Introduction

Healthcare-associated infections (HAIs) are associated with morbidity, mortality, and prolonged hospitalization, and represent a serious threat to patient safety. Hospitalized children are a particularly vulnerable population [1]. The incidence of HAI in adult and pediatric intensive care units (PICUs) is high. This is due to the many invasive procedures and frequent antibiotic use, which put the patients at risk for infection and promote the emergence of multidrug-resistant organisms [2]. The use of invasive devices in PICUs, such as central vascular lines and mechanical ventilation, is similar to adult intensive care and thus the burden of ventilator-associated pneumonia (VAP) and other HAIs is also similar [3]. In this review, we describe the epidemiology of VAP, summarize risk factors, and discuss effective prevention measures in PICUs and neonatal ICUs (NICUs).

Review

Literature search and selection strategy

A Medline search was performed for publications prior to 1 May 2014 using the following search (MeSH) terms: “pneumonia, ventilator associated” AND (child* OR neonat* OR infant* OR pediatr* OR paediatr*) and also pneumonia AND (nosocomial OR “healthcare-associated” OR “healthcare associated” OR “health care associated”) AND (ventilat* OR intubat* OR respirat*) AND (child* OR neonat* OR infant* OR pediatr* OR paediatr*). Cross-referencing from retrieved publications was used to complete the search, including manual searches of cited references and relevant abstracts. Publications were eligible to be analyzed if they addressed VAP in any inpatient pediatric population. A total of 443 titles and abstracts were screened; 95 were retained for discussion in this review.

Definitions

A uniform definition of VAP needs to have the capacity to be relevant for clinical trials, while balancing the risks of experimental therapy and sampling procedures with potential benefits for study patients [4]. If the definition of VAP is already controversial for adults, it is even more challenging for children, in particular for ventilated neonates. The starting point of the recent United States (US) Centers for Disease Prevention and Control (CDC) definitions for adults is a ventilator-associated complication (VAC), which is further narrowed towards infectious VAC and then towards possible or probable VAP, according to additional diagnostics [5]. However, It is not clear whether this algorithm can be applied to children in different age groups and, thus, the conventional CDC definitions of hospital-acquired pneumonia for children and neonates remain valid for the time being [6]. These definitions do not specify between “ventilated” or “non-ventilated” and the use of the term “VAP” depends on the time on ventilation (48 h or longer). The German national nosocomial infection surveillance system (Krankenhaus Infektions Surveillance System [KISS]) offers a definition for very low birth weight infants in their “Neo-KISS” module [7]. A Dutch study group established their own definition for VAP in neonates, which are more inclusive than the CDC definitions [8]. Table 1 summarizes the definitions of hospital-acquired pneumonia by stratifying age groups into neonates, infants (≤1 year), and children (>1 year to ≤ 16 years). All definitions combine clinical and radiologic signs. In addition, the CDC and the European Centre for Disease Prevention and Control (ECDC) definitions further distinguish between definite, probable, and possible healthcare-associated pneumonia, based on microbiologic findings (Table 2) [9]. Clinical and radiologic findings lack sensitivity and specificity. However, tracheal aspirate cultures have also low sensitivity (31-69%) and specificity (55-100%). A positive tracheal culture alone does not discriminate between bacterial colonization and respiratory infection. Bronchoalveolar lavage (BAL) provides better results, but the range of sensitivity (11-90%) and specificity (43-100%) is large.

Table 1 Case definitions of hospital-acquired pneumonia in children stratified by different age groups
Full size table
Table 2 Classification of hospital-acquired pneumonia in children based on microbiological results
Full size table

Clinical criteria

Clinical criteria for healthcare-associated pneumonia include fever, leukocytosis or leucopoenia, purulent secretions, new or worsening cough, dyspnoea, tachypnoea, crackles or bronchial breath sounds, and worsening gas exchange. These criteria are nonspecific and their sensitivity and specificity relative to the underlying pathology is poor [2]. Clinical findings must be combined with radiologic and microbiologic findings. In a study of 70 children with VAP, the modified clinical pulmonary infection score (mCPIS) of six or higher had a sensitivity of 94%, a specificity of 50%, a positive predictive value of 64%, a negative predictive value of 90%, and positive and likelihood ratios of 1.9 and 0.1, respectively [10].

