- Email: [email protected]
A UK Multicentre Study of the Antimicrobial Susceptibility of Bacterial Pathogens Causing Urinary Tract Infection
Journal of Infection (2003) 46: 94±100 doi:10.1053/jinf.2002.1091, available online at http://www.idealibrary.com on
A UK Multicentre Study of the Antimicrobial Susceptibility of Bacterial Pathogens Causing Urinary Tract Infection D. J. Farrell*, I. Morrissey, D. De Rubeis, M. Robbins and D. Felmingham GR Micro Limited, London, UK Objectives: To determine the prevalence of resistance amongst urinary tract pathogens against antimicrobials used to treat urinary tract infections (UTIs) in the UK to provide data to help direct empirical therapy. Method: During 1999±2000, a total of 1291 bacterial isolates causing UTI were collected from 8 centres in the UK. Isolates were cultured from patients with (1) community-acquired UTI in those less than 65 years old (397), (2) hospital-acquired UTI other than those admitted with pyelonephritis (394), (3) pyelonephritis (108) and (4) community-acquired UTI in those greater than 65 years old (392). After re-identification, MICs for a range of antimicrobials were determined and interpreted using NCCLS procedures and interpretive guidelines. Results: Escherichia coli was the predominant pathogen in all categories but the total percentage for each category varied (56.3±77.3%). The next three pathogens of importance were Enterococcus faecalis, Klebsiella pneumoniae and Proteus mirabilis which varied in prevalence slightly from category to category. The activity of amoxycillin against E. coli (51.3% susceptible) was greatly reduced as a result of b-lactamase production and only partially restored by the addition of clavulanic acid (78.8% susceptible). Cefuroxime was very active against E. coli using parenteral form breakpoints (97.1% susceptible) but less so using oral form breakpoints (68.6% susceptible). Cefuroxime was inactive against Enterococcus spp. and Pseudomonas spp. Nitrofurantoin was very active against isolates of E. coli (96.3% susceptible) and E. faecalis but not against K. pneumoniae, P. mirabilis or Pseudomonas aeruginosa. Overall susceptibility to trimethoprim ranged from 58.1% to 84.5% for the most prevalent pathogens. Ciprofloxacin was highly active against the UTI pathogens examined in this study with susceptibilities of between 88.6% and 97.7% for the most prevalent pathogens (E. coli, n 864, 97.7% susceptible) and was the only oral agent tested with activity against Pseudomonas spp. Conclusion: These data provide much needed information on the prevalence of antimicrobial resistance amongst pathogens currently causing UTI in the UK. # 2002 The British Infection Society. Published by Elsevier Science Ltd. All rights reserved.
Introduction Urinary tract infections (UTIs) are amongst the most common infections described in the outpatient setting [1,2]. UTI are also the most common infection in acute and long-term care hospital patients and in patients in nursing homes [3,4]. The most commonly isolated bacterial pathogen, whether in community-acquired or hospitalized patient UTI, is Escherichia coli [5±7]. Increasing resistance in bacterial pathogens is of worldwide concern. The prevalence of antimicrobial resistance in both out- and hospital patients with UTI is increasing and can vary according to geographical and * Please address all correspondence to: D. J. Farrell, GR Micro Ltd, 7-9, William Road, London, NW1 3ER, UK. Tel.: 44(0)207-388-7320; Fax: 44(0)207-388-7324; E-mail address: [email protected] (D. J. Farrell). 0163±4453/03/$35.00
regional location [8,9]. In almost all cases of UTI, whether outpatient or hospital patient, antimicrobial therapy is initiated empirically before the results of urine culture are available. Hence, there exists a great need for antimicrobial resistance surveillance at local, national and international levels. The aim of this study was to determine the distribution of uropathogens and their antimicrobial susceptibility patterns in a multi-site study in the UK. To reduce the bias due to different pathogen distribution and antimicrobial susceptibility profile associated with the type of UTI and in/outpatient status, patients were assigned to 1 of 4 categories according to age and clinical diagnosis. These categories were: (1) community-acquired UTI in those less than 65 years old, (2) hospital-acquired UTI other than those admitted with pyelonephritis, (3) pyelonephritis, and (4) community-acquired UTI in those greater than 65 years old.
