Catheter-Associated Urinary Tract Infections: Epidemiology, Pathogenesis, and Prevention WALTER

E. STAMM,

M.D.,

Seattle,

Washington

Catheter-associated urinary tract infections (UTI,) remain the most common nosocomial infection. Although usually benign, UTI, cause bacteremia in Z-4% of patients and have been associated with a case fatality rate three times as high as nonbacteriuric patients. Risk factors for UTI, identified in multivariate analyses include increasing duration of use, female sex, absence of systemic antibiotics, and disconnection of the catheter-collecting tube junction. Recent studies suggest that most episodes of low colony count bacteriuria ( 102-lo4 cfulml) rapidly progress to high ( r 105/ml) colony counts within 24-48 hours. In persons with long-term catheterization, bacteriuria inevitably develops and the infecting strains change frequently. In this setting, Proteus and Morganella species produce catheter encrustations and persistent bacteriuria. Routes of bacterial entry have been well defined and differ by gender, with the periurethral route predominating in women and the intraluminal route in men. Growth of bacteria in biofilms on the inner surface of catheters promotes encrustation and may protect bacteria from antimicrobial agents. Bacterial virulence factors have not been well characterized in UTIc, but fimbrial adhesins have been associated with bacterial persistence in the catheterized urinary tract, and urease production has been associated with stone formation and catheter encrustation. Recent efforts to prevent UTI, have focused mainly on preventing bacterial entry to the urinary tract or eradicating bacteriuria after its onset and have been largely unsuccessful. Systemic antimicrobials, sealed tubing and catheter junctions, silver ion-coated catheters, and antiseptics in the collecting bag have all been effica-

Harborview

Medical Center,

University

of Washington,

Division of Infectious Diseases ZA-89, Harborview 9th Avenue, Seattle, Washington 98104.

Seattle, Washing-

Medical Center,

September

16,

325

1991

cious in one or more controlled trials. Failure to stratify patients by major risk factors, especially gender, antimicrobial exposure, and catheter duration, makes interpretation of many trials difficult. Further research in the areas of innovative catheter system design, bacterial-host epithelial cell interaction, and targeted antimicrobial prophylaxis seem the most likely approaches to controlling UTI, in the future.

atheter-associated urinary tract infections (UTI,) remain the most common of all nosocomial infections, accounting for approximately 40% of infections in most hospitals. 111. In the 198Os, considerable research has furthered our knowledge of the epidemiology, pathogenesis, and prevention of these infections. These advances will be reviewed in this article.

C

EPIDEMIOLOGY In the absence of national reporting of nosocomial infection data, it is difficult to assess trends in incidence or prevalence of a given nosocomial infection. Within the National Nosocomial Infections Surveillance (NNIS) hospital reporting system, there has been no apparent change in the proportion of all nosocomial infections attributable to the urinary tract [21. However, the NNIS hospitals are not a representative sample of United States hospitals and have changed their composition considerably over the last decade! Interestingly, the prevalence of UTI, in prospective studies between 1966 and 1990 reveals a general trend toward decreasing occurrence of bacteriuria in study populations of a similar nature (Table I> [3-lo]. All of these studies were done in general hospital populations at large academically affiliated hospitals. Although such a comparison of studies done in different locales by different investigators at different times is far from conclusive, there is an apparent trend toward a reduced prevalence of bacteriuria over the time period shown. If real, many factors could account for this decrease, including increased use of systemic antimicrobials in catheterized patients, decreasing lengths of hosThe American

Journal

of Medicine

Volume

91 (suppl3B)

38-658

CONFERENCE

ON NOSOCOMIAL

INFECTIONS

/ STAMM

TABLE I Prevalenceof Catheter-AssociatedUrinary Tract Infections: Prospective Studies, 1966-1990 Kunin and McCormack [31 Finkelberg [41 Garibaldi et a/ El Warren eta/ [71 PlattetalL Thompson et al [81 Johnson et a/ [IO1

