Vol. 48, No. 6APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Dec. 1984, P. 1159-11650099-2240/84/121159-07$02.00/0Copyright © 1984, American Society for Microbiology

Microbial Colonization of Human Ileal ConduitsRAPHAEL C. Y. CHAN,' GREGOR REID," 2 ANDREW W. BRUCE,1 AND J. WILLIAM COSTERTON2*

Division of Urology, Department of Suirgery, University of Toronto, Toronto General Hospital, Toronto, Ontario, M5G1L7,1 and Department of Biology, University of Calgary, Calgary, Alberta T2N IN4,- Canada

Received 11 April 1984/Accepted 7 September 1984

Morphological and microbiological techniques were used to locate and identify the microorganisms thatcolonized the human ileal conduits in 17 different patients from 5 days after surgery up to as many as 16 years

of service as a urine conduit. The ecological sequence of this colonization assumes some practical importancebecause the ascending growth of pathogenic organisms in this essentially open, unvalved urinary tract diversionsystem leads to the development of life-threatening pyelonephritis. Extensive examination of the microvillussurfaces of the ilea of five accident victims by both transmission and scanning electron microscopy showed thatthese tissue surfaces were not colonized by bacteria, even in the absence of prophylactic antibiotic therapy, andthat these surfaces were not occupied by adherent microorganisms after several years of service as a urineconduit, even when the skin surface stoma and the conduit contents were heavily colonized by bacteria andyeasts. During the initial period (10 days) of postoperative antibiotic therapy, the mucus and urine within theconduit were largely colonized by yeasts. A mixed population of yeasts and gram-positive cocci subsequentlydeveloped in the conduit itself, and gram-positive cocci were seen to be avidly adherent to epidermal cells at thestoma. As antibiotic protection was gradually withdrawn, gram-negative organisms became a part of the mixedmicrobial flora of the conduit contents, and some of the potentially pathogenic organisms of this group (e.g.,Escherichia spp., Proteus spp., Pseudomonas spp., etc.) were isolated from patients with pyelonephritis thatappeared to come from the ileal conduit. This study of the natural microbial ecology of this surgical urinarytract diversion is intended to provide a rational basis for the manipulation of the ileal conduit to make use ofcompetitive microbial exclusion to protect patients with ileal conduits from ascending infections.

When a bladder is removed because of cancer or neuro-genic disease, the urine is diverted from the ureters to theskin surface with a short section of the ileum used toconstruct an ileal conduit (3, 24). The ureters are connectedto the proximal end of a blind tube (6 to 8 in. [ca. 15.2 to 20.3cm]) constructed of an excised section of the ileum joined tothe skin at its distal end. Surgical procedures cannot pre-clude the reflux of urine from the ileal conduit to the kidneysvia the ureters and, because 80% of patients with ilealconduits develop bacteriuria (23), a significant number de-velop pyelonephritis. Laboratory animals in which ilealconduits were surgically constructed showed an 83% inci-dence of pyelonephritis (19). In humans, these microbialinfections are life threatening (1) and may also indirectlycause stenosis of the conduit (10) or renal deterioration (21)or both.The autochthonous bacterial population of the ileum is

controlled, before surgery, by empirical antibiotic therapy,and the ileal conduit is colonized postsurgically, via thecutaneous opening (stoma), as it is converted from itsdigestive function into a urine conduit. This ascendingcolonziation is influenced by the postoperative administra-tion of antimicrobial agents, which affect the nature of theorganisms entering the conduit. Thus, the organisms thatpersist in the ileal conduit are influenced by ecologicalfactors and by the selective pressure of antimicrobial thera-PY.Because the incidence of bacteriuria and pyelonephritis is

high in patients with ileal conduits, despite intermittentantimicrobial therapy, we resolved to study the developmentand persistence of autochthonous and pathogenic microbial

* Corresponding author.

populations in ileal conduits in 17 patients immediately aftersurgery and up to 16 years after surgery.