# 2002 The British Infection Society. Published by Elsevier Science Ltd. All rights reserved.
UK Surveillance of UTI Pathogens
Materials and Methods
95
Results
Participating centres During 1999±2000, 8 centres geographically spread throughout the UK took part in the study: St. Mary's Hospital (London), Southmead Hospital (Bristol), Leicester Royal Infirmary (Leicester), Leeds General Infirmary (Leeds), the Royal Victoria Hospital (Newcastleupon-Tyne), the Southern General Hospital (Glasgow), Raigmore Hospital (Inverness) and the Royal Hospitals (Belfast). Patients and bacterial isolates Each centre was required to submit 200 bacterial uropathogens, cultured from the urine of patients categorized into one of four infection types (i.e. 50 isolates from each category), to the central testing laboratory (GR Micro Ltd, London). Repeat isolates from the same patient, cultured within 2 weeks of the first isolation, were excluded from the collection. Each isolate was designated as a `pathogen' at the discretion of the centre performing urinalysis. Isolates were stored at 70 C at the centre until a batch of sufficient size was ready for transportation to the Central Testing Laboratory. Identification and susceptibility testing Bacterial isolates were identified by biochemical profiling using API systems (bioMeÂrieux, Basingstoke, UK). MICs were determined using the National Committee for Clinical Laboratory Standards (NCCLS) agar dilution method [10]. Breakpoints used for interpretation of MIC results were those specified by the NCCLS [11].
The overall female to male ratio was 4.1:1, ranging from 2.6:1 in hospitalized patients to 6.8:1 in patients less than 65 years old with community-acquired UTI (Table I). The mean age was 47 years for patients with pyelonephritis compared to 27 years for patients less than 65 years old with community-acquired UTI (Table I). In total, 1291 bacterial isolates were examined in the study. E. coli was the most common urinary pathogen identified. The relative frequency of E. coli was lowest in non-pyelonephritis hospital patients (56.3%), presumably due to the increased frequency of a wide range of other nosocomial pathogens, and highest in patients less than 65 years old with community-acquired UTI (Table II). Enterococcus faecalis, Klebsiella pneumoniae, Proteus mirabilis and Pseudomonas aeruginosa were the next most common species isolated with the rank order varying between infection categories (Table II). Other isolates of importance included Staphylococcus saprophyticus (responsible for 2.0% of infections in patients less than 65 years old with community-acquired UTI) and Klebsiella oxytoca (2.8% of patients with pyelonephritis). The activity of amoxycillin against E. coli (51.3% susceptible) was greatly reduced as a result of blactamase production and only partially restored by the addition of clavulanic acid (78.8% susceptible) (Table III). As expected, K. pneumoniae isolates were almost fully resistant in vitro to amoxycillin. Cefuroxime was very active against the E. coli using parenteral form breakpoints (97.1% susceptible) (Table IV) but not using oral form breakpoints (68.6% susceptible). As expected, cefuroxime was inactive against Enterococcus faecalis or Pseudomonas aeruginosa. (Table III). The activities of cephradine, tetracycline, sulphamethoxazole and
Table I. Gender and age distribution by category. Infection category
Community-acquired UTI (patients less than 65 years old) UTI in hospital patients other than those admitted with pyelonephritis Pyelonephritis Community-acquired UTI (patients greater than 65 years old) All categories combined
Number of isolates
Female/male (ratio)
Age range (years)
Mean age (years)
397
345/51* (6.8:1)
<1±64
27
394
286/108 (2.6:1)
<1±99
33
108 392
87/21 (4.1:1) 321/71 (4.5:1)
<1±92 65±97
47 78
1291
1039/251 (4.1:1)
<1±99
57
*One patient of unknown gender is omitted from the analysis.