1966 1969 1974 1978 1983

23%

21% 23%

17% 190; 10%

%i

pitalization and catheterization, and perhaps improved infection control efforts. Likewise, it is difficult to be certain whether the relative frequency of individual pathogens causing nosocomial UTIs has changed in the last decade. However, data from the NNIS system support a probable increase in candidal UTIs over the last decade [21. Increases in enterococcal and candidal infections of the catheterized urinary tract have been noted in individual hospitals as well. A number of prospective studies over the last decade have analyzed the major risk factors associated with UTI, [3-101. In contradistinction to previous analyses, logistic regression has been utilized to assess the independence of various potential risk factors [lO,ll]. Four factors have been repeatedly shown in representative studies to be major risk factors for UTI: female gender, duration of catheterization, absence of systemic antibiotics, and catheter care violations. Although other risk factors have been identified in specific populations and individual studies, these four risk factors are of sufficient reproducibility that investigators should adjust analyses for these factors when evaluating new preventive methods for UTI,. Finally, an important study by Platt and colleagues [61demonstrated that UTI, was associated with a case fatality rate nearly three times as high as that found in nonbacteriuric patients. This increased relative risk of mortality in patients with UTI, remained after adjustment for various potentially confounding factors, such as age, severity of illness, hospital service, duration of catheterization, creatinine, or person inserting the catheter. The mechanism accounting for such an increased mortality in catheterized patients would presumably be secondary bacteremia and septicemia. Of the 25 patients who died, two were recognized as having gram-negative sepsis and 10 others were febrile or had a peripheral blood leukocyte count >20,000/mm3. Thus, only a minority of patients had recognized gram-negative septicemia, but other patients may have had unrecognized infection contributing to their deaths. In an attempt to provide further evidence for the hypothesis that catheter-associated bacteriuria con38-668

September

16, 1991

The American

Journal

of Medicine

tributes to mortality, Platt and colleagues WI conducted a randomized controlled trial in which sealed catheter junctions were utilized as an intervention to prevent bacteriuria. It was found that the sealed junction catheters significantly reduced the risk of UT1 in catheterized patients and also reduced the risk of death in the subset of persons not receiving systemic antimicrobials. The implications of this important set of studies are farreaching in that much greater importance is attributable to UTIs associated with catheterization if they truly contribute to mortality via unrecognized septicemia. Further studies to support this hypothesis are needed. Of interest is a recent study by van Deventer and colleagues [131 in which the occurrence of bacteremia and endotoxemia was evaluated in 76 hospitalized patients with gram-negative rod bacteriuria and fever. Approximately half were catheterized patients with nosocomial urinary tract infection. Of those who had suspected sepsis defined on clinical grounds, the majority had positive blood cultures and endotoxemia. Interestingly, however, of those who had fever alone in addition to gram-negative rod bacteriuria, the prevalence of positive blood cultures was 27 (44%) of 62 and an additional four patients were endotoxemic. These results are consistent with the hypothesis that unrecognized bacteremia and septicemia may occur in hospitalized patients with gram-negative rod bacteriuria. Aside from mortality, it is important to emphasize that the sequelae of UTI, remain poorly understood and should be further investigated. For example, criteria have not been established for differentiating asymptomatic colonization of the urinary tract from asymptomatic infection. Of those patients who have catheter-associated bacteriuria and lack symptoms, how many actually have invasive infection? The presence of pyuria and hematuria in this setting or the presence or absence of local or systemic antibody could be studied to determine whether such variables could differentiate colonization from asymptomatic infection. Likewise, localization of infection within the urinary tract remains poorly understood in the setting of catheter-associated infection. How often does bacteriuria in a catheterized patient, despite the absence of symptoms, represent renal infection? Certainly, the latter could have important implications for approaches to therapy. Finally, the incidence of febrile episodes, acute pyelonephritis, and epididymitis due to UTI, remains poorly understood and deserves further evaluation. The best understood sequelae of UTI, remain the most serious ones, bacteremia and death. Prospective studies in the last decade continue to Volume

91 (suppl

38)

CONFERENCE

show that l-4% of patients with UTI, develop bacteremia and that the case fatality rate in such bacteremic patients is 13-30% [14-171. Although clinically recognized bacteremia is thus not common among patients with catheter-associated infection, the large number of patients with catheterassociated bacteriuria make the infected urinary tract the most common source of gram-negative sepsis in hospitalized patients [ 171.

PATHOGENESIS It has long been known that bacteria take one of two routes of entry into the catheterized urinary tract, the periurethral route or the intraluminal route. Studies in the last decade have clarified that there are important differences in route of entry by gender. In women, approximately 70% of episodes of bacteriuria occur through the periurethral entry [181. In most instances, bacteria infecting the urinary tract in women emanate from rectal flora [ 181. Thus, in women the sequence of events in the catheterized urinary tract is similar to that in noncatheterized patients in which fecal strains colonize the periurethral zone and subsequently enter the urinary tract [191. In males, however, the opposite is true in that the majority of infections occur via the intraluminal route 1181. Infecting strains in such cases are, therefore, not of rectal origin and often result from cross-infection. These differences are illustrated in a study by Daifuku and Stamm [la] in which patients in an intensive care unit were prospectively followed from the day of hospitalization through the day of catheter removal. Cultures of the periurethral zone, fecal flora, the catheter, and urinary tract were obtained on a daily basis. Of 35 infectious episodes, 18 occurred in women and 17 in men. Fourteen of 18 women had antecedent rectal colonization with the infecting strain and 12 of 18 women had antecedent urethral colonization. In contradistinction, only 5 of 17 infectious episodes in men were preceded by rectal and urethral colonization. These results have important implications in that preventive measures for men and women may well need to be different. Studies in the last decade have also defined with greater precision the routes and timing of bacterial migration and growth within the catheterized urinary tract. Both human and animal studies have demonstrated that bacteria entering the drainage bag can be found approximately 24-48 hours later in the bladder 120,211. This occurs with great regularity in patients not receiving antimicrobials. On the other hand, periurethral colonization is not followed invariably by bacteriuria, and when bacteriuria occurs it is often greater than 72 hours in September

