Catheter-associated urinary tract infections: prevalence of uropathogens and pattern of antimicrobial resistance in a UK hospital (1996–2001)

Authors


Abstract

OBJECTIVE

To assess the change in the bacterial profile and pattern of antibiotic resistance of catheter-associated urinary tract infections (CAUTIs, the most prevalent form of nosocomial infections) between 1996 and 2001.

MATERIALS AND METHODS

Catheter samples of urine (CSUs) submitted in 1996, 1998 and 2001 (2451, 2460 and 3349 specimens, respectively) were analysed. The distribution of different uropathogens in bacterial CAUTIs and their in vitro antimicrobial resistance was evaluated over the study interval. The likelihood ratio test was used to assess whether there was a linear trend according to calendar year.

RESULTS

Escherichia coli was the most frequently isolated pathogen in all years, but its frequency declined over time (35.6%, 32.5% and 26.6%, respectively). Enterococcus was the second most frequent overall, with a significant increase in frequency with time (11.8%, 15.3% and 22.0%, respectively). There was also a considerable change in resistance patterns to antibiotics. As a result, in 1996, CAUTIs were least often resistant to ciprofloxacin (8.0%) followed by co-amoxiclav (18.5%) and cephalexin (25.4%). In 2001, CAUTIs were least often resistant to co-amoxiclav (22.5%), followed by ciprofloxacin (27.2%) and nitrofurantoin (28.8%).

CONCLUSION

The types of organisms associated with CAUTI have changed over the last 5 years in a UK institution, as have the patterns of antibiotic resistance. Currently, the most appropriate agents for the empirical management of CAUTIs seem to be co-amoxiclav, ciprofloxacin and nitrofurantoin.

INTRODUCTION

Up to 25% of hospitalized patients undergo urinary catheterization [1,2]; a similar proportion of patients cared for in residential homes will have long-term indwelling catheters [3]. Although often a necessary intervention, indwelling urinary catheters are a leading cause of nosocomial infection and have been associated with both morbidity and mortality [1–10].

Up to 30% of catheterized patients can have genitourinary or systemic symptoms related to catheter-associated UTI (CAUTI) [1,8], and up to 4% may develop catheter-related bacteraemia [1,7,11]. Empirical antibiotic treatment is usually started when a symptomatic CAUTI is suspected and the result of urine culture is not yet available.

Once the catheter has been removed some patients with asymptomatic CAUTI continue to have bacteriuria or become symptomatic [12–14]. In one study, over a quarter of patients with asymptomatic CAUTI developed symptoms of UTI after the removal of their catheter [12]. To prevent or reduce this type of catheter-related morbidity, many clinicians have a policy of administering a short course of prophylactic antibiotics on catheter withdrawal for all or selected groups of patients.

The aim of the present study was to assess the bacterial profile for CAUTI and the antimicrobial resistance to the most commonly used antibiotics over the last 5 years. The results can be used to inform the choice of antibiotics used in the therapeutic or prophylactic settings before the results of the urine culture are available.

MATERIALS AND METHODS

A study of all catheter specimens of urine (CSUs) processed by the laboratory during 1996, 1998 and 2001 was conducted at one teaching hospital in the UK. All specimens with bacteriuria of> 103 colony-forming units (cfu)/mL urine (which defines CAUTI [1]) of one or two organisms were analysed to determine their causative uropathogens and their antibiotic susceptibilities.

Quantitative urine cultures were assessed using a standard loop method. Colonies were picked and plated onto blood agar and cystine lactose-electrolyte-deficient medium, and incubated overnight at 37 °C. After incubation at 37 °C for 24 h, the organisms were identified by standard tests.

In vitro susceptibility tests to antibiotics were conducted using the Stokes comparative disk method. If three or more types of organisms were present, the result was considered as a ‘mixed growth’ and susceptibility to antibiotics was not tested. Pathogens (except Pseudomonas spp) were tested against amoxycillin, co-amoxiclav (augmentin), cephalexin, ciprofloxacin, nitrofurantoin and trimethoprim. For Staphylococci, they were not tested against ciprofloxacin or cephalexin but against flucloxacillin (methicillin) instead. Pseudomonas spp were tested against ciprofloxacin, ceftazidime, gentamicin, piperacillin and meropenem.