96
D. J. Farrell et al. Table II. Relative frequency (% of total) of isolation of species causing various types of UTI. Community-acquired UTI (patients less than 65 years) (n 397)
Hospital patient UTI (excluding pyelonephritis) (n 394)
Pyelonephritis (n 108)
Organism
%
Organism
%
Organism
%
Organism
%
E. coli P. mirabilis E. faecalis K. pneumoniae S. saprophyticus P. aeruginosa S. agalactiae Other species
77.3 4.3 3.8 3.5 2.0 1.8 1.5 5.8
E. coli E. faecalis K. pneumoniae P. mirabilis P. aeruginosa E. cloacae S. aureus E. faecium K. oxytoca Other species
56.3 8.4 6.9 6.3 3.8 2.5 2.5 2.0 2.0 9.1
E. coli K. pneumoniae E. faecalis P. mirabilis K. oxytoca Other species
68.5 6.5 5.6 2.8 2.8 13.9
E. coli E. faecalis K. pneumoniae P. mirabilis P. aeruginosa E. cloacae K. oxytoca Other species
66.6 5.9 5.9 4.3 3.1 2.8 1.8 9.7
trimethoprim were low against E. coli and these agents provided no, or limited, coverage against some of the other prevalent pathogens (Table III). Nitrofurantoin was very active against isolates E. coli (96.3% susceptible) and E. faecalis (100%) but not against K. pneumoniae, P. mirabilis or P. aeruginosa. Ciprofloxacin was highly active against the UTI pathogens examined in this study with susceptibilities of between 87.1% and 97.7% for the most prevalent pathogens (E. coli, n 864, 97.7% susceptible). For all pathogens, regardless of species, ciprofloxacin had the lowest MIC50 (0.015 mg/L) and MlC90 (1 mg/L) for all of the orally available antibiotics (Table III). Ciprofloxacin was the only oral agent tested with activity against P. aeruginosa (88.6% susceptible). All of the parenterally available antibiotics had excellent activity (87.3±100%) against the most prevalent Enterobacteriaceae (Table IV). Escherichia coli was the only species to have sufficient numbers to obtain valid percent susceptible data by infection category (Table V). Susceptibility patterns were similar between the categories, however susceptibilities for all oral agents were lower in isolates from hospital patients (category 2) than outpatients. Also, the susceptibility of E. coli to trimethoprim was considerably lower in patients with pyelonephritis compared to the other categories of infection (64.9% susceptible compared to 72.8±76.5% for the other 3 categories). Within the whole population, 19 multiresistant isolates were found: 12 Enterobacteriaceae extendedspectrum beta-lactamases (ESBLs), two vancomycinresistant enterococci (VRE) (both vanA phenotype) and five methicillin-resistant S. aureus (MRSA). All but one isolate were obtained from hospitalized patients, however one MRSA isolate was from an elderly patient in the community. The multidrug-resistant Enterobacteriaceae
Community-acquired UTI (patients greater than 65 years) (n 392)
(K. pneumoniae 6, E. coli 4, K. oxytoca 1, E. cloacae 1) were all extremely resistant to most oral antibiotics, except 4 isolates which were susceptible to ciprofloxacin and 2 isolates which were susceptible to amoxycillin/ clavulanate. All 12 were susceptible to imipenem and amikacin but not ceftazidime, gentamicin or cefuroxime when using cefuroxime sodium breakpoints. One VRE was susceptible to amoxycillin and imipenem, the other resistant to both agents. Both VRE were resistant to all other agents tested. All 4 MRSA were susceptible to gentamicin, amikacin, trimethoprim, tetracycline and sulphamethoxazole but resistant to all other agents. A total of 37/40 (92.5%) of S. saprophyticus isolates were found in women, with 31/37 (83.8%) in patients less than 65 years of age with uncomplicated UTI or pyelonephritis, and 24/37 (64.9%) of the patients aged between 16 and 35 years old.
