Biofilm formation on three different endotracheal tubes: a prospective clinical trial

We included patients 18 years and older who were admitted to our intensive care unit (ICU) and were expected to require invasive mechanical ventilation for at least 24 h. Patients were allowed to participate only once and were included during six separate time periods from February 2014 to April 2017. Depending on the period, patients were intubated on clinical indications with one of the three different types of ETTs tested in our study. Each type of ETT was used in two of the six periods. The use of the different ETTs according to study period rather than by randomization was done for logistical reasons. Patients in study periods one and four received an uncoated PVC ETT (Oral/Nasal Endotracheal Tube, Mallinckrodt™, Medtronic, Dublin, Ireland), which is standard at our hospital; patients in periods two and six received an SC ETT (Siliconized PVC, Oral/Nasal Soft Seal® Cuffed Tracheal Tube, Portex™, Smith’s Medical, Kista, Sweden); and patients in periods three and five received an NbMC PVC ETT with a thin noble metal alloy coating consisting of gold, silver, and palladium (Bactiguard Infection Protection Endotracheal Tube, Bactiguard®, Tullinge, Sweden). Each of the six study periods was planned to last for about 3 months, and the goal was to include a minimum of 15 patients in each period. Bundles of maneuvers to prevent VAP (e.g., placing patients in a semi-recumbent position and controlling cuff pressure every 8 h to ensure a pressure of 20–30 mmHg) are standard at our facility. With the exception of use of different ETT materials during the study periods, all patients received standard intensive care according to their diagnosis and the clinical decisions of the responsible physicians. Furthermore, all patients were monitored daily for culture response, and a specialist in infection medicine who was blinded to the study evaluated antibiotic treatment daily based on culture response and proven experience.

Samples for surveillance cultures (i.e., oropharyngeal swabs and endotracheal aspirates) were collected from all included patients on days 1, 2, 3, 5, 7, 14, and 21 and thereafter once a week. This was also done on the day of extubation if not previously scheduled. All oropharyngeal swabs and endotracheal aspirates were processed in the same manner at the Department of Clinical Microbiology by standardized, extended microbiological procedures [18]. For details, please see the online supplementary material.

Age, sex, body mass index (BMI), days with invasive ventilation, antibiotic and/or proton pump inhibitor (PPI) use during mechanical ventilation, Simplified Acute Physiology Score 3 (SAPS 3), and main diagnosis on ICU admission were documented for all patients. Data on the occurrence of VAP and on microorganisms isolated in surveillance cultures and ETT biofilms were collected for all patients.

Patients were extubated at the discretion of the treating physician or in the case of a patient’s death. After extubation, the ETTs were collected, avoiding contamination from other than oropharyngeal flora, and rinsed (inside and outside) with 1 L of sterile saline to eliminate excess mucus. Thereafter, the distal tip of the ETT was divided into four pieces for scanning electron microscopy (SEM; two pieces) and microbial cultures (two pieces). Finally, the ETT tip was cut in a cross-sectional manner 1.5 cm above the distal tip. Pieces of the ETT for microbial cultures were sonicated in phosphate-buffered saline (PBS) at 47 kHz for 90 s to dislodge biofilm microorganisms. The solution was then homogenized by vortex mixing and subsequently cultured using the same procedures as applied for the oropharyngeal swabs and endotracheal aspirates. An unpublished pilot study comparing different methods for processing of the ETTs had indicated that the method outlined above was optimal for removing the biofilm and for dislodging the microorganisms in the film before culturing. For SEM, pieces of an ETT were fixed in a solution of 4% formaldehyde diluted with PBS at room temperature for 30 min and then dehydrated with crescent ethanol concentrations, air-dried overnight, and sputter-coated with a thin (15-nm) Au/Pd film (Gatan PECS Mod 682, Gatan, Inc., Pleasanton, CA, USA) to prevent charging during SEM analysis.

