Bacterial biofilms and catheters : A key to understanding bacterial strategies in catheter-associated urinary tract infection

J C NICKEL, JW CosTERTON. Bacterial biofilms and catheters: A key to understanding bacterial strategies in catheter-associated urinary tract infe c t ion_ Can J Infect Dis 1992;3(5):261-267. Des pile major technological improvements in catheter drainage systems. the indwelling Foley catheter remains the most common cause of nosocomial infection in medical practice. By approaching this common complicated urinary tract infection from the perspective of the biofilm strategy bacteria appear to use to overcome obstacles to produce bacteriuria. one appreciates a new understanding of these infections . An adherent bionlm of bacteria in their secretory products ascends U1e luminal and external surface of the catheter and drainage system from a contaminated drainage spigot or urethral meatus into the bladder. If the intraluminal route of bacterial ascent is delayed by strict steiile closed drainage or addition of internal modifications lo the system. U1e exiraluminal or u rethral route assumes greater importance in U1e development of bacteriuria. but takes significantly longer. Bacterial growth within these thick coherent bioftlms confers a large measure of re lative resistance to antibiotics even U1ough the individual bacte1ium remains sensitive, thus accounting for the failure of antibiotic U1empy. With disruption of the protective mucous layer of U1e bladder by mechanical irritation. U1e bacteria colonizing the caU1eter can adhere to the bladder's mucosal surface and cause infection. An appreciation of U1e role of bacterial biofilms in these infections should suggest future directions for research that may ultimately reduce the Iisk of catheter-associated infection.

T HE FORMATI01 OF BACTERIAL BIOFILMS ON SURFACES appears to be a universal bacterial strategy for survival in bolh natu re and d isease (1).Recen t evidence indicates lhal bacterial bio fil ms might a lso be invo lved in biomaterial-related bacterial infections (1).Th is recent im proved u n der landing of U1e palhogen esis of catheter-a sociated urinary tract in fections may h elp explain why calheteri zation remains U1e most comm on cause of nosocomial infection in med ical practice (2).despite major technoloaical chancres in catheter material.design and collectincr sy terns (3).Over U1e past decade lhe aulhors• research group has extensively tudied the role of lhis bacterial mode of growU1 in the etiology of biomaterial -related urinary tract inf ction.In lhis review lhe data generated by the aulhors• study of bacterial biofilm growth are summarized and synlhesized as they pertain to an understanding of the palhogenesis.prevention and treatment of catheterassociated infections.

