Culture of the Surfaces of Urinary Catheters to Sample Urethral Flora and Study the Effect of Antimicrobial Therapy

Transcription

1 JOURNAL OF CLINICAL MICROBIOLOGY, June 985, p /85/ $0.00/0 Copyright C 985, American Society for Microbiology Vol., No. 6 Culture of the Surfaces of Urinary Catheters to Sample Urethral Flora and Study the Effect of Antimicrobial Therapy CALVIN M. KUNIN* A CHERRILEE STEELE The Department of Medicine, Ohio State University, Columbus, Ohio 30 Received 3 December 98/Accepted February 985 At the time indwelling urinary catheters were removed, the urethral flora was sampled by rolling the catheters onto culture plates. Cultures were obtained from 398 patients with sterile urine. Catheter surfaces were colonized less often in males than in females (6.8 and 67.0%, respectively) and yielded correspondingly fewer bacterial species per catheter. In both sexes, gram-positive species were isolated more frequently than gram-negative species (8 and 3 isolates, respectively). The bacteria most commonly isolated were Staphylococcus epidermidis and Streptococcusfaecalis. Staphylococcus saprophyticus was not recovered despite the large proportion of young, sexually active women in the study population. Gram-negative bacteria, mainly Escherichia coli, gradually colonized the catheters over time but at all times were recovered in lower amounts than gram-positive organisms and less often in males than in females. Slime production by coagulase-negative staphylococci did not relate to the density of organisms recovered from the catheters or influence the presence of gram-negative bacteria. Despite frequent use of antibiotics in this population, the urethral flora remained highly susceptible to these drugs. Antibacterial activity was not detected on the catheters despite high concentrations in the urine. This suggests that systemic antibiotics have a very limited effect on resident urethral flora. Despite relatively small numbers in the urethra, gram-negative bacteria readily produce urinary tract infections in catheterized patients. In patients with indwelling urinary catheters, microorganisms may be introduced into the bladder by three major routes. These are (i) inoculation of microorganisms into the bladder at the time the catheter is inserted, (ii) ascent of microorganisms into the bladder through the catheter lumen from a contaminated drainage system, and (iii) migration into the bladder of microorganisms present in the urethra around the catheter. Although a small number of urethral bacteria may be introduced into the bladder when the catheter is inserted, they appear to be rapidly washed out by urine flow or are unable to attach and multiply in the bladder. This probably accounts for the relatively low frequency of bacteriuria after single or intermittent catheterization (). Occurrence of the second mechanism, ascent from a contaminated drainage system, has been reduced markedly by the use of closed methods of catheter drainage and improved methods of management of the system (). Despite these measures, development of bacteriuria occurs in patients with indwelling catheters left in place for to weeks or longer. The current concept concerning the development of catheter-induced bacteriuria is that the urethra becomes colonized with gram-negative rods and enterococci derived from the fecal flora. These bacteria enter the bladder through the mucosal sheath around the catheter. It is likely that mechanical irritation of urethral and bladder mucosa by the catheter enhances the susceptibility of these structures to invasion by microorganisms. The evidence which supports the intraurethral source of infection is based on the observations of Kass and Schneiderman (9), who showed that Serratia marcescens applied to the perineum of catheterized females colonize the bladder urine. More recently, Garibaldi et al. (8), Daifuku and Stamm (7), and Schaefer and Chmiel (0) found that antecedent urethral colonization with enteric gram-negative rods * Corresponding author. 