Characterization of bacterial biofilms on tracheostomy tubes

To characterize the structure and microbial content of biofilms found on tracheostomy tubes. To determine the correlation between the patients' clinical condition and biofilm content.


INTRODUCTION
Biofilms are complex three-dimensional structures composed of bacteria living in an extracellular matrix rich in polysaccharides, nucleic acids and proteins. Biofilm formation occurs in a well-defined series of steps. The bacteria first adhere to a surface. This surface can be either artificial, such as an implant, or natural, such as a tooth surface. Following attachment, bacteria aggregate and replicate to form microcolonies that produce the extracellular matrix resulting in a mature biofilm.
Within mature biofilms bacteria may enter a slow growth or quiescent state, allowing for prolonged survival. 1 Some bacteria within biofilms will go through cycles of active shedding, whereas others are shed by shear forces. These intermittently shed bacteria could potentially lead to active infection and may be the cause of some recurring infections despite treatment with multiple courses of antibiotics. 2 Biofilms have been implicated in recurrent septicemia in patients with indwelling devices, recurrent acute otitis media, recurrent Pseudomonas pneumonias in cystic fibrosis patients, and in infectious endocarditis. 3,4 Resistance mechanisms afforded by the presence of biofilms include decreased rates of bacterial division, which make antibiotics less effective. A negatively charged biofilm surface can repel positively charged antibiotic molecules. Differences in the microenvironment within the biofilm allow for differential gene expression that permit bacteria to better survive in high-stress, low-nutrient environments. Changes in membrane structure, such as increased expression of efflux pumps, decrease the accumulation of antibiotics in bacterial cells. Finally, the large size of the biofilm matrix makes it difficult for phagocytes to ingest them. 2 These and other mechanisms make bacteria in biofilms resistant to antibiotics even at concentrations several thousand times greater than the minimum inhibitory concentration. 5 The presence of biofilms on indwelling biomedical devices, such as pressure equalization tubes, orthopedic prostheses, urinary catheters, intravascular catheters, and dental implants, has been increasingly recognized over the past decade. 6,7 Biofilm formation on tracheostomy tubes has been less well studied. Jarrett et al. observed biofilm formation on tracheostomy tubes in vitro after inoculating tracheostomy tubes made of four different materials with Pseudomonas aeruginosa and Staphylococcus epidermidis. 8 Perkins et al. noted an increasing density of biofilm formation as one progressed toward the tip of tracheostomy tubes obtained from nonventilated pediatric inpatients. 9 We undertook this study to characterize biofilm formation on tracheostomy tubes in adults by determining their structure using confocal microscopy, specifying the types and quantifying the number of bacteria present, and correlating this with the patients' clinical condition at the time of tube change.

MATERIALS AND METHODS
After obtaining approval from our university's institutional review board, both inpatients and outpatients at Temple University Hospital and the Department of Otolaryngology-Head and Neck Surgery requiring tracheostomy tube replacement were asked to enroll in this study. The patients' medical and social history was reviewed and recorded (Table I). The type of tracheostomy tube, time since last change, and appearance of the stoma were also recorded at the time of tube change (Table II). The tube removed was then collected, stored in a sterile plastic bag at 4 C and processed within 24 hours.
The exterior of the tracheostomy tubes was first wiped with ethanol to remove and kill any bacteria present on the outside surface. The tube was then sterilely sectioned into 2-mm slices in a tissue culture hood to prevent contamination; two adjacent sections were used for analysis. One section was fixed in 3.7% formaldehyde, stained with Syto7 (Molecular Probes Inc., Portland, OR), which fluorescently labels DNA, and then imaged using an inverted Leica TS5 confocal microscope (Leica Microsystems, Wetzlar, Germany). The adjacent section of the tube was placed in 10 mL of phosphate buffered saline (PBS) and sonicated to release bacteria using a Fisher Scientific Sonic Dismembrator Model 500 (Thermo Fisher Scientific Inc., Waltham, MA) at 60% for three 10-second pulses. One hundred microliters of this fluid was then diluted 1:1000 in PBS, plated on tryptic soy agar (TSA) or TSA supplemented blood agar (BAP) and incubated overnight at 37 C. On the following morning, the plates were visually inspected for number and types of colonies present. Images of the plates were then captured using an Epson digital scanner (Epson America Inc., Long Beach, CA) for examination and quantification of colonies.
The number of colony forming units (CFUs) was then calculated for each 2 mm section, with 10 6 being the lower limit of detection. Representative colony types were chosen at random from at least three independent plates for inoculation in tryptic soy broth. After overnight incubation at 37 C, samples were removed, mixed with 30% glycerol and stored at À70 C.
Precise determination of all bacterial species present based upon plating alone was difficult. Although some bacterial colonies such as Proteus vulgaris had a very distinct morphology, others, including coagulase-negative staphylococci and Staphylococcus aureus, could not be reliably distinguished by morphology alone on TSA or BAP without further diagnostic testing. Colonies were selected for speciation from 14 of the 19 specimens. Because the difficulty in speciating bacteria based on colony morphology alone was noted late in the study, bacterial stocks did not exist for five of the 19 positive specimens, and not all of the bacteria species from a single patient were available as stocks.
The 14 stored stocks were then further purified by serial plating, Gram stained, and speciated by standard clinical microbiology laboratory methods using pure cultures. Gram-positive organisms were evaluated by catalase testing (Fisher Chemicals, Fair Lawn, NJ) and slide coagulase testing (Pastorex Staph Plus, Bio-Rad, Marnes la Coquette, France). Streptococcus species were identified by the RapID STR System (Remel, Inc., Lenexa, KS). Enterococcus species were identified by PYR