MATERIALS AND METHODSPatients. A total of 17 patients (12 males, 5 females) who

had undergone urinary tract diversion within a period of upto 16 years before this investigation were studied, andsamples were taken from their conduits as described below.The patients were originally operated on for the treatment oftransitional cell carcinoma, bladder cancer, or neurogenicbladder disease. Five accident victims were also included inthe investigation as a source of "normal" ileal tissue andwere not subjected to antibiotic therapy.Specimen collection and processing. Urine samples, mucus

samples, and cup biopsy specimens of ileal conduit tissueswere taken from four patients at 5- to 10-day intervals for upto 40 days after the operation. Another group of 13 patientswas examined on one occasion; a swab was taken from thestoma, and urine samples and cup biopsy specimens weretaken from superficial and deep areas of the conduit by adouble-lumen-catheter technique (2). The mucus and urinesamples were cultured by standard microbiological tech-niques, and organisms were isolated and identified. Mucussamples and cup biopsy specimens were processed fortransmission electron microscopy (TEM) by being fixed in5% glutaraldehyde in cacodylate buffer (0.1 M, pH 7.2) with0.15% ruthenium red for 2 h at room temperature and thenprocessed as described by Costerton (8) and Chan and Bruce(5). The sections were examined with a Hitachi 500 electronmicroscope at an accelerating voltage of 60 kv. For scanningelectron microscopy (SEM), the specimens were fixed andprocessed as described by Chan and Bruce (5) and examinedwith a Hitachi 450 microscope at an accelerating voltage of20 kV.

1159

on June 11, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

.1VAI? I~

t

1160 CHAN ET AL.

on June 11, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

MICROBIOLOGY OF ILEAL CONDUITS 1161

TABLE 1. Microbial isolates from stoma and urine samples collected from the ileal conduits of 13 patients over 1 month after surgery, asreported by the Toronto General Hospital Diagnostic Laboratory

PatientTime after Bacteria found in:Patient Tm fe

surgery Stoma samples Deep and superficial urine samples

1 1 mo Proteus mirabilis, E. coli 1, E. coli 2, enterococci, Candida E. coli 1, E. coli 2, P. mirabilis, C. albicansalbicans, Staphylococcus aureus, Enterobacter cloacae,Morganella spp., Citrobacter spp., Klebsiella oxytoca,Pseudomonas aeruginosa

2 4 mo P. vulgaris, E. coli, K. oxytoca, Morganella morganii, P. rettgeri, M. morganii, enterococci, K.enterococci, Streptococcus viridans oxytoca, P. vulgaris

3 4 mo Streptococcus spp. Enterococci, gram-positive and gram-negativebacilli, C. albicans

4 4 mo E. coli, K. pneumoniae, enterococci, S. epidermidis E. coli

5 6 mo E. coli E. coli, gram-positive bacilli

6 6 mo Staphylococcus spp. Staphylococcus spp.

7 7 mo Staphylococcus spp. Staphylococcus spp.

8 17 mo Staphylococcus spp. Streptococcus spp., gram-positive bacilli E. coli 1, E. coli 2, K. pneumoniae,enterococci

9 20 mo E. coli, P. mirabilis, enterococci E. coli

10 33 mo M. morganii, P. aeruginosa, K. pneumoniae, enterococci K. pneumoniae, K. oxvtoca

11 43 mo E. coli, K. oxytoca, Providencia rettgeri, P. aeruginosa, Noneenterococci

12 10 yr E. coli E. coli

13 16 yr E. coli, gram-positive cocci E. coli, gram-positive cocci

RESULTS (Fig. 3), but this histological change is not universal whenthe ileum becomes a urine conduit; well-developed microvilli

