfaces do not have the usual protective microflora of older infants and adults [47]. Immature, fragile skin of the VLBW neonate possibly does not serve as an adequate protective barrier against pathogens that colonize the skin and have the capacity to cause invasive disease [48].
Extrinsic Risk Factors
Many of the extrinsic HAI risk factors in the nursery are device- or environmentally related, as is observed in adult
P.421
ICUs. Extrinsic risk factors can be categorized as (1) medical devices and equipment, (2) medical treatments, (3) behavioral interventions, and (4) administrative and structural.
Medical Devices and Equipment
CVCs including peripherally inserted central catheters (PICC) are associated with increased risk of BSI caused by bacteria or fungi [4,5,49,50,51,52]. In fact, an increased risk of candidemia has been associated with each day of catheter use [52]. Various medical devices used for respiratory support (e.g., ventilators and noninvasive ventilatory devices including CPAP, ECMO, and Vapotherm 2000i™ [Vapotherm Inc, Stevensville, MD] oxygen delivery device) are associated with increased risk of infection [53,54]. The need by the neonate's with respiratory distress for increased humidification provides direct exposure of the respiratory mucosa to water that could be contaminated; therefore, the use of sterile water in such situations is recommended [55]. For example, contamination of the Vapotherm™ device with Ralstonia sp. during the manufacturing process that could not be eradicated with various disinfection procedures led to clinical infections and the withdrawal of this product from the market [56].
The use of umbilical catheters and the feeding of breast milk are unique to the nursery, and specific guidelines are required to care for the umbilicus and store breast milk for those infants who are too sick to suckle directly from the mother's breast. For example, breast milk from a mother who was infected with Group B Streptococcus (GBS) or S. aureus (MSSA or [MRSA]) has been implicated as the vehicle of transmission of these pathogens to infants and resulting severe sepsis [57,58,59]. Neonatal sepsis caused by Klebsiella pneumoniae has been associated with contaminated breast milk resulting from contamination of a component of a breast pump [60], and contamination of a milk bank pasteurizer was associated with an NICU outbreak of Pseudomonas aeruginosa infections [61]. Furthermore, viral agents that could be transmitted to infants in breast milk or in the blood contact associated with breast feeding from dry, cracked nipples include hepatitis B virus (HBV), human immuno-deficiency virus (HIV), and human T-lymphotrophic virus type 1 (HTLV-1). Therefore, in countries where safe formula for bottle feeding is readily available, breast feeding is contraindicated for mothers known to be infected with HIV and HTLV-1. HBV vaccine given to neonates is protective against transmission by breast feeding. Although cytomegalovirus (CMV) is transmitted in breast milk, maternal antibody is protective against clinically significant disease. The CDC's strict guidelines developed for banking human milk obtained from unrelated donors include screening all donors for HIV, HTLV-1, and HBV surface antigen and pasteurization (62.5°C for 30 minutes) of all milk specimens. Bacterial counts of 104 colony-forming units (cfu)/ml or more of nonpathogenic organisms and the presence of gram-negative bacteria (GNB), S. aureus, or α- or β-hemolytic streptococci preclude the use of milk the specimens come from [62]. Established guidelines for human breast milk banks (www.hmbana.org), personal hygiene, and handling and decontaminating the components of breast milk pumps [62] should be followed.
Medical Treatments
The use of steroids for treatment of chronic bronchopulmonary dysplasia has been associated with an increased risk of infection [63,64,65]. The finding of an increase in the incidence of disseminated Candida spp. infections associated with the unique practice of single-dose steroid administration in infants with prolonged hypotension shortly after birth further supports the role of steroids as an independent risk factor [64]. The risk–benefit ratio must be carefully considered before initiating a course of steroids in HRN infants. The use of H2 blocker/proton pump inhibitor therapy has been associated with an increased incidence of necrotizing enterocolitis [66], with sepsis associated with GNB in VLBW infants [50], and with candidemia [67]. An increased risk of invasive disease caused by extended-spectrum beta-lactamase (ESBL)-producing Klebsiella spp. [51] or Candida spp. [68] has been associated with the use of third-generation cepahlosporins. Finally, topical petrolatum applied to the skin for improved moisturizing effect was associated with an increased risk of fungal infections [47].
Behavioral Interventions
There are theoretical concerns that infection risk can increase in association with the innovative practices of cobedding [69] and kangaroo care [70] used in the NICU to improve developmental outcomes as a result of increased opportunity for skin-to-skin exposure of multiple gestation infants to each other and to their mothers, respectively. However, the infection risk is actually reduced with kangaroo care. Although toys in NICU beds have been found to be contaminated with pathogenic bacteria [71], their role in transmission of infection has not been established.
Administrative and Structural Issues
Studies of outbreaks in the newborn nursery and NICUs were some of the first to demonstrate the relationship between rates of late onset infection and overcrowding and understaffing [26,72,73,74,75,76]. One study that evaluated staffing levels of registered nurses specifically found a significant reduction in the BSI risk when registered nurse staffing hours were increased in one NICU [76]. A movement toward constructing new NICUs with all single-patient rooms is emerging; its purpose is to improve the neurodevelopmental effects of lighting and sound, facilitating family-centered care, breast feeding, kangaroo care, and
P.422
isolation precautions when needed, and thus reduce HAI rates [77,78]. Although data are not yet sufficient to support an evidence-based recommendation for single-patient rooms in NICUs, the most recent publication of recommended standards for NICU design and of experience suggests that this design will become more prevalent in the future. Finally, exposure to construction dust or to spores during transport could result in cutaneous or invasive aspergillosis in the neonate [79]; thus, dust containment and air filtration during construction, renovation, and any disruption of the integrity of the environment [80] are especially important for the NICU because the VLBW infant is at increased risk of developing disease after exposure.
Sites of Infection
The sites of HAIs in the HRN differ from those in adults [5]. Primary BSIs account for 30–50% of episodes in neonates, depending on birth weight, and SSIs and UTIs are rare. In contrast, >40% of HAIs in adults are UTIs, and 20% are SSIs (see UTI and SSI and ICU Chapters 24A, 24B, 30 and 35). Cutaneous sites are more likely to be involved in neonates. Clinical manifestations of MSSA and MRSA in otherwise healthy term neonates [81] or in those in the NICU [82] in the current era have been reviewed. Pustules, bullous impetigo, subcutaneous abscesses, scalded skin syndrome, and toxic shock syndrome associated with either MSSA or MRSA can be seen in outbreaks in term nurseries (often presenting after discharge) and NICUs. Omphalitis is rare (0.7%) in developed countries where deliveries are performed aseptically and cord care to prevent infection is performed routinely. However, occurence of omphalitis can involve serious complications including sepsis, superficial or deep abscesses, necrotizing fasciitis, peritonitis, and hepatic vein thrombosis [83]. Congenital mucocutaneous candidiasis in term infants generally is not associated with invasive disease whereas fungal dermatitis caused by Candida albicans is considered a manifestation of systemic disease when it occurs during the second week of life in extremely low birth weight (ELBW) infants after vaginal delivery, postnatal administration of steroids, or hyperglycemia [84]. Pulmonary infections result from exposure to respiratory viruses circulating in the community, and complications of respiratory support, gastroenteritis, and colitis can also occur following exposure to viruses and/or bacteria circulating in the community. Osteomyelitis/septic arthritis and conjunctivitis are other less common manifestations. Meningitis and brain abscess can also occur in neonates.
Necrotizing enterocolitis (NEC) is one of the most common gastrointestinal emergencies among neonates. In fact, >90% of episodes occur in infants who are born preterm, and the NEC risk is inversely related to birth weight and gestational age. Overall, there is about a 7% occurrence among VLBW infants with substantial variation over time and from center to center [85]. Factors that contribute to the development of NEC include developmental immaturity of the gastrointestinal tract function including circulatory regulation, hypoxic/ischemic injury, abnormal bacterial colonization, and early feeding of formula milk. The role of specific inflammatory cytokines in the pathogenesis of NEC is under investigation. Sporadic episodes and clusters can occur. Several different bacteria (e.g., E. coli, Klebsiella pneumoniae, Enterobacter cloacae, Clostridium sp.) and viruses (e.g., rotavirus, coronavirus, enterovirus) have been associated with NEC in some reports of clusters [86]. Outbreaks can be controlled by implementing infection control measures, including hand hygiene, contact precautions, cohorting of infants and staff, and restriction of HCWs with signs of gastrointestinal tract illness from duty until resolved [86].
Clinical manifestations include feeding intolerance; delayed gastric emptying; localized signs of abdominal distension, tenderness, cellulitis, peritonitis, occult, and grossly bloody stools; radiographic evidence of pneumatosis intestinalis or portal vein gas; and nonspecific systemic signs of sepsis with metabolic acidosis, thrombocytopenia, or disseminated intravascular coagulation (DIC). Mortality rates vary from 15–30%. When intestinal necrosis occurs, resecting is necessary, leaving the infant with short gut syndrome and totally parenteral nutrition dependent.
Etiology, Clinical Manifestations, and Epidemiology
Bacterial Outbreaks
Trends
Infections acquired in normal newborn nurseries are most frequently not invasive, usually involve the skin or mucous membranes, and result from HCW hand carriage, contaminated equipment, and medications. Impetigo (skin pustules), conjunctivitis [87], omphalitis, and soft tissue abscess are the most frequently observed clinical manifestations. S. aureusremains the most frequently isolated pathogen from such infants with MRSA causing infection more often than MSSA in communities with high prevalence of community-associated MRSA (CO-MRSA) infections [81]. Outbreaks of group A streptococcal [18,88] and of diarrhea caused by bacterial pathogens (e.g., Salmonella spp. or Shigella spp.) could occur in both term and preterm nurseries [17,89,90] but have been reported less frequently in recent years (see Chapters 34 and 41). CONS rarely cause early onset disease in otherwise healthy term neonates without any devices in place.
