Early development of bacterial community diversity in emergently placed urinary catheters
© Foxman et al.; licensee BioMed Central Ltd. 2012
Received: 10 April 2012
Accepted: 31 May 2012
Published: 27 June 2012
Approximately 25% of hospitalized patients have a urinary catheter, and catheter associated urinary tract infection is the most common nosocomial infection in the US, causing >1 million cases/year. However, the natural history of the biofilms that rapidly form on urinary catheters and lead to infection is not well described.
We characterized the dynamics of catheter colonization among catheters collected from 3 women and 5 men in a trauma burn unit with different indwelling times using TRFLP and culture. All patients received antibiotic therapy. Results: Colony-forming units increased along the extraluminal catheter surface from the catheter balloon to the urethra, but no trend was apparent for the intraluminal surface. This suggests extraluminal bacteria come from periurethral communities while intraluminal bacteria are introduced via the catheter or already inhabit the urine/bladder. Richness of operational taxonomic units (OTUs) increased over time on the intraluminal surface, but was constant extraluminally.
OTU community composition was explained best by time rather than axial location or surface. Our results suggest that catheter colonization can be very dynamic, and possibly have a predictable succession.
KeywordsUrinary tract infection Microbial ecology Biofilms Urinary catheter
Catheter associated urinary tract infection is the most common nosocomial infection, causing >1 million cases annually in the United States. All types of indwelling urinary catheters, including silver-coated and those impregnated with antibiotics, are colonized by biofilms -- an essential pathogenic feature of catheter infections. Biofilms tend to be resistant to antibiotic treatment and may enhance the development of further antibiotic resistance. The majority (88%) of urinary catheter biofilms contain 3 or more microbial species. It is assumed that a biofilm on a catheter pre-disposes to bacteriuria, but it is uncertain if the biofilm is the source of bacteriuria. Among catheterized patients, the risk of bacteriuria is ~ 3-8% per day. Almost all will be colonized after one month.
Bacteria are introduced with the urinary catheter, or ascend the extraluminal surface from the urethra to the bladder after the catheter is inserted. Intraluminal colonization can occur if there are breaks in infection control, leading to contamination of the catheter bag, or failure of the closed drainage system. Information on the bacterial community structure and dynamics of catheter biofilm development is very limited. In a study of 34 urology patients, Barford et al. showed the number of colony forming units (CFUs) recovered from the urinary catheter surface increases with duration of catheterization, and the number of CFUs increases more rapidly on the extraluminxal than intraluminal surface. Barford et al. and a study by Frank et al., following 14 patients catheterized for two weeks following total prostatectomy, also reported on differences in community structure between catheter intraluminal and extraluminal surfaces, finding more different strains on extraluminal than intraluminal surfaces. Frank et al. further reported that each catheter had unique biofilm communities and intraluminal and extraluminal communities on a given catheter were often substantially different. The faster colonization on the extraluminal than intraluminal surface, and its distinct composition, suggests that the extraluminal surface is inoculated from the periurethral skin upon insertion and the intraluminal surface is either colonized by existing bladder inhabitants or those that colonize the urine following catheter insertion. Presumably an analogous mechanism occurs in other devices, such as percutaneous nephrostomy tubes and urethral stents, placed in the upper urinary tract. With the exception of Barford et al., we found no studies actually following the dynamics of catheter biofilms over time, and the dynamics in Barford et al. are limited to changes in total CFUs and not community composition. Understanding biofilm dynamics is essential for identifying ways to manipulate biofilm growth on catheters and prevent catheter-associated infection.
We used terminal restriction fragment length polymorphism analysis (TRFLP) and CFU counts to characterize indwelling urinary catheter microbial community dynamics as a function of catheterization duration, intra- or extraluminal device surface, and distance from the catheter balloon to the external environment. TRFLP is a method for profiling microbial communities based on DNA. We were particularly interested in whether a predictable succession in microbial community structure was observed and whether differences in dynamics between catheter surfaces and distance from the catheter balloon could be used to infer the sources of the microbes colonizing the catheter.
