Introduction

Neonates requiring intensive care represent a particularly vulnerable population because of their frequent need for intubation (as part of resuscitation), prolonged mechanical ventilation, and susceptibility to hospital-acquired infections. Hospital-acquired infections in neonates increase costs, prolong hospitalization, and are major causes of morbidity and mortality. In pediatric intensive care patients and adults, ventilator-associated pneumonia (VAP) is the second most commonly occurring health care–associated infection. However, the rate of VAP in infants and neonates, especially those with underlying pulmonary disorders, is not as clear. The diagnosis of VAP is problematic because it occurs in ventilated infants who are likely to have other reasons for respiratory decompensation (e.g., atelectasis or heart failure secondary to a patent ductus arteriosus). Moreover, procedures commonly used to diagnose VAP in adults (e.g., bronchoscopy, lung biopsy, protected brush specimen, and bronchoalveolar lavage) are rarely used in the neonatal population.

The current definition used by the National Healthcare Safety Network (NHSN) of the Centers for Disease Control and Prevention (CDC) for the diagnosis of VAP in infants (this definition is not specific for neonates but is applied to them) requires new and persistent radiographic infiltrates and worsening gas exchange in infants who are ventilated for at least 48 hours and who exhibit at least three of the following criteria: temperature instability with no other recognized cause, leukopenia, change in the characteristics of respiratory secretions, respiratory distress, and bradycardia or tachycardia. The strictness of this definition makes it likely that the diagnosis of VAP is underreported in the United States. The 2013 data from the NHSN report an incidence of 1.0 to 1.1/1000 ventilator days for infants weighing <1000 g in level III neonatal intensive care units (NICUs). Rates of VAP/1000 ventilator days in the International Nosocomial Infection Control Consortium are eightfold higher.

In 2011, the CDC convened a working group consisting of critical care organizations, infection control groups, and epidemiology organizations to address the lack of precision in the diagnosis of VAP. As a result, a new term was developed, ventilator-associated event, or VAE. Within the VAE surveillance algorithm, three definition tiers were recognized. A ventilator-associated condition is defined by worsening oxygenation in a patient ventilated for at least 2 days (etiology not specified). The second tier is termed IVAC (infection-related ventilator-associated complication). Patients with an IVAC exhibit hypoxemia accompanied by general and objective evidence of inflammation/infection. To meet this definition antibiotics must be administered for at least 4 days. The third tier is possible or probable VAP, which requires additional laboratory evidence of white blood cells on Gram stain and/or positive cultures. The new definition is meant to improve epidemiologic surveillance, but it is clearly not easily adaptable to the NICU setting.

This chapter reviews the epidemiology, pathogenesis, diagnosis, prevention strategies, and treatment of VAP in the newborn infant.

Epidemiology

The most recent surveillance data as of this writing suggest that the incidence of VAP may be decreasing. Patrick et al. examined the incidence of health care–associated infections among critically ill children in 173 U.S. hospitals from 2007 to 2012 and noted the rate of VAP had decreased from 1.6 to 0.6/1000 ventilator days. An NHSN report indicates a pooled mean incidence rate varying from 0.3 to 1.6/1000 ventilator days ( Table 30-1 ) in U.S. level II/II NICUs. This represents a greater than 50% decrease from that reported in 2004. Cernada et al. note that regional economic development has a major influence on the incidence of VAP. Reported rates in developed countries range from 2.7 to 10.9/1000 ventilator days, whereas those in developing nations are as high as 37.2/1000 ventilator days. Variability in the diagnostic criteria used also contributes to the wide range of reported incidence rates in the literature.