HCWs are rarely the source of outbreaks HAIs caused by bacteria and fungi, especially MRSA, but when they are, factors are usually present to increase the transmission of infectious agents to others (e.g., sinusitis, draining otitis externa, chronic otitis; respiratory tract infections, dermatitis, onychomycosis, and artificial nails) [91,92,93,94,95,96,97]. An HCW colonized with an epidemic strain of S. aureus
P.423
rarely has been identified; when it has, the outbreak has been controlled by removing that individual from direct patient care [98]. Individual wearing artificial nail who have direct patient contact have been implicated in outbreaks of Pseudomonas aeruginosa [95,96] and ESBL-producing Klebsiella pneumoniae [97] in NICUs where molecular typing demonstrated that HCW and patient isolates were indistinguishable. These studies contributed substantially to the recommendation not to wear artificial nails or extenders when having direct contact with high-risk patients.
Unexplained shifts in the predominant etiology of bacterial infections in high-risk infants have been observed over time [99,100]. Invasive strains of S. aureus were predominant in the 1950s. For unexplained reasons, GNB, especially Pseudomonas aeruginosa, Klebsiella species, and Escherichia coli strains prevailed in the 1960s but were replaced by GBS in the 1970s. GBS remained the major pathogens of early onset disease throughout the 1980s and 1990s but manifested as late onset disease, most frequently meningitis, less commonly as osteomyelitis/septic arthritis, and rarely event [101] as horizontal transmission within the NICU. However, in the 1980s, MRSA and CONS emerged as prevailing HAI pathogens in the HRN. Although CONS continue to account for 40–50% of late onset infections in most NICUs in the 1990s, one NICU reported a shift to predominant GNB, especially ceftazidime-resistant Enterobacter sp., as the pathogen from 1996–2001 [102], and another reported an increase the number of commensals [100], similar to the experience of the National Institute of Child Health and Human Development Neonatal Research Network [6]. The following three pathogens associated with <15–20% of HRN infections are particularly problematic because of the difficulty in treatment: (1) enterococci, especially vancomycin-resistant enterococcus (VRE), (2) multidrug-resistant GNB, especially Enterobacter spp. and ESBL-producing Klebsiella spp, and (3) fungi, predominantly Candida spp., especially nonalbicans species.
Group B Streptococcus
From the late 1970s through the mid-1990s, GBS was the most frequently isolated pathogen from term infants with early onset disease, accounting for ~70% of episodes [6]. This pathogen was acquired from the mother in the peripartum period; in ~70% of infants <2,000 g at birth, GBS infections were acquired in utero with positive blood cultures obtained at birth. However, following the development and implementation of the evidence-based recommendations published collaboratively by CDC, American College of Obstetrics and Gynecology (ACOG), and the American Academy of Pediatrics (AAP) for the administration of chemoprophylaxis at the onset of labor to colonized women, those delivering prematurely, and those who have other risk factors that were published in 1996, there was a 65% reduction in the incidence of early onset GBS disease from 1993 to 1998 and a plateau in 1999–2001 [103]. The guidelines were updated to include a recommendation to screen all women at 35–37 weeks gestation based on a population-based study that demonstrated a greater reduction in disease associated with a screening-based strategy compared with a risk-based strategy [104]. This resulted in a further reduction in 2003–2004 to a rate of 0.34 per 1,000 live births reported by the CDC's active bacterial core (ABCS) surveillance network and a narrowing of the racial disparity in the rates of disease [105]. These data represent an 80% reduction from a rate in 1,000 live births of 1.7 in 1993 to 0.34 in 2004. A persistent reduction in early onset GBS disease has been reported also in the VLBW infants in the National Institute of Child Health and Human Development Neonatal Research Network [106] and by other groups who have been tracking rates of disease. Following publication of the GBS chemoprophylaxis guidelines, increased the rates of ampicillin-resistant E. coli as a cause of early onset sepsis in VLBW infants have, but no association between intrapartum antibiotic exposure and overall or ampicillin-resistant E. coli sepsis has been found [106,107]. Continued population-based surveillance is needed.
Coagulase Negative Staphylococci
CONS account for nearly 50% of late onset HAIs in most recent reports [4,5,6,41,42,108]. Several reasons for the increased recognition of this organism as a neonatal pathogen follow:
The epidemiology of neonatal CONS BSI has been studied extensively using pulsed-field gel electrophoresis (PFGE), ribotyping, DNA-DNA hybridization, and restriction endonuclease analysis in addition to the traditional methods of speciation, phage typing, and plasmid analysis. PFGE is the most reliable method for confirming identity of strains. Distinct clones of both Staphylococcus haemolyticus [109,110] and Staphylococcus epidermidis [111,112,113] can become endemic in HRNs and cause clusters of infections over periods of 6 months to as long as 10 years. At the same time, many completely unrelated strains can be isolated from infants within the same unit. Some HRNs possibly have no related strains identified during a specific period of time [114,115]. Eastick et al. [116] have reported relatively stable reservoirs of CONS in the feces, around the ear, and in the axillae and nares but small, unstable
P.424
numbers on the skin of the forearm and leg. Thus, cross-contamination among sites on the same infant as well as horizontal spread among infants is an important mode of transmission. Infusion of parenteral fluids contaminated with CONS is a rare source of BSI in the HRN [117] (see Chapter 44).
The clinical manifestations of CONS-BSI are most often nonspecific, and this pathogen is rarely considered a cause of death [118]. The nonspecific signs of sepsis observed most frequently are fever, apnea and bradycardia, feeding intolerance, and lethargy. Temperature instability, thrombocytopenia, abdominal distention, and persistent BSI in the absence of a CVC have been associated with disease caused by CONS [119,120]. Specific delta toxin-producing strains have been found in pure culture in the stool or in the blood or peritoneal fluid of patients with a mild form of necrotizing enterocolitis [121,122]. Focal infections associated with these pathogens include neck abscess, omphalitis, wound abscess, and mastitis [123]. Right-sided endocarditis caused by CONS must be considered in the presence of a CVC, persistent BSI (>48 hours), and thrombocytopenia during appropriate antimicrobial therapy for BSI [124]. Physical examination, could find no other abnormalities but an echocardiogram could demonstrate vegetations in the right atrium or on the tricuspid valve. Removal of the CVC is required for clearance of CONS if BSI persists >4 days [125].
Staphylococcus Aureus
After CONS, S. aureus was the second most frequently isolated pathogen from HRN infants in the 1986–1993 NNIS report [4] and in studies from the Neonatal Research Network [6]. In contrast, the point prevalence study of 29 NICUs performed in 1999 reported that enterococci were second after CONS, accounting for 15.5% of BSIs, and S. aureus accounted for only 3.4% of BSIs [5]. Since the late 1980s, S. aureus outbreaks in the HRN have been associated with both MSSA [73,98,126,127,128,129] and MRSA [26,130,131,132,133,134], and reports of outbreaks of MRSA have surpassed those of MSSA in recent years. Nurseries situated in large general hospitals that share services (laboratory, radiology) and HCWs (nurses, respiratory therapists, consulting physicians) are especially vulnerable to the acquisition of virulent strains from geographically distant foci of healthcare-associated MSSA or MRSA infection. A single or multiple virulent strains can be introduced into the nursery by a colonized infant, a visiting family member, and—rarely—an HCW [135,136,137]. However, the principal mode of transmission is horizontal via hands of HCWs who fail to follow recommendations for hand hygiene between patient contacts. PFGE is the most useful method to determine whether a single or multiple strains caused an outbreak [138]. High rates of colonization (30–70%) can become established before clinical disease is recognized. Environmental sources or chronic carriers have rarely been implicated in nosocomial transmission in the NICU, but environmental contamination with S. aureus does occur and contributes to the overall burden of pathogens in high-risk units. Unidentified virulence factors and environmental conditions determine the persistence of the organism in the HRN. In Parkland Health and Hospital System's crowded HRN, MRSA persisted for a 3-year period from 1988 to 1991 [26]. Neither susceptible S. aureus and MRSA nursery outbreaks were controlled until conditions of overcrowding and understaffing were corrected [26,73]. Although MRSA is an important pathogen in NICUs, it is notable that in 2002, only 25% of S. aureus isolates from NICU patients in the NNIS system were resistant to methicillin compared with nearly 60% of S. aureus isolates from the NNIS ICU isolates overall [Fridkin, S, personal communication]. The prevalence of CO-MRSA strains that have antibiotic susceptability and molecular profiles (staphylococcal chromosomal cassette [SCC] mec types IV or V) distinct from those of the traditional healthcare-associated MRSA strains [138] have increased in prevalence and transmission of those CO-MRSA strains have been reported within healthcare facilities [139,140,141]. Because CO-MRSA infections have been described in obstetrical patients [142] and 13/14 MRSA strains isolated from vaginal-rectal prenatal GBS screening cultures were CO-MRSA by SCCmec-typing [143], a likely source for neonatal infections is the colonized or infected mother. Trends in CO-MRSA strains as the etiology of HAIs are evolving.
Skin pustules, bullous impetigo, scalded skin syndrome [126,127,128], soft tissue abscesses, mastitis, conjunctivitis, pneumonia with or without empyema, osteomyelitis, septic arthritis, and SSIs are the most frequent manifestations of both MSSA and MRSA infections in neonates, and the case-fatality rate associated with invasive disease can be 15–30% [81,82]. BSIs with multiple foci of disease are characteristic of invasive S. aureus disease. Optimal diagnostic evaluation and treatment of neonates with only minor manifestations (e.g., pustules) of CO-MRSA are under investigation at the time of this publication.