Ileal tissues were obtained from five accident victims who were often seen in ileal conduits 2 to 16 years after surgery.had no recent history of exposure to antibiotics and from We undertook an intensive study of 17 patients, rangingseven patients who had undergone urinary tract bypass from 1 day to 16 years after urinary tract diversion, tosurgery and had been given preoperative antibiotic therapy. determine the extent of microbial colonization of their ilealExamination of very large areas (at least 4 mm2 in each of the conduits. Swab samples were taken at the stoma, and urinefive accident victims) of the surfaces of these ileal tissues by samples and cup biopsies were taken from a superficial site 3SEM showed intact villus structures and extensive remnants cm inside the stoma and from a deep site near the conduit-of the mucus blanket (Fig. 1), but no bacterial cells could be ureter junction. The keratinized epithelial cells obtaineddiscerned either on the tissue surface or embedded in the from stoma samples taken from patients 2 weeks aftermucus. Similarly, TEM of these ileal tissues showed intact, surgery were heavily colonized by gram-positive cocci (Fig.well-developed microvilli and a partial retention of the 4); samples collected over 1 month after surgery contained amucus blanket, but no bacteria or other microorganisms great variety of gram-positive and gram-negative bacteriacould be seen either on the tissue surface or within the (Table 1). Detailed TEM examination of large areas of themucus (Fig. 2). At least 6.0 x 10' ,um of the tissue surface surfaces of the deep and superficial biopsy specimens fromwas examined in this study. TEM of a biopsy specimen all 17 patients showed the complete absence of adherenttaken from the wall of an ileal conduit 40 days after surgery bacteria on the surfaces of the microvilli of the ileal conduitshowed that the microvilli were truncated and degenerated tissues (Fig. 3). Mucus collected from the deep and superfi-

FIG. 1. SEM of a critical-point-dried, thiocarbohydrazide-treated preparation of ileal tissue from an accident victim with no immediatepast history of antibiotic usage. The villi can be clearly seen, as can remnants of the mucus blanket, and careful examination of very large ar-eas (at least 4 mm2 in each of the five accident victims) of the surfaces of the microvilli of this tissue at a higher magnification did not revealany bacteria or protozoa. Bar, 50 p.m.

FIG. 2. TEM of a section of the same ruthenium red-stained preparation seen in Fig. 3 showing the long and well-developed microvilli andthe extensive fibrous glycocalyx that surrounds these finger-like projections. Bar, 0.1 p.m.

FIG. 3. TEM of a section of a ruthenium red-stained preparation of a biopsy taken from an ileal conduit 40 days after surgery. Therelocation of this tissue and its use as a urine conduit altered the surface structure of the tissue; the microvilli became shorter and degeneratedin their organization. This effect was seen throughout this particular sample, but was not seen in all patients. Bar, 0.1 ,um.

VOL. 48, 1984

on June 11, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

ItiI

_

*._

___ _f . ._._JF t !_ _s

s

t_ ',f~,

IIS-

%, ^-V

. .4D

-a

a,M

IT Is*h. -A

N N%ItA'*P

I

C

-t.4

.e ~,

'a

A-il4

*," _-

o-

r-<49

-t-Iv -bi

.*I414.'

'.4k .., .4t I.p"4 .,,A

.ll.- I: 5,16. -1

FIG. 4. TEM of a section of a ruthenium red-stained preparation of a stoma sample taken 2 weeks after surgery. During this period ofintensive postoperative antibiotic therapy, most of the surfaces of keratinized epidermal cells were very heavily colonized by gram-positivecocci. Bar, 1 p.m.FIG. 5. TEM of a section of a ruthenium red-stained preparation of mucus recovered with urine from a deep site near the conduit-ureter

junction of an ileal conduit 13 days after surgery. At this stage in postoperative antibiotic therapy, this material was heavily colonized by amixed population of yeasts and gram-positive cocci, and this colonization followed an earlier stage in which only yeasts were found in themucus from the deep sampling site. Bar, 1 ,um.

1162

. r

Ww'.Ilo

A

-: V)%,r 111#A.

I

I

*.bp

%low

-14011L.AIt

.1

. :.. ,- I J"X 4-- >-rt,

.'-i-

i. i

'c .., ,

1,o .

-.wk

-AkWI A&

OPW

.;. "!017...-WAW,-6 .01

.. I

-imm -s-o -4w, 71W.ll -

on June 11, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

A,i%1- If -

A.',s ,# 4. 0, e4s t

kV0-s + #+

>{ 8 :aso+

- :r' -

>f t

Jr#r

''4-S,' T

i*istr. -^*.,4 e.u 1 t ,*

FIG. 6. SEM of a critical-point-dried, thiocarbohydrazide-treated preparation of a deep-site biopsy of an ileal conduit 40 days aftersurgery. A huge ball of microbes occupied the mucus at the tissue surface, and this mass was composed of large numbers of spherical bacterialcells (arrows), smaller numbers of large, spherical yeast cells, and occasional hyphal elements of the colonizing yeasts. Careful examination ofthe tissue surface (T) at a higher magnification revealed no adherent bacterial or fungal cells. Bar, 5 p.m.