Critically ill neonates are at high risk for hospital-acquired infections. Extrinsic factors such as inconsistent hand hygiene practices and overcrowding in NICUs contribute to the high incidence of health care–associated infections. Increased susceptibility of the neonatal host is multifactorial in nature and includes immaturity of the immune system (particularly infants weighing less than 1500 g) and the need for invasive monitoring, central catheters, prolonged intravenous alimentation, and mechanical ventilation. In a large prospective surveillance study, Van der Zwet et al. identified low birth weight (odds ratio (OR) 1.37; confidence interval (CI) 1.01, 1.85) and mechanical ventilation (OR 9.69; CI 4.60, 20.4) as risk factors for hospital-acquired pneumonia. Hentschel et al. observed a difference in the hospital-acquired pneumonia rates between intubated infants and those receiving nasal continuous positive airway pressure (NCPAP) (12.8 vs 1.8/1000 ventilator or NCPAP days).

A number of studies have examined specific risk factors for VAP among critically ill neonates. Yuan et al. conducted a retrospective cohort study in 259 patients who were ventilated more than 48 hours. By logistic regression analysis, the following variables independently predicted VAP: reintubation (OR 5.3; CI 2.0, 14.0), duration of mechanical ventilation (OR 4.8; CI 2.2, 10.4), treatment with opiates (OR 3.8; CI 1.8, 8.5), and frequency of endotracheal suctioning (OR 3.5; CI 1.6, 7.4). Low gestational age and/or birth weight have also been identified as risk factors for VAP. Data from the CDC’s NHSN (2006-2008) at 304 participating hospitals revealed a VAP rate of 2.36/1000 ventilator days among neonates weighing less than 750 g and 0.72/1000 ventilator days in those >2500 g. A prospective cohort study by Apisarnthanarak et al. reported a VAP incidence of 6.5/1000 ventilator days in infants less than 28 weeks’ gestational age, compared with 4/1000 ventilator days in those of ≥28 weeks’ gestation ( Fig. 30-1 ). The same authors found that after adjustment for the duration of endotracheal intubation, a preceding bloodstream infection (with an unrelated organism) was an independent risk factor for VAP (OR 3.5; CI 1.2, 12.3). This finding may reflect a subpopulation of infants with compromised immune function. Interestingly, the occurrence of VAP in this study was strongly associated with an increased likelihood of mortality (OR 3.0; CI 1.2, 12.3). A 2014 meta-analysis published by Tan et al. identified the following risk factors for the development of VAP: length of stay in NICU (OR 23.45), reintubation (OR 9.18), enteral feeding (OR 5.59), mechanical ventilation (OR 4.04), transfusion (OR 3.32), low birth weight (OR 3.16), prematurity (OR 2.66), parenteral nutrition (OR 2.30), bronchopulmonary dysplasia (OR 2.21), and tracheal intubation (OR 1.12). Additional risk factors for VAP have been described in the pediatric ICU literature (presence of a genetic syndrome, bronchoscopy, and the use of steroids, antibiotics, and histamine-type 2 receptor blockers). However, their association with VAP in the neonatal population remains unclear.

Distribution of ventilator-associated pneumonia ( VAP ) rates per 1000 ventilator days among neonates with gestational ages of 22 to 33 weeks.

(From Apisarnthanarak A, Holamann-Pazgal G, Hamvas A, et al. Pediatrics . 112:1283-1289, 2003.)

Pathogenesis

VAP occurs when bacterial, fungal, or viral pathogens gain entrance to the normally sterile lower respiratory tract. Only rarely does the organism gain entry to the lung through hematogenous dissemination or by bacterial translocation from the gastrointestinal tract. Pathogens responsible for VAP originate from exogenous sources (hands of health care workers, ventilator circuit, biofilm of endotracheal tube) or endogenous sources (colonized oropharyngeal, tracheal, and gastric secretions) ( Fig. 30-2 ). The organism gains entry to the respiratory tract by colonizing the endotracheal tube and the upper airway, by tracheal suctioning, or by direct aspiration of gastrointestinal contents. Cuffed endotracheal tubes are not generally used in the NICU. This practice provides easier access for microorganisms to the lower respiratory tract of neonates. Furthermore, microscopic aspiration may be more common than previously appreciated. Farharth et al. quantified pepsin, a marker of gastric contents, in tracheal aspirate samples from 45 ventilated newborn infants. Pepsin was detected in 92.8% of tracheal aspirate samples. The mean concentration of pepsin was significantly lower when the infants were unfed ( Fig. 30-3 ). Methylxanthines increased tracheal aspirate pepsin levels, and infants who developed bronchopulmonary dysplasia (BPD), or developed BPD or died before 36 weeks’ gestation, had significantly higher levels ( Fig. 30-4 ). Pepsin levels were also higher in infants who developed severe BPD versus those with moderate BPD ( Fig. 30-5 ).