Enterococcus
Enterococci have been recognized as important pathogens in the NICU since 1979 [144,145]. The incidence [4,5,100,146] and number of reports of nursery outbreaks associated with susceptible Streptococcus faecium [147], Streptococcus faecalis [148], and, more recently, VRE [146,149,150,151] have increased substantially since the 1980s, but enterococci still account only for a relatively small proportion of neonatal infections. In regions of the United States where VRE have become problematic HAI pathogens, spread to the NICU has occurred [146]. Repetitive sequence polymerase chain reaction (rep-PCR) DNA fingerprinting is useful to distinguish outbreak strains from endemic strains and “background” nursery flora and to identify distinct
P.425
clusters [150]. Invasion occurs most often from the infant's endogenous flora. In enterococcal outbreaks, however, environmental contamination, exposure to broadspectrum antimicrobials, horizontal transmission by HCW hands and, for VRE, repeated introduction by infants transferred to a regional NICU from a birth hospital with endemic VRE are important factors in the perpetuation of outbreaks.
Enterococci are associated only rarely with early onset sepsis. Most often, enterococci are isolated from the blood of low and VLBW infants with serious underlying conditions who have been in the NICU for >30–60 days. Prolonged use of CVCs, necrotizing enterocolitis, and bowel resection in addition to exposure to antibiotics are associated findings. Polymicrobial BSI is especially likely to occur in association with intra-abdominal disease, supporting the intestine as a point of invasion of enterococci. Clinical manifestations of enterococcal infections are nonspecific. The unusual episodes of meningitis [147] and endocarditis require prolonged (≥3 weeks and ≥6 weeks, respectively) treatment with combination therapy using a cell wall active agent combined with an aminoglycoside for cure. Successful treatment of endocarditis in a neonate with linezolid has been reported [152], but no experience with daptomycin in neonates has been published at this time. Linezolid should be restricted only when no other drugs are available to reduce the risk of emergence of resistance.
Gram-Negative Bacilli (GNB)
Infections caused by GNB are associated with the highest case-fatality rates: 40% (range, 24–62%) but approaching 90–100% in earlier studies. GNB infections in neonates have increased in many NICUs [6,100,102,106,153]. E. coli and Klebsiella or Enterobacter spp. are the most frequently isolated GNB in the absence of an outbreak [4,5,6,99,100,102,106]. ESBL-producing isolates of these species that have been reported by NICUs in recent years are especially problematic because of their resistance to the antimicrobial agents used routinely for empiric gram-negative coverage [51,94,97,154]. Of note, multispecies outbreaks of ESBLs can represent transfer of the same plasmid among species with person-to-person transmission [153]. Other GNB responsible for temporally related clusters in the NICU include P. aeruginosa [61,92,95,96], Serratia spp. [155,156,157,158], Citrobacter diversus [159,160,161], Salmonella spp. [17,162], Acinetobacter spp. [16], Chryseobacterium (Flavobacterium) meningosepticum [163], Ralstonia pickettii [56,164], and Burkholderia cepacia [164].
Molecular typing techniques (e.g., PFGE, ribotyping, polymerase chain reaction (PCR)-based methods) are valuable tools for defining the epidemiology of GNB infections in the NICU during both epidemic and nonepidemic periods [165,166] and should be used during outbreaks to determine whether horizontal transmission or selective antibiotic pressure is the predominant mechanism of spread and to guide the development of effective interventions (see Chapter 16). Although nursery outbreaks of GNB infections have been associated with environmental contamination (e.g., antiseptic solutions [157], intravenous medications and solutions [157,158,164,167], human breast milk [58,59,60], sinks [158,163], respiratory therapy and resuscitation equipment [17,56], and bandages [16]), such contamination is implicated less frequently in well-developed countries with clean water supplies and well-defined protocols for sterilization of equipment and single use of disposable items (see Chapter 20). Some outbreaks of Serratia marcescens have been controlled by enforcing recommended general infection control measures, even when a source for the Serratia spp. cannot be determined or when isolates are unrelated by molecular typing [156]. In the absence of an environmental source, it is likely that a virulent epidemic strain is acquired from the infected mother or—rarely—from an colonized or infected HCW or emerges from the infant's endogenous flora and then is transmitted horizontally to HCW hands. Identical strains of C. diversus have been identified in maternal infant pairs [160] and in HCWs and infected neonates [161]. Prolonged rectal colonization in addition to hand colonization has been implicated in outbreaks of C. diversus meningitis that were not controlled until colonized HCWs were removed from the nursery [160,161]. In the absence of outbreaks, clones of resistant GNB have been shown to be acquired but cleared rapidly from the infants' flora after initial colonization. However, horizontal transmission of these organisms occurs and occasionally results in clinical disease [165]. Clones of multidrug-resistantEnterobacter spp. transmitted in the NICU also were found in geographically distinct areas of a children's hospital [166].
Clinical manifestations characteristic of GNB infections include necrotizing ophthalmitis, pneumonia, ecthyma gangrenosum, cardiovascular collapse, and meningitis. Although ecthymatous lesions are associated most frequently with P. aeruginosa invasive disease, they can be associated with other GNB pathogens or fungi. Biopsy and culture of such lesions are helpful in isolating the specific pathogen. Salmonella spp. have a proclivity for causing osteomyelitis and septic arthritis, and a meningitis that is especially difficult to cure even with prolonged courses of antibiotics that are highly active in vitro against the infecting strain. C. diversus meningitis is notable for occurring in clusters in NICUs and has an association with brain abscess in 77% of patients compared to 7% of patients with meningitis caused by other GNB [160]. Flavobacterium meningosepticum, an unusual GNB resistant to most antimicrobial agents used for empiric treatment of gram-negative meningitis in the newborn and is treated successfully by the combination of vancomycin and rifampin [168], is a less frequent cause of epidemic GNB meningitis [163].
The importance of antibiotic use in selecting multidrug-resistant strains of GNB in the NICU for aminoglycosides has been well documented [169,170] and for
P.426
third-generation cephalosporins [51,171,172]. Monitoring for the development of resistance to the first-line aminoglycoside, usually gentamicin or tobramycin, is helpful to guide changes in antibiotic prescribing patterns. In the absence of multidrug-resistant strains, amikacin, third-generation cephalosporins, and meropenem are best reserved for rare individuals whose infection resistant routine treatment regimens. The combination of an aminoglycoside and third-generation cephalosporin is recommended to treat GNB meningitis even with organisms that are fully susceptible to aminoglycosides [62].
Bordetella Pertussis
Transmission of B. pertussis has been reported in normal newborn and intermediate care nurseries [173]. The source is usually an undiagnosed HCW [174] or visitor [175]; therefore, screening and administering the adult Tdap vaccine according to published recommendations [173] are especially important for family members likely to visit infants in the nurseries and for HCWs caring for such infants. Administration of adult Tdap vaccine to family members of infants in the NICU has been suggested as a method to improve vaccine uptake as demonstrated for influenza vaccine [176]. Healthcare facilities should provide the adult Tdap vaccine for HCWs.
Clostridium Difficile
Outbreaks of Clostridium difficile-associated colitis do not occur in the newborn nursery. In contrast to healthy adults who have an asymptomatic colonization rate of <5%, toxin-producing strains of C. difficile could be recovered from the stools of as many as 55% of asymptomatic neonates [177]. Disease usually does not occur because the immature intestinal mucosa of the neonate lacks receptors for the C. difficile toxin. This organism does not play a role in the pathogenesis of necrotizing enterocolitis, but an association with severe enterocolitis in infants with Hirschsprung's disease has been reported [178].
Fungal Infection
Of the subspecialty services in NNIS hospitals conducting hospitalwide surveillance, the HRN had one of the highest nosocomial fungal infection rates for the period 1986–1990: 7.6 per 1,000 discharges, after burn/trauma (16.1), cardiac surgery (11.2), and oncology (8.6) [179]. During this 5-year period, the fungal infection rate in the HRN increased from 4.7 to 9.6 per 1,000. In one study of candidemia in a pediatric population from 1988 to 1992, 25% of patients were premature infants [180]. In the 1986–1994 NNIS surveillance period, Candidaspp. accounted for 7% of all BSIs in the HRN: it was fourth after CONS, GBS, and S. aureus [4]. The most recent data from 128 NICUs participating in the NNIS system showed a 24% reduction in the number of infections per 1,000 patient-days among neonates <1,000 g from 3.5 during 1995–1999 to 2.68 (p < .01) during 2000–2004 with a stable rate in the larger neonates [181]. The decreases occurred among both C. albicans and C. parapsilosis BSI. There was no increase in species (e.g., C. glabrata or C. krusei) that would be expected to demonstrate resistance to fluconazole. NICU-specific rates of hospital-acquired candidemia for infants <1,000 g varied substantially: The median attack rate was 7.5%, but 25% of NICUs reported rates >13.5%; this is similar to the variation from 2.4–20.4% reported from the smaller National Institute of Child Health and Human Development Neonatal Network. However, data to identify practices in the NICUs with lower rates were insufficient.
The most consistent risk factors for candidemia in NICU infants reported in case-control studies using multivariate analyses are CVC use; previous bacterial BSIs; gastrointestinal pathology; abdominal surgery; gestational age <26 weeks; colonization, especially in multiple sites; and prolonged courses of broad-spectrum antibiotics (e.g., primarily third-generation cephalosporins) [67,68,182,183,184,185,186]. The importance of other risk factors, including infusion of hyperalimentation fluids, especially those containing lipid emulsions, delayed feedings, endotracheal intubation, administration of histamine-2 (H2) blocking agents, and corticosteroids as independent risk factors identified in some studies is not well established. In most instances, candidemia develops as a consequence of endogenous colonization rather than as a result of cross-contamination. Baley et al. [187] reported a 26.7% fungal colonization rate in infants weighing <1,500 g at birth who were followed prospectively. Two-thirds of these infants were colonized within the first week of life, probably reflecting maternal transmission during labor and delivery. Systemic disease developed in 7.7% of colonized infants. Although the National Epidemiology of Mycosis Survey (NEMIS) study group did not identify colonization as an independent risk factor in the six participating NICUs in 1993–1995, gastrointestinal (GI) tract colonization preceded candidemia in 43% of case-patients [67]. Once colonized, invasion arises most often from the GI tract. Early colonization is more likely to be associated with Candida albicans,whereas other species, such as C. parapsilosis and C. tropicalis, are more likely to be associated with late colonization that can result from horizontal transmission among infants or from HCWs.