Radiologic criteria

Radiologic criteria include the presence of new or progressive pulmonary infiltrates, adhesions or fluid in lobar fissures/pleura, cavitations, air bronchograms, or pneumatoceles on chest x-rays. The presence of air bronchograms has a higher sensitivity (58–83%) than “evolving infiltrates” (50–78%) [2]. Sequential chest x-rays (days -3, 0, 2, 7) help to confirm healthcare-associated pneumonia in complex cases, such as children with underlying cardiac or pulmonary disease. Onset and progression of pneumonia in imaging is fast, but improvement takes time.

Microbiologic criteria

Respiratory cultures are obtained by tracheal aspirates, bronchoalveolar lavage (BAL), non-bronchoscopic BAL, or protected brush specimens (PBS) [10]. Thresholds are summarized in Table 2.

Epidemiology

Healthcare-associated pneumonia was the most common HAI in five studies [11–15], and second only to bacteremia in another two reports [16, 17]. The range of VAP incidence density rates in both children and neonates is large. Rates as low as 1/1000 ventilator-days and as high as 63/1000 ventilator-days have been reported (Table 3). The incidence follows a geographical distribution and depends on the type of hospital and the country income level. A surveillance study from the International Nosocomial Infection Control Consortium (INICC) identified higher VAP rates in academic compared to non-academic hospitals [18]. The same study reported higher rates in lower-middle-income compared to upper-middle-income countries. Extreme PICU rates have been reported from India (36.2%) [19] and Egypt (31.8/1000 ventilator-days) [20]. Surveys in the USA and Germany found consistently lower rates (Table 3) [21–23]. However, high rates were reported also by high-income countries. A European multicenter study found that 23.6% of children admitted to a PICU developed VAP [24]. An Italian study identified 6.6% children with VAP among 451 on mechanical ventilation [25], and a mixed PICU in Australia identified 6.7% children with VAP among 269 on mechanical ventilation [26].

Table 3 Incidence densities and proportions of ventilator-associated pneumonia in pediatric settings
Full size table

VAP is also common in the NICU and proportions between 6.8% and 57.0% of HAIs have been reported [34, 60–66]. A Spanish study identified VAP in 9.1% of 198 neonates on mechanical ventilation [67]. In a Taiwanese NICU, 11.4% of 528 neonates had one or more HAIs, with VAP contributing to 18.6% [68]. An INICC survey summarizing results from 30 NICUs in 15 countries reported significantly higher VAP rates in academic compared to non-academic institutions [69]. VAP incidence densities in an Iranian and Turkish NICU were 13.8/1000 and 11.6/1000 ventilator-days, respectively [27, 35]. A higher incidence was reported in another Iranian study with 42% of 38 neonates on mechanical ventilation [70]. Table 4 summarizes birth weight-dependent numbers from different studies [8, 21–23, 42–44, 49, 71].

Table 4 Incidence densities of ventilator-associated pneumonia in neonatal intensive care units stratified by birth weight
Full size table

Several studies from the USA, Italy, and Iran found that VAP prolonged mechanical ventilation by approximately 8–12 days [25, 70, 72, 73], and this may even be as high as 56 days in extremely preterm neonates [46]. Prolonged length of stay was the main driver of attributable costs of up to US$ 1040 in Iran and US$ 51,157 in the USA [70, 73]. There are no data on the attributable mortality of VAP. The mortality of HAI in the PICU is estimated to range between 5-14% [27, 44], to which VAP may significantly contribute (P = 0.04) [25].