Pathogenesis of bacterial hospital-associated and ventilator-associated pneumonia.

(Data from Garland JS. Neoreviews . 15:e225-e235, 2014.)

The levels of pepsin when infants were unfed and during the feed. Median concentration of pepsin was significantly lower in the unfed state compared with the concentration during feeding.

(From Farhath S, Aghai ZH, Nakhla T, et al. J Pediatr Gastroenterol Nutr . 43:336-341, 2006.)

Mean pepsin concentration from tracheal aspirates in infants with no bronchopulmonary dysplasia ( BPD ) and from infants with BPD, or with BPD or who died before 36 weeks’ gestation.

(From Farhath S, Zhaoping H, Nakhla T, et al. Pediatrics . 121:e253-e259, 2008.)

Mean pepsin concentration in preterm infants with moderate bronchopulmonary dysplasia ( BPD ) and severe BPD.

(From Farhath S, Zhaoping H, Nakhla T, et al. Pediatrics . 121:e253-e259, 2008.)

Microbiology

In NHSN surveys of adult and pediatric hospitals, Staphylococcus aureus and gram-negative organisms, primarily Pseudomonas and Klebsiella species, have remained the predominant pathogens over time. Apisarnthanarak et al. recovered gram-negative microorganisms from the respiratory secretions in 94% of VAP episodes. Staphylococcus aureus was recovered in about one-quarter of infants with VAP, and multiple organisms were recovered from the airway in 58% of episodes. Yuan et al. recovered gram-negative bacteria in the majority of infants with VAP. Table 30-2 shows the predominant pathogens, stratified by birth weight, associated with VAP among NICU patients from the CDC’s NHSN surveillance system from 2006 through 2008. Historically most health care–associated pneumonia is felt to be polymicrobial in nature. However, this may be reflective of sampling techniques, which are typically noninvasive and nonspecific in neonates, such as endotracheal tube suctioning. Cernada et al. found that when targeted techniques for sample collection such as bronchoalveolar lavage are performed, polymicrobial VAP represented merely 16.7% of episodes.

Diagnosis

The diagnosis of VAP is problematic. As noted above, the NHSN and CDC definitions are of value for epidemiologic studies, but of limited help for identifying infants with VAP. Complicating this issue is that most infants with suspected VAP have underlying lung disease that predisposes them to atelectasis and episodes of clinical deterioration. Additionally, it is not uncommon for general radiologists to report the presence of an infiltrate on a chest radiograph in an otherwise asymptomatic infant.