Discussion This study shows the distribution of microbial species causing UTIs in the UK during 1999±2000 and their susceptibility patterns to the most commonly used oral and parenteral antimicrobial agents. To reduce bias in the data, UTIs were categorized into 4 categories of infection based on simple or complicated UTI, hospital status, patient age and a clinical diagnosis of pyelonephritis. To reduce technical variation, all isolates were re-identified and had susceptibility tests performed at a central reference laboratory using NCCLS methodology and interpretation [10,11]. Overall, 74 different species of bacteria were identified in the 1291 patients with UTI with between 29 and 49 species being found in each category. This variation in
2
1
NA 128 128 16±>64 NA 8 32 0.015±>64
NA 16 >64 0.015±>64
96.8 1 8 0.5±32
72.6 1 128 0.5±64
NA >64 >64 16±>64
84.5 2 16 1±32
100 0.25 0.5 0.25±1
100 0.25 0.5 0.25±1
1.4 >64 >64 1±128
78.8 8 16 1±64
51.3 8 >64 1±>64
NA 8 128 0.015±>128
NA >128 >128 All > 128
4.8 16 32 8±128
NA 4 32 0.015±>128
NA >128 >128 All > 128
95.2 2 4 0.5±>128
81.7 4 32 1±128
NA 16 >128 2±>128
NA2 16 128 16±128 71.8 8 32 4±>128
68.6 4 8 0.25±>128
Cefuroxime1
65.3 8 16 2±>128
Amoxycillin Amoxycillin/ Cephradine clavulanate
Using NCCLS breakpoints for the oral form of cefuroxime (cefuroxime axetil). NA not applicable. No NCCLS breakpoint available. n Number of isolates.
P. aeruginosa (35, 2.7%) % susceptible MIC50 MIC90 Range All isolates combined % susceptbile MIC50 MIC90 Range
K. pneumoniae (71, 5.5%) % susceptible MIC50 MIC90 Range P. mirabilis (62, 4.8%) % susceptible MIC50 MIC90 Range
E. coli (864, 66.9%) % susceptible MIC50 MIC90 Range E. faecalis (77, 6.0%) % susceptible MIC50 MIC90 Range
Organism (n, % of UTI)
NA 2 >64 0.12±>64
0 32 32 16±>64
0 64 >64 32±>64
81.7 2 >64 1±>64
29.9 64 >64 0.25±>64
74.4 2 >64 0.5±>64
NA 32 >512 0.5±>512
57.1 256 512 64±>512
87.1 16 >512 4±>512
84.5 32 >512 8±>512
NA >512 >512 All > 512
62.3 32 >512 1±>512
NA 0.5 >64 0.12±>64
NA >64 >64 32±>64
58.1 4 >64 1±>64
84.5 1 >64 0.25±>64
NA 0.5 >64 0.12±>64
73.3 0.5 >64 0.12±>64
NA 32 128 2±>512
NA >512 >512 All > 512
0 128 128 64±512
15.5 128 256 16±512
100 8 8 4±16
96.3 32 32 2±256
NA 4 >64 1±>64
NA >64 >64 All > 64
83.9 4 >64 2±>64
81.7 4 >64 2±>64
NA >64 >64 All > 64
94.7 2 4 1±>64
NA 0.015 1 0.002±>64
88.6 0.25 2 0.03±16
87.1 0.03 2 0.015±4
94.4 0.03 0.5 0.008±32
NA 1 64 0.25±64
97.7 0.015 0.015 0.002±64
Tetracycline Sulphamethoxazole Trimethoprim Nitrofurantoin Nalidixic acid Ciprofloxacin
Table III. Percent susceptible, MIC50, MIC90 and range (in mg/L) for orally administered antimicrobial agents against the most prevalent uropathogens.