90 is followed by catheter-associated bacteriuria with the same organisms. In those studies females acquired both meatal colonization and catheter-associated bacteriuria more readily than did males. The current study was designed to enable us to more fully understand the ecosystem of aerobic urethral flora in patients with indwelling catheters and to characterize the effects on urethral flora of duration of catheterization, age, sex, and concurrent antimicrobial therapy. MATERIALS A METHODS Patient population. The study was conducted on the obstetrics, gynecology, urology, and medical intensive care units at the Ohio State University Hospital from November 983 to June 98. Catheters were inserted after the periurethral area was cleansed with a solution of povidone-iodine. Routine catheter care at this hospital consists of attaching the catheter to a closed drainage system and cleansing the periurethral area with soap and water daily. The catheter lubricant (Surgilube; Bike Golden Inc., Melville, N.Y.) did not contain antimicrobial agents. Most of the catheters used were made of Teflon-latex obtained from a variety of manufacturers. Before catheter removal, the nurses cleansed an area of the catheter with an alcohol swab, inserted a sterile needle attached to a sterile syringe, aspirated 5 to 0 ml of urine from the catheter, and placed the specimen in a sterile tube. The catheter was then removed and wrapped in a clean, plastic-lined pad. Urine and catheter specimens were immediately placed in a refrigerator and were cultured within h of collection. Information obtained included the name of the patient, hospital identification number, age, sex, times of insertion and removal of the catheter, and current antimicrobial therapy. Bacteriologic studies. Urine specimens were cultured quantitatively on sheep blood agar plates (GIBCO Laboratories, Grand Island, N.Y.) with a disposable 00-pI loop and a plate-revolving device to ensure even spread of the inocu- Downloaded from on November, 08 by guest

2 VOL., 985 EFFECTS OF THERAPY ON URINARY CATHETER FLORA 903 TABLE. Age and sex distribution of the population with sterile urine Age group No. (%) of patients (yrs) Male Female <0 3 (3.7) 3 (0.6) -0 (7.) (70.9) -60 (9.3) 35 (.) (6.3) (6.6) >80 3 (3.7) (0.6) lum. Plates were read after incubation for 8 h at 37 C without CO. Sterile gloves and scissors were used to culture the catheters. Beginning at the tip, the first 8 in. (ca. 0.3 cm) of the catheter was cut into -in. (.5-cm) segments. Each segment was then rolled onto prelabeled sections of plates containing Columbia colistin-nalidixic acid and MacConkey agar (GIBCO), and incubation took place at 37 C for h. Catheters obtained from patients whose urine showed no growth were studied further. The density of growth on plates on which the catheter had been rolled was measured semiquantitatively by the scheme used for dipslides (6). Density extended from 0 or fewer colonies on the plate (0) to to 50 (+) to heavy confluent growth (+). One to five representative colonies were picked from each inoculum site and streaked onto blood agar plates until pure isolates were obtained. Gram-positive bacteria were identified and classified by the catalase slide test (3% H0), coagulase latex agglutination slide test (Staphlatex agglutination kit; American MicroScan, Lexington, Ky.), and coagulase tube method with pig plasma. A bile-esculin plate was used to help identify enterococci. Susceptibility to novobiocin was tested with 5-pug disks, bacitracin with 0.0-U disks, and ethyl hydrocupreine hydrochloride with 5-,ug disks (BBL Microbiology Systems, Cockeysville, Md.). Strains of coagulase-negative staphylococci were identified to species level by the scheme of Kloos and Schleifer (). Gram-negative bacteria were separated by the oxidase test (Marion Scientific, Div. Marion Laboratories, Inc., Kansas City, Mo.). Nonfermenting organisms were then identified with Uni-N/F Tek plates (Flow Laboratories, Inc., McLean, Va.) and triple sugar iron agar tubes (Difco Laboratories, Detroit, Mich.). Enteric organisms were identified with biochemical sensitivity plates (Ohio State University, Columbus). Antimicrobial susceptibility tests were performed with serially diluted, dehydrated drugs prepared by the Clinical Microbiology Laboratory at Ohio State University Hospital. The inoculum was adjusted with a McFarland standard. Slime production by staphylo- TABLE. Agent Antimicrobial agents used in the population with sterile urine' No. (%) of patients Cephalosporins... 