Materials and methods
We collected urinary catheters from a convenience sample of adult polytrauma patients with multisystem injury without hemodynamic instability with or without instability or shock resuscitated in the emergency department of a Level I trauma center and admitted to the trauma and burn intensive care unit. Catheters were inserted using aseptic procedures including cleaning of the urethral meatus with a betadyne solution. All catheters were silver impregnated. Catheters were collected for analysis when the treating team believed it appropriate for their removal. Catheters were transferred in sterile containers to the laboratory, and refrigerated until processing, which occurred within 1 hour following removal. No modifications to standard critical care practice were imposed. Participants were likely treated with antibiotics during their hospital stay, but we did not have access to that information. The proposed use of catheters and urine was reviewed by the University of Michigan institutional review committee on human subjects protection and considered exempt because we used discarded materials that were not identifiable.
Sample collection and treatment
Catheters were marked, at the time of removal, to indicate the point of entry into the urethra. Catheters collected from male patients were cut into four axial segments. The most proximal ‘bladder’ section included the inserted tip of the catheter and the associated balloon. Sections 2 and 3 were subsequent 10 cm segments comprising the ‘deep urethral’ and ‘superficial urethral’ regions. The fourth segment was an additional 10 cm ‘external’ segment. Catheters collected from female patients were divided into a bladder segment, a 10 cm urethral segment, and an external segment. Each section was bisected lengthwise to allow sampling of the intraluminal and extraluminal device surfaces separately, thus organizing each device into 8 and 6 spatial zones for male and female patients, respectively. Bacteria were collected from each zone using a sterile cotton swab which was placed into 2 ml of 1 x PBS buffer, then recovered using the Swab Extraction Tube System (Fisher Scientific). Cells were resuspended in 1 ml of PBS. Urine (20 mL) was collected from the catheter port of 5 participants (three male patients, two catheterized for one day, and one for two days; and two female patients, one catheterized for two days, the other for nine days) at time of catheter removal. Urine was centrifuged at 10,000× g for 10 min at 4 °C and the pellets were resuspended in 1 ml of PBS. 100 μl of cell suspensions from the catheters and urine was used for quantitative culture on tryptic soy agar and blood agar. The plates were incubated at 37 °C for two days or until colonies showed. The remaining 900 μl of cell suspension from the catheters was used for DNA extraction. For the analysis, the number of CFUs for the two plates were averaged.
Terminal restriction fragment length polymorphism (TRFLP) analysis
Bacterial community fingerprints were obtained using TRFLP. DNA was extracted from bacterial cells from urine samples and each catheter zone using the MasterPure™ DNA Purification Kit (Epicentre Biotechnologies, Madison, WI). Polymerase chain reaction (PCR) amplification was performed in triplicate with the primer pairs 63F: 5-CAGGCCTAACACATGCAAGTC -3′ and 1389R: 5- ACGGGCGGTGTGTACAAG -3′ targeting general bacterial 16S rRNA gene. The forward primer was 5’-labeled with the fluorescent dye 6-FAM (6-carboxyfluorescein) (IDT, Coralville, IA). To minimize amplification of chimeras and pseudo-terminal restriction fragments, 25 cycles were conducted with extended elongation time of 3 minutes. After amplification, the triplicate PCR products were pooled and purified using a QIAquick PCR purification kit (Qiagene, Valencia, CA). The purified PCR products were digested with restriction enzymes Hha I or Msp I (NEB, Ipswich, MA) for 4 hours at 37°C. Following digestion, samples were loaded into a 96-well GS1000 ROX1000 CE plate and sent to the DNA Core at University of Michigan for T-RFLP analysis. Peaks in each profile were related to specific fragment lengths based on a size marker (50-1000 MapMarker, GeneScan™, Applied Biosystems). Data were retrieved using Peak Scanner software v1.0 (Applied Biosystems) and normalized using T-Align before further statistical analysis. Each distinct 16s fragment length produced was considered an operational taxonomic unit (OTU).