UK Surveillance of UTI Pathogens 97
98
D. J. Farrell et al.
Table IV. Percent susceptible, MIC50, MIC90 and range (in mg/L) for parenterally-administered antimicrobial agents against the most prevalent uropathogens. Organism (n, % of UTI)
Ceftazidime
Imipenem
Gentamicin
Amikacin
Cefuroxime1
E. coli (864, 66.9%) % susceptible MIC50 MIC90 Range
99.3 0.12 0.25 0.06±>128
100 0.12 0.25 0.06±1
98.8 0.25 0.5 0.12±>64
100 1 2 0.25±16
97.1 4 8 0.25±>128
E. faecalis (77, 6.0%) % susceptible MIC50 MIC90 Range
NA2 64 >128 8±>128
NA 1 2 0.25±4
NA 8 >64 1±>64
NA >64 >64 8±>64
NA 16 >128 2±>128
K. pneumoniae (71, 5.5%) % susceptible MIC50 MIC90 Range
91.6 0.12 1 0.06±>128
100 0.12 0.25 0.06±0.5
91.6 0.25 0.5 0.12±64
100 1 1 0.5±16
87.3 4 32 1±128
P. mirabilis (62, 4.8%) % susceptible MIC50 MIC90 Range
100 0.06 0.12 0.06±0.25
100 0.5 2 0.06±4
100 0.5 1 0.12±1
100 2 4 0.5±4
98.4 2 4 0.5±>128
P. aeruginosa (35, 2.7%) % susceptible MIC50 MIC90 Range
60.0 8 32 1±>128
97.1 2 2 0.12±16
97.1 2 4 0.25±32
100 4 8 0.5±8
NA >128 >128 All > 128
All isolates combined % susceptible MIC50 MIC90 Range
NA 0.25 16 0.06±>128
NA 0.12 1 0.008±>64
NA 0.25 4 0.03±>64
NA 1 4 0.12±>64
NA 4 32 0.015±>128
1 2
Using NCCLS breakpoints for the parenteral form of cefuroxime (cefuroxime sodium). NA not applicable. No NCCLS breakpoint available.
the number of species involved in UTI is greater than a recent report of 32 species identified in a study of 1510 hospital patients (43 species were found in hospitalized patients in our study with a population size of 394 patients) [8]. E. coli was the most frequent uropathogen in all categories of infection. The frequency of E. coli isolated (66.9% in total, category range 56.3±77.3%) is lower than a recently reported frequency for the UK of 81.2% [12]. E. faecalis was the second most common isolate causing 6.0% of infections although the frequency of the most prevalent Enterobacteriaceae combined were greater. E. faecalis was more frequently isolated in hospitalized patients than non-hospitalized. Staphylococcus saprophyticus is a known opportunistic pathogen predominantly in young, sexually active females [13]. Recent reports have found this pathogen to be present in 9.7% (n 5739) and 4.0% (n 4342) of women with uncomplicated cystitis [9,13]. In our study, S. saprophyticus infection in women less than
65 years old with community-acquired UTI, therefore, is demographically similar but lower in frequency (2.0%) than previous reports [9,13]. A recent European study reported an ampicillin susceptibility of 58.9% for E. coli from 246 patients in the UK [12]. The rate of susceptibility to amoxycillin in our study (total 51.3%, category range 48.5±55.6%) is slightly lower, but similar to this percentage. The same study reported an amoxycillin/clavulanate susceptibility of 96.3% for UK patients [12]. This rate is vastly different from our rate (78.8%) but similar to recently reported data from Canada (95%, n 694) [14]. Possible reasons for this difference are geographical variation due to a higher prevalence of variants of b-lactamase resistant E. coli (such as hyper-b-lactamase producers with TEMmediated resistance) or methodological discrepancies. For example, the study reporting low levels of amoxycillin/clavulanate resistance in the UK did not perform MICs using NCCLS methodology, as in our study, but
307 222 74 261 864
1 2 3 4 All categories
Ceftazidime
100 97.3 100 100 99.3
100 100 100 100 100
Imipenem
Parenteral agents
Gentamicin 100 97.3 97.3 99.2 98.8
Amikacin 100 100 100 100 100
Cefuroxime1 97.1 93.7 98.6 98.1 96.6
Oral agents
48.5 48.6 55.4 55.6 51.3
Amoxycillin
1
n Number of isolates. Using NCCLS breakpoints for the parenteral form of cefuroxime (cefuroxime sodium). 2 Using NCCLS breakpoints for the oral form of cefuroxime (cefuroxime axetil). 3 NCCLS breakpoints for the parenteral and oral forms of ciprofloxacin are identical.