FIG. 7. TEM of a section of a ruthenium red-stained preparation of mucus recovered with urine from a deep site in the ileal conduit of a pa-

tient who had suffered six incidents of pyelonephritis in the 16 years after his surgery. Note that both gram-positive and gram-negativebacterial cells grew in glycocalyx-enclosed microcolonies within this mucus. Bar, 1 pLm.

1163

- -1

on June 11, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

cial areas of the conduit during the first 10 to 13 days aftersurgery was heavily colonized, initially by yeast cells andthen by a mixture of yeasts and gram-positive cocci (Fig. 5).SEM of mucus specimens showed extensive microcoloniesof both yeasts and cocci. SEM of deep and superficial biopsyspecimens often showed huge aggregates of bacteria andyeasts in mucus separated from the tissue surfaces (Fig. 6),although areas not covered by mucus were entirely devoid ofadherent bacteria (Fig. 6). The very large microbial aggre-gates seen in the mucus were composed of both coccoidbacteria (ca. 1 p.m) and spherical and hyphal yeast cells (ca.3 p.m). Thus, both TEM and SEM showed that the tissuesurfaces of the ileal conduit are not colonized by bacteria oryeasts in patients examined from 1 day to 16 years aftersurgery.As time after surgery increased and as patients were

removed from postsurgical antibiotic regimens, a large vari-ety of pathogenic bacteria (Table 1) was commonly found at>106 cells per ml of urine. A large proportion of thesepathogenic organisms were gram-negative species (Table 1),and direct examination of urine specimens by TEM showedlarge numbers of gram-negative bacteria in glycocalyx-en-closed microcolonies in the fibrous mucus matrix. Onepatient (number 13 in Table 1) had had six incidents ofpyelonephritis, caused by Escherichia coli, which had beencontrolled by extensive antibiotic therapy, and the urinespecimens taken from his ileal conduit 16 years after surgerystill showed a preponderance of gram-negative bacteria inextensive microcolonies in the fibrous mucus matrix and asmaller number of gram-positive cocci (Fig. 7).

DISCUSSIONPatients who have undergone cystectomy for bladder

cancer are at a very high risk of acquiring upper urinary tractinfections, because the microbiological barrier function ofthe autochthonously colonized urethra (11, 12) and thedefense mechanisms of the normal bladder (9) are no longerprotective and because ileal conduits cannot be constructedwith valves to prevent reflux through the ureters to thekidneys (22). Surgeons note (22) that ileal conduits of anoptimum length (6 to 8 in. [ca. 15.2 to 20.3 cm]) conducturine satisfactorily and offer some protection against ascend-ing colonization and infection. Nevertheless, the majority ofpatients with urinary tract diversion develop bacteriuria and,often, obstruction of the stoma (21), loop stenosis (10),pyelonephritis (13), and renal deterioration (14).

It is known that the human ileum is not heavily colonizedby bacteria, and this observation was confirmed by ourdirect electron microscopic observation of very large areasof the surfaces of ileal samples from patients either treated ornot treated with antibiotics. Although the mucus blanket waslargely retained by novel handling (20) of these specimens,neither bacteria nor protozoa were seen on the tissues or inthe mucus. This lack of an autochthonous microbial floracontrasts with th.- heavily colonized ilea of ruminants (7) androdents (20). Because natural urine conduits such as thehuman urethra are colonzied by autochthonous bacteria (12)and because these organisms are avidly adherent to uro-epithelial cells (4, 5, 11, 18), we anticipated that bacteriaentering the stoma would readily colonize the surfaces of themicrovilli of the ileal conduit. However, a very thoroughexamination of the ileal tissue surfaces obtained from pa-tients 1 day to 16 years after surgery showed that they werevirtually devoid of adherent microorganisms. The flowingurine, on the other hand, contained large numbers of yeastcells during the initial 1- to 10-day period of intensive