P.427
as a suppository [189]. The presence of onychomycosis in HCWs has been suggested as a possible source of nosocomial candidiasis spread by hand carriage [194]. Thus, it is likely that wide variation in rates of Candida BSIs is a marker for management choices and infection control practices [195].
Clinical manifestations of disseminated candidiasis usually are nonspecific. Neonates with a maternal history of vaginal moniliasis rarely contract congenital mucocutaneous candidiasis [84]. At birth, such infants have an intensely erythematous maculopapular eruption of the trunk and extremities that rapidly becomes vesicular and pustular and then resolves with extensive desquamation. Palms and soles are almost always affected. These infants who are otherwise asymptomatic do not have systemic involvement and respond well to topical antifungal therapy. In contrast, the development of a diffuse, erythematous, scaling, burnlike dermatitis in an infant <1,500 g birth weight is more likely to be a manifestation of invasive disease and requires systemic therapy [196].
Candida spp. are less frequently recovered from blood cultures of infants with invasive candidiasis than from infants with invasive bacterial disease. Candidemia is present most often in association with a CVC. In such patients, the CVC should be removed as soon as possible to facilitate clearance of the bloodstream. Persistently positive cultures of blood forCandida sp. is a risk factor for focal complications, and the risk increases with the increasing duration of positivity [197]. A minimum of two weeks of antifungal therapy is recommended when the CVC has been removed in the absence of other sites of focal infection to prevent the development of other foci. True candiduria can reflect disseminated disease or localized cystitis; thus, isolation of Candida sp. from the urine indicates the need to evaluate other sites of infection. Echocardiogram can best detect Candida spp. endocarditis associated with a CVC in an infant with a structurally normal heart. Large vegetations can be present in the absence of a murmur and signs of congestive heart failure [198]. At autopsy, fungal vegetations have been found in infants with endocarditis unsuspected before death. Osteomyelitis, septic arthritis, meningitis, and brain abscess are other foci of infection that can be present with relatively few specific physical signs. The most severe episodes of meningitis and endocarditis require prolonged treatment courses of amphotericin B, up to 40 to 50 milligrams per kilogram (mg/kg) total dose or at least 6 weeks of a liposomal amphotericin preparation, usually in combination with a second antifungal agent. At the time of publication, experience with the newer azoles (e.g., voriconazole) and echinocandins (e.g., caspofungin, micafungin) in neonates is not sufficient to determine whether these agents offer for treating any advantage serous invasive candidiasis. Medical therapy of VLBW infants with Candida spp. endocarditis with a combination of two antifungal agents has been successful and is often preferred due to the high risk of surgical excision of vegetations in such infants. A review of the literature found little difference in survival between those neonates treated with antifungal therapy only (65%) compared with those who were treated with antifungal therapy and surgery (60%), p = 1.0 [199].
Other fungi associated with HAI in the NICU include Aspergillus spp. [199,200], Malassezia spp. [202,203], Rhizopus spp. [204], and Trichosporon beigelii [205]. Aspergillus, Rhizopus,and Trichosporon spp. are acquired from the environment via exposure to either construction dust or contaminated medical supplies. Infants with aspergillosis are notably not neutropenic but are premature and do not have normal chemotaxis and phagocytosis. Cutaneous aspergillosis can occur as the initial manifestation of disease or as one site of involvement in disseminated disease. Because of the risk of dissemination from primary cutaneous lesions and the difficulty in distinguishing primary from secondary lesions, aggressive systemic antifungal therapy is always indicated. Successful systemic voriconazole therapy of severe primary cutaneous aspergillosis that was refractory to amphotericin B in VLBW premature infants has been reported [201].
Both Malassezia furfur [202] and Malassezia pachydermatis [203] have been associated with BSI clusters in high-risk premature infants who have received intralipids through CVCs.Malassezia spp. isolates are most frequently isolated from blood obtained through a CVC and are rarely recovered from peripheral blood. M. furfur skin colonization rates of 25–84% have been reported for infants during prolonged NICU admissions, whereas >5% of infants hospitalized in a non-NICU setting or attending a well-baby clinic were colonized with this organism [202]. Malessezia spp. infections can manifest as a self-limiting condition, neonatal cephalic pustulosis on the face, scalp, or neck, and as a more severe clinical sepsis in neonates with CVCs. When temporally related episodes occur, these organisms most likely are carried from patient to patient on HCW hands. However, in one outbreak of Malasseziapachydermatis, transmission from pet dogs to HCWs and then to infants in the NICU was likely [203]. Because M. furfur requires an exogenous source of lipids to support its growth, the clinical microbiology laboratory inoculates the specimen on a solid agar plate that is then covered with a layer of sterile olive oil. The Isolator™ (Dupont Company, Wilmington, DE) blood culture system is the most convenient system available for isolating these organisms.
Viral Infections
Outbreaks of viral infections in the NICU have been recognized with increased frequency with the advent of improved techniques for viral isolation and identification of viral antigens using direct fluorescent antibody, enzyme-linked immuno assay (ELISA), and PCR technology (see Chapter 9). In nursery outbreaks of viral infections, the virus can be introduced by a staff member [206], and rates of infection of staff members could be high [207,208,209,210,211].
P.428
Because staff members have mild illness, they continue to work and to spread the pathogen by direct inoculation of their infected secretions. Respiratory syncytial virus (RSV) and rotavirus are the most frequently identified viral agents associated with nosocomial transmission [208,210] (see Chapters 32 and 34). Other respiratory viruses, notably influenza [211], parainfluenza [75], and adenoviruses [206] can be associated with clusters of respiratory infections having clinical manifestations similar to these of RSV. Manifestations of RSV infections in infants <3 weeks of age are atypical and more likely to include apnea, lethargy, and poor feeding in the absence of respiratory symptoms. Beyond 3 weeks of age, bronchiolitis and pneumonia are characteristic of RSV infections; apnea can precede the onset of respiratory symptoms. Infants with bronchopulmonary dysplasia requiring oxygen, congenital heart disease with pulmonary hypertension, and congenital immune deficiency syndromes are at the greatest risk for developing severe disease after infection with RSV. Substantial medical and economic impact of an RSV outbreak in an NICU has been quantified [212]. Several different types of adenovirus have been associated with nursery outbreaks [206,213]. The most common clinical manifestations are conjunctivitis and pneumonia. Direct inoculation by ophthalmologic instruments and transmission by the droplet route have been documented [213]; specific guidelines for retinopathy of prematurity (ROP) ophthalmologic examinations to prevent contact transmission include using gloves with change between patients, soaking instruments in 70% alcohol solution for 5–10 minutes, and changing alcohol solution twice daily have been issued [206]. Neonates can manifest multiorgan involvement with cardiovascular collapse and a bacterial sepsislike clinical syndrome with a case-fatality rate of 84%. Severity of illness can be determined by identifying the specific type of adenovirus causing the infections. The role of emerging respiratory viruses detected in infants and young children (e.g., human metapneumovirus, bocavirus) for neonates and potential for nursery outbreaks remains to be identified.
Nursery outbreaks of viruses with a gastrointestinal reservoir have been described, and those of coxsackie virus and echovirus infections have been reviewed [214]. In most outbreaks, the source patients acquired infection from their mothers and had severe disease. Case-fatality rates associated with hepatitis can be as high as 83%. Infants who acquire infection by nosocomial transmission have milder disease, with case-fatality rates of 12% or less. HCWs also can be vectors for spreading this group of viruses. The most common clinical manifestations of neonatal coxsackie virus and echovirus infections are hepatitis, meningoencephalitis, myocarditis, and pneumonia and are similar to the disease associated with neonatal herpes simplex encephalitis and disseminated infections. Specific diagnostic testing using sensitive and specific PCR assays is essential to distinguish these agents because no antiviral agents are available to treat coxsackie virus and echovirus infections.
Rotavirus and hepatitis A virus are transmitted by the fecal–oral route. Rotavirus transmission in the nursery is well documented; infection can be asymptomatic or associated with mild to moderate or severe diarrhea. Rotavirus-associated outbreaks of necrotizing enterocolitis have been reported [208]. HAI outbreaks of hepatitis A virus are extremely rare because of the relatively brief duration and low titer of viral shedding within the stool [215,216,217] (see Chapter 42). However, immunologically immature preterm infants can excrete hepatitis A viral antigen and RNA for as long as 4–5 months after the acute infection [216]. The source infants in reported outbreaks acquired hepatitis A virus either by vertical transmission from the mother before or during delivery [215] or by blood transfusion [216,217].
The following viruses are not transmitted horizontally in the nursery: human immunodeficiency virus (HIV), HBV, hepatitis C, and cytomegalovirus (CMV). Inadvertent exposure to expressed breast milk from a different mother who is infected with HIV has the theoretical risk of transmitting it. The source mother and the exposed infant should be tested for HIV under these circumstances [62]. Observing standard precautions and screening blood for transfusion and for CMV as well as using filtered blood or leukocyte-poor red blood cells have prevented transmission of infection (see Chapter 13). There is no significant increase in the excretion of CMV in infants in the nursery compared with patients in other areas of a pediatric hospital [218], nor is there a significant increase in the risk of acquisition of CMV infection among HCWs on pediatric or neonatal units compared with similar adults without hospital exposure [219,220]; thus, pregnant HCWs are not restricted from caring for infants known to excrete CMV [220]. Transmission of either herpes simplex [221,222] or varicella-zoster virus [223] in the nursery is extremely rare. Although transmission from oral lesions of nursery HCWs has been reported [222], the risk is so low that such individuals are no longer excluded from patient care as long as the lesion can be covered until dried [62]. In contrast, HCWs with herpetic whitlow are excluded from direct patient contact in the nursery. With widespread use of the varicella virus vaccine licensed in 1995, the number of susceptible or infected neonates, mothers, visitors, and HCWs has decreased substantially.