Logistic regression was used to investigate whether the prevalence of bacterial uropathogens or their in vitro resistance patterns had changed with time, with calendar year as the only independent variable in the model. The likelihood ratio test was used to assess whether there was a linear trend according to calendar year; results with P < 0.05 were considered to be statistically significant.

RESULTS

The microbiological laboratory in our institution received 2541, 2460 and 3349 CSUs in 1996, 1998 and 2001, respectively (Table 1). In bacterial CAUTIs, there was an increase in the percentage of CAUTIs with two isolated organisms (polymicrobial) over time (P < 0.001). Escherichia coli was the most frequently isolated uropathogen in all years (Table 1), but there was a gradual decline in its frequency over time (P < 0.001). Enterococci were the second most frequent overall and conversely, there was a significant increase in their frequency over time (P < 0.001). Proteus spp, Coliforms (Klebsiella, Enterobacter and Serratia spp) and Pseudomonas spp were the third, fourth and fifth most frequent overall pathogens, with no evidence of a change with time. Staphylococci spp were the sixth most frequent and there was a gradual reduction in their frequency with time (P = 0.02).

Table 1. The results of CSUs, the frequency of bacterial pathogens, and the frequency of CAUTIs resistant to antibiotics by calendar year
Variable199619982001All years
  1. a ≥3 pathogens identified (no quantitative culture); b bacteriuria with two pathogens each of > 103 cfu/mL urine (two isolates); c 103 cfu/mL urine of one pathogen (one isolate). * P  < 0.05; P  < 0.01,P < 0.001 for values with change over time; n, number of CAUTI isolates with a specific organism per year. K, S, E, Klebsiella, Enterobacter, Serratia spp.

CSU
No growth1513148921075109
Mixed growtha  432  440  4981370
Bacterial CAUTIs  535  471  6711677
 polymicrobial, n (%)b    30 (6)    62 (15)  153 (23)  245 (15)
 monomicrobial, n (%)c  505 (94)  409 (85)  518 (77)1432 (85)
Candida CAUTIs    61    60    72  193
Total CSUs2541246033408341
Frequency of pathogens, %
No.  565  533  8241922
E. coli    35.5    32.4    26.6    30.9
Enterococci spp    11.8    15.3    22.0    17.2
Proteus spp    14.1    14.6    17.1    15.6
Coliforms (K, S, E)     11.8    13.8     11.9    12.4
Pseudomonas sp.     11.8     11.4    10.8     11.2
Staphylococci sp.    12.0      9.1      7.9*      9.5
Others      2.6      3.0      3.1      2.9
Frequency of CAUTIs resistant to antibiotics, %
E. coli, n  201  173  220  594
Co-amoxiclav    10.4    19.0    18.1*    15.8
Amoxycillin    57.2    57.2    56.3    57.0
Cephalexin      3.4      3.4      7.2      5.0
Ciprofloxacin      3.4      4.6      9.0*      5.9
Nitrofurantoin      5.9      5.7      5.4      5.7
Trimethoprim    37.8    35.8    34    35.8
Enterococci spp, n    67    82  182  331
Co-amoxiclav      1.4      6.1      4.9      4.5
Amoxycillin      2.9      2.4      4.9      3.1
Cephalexin    98.5  100    99.4    99.3
Ciprofloxacin    40.2    48.7    84.0*    66.4
Nitrofurantoin      5.9      3.6      1.0      2.7
Trimethoprim    40.2    39.0    37.3    38.3
Proteus spp, n    77    78  142  297
Co-amoxiclav    15.5    23.0    21.1    20.2
Amoxycillin    51.9    48.7    39.4    45.1
Cephalexin    15.5    21.7    18.0    18.5
Ciprofloxacin      1.3      5.1      4.2      3.7
Nitrofurantoin    98.7  100  100    99.6
Trimethoprim    54.5    47.4    45.8    48.4
Coliforms (K, S, E), n    66    77    99  242
Co-amoxiclav    24.2    29.7    29.2    27.6
Amoxycillin    92.4    94.5    95.9    93.3
Cephalexin    21.2    27.0    31.3    26.8
Ciprofloxacin      3.03      4.0      4.0      3.7
Nitrofurantoin    48.4    50.0    38.3    44.2
Trimethoprim    18.1    22.9    31.3*    24.7
Pseudomonas spp, n    67    61    89  217
Ciprofloxacin      4.47      3.2      4.4      4.1
Ceftazidime      0      1.6      1.1      0.9
Gentamicin      0      1.6      2.2      1.3
Piperacillin      0      6.5      6.0      4.1
Meropenem      1.4      0      0      0.4
Staphylococci spp, n    68    49    68  185
Co-amoxiclav    52.9    53.0    66.1    59.4
Amoxycillin    95.5    81.6    94.1    91.3
Flucloxacillin    61.7    65.3    76.4    68.1
Nitrofurantoin      0      0      1.4      0.5
Trimethoprim    57.3    51.0    30.8*    45.9