For number of CFUs and OTU richness, the effects of time, axial location, and intraluminal/extraluminal surface type, as well as interactions between axial location, time, and surface type were tested with a repeated-measures, mixed-effects model in R. CFU was log transformed prior to analysis. For OTU richness, because two different digests were performed, both restriction digests were combined in one model and digest was included as a main effect. Study subject was modeled as a random effect. For all analyses, the significance of fixed effects (factors of interest that are not a random sample from a larger population) was tested with F tests, and the significance of random effects (factors whose levels are a random sample from a larger population) and the proper covariance structure for the repeated measures were tested using likelihood ratio tests. In this study of dynamics, we treated time as a fixed factor because it was the central factor of interest. Axial location was modeled as a continuous variable (1, 2, 3, 4 for males and 1, 2, 3 for females), because AIC scores were lower for both CFU and OTU richness when models with axial location as a continuous vs. a categorical (3 categories, balloon, middle, and air) variable were compared.
To analyze whether subject, time, axial location, and surface affected OTU community composition, we calculated pairwise Bray-Curtis dissimilarity indices among all samples separately for each of the Hha I and Msp I digests. Bray-Curtis ranges between 0 and 1: a value of 0 indicates samples have the same species composition and 1 indicates samples have no species in common. We used the quantitative form of Bray-Curtis which takes into consideration amount, not just presence/absence, of each OTU:, where dj,k is the dissimilarity between samples j and k, and xi is the abundance (standardized peak area) of OTU i. This resulted in one dissimilarity matrix for each digest. We then tested the significance of subject, time, length, and surface for degree of community dissimilarity using permutation tests in R for each digest (package vegan, functions vegdist and adonis). Because subject and time were completely confounded (each subject was sampled once, at a given time) and because adonis and other multivariate permutation programs do not have the capacity for complex random effects structure in the analysis, we first tested the effect of subject with a model including only subject as the independent variable. We then tested the effect of time in a model including only time as the independent variable. Lastly, we tested the effect of axial location and surface in the same model with permutations restricted within subjects.
To visualize OTU community composition for the Hha I digest, an NMDS (non-metric multi-dimensional scaling) plot was constructed based on the Bray-Curtis dissimilarity matrix. For ease of visualization, for each subject, all samples from the extraluminal surface were combined and all samples from the intraluminal were combined. For five of the subjects, urine samples were also obtained; these samples were included in the NMDS.
Urinary catheters placed as part of the resuscitation of acutely injured outpatients were collected from 3 women and 5 men, and had been in place an average of 3 days (range 1 to 16) before removal.
Colony forming units on tryptic soy and blood agars
Effect of time, extra- versus intraluminal location, and axial location on number of colony forming units (CFU) in urinary catheters
Surface × time
Surface × axial location
Effects of digest, time, surface, and axial location on OTU richness in urinary catheters
Surface × duration
Surface × axial location
OTU community composition
Analyses of variance using community dissimilarity matrices with permutation tests for significance (function adonis in R package vegan) for digests from microbial communities from urinary catheters
Culture and TRFLP analysis of 8 urinary catheters placed for acute trauma resuscitation and left in place between 1 and 16 days showed rapid development of complex microbial community structures that varied by location within the device. The key findings of the current report are these. First, the cultivable burden of organisms attached to the devices was higher on the extraluminal compared to intraluminal surface particularly toward the external portion of the catheter (Figure1). The number of colony-forming units tended to increase from the bladder to outside the body on the extraluminal surface while remaining similar along the inner surface. This is not surprising given the differences in potential sources of colonists found along the extraluminal surface (from bladder to urethra to patient bed) and the more similar potential sources along the intraluminal surface (urine), and that the urine likely included antibiotics. However, it is possible that - despite using the same methods - that recovery was greater on the extraluminal surface, or that the external catheter surface was contaminated during catheter removal. We believe that unlikely, as catheters were processed rapidly following removal, and our findings are consistent with other, similar studies. In a study of catheters from 34 urology patients all segments of the extraluminal surface were colonized by day 1 but it was not until day 4 that all axial locations along the intraluminal surfaces were colonized.