The inner and outer surfaces of the ETTs were examined by SEM (Zeiss Supra 40VP, Carl Zeiss Microscopy GmbH, Jena, Germany). This analysis was performed using a secondary electron detector set at a working distance of 8–10 mm and electron acceleration of 3.68–4.05 kV. Low magnification (100× to 1000×) was used to rank biofilm coverage as follows: 0, no biofilm; 1, scarce coverage of < 10%; 2, clusters with 10–70% coverage; 3, confluent film with > 70% coverage. High magnification (10,000× to 50,000×) was used to evaluate biofilm density (0, no biofilm; 1, low/very porous; 2 medium; 3, high/compact) and level of thickness (0, no biofilm; 1, thin 0.1–1.0 μm; 2, medium 1.1–7 μm; 3, thick > 7 μm). Film thickness was estimated as the difference in focus distance between the outer surface of the biofilm and the ETT surface. When measuring the thickness, the mean was calculated based on multiple representative points in the sample. The final grade of the biofilm was then calculated by adding together the scores from coverage, density, and thickness to give a total score of 0 to 9. A high/advanced biofilm grade was defined as having a score of ≥ 7. The grading system is summarized in Table 1. Grading of the biofilm was performed by a researcher at RISE who was blinded to all patient information including type of ETT analyzed.

Surveillance cultures and ETT tip cultures with growth of at least two species of bacteria commonly found in the oral cavity were classified as having normal flora (negative cultures). Species not normally found in the oral cavity (e.g., pathogens, gut flora, or overgrowth of normal oral flora) were classified as abnormal flora (positive cultures). All culture results were reviewed by a microbiologist to ensure correct classification. The diagnosis of VAP was made by two independent physicians evaluating the cases based on clinical and radiological examinations. VAP was defined by the following: (1) new or progressive lobar infiltrate > 48 h after intubation; (2) two or more of the minor criteria fever, leukocytosis/leukopenia, and purulent respiratory secretions; and (3) microbiological confirmation in endotracheal aspirate [19]. Colonization with common VAP pathogens included cultures with Enterococcus faecium, E. faecalis, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, Streptococcus pneumoniae, Haemophilus influenzae, and other Enterobacteriaceae species.

VAP relapse was defined as previously reported [12]. Microbial persistence was defined as persistence of the causative agent of VAP in at least two surveillance cultures despite 48 h of appropriate antibiotic therapy [20].

Results were expressed as median (interquartile range [q1–q3]) for continuous variables and as numbers (percentages) for dichotomous variables. A p value of < 0.05 was considered significant, and all statistical tests were two-tailed. Fisher’s exact test for categorical variables was used for groupwise comparison of ETT cultures. Multivariable logistic regression analysis was applied to identify independent factors associated with the formation of high-grade biofilm on the ETT, that is, a score of ≥ 7 in the scoring system described above, and this cut-off was chosen based on data from an earlier study [11]. Univariable logistic regression analyses were conducted to evaluate patients’ characteristics that could be associated with the development of VAP. Multivariable regression analysis of the outcome VAP was not performed due to too few cases of the condition (n = 12). The Hosmer-Lemeshow test was used to determine goodness of fit for all logistic regression analyses. Clinical studies comparing the degree of biofilm formation on ETTs using the same scoring system applied in our investigation are lacking, and therefore, we based the present power calculations on previously published laboratory studies [21, 22]. Given a power of 0.8 and an alpha level of 0.05, a sample size of 27 was needed in each group. We carried out all analyses with SPSS 26 (SPSS Inc., Chicago, IL, USA).

Biofilm was present on the ETTs of 97% (n = 103) of the patients and could be seen on the surface of the tubes after 24 h of intubation. The biofilm formation varied from low/porous (Fig. 2a) to confluent/abundant biofilm matrices (Fig. 2b). Colonies of different microorganisms embedded in the biofilm matrix were often recognized by SEM (Fig. 2c).

When using multivariable logistic regression analysis to assess possible predictors of high-grade biofilm formation (score ≥ 7) on the ETTs, the SC ETTs and the NbMC ETTs were independently associated with reduced high-grade biofilm formation compared to the uncoated PVC tubes (odds ratio [OR] 0.18, 95% confidence interval [CI] 0.06–0.59, p = 0.005 for the former, and OR 0.34, 95% CI 0.13–0.93, p = 0.036 for the latter). There was no significant difference between SC and NbMC ETTs (OR 0.54; 95% CI 0.17–1.65, p = 0.278). Age, sex, days with invasive ventilation, and colonization with common VAP pathogens in surveillance cultures did not predict higher biofilm formation in univariable or multivariable analyses (Table 3).