16, 1991

ON NOSOCOMIAL

INFECTIONS

/ STAMM

duration after the establishment of periurethral colonization [18,22,23]. Interestingly, some patients have persistent periurethral colonization but never develop bacteriuria [231. Finally, once in the bladder, small numbers of microorganisms (for example, 100 cfu/ml) increase to large numbers (>100,000/m1) in less than 24 hours. This was demonstrated most clearly in a study by Stark and Maki 1241 in which sequential cultures were obtained from catheterized patients in an intensive care unit. Almost all patients showed a rapid increase in the number of bacteria or Candida from low colony counts to high colony counts within 48 hours of initial bladder colonization. Studies in the last decade have demonstrated the important role of attachment and growth of bacteria on the inner surface of the catheter in the pathogenesis of UTI, [20,25-281. It has become increasingly clear that two populations of bacteria exist in the catheterized urinary tract, those growing within the urine itself (planktonic growth) and those growing on the surface of the catheter (biofilm growth). The growth of a bacterial biofilm on the urinary catheter progresses according to a well-defined sequence of events: bacteria attach to the urinary catheter, initiate a biofilm form of growth in which sheets of organisms coat the catheter and secrete an extracellular matrix of bacterial glycocalyces in which they become embedded. In addition, host urinary proteins such as Tamm-Horsfall protein and urinary salts are often incorporated into this biofilm growth on the catheter. Subsequently, this material leads to encrustation of the inner surface of the catheter. Examination of catheters by electron microscopy demonstrates this biofilm growth on the inner surface of catheters removed from both patients and animals 125-281. Certain genera of bacteria, particularly Proteeae and Pseudomonas, are highly associated with a propensity for biofilm growth and catheter obstruction 1291. In general, biofilm seen on the inner surface of the catheter is much thicker and of a different nature than that seen on the external catheter surface 1201. On the external surface of the catheter, small numbers of microcolonies are seen adherent to the catheter but a thick well-developed biofilm with glycocalyces is less frequently noted [201. The realization that biofilm growth occurs on catheters has important implications. From the standpoint of prevention, catheter materials that retard bacterial adherence and biofilm growth should be sought 1251. Second, urine cultures obtained from the catheter may not reflect bladder bacteriuria in patients who have organisms in a biofilm on the inner surface of the catheter 1301.

The American

Journal

of Medicine

V&me

91 kuppl3B)

38-6733

CONFERENCE

ON NOSOCOMIAL

INFECTIONS

/ STAMM

Rather, bladder urine may be sterile but organisms from the catheter biofilm may contaminate the aspirated urine culture. Third, biofilms have been demonstrated to retard the activity of antimicrobial agents when compared with activity of the same antimicrobial against organisms growing planktonically [27]. Thus, patients who are treated with antimicrobials for UTI, may fail therapy when organisms are present within biofilms on the inner surface of the catheter. This implies that catheters (particularly those in place longer than a week) should perhaps be replaced if bacteriuria is to be treated. Studies of this issue should, however, be undertaken. In patients without urinary catheters who develop urinary tract infections, it has been demonstrated that a small number of “urovirulent clones” of Escherichia coli produce most serious infections of the urinary tract [19]. These clones are defined on the basis of a limited number of 0, K, and H serotypes and by the presence in these clones of specific virulence genes for P fimbriae, hemolysin, and siderophores. Studies to determine whether similar strains frequently infect the catheterized urinary tract have shown that the urovirulent strains seen in community-acquired UTIs rarely cause bacteremic UTI, [31,321. The diversity of species causing UTI, implies that impaired host defenses are probably more important than specific bacterial virulence factors in the genesis of these infections [331. However, specific bacterial properties may be of importance in some settings. It has been demonstrated, for example, that specific strains of bacteria adhere more avidly than do other strains to either catheter materials themselves or to uroepithelial cells from patients’ bladders [341. In a study by Daifuku and Stamm [341 uroepithelial cells from catheterized patients were incubated in vitro with potential uropathogens and large differences in the ability of specific species to attach to the exfoliated cells of catheterized patients was observed. In addition, the receptivity of epithelial cells to attaching bacteria was not uniform over a given patient’s hospitalization. Interestingly, the avidity of uroepithelial cells from patients who were being followed prospectively and who had cells collected on a daily basis was markedly increased in the 2 to 3-day period prior to onset of infection. Thus, it is possible that alterations in uroepithelial cell receptivity predispose catheterized patients to UTI. If so, better understanding of this phenomenon could lead to novel approaches to preventing UTI,. PREVENTION In females, a logical approach to preventing UTI, would seem to be the use of antimicrobials to block 38-683