THE ASCENDING BACTERIAL BIOFILM
It is generally accepted U1al bacteria gain entrance to lhc bladder from retrocrrad intraluminal ascent of organisms from contaminated op n collection vessels in lhe early days (4).from lhe collecting bag or disconnected catheter drainage tube junction since the introduction of the closed urinary drainage system (5)(6)(7)(8)(9).and cxlraluminally from a colon ized urethral meatus if strict sterile closed drainage is maintained (6. 7. 1 0. 1 1).The retrograde intraluminal spread of organisms was thought to arise from bacteria migrating up the urine column.reflux of infected urine into the bladder.or bacterial transport involving air bubbles (12).However. the addition of vents.flutter valves.air locks and drip chambers to drainage systems did not s ignificantly alter the rates of calheler-a sociatcd urinary tract infection (5.6.8).Observations in animal models of the clo eel catheter drainage ystem have d isclosed that bacteria fom1 lhick coherent b iofilms adherent to experimentally contaminated drainage spouts extending proximally into lhe drainacre bag and subsequently into the calheter (13.14).Employing a bacteriologically stre sed animal model of short term caU1eterization (fewer than seven clays).contamination of the dra inage spout or accidental disconnection of lhe drainage lube resulted in bacteriuria wi th in a short lime (32 to 48 h).If a strict sterile closed d ra inage system was maintained and lhe urethral meatus-catheter junction was inoculated.U1e extralum inal route wou ld assume greater importance in lhe development of bacteriuria: however.this palhway was considerably lower (72 to 168 h).These findings regarding lhe relative importance of the intraluminal and extraluminal periureU1ra l routes were confirmed in furlher animal model ludies employing a microbicidal hurd le or barrier in the oullel tube of the drainage bag (15. 16).Intraluminal colonization by exogenous bacteria was precluded in the majority or animals during the eight day period.and in lhe 25% where intralu m inal colonization occu rred it wa s icrn ificanlly delayed (7.2± 1 days).However.downstream colonization of drainage systems, by mixed populations of bacteria lhat had migrated in to lhe b la dder via U1e exlraluminal periurethral route.developed in a lmo t one-half of Lhe anin1als.Sim ilar bacteria l biofilms have be n described adh erent to urinary d rainage bags (17).Fo ley catheters recovered from patients with catheter as ocialed infections (18.19) and even caU1eters removed from patients wil11 no bacteriu ria (20) (Figure 1).Examination of 10 urinary caU1eters associated with urinary tract infections by scanning electron microscopy confirmed that lhe bacteria associated wilh lhese infections grew in glycocalyx-enclosed b iofil ms on U1e s u rface of lhe calhelers (19) .
The bacterial popu lations demonstrated a heterogeneity U1at was not evident from culture re ults.and it was demonstrated U1al only a small propor tion of lhe microorganisms -includ ing funcri -identified morphological ly by scanning or transm ission electron m icroscopy were recovered by routine culture meU1od .
Emp loying an in vitro system. it could be clearly demonstrated lhal U1e bacteria were ascending lhe surface of lhe calheter in a ere ping coherent biofilm contain ing bacteria l cells in U1eir secretory p roducts or glycocalyx (21).In Lh e absence of antibiotics it appeared Lhal U1e ascending bacteri a l biofil m was moving by two mechanisms: rapidly dividing bacterial cells spread ing along lhe caU1eter surface wilhin the glycocalyx materia l of the biofilm.and plankton ic or floating bacterial cells wilhin the urine column leapfrogging ju t ahead of lhe adherent biofilm.perhaps assisted by lhe turbu lence caused whe n lhe u ri ne flow meets lhc biofilm front.Th is •saltato ry• bacterial movement may al low some bacteria t.o e lab lish adh erent microcol-onies a head of th e ascending biofilm .which expand and coalesce wiU1 the main bacterial aggregate into the ascend ing coh e rent biofilm (21 ) (Figure 2).

BLADDER DEFENCES AND THE BACTERIAL BIOFILM
Planktonic bacte ria being released from the b iofilm a dh e re nt lo the Foley cath eter can be easily demonstrated in aspirated urine cultures; h owever.al this point the bacteria are colonizing only the cath ete r s urface and have not yet caused cystitis.The int.rave ical segment of U1e Foley catheter eventually becomes covered with a much thicker colonizing ad herent bacte rial aggregate enclosed within the bacterial slime matrix.This macroscopic bacterial biofilm (Figure 3) can create flow kinetic problems by partially blocking cathete r islets and reducing the tubular diameter of U1e catheter lum en.This biofilm-induced disruption of effective urin e flow may increase the volume a nd perhaps pressure of the residual urine U1al is always present in catheterized bladders (22) .Planktonic bacteria shed from the colonized ca theter inhabit this intravesical urine.but the ste p from asymptomatic bacteriuria lo symptomatic catheter-associated cystitis involves actual bacterial adherence lo the bla dder s urface (23).
A thin blanket. of mucus or glycosaminoglycan coals the bladder mucosal surface (24) (Figure 4) and ap pears to inhibit bacterial a dherence to the uroepiU1elium.Th e indwelling Foley caU1eter appears lo disrupt thi bladder mucus or glycosan1inoglycan layer (25) a nd causes mechanical irritation and even erosion of the bla dder mu cosa, exposing surfaces that.a llow bacterial adherence.Particular strains of bacteria are more a dapted lo producing symptomatic caU1 eter-associat.edcystitis.The a bility of strains of Escherichia coli lo establish mucosal inflammation seems to correlate wilh the presence of fimbriae (26).whi ch act as adhesions lo uroepithelial cells (27) .When s uffi cie nt damaO"e is ca u sed to the bladder defen ce mecha n ism.and U1e bacterial inoculum is of critical size and paU1oge nicily.bacterial adherence to U1e mucosa or exposed mu cosal cell recep tors occur .a nd the patient deve lops symptomatic cysti tis .Patients who proO"ress from asymptomatic ca U1eter-associat.edbacteriuria lo symptom atic cath e te r-associated infec tion complain of bladder irri labili ly, pain and fever secondary lo bacterial a dhe ren ce lo. or invasion of.U1e damaged bladder mucosa.or lo subsequent upper tract.infec tion .