9 (.6) (0.) Aminoglycosides (3.7) Clindamycin... 7 (.3) Other penicillins... 7 (3.)... 6 (.7) Others... 6 (7.3) a Courses per patient receiving antimicrobial agents:.. cocci was determined by the tube method described by Christensen et al. (5). Detection of antimicrobial activity in urine and on catheter surfaces was accomplished with Escherichia coli 6, which is susceptible to jig of ampicillin and,ug each of cephalothin and gentamicin per ml. Urine was assayed for antimicrobial activity by the standard twofold dilution method in Trypticase soy broth (BBL) with an inoculum of an equal volume of a :,000 dilution of an overnight culture. For studies of antimicrobial activity on catheters, a :00 dilution of an overnight culture was swabbed onto the surface of a Trypticase soy agar plate. Starting at the tip, the first four -in. (ca..5 cm) segments of the catheters were placed on the plates (after culturing) and incubated for h at 37 C. Plates were then examined for zones of inhibition around the catheters. Fresh, unused catheters were examined by these methods for antimicrobial activity. RESULTS Characteristics of the population and use of antibiotics. The patient population was made up of 8 males (0.6%) and 36 females (79.%). A total of 9 patients (55.0%) used antibiotics, 37 (3.%) did not, and it was unknown whether antibiotics were used in the remaining patients (0.6%). The patients were housed in the obstetrics (66 [.7%]), gynecology (07 [6.9%]), urology-surgery (08 [7.%]), and intensive care (7 [.3%]) units. The distribution of the patient population according to age group and antibiotics used is shown in Tables and. Most of the patients were young adult females, a fact that reflects the frequent use of catheters in obstetrics and gynecology. Antimicrobial agents were commonly used in this population. The mean duration of catheterization was relatively short (3.3 and.8 days in males and females, respectively). The population included males and 9 females who were catheterized for or more days. Recovery of bacteria from catheter surfaces. The bacteria were distributed fairly evenly in the -in. (.5-cm) sections. The density was lowest at the tip and proximal end and greatest at in. (ca. 0 cm) distal to the balloon. The number of bacterial strains isolated from male and female patients is presented in Table 3. Surfaces of catheters from males were colonized less often than those from females (6.8 versus 67.0%), and proportionately fewer catheters from males contained multiple strains (P < 0.00). Bacteria were present in urinary catheters of 7 males (30.9%) and females (75.6%) who used antimicrobial agents. Bacteria were not present in 38 males (69.%) or 0 females (.%) using the drugs. In patients not using antibiotics, bacteria were present for 7 males (3.8%) and females (9.%) and not present for 9 males (56.%) and 7 females (5.9%). Although catheters from patients receiving antibiotics tended to be colonized less often than those from TABLE 3. Detection of microorganisms on catheter surfaces" No. (%)b of patients from whom the following no. of strains were Sex of isolated: patients Male 55 (67.0) 6 (9.5) 6 (7.3) 3 (3.7) (.5) (.5) Female 53 (6.8) 06 (33.5) 83 (6.) 5 (6.5) 5 (.8) 7 (.) a Difference between males and females for presence of organisms on catheters, chi square = 83.3, P < Number of strains per person: 5 (male) and.66 (female). " Percentage by sex of strains isolated. Downloaded from on November, 08 by guest

3 90 KUNIN A STEELE TABLE. Frequency of gram-positive bacteria recovered from catheters according to sex Bacterium Male No. (%) of patients Female Coagulase-negative staphylococcia 9 (5.) 76 (.8) Group D streptococci" (3.3) 67 (.5) Other streptococci 0 (0.0) 8 (.6) S. aureus 0 (0.0) 6 (.) Diphtheroids 3 (8.6) (3.6) Others (.8) (0.6) " Of these, 70.% were S. epidermidis. "Of these, 9.% were S. faecalis. 