RESULTS
Twenty-one tracheostomy tubes were collected from 14 patients between May and October 2008. Two of the tracheostomy tubes were collected from hospital inpatients; the remaining 19 tracheostomy tubes were collected from 12 different patients in the outpatient setting. The age of the patients enrolled in this study ranged from 21 years to 83 years with a mean of 55 years and a median of 60 years. There were seven males and seven females.
Four patients were mechanically ventilated at some point during the day. Three patients were active tobacco smokers and four of the patients were active alcohol drinkers. Seven patients had hypertension, five had diabetes, two patients had obstructive sleep apnea, two had previous cerebrovascular accidents, two had gastroesophageal reflux, one had asthma, one had a seizure disorder, one was schizophrenic, and one had a past myocardial infarction.
Indications for tracheostomy in the patient population studied were ventilator-dependent respiratory failure in four patients, tracheal stenosis in two patients, laryngeal stenosis in two patients, laryngeal cancer in one patient, obstructive sleep apnea in one patient, chondromalacia in one patient, and chronic aspiration following cerebrovascular accident in one patient. Four patients were mechanically ventilated at some point during the day.
Of the 21 tracheostomy tubes collected, 17 were uncuffed Shiley tubes, two were cuffed Shiley tubes, one was an uncuffed Bivona Hyperflex tube, and one was an uncuffed Portex tracheostomy tube. Stomal edema, induration, or granulation tissue was noted in six patients at eight of the 21 tracheostomy tube changes performed.
Patients enrolled in this study reported changing their inner cannula, if present, from one to seven times per day. Recent sputum culture results were present in three of 21 patients at the time of tracheostomy tube change. In two cultures, methicillin resistant Staphylococcus aureus was present and Candida albicans was present in another.
Biofilm formation on the tracheostomy tube sections was observed using confocal microscopy ( Figs. 1 and 2). Biofilms appeared to form both on the tube surface and in the mucus layer coating the inside of the tube. In the latter, the biofilms often appeared as discrete microcolonies that appeared to be monospecies biofilms even though there were multiple types of bacteria present on the tracheostomy tube. Biofilm formation was detected as early as 1 week after tube insertion, which was the earliest time point evaluated. Bacteria were cultured from 19 of the 21 tracheotomy tubes collected. There were between 1 Â 10 6 and 1 Â 10 10 CFUs present in each of the 2-mm sections. The numbers were evenly distributed across the range. Patients 6 and 13 both had no detectable bacteria on their tracheotomy tubes with 1 Â 10 6 CFUs being the lower limit of detection. The number of bacteria isolated and the CFUs calculated varied in tubes obtained from the same patient at different times.
Using a univariate regression analysis, an inverse correlation was noted between the number of CFUs present in each 2 mm tracheostomy tube section and the number of times the tracheostomy inner cannula was changed daily (P ¼ .0137). None of the other aspects of the past medical history nor tracheostomy care-related factors had a statistically significant correlation (Table IV).