Surveillance cultures were obtained for all patients, but cultures that were missing according to the predefined culture scheme represented 5% (n = 41) of all planned cultures. For all oropharyngeal (n = 376) and endotracheal (n = 381) cultures, 66% (n = 248) and 64% (n = 242), respectively, turned positive and about 22% of those were noted as polymicrobial. A majority of the patients developed abnormal oropharyngeal flora (82%) and became colonized in the lower airways (79%) during invasive ventilation. Colonization with common VAP pathogens was found in oropharyngeal cultures from 33% (n = 35) of the patients and in endotracheal cultures from 27% (n = 29). In colonized patients, the median time from intubation to culture positivity was 1 day for both oropharyngeal and endotracheal cultures. In patients colonized in oropharyngeal or in endotracheal cultures (all positive cultures), the same microorganism was found on the ETT in 48% and 49% of the cases, respectively. In patients colonized with common VAP pathogens, the same microorganism was found on the ETT in 29% (10 out of 35) of the oropharyngeal cultures and 38% (11 out of 29) of the endotracheal cultures.

Microorganisms isolated from surveillance cultures in the three study groups are listed in Table 4. The surveillance cultures turned positive for several different microorganisms, most often Candida albicans and other candida species. Among bacteria, Enterococcus faecalis, E. faecium, Staphylococcus aureus, Klebsiella species, Stenotrophomonas maltophilia, Pseudomonas aeruginosa, and Haemophilus influenzae occurred most frequently (Table 4).

At extubation, 99% (n = 105) of the ETTs were cultured and 47% (n = 49) turned positive. No significant difference in positive ETT cultures was observed between the groups (uncoated PVC 41% [n = 14], SC PVC 45% [n = 13], and NbMC PVC 52% [n = 22]; p = 0.61). An ETT culture positive for common VAP pathogens at extubation was predicted by colonization with common VAP pathogens in oropharyngeal cultures (OR 17.6 [95% CI 3.8–81.6], p < 0.001) and in endotracheal cultures (OR 10.3 [95% CI 3.1–34.2], p < 0.001). If the ETT tip turned positive, the microorganism was found in oropharyngeal cultures in 86% (n = 42) of the patients and in endotracheal cultures in 88% (n = 43). The microorganisms found on the ETTs at points of extubation could be detected in the first oropharyngeal culture in 60% (n = 29) of the patients and in the first endotracheal cultures in 58% (n = 28). Candida albicans (39%) and other candida species (11.5%) were the microorganisms that occurred most frequently on the ETTs, followed by Staphylococcus aureus (8.2%), Enterococcus faecalis (8.2%), Pseudomonas aeruginosa (6.6%), and Stenotrophomonas maltophilia (6.6%). Microbial isolations from the different ETTs are presented in Table 5.

Twelve patients developed 15 episodes of VAP during their stay in the ICU, and three of those episodes were VAP relapses. The most common pathogens involved in VAP were Enterococcus faecium, E. faecalis, Staphylococcus aureus, Klebsiella spp., and Acinetobacter spp. High-grade biofilm formation (score ≥ 7) on the ETTs, days of invasive ventilation, and age were significantly associated with the development of VAP (OR 4.17, 95% CI 1.14–15.3, p = 0.031; and OR 1.11, 95% CI 1.01–1.22, p = 0.026; and OR 0.96, 95% CI 0.92–0.99, p = 0.046, respectively). ETT material, sex, and colonization with common VAP pathogens in surveillance cultures were not associated with development of VAP (Table 6).

Microbial persistence in surveillance cultures could be evaluated in seven of 12 patients who developed VAP, and it occurred in five patients (71%) after appropriate antibiotic treatment for VAP. At extubation, the microorganisms previously causing VAP could be found in the ETT biofilm in 56% of the cases (5 out of 9 patients) despite appropriate antibiotic therapy. The microorganisms most often involved in microbial persistence were Klebsiella spp., Candida parapilosis, Enterococcus faecium, and E. faecalis. In about half of the VAP cases, an evaluation of microbial persistence in surveillance cultures (n = 5) or in the ETT biofilm (n = 3) could not be done, because the course of antimicrobial treatment (primarily antifungal treatment < 7 days) was too short, the pathogens were resistant to initial treatment (inappropriate antibiotic therapy), or the patient received a tracheostomy or died.