Microbial colonization of the lower airways a ft er insertion of a cu ff ed endotracheal tube in pediatric patients

Background. Ventilator-associated pneumonia (VAP) still remains a common device-associated hospital acquired infection in pediatric and adult intensive care units. The aim of our study was to determine ways of microbial transmission to the lower airways in intubated patients admitted to a single tertiary-care pediatric intensive care unit. Methods. This was a prospective observational study. A total of 284 sample sets (oropharyngeal swabs, swabs from the lumen of the proximal tip of an endotracheal tube, and bronchoalveolar lavage samples) were collected from 62 consecutive pediatric patients intubated for > 24 hours. Pulsed-field gel electrophoresis was performed on all isolated pathogens, which were later identified by MALDI biotyper (MALDI-TOF mass spectrometry). Results. Overall colonization rates were high and did not differ significantly at different time points in the oropharynx (75%–100%) and the lower airways (50%– 76.5%). The endotracheal tube was colonized at lower rates: on day 1–3 (28.8%), on day 4–6 (52.7%), on day 7–9 (61.8%) and on day 10-12 (52.9%) (P < 0.001). A total of 191 matched sample sets from the lower airways and at least one site above were collected from 46 (74.2%) patients. In the oropharynx-lower airways group, Candida spp. (76.9%) and upper airway bacteria (63.2%); in the endotracheal tubelower airway group, S. aureus (15.7%) and upper airway bacteria (21.1%); in the oropharynx-endotracheal tube-lower airway group, Enterobacteriaceae (70.8%) prevailed (P < 0.001). The mean survival (entrance) time to lower airways for the Acinetobacter / Pseudomonas / Stenotrophomonas group was 8.28 ± 0.81 days; for the Enterobacteriaceae group, 5.63 ± 0.41; and for Candida spp. group, 3.00 ± 0.82 days (P < 0.005). Conclusions. Oropharyngeal contamination of the lower airways is the most important route of colonization. Different pathogens enter the lower airways at different time intervals from the insertion of an endotracheal tube.


INTRODUCTION
Ventilator-associated pneumonia (VAP) still remains a common device-associated hospital acquired infection in pediatric and adult intensive care units (ICUs).During the last decade, a remarkable decline in the VAP incidence has been documented in economically developed countries; however, the incidence in developing countries is decreasing as well (1).Despite signifi-cant improvements, there is still a lot of controversy on how to achieve minimal incidence, even knowing the pathogenesis of colonization of the tracheal bronchial tree (2)(3)(4).Sometimes the implementation of VAP prevention bundles in the pediatric population is problematic.Likewise, elevation of the head by 30o-45o angle for an infant is even impossible, and deep vein thrombosis prophylaxis in the pediatric population does not work in terms of VAP prevention.Education and training of nursing staff and implementation of adjusted "pediatric" bundles have a significant impact on reduction in the VAP incidence (5).However, having a minimal incidence, there is still a question whether we can do something more.Therefore, the aim of our study was to determine ways of bacterial transmission to the lower airways (LA) and associated risk factors in intubated patients after the implementation of VAP prevention bundle in a tertiary-care pediatric intensive care unit (PICU).