n
Category
78.8 74.3 86.5 80.5 78.8
Amoxycillin/clavulanate
Table V. Percent susceptibility of E. coli isolates from each category of infection.
Cephradine 68.4 58.1 78.4 64.0 65.3
Cefuroxime2 71.0 60.8 78.4 69.7 68.6
Tetracycline 74.9 73.0 75.7 75.1 74.5
Sulphamethoxazole 65.5 57.7 59.5 63.2 62.3
Trimethoprim 76.5 72.1 64.9 72.8 73.3
Nitrofurantoin 97.7 95.9 95.9 95.0 96.3
Nalidixic acid 96.7 91.9 90.5 95.8 94.7
Ciprofloxacin 3 99.3 91.9 93.2 98.5 95.8
UK Surveillance of UTI Pathogens 99
100
D. J. Farrell et al.
used disc susceptibility testing [12]. The decision to prescribe amoxycillin/clavulanate empirically, therefore, requires consideration of current local susceptibility data. Comparative studies between different amoxycillin/clavulanate susceptibility testing methodologies are needed to increase confidence in the reliability of these results. The two oral cephalosporins (cephradine and cefuroxime), trimethoprim, sulphamethoxazole, and tetracycline did not have good in vitro coverage for many of the uropathogens isolated in this study. Because of this, they are of little use as empirical therapeutic options. In general, the parenteral agents showed excellent in vitro coverage of the uropathogens isolated in this study. This information will be useful in choosing empiric therapy for seriously ill hospitalized patients with suspected urosepsis. Although nitrofurantoin demonstrated excellent in vitro activity against E. coli and E. faecalis (96.3% and 100% susceptible, respectively) it demonstrated little or no activity against other prevalent Enterobacteriaceae and P. aeruginosa. Also, although debateable, nitrofurantoin is not traditionally used in urinary tract infections with any significant tissue involvement (such as pyelonephritis). Nalidixic acid demonstrated excellent in vitro activity against E. coli (94.7% susceptible), good activity against other prevalent Enterobacteriaceae but no activity against Gram-positive organisms or P. aeruginosa. In addition, as with nitrofurantoin, nalidixic acid has poor tissue penetration. Overall, ciprofloxacin demonstrated the greatest in vitro activity against the widest range of uropathogens for all of the oral agents tested with susceptibilities of between 87.1% and 97.7% for the most prevalent pathogens (E. coli, n 864, 97.7% susceptible). This activity was consistent across each of the infection categories. Importantly, ciprofloxacin was the only oral agent tested with activity against Pseudomonas spp. Although no NCCLS breakpoints exist for ciprofloxacin against E. faecalis, ciprofloxacin demonstrated activity against this species (MIC50 1 mg/L). In conclusion, these data provide much needed information on the prevalence of antimicrobial resistance amongst pathogens currently causing UTI in the UK. This data was obtained using standardized (NCCLS) methodology and has been divided into infection type categories to remove bias. Overall, ciprofloxacin demonstrated the most complete in vitro coverage amongst the antimicrobials tested and hence emphasizing its great utility as a choice for empiric therapy of UTI in the UK. As with other agents, increased use results in concern of resistance development. Prudent prescribing, establishment and use of infection control guidelines,
local monitoring of resistance and longitudinal surveillance studies are necessary to minimized and monitor the development of resistance. We hope to repeat this study 2002/2003 to monitor the development of resistance in the UK. Acknowledgments We thank Bayer AG for sponsoring this study and, as always, we are grateful to the collecting centres for their participation in the study.