50 r 0O % 30 OF STRAINS 0 [ 0 O L J. CLIN. MICROBIOL. Strep. fueco/is ( 63) S ourus ( 6 ) R osepidermidis ( 37 ) ---E:co/i(7coi () patients not receiving antibiotics, these differences were not statistically significant at the 0.05 level by the chi-square test (males: chi square =.6, P > 0.; females: chi square =.7, P = 0.). Characterization of the bacteria recovered from catheter surfaces. Gram-positive bacteria were isolated more commonly than gram-negative bacteria (8 versus 3, respectively [Table and 5]). In addition, gram-positive bacteria populated catheter surfaces more densely than did gram-negative bacteria (Fig. ). Among the gram-positive bacteria, coagulase-negative staphylococci and group D streptococci were most frequently isolated, accounting for 88.5 and 87.3% of organisms for males and females, respectively. Most of the coagulase-negative staphylococci were Staphylococcus epidermidis (70.%) (Table 6). The remaining staphylococci were mainly Staphylococcus hominis and Staphylococcus haemolyticus. Staphylococcius saprophyticus was not isolated. Most of the group D streptococci were Streptococcus faecalis (9.%). E. coli accounted for more than half of the gram-negative bacteria, shown in Table 5. The remainder consisted of enteric organisms commonly encountered in urinary tract infections. The effect on colonization of the time the catheters were in place is shown in Fig.. Virtually all catheters in females became colonized by or more days, compared with only about half in males. The acquisition of gram-negative bacte- TABLE 5. Frequency of gram-negative bacteria recovered from catheters Bacterium No. M of isolates (no. [%J of patients)no(%ofilae E. coli (73 [55.7]) Proteus spp. (6 [.]) P. mirabilis.3 (8.3) P. vulgaris. (.5) Klebsiella spp. (8 [3.7]) K. pneumoniae..5 (83.3) K. oxytoca.3 (6.7) Enterobaccter ( [9.]) E. cloacae.7 (58.3) E. aerogenes. (33.3) E. agglomerans. (8.3) Others' ( [9.]) aincludes isolates of six Pseudomonas, two Morganella, three Citrobacter, and one Acinetobacter species. ' + I DENSITY ON CATHETER FIG.. Relation between bacterial species according to their density on the surfaces of urethral catheters. ria increased with the time the catheter was in place, but in both males and females it was lower than for gram-positive bacteria at all time periods. Relation between ability of staphylococci to form slime and colonization of catheters. Coagulase-negative staphylococcal strains tested for formation of slime were examined in relation to their relative abundance on catheter surfaces. Heavy growth occurred in isolates from 6 catheters that had slime production (3.0%) and 3 that did not (65.9%); light growth occurred in (0%) of isolates that were positive for slime production and 6 (75.0%) of isolates that were negative (chi square =.3, P > 0.0). Heavy growth on surfaces was determined by densities corresponding to 00,000 or more by the dip slide criterion. There was no association between slime production and magnitude of catheter colonization. Slime production by coagulase-negative staphylococci also did not affect the adherence of E. coli to catheters. Heavy growth occurred in 5 (3.8%) slime-positive isolates and in 9 (8.0%) slime-negative isolates; slight or no growth occurred in 6 (76.%) slime positive isolates and in (8.0%) slime-negative isolates (chi square = 0.3, P = 0.70) MALES Colonized m (+) Gm (-) DAYS lor< -3 or: N= (33) (5) () FEMALES Ioonized lors -3 0r: (6) (35) (9) Gm(+) Gm (-) DAYS CATHETER IS IN PLACE FIG.. Effect that duration urethral catheters were in place had on colonization in males and females with gram-negative and -positive bacteria. Downloaded from on November, 08 by guest

4 VOL., 985 EFFECTS OF THERAPY ON URINARY CATHETER FLORA 905 TABLE 6. Frequency of species of coagulase-negative staphylococci recovered from catheters Species No. (%) of catheters S. epidermidis (70.) S. hominis (6.9) S. haemolyticus... (0.8) S. cohni... () S. warnii... (.0) Unidentified... () Antimicrobial susceptibility of bacteria recovered from catheter surfaces. Because of the common use of antibiotics in this population, it was of interest to determine the antibiotic susceptibility of organisms recovered from catheters at ranges readily achievable in urine. These data are summarized in Tables 7 and 8 for gram-positive and -negative bacteria, respectively. Despite heavy use of antibiotics in the patient population, the predominant staphylococcal species were highly susceptible to the most commonly