DISCUSSION
Biofilms are complex, organized, three-dimensional structures that may form on indwelling medical devices as quickly as 2 hours after insertion, although some bacteria take weeks to form biofilms on these devices depending upon environmental factors. 10 In our study, biofilms were present at the earliest time point evaluated, 7 days after placement of a tracheostomy tube. These findings were consistent with those of Jarrett et al., where biofilm was noted on tracheostomy tubes after 6 days of intubation in a laboratory setting. 8 The presence of biofilms on 19 of 21 tracheostomy tubes is in keeping with Perkins, et al.'s findings on pediatric tracheostomy tubes in which 10 of 11 pediatric tubes were noted to have biofilms present by confocal microscopy with concurrent bacterial staining. 9 Unique to our study, however, was that the biofilms in the mucus layer appeared to be monospecies biofilms. When multiple bacterial species are present in a biofilm, the bacteria will grow either intermixed or in close proximity to each other. In the mucus layer, the biofilms seemed to be separate monospecies biofilms. This may be due to factors unique to biofilm formation in mucus or properties of the bacteria forming the biofilms.
The absence of detectable bacteria in two of our patients may well be related to excellent local care. Both patients who did not have biofilms identified in their tracheostomy tubes at the time of collection reported being fastidious with their tracheostomy care, including removing and cleaning the entire tube multiple times per week.
The organisms identified in our study include many of the common flora of the upper aerodigestive tract and have been isolated from biofilms on endotracheal tubes and laryngeal stents. 11 Though the speciation for each patient may not be complete, some trends were  identified. The species of bacteria isolated were variable from patient to patient and over time in the same patient. Coagulase negative staphylococci were isolated from a number of patients. The predominant coagulase negative staphylococci is Staphylococcus epidermidis. Staphylococcus epidermidis is often a member of the normal flora and an excellent biofilm former, which may account for its presence in six of the 12 tracheostomy tubes on which speciation was completed. Staphylococcus aureus was present in four patients in whom bacterial species were identified, and three of these patients had evidence of stomal granulation. Grillo et al. observed stomal granulation tissue in the presence of nonabsorbable suture and correlated this with the presence of Pseudomonas aeruginosa and Staphylococcus aureus on culture. 12 The presence of Escherichia coli, which is a normal fecal flora, in two patients confirms that hygiene may contribute to colonization and subsequent biofilm formation.
One patient exhibited beta-hemolytic group B streptococci, suggesting that it can form biofilms in mucus or on tracheostomy tubes. Although Streptococcus agalactiae biofilm formation is believed to occur during bovine mastitis, there has been only one report of group B streptococci growing in a biofilm. 13 This was present on three different tubes collected from this patient and appeared to have a two to three logarithm increase in the number of colony forming units at one point, suggesting a possible infection.
Several patients in our study had Pseudomonas aeruginosa. Although Pseudomonas aeruginosa is often associated with ventilator assisted pneumonia (VAP), no patient in our study had an acute pulmonary infection. The question remains whether bacterial biofilms within endotracheal tubes and tracheostomy tubes are causative agents for VAP, or if the bacteria present in these biofilms are merely colonizing tracheostomy tubes after an active infection has cleared.
Prior to this study, quantification of bacterial CFUs within tracheostomy tube biofilms had not been reported. The number of CFUs ranged from 1 Â 10 6 to 1 Â 10 10 in those tracheostomy tubes that had biofilms present. Of note, in one patient, an increase in the CFU count was associated with a hospital admission and correlated with a bacterial tracheitis, later confirmed to be due to methicillin resistant Staphylococcus aureus. The univariate regression analysis performed was also able to show a negative correlation between the number of CFUs present and the number of times an inner cannula was changed daily, again confirming the importance of local care.

CONCLUSION
Bacterial biofilms are present in greater than 90% of tracheostomy tubes collected in both the inpatient and outpatient setting. Although a variety of bacteria were identified in the biofilm, they often appeared as discrete microcolonies that appeared to be monospecies biofilm on confocal microscopy. The number of colony forming units was between 1 Â 10 6 and 1 Â 10 10 per 2-mm section, which inversely correlated with the frequency of the inner cannula change (P < .05).