References 1 Stamm WE, Hooton TM. The management of urinary tract infection. N Engl J Med 1993; 329: 1328±1334. 2 Warren JW, Abrutyn E, Hebel JR, Johnson JR, Schaeffer AJ, Stamm WE. Guidelines for antimicrobial treatment of uncomplicated acute bacterial cystitis and acute pyelonephritis in women. Clin Infect Dis 1999; 29: 745±758. 3 Gastmeier P. Nosocomial urinary tract infections: many unresolved questions. Clin Micro Infect 2001; 7: 521±522. 4 Nicolle LE, Strausbaugh LJ, Garibaldi RA. Infections and antibiotic resistance in nursing homes. Clin Microbiol Rev 1996; 9: 1±17. 5 GruÈneberg RN. Changes in urinary pathogens and their antibiotic sensitivities. 1971±1992. J Antimicrob Chemother 1994; 33(Suppl. A): 1±8. 6 MacGowan AP, Brown NM, Holt HA, Lovering AM, McCulloch SY, Reeves DS. An eight-year survey of the antimicrobial susceptibility patterns of 85,971 bacteria isolated from patients in a district general hospital and the local community. J Antimicrob Chemother 1993; 31: 543±557. 7 Barrett SP, Savage MA, Rebec MP, Guyot A, Andrews N, Shrimpton SB. Antibiotic sensitivity of bacteria associated with community-acquired urinary tract infection in Britain. J Antimicrob Chemother 1999; 44: 359±365. 8 Mathai D, Jones RN, Pfaller MA. Epidemiology and frequency of resistance among pathogens causing urinary tract infection in 1,510 hospitalized patients: a report from the SENTRY Antimicrobial Surveillance Program (North America). Diag Microbiol Infect Dis 2001; 40: 129±136. 9 Karlowsky JA, Jones ME, Thornsberry C, Critchley I, Kelly L, Sahm D. Prevalence of antimicrobial resistance among urinary tract pathogens isolated from female outpatients across the US in 1999. Int J Antimicrob Agents 2001; 18: 121±127. 10 NCCLS (2000). Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically; Approved StandardÐ5th edn. NCCLS document M7-A5 (ISBN 1-56238-394-9). NCCLS, Wayne, Pennsylvania 19087-1898, USA. 11 NCCLS (2002). Performance standards for antimicrobial susceptibility testing; Twelfth informational supplement. NCCLS document M100-12 (ISBN 1-56238-454-6) NCCLS, Wayne, Pennsylvania 19087-1898, USA. 12 Kahlmeter G. The ECO-SENS project: a prospective, multinational, multicentre epidemiological survey of the prevalence and antimicrobial susceptibility of urinary tract pathogens. J Antimicrob Chemother 2000; 46(Suppl. S1): 15±22. 13 Gupta K, Scholes D, Stamm WE. Increasing prevalence of antimicrobial resistance among uropathogens causing acute uncomplicated cystitis in women. JAMA 1999; 281: 736±738. 14 Blondeau JM, Yaschuk Y, Suter M, Vaughan D, and the Canadian Antimicrobial Study Group. In-vitro susceptibility of 1982 respiratory pathogens and 1921 urinary tract pathogens against 19 antimicrobial agents: a Canadian multicentre study. J Antimicrobial Chemotherapy 1999; 43(Suppl. A): 3±23.