used drugs. The only exception was strains of S. haemolyticus. In this group, 7 of strains were highly resistant to penicillin G. Resistance of this species was somewhat lower than expected for nosocomial isolates from this hospital. Perhaps the differences can be explained by the fact that these isolates were obtained from otherwise healthy young women. Isolates of S. faecalis, although for the most part resistant to cephalosporins, were highly susceptible to penicillins and, to a lesser extent, to gentamicin. The susceptibility of Streptococcus faecium was similar to that of S. faecalis. Other streptococci were also highly susceptible to antibiotics. Strains of E. coli, Klebsiella pneumoniae, and Proteus mirabilis were for the most part susceptible to many of the antibiotics used. The major exceptions were the other enteric gram-negative bacteria such as the Klebsiella, Enterobacter, and Pseudomonas species commonly encountered in hospital environments. Penetration of antibiotics into the urethra. Antimicrobial therapy did not appear to have a significant effect on recovery from catheters of bacteria highly susceptible to the antibiotics used. It therefore seemed likely that the antibiotics penetrate poorly into the urethral mucosa. To evaluate this possibility, we placed catheters removed from patients on the surface of agar plates seeded with a strain of E. coli highly susceptible to ampicillin, cephalothin, and gentamicin. Fresh, unused catheters showed no antimicrobial activity on these plates. Urine obtained at the same time was assayed for antimicrobial activity. A total of 60 catheter and urine pairs were tested. Thirty-five of the patients were receiving antibiotics. Of these, 5 followed courses of penicillins (mostly ampicillin), 6 of the cephalosporins (mostly cefazolin), 7 of aminoglycosides (mostly amikacin), and of other agents, including trimethoprim-sulfamethoxazole (), doxycyline, and clindamycin. Urine obtained from patients receiving antibiotics inhibited the test strain at a dilution of :3 or greater in 7.3% of individuals. The geometric mean dilution of urine showing activity was between :6 and :8. Antimicrobial activity at a :3 dilution, but no greater, was observed in 8.0% of patients not receiving antibiotics. The remaining urine specimens showed activity at a dilution of : or lower. Zones of inhibition were observed around the catheter tip and occasionally around the balloon in 7 of the 35 patients receiving antibiotics. The urine from most of these patients had high levels of antimicrobial activity, in the range of :56 or greater. No inhibition was detected from catheters removed from patients not on therapy. In contrast to these findings, there was no activity in portions of the catheter distal to the balloon from any patient, regardless of receipt of antibiotics or presence of antibacterial activity in urine. Growth of coagulase-negative staphylococci in urine and culture media. The much greater frequency and larger numbers of coagulase-negative staphylococci than of gram-negative bacteria recovered from the surfaces of catheters raised the question of why staphylococci do not more commonly produce urinary tract infections in patients with indwelling catheters. To answer this question, strains of E. coli and Staphylococcus aureus, S. epidermidis, S. haemolyticus, and S. saprophyticus were grown overnight in Trypticase soy broth. TABLE 7. Antimicrobial susceptibility of gram-positive bacteria recovered from urethral catheters Organism (no. of MIC (.g/ml)' % Suscepisolates) and drug 50% 90% tibleb S. epidermidis (0) S. hominis (3) Gentamycin S. haemolyticus () S. aureus (7) d S. faecalis (56) a 50%o and 90%o, MICs for 50 and 90%o of the strains, respectively. b Susceptible to c pg/ml. c -, End point not reached. d, Test not done. Downloaded from on November, 08 by guest