September

16, 1991

The American

Journal

of Medicine

entry of bacteria through the periurethral space 1351. In the late 1970s and early 1980s extensive studies by Burke and colleagues [36,371 unfortunately demonstrated little success with various regimens for preventing meatal entry, including soap and water washing of the periurethral zone once daily, application of an iodophor solution and ointment twice daily, and application of a polyantibiotic ointment twice daily. The reasons for the failure of these approaches are not clear, but could result from an insufficient dose, duration, or mode of application of these various compounds. More recently, the concept of incorporating antimicrobial substances into the catheter itself has been revived. Studies by Butler and Kunin [381 in 1968 showed no efficacy when a catheter impregnated with antibiotics was studied in a small number of patients. However, technologic developments have now permitted the coating or impregnation of catheters with silver ion, and silver-coated or impregnated catheters have recently been developed. This approach has an appealing logical basis in that silver ions are bactericidal, can be applied to catheters, are nontoxic, and when used topically have been effective in other settings, such as controlling burn wound infections [391. Schaeffer and colleagues 1401 demonstrated that silvercoated catheters (in conjunction with a silver catheter tubing junction connector and an antisepticfilled drainage bag) reduced the incidence of UTI, and delayed their onset in chronically catheterized patients. In a recently published trial by Liedberg and Lundeberg [411, excellent results were also obtained in short-term catheterized patients utilizing the silver catheter. They performed a randomized trial in a population of patients who had recently been catheterized after surgery. Most of the patients were men and none were on antibiotics. They demonstrated that 10% of the patients catheterized with the silver-coated latex catheter had UTIs versus 37% of those receiving a Teflon control catheter (p < 0.01). However, other studies of the silver-coated catheter in general hospital populations have demonstrated less benefit. Johnson and colleagues [lo] performed a randomized trial of UTI, in a large general hospital population. Overall, the rate of bacteriuria was 10% in a control population and 9% in recipients of the silver-coated catheter. When patients were stratified by gender and use of antimicrobials, benefit associated with use of the silver-coated catheter was observed in females who had not received systemic antimicrobials. In all other groups, there was no apparent preventive value associated with use of the silver catheter. Burke and colleagues (personal communication) have also found no benefit associated with the silver catheter in a general Volume

91 (suppl36)

CONFERENCE

hospital population. Further study of these catheters appears warranted [9]. Junctional disconnection or other errors associated with care of the catheter and drainage system have been identified as risk factors for UTI, in several prospective studies [5-7,101. Interestingly, such errors have been extremely frequent, occurring in 25-50% of patients in some studies. Thus, it would appear that staff education and behavior modification could play a role in reducing occurrence of bacteriuria in hospitalized patients. An alternative approach is the use of sealed junctions. In one prospective study, sealed junctions reduced the risk of UTI, in patients who were not receiving antimicrobials [12]. The cost effectiveness and value of this approach in an individual hospital will be dependent on the prevalence of disconnection and whether or not antimicrobials are frequently used in the patient population being catheterized. Considerable research in the last decade has also focused on bag disinfection [8,42-441. The rationale for bag disinfection seems clear, namely, that the introduction of an antimicrobial agent will inhibit bacteria that are allowed to contaminate the bag at the point of drainage. Hydrogen peroxide, iodophors, chlorhexidine, and trichloroisocyanuric acid have been used for this purpose. Studies done to date demonstrate that, without question, use of these agents reduces the occurrence of bag contamination [8,42-441. However, the effect of these approaches on reducing the overall rate of UTI, is often small due to the relative infrequency of bag source as the point of infection. This has been demonstrated to be between 0% and 20% in various studies. In addition, the use of systemic antimicrobials in more than one-half of patients with urinary catheters in some studies probably makes the addition of antimicrobials to the collecting bag superfluous in such patients. Systemic antimicrobials have repeatedly been demonstrated to decrease the occurrence of UTIs in the first 4-5 days of catheterization [5,10,121. They have also been effective in preventing bacteriuria in controlled trials [451. However, the considerations of cost, adverse reactions, and emergence of drug resistance have prevented their routine use. It may, however, be reasonable to consider systemic antimicrobials as a preventive measure in high-risk patients who have an anticipated short duration of catheterization. Similarly, a very interesting recent study has been published demonstrating the value of selective decontamination of the gut in preventing catheter-associated UTIs [461. As noted earlier, in women, strains that subsequently infect the urinary tract generally emanate from the fecal bacterial population. Thus, selective decontamination of the gut as has been used for September