Study population and data collection
This prospective observational study was conducted in a single tertiary-care PICU (8 beds, around 50 patients intubated for > 24 hours annually) of the Hospital of Lithuanian University of Health Sciences Kauno Klinikos from February 2012 to June 2013.All consecutive patients aged from 1 month to 18 years and intubated for >24 hours were eligible for inclusion into the study.Exclusion criteria were multiple congenital abnormalities, chronic infection (e.g.cystic fibrosis), and mental disorders.Patients exited the study at time of extubation, tracheostomy, and death.None refused to participate in the study.
The study protocol was approved by Kaunas Regional Biomedical Research Ethics Committee and written informed consent was obtained on January 9, 2012 (registration No. 8/2012).According to the protocol, microbiological specimens were collected from three sites: oropharynx (OPX) (swabs), lumen of the proximal tip of an endotracheal tube (ETT) (swabs), and LA (bronchoalveolar lavage [BAL] aspirate).The blind BAL sample was taken by inserting a single-lumen regular endotracheal suction catheter of appropriate size through an orally inserted endotracheal tube.Patients were preoxigenated with 100% oxygen.Then a suction catheter was inserted to a wedge position, instilling 1 mL/kg saline for <20 kg and 20 mL saline for >20 kg patient, and then immediately withdrawing the fluid (6)(7)(8).Any count of pathogens (cfu/mL) in LA was considered as positive, and BAL aspirate with ≥105 cfu/mL was defined as heavy colonization (9).The oropharynx was chosen because its crossroad position for the nasopharynx, oral cavity, and hypopharynx and potential endogenous source, and proximal tip of the ETT because of its potential exogenous source of contamination and colonization of the tracheal bronchial tree (1;3;4;10) .Sets of three microbiological samples were collected on the first day and every third consecutive day until day 18 and later every fifth day until day 28.Extra sample sets were taken in the following seven circumstances: 1) before extubation, 2) after bronchoscopy, 3) gastrointestinal endoscopy and after patient transportation for investigation such as 4) computed tomography (CT), 5) magnetic resonance imaging (MRI), and 6) other or 7) surgery outside the PICU area.A total of 284 sample sets (284 × 3 samples) were collected for 62 patients.

Microbiologic methods
Study samples were coded and sent to the microbiology laboratory for identification.The samples were inoculated directly onto 5% sheep blood agar (BBL, USA), chocolate agar (BBL, USA), and MacConkey agar plates (Oxoid, UK).Sheep blood and chocolate agar plates were incubated at 35°C in an atmosphere containing 5% CO2 and MacConkey agar plates, at 35°C for 18-24 hours.If a culture was negative on the first observation, sheep blood and chocolate agar plates were re-examined after the second 24-hour incubation.Pulsedfield gel electrophoresis (PFGE) was performed on all isolated pathogens by using the modified procedures published by Barth and Pitt as well as Grothues and Tümmler (10;11).Isolated pathogens, potentially causing VAP, and others (S. pneumoniae, H. influenzae, S. aureus, E. coli, K. pneumoniae, Enterobacter cloacae, Enterococcus spp., P. aeruginosa, Acinetobacter spp., Stenotrophomonas maltophilia, Candida spp., beta-hemolytic streptococci) were identified using matrix-assisted laser desorption ionization-time of flight (MALDI-TOF, Bruker) mass spectrometry (MS).Disk diffusion susceptibility testing was performed, and zone diameters of inhibition were interpreted according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) recommendations.

VAP definition and prevention in the PICU
VAP was identified during the routine surveillance procedure by using a combination of imaging, clinical, and laboratory criteria (pathogen concentration of ≥104 cfu/mL in a blind BAL aspirate).Ventilator-associated tracheobronchitis (VAT) diagnosis was based on the absence of clinical and radiographic evidence of pneumonia and the following criteria: positive culture obtained by deep tracheal aspirate and 2 signs or symptoms with no recognizable cause (fever [>38.5°C],cough, new or increased sputum production, rhonchi, or wheezing) (12;13).In the period from 2006 to 2007, when the multimodal intervention was designed (education of the PICU staff about VAP prevention and correction of daily care of a patient according to the evidence-based recommendations, feedback communication with the PICU staff in the postintervention period and implementation of new daily care protocols), there was a sharp decrease in the VAP incidence (21.8 versus 8.8 per 1000 ventilator-days) (14).Later, the implementation of VAP prevention bundle (semi-recumbent 30°-45° position of the head, daily evaluation of readiness to wean, comprehensive oral care protocol, periodical check for gastric overdistention, ETT cuff pressure between 20 and 30 cm of water, stress ulcer prophylaxis, periodic drainage and absence of tubing condensate and annual reporting of surveillance results in the PICU) led to a further decline in the VAP incidence curve (2.0 per 1000 ventilator-days in 2012; surveillance data were submitted to the INICC registry) (1).