For uropathogen resistance to antibiotics, the trends are shown in Table 1. Over the 5-year period, only a small percentage of E. coli were resistant to nitrofurantoin, cephalexin, ciprofloxacin and co-amoxiclav. However, over this period resistance to co-amoxiclav (P = 0.044) and ciprofloxacin (P = 0.014) increased, with no evidence of change in resistance to other antibiotics.

Enterococci spp were infrequently resistant to nitrofurantoin, amoxycillin and co-amoxiclav. The resistance to ciprofloxacin increased significantly over time (P < 0.001) with no evidence of change in resistance to other antibiotics.

Proteus spp were infrequently resistant to ciprofloxacin and more frequently resistant to cephalexin and co-amoxiclav. There was no evidence of change in resistance to antibiotics over time.

Coliforms were infrequently resistant to ciprofloxacin only. The relatively frequent resistance to trimethoprim increased further over the study period (P = 0.045). There was no evidence of change in the resistance to other antibiotics.

Pseudomonas spp were rarely resistant to any of the antibiotics tested and there was no evidence of a change over time.

Staphylococci spp were hardly ever resistant to nitrofurantoin. The frequency of resistance to trimethoprim reduced over time but remained high (P = 0.002). There was an increase in resistance to flucloxacillin (methicillin) but it was not statistically significant (P = 0.062). There was no evidence of change in resistance to other antibiotics.

The overall bacterial resistance trends are shown in Table 2. In 1996, CAUTIs were least often resistant to ciprofloxacin followed, in ascending order, by co-amoxiclav and cephalexin. However, in 2001, CAUTIs were least often resistant to co-amoxiclav followed by ciprofloxacin and nitrofurantoin. This change in resistance patterns was the result of the increase in resistance to ciprofloxacin and cephalexin (both P < 0.001) with no change in resistance to co-amoxiclav or nitrofurantoin over time. CAUTIs were frequently resistant to trimethoprim with no evidence of change over time. In contrast, the overall resistance to amoxycillin declined with time (P < 0.001) but remained high.

Table 2. The pattern of total CAUTIs (%) resistant to antibiotics
Antibiotic199620011998 P All years
  • *

    Pseudomonas isolates excluded;

  • Staphylococci isolates excluded;

  • Staphylococci and Pseudomonas isolates excluded.

Co-amoxiclav*18.522.522.90.1521.4
Ciprofloxacin  8.0527.213.2< 0.00117.8
Nitrofurantoin*27.628.828.60.6428.4
Trimethoprim*39.837.537.90.4738.3
Cephalexin25.440.230.7< 0.00134.3
Amoxycillin*58.949.154.8< 0.00153.5

DISCUSSION

There was a significant change in the bacterial spectrum and antimicrobial resistance of CAUTIs over 5 years (1996–2001) in a UK institution. First, the frequency of polymicrobial (two isolated pathogens) CAUTIs has increased, and second, the frequency of E. coli and Staphylococci CAUTIs has decreased with time. Enterococci have become more prevalent.