ANTIB IOTIC RESISTANCE IN CATH ETER-ASSOC IATED
BACTERIAL BIOFILMS Histori cally.physicians have used antibiotics boU1 lo control and prevent catheter-associated infec tions with va1ying and conl1icting degrees of success.In a n in vitro model it was shown U1al bacterial growth within U1ick biofilms adherent.to urinary cath eter material in an arti fi cia l urin e milie u confe rred a m eas ure of antib iotic resistance on the sessil e bacterial cell within the bio- fi lm (28.29) (Figure 5).In U1is experimen t. disks of urinary caU1 ele r material were exposed to the fl ow of a rtifi cia l urine conta ining cells of Pseudomonas aerug inosa forming a thi ck adherent biofilm composed of these bacteria and U1eir exopo lysacch ari de products.After U1is colonization .sterile urin e conta ining Lobramycin 1000 pg/mL was l1ovved past the established biofilm.and a s ign ificant propo rtion of the bacterial cells vviU1in U1e biofilms were found still viable after 12 h of expos ure to this very high oncentration of a n1in oglycoside an tibiotic.Planktonic or floa ting cells taken from the Lest system just before exposure of th e biofilm to U1e a ntibiotic were completely killed by 50 pg/mL Lobramycin.However.this resistance did not transfer to U1 ese planktonic floa ting cells in U1e urine, and 100% bactericidal ac tivity of high dose tobramycin was obe rved in tili population.The minimal inhibitory conce ntrations of tobramycin for cells taken from U1e seeding cu ltures before colonization of U1e ca tileter material a nd tilat for dispersed surviving cells recovered dircclly from tile tobramycin-treated biofilm we re found to be tile sam e (0.4 pg/mL).This indicates that growth within th ick a dhe rent biofilms confers a relative measure or a nlibioli c resista n ce on cells of Ps aerug[nosa.possibly U1e resu lt of poor a ntibioti c penetration of U1e biofilm matrL...: (29).Using physioloO"ical techniqu es to study tile metabo lic activity (radiorespiromelri c assay) of uro-patho~enic bacte ria gro\ving in biofilms in urine-con -taining systems.U1e present authors and colleagues have been able to determine U1at U1ere is a significant reduction in mineralization activity (evolution of carbon dioxide) during antibiotic challenge (30).Respiratory activity ceased almost immediately for planktonic samples.but residual metabolic activity.albeit very much reduced.was detected in sessile caU1eter-associaled bacte1ial biofilms exposed for more U1an 18 h to high antibiotic concentrations.Antibiotic-stressed. caU1eter-associated bacteria transferred to a post exposure enrichment.broth showed an ability to re-establish respiratory activity not observed for the planktonic samples.This decreased metabolic activity of bacteria deep wiUlin the biofilm during periods of microbicidal threat may also h elp to explain the relative resistance of tilese bacte1ia to treatment and the observation that catheter-associated bacteria can survive exposure to antibiotics in U1e clinical setting (31).
When antibiotics are added to an in vitro system assessing bacterial biofilm movement.following a lag phase prolonged for almost 12 h.the biofilm can ascend the catheter surface (21).However, in the experimenta l set.Ung U1e biofilm movement was retard ed to a rate of a bout 0.2 to 0.3 em/h.compared to bacterial bio!Hm ascens ion along U1e same catheter surface against an identical flow of artificial urine without antibiotics (1 to 2 em/ h) .Bacterial biofilm ascent along tile external and luminal surface of U1e catheter appears t.o be significantly s lower in the presence of antibiotics because tileoretically antibiotics appear to negate any saltatory planktonic movement of the biofilm.The antibiotic also reta rds the rapid division of bacteria in tile surface areas of t.he biofilm by either killing U1em or causing a decrease in metabolic activity: therefore.the biomovemen t is dependent on U1e division of bacterial cells deep wifuin the protected matrix of the biofilm (21).The present authors• clinical ex.rperience confirms U1ese experimental data that cafueter-associated infection remains a problem despite fue preventative use of antibiotics (31).
In animal studies.incorporating a microbicidal hurdle or barrier to intraluminal ascending bacterial migration can preclude, for a short time.fue intraluminal ascent of fue bacterial biofilm; however.U1e extraluminal or periurefural surface U1en becomes t he predominant route of bacterial entry into U1e bladder (15. 16).OU1er studies have similarly shown U1at the a ddition of antimicrobial solutions to tile drainage bag on ly precludes the intraluminal route of infection (7.32.33).It would U1erefore be expected U1at local a ntimicrobial treatment of U1e urethra l meatus should decrease U1e rate of catheter-associated infection in pa tie nts with a properly maintained closed urinary dra inage system.but clinical studies do not confirm th is (34).
The use of antibiotics for U1e prophylaxis of infection has bee n further examined .In a study designed to examine in detail the local effects of different doses of an a ntibiotic on the establishment of bacterial biofilms in catheter-associated infections. it was discovered that extremely high doses of amdinocillin can destroy and sterilize an Escherichia coli biofilm colonizing U1e Foley caU1e ter in a rabbit model (35).Unfortunately. the doses required are not clinically practical.
At lowe r .more clinical closes the bladder wall fo llowed by the bla dder urine could be cleared of infecting organisms .Perhaps the natural host defence mecha nis ms of the bladder mucosa and mine work synergis tically with the antibiotic to clear these bacterial populations.
The experimental data suggest that alU1ough antibiotics cannot be used to clear bacterial biofilm completely from U1e surface of the Foley catheter.they do s low down the ascent of the bacterial biofilm and can eradicate.for at least a short time.the bacterial populations adherent to the bladder mucosa that create the symptom s of cystitis.Clinical studies h ave s ubseq uen tly a lso s u ggested U1at antibiotics used in s hort term ca thete rization may reduce the serious seq u elae asociated with catheter-related infection (36): however.it is obvious that an tibiotics would be counterprodu ctive in any form of long term Foley caU1eterization beca u se of the high possibility of emergence of specific a ntibiotic-resistant bacterial strains (31) .The a uU1ors• tudies in postoperative short term catheterization (two days) (37) suggest that antibiotics and/ or expensive bacterial hurdles a re not indicated.but rather strict maintenance of a sterile closed system is a ll U1at is required.Unfortunately, the definitive data re<l'arding th e use of antibiotics in catheterization lasting between two and seven days are not available at this lime.