Statistical analysis
First of all, all the microbiological samples were analyzed and later, matched sample sets.A matched sample set was defined if there was a growth of matching pathogens from the LA and at least one of the other upper airway sites (OPX or ETT) at the moment of sample collection.For statistical analysis, all the pathogens were divided into five groups: Acinetobacter, Pseudomonas and Stenotrophomonas (APS); Candida spp.; Enterobacteriaceae; S. aureus; and bacteria of the upper airway (UA) with the oral cavity.The descriptive analysis methods were used (mean, standard deviation, median, interquartile range (IQR), proportion).Pediatric index of mortality (PIM2) was used as independent descriptive variable for assessment of status of a patient at the time of his admission to PICU or the first face-to-face contact with PICU physician (15).The crosstabulation method and the chi-square test with an estimate of adjusted residual (≥ |2.0|) were used to compare categorical variables among different subgroups.In addition, matched samples were tested using Kaplan-Meier survival analysis and the Breslow test for pairwise comparisons.Survival time was defined as the time period between ETT placement and entry of a matching pathogen from the OPX or the ETT to the LA.For all analyses, P < 0.05 (2-tailed) was considered statistically significant.
The occurrences of matched samples were not associated with eight circumstances (one planned and seven extra) for sample collection (χ2 = 22.11, df = 14, P > 0.05).The overall distribution of the first matched cases in matched sample groups by eight reasons was also homogeneous (χ2 = 20.77,df = 12, P = 0.05), however, the occurrence of the first OPX-LA case was significantly associated with transportation for a CT scan outside the PICU area (n = 6, 100%, AR = 2.1).
Overall survival analysis of matched samples showed that the mean survival time was the longest for APS (8.28 (4;12;18).Risk factors for VAP are well described and preventive bundles have been proposed based on these data (20)(21)(22).However, leakage near the endotracheal cuff remains the main bridge between the OPX and the LA, causing colonization, VAT, and VAP.Prevention of leakage still remains a big challenge in the care of a critically ill patient (5).In the PICU, cuff pressure measurements are crucial (5;23;24), because an application of subglotic secretion drainage (SSD) is limited due to smaller tube sizes used (ETT tubes <5 mm in diameter with a SSD port are not produced), despite evidence about SSD benefit in VAP prevention.The clinical benefit of ETT with taper-shaped cuffs and silver-coated ETT still needs to be proved (5;25).Interestingly, in our study we have not detected highly resistant or multidrug-resistant pathogens in the LA, and an overall antimicrobial susceptibility picture was even better than we expected.Limitation of our study is that we have not investigated the influence of antimicrobials on the colonization of OPX, ETT, and LA.However, we are sure that our colonization data cannot be biased by antimicrobials, because OPX and LA colonization rates remained steady over a study time and antimicrobial susceptibility rates were high and mortality rate was low, indicating efficacy of the antimicrobial stewardship program in the PICU.
In conclusion, oropharyngeal contamination of the lower airways is the most important route of colonization.Different pathogens enter the lower airways at different time intervals from the moment of endotracheal tube placement.Reduction of leakage near the endotracheal cuff remains challenging in the PICU, and meticulous care of the oropharynx and endotracheal cuff pressure monitoring are advised before transportation of an intubated patient.

Table 1 .
Clinical characteristics of patients

Table 2 .
Colonization rates by site and time period

Table 3 .
Antimicrobial susceptibility of isolates in matched sample sets

Table 4 .
Colonization rates by pathogen groups and matched sample groups

Table 5 .
Colonization rates by time period and matched sample groups AR, adjusted residual.Figure.Survival analysis of matched samples by pathogen groups