Tuberculosis
Congenital tuberculosis (TB) is extremely rare but can occur even when the mother is asymptomatic. Infants in the NICU who have TB are unlikely to transmit Mycobacterium tuberculosis to other infants and visitors but can transmit it to HCWs with unprotected close exposure [224]. However, M. tuberculosis has been transmitted to exposed neonates in the maternity unit, newborn nursery, and in recent years NICU and, to visitors and HCWs with active pulmonary TB in countries where TB is considered endemic [225] and
P.429
in the United States [226,227,228,229]. The neonate is at great risk for developing a severe disseminated disease, including meningitis, once exposed to an individual with active TB. Thus, recognition of an exposure must be followed by a comprehensive evaluation and initiation of chemoprophylaxis when the exposure is significant. A decision analysis has been developed to assist neonatologists and infection control personnel in managing of M. tuberculosis exposures [229]. When a TB diagnosis is considered in a young infant, a liver biopsy that shows granulomas and acid fast bacilli is diagnostic.
HAI Treatment
Physicians treating specific HAI can consult their local nursery guidelines, the AAP Redbook [62], and Nelson's Pocket Book of Pediatric Antimicrobial Therapy [230] for specific recommendations concerning choice of agent, dosing regimen, and duration of therapy. Empiric therapy for early onset disease consists of ampicillin combined with an aminoglycoside. Ampicillin provides activity against GBS, enterococci, and Listeria monocytogenes. If S. aureus is suspected, methicillin, oxacillin, or—in areas with high prevalence of MRSA—vancomycin is recommended for gram-positive coverage. The choice of aminoglycoside is based on susceptibility patterns in the nursery, but amikacin usually is reserved because of its greater activity against some GNB resistant to gentamicin and tobramycin as well as its higher cost. In one study, empiric use of ampicillin and cefotaxime during the first three days of life was associated with an increased risk of death compared with ampicillin and gentamicin, but the mechanism is unknown [231].
When treating suspected late onset sepsis, the choice of antimicrobial agents varies substantially neonatologists [232]. Oxacillin can be used for gram-positive coverage safely in NICUs that do not have MRSA as a prevalent pathogen and in infants with CVCs who are minimally symptomatic. Blood cultures from two separate sites are helpful in distinguishing CONS that are skin contaminants from true pathogens. The advantage of not using vancomycin routinely is the reduced risk of emergence of VRE, vancomycin-resistant staphylococci, and GNB infections [232,233]. Third-generation cephalosporins are not recommended for routine empiric use because of the rapidity of emergence of resistance in an individual patient during therapy and in microflora of the unit [51,171,172,234] as well as the association with an increased risk of candidiasis in VLBW infants [66]. However, cefotaxime is added to the aminoglycoside if there is a strong suspicion of GNB meningitis. Once the identification and susceptibility of the GNB-causing meningitis is known, cefotaxime in combination with an aminoglycoside is continued for at least 10 days and a minimum of a 21-day course is completed. Cefatazidime, cefipime, piperacillin-tazobactam, and meropenem can be substituted for treatment of susceptible strains of Pseudomonas and Stenotrophomonas spp. and other multidrug-resistant GNB. Ceftriaxone is not used in the neonate because of its displacement of bilirubin from albumin-binding sites [234] and its strong suppressive effect on the neonate's developing gastrointestinal flora [236]. Monitoring susceptibility patterns for GNB and antibiotic use is necessary to ensure that the currently recommended agents will be effective against the current pathogens (see Role of Microbiology in Chapter 15). With the exception of CONS-BSI, meningitis can be present in 10–20% of BSIs and in the absence of BSI or abnormalities of the cerebrospinal fluid (CSF) cell counts, glucose, and protein in 30–40% of infants with late onset sepsis. Therefore, a lumbar puncture for CSF culture and analysis is essential to diagnose meningitis [237,238]. Meningitic dosages of antimicrobial agents are recommended until meningitis has been ruled out.
Specific anaerobic coverage rarely is required in neo-nates. The most frequent indications are intra-abdominal sepsis, necrotizing enterocolitis, and intestinal perforation. Clindamycin, metronidazole, piperacillin-tazobactam, and meropenem provide activity against most anaerobes. Now that safer drugs for treating GNB infections are available, chloramphenicol is not recommended for neonates because of its toxicity and unpredictable pharmacokinetics, and trimethoprim-sulfamethoxazole is not used because of the rare idiosyncratic hepatotoxity secondary to sulfa component.
Removing the catheter as soon as CLA-BSIs associated with possible to clear the fungemia and prevent seeding other foci is essential for Candida sp. Traditionally, Amphotericin B and fluconazole have been the only treatment options for invasive fungal disease in the neonate. Although flucytosine was added to amphotericin B for enhanced activity in the most severe disease (e.g., endocarditis), its gastrointestinal toxicity prohibited its use for a prolonged course. While C. albicans had been nearly universally susceptible to fluconazole, the increasing frequency of infections caused by more resistant Candida spp. and the toxicity associated with amphotericin B have made antifungal treatment more difficult. Although most studies of fluconazole prophylaxis of high-risk neonates have not found emergence of resistance, one report described the emergence of fluconazole-resistant subclones of a single strain of C. parapsilosis that was a major cause of candidemia during a 12-year period when fluconazole prophylaxis was used [239]. Fortunately, The development of antifungal agents with less toxicity and greater activity against the more difficult to treat fungi has recently surged [240]. The new azole drug voriconazole now is preferred for the treatment of invasive aspergillosis. This drug has not been studied in neonates and there is theoretical concern about the effect on the developing retina because of the visual adverse events reported in adults and older children. The echinocandins represent a class of agents that interfere with fungal cell wall biosynthesis by inhibition of an enzyme present in
P.430
fungi but absent in mammalian cells. These drugs are not substrates for the hepatic cytochrome p450 isoenzyme system or the intestinal glycopeptides, thereby reducing drug interactions. The echinocandins have excellent fungicidal activity against most Candida spp. and are fungistatic against Aspergillus sp. Caspofungin is the first in this class of agents to be studied. Although no large studies of the use of in neonates have been performed, two case-series [241,242] and individual case-reports indicate that it successfully treats refractory candidiasis in VLBW neonates and is well tolerated. Micafungin is another agent in this class that has undergone pharmacokinetic studies in neonates [243] and is now being used in larger neonatal studies.
Acyclovir is the preferred treatment for Herpes simplex and Varicella-Zoster viral infections. In a large, multicenter randomized controlled trial of ganciclovir therapy of infants with congenital CMV infection and central nervous system involvement, treatment prevented deterioration of hearing [244]. Studies to identify the patients who will experience the greatest benefit of treatment and to determine the optimal treatment regimen are ongoing. Treatment of RSV infections is primarily supportive. Recommendations for the use of aerosolized ribavirin therapy in RSV infections have been modified since the drug's licensure in 1986 because of the continuing questions concerning efficacy. Decisions concerning ribavirin use must be individualized and made with the knowledge of conflicting data concerning efficacy. Ribavirin can be considered for use in the infected, high-risk, low-birth-weight infant with chronic bronchopulmonary dysplasia, congenital heart disease, congenital immunodeficiency syndrome, and other serious underlying conditions such as neurologic and metabolic disease and multiple congenital anomalies [62]. Ribavirin must be initiated early in the course of disease for optimum effectiveness. Some experts recommend the addition of the monoclonal antibody palivizumab to treat severe disease in high-risk patients based on anecdotal experience in adult bone marrow transplant patients [245]. No efficacy data in human infants have been published; therefore, further studies are needed.
|
TABLE 25-3 |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
HAI Prevention
Nursery Design and Staffing
The AAP and the ACOG collaborate to develop guidelines for all aspects of perinatal care, including infection control in the nursery [246]; they revise these guidelines at regular intervals; the next revision is scheduled for publication in 2007. Table 25-3 summarizes the recommended nurse/infant ratios based on the acuity of the medical condition and the amount and type of nursing care and support equipment needed. Nurse/infant ratios below those recommended have been associate with increased rates of bacterial invasive disease caused by MSSA [73], MRSA [26], Enterobacter cloacae [74], and other bacteria [72,76] and viral respiratory infections [75] in the nursery and NICU. Spatial requirements are defined in the 2006 Guidelines for Design and Construction of Health Care Facilities published by the American Institute of Architects (AIA) (www.aia.org/aah_gd_hospcons) and by the Consensus Committee on Recommended Standards for Newborn Intensive Care Unit Design [247] (Table 25-3). Single-patient rooms in the NICU can offer an infection control advantage,
P.431
but more data are needed to support a universal recommendation.