postoperative antibiotic therapy and a mixed population ofyeasts and gram-positive cocci during the 10- to 13-daypostoperative period. Despite the presence of this vigorousluminal population of yeasts and gram-positive cocci anddespite the avid adhesion of gram-positive cocci to epithelialcells at the stoma, these organisms did not develop anadherent microbial population on the tissue surfaces of theileal conduit.The mucus present in the conduit is produced primarily by

the ileal goblet cells, although it is likely that the Tamm-Horsfall glycoprotein secreted by renal tubular cells is alsopresent. The latter has been shown to adhere to uropatho-genic bacteria which have a mannose-sensitive adhesin andto block the adherence of these organisms to uroepithelialcells (16, 17). However, bacteria which have a mannose-resistant adhesin have been found to adhere to uroepithelialcells regardless of the presence of the Tamm-Horsfall uri-nary mucus (17) or of mucopolysaccharides which coat thesurfaces of uroepithelial cells (18). In the present study, alarge variety of uropathogenic bacteria, some of whichcontained mannose-resistant adhesins or mannose-sensitiveadhesins or both, adhered to mucus within the conduit anddeveloped into dense microcolonies even in patients whohad been treated with antibiotic therapy for various periodsof time. An earlier study by Needham et al. (15) alsodemonstrated the presence of large numbers of uropatho-gens in the urine of patients with ileal conduits.Because the urinary tract diversion patient has an open

urinary tract system, with reflux between the conduit andthe kidneys, we would expect a rate of pyelonephritis evenhigher than that observed. The outward flow of the mucus,coupled with its capacity to bind bacterial cells, may contrib-ute to this apparent protection and may account for thefailure of luminal microorganisms to colonize the ileal tissuesurfaces. Other factors, such as urine pH, the lack of specificreceptor sites, and local immune factors, may enhance thisputative protective function of the mucus.

This pattern of microbial colonization of the ileal conduitsuggests that the principle of competitive exclusion (6) couldbe used to develop a vigorous autochthonous bacterialpopulation both in the mucus-filled lumen and on the tissuesurfaces of the ileal conduit and thus to preclude ascendingpathogenic colonization. We shall examine the early postop-erative inoculation of the ileal conduit with autochthonousbacteria (e.g., Lactobacillus spp. of vaginal origin) in ananimal model to assess the efficacy of a vigorous adherentnonpathogenic bacterial population in the prevention ofpathogenic bacterial invasions of-the upper urinary tract.

ACKNOWLEDGMENTSThis work was supported by the Kidney Foundation of Canada

and by a fellowship awarded to G.R. by the Alberta HeritageFoundation for Medical Research.The capable technical assistance of Joyce Nelligan, Ian Lee, and

Sheila Costerton is gratefully acknowledged, as is the managerialassistance of Lee Watkins.

LITERATURE CITED1. Bergman, B. 1978. Studies on patients with ileal conduit diver-

sion with special regards to infection. Scand. J. Urol. Nephrol.Suppl. 47:1-32.

2. Bishop, R. F., E. Durham Smith, and M. Gracey. 1971. Bacterialflora of urine from ileal conduit. J. Urol. 105:452-455.

3. Bricker, E. M. 1956. Substitution for urinary bladder by use ofisolated ileal segments. Surg. Clin. N. Am. 35:1117-1130.

4. Bruce, A. W., R. C. Y. Chan, D. Pinkerton, A. Morales, and P.

1164 CHAN ET AL.

on June 11, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

MICROBIOLOGY OF ILEAL CONDUITS 1165

Chadwick. 1983. Adherence of Gram-negative uropathogens tohuman uroepithelial cells. J. Urol. 130:293-298.

5. Chan, R. C. Y., and A. W. Bruce. 1983. The influence of growthmedia on the morphology and in vitro adherence characteristicsof Gram-negative urinary pathogens. J. Urol. 129:411-417.

6. Chan, R. C. Y., A. W. Bruce, and G. Reid. 1984. Adherence ofcervical, vaginal and distal urethral normal microbial flora tohuman uroepithelial cells and the inhibition of adherence ofGram-negative uropathogens by competitive exclusion. J. Urol.131:596-601.