Second, using non-culture methods, we observed an increase in the richness of OTUs over time on the intraluminal surface, but constant richness extraluminally and with axial location. The OTU composition of the communities changed linearly over time, but surface had only marginal effects on the OTU community composition. This contrasts with Barford et al.’s study of 34 urology patients, which found significantly more strains and genera on extraluminal than intraluminal surfaces, although they did not present data by time since catheterization. We found no other reports for comparison. Both our and Barford’s results are consistent with the hypothesis that the extraluminal surface is inoculated during catheter insertion.
Finally, results from our analysis of species composition on the catheters suggest that there may be a predictable succession in microbial communities over time ( Additional file1: Figure S2). We found no similar analyses in the literature to support or refute this observation. Most catheter-associated infections of the urinary tract are asymptomatic and do not progress to symptomatic infection, but colonization is the first step towards symptomatic infection. An outstanding question - which we could not directly address - is the extent to which the catheter biofilm is the source of asymptomatic bacteriuria and catheter-associated urinary tract infection. A study of 86 patients with urethral catheters in Japan, found that urinary and intraluminal catheter isolates were not always the same, independent of antibiotic therapy. A 1999 study of 1,497 newly catheterized patients conducted at Mayo Clinic compared paired daily urine cultures from the catheter specimen port and collection bag, and assumed that extraluminal infections would be first detected in the specimen port and intraluminal infections in the collection bag. They concluded that 66% of infections were extraluminally acquired. However, they assumed that intraluminal infections were caused by microorganisms ascending the catheter, an assumption not supported by our study or that of Barford et al. which both suggest that bacteria colonizing the intraluminary surface of the catheter and urine have a common source. Further, for the 5 urine samples we tested, the urinary communities tended to be more similar to intraluminal than extraluminal communities (Bray-Curtis dissimilarity of 0.334 (SE 0.018) vs. 0.435 (SE 0.041)). It is possible that the bacteria that grow best on catheter surfaces are not the same as those that grow well in planktonic form in the urine or that bind well to urinary tract tissues thereby interacting with the host to cause infection. Notably, in a study of 14 patients catheterized for 2 weeks following total prostatectomy, the most prevalent species found in catheter biofilms (either surface) were Pseudomonas aeruginosa (64%), Klebsiella pneumoniae (63%) and Escherichia coli (50%). The bacterial species that most often cause catheter-associated UTI is Escherichia coli. As bacteriuria from catheterized patients is often polymicrobial - especially in patients where catheterization is of longer duration - it is also possible that the microbial community structure may mediate interaction with the host.
This pilot study is a first step toward understanding the dynamics of catheter colonization. Like all studies conducted in real populations, there were inevitable compromises in data collection: we had no control over the treatment of patients with antibiotics, time of catheter removal, and reason for hospitalization. Nonetheless, we detected significant trends and patterns that are unlikely to be the result of systematic error. Future, larger, studies, with greater resolution of species present, are needed to identify how changing microbial communities are related to the development of asymptomatic bacteriuria and symptomatic infection.
The authors gratefully acknowledge the nurses, residents and staff of the trauma burn center for their assistance in conducting this study. This work was supported by the Office of the Vice-President for Research and the Center for Molecular and Clinical Epidemiology of Infectious Diseases at the University of Michigan. J. W. is supported by an internal grant from the School of Public Health to C. Xi.