Many experts recommend central high-efficiency particulate air (HEPA) filtration when constructing a new NICU due to the vulnerability of VLBW infants to infection caused by airborne spores and because surgical procedures (e.g., ECMO cannulation/decannulation, exploratory laparotomy in infants with necrotizing enterocolitis) are often performed in the NICU when high-risk infants are too unstable to withstand transport to the operating suite.
|
TABLE 25-4 |
||
|
HAI Surveillance
An active surveillance program is an essential component of HAI prevention (see Chapter 5). In most hospitals, surveillance of positive clinical culture results is performed by infection control personnel who are responsible for several different units in the hospital. The CDC recommends the following: [1] prospective surveillance on a regular basis by trained infection control professionals using standardized definitions, [2] analysis of infection rates using established epidemiologic and statistical methods (e.g., calculating rates using appropriate denominators that reflect duration of exposure and using statistical process control charts for trending rates), [3] regularly using data in decision making, and [4] employing an effective and trained healthcare epidemiologist who develops infection control strategies and policies and serves as a liaison with the medical community and the administration (2–22;248,249). For the NICU, birth-weight categories are used for risk stratification. A working definition of epidemiologically important organisms has been developed to assist the infection control team in recognizing pathogens that require further investigation and preventive measures; see Table 25-4 [250]. When infection control personnel identify temporally related clusters of clinical infections and/or epidemiologically important pathogens, especially those that are multidrug-resistant, they collaborate with the nursery staff to develop a prevention program (see Chapter 2). As soon as a cluster or outbreak is suspected, infection control personnel should notify the microbiology laboratory of the need for active surveillance cultures and to save isolates should molecular fingerprinting studies be indicated (see Chapter 6). Designating a nursery staff member who understands the psychology and operational logistics of the unit as the infection control liaison facilitates education and adherence to policies for hand hygiene, isolation precautions, cohorting of patients and staff, proper cleaning, disinfection and sterilization of medical equipment, and other aseptic practices [26,251,252]. Participation of nursery staff members in the design of prevention programs increases adherence and success. feedback of positive results is most important to the staff when they have succeeded in controlling an outbreak; guidelines to prevent a similar outbreak is essential.
P.432
Routinely culturing neonate body surfaces (i.e., skin, umbilicus, mucous membranes, tracheal aspirates, rectal swabs) is not recommended in nonoutbreak settings because they do not predict which infants are at risk for sepsis and are very costly [253,254]. In contrast, body surface cultures are helpful during an outbreak to identify infants who are colonized with the target pathogen in addition to infants who have positive clinical cultures for target pathogens (e.g., MRSA, VRE, and multidrug-resistant GNB). Newly admitted infants must be kept in cohorts separate from colonized and infected infants to limit horizontal transmission of the outbreak organism. In addition, it can be useful to take active MRSA and VRE surveillance cultures at regular intervals (e.g., 3, 6, or 12 months) in NICUs that have no clinical episodes of MRSA or VRE to detect the presence of target multidrug-resistant organisms (MDROs) establishing high colonization rates and clinical disease [130,250]. Regional collaboration for NICU surveillance and control of MRSA has proven particularly useful [130].
Isolation Precautions
The 2006 AAP Redbook [62] has incorporated the recommendations for standard and transmission-based precautions published by the CDC in 1996 [255] (revised for anticipated publication in 2007); the recommendations should be consulted for managing specific infections. Standard precautions remain the foundation in infection control to which other categories of precautions are added. Standard precautions have been expanded to include recommendations for respiratory hygiene/cough etiquette for individuals with symptoms of respiratory tract infection. However, such individuals should be restricted from entering the nursery. See the Tables 25-5 and 25-6, for the components of standard precautions and transmission-based precautions respectively. Table 25-7 summarizes the recommended precautions for the most frequently encountered neonatal infections that require contact, airborne, or droplet precautions in addition to standard precautions. In NICUs with single-patient rooms, personal protective equipment (PPE) needed for contact and droplet precautions should be donned upon entry into the room. In NICUs with multibed pods or bays, the space around the isolette of the infant who requires contact and/or droplet precautions is usually designated with tape on the floor and signs on the isolette. A separate room is not required for infants on contact or droplet precautions, but spatial separation from uninfected infants is preferred. Although enclosed isolettes provide a limited amount of barrier protection relative to open warmers or bassinets [129], they cannot be relied on to prevent pathogen spread to other infants by HCW hand carriage. Neonates usually are unable to generate large-particle droplets spontaneously, but endotracheal suctioning or administration of aerosol treatments can generate infectious droplets. According to standard precautions, masks are indicated if splatter of respiratory secretions is anticipated (e.g., endotracheal suctioning or intubation). At least one airborne infection isolation room (AIIR) with negative-pressure ventilation that meets standard requirements is recommended for every nursery/NICU to isolate neonates with perinatal exposure to maternal varicella and suspected or confirmed TB. Most other infections do not require special isolation rooms.
During outbreaks, the most effective method for preventing horizontal transmission is by cohorting infants who are colonized or infected with epidemiologically important pathogens away from newly admitted patients and ideally with dedicated personnel who do not care for newly admitted infants or infants who are not colonized or infected [26,130,256,257]. If transmission continues in the presence of strict cohorting, continual introduction of the epidemic organisms from a carrier or multiple different sources or from ineffective hand antisepsis is likely [159].
Hand Hygiene and Gloves
Performing hand hygiene between patient contacts is the single most important measure for preventing HAIs in the nursery [258](see Chapter 13). The scientific information documenting the role of transmission via the hands of HCWs, the efficacy for hand decontamination with either antimicrobial-containing soap or alcohol-based hand rubs, and the recommended use of gloves is summarized in the CDC/HICPAC Guideline for Hand Hygiene in Health-Care Settings [258]; an implementation guide has been developed through the collaboration of the Institute for Healthcare Improvement (IHI), CDC, Society of Healthcare Epidemiologists of America (SHEA), and the Professionals in Infection Control and Epidemiology (APIC) (www.IHI.org). The Guideline for Hand Hygiene summarizes nine studies, three of which were conducted in the NICU, newborn nursery, or general pediatrics units, that demonstrate a temporal relationship between the introduction of new hand-hygiene products and improved hand-hygiene practices and decreased MRSA infection rates when added to the other control measures that had been in place, including obtaining weekly active surveillance cultures and contact precautions. The important components of hand hygiene are consistency, duration of exposure, and antimicrobial content of the soap or waterless hand rub [258]. Alcohol-containing antiseptic hand gels and antimicrobial soap and water are preferred in nurseries and NICUs. A hands-free hand washing station should be provided in each single-patient room. Every infant bed in multi-bed rooms should be within 20 feet of a hand-washing station but should be no closer than 3 feet to other beds [247]. Having an alcohol gel dispenser mounted at each bedside provides the best opportunity for using hand hygiene consistently.
P.433
|
TABLE 25-5 |
||||||||||||||||||||||||||||
|
A 2-minute scrub from hands to forearms with an antiseptic-containing soap is no longer recommended for nursery staff at the beginning of a work shift [258]. However, published recommendations in the AAP/ACOG Guidelines for Perinatal Care for the number of nursery scrub areas are to have (1) one scrub area at the entrance to each nursery with faucets operated by foot or knee controls and (2) one scrub sink for at least every 6–8 patient stations in the normal newborn nursery and for every 3–4 patient stations in the admission/observation, continuing care, intermediate care, and intensive care areas [246]. While some scrub sinks are required because surgical procedures are performed in the NICU, their number can likely be reduced. Hands should be washed thoroughly with an antimicrobial soap and water on entry into the nursery/NICU, and hand hygiene should be performed using either an antiseptic alcohol hand rub or antimicrobial soap and water between patient contacts. As part of good hand-hygiene practices, HCWs who have direct contact with high-risk patients should have well-groomed and short natural nails [258]. The evidence base to support the recommendations for HCWs with direct contact with patients in ICUs includes four studies performed in NICUs [94,95,96,97]. Novel strategies to improve adherence to recommendations for hand hygiene and monitoring adherence are needed in the nursery/NICU as in all other areas of healthcare facilities.
P.434
Gloves must be worn when contact with blood or body fluids is anticipated or when handling anything on the transmission-based precautions list that has come in contact with an infant. Gloves are never washed between patients but must be changed between patient contacts. Hand hygiene is performed immediately upon glove removal.
|
TABLE 25-6 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gowns, Caps, and Masks
The use of gowns on entrance into the nursery is a longstanding ritual that many nurseries are hesitant to relinquish. Studies of varying design conducted in the nursery and NICU confirm gown's lack of efficacy for preventing HAIs [259,260]. In alternate 2-month gowning and no-gowning cycles, Pelke et al. [260] demonstrated no significant differences in the rates of bacterial colonization, HAIs including RSV and necrotizing enterocolitis, and mortality. In addition, compliance with hand washing was not increased during the gowning cycles, nor was traffic into the unit changed. Thus, in a nonoutbreak setting, gowns are not required for staff or visitors upon entrance to the nursery but are indicated when there is soiling with blood or body fluids is anticipated, following contact precautions is necessary, clustered infections with epidemiologically important organisms that are considered to have been transmitted by the contact route are present, and parental concern exists for excessively soiled clothing. Some nurseries have continued to use gowns for handling newborn infants (e.g., cuddling, feeding).
Caps and masks are indicated when performing sterile procedures, including CVC placement. Masks also are used as part of standard precautions and droplet precautions to protect HCWs and an infant from an individual HCW with respiratory tract infection who is considered indispensable and cannot be removed from the nursery/NICU. Respirators of N95 or higher are indicated for HCWs having contact with a patient with suspected or confirmed TB.