5 906 KUNIN A STEELE They were inoculated as 0. ml of a 0-6 dilution into 0 ml of urine (ph 7.0), Trypticase soy broth, or medium 99. Each culture tube was plated quantitatively at 0 and 6 h. There was only slight growth of the coagulase-negative staphylococci at 6 h in urine or in the culture media. In contrast, the strains of E. coli and S. aureus grew well in both urine and the media. They increased about 00-fold or more by 6 h. TABLE 8. Antimicrobial susceptibility of gram-negative bacteria recovered from urethral catheters Organism (no. of MIC (,ug/ml)a % Suscepisolates) and drug 50% 90% tibleb E. coli (7) K. pneumoniae (8) P. mirabilis Othersg d a 50% and 90%, MICs for 50 and 90% of the strains, respectively. b Susceptible to s,ug/ml. -, End point not reached. d, Test not done. e Susceptible to s6,ug/ml. f Susceptible to s6,ug/ml. Includes K. oxytoca, 7 E. cloacae, E. aerogenes, E. aglomerans, 3 Citrobacter diversus, Pseudomonas aeruginosa, Morganella morgagni, P. vulgaris, and Acinetobacter isolates. DISCUSSION J. CLIN. MICROBIOL. The aerobic flora of the entire length of the urethra in catheterized patients was sampled by using the catheter to obtain urethral cultures before acquisition of bacteriuria. In using this method, we assumed that bacteria growing in the urethra would be caught in the mucosal slime and adhere to the catheter. Sugarman (3), using radiolabeled E. coli and Klebsiella, demonstrated that these organisms adhere readily to latex and latex-teflon and less well to silicone catheters. He found that adherence is independent of temperature. It was reasonable to assume, therefore, that the catheter may serve as an intraurethral swab. The urethral flora of healthy women contains large numbers of both anaerobic and aerobic bacteria, but Enterobacteriaceae and other gram-negative rods are rarely found (7, 9, ). It has been proposed by Stamey et al. () that urethral colonization with gram-negative rods, enterococci, and S. saprophyticus is a major prerequisite for acquisition of urinary tract infections. This thesis is generally accepted, although there remains some debate concerning the susceptibility of women without previous urinary tract infections to acquire gram-negative rods in the periurethral zone (, 3). Colonization of the urethral meatus with gram-negative rods and enterococci appears also to be an important first step in acquisition of catheter-associated infections. Garibaldi et al. (8) demonstrated that catheter-associated bacteriuria is more frequently acquired by subjects who are female, are 50 years of age or older, have not received antibiotics, or who had conditions left untreated by surgery. They found that the risk of meatal colonization at the time the catheter is inserted is also greater in this high-risk group. In prospective studies in catheterized patients, Schaeffer and Chmiel (0) and Daifuku and Stamm (7) demonstrated that gram-negative rods become, over time, increasingly prevalent on the urethral meatus. Both studies were able to relate meatal colonization to acquisition of bacteriuria. The composition of the urethral flora of healthy females of various physiologic age groups and that of females with recurrent urinary tract infections were compared quantitatively by Marrie et al. (7). Aerobes accounted for twothirds to three-quarters of the flora of the premenarchal and reproductive age groups, whereas anaerobes predominated in the postmenopausal age groups. In that study, the aerobes consisted of lactobacilli, coagulase-negative staphylococci, corynebacteria, and streptococci. Aerobic gram-negative rods were not isolated from the urethral wash-out urine obtained from premenarchal and reproductive age groups, but only from subjects with recurrent urinary tract infections and some of the postmenopausal women. The results of this study are in accord with those of Marrie et al. (7) and others which showed that the predominant aerobic microbial flora of the urethra in females consists of gram-positive bacteria. We found that the aerobic flora in the male urethra was similar to that in the female, but the urethra in the male tended to be colonized less often. Gram-negative rods were acquired in increasing frequency in both sexes over the time the catheter was left in place, but at a lower rate in males than in females. This may explain, in part, the lower rate of acquisition of catheter-associated bacteriuria in males. The major difference between the results in this study and those of others was in recovery of lactobacilli and diphtheroids. These organisms accounted for about 50% of the aerobic bacteria isolated by Marrie et al. (7) from the urethral wash-out specimens in females. They were also Downloaded from on November, 08 by guest