16, 1991

ON NOSOCOMlAL

INFECTIONS

/ STAMM

prevention of infections in neutropenic patients may be of value in preventing UTI,. According to Vollaard and colleagues [461, who performed a trial in which oral norfloxacin plus oral amphoteracin B was administered to patients with urinary catheters, this regimen decreased the occurrence of UTI, and delayed the onset of infection. The regimen was effective in decontaminating the gut of aerobic gram-negative rods and was associated with high urinary concentrations of antimicrobials that had broad spectrum of activity. This approach may be very useful for patients who are anticipated to have short-term, high-risk catheterizations and would probably be most likely to be successful in females. LONG-TERM CATHETERIZATION The second major population of patients who develop UTI, are those in whom catheterization is long term ( 2 30 days) or, in some cases, lifelong. Some of these patients are hospitalized but many are encountered in nursing homes or other chronic care facilities. Considerable information regarding the epidemiology, pathogt2neaiq and prevention of infection in these patients has been gathered in the last 10 years, much of it by Warren and Mobley and their colleagues at the University of Maryland [47-491. Studies have demonstrated that the incidence of bacteriuria in these patients is approximately 8-10% per day and thus similar to that seen in the acutely catheterized patient [47]. Due to the chronic nature of the catheter’s presence, however, the prevalence of bacteriuria in this population essentially is 100% 147,481. Prospective studies have shown that in this population, bacteriuria generally becomes chronic [481. However, the infecting strains present often cycle with new strains entering and old strains leaving the catheterized urinary tract. Frequently, bacteriuria is polymicrobial [48]; 85% of patients have more than two strains and 10% have more than five strains present. Some strains, specifically Proteeae, Pseudomonas, and enterococci, appear to be persistent within the urinary tract once present, whereas other strains frequently cycle in and out of the urinary tract [47,481. This may be due to the fact that those that persist are able to adhere more avidly to the catheter. For Providencia stuartii, specific fimbria (the MR/K fimbriae) have been shown to promote adherence to the catheter and persistence in the urinary tract [491. On the other hand, E. coli type 1 fimbriae have been shown to promote persistence in the urinary tract by increasing avidity of adherence to uroepithelial cells [491. Thus, there may be two populations of organisms in such patients, namely, those that attach to the catheter and those that attach to bladder uroepithelial cells. Alternatively, Proteeae has been found as

The American

Journal

of Medicine

Volume

91 kx~ppl3B)

3B-69s

CONFERENCE

ON NOSOCOMIAL

INFECTIONS

/ STAMM

persistent colonizers of the groin in nursing home patients, and such colonization may provide a reservoir from which infection originates 1501. A recurrent problem in these patients is that of catheter encrustation and eventual obstruction. The pathogenesis of catheter encrustation is similar in many ways to that of an infection stone [471. Bacteria adhere to the catheter and produce encrustation as described before. The encrustation consists of bacteria, their glycocalyces, host proteins such as Tamm-Horsfall protein, and eventually urinary crystals such as struvite and apatite [20,26,51,52]. This process is accelerated byureaseproducing bacteria, especially Proteus mirabilis, which has a very potent urease [53,541. The urease alkalinizes the urine, promotes crystallization of struvite and apatite, and promotes the encrustation and obstruction process. Patient-related factors may play a role here as well in that some patients’ urine leads to catheter obstruction more rapidly than others. Further study of the process of encrustation and catheter blockage would be of great interest in these patients. Prevention of bacteriuria and the complications of bacteriuria in long-term catheterized patients has been largely unsuccessful. Intermittent urethral catheterization has been shown to be effective compared with long-term indwelling catheterization in studies utilizing historical controls [47,55,561. Interestingly, no controlled trials of intermittent catheterization versus continuous catheterization have been performed, but it seems apparent from data in spinal cord injury units that this approach is of clear benefit. In patients on intermittent catheterization, a variety of antimicrobials may be useful in reducing the occurrence or duration of bacteriuria [4,57-611. Antimicrobials used in this manner include neomycin-polymxin solution, methenamine, nitrofurantoin, and trimethoprim-sulfamethoxazole. It has not been possible to demonstrate that any prophylactic regimen is effective in patients with chronically indwelling catheters 1471. Likewise, treatment of asymptomatic bacteriuria does not appear to be of any benefit in reducing the complications of bacteriuria in chronically catheterized patients [481.