A UK Multicentre Study of the Antimicrobial Susceptibility of Bacterial Pathogens Causing Urinary Tract Infection D. J. Farrell*, I. Morrissey, D. De Rubeis, M. Robbins and D. Felmingham GR Micro Limited, London, UK Objectives: To determine the prevalence of resistance amongst urinary tract pathogens against antimicrobials used to treat urinary tract infections (UTIs) in the UK to provide data to help direct empirical therapy. Method: During 1999±2000, a total of 1291 bacterial isolates causing UTI were collected from 8 centres in the UK. Isolates were cultured from patients with (1) community-acquired UTI in those less than 65 years old (397), (2) hospital-acquired UTI other than those admitted with pyelonephritis (394), (3) pyelonephritis (108) and (4) community-acquired UTI in those greater than 65 years old (392). After re-identification, MICs for a range of antimicrobials were determined and interpreted using NCCLS procedures and interpretive guidelines. Results: Escherichia coli was the predominant pathogen in all categories but the total percentage for each category varied (56.3±77.3%). The next three pathogens of importance were Enterococcus faecalis, Klebsiella pneumoniae and Proteus mirabilis which varied in prevalence slightly from category to category. The activity of amoxycillin against E. coli (51.3% susceptible) was greatly reduced as a result of b-lactamase production and only partially restored by the addition of clavulanic acid (78.8% susceptible). Cefuroxime was very active against E. coli using parenteral form breakpoints (97.1% susceptible) but less so using oral form breakpoints (68.6% susceptible). Cefuroxime was inactive against Enterococcus spp. and Pseudomonas spp. Nitrofurantoin was very active against isolates of E. coli (96.3% susceptible) and E. faecalis but not against K. pneumoniae, P. mirabilis or Pseudomonas aeruginosa. Overall susceptibility to trimethoprim ranged from 58.1% to 84.5% for the most prevalent pathogens. Ciprofloxacin was highly active against the UTI pathogens examined in this study with susceptibilities of between 88.6% and 97.7% for the most prevalent pathogens (E. coli, n 864, 97.7% susceptible) and was the only oral agent tested with activity against Pseudomonas spp. Conclusion: These data provide much needed information on the prevalence of antimicrobial resistance amongst pathogens currently causing UTI in the UK. # 2002 The British Infection Society. Published by Elsevier Science Ltd. All rights reserved.
Introduction Urinary tract infections (UTIs) are amongst the most common infections described in the outpatient setting [1,2]. UTI are also the most common infection in acute and long-term care hospital patients and in patients in nursing homes [3,4]. The most commonly isolated bacterial pathogen, whether in community-acquired or hospitalized patient UTI, is Escherichia coli [5±7]. Increasing resistance in bacterial pathogens is of worldwide concern. The prevalence of antimicrobial resistance in both out- and hospital patients with UTI is increasing and can vary according to geographical and * Please address all correspondence to: D. J. Farrell, GR Micro Ltd, 7-9, William Road, London, NW1 3ER, UK. Tel.: 44(0)207-388-7320; Fax: 44(0)207-388-7324; E-mail address: [email protected] (D. J. Farrell). 0163±4453/03/$35.00
regional location [8,9]. In almost all cases of UTI, whether outpatient or hospital patient, antimicrobial therapy is initiated empirically before the results of urine culture are available. Hence, there exists a great need for antimicrobial resistance surveillance at local, national and international levels. The aim of this study was to determine the distribution of uropathogens and their antimicrobial susceptibility patterns in a multi-site study in the UK. To reduce the bias due to different pathogen distribution and antimicrobial susceptibility profile associated with the type of UTI and in/outpatient status, patients were assigned to 1 of 4 categories according to age and clinical diagnosis. These categories were: (1) community-acquired UTI in those less than 65 years old, (2) hospital-acquired UTI other than those admitted with pyelonephritis, (3) pyelonephritis, and (4) community-acquired UTI in those greater than 65 years old.
# 2002 The British Infection Society. Published by Elsevier Science Ltd. All rights reserved.
UK Surveillance of UTI Pathogens
Materials and Methods
95
Results
Participating centres During 1999±2000, 8 centres geographically spread throughout the UK took part in the study: St. Mary's Hospital (London), Southmead Hospital (Bristol), Leicester Royal Infirmary (Leicester), Leeds General Infirmary (Leeds), the Royal Victoria Hospital (Newcastleupon-Tyne), the Southern General Hospital (Glasgow), Raigmore Hospital (Inverness) and the Royal Hospitals (Belfast). Patients and bacterial isolates Each centre was required to submit 200 bacterial