Prevention of Transmission of Multidrug Resistant Organisms (MDROs)
Control of MDROs requires three groups of interventions: (1) prudent antimicrobial use to prevent emergence of resistance, (2) implementation of bundled practices to prevent device-related and SSIs (www.ihi.org/IHI/Programs/Campaign), and (3) infection control measures to prevent transmission within the NICU [250]. Based on the voluminous number of published studies of control of MDROs in NICUs and in other healthcare settings, it is clear that transmission of MDROs in healthcare settings can be controlled, but the single most effective strategy for all settings has not
P.435
been established. Seven categories of measures are critical for preventing MDRO transmission [250]:
The most effective measure for MRSA and VRE control include a risk assessment, active surveillance testing for colonized patients, contact isolation for infected/colonized patients, and hand hand hygiene. Sole focus on MRSA or VRE may increase the possibility that multidrug-resistant GNB may emerge [261]. Expanding the program to include resistant GNB can be especially relevant for the NICU. The recommendations for implementing the baseline of these 7 measures in all healthcare facilities, especially in high-risk units within those facilities and defining target MDROs within a specific unit, assessing trends, and implementing a more intensified set of measures if rates of MDROs are not decreasing serve as a template for MDRO control in the NICU and include many studies of outbreak control in NICUs in the evidence base [250]. The NICU must participate in the overall facility strategies for MDRO control and interventions specific to its conditions.
|
TABLE 25-7 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Skin, Eye, and Cord Care
Initial cleansing of skin after birth is delayed until the neonate's temperature has stabilized. Warm water alone or warm water and a mild, nonmedicated soap are recommended when the skin is cleaned [246]. Hexachlorophene specifically is no longer recommended for routine daily bathing of neonates because of the neurotoxicity demonstrated previously when it is absorbed in large concentrations. However, chlorhexidine gluconate is poorly absorbed through intact skin and is therefore a suitable agent. When
P.436
intramuscular injections are given in the delivery room as part of prophylaxis regimens (e.g., penicillin G to prevent early onset GBS disease or gonococcal ophthalmia, ceftriaxone to prevent gonococcal ophthalmia, and vitamin K to prevent hemorrhagic disease of the newborn), the injection site must first be cleansed well with alcohol. This prevents the introduction of microorganisms such as HIV, HBV, and herpes simplex virus in maternal blood and body fluids that could be contaminating the infant's skin.
A single application within 1 hour of delivery of topical tetracycline (1%) or erythromycin (0.5%) ophthalmic ointment is preferred to prevent gonococcal ophthalmia. The use of 1% silver nitrate drops is discouraged because of the associated chemical irritation. The eyes should not be irrigated after instillation of any of these agents. Single-use tubes or vials must be used to prevent cross-infection. A single dose of ceftriaxone, 125 mg (25 to 50 mg/kg for low-birth weight infants) intramuscularly or intravenously, is recommended for infants born to mothers with active gonorrhea at delivery [62].
The umbilical cord has been reported to become colonized with S. aureus in up to 70% of infants within 48 hours after birth and can potentially serve as a point of entrance for pathogens to cause invasive disease. High colonization rates in the nursery are associated with an increased rate of postdischarge infection for term infants and longer hospital stay for low-birth weight infants. Therefore, most nursery protocols include antiseptic treatment. Certainly, any nursery experiencing an increased rate of S. aureus infections uses topical antiseptics for cord care. In this era of increasing prevalence of CO-MRSA in many communities, cord care to prevent MRSA colonization could be even more important. The role of antimicrobial applications to the umbilical cord to prevent bacterial colonization and infection has been reviewed [262,263]. These reviews conclude that there is evidence that applying antiseptic to the cord reduces bacterial colonization, but there is insufficient evidence to determine whether any agent is preferred. The delay in time to cord separation >7 days in infants treated with topical antiseptics is not significant clinically and should not deter the use of antibacterial products. Use of dry cord care and isopropyl alcohol has been associated with higher colonization rates than of the following antiseptic agents: (1) triple dye, a combination of brilliant green (2.29 mg/mL), proflavine hemisulfate (1.14 mg/mL), and gentian violet (2.29 mg/mL), (2) bacitracin ointment, (3) chlorhexidine, and (4) silver sulfadiazine cream (1% Silvadene™). However, consistently greater efficacy has not been demonstrated for any one agent [262,263]. Iodine-containing agents are not recommended because of the possibility of transcutaneous absorption and suppression of neonatal thyroid function. Short-term use of mupiricin (Bactroban™) ointment to control an outbreak associated with susceptible S. aureus or MRSA may be considered, but mupiricin is not recommended routinely because of the emergence of resistant strains after its frequent or prolonged use [264]. If a nursery discontinues antimicrobial care of the umbilicus in favor of dry cord care, it is important to conduct postdischarge surveillance to ensure the absence of adverse effects.
Traditionally, triple dye has not been used in the NICU because of theoretical concern about increasing systemic absorption and rendering the umbilical stump unsuitable for vessel catheterization. However, adverse effects have been observed since a routine single application of triple dye to the umbilical stump was extended to infants in the intermediate care area of Parkland Health and Hospital System in 1988 and then to the NICU area in 1991. In the NICU, the application of triple dye after umbilical vessel catheterization in the first 12 hours after birth was instrumental in controlling a prolonged MRSA outbreak [26].
Intravascular Catheters and Respiratory Therapy Equipment
Few controlled investigations of care practices of intravascular catheter or respiratory therapy equipment specifically in the newborn nursery have been conducted. Consequently, nurseries usually follow guidelines for the care of catheters and respiratory equipment based on studies in older children and adults [55,265] (see Chapters 31 and 37). PICC and surgically placed tunneled CVCs are used in the NICU, which have no significant differences in the complication rate [266]. Application of the bundled practices for preventing CLA-BSIs in children (www.chca.com/news/camapign.html) have been applied to neonates and have reduced rates of CLA-BSI significantly. The bundled practices include (1) assessing daily the need for the catheter and goals for removal, (2) using chlorhexidine to prepare the insertion site and for maintenance, (3) using maximal sterile barrier precautions for catheter insertion, (4) consistently using recommended hand hygiene, and (5) observing insertion and maintenance practices. Femoral lines must be protected against contamination from urine and stool. Although chlorhexidine has not been approved for use in infants <2 months of age, ample experience indicates that it can be used safely in neonates. Minimizing catheter manipulation and paying specific attention to aseptic technique when catheters are manipulated (e.g., hub, exit site, blood sampling) are important for preventing CLA-BSIs [267]. Practices for which some but not enough data suggest efficacy in support of recommendations for routine use include chlorhexidine sponge dressings (Biopatch)™ [268] and vancomycin-heparin lock solution [269]. The chlorhexidine sponge dressing should not be used in neonates <1,000 g or <7 days old or of gestational age <26 weeks because of reports of exudative-type local reactions and pressure necrosis occurring under the patch in these very premature infants [265,268]. Unique to the nursery is the use of
P.437
umbilical arterial and venous catheters for which many care issues remain unresolved. The rates of colonization and associated BSIs are similar for umbilical arterial and venous catheters. The umbilical insertion site must be cleansed with an appropriate antiseptic before catheter insertion. Tincture of iodine is not used because of the potential effect on the neonatal thyroid [246]. Umbilical arterial catheters should not be left in place >5 days. Removing umbilical venous catheters after 14 days is recommended, but a recent randomized trial reported no increased adverse events associated with umbilical venous catheters in neonates with birth weights <1,251 g for up to 28 days compared with removal of the umbilical venous catheter after 7–10 days and its replacement with a PICC [270].
Mouth-suctioning devices, such as the De Lee suction trap, are no longer used because of the risk of exposure to potentially infectious aspirated material that could enter the HCW's mouth. When a mechanical suction apparatus is used, the negative pressure should be no more than 100 millimeters of mercury when the suction tubing is occluded [246]. The optimal bundled practices for preventing ventilator-associated pneumonia in the NICU and PICU have not been determined, but the options for modifying the adult bundles and experiences in two institutions have been reviewed [271]. Only sterile water should be used for any device that provides humidification to the neonate, and all equipment must be cleaned and disinfected according to manufacturers' recommendations [55]. Isolation of pathogens such as Burkholderia cepacia and Ralstonia sp. from respiratory secretions should alert NICU personnel to the possibility of contaminated equipment [56,272].
Each nursery should develop protocols for care of the specific intravascular catheters and respiratory therapy equipment that are used in the specific unit. These protocols should be consistent with the current AAP or CDC guidelines.
Immunoprophylaxis
As described, the administration of both standard IVIG preparations [40,41,42] and high titer staphylococcal immunoglobulin products [44,45,46] either to treat or prevent neonatal sepsis has not been efficacious, even when serum IgG levels are maintained >400 millimeters per deciliter The largest multicenter, controlled trial, which enrolled 2,416 infants, reported lot-to-lot variation in antibody profile despite the use of thousands of donors for processing each batch of IVIG [42]. In contrast, palivizumab, a humanized mouse monoclonal antibody that neutralizes RSV and prevents viral binding to cells is recommended for monthly intramuscular injection during the RSV season for high-risk children to prevent severe disease and RSV-related hospitalizations [273]. High risk groups include (1) infants and children <2 years old with chronic lung disease who have required medical therapy for chronic lung disease within 6 months before the anticipated start of the RSV season, (2) infants born at ≤32 weeks gestation even if they do not have chronic lung disease, (3) infants born at 32–35 weeks gestation with ≥2 risk factors for RSV-associated hospitalization, and (4) children ≤24 months of age with hemodynamically significant cyanotic and acyanotic congenital heart disease. Although efficacy and safety have been demonstrated, cost effectiveness has not; therefore, targeting only those for whom efficacy has been shown is important [274]. A large multicenter trial of Numax™, a novel recombinant humanized IgG1 monoclonal antibody derived from palivizumab, has been completed, but results a not yet have been published. It is possible that this product could has improved efficacy.
A single dose of palivizumab is recommended for high-risk neonates at the time of discharge during the RSV season. However, because very premature infants do not sustain protective serum concentrations until after the second dose, it has been suggested that such infants should receive a dose at 1 month before discharge from the NICU in addition to the dose at discharge [275]. The use of palivizumab to control an RSV outbreak has not been studied. However, there is one report of the successful use of palivizumab administration to all infants in an NICU during an RSV outbreak that was not controlled by standard infection control measures [276]. It is important to note that RSV transmission can be well controlled [210] by following contact precautions, screening visitors and cohorting infants.