FUTURE DIRECTIONS Studies undertaken in the 1980s suggest various new avenues for future study of catheter-associated bacteriuria in the 1990s. The use of targeted systemic antimicrobial prophylaxis or of selective gut decontamination offers promise for prevention of bacteriuria in females undergoing short-term catheterization. Similarly, antimicrobial catheters utilizing silver or other compounds may be benefi3B-70s

September

16, 1991

The American

Journal

of Medicine

cial in this group. Epidemiologic studies are required to clarify who develops serious sequelae of UTI,. Other important unanswered questions include the following: Can asymptomatic colonization be distinguished from asymptomatic infection? Does this have implications for development of subsequent complications? Can patients who win develop bacteriuria, upper tract infection, and death be identified earlier? Is there a substantial occurrence of occult bacteremia and endotoxemia in such patients? Can catheter materials be developed that will retard bacterial adherence and reduce the occurrence of biofilm growth, or can other more innovative ways to build catheters be developed? Further studies of the interaction of bacteria and bladder epithelial cells may be of value. If specific adhesins are identified, the catheterized urinary tract might be an ideal location for the use of competitive inhibitors, which could be placed in the catheterized urinary tract via infusion and competitively inhibit attachment of bacteria. In long-term catheterized patients, prevention of bacteriuria at the moment seems unlikely. Thus, it may be more fruitful to focus on reducing the occurrence of complications related to catheterization, most importantly encrustation and obstruction. Urease inhibitors may be of benefit in this regard and it may be possible to inhibit bacterial attachment if such attachment is mediated by specific fimbriae as demonstrated in recent studies.

REFERENCES 1. Haley R, Culver D, White J, et a/. The nationwide nosocomial infection rate: a new need for vital statistics. Am J Epidemiol 1985; 121: 159-67. 2. Gaynes R, Culver DH, Emori G, et a/. The National Nosocomial Infections Surveillance system: plans for the 1990s and beyond. Am J Med 1991; 91(Suppl36): 116-120. 3. Kunin CM, McCormack RC. Prevention of catheter-induced urinary-tract infections by sterile closed drainage. N Engl J Med 1966; 274: 1155-61. 4. Kunin CM, Finkelberg Z. Evaluation of an intraurethral lubricating catheter in prevention of catheter-induced urinary tract infections. J Urol 1971; 106: 928-30. 5. Garibaldi RA, Burke JP, Dickman ML, Smith CB. Factors predisposing to bacteriuria during indwelling urethral catheterization. N Engl J Med 1974; 291: 215-9. 6. Platt R, Polk BF, Murdock B, Rosner B. Mortality associated with nosocomial urinary tract infection. N Engl J Med 1982; 307: 637-42. 7. Warren JW, Platt R, Thomas RJ, et al. Antibiotic irrigation and catheterassociated urinary-tract infections. N Engl J Med 1978; 299: 570-3. 8. Thompson RL, Haley CE, Searcy MA, et a/. Catheter-associated bacteriuria. Failure to reduce attack rates using periodic instillations of a disinfectant into urinatydrainagesystems. JAMA 1984; 251: 747-51. 9. Classen DC, Stevens LE, Bass SA, Burke JP. Lack of efficacy of a silver oxide-coated urinary catheter in the prevention of catheter-associated bacteriuria: a large randomized clinical trial. Interscience Conference on Antimicrobial Agents and Chemotherapy, Atlanta, October 22-26, 1990. 10. Johnson JR, Roberts PL, Olsen RJ, Moyer KA, Stamm WE. Prevention of catheter-associated urinary tract infections with a silver oxide-coated urinary catheter: clinical and microbiological correlates. J Infect Dis 1990; 162: 1145-50. 11. Platt R, Polk BF, Murdock B, Rosner B. Riskfactorsfor nosocomial urinary tract infection. Am J Epidemiol 1986; 124: 977-85.