There has been a change in the antimicrobial resistance profile of various organisms. E. coli resistance to co-amoxiclav and ciprofloxacin has increased, and Enterococcal resistance to ciprofloxacin has doubled. The resistance of total CAUTIs to ciprofloxacin and cephalexin has increased, but to co-amoxiclav and nitrofurantoin remained unchanged over time. As a result of all these changes, CAUTIs in 2001 were more frequently sensitive to co-amoxiclav, with a sensitivity rate of 77.5% followed by ciprofloxacin (72.8%) and nitrofurantoin (71.2%).

METHODOLOGICAL LIMITATIONS

First, we could not accurately characterize the patient population; information on the details of the patients’ conditions, their comorbidities and the duration of their catheterization would have been useful, but these characteristics can only be captured in a prospective study. The second concern relates to generalisability; the criteria for defining CAUTI and antimicrobial testing might differ among institutions. Prescribing habits and case-mix will undoubtedly influence the patterns of bacterial resistance. The extent to which this is the case can only be confirmed once similar studies in other settings have reported their findings. A comparable study was carried out by Bonadio et al.[15] in an Italian teaching hospital covering the period 1996–99. The study findings were similar to the present in some aspects but different in others. The frequency of E. coli CAUTI in their institution was 30%, which is similar to the present overall frequency (31%). However, the overall frequency in their study was higher for Pseudomonas spp and lower for Enterococci (24% vs 11% and 13% vs 17%, respectively). The pattern of overall antimicrobial resistance was very similar for amoxicillin and ciprofloxacillin, but overall resistance to nitrofurantoin was less frequent than in the present study (11% vs 28%). Unless prospective multicentre studies are conducted it is difficult to explore the reason behind these differences, because of the methodological limitations described above.

CLINICAL IMPLICATIONS

The increase in polymicrobial infections could make CAUTIs more difficult to treat, especially while catheters remain in situ[16]. It might require either an antibiotic to which the two cultured organisms are sensitive or two different antibiotics might be needed. The changes in resistance over time means that currently, according to the present results, the most appropriate agents for CAUTIs are co-amoxiclav, ciprofloxacin and nitrofurantoin. About three-quarters of all CAUTIs are sensitive to these antibiotics. However, the increase in resistance of E. coli and Enterococci to ciprofloxacin (the most commonly used fluoroquinolone) raises concern about the appropriateness of its use in the empirical management of all CAUTIs. At present, Enterococci are the causative agent of about a quarter of all CAUTIs and they are almost always resistant to ciprofloxacin.

This increase in resistance to fluoroquinolones has been reported in other studies in different countries [16–18]. We believe (as do others) that the overuse of fluoroquinolones in the last few years has contributed to this rise in resistance, which is likely to increase further in the future as use increases [16–18]. However, a recent meeting organized by the National Institute of Allergy and Infectious Diseases in the USA seemed to promote the use of fluoroquinolones as the first empirical treatment for all complicated UTIs, including CAUTIs [19,20]. Whilst this policy might be practical, careful monitoring will be required to ensure that the resistance of CAUTIs to fluoroquinolones does not increase.

The frequency of E. coli and Enterococci is high in short-term catheterization [21]; nitrofurantoin has maintained its effectiveness over time against both organisms. From the present results, nitrofurantoin seems therefore a good choice for the empirical management of CAUTIs in a hospital setting where short-term catheterization is the norm. Proteus spp are frequently encountered in patients with long-term catheters [21]. Ciprofloxacillin, which is highly effective against this pathogen, seems to be a good choice for the empirical management of CAUTIs in these patients.

Abbreviations
CAUTI

catheter-associated urinary tract infection

CSU

catheter specimens of urine

cfu

colony-forming unit.

Ancillary