Risk factors

Ventilation was the most important identified risk for HAI in a prevalence study of 21 hospitals in Mexico (odds ratio [OR], 2.3; 95% confidence interval [CI], 1.2-4.1) [74]. Reintubation (OR, 2.7; CI, 1.2-6.2) and transport out of the PICU (OR, 8.9; CI, 3.8-20.7) were significant risk factors identified in a US PICU [74]. Other extrinsic risk factors include prior antibiotic therapy (OR, 2.89; CI, 1.41-5.94), bronchoscopy (OR, 4.48; CI, 2.31-8.71), immunosuppressive drugs (OR, 1.87; CI, 1.07-3.27), and the use of enteral feeding (OR, 8.78; CI, 2.13-36.20) [75–77]. A number of intrinsic factors predisposing for VAP have been reported, such as young age (<12 months) [75, 78], subglottic or tracheal stenosis (P = 0.02), trauma (P = 0.02), tracheostomy (P = 0.04) [72], gastroesophageal reflux [79], immunodeficiency [28], neuromuscular blockade [28, 75, 80], genetic syndromes (OR, 2.04; CI, 1.08-3.86) [46, 76], and gender (female: OR, 10.32; CI, 2.9-37.2) [77].

In neonates, the main risk factors are low birth weight (hazard ratio [HR], 1.37; CI, 1.0-1.9]) and mechanical ventilation (HR, 9.7; CI, 4.6-20.4) [8]. Time of mechanical ventilation was a main factor in Spanish (OR, 1.1; CI,1.1-1.2) [67], Chinese (OR, 4.8; CI, 2.2-10.4) [30], and Iranian studies (P <0.001) [70]. Reintubation, absence of tube feeding, and absence of stress ulcer prophylaxis were risk factors in Australia [26]. In an Italian study, reintubation (P < 0.001), tracheostomy (P = 0.04), and enteral feeding (P = 0.02) were associated with VAP [25]. Risk factors for VAP are summarized in Table 5.

Table 5 Risk factors for ventilator-associated pneumonia in pediatric and neonatal settings
Full size table

Microorganisms

The microorganism type and antibiotic susceptibility are variable according to the geographical region (Figure 1). Gram-negative pathogens predominate, but their contribution is exceptionally high in Asia. Overall, the most common pathogens are Pseudomonas aeruginosa, Acinetobacter baumannii, and Enterobacteriaceae. In Europe and North America Staphylococcus aureus predominate [8, 77, 81]. In Asia, most pathogens are multidrug-resistant [82–84]. A Greek group reported 65 children with 71 infections (20 VAP) due to carbapenem-resistant Gram-negative pathogens [85]. Isolates included Pseudomonas spp. (41.1%), Acinetobacter spp. (39.7%), and Klebsiella spp. (19.2%).

Figure 1
figure 1

Geographical distribution of pathogens causing ventilator-associated pneumonia in children.

Full size image

Prevention

Many interventions in different combinations have been shown to play a role in VAP prevention: hand hygiene, preferably with alcohol-based handrub; glove and gown use for endotracheal tube manipulation; backrest elevation of 30° to 45°; oral care with chlorhexidine; stress ulcer prophylaxis; cuff pressure maintenance; use of orogastric tubes; avoidance of gastric overdistension; and elimination of nonessential tracheal suction [86]. Oral care with chlorhexidine compared to placebo in 96 children on mechanical ventilation was not effective in reducing VAP in a Brazilian study [87]. Similar results were reported in a placebo-controlled study with high VAP rates in North India [88] and a randomized trial among children undergoing cardiac surgery in Brazil [89]. Gastroesophageal reflux is a constant incident in mechanically- ventilated children, with alkaline reflux more common than acidic reflux [79]. Thus, stress ulcer prophylaxis is rather unlikely to prevent VAP and, consequently, neither sucralfate nor ranitidine were effective in VAP prevention in a small study [90]. Two studies showed that VAP rates are lower in neonates undergoing nasal continuous positive airway pressure compared to the use of mechanical ventilation [21, 36].