Volume

91 (suppl

38)

CONFERENCE 12. Platt R, Murdock B, Polk BF, Rosner 6. Reduction of mortality associated with nosocomial urinarytract infection. Lancet 1983; 1: 1893-7. 13. van Deventer SJH, de Vries I, van Eps LWS, eta/. Endotoxemia, bacteria, and urosepsis. in: Bacterial endotoxins: pathophysiological effects, clinical significance, and pharmacological control. Alan R. Liss, 1988: 213-224. 14. Bryan C, Reynold K. Hospital-acquired bacteremic urinary tract infection: epidemiology and outcome. J Urol 1984; 132: 494-8. 15. Gordon D, Bune A, Grime B, eta/. Diagnosticcriteria and natural history of catheter-associated urinary tract infections after prostatectomy. Lancet 1983; 1: 1269-71. 16. Stamm WE, Martin SM, Bennett JV. Epidemiology of nosocomial infections due to gram-negative bacilli: aspects relevant to development and use of vaccines. J Infect Dis 1977; 136: S151-60. 17. Kreger BE, Craven DE, Carling PC, McCabe WR. Gram-negative bacteria. Ill. Reassessment of etiology, epidemiology and ecology in 612 patients. Am J Med :980; 68: 332-55. 18. Daifuku R, Stamm W. Association of rectal and urethral colonization with urinary tract infection in patients with indwelling catheters. JAMA 1984; 252: 2028-30. 19. Stamm WE, Hooton TM, Johnson RJ, et a/. Urinary tract infections: from pathogenesis to treatment. J Infect Dis 1989; 159: 400-6. 20. Nickel JC, Grant SK, Costerton JW. Catheter-associated bacteriuria, an experimental study. Urology 1985; 26: 369-75. 21. Schaeffer AJ. Catheter-associated bacteriuria. Urol Clin North Am 1986; 4: 735-58. 22. Garibaldi RA, Burke JP, Britt MR, et al. Meatal colonization and catheter-associated bacteriuria. N Engl J Med 1980; 303: 316-8. 23. Schaeffer AJ, Chmiel J. Urethral meatal colonization in the pathogenesis of catheter-associated bacteriuria. J Urol 1983; 130: 1096-g. 24. Stark RP, Maki DG. Bacteriuria in the catheterized patient. What quantitative level of bacteriuria is relevant? N Engl J Med 1984; 311: 560-4. 25. Cox AJ, Hukins DWL, Davies KE, lrlam JC, Sutton TM. An automated technique for in vitro assessment of the susceptibility of urinary catheter materials to encrustation. Eng Med 1987; 6: 37-41. 26. Nickel JC, Gristina P, Costerton JW. Electron microscopic study of an infected Foley catheter. Can J Surg 1985; 28: 50-2. 27. Nickel JC, Ruseska I, Wright JB, Costerton JW. Tobramycin resistance of Pseudomonas aeruginosa cells growing as a biofilm on urinary catheter material. Antimicrob Agents Chemother 1985; 27: 619-24. 28. Cox AJ, Hukins DWL, Sutton TM. Infection of catheterized patients: bacterial colonization of encrusted Foley catheters shown byscanningelectron microscopy. Urol Res 1989; 17: 349-52. 29. Mobley HLT, Warren JW. Urease-positive bacteriuria and obstruction of long-term urinary catheters. J Clin Microbial 1987; 25: 2216-7. 30. Bergquist D, Bronnestam R, Hedelin H, Stahl A. The relevance of urinary sampling methods in patients with indwelling Foley catheters. Br J Urol 1980; 52: 92-5. 31. Johnson JR, Moseley SL, Roberts PL, Stamm WE. Aerobactin and other virulence factor genes among strains of E. co/i causing urosepsis-association with patients’ characteristics. Infect lmmun 1988; 56: 405-12. 32. Johnson JR, Roberts PL, Stamm WE. P-fimbriae and other virulence factors in E, co/i urosepsis-association with patients characteristics. J Infect Dis 1987; 156: 225-9. 33. Stamm WE, Martin SM, Bennett JV. Epidemiology of nosocomial infections due to gram-negative bacilli: aspects relevant to the development and use of vaccines. J Infect Dis 1977; 136: S151-60. 34. Daifuku R, Stamm W. Bacterial adherence to bladder uroepithelial cells in catheter-associated urinary tract infection. N Engl J Med 1986; 314: 1208-13. 35. Burke JP, Larsen R, Stevens L. Nosocomial bacteria--estimating the potential for prevention by closed sterile drainage systems, Infect Control 1986; 7: 96-9. 36. Burke JP, Garibaldi RA, Britt MR, et a/. Prevention of catheter-associated urinary tract infections. Efficacy of daily meatal care regimens. Am J Med 1981; 70: 655-85. 37. Burke J, Jacobson J, Garibaldi R, et a/. Evaluation of daily meatal care with poly-antibiotic ointment in prevention of urinary catheter-associated bacteriuria. J Ural 1983; 129: 331-4.