- Klevens RM, Edwards JR, Richards CL, et al: Estimating health care-associated infections and deaths in U.S. hospitals, 2002. Public Health Rep. 2007, 122 (2): 160-166.PubMedPubMed CentralGoogle Scholar
- Stickler DJ: Bacterial biofilms in patients with indwelling urinary catheters. Nat Clin Pract Urol. 2008, 5 (11): 598-608.PubMedView ArticleGoogle Scholar
- Francolini I, Donelli G: Prevention and control of biofilm-based medical-device-related infections. FEMS Immunol Med Microbiol. 2010, 59 (3): 227-238.PubMedGoogle Scholar
- Hola V, Ruzicka F, Horka M: Microbial diversity in biofilm infections of the urinary tract with the use of sonication techniques. FEMS Immunol Med Microbiol. 2010, 59 (3): 525-528.PubMedGoogle Scholar
- Matsukawa M, Kunishima Y, Takahashi S, et al: Bacterial colonization on intraluminal surface of urethral catheter. Urology. 2005, 65 (3): 440-444. 10.1016/j.urology.2004.10.065.PubMedView ArticleGoogle Scholar
- Hooton TM, Bradley SF, Cardenas DD, et al: Diagnosis, prevention, and treatment of catheter-associated urinary tract infection in adults: 2009 International Clinical Practice Guidelines from the Infectious Diseases Society of America. Clin Infect Dis. 2010, 50 (5): 625-663. 10.1086/650482.PubMedView ArticleGoogle Scholar
- Barford JM, Anson K, Hu Y, Coates AR: A model of catheter-associated urinary tract infection initiated by bacterial contamination of the catheter tip. BJU Int. 2008, 102 (1): 67-74. 10.1111/j.1464-410X.2008.07465.x.PubMedView ArticleGoogle Scholar
- Saint S, Chenoweth CE: Biofilms and catheter-associated urinary tract infections. Infect Dis Clin North Am. 2003, 17 (2): 411-432. 10.1016/S0891-5520(03)00011-4. ReviewPubMedView ArticleGoogle Scholar
- Frank DN, Wilson SS, St Amand AL, et al: Culture-independent microbiological analysis of foley urinary catheter biofilms. PLoS One. 2009, 4 (11): e7811-10.1371/journal.pone.0007811.PubMedPubMed CentralView ArticleGoogle Scholar
- Liu WT, Marsh TL, Cheng H, et al: Characterization of microbial diversity by determining terminal restriction fragment length polymorphisms of genes encoding 16S rRNA. Appl Environ Microbiol. 1997, 63 (11): 4516-4522.PubMedPubMed CentralGoogle Scholar
- Osborn AM, Moore ER, Timmis KN: An evaluation of terminal-restriction fragment length polymorphism (T-RFLP) analysis for the study of microbial community structure and dynamics. Environ Microbiol. 2000, 2 (1): 39-50. 10.1046/j.1462-2920.2000.00081.x.PubMedView ArticleGoogle Scholar
- Egert M, Friedrich MW: Formation of pseudo-terminal restriction fragments, a PCR-related bias affecting terminal restriction fragment length polymorphism analysis of microbial community structure. Appl Environ Microbiol. 2003, 69 (5): 2555-2562. 10.1128/AEM.69.5.2555-2562.2003.PubMedPubMed CentralView ArticleGoogle Scholar
- Smith CJ, Danilowicz BS, Clear AK, et al: T-Align, a web-based tool for comparison of multiple terminal restriction fragment length polymorphism profiles. FEMS Microbiol Ecol. 2005, 54 (3): 375-380. 10.1016/j.femsec.2005.05.002.PubMedView ArticleGoogle Scholar
- R Development Core Team: R: A language and environment for statistical computing. 2008, Vienna: R Foundation for Statistical ComputingGoogle Scholar
- West BT, Welch KB, Galecki AT: Linear mixed models: a practical guide using statistical software. 2007, Boca Raton: Chapman & Hall/CRC PressGoogle Scholar
- Tambyah PA, Halvorson KT, Maki DG: A prospective study of pathogenesis of catheter-associated urinary tract infections. Mayo Clin Proc. 1999, 74 (2): 131-136. 10.4065/74.2.131.PubMedView ArticleGoogle Scholar