7. Cheng, K.-J., R. T. Irvin, and J. W. Costerton. 1981. Autoch-thonous and pathogenic colonization of animal tissues by bacte-ria. Can. J. Microbiol. 27:461-490.

8. Costerton, J. W. 1980. Some techniques involved in the study ofadsorption of microorganisms to surfaces, p. 403-432. In G.Bitton and K. C. Marshall (ed.), Adsorption of microorganismsto surfaces. John Wiley & Sons, Inc., New York.

9. Cox, C. E., and F. Hinman, Jr. 1961. Experiments with inducedbacteriuria, vesical emptying and bacterial growth on the mech-anism of bladder defense to infection. J. Urol. 66:739-747.

10. Hendren, W. A., and G. A. McLorie. 1983. Late stricture ofintestinal ureter. J. Urol. 129:584-590.

11. Marrie, T. J., J. Lam, and J. W. Costerton. 1980. Bacterialadhesion to uroepithelial cells: a morphologic study. J. Infect.Dis. 142:239-246.

12. Marrie, T. J., C. A. Swantee, and M. Hartlen. 1980. Aerobic andanaerobic urethral flora of healthy females in various physiolog-ical age groups and of females with urinary tract infections. J.Clin. Microbiol. 11:654-659.

13. Mebust, W. K., J. D. Foret, and W. L. Valk. 1969. Fifteen yearsof experience with urinary diversion in myelomeningocele pa-tients. J. Urol. 101:177-182.

14. Middleton, A. W., Jr., and W. H. Hendren. 1976. Ileal conduitsin children at the Massachusetts General Hospital from 1955 to

1970. J. Urol. 115:591-595.15. Needham, R. E., M. M. Smith, and J. M. Matsen. 1970. Differ-

ences in the bacteriology of intestinal loop urinary diversions. J.Urol. 104:831-833.

16. 0rskov, I., A. Ferenz, and F. 0rskov. 1980. Tamm-Horsfallprotein or uromucoid is the normal urinary slime that traps type1 fimbriated Escherichia coli. Lancet i:887.

17. 0rskov, I., F. Orskov, and A. Birch-Andersen. 1980. Compari-son of Escherichia coli fimbrial antigen F7 with type 1 fimbriae.Infect. Immun. 27:657-666.

18. Reid, G., H. J. L. Brooks, and D. F. Bacon. 1983. In vitroattachment of Escherichia coli to human uroepithelial cells.Variation in receptivity during menstrual cycle and pregnancy.J. Infect. Dis. 148:412-421.

19. Richie, J. P., D. G. Skinner, and J. Waisman. 1974. The effect ofreflux on the development of pyelonephritis in urinary diver-sion: an experimental study. J. Surg. Res. 16:256-262.

20. Rozee, K. R., D. Cooper, K. Lam, and J. W. Costerton. 1982.Microbial flora of the mouse ileum mucous layer and epithelialsurface. Appl. Environ. Microbiol. 43:1451-1463.

21. Shapiro, S. R., R. Lebowitz, and A. H. Colodny. 1975. Fate of 90children with ileal conduit urinary diversion a decade later:analysis of complications, pyelography, renal function andbacteriology. J. Urol. 114:289-295.

22. Skinner, D. G., and J. P. Richie. 1979. Ureterointestinal diver-sion, p. 2211-2230. In J. H. Harrison, R. F. Gittes, A. D.Perlmutter, T. A. Stamey, and P. C. Walsh (ed.), Campbell'surology, 4th ed., vol. 3. The W. B. Saunders Co., Philadelphia.

23. Spence, B., W. Stewart, and A. S. Cass. 1972. Use of a doublelumen catheter to determine bacteriuria in intestinal loop diver-sion in children. J. Urol. 108:800-801.

24. Williams, D. I. 1972. Urinary diversion by sigmoid conduit, p.294. In R. Scott (ed.), Current controversies in urologic manage-ment. The W. B. Saunders Co., Philadelphia.

VOL. 48, 1984

on June 11, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from