September

16,

1991

ON NOSOCOMIAL

INFECTIONS

/ STAMM

38. Butler KH, Kunin CM. Evaluation of polymyxin catheter lubricant and impregnated catheters. J Urol 1968; 100: 560-8. 39. Akiyama H, Okamoto S. Prophylaxis of indwelling urethral catheter infection: clinical experience with a modified Foley catheter and drainage system. J Urol 1979; 121: 40-2. 40. Schaeffer AJ, Stony KO, Johnson SM. Effect of silver oxide/trichloroisocyanuric acid antimicrobial urinary drainage system on catheter-associated bacteriuria. J Urol 1988; 139: 69-73. 41. Liedberg H, Lundeberg T. Silver alloy coated catheters reduce catheterassociated bacteriuria. Br J Urol 1990; 65: 379-81. 42. Gillespie W, Jones J, Teasdale C, et al. Does the addition of disinfectant to urine drainage bags prevent infection in catheterized patients? Lancet 1983; 1: 1037-g. 43. Sweet DE, Goodpasture HC, Holl K, et al. Evaluation of H,O, prophylaxis of bacteriuria in patients with long-term indwelling Foley catheters: a randomized controlled study. Infect Control 1985; 6: 263-6. 44. Maizels M, Schaeffer AJ. Decreased incidence of bacteriuria associated with periodic instillations of hydrogen peroxide into the urethral catheter drainage bag. J Urol 1980; 123: 841-5. 45. Britt MR, Garibaldi RA, Miller WA, et al. Antimicrobial prophylaxis for catheter-associated bacteriuria. Antimicrob Agents Chemother 1977; 11: 240-3. 46. Vollaard EJ, Clasener HAL, Zambon JV, Joosten HJM, van Griethuysen AJA. Prevention of catheter associated gram-negative bacilluria with norfloxacin by selective decontamination of the bowel and high urinary concentration. J Antimicrob Chemother 1989; 23: 915-22. 47. Warren JW. Catheter-associated urinary tract infections. Infect Dis Clin North Am 1987; 1: 823-55. 48. Warren JW, Tenney JH, Hoopes JM, Muncie HL. A prospective microbiologicstudy of bacteriuria in patients with chronic indwelling urethral catheters. J Infect Dis 1982; 146: 719-23. 49. Mobley HLT, Chippendale GR, Tenney JH, et al. MR/K hemagglutination of Providencia stuarti correlates with catheter adherence and with persistence in catheter-associated bacteriuria. J Infect Dis 1988; 157: 264-71. 50. Ehrenkranz NJ, Alfonso BC, Eckert DG, Moskowitz LB. Proteeae species bacteriuria accompanying Proteeae sp. groin skin carriage in geriatric outpatients. J Clin Microbial 1989; 27: 1988-91. 51. Hedelin H, Eddeland A, Larsson L, et a/. The composition of catheter encrustations, including the effects of allopurinol treatment. Br J Urol 1984; 56: 250-4. 52. Ladd TI, Schmiel D, Nickel JC, Costerton JW. Rapid method for detection of adherent bacteria on Foley urinary catheters. J Clin Microbial 1985; 21: 1004-6. 53. McLean R, Nickel JC, Noakes VC, Costerton JW. An in vitro ultrastructural study of infectious kidney stone genesis. Infect lmmun 1985; 49: 805-l 1. 54. Mobley HLT, Warren JW. Urease-positive bacteriuria and obstruction of long-term urinarycatheters. J Clin Microbial 1987; 25: 2216-7. 55. Anderson RU. Prophylaxis of bacteriuria during intermittent cathetenzation of the acute neurogenic bladder. J Urol 1980; 123: 364-6. 56. Donovan WH, Stolov WC, Clowers DE, Clowers MR. Bacteriuria during intermittent catheterization following spinal cord injury. Arch Phys Med Rehabil 1978; 59: 351-7. 57. Rhame FS, Perkash I. Urinary tract infections occurring in recent spinal cord injury patients on intermittent catheterization. J Urol 1979; 122: 669-73. 58. Maynard F, Diokno A. Urinary infection and complications during clean intermittent catheterization following spinal cord injury. J Urol 1984; 132: 943-6. 59. Kuhlemeier K, Stover S, Lloyd L. Prophylactic antibacterial therapy for preventing urinary tract infections in spinal cord injury patients. J Urol 1985; 134: 514-7. 60. Krebs M, Halvorsen R, Fishman I, Santos-Mendoza N. Prevention of urinary tract infection during intermittent catheterization. J Urol 1983; 131: 82-5. 61. Warren JW, Hoopes JM, Muncie HL, Anthony WC. Ineffectiveness of cephalexin in treatment of “cephalexin-resistant” bacteriuria in patients with chronic indwelling urethral catheters. J Urol 1983; 129: 71-3.

The American

Journal

of Medicine

Volume

91 (suppl3B)

3B--71s