CRYSTALLIZATION WITHIN THE CATHETER-ASSOCIATED BIOFILM
If ilie bacterial biofilm adherent to the Foley ca U1eler is inhabited by urease-producing bacteria such as Proteus species.s usceptible patients can d evelop catheter encrustatio n.This can cause major difficulties in patients requiring long term Foley catheterization.The urease produced by these bacteria hydrolyzes urea lo ammonia and ra ises the pH of the urine .particula rly in U1e microenvironment or th e bacterial biofilm on the catheter surface.a llowing calcium .phosp hate a nd struvite crystals to form within U1e urine a nd.more importantly.within the biofilm on the catheter s urface (38) (Figure 6) .The bacterial biofilm m atrix entraps th e crystals on the caU1eter surface.which aggregate a nd coalesce to become macroscopic e ncrustation s.These encru stations a re produced in a manner a n a logo us lo deposition of struvile calculi wiU1in the kidn ey lo crea te a slagh om calc ulus.On the caU1eter.slruvile c1ystals.matrix glycoprotein and uromu coid compounds can and will eventu a lly block U1e is lets and lum en of th e catheter (19. 39).Irrigatio n of catheters with acid solu - lions is som etimes used to prevent and dissolve encrusta tions.but th e actu al effi cacy and h azards of U1is procedure perhaps outweigh its benefits.There appears lo be a high er risk of bacteremia and sepsis in patients with encrusted caU1 eters secondary to both blockage a nd U1e clini cal manipu lation and irriga tion required to clea r t h e blockage.Th e bladder surface glycosaminoglycan or mu cou s layer is extrem ely acid sensitive (24).an d irrigatio n f1uids may in fact disrupt further U1e bladder mu cosal surface, allowing for s ubsequent bacte ri a l ad h esion a nd symptomatic cystitis.There is some evidence.U1ough very limited.that U1e composition of the cathe ter an d s urface characteristics m ay inhibit U1e deposition of struvite.There is perhaps less encru station associated with silicone versus latex caU1eters (40).but not between s ilicone and hydrogel-coated la tex catheters (41).There a lso appears lo be significant individua l patient variation between so-called 'blockers ' and •non bloc kers' (42) .likely secondary lo fueir inhe rent urine composition.Some individuals a ppear to block their ca U1eters repeatedly.while others do not.even when infected wiU1 the same species of organism.