Varicella-zoster immune globulin (VZIG™ [Massachusetts Public Health Bilogics Laboratory, Boston, MS], VariZIG™ [Cangene Corp., Winneped, Canada]) is indicated for susceptible high-risk individuals exposed to varicella [277]. After an exposure in the NICU, VZIG or VariZIG 125 units are administered to premature infants of >28 weeks' gestation whose mothers have no prior history of varicella or varicella immunization and to all premature infants <28 weeks' gestation or ≤1,000 g birth weight regardless of maternal history due to lack of placental transfer of antibody in earlier stages of pregnancy [62,277]. Most premature infants of ≥28 weeks' gestation whose mothers are immune have acquired sufficient maternal antibody to protect them from severe disease and complications. However, chronologic age >2 months and seven or more transfusions of packed red cells can be associated with increased rates of seronegativity in infants whose mothers are immune [278]. Neither VZIG or VariZIG is recommended for healthy, term infants postnatally exposed even if their mothers have a negative history of varicella. Before administrating VZIG or VariZIG in a nursery exposure, obtaining serum from the infants to confirm susceptibility is helpful. If antibody determinations are available within 72 hours of the exposure, VZIG or VariZIG administration can be delayed until results are available. If antibody is present, the infant does not require isolation during the 10–28 days after exposure. VZIG is not indicated if an infant has received an infusion of IVIG for other indications within the previous 3 weeks.
P.438
It is prudent to provide influenza vaccine to the visiting family members to protect them and the infants in the NICU [62,211]. The NICU also provides an excellent opportunity to counsel adolescent and adult family members about the importance of receiving Tdap to protect their own infants and others in the NICU and to provide it to them [176].
Chemoprophylaxis
Maternal screening and chemoprophylaxis have dramatically reduced the risk of early onset GBS disease [105]. However, the use of antimicrobial agents for HAI prevention in the NICU is strongly discouraged because of the risk of the emergence of resistant microorganisms that will require more broad-spectrum and potentially more toxic antimicrobial agents for treatment (see Chapters 14 and 15). Two groups of investigators have reported the efficacy of low-dose vancomycin, 25 µg/mL of total parenteral nutrition fluid, to decrease catheter colonization and BSI caused by CONS. In two different, prospective, randomized, controlled trials of 70 and 150 VLBW infants, CONS-BSI was decreased in infants weighing <1,500 g from 34% to 1.4%, respectively, and in infants weighing <1,000 g from 26–2.8%, respectively [279,280]. Despite the beneficial effect, neither set of investigators and an accompanying editorial recommends the routine use of this regimen because of the risk of the emergence of vancomycin-resistant organisms [281]. Continued exposure of the normal flora to low concentrations of vancomycin creates especially favorable conditions for resistance to occur. In addition, morbidity and mortality associated with CONS infection is not severe enough to justify the risks. In another randomized study of 148 infants weighing <1,500 g with percutaneous CVCs, the use of amoxicillin 100 mg/kg/day intravenously in three divided doses had a negligible effect on the incidence of septicemia because the rate in the control group was so low, 2.7% [282]. In conclusion, strict adherence to the recommended bundled practices for CVC insertion and maintenance remains the preferred method of prevention of CLA-BSIs.
Far more controversial is the role of fluconazole for prophylaxis of invasive candidiasis in the high-risk VLBW infant. Many studies of fluconazole prophylaxis have been published [283,284,285,286,287,288], and the results and cautions to neonatologists are best summarized in editorials by Long and Stevens [195] and by Fanaroff [284]. Published studies of fluconazole prophylaxis have demonstrated wide center-to-center variation in rates of invasive candidiasis before initiating prophylaxis, have targeted VLBW infants <1,500 g and ELBW infants <1,000 grams and/or gestational age <30 or 32 weeks, and have administered fluconazole (1) daily for the first 30 days of life (2) daily during periods of administration of antibiotics for >3 days [288], (3) every third day for 2 weeks followed by every other day for 2 weeks and then daily for 2 weeks [284,285], (4) twice weekly for 6 weeks [284], or (5) every third day for 1 week followed by daily for 3 weeks [286,287]. All studies have been single-center studies, and most have been prepost intervention studies. Two small prospective, randomized placebo-controlled studies were published in 2001 [289,290]; the more recent prospective randomized, double-blind clinical trial compared two different dosing schedules but did not have a control group [284]. The conclusion from these studies is that fluconazole prophylaxis does prevent Candida spp. BSIs and reduce Candida-related mortality in ELBW infants. Unfortunately, the recent publication from the NNIS HRNs that demonstrated a decreased incidence in Candida spp. BSIs in ELBW infants in 2000–2004 compared with 1995–1999 could not determine whether the reduced rate was associated with the use of fluconazole [7]. The concern is that there is evidence of the emergence of nonalbicans Candida strains that are resistant to fluconazole, there has been no multicenter trial, and Candida spp. BSIs are markers of management choices and infection control practices. Evidence-based practices relating to management of CVCs and prudent use of antimicrobials are more likely to be long-term safe and effective measures than the use of an antifungal agent for prophylaxis. Hence, caution is urged until more is known.
Visitation of Siblings and Others
Because the acquisition of a seemingly innocuous viral infection in high-risk neonates can result in unnecessary evaluation and empirical therapy for septicemia and serious life-threatening disease, special visitation policies are required in nurseries and NICUs. All visitors with signs or symptoms of respiratory or gastrointestinal tract infection should be restricted from visiting any patients in healthcare facilities. During the influenza season, it is preferred that all visitors have received influenza vaccine. Increased restrictions could be needed in the midst of a community outbreak (e.g., severe acute respiratory syndrome, influenza). For infants requiring contact precautions, the use of PPE by visitors is determined by the nature of the interaction with the patient and the likelihood that the visitor will frequent common areas in the nursery/NICU area or interact with other infants' family members. Although the neonatology staff encourage visits by siblings in the NICU, the medical risk must not outweigh the psychosocial benefit. Studies demonstrate that parents favorably regard sibling visitation [291] and that bacterial colonization [292,293] or subsequent infection [294] does not increase in the neonate who has been visited by siblings, but these studies are limited by small numbers. Strict guidelines for sibling visitation should be established and enforced to maximize visitation opportunities and minimize risks of transmission of infectious agents. The following visitation recommendations can guide policy development:
P.439
Occupational and Employee Health
All HCWs in the all nursery areas must be immune to vaccine-preventable diseases. All nursery staff should be screened by history and, when indicated, serology for susceptibility to rubella, rubeola, varicella, and HBV (see Chapter 3). Appropriate immunizations must be provided for those who are seronegative. Annual influenza immunizations also should be administered to staff members, including pregnant women, in October and November of each year according to the CDC recommendations [295]. The cold-adapted, live-attenuated influenza vaccine (FLU-MIST™) can be administered safely to HCWs in the nursery without contraindications; because the amount of attenuated virus shed is below the infectious dose, the vaccine virus is unable to replicate at the higher temperatures of the lower respiratory tract, and no adverse effects have been reported in contacts of recipients of this vaccine [296]. This vaccine also offers the advantage of improved protection against drifted strains not contained in the vaccine as compared with the killed vaccine. The newest vaccine recommended for administration to HCWs in contact with young infants is the adult pertussis vaccine (Tdap) that should be administered as a single dose if ≥2 years have elapsed since the most recent Tetnus-diptheria vaccine (Td) [173]. Shorter intervals may be used in the midst of an outbreak.
Guidelines for removing HCWs with highly contagious conditions from direct patient contact in the nursery should be consulted for specific recommendations [246,297]. Decisions concerning the removal of personnel with respiratory, gastrointestinal, or mucocutaneous infections must be made on an individual basis. Removal of all individuals with mild illnesses could be impractical in an overcrowded, understaffed nursery. Therefore, specific instructions concerning precautions to prevent transmission of infection to patients must be given. Individuals with pertussis, active TB, varicella, exudative skin lesions, or weeping dermatitis must be removed from direct patient contact until they are no longer infectious. HCWs with herpes labialis (“cold sores”) are no longer excluded from the nursery because the transmission risk is so low. Such individuals are instructed to cover the lesions, not to touch the area surrounding the lesions, carefully observe hand-washing procedures, and not to kiss or cuddle neonates under their care. The role of topical penciclovir or oral acyclovir is not established but decreases the quantity and duration of viral shedding in treated individuals. HCWs with herpetic whitlow must be restricted from contact with neonates until the lesions are completely crusted. HCWs who are known to be carriers of hepatitis B surface antigen or infected with HIV are managed and counseled individually (see Healthcare worker and HIV Chapters 42 and 43). For those with HIV, continued patient contact is determined by the stage of disease and the absence of potentially transmissible infections and is governed by state law. Percutaneous and mucocutaneous exposures to blood-borne pathogens are managed according to standard protocols (see Chapter 44).
In the nursery, there has been much concern about the exposure of pregnant HCWs to neonates with congenital infections, especially CMV [220]. Frequently, the most anxiety is generated over infections that pose the least risk. Several epidemiologic studies in hospitals and day care centers have established that the nursery staff do not have an increased risk of acquiring CMV from their patients and that exposure to toddlers in day-care centers is associated with a significantly higher risk of seroconversion [218,219,220]. Therefore, pregnant women are not restricted from caring for infants who are identified as infected with CMV. All female HCWs in the childbearing age group must be taught to adhere strictly to standard precautions, especially hand hygiene and the use of gloves when contact with urine, saliva or blood is likely and that prepregnancy immunization against HBV, rubella, rubeola, and varicella is the most efficacious method of protecting themselves and their unborn children. Table 25-8 summarizes the relevant facts concerning infections that create the greatest concern among pregnant women in direct contact with young infants and as well as additional information [220]. We have found that pregnant HCWs are reassured greatly by reviewing this information with them.
The only restriction that is applied to pregnant HCWs is to avoid exposure to ribavirin. Although there are no data support the theoretic risks of teratogenicity in humans,
P.440
P.441
hospitals should follow the manufacturer's recommendation to restrict exposure [62].
|
TABLE 25-8 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
References
P.442
P.443
P.444
P.445
P.446
P.447