In adults with fever, pulmonary infiltrates, and clinical criteria for VAP, only 42% had a definitive diagnosis of pneumonia and 66% had a noninfectious etiology. Moreover, postmortem examinations demonstrated that only 35% of patients with a new or progressive infiltrate had histopathologic evidence of a pneumonic process. In ventilated adults, invasive techniques have been used to quantify the bacterial load as a way to distinguish infection from colonization. The airway of the newborn infant is colonized soon after intubation with a variety of potential pathogens, and a specimen taken from the endotracheal tube will not differentiate colonization from infection. Therefore, cultures of tracheal aspirates have low sensitivity, specificity, and positive predictive accuracy for the diagnosis of VAP. In a national survey from the United Kingdom regarding the diagnosis and management of VAP, only 57.8% obtained an endotracheal aspirate before starting empirical antibiotics. The American Thoracic Society and Infectious Diseases Society of America guidelines suggest using, in adult patients, a threshold of 10 ³ for quantitative culture from a protected specimen brush sample, 10 ⁴ for quantitative culture of bronchoalveolar lavage (BAL) fluid, and 10 ⁵ or 10 ⁶ for quantitative culture of tracheal aspirates. A meta-analysis of 23 studies in adult patients of quantitative BAL cultures and 18 studies of protected specimen brush cultures suggested the diagnostic value of these methods. However, a Cochrane review published in 2008 found no evidence that the use of quantitative cultures (vs qualitative cultures) reduced either mortality or time spent in the ICU. Furthermore, the use of quantitative cultures has been criticized because it may delay initiation of antibiotic therapy and there is the possibility of a false negative test secondary to preexisting use of antibiotics. Most recently, some pediatric intensive care units have begun to use a modified pulmonary infection score, mCPIS (using temperature, leukocyte count, chest radiography, pulmonary secretions, PaO ₂ /FiO ₂ (mm Hg), and cultures of nonbronchoscopic BAL), in the diagnosis of VAP. In a 2014 study of children who met the CDC criteria for VAP, a score of ≥6 had a sensitivity of 94%, a specificity of 50%, a positive predictive accuracy of 64%, and a negative predictive accuracy of 90%. Therefore, the score was best at identifying children who were unlikely to have VAP, but not nearly as good at identifying patients with VAP. There are no studies using the mCPIS in neonates.

A few studies in neonates have evaluated the value of nonbronchoscopic BAL in neonates. In this technique, a suction catheter is placed into the endotracheal tube until resistance is met and then a small amount of sterile saline is placed and then suctioned back. Köskal et al. obtained BAL specimens from 145 intubated newborn infants and did quantitative counts and smears for white blood cells on the BAL fluid. Using CDC criteria for VAP, 44 infants (30%) were diagnosed as infected and 90% of those infants ( n = 40) had positive BAL cultures. The percentage of neutrophils containing intracellular bacteria was significantly higher in infants with VAP (vs colonized, asymptomatic infants), as was the presence of leukocytes in BAL fluid (84% vs 26%). Quantitative cultures (greater than 10 ⁵ cfu/mL) also distinguished infants with VAP from colonized asymptomatic infants. The sensitivity and specificity of intracellular bacteria and quantitative cultures were 94% and 90% respectively. Cernada et al. used a nonbronchoscopic BAL technique to identify 18 episodes of VAP in 16 neonates. There were no complications with the invasive, blind BAL technique; however, four patients considered too unstable were excluded from the analysis. The presence of purulent tracheal aspirates may not be a reliable marker of VAP. Cordero et al. demonstrated that the majority (71%) of infants with purulent “tracheal aspirates” were asymptomatic. Furthermore, radiologically documented VAP occurred in 7% of very low birth-weight infants who never had a purulent tracheal aspirate and in 5% who did. Purulence on a tracheal smear was directly related to the duration of endotracheal intubation. In a 2015 study by Seligman et al., the absence of gram-positive cocci on samples obtained from an endotracheal aspirate was of value in excluding Staphylococcus as a potential pathogen. However, the Canadian Critical Care Trials group found no benefit to Gram stain of endotracheal aspirates and BAL specimens, with only a limited role in guiding therapy. Similarly, a meta-analysis published in 2012 noted a limited agreement between Gram stain and culture (mostly BAL specimens) and concluded that the Gram stain should not be used to narrow antiinfective therapy until culture results become available. In the only study of preterm infants with VAP the authors concluded that a Gram stain from a tracheal aspirate was useful in predicting classes of culturable microorganisms and for guiding appropriate initial antibiotic therapy. The use of biochemical markers (e.g., C-reactive protein and procalcitonin) has been investigated in adults, but has not been of much value. Traditional markers (neutrophil indices and acute-phase reactants) are of value in infants who are bacteremic, but that represents a minority of the infants with VAP.