FUTURE DIRECTIONS
ll is hoped U1al furth er advances in m ed ical technology will a llow modification of ca U1eterizaUon procedures.duration a nd need for catheterization.and provide improvem ents in the design of cailieter urina1y drainage system s.Studies have demonstrated U1at the internal lumin al route of ca U1eter-associaled infections ca n be almost completely negated.at leas t for a short Lime .by the use of a strictly maintained sterile closed drainage system .with possible a ddi tion of a bacterial ban•ier or hurdle.The design of such a device or mechan ism to slop ili e periurethra l route would be very helpful.Biomaterial research is an exploding n ew science .and research musl continu e vvilh the e n ew materials in respect to mu cosal biocompatibilily a nd effectiven ess in redu cing bacterial biofilm a ltachmenl.It is anticipated that n ew biomaterials will eventu a lly reduce bacterial adh erence a nd biofilm form ation a nd s ubsequenUy decrease U1 e rate of caU1eter-associated infection.New antibiotics being d eveloped m ay be able to penetrate th e bacterial biofilm a nd may be more effective in this and other prosthesis-related infections.Further studies are requ ired to rationalize U1e use of

Figure 1 )
Figure 1) Scanning electron micrograph of the luminal swjace of a colonized catheter demonstrating the bacterial cells within coherent biojl.lmsalmost completely bw•ied in a condensed sl.imelilce glycocalyx.In some areas the glycocalyx has been lost dur ing preparation.disclosing I he bacterial nature of this Lh iclc biqfilm (lower section) as well as the surface of the catheter (lower left comer).Bar 5 J.lm. (Reproduced with pem1issionji-om r~{erence 19)

Figure 4 )IFigure 5 )
Figure 4) Scanning electron micrograph clearly showing a specially prepared antibody-stabilized mucous layer covering the uroepithe liLLn1 as a continuous blanket.This mucus or glycosaminoglycan layer acts as a barrier to bacterial adherence.Mechanical or microbial irritation dismpts this protective layer.allowing bacterial adherence and SLLbsequen t symptomatic cystitis.The bladder mucosa can be observed through a crescent -shaped defect in the mucus layer.(x-+2)

Bacterial biofilms in catheter infection Figure 6)
Scanning electron micrograph of the swface of an encrusted catheter shows rod-s haped and coccoid bacte1ial cells among dehydration condensed residue of the enveloping ex• opolysaccharoid g lycocalyx and uromucoid •slime•.Discrete crystals are noted within this coherent biojilm.The association of bacteria.slime and crysta ls leads to eventua l obstruction in the catheter.Bar 5 ,,m.(Reproduced w ith permission from r~{er•