6 VOL., 985 EFFECTS OF THERAPY ON URINARY CATHETER FLORA 907 commonly isolated by Maskell et al. (8) and Brumfitt et al. (3) from the urethras of females. Corynebacteria and diphtheroids were isolated on only three occasions in the current study, and lactobacilli were not recovered. These differences cannot be explained by culture methods, because the blood agar plates used in this study have been known to recover these organisms in this institution. It is doubtful that use of antimicrobial agents can explain the failure to isolate lactobacilli, because at least one-third of the patients did not receive antibiotics. It is possible that they either do not adhere well to catheters or are overgrown by other organisms. Slime production by staphylococci appears to have an important role in adherence to intravascular devices (5). It did not account for either adherence of staphylococci or inhibition of attachment of E. coli to urinary catheters in the current study. Despite the mnuch higher frequency of isolation of grampositive bacteria from catheters and their much greater density than gram-negative bacteria on the catheters, grampositive bacteria (other than enterococci) produced catheterassociated urinary tract infections much less often. They appeared to have a competitive advantage over gram-negative bacteria in the urethra, but not in the bladder urine. We confirmed the observation of Stamey and Mihara () that coagulase-negative staphylococci grow slowly in urine. We found that they tend to grow slowly in Trypicase soy broth and medium 99 as well. This property may explain, in part, their infrequent role in urinary tract infections. This concept is not, however, entirely satisfactory, because the strain of S. saprophyticus that was tested also grew poorly in urine and culture media yet is a common cause of urinary tract infections in young females (,, 5, 6, 5). S. saprophyticus was not isolated from catheters. This is surprising because a large proportion of the study population consisted of young, sexually active females. Possible explanations for absence of this organism in the current study are that it occurs more often in sexually promiscuous individuals or that its occurrence is seasonally related. Wallmark et al. (5) and Latham et al. () observed higher rates of infection with this organism among young adult female outpatients during the summer and fall. The organism is rarely recovered from hospitalized men and women with bacteriuria (5, 6, 5) and even more rarely from males (0). The current study was begun in the late fall and ended in June, times when this organism is least likely to be found. Antimicrobial agents are commonly used in patients who have indwelling urinary catheters. The agents tend, at least temporarily, to reduce the frequency of acquisition of bacteriuria (). Despite their frequent use in the patient population studied, the antibiotics did not exert a marked effect on urethral flora. Most of the strains of coagulase-negative staphylococci isolated from the catheters were highly susceptible to the antibiotics used, yet they also were the most abundant bacteria recovered. We did not detect any antibacterial activity diffusing from the catheters. It appears, therefore, that antibiotics penetrate into urethral secretions poorly despite high concentrations achieved in the urine. Trimethoprim was not commonly used in the patient population studied, and therefore no statement can be made about its potential effect on the urethral flora. The major biological question left unanswered by this study and those of other workers is why the urethra favors the growth of gram-positive bacteria and tends to exclude gram-negative bacteria. It is possible that bacteriocins or epidermidins produced by gram-positive bacteria inhibit gram-negative organisms. Beaudet et al. (), for example, have described a gonococcal growth inhibitor produced by coagulase-negative staphylococci. In this study, only a relatively small number of gram-negative bacteria colonized the urethra, even in the presence of the catheter. This inoculum seemed sufficient, however, to reach the bladder and to subsequently grow in urine. Although gram-positive bacteria are able to produce urinary tract infections, they are clearly less pathogenic despite their relatively greater abundance in the urethra. We wish to emphasize that the results of this study should not be misconstrued to imply how often urethral catheters should be changed, nor do we recommend that cultures of catheters be done routinely. LITERATURE CITED. Bailey, R Significance of coagulase-negative Staphylococcus in urine. J. Infect. Dis. 7: Beaudet, R., J. G. Bisaillon, S. A. Saheb, and M. Sylvestre. 