A prevention bundle reduced VAP from 7.8/1000 to 0.5/1000 ventilator-days (P < 0.001) in a US PICU with an estimated economy of 400 hospital-days and cost-savings of US$ 2,353,222 [73]. In another PICU, a bundle adapted to local needs by plan-do-study-act cycles reduced VAP rates in a similar manner [72]. The bundle addressed handling of ventilator circuits and oral suctioning, hand hygiene, regular oral care with chlorhexidine, and backrest elevation. By applying a multimodal intervention, three PICUs reduced the incidence of hospital-acquired pneumonia from 5.6 per 100 patients at baseline to 1.9 in the intervention (P = 0.016) [91]. An educational program targeting resident physicians and nurses in a PICU of a lower-middle-income country resulted in a non-significant VAP reduction of 28% (P = 0.21) [92]. A quality improvement intervention targeting hand hygiene and establishing quality practices decreased VAP from 28.3/1000 to 10.6/1000 ventilator-days (P = 0.005), which was sustainable over a long-term, follow-up period [93]. In a before-after study in eight PICUs of five developing countries, the efficacy of a multidimensional infection control program including education, outcome surveillance, process surveillance, and feedback on VAP rates and performance reduced VAP from 11.7/1000 to 8.1/1000 ventilator-days (P = 0.02) [94]. The institution of a purpose-designed bundle by a nurse-led VAP surveillance program addressed backrest elevation; oral care using chlorhexidine; clean suctioning practice; ranitidine for all children not on full feeds; and four-hourly documentation [95]. After bundle implementation, no VAP was recorded over a 12-month period. The baseline ventilator-associated tracheobronchitis rate of 3.9/1000 ventilator-days was reduced to 1.8/1000 (P = 0.04) by implementing a multidisciplinary quality improvement initiative in another US PICU [96].

A strategy combining care practices with empowering the bedside nurse to lead bundle implementation in a NICU encouraged personal ownership and compliance with the bundle and finally reduced VAP by 31%, resulting in savings of 72 hospital-days and US$ 300,000 [97]. The INICC multidimensional infection control program was associated with significant reductions of VAP rates in the NICUs of 15 cities from 10 developing countries [98]. VAP rates at baseline and intervention were 17.8/1000 and 12.0/1000 ventilator-days, respectively [98]. Of 491 patients receiving mechanical ventilation in a Chinese NICU, the rate of VAP decreased from 48.8/1000 to 25.7/1000 ventilator-days and further diminished to 18.5/1000 after hospital relocation and establishing a bundle of comprehensive preventive measures (P < 0.001) [99].

Conclusion

VAP is common in mechanically-ventilated children with a wide variation of incidence density rates across geographical regions. Surveillance definitions are challenging in pediatric settings because the combination of clinical and radiologic signs leaves too much room for interpretation. This is particularly important in neonates, where CDC and INICC guidelines, and the German KISS program follows mainly the rationale of the definitions for older children. Gram-negative pathogens are the most common microorganisms, particularly A. baumannii and P. aeruginosa. However, there is a geographic variation with Gram-positive organisms more frequently observed in high-income compared to low- and middle-income countries. Similar to the evidence base of adult settings, a number of studies reported effective VAP prevention strategies. Successful programs combined multiple interventions, such as hand hygiene, glove and gown use for endotracheal tube manipulation, backrest elevation, oral care with chlorhexidine, stress ulcer prophylaxis, cuff pressure maintenance where appropriate, use of orogastric tubes, avoidance of gastric overdistension, and elimination of nonessential tracheal suction. When applied as a multimodal strategy by an interdisciplinary team, these interventions are most likely to be successful among neonates, infants, and children, and have proven effectiveness in high-, as well as in low- and middle-income countries.