98. Production, purification, and preliminary characterization of a gonococcal growth inhibitor produced by coagulase-negative staphylococcus isolated from the urogenital flora. Antimicrob. Agents Chemother. : Brumfitt, W., J. M. T. Hamilton-Miller, H. Ludlam, and A. Gooding. 98. Lactobacilli do not cause frequency and dysuria syndrome. Lancet ii: Cattell, W. R., M. A. McSherry, A. Northeast, E. Powell, H. J. L. Brooks, and F. O'Grady. 97. Periurethral enterobacterial carriage in pathogenesis of recurrent urinary infection. Br. Med. J. : Christensen, G. D., W. A. Simpson, A. L. Bisno, and E. H. Beachey. 98. Adherence of slime-producing strains of Staphylococcus epidermidis to smooth surfaces. Infect. Immun. 37: Cohen, S., and E. H. Kass A simple method for quantitative urine culture. N. Engl. J. Med. 77: Daifuku, R., and W. E. Stamm. 98. Association of rectal and urethral colonization with urinary tract infection in patients with indwelling catheters. J. Am. Med. Assoc. 5: Garibaldi, R. A., J. P. Burke, M. R. Britt, W. Miller, and C. Smith Metal colonization and catheter associated bacteriuria. N. Engl. J. Med. 303: Kass, E. H., and L. J. Schneiderman Entry of bacteria into the urinary tracts of patients with inlying catheters. N. Engl. J. Med. 56: Kauffman, C. A., C. S. Hertz, and J. N. Sheagren Staphylococcus saprophyticus: role in urinary tract infections in men. J. Urol. 30: Kloos, W. E., and K. H. Schleifer Simplified scheme for routine identification of human Staphylococcus species. J. Clin. Microbiol. : Kunin, C. M., and R. C. McCormack Prevention of catheter-induced urinary tract infections by sterile closed drainage. N. Engl. J. Med. 7: Kunin, C. M., F. Polyak, and E. Postel Periurethral bacterial flora in women. Prolonged intermittent colonization with Escherichia coli. J. Am. Med. Assoc. 3: Latham, R. H., K. Running, and W. E. Stamm Urinary tract infections in young adult women caused by Staphylococcus saprophyticus. J. Am. Med. Assoc. 50: Marrie, T. J., and C. Kwan. 98. Antimicrobial susceptibility of Staphylococcus saprophyticus and urethral staphylococci. Antimicrob. Agents Chemother. : Marrie, T. J., C. Kwan, M. A. Noble, A. West, and L. Duffield. 98. Staphylococcus saprophyticus as a cause of urinary tract infections. J. Clin. Microbiol. 6: Marrie, T. J., C. A. Swantee, and M. Hartlen Aerobic and anaerobic urethral flora of a healthy females in various physiological age groups and of females with urinary tract infections. J. Downloaded from on November, 08 by guest

7 908 KUNIN A STEELE Clin. Microbiol. : Maskell, R., L. Pead, and J. Allen The puzzle of the urethral syndrome: a possible answer? Lancet : Pfau, A., and T. Sacks. 98. The bacterial flora of the vaginal vestibule, urethra and vagina in premenopausal women with recurrent urinary tract infections. J. Urol. 6: Schaefer, A. J., and J. Chmiel Urethral meatal colonization in the pathogenesis of catheter-associated bacteriuria. J. Urol. 30: Stamey, T. A., and G. Mihara Observations on the growth of urethral and vaginal bacteria in sterile urine. J. Urol. J. CLIN. MICROBIOL. : Stamey, T. A., M. Timothy, M. Millar, and G. Mihara. 97. Recurrent urinary infection in adult women: the role of intoital enterobacteria. Calif. Med. 5: Sugarman, B. 98. Adherence of bacteria to urinary catheters. Urol. Res. 0: Turck, M., B. Goffe, and R. G. Petersdorf. 96. The urethral catheter and urinary tract infection. J. Urol. 88: Wallmark, G. I., I. Anemark, and B. Telander Staphylococcus saprophyticus: a frequent cause of urinary tract infections among female outpatients. J. Infect. Dis. 38: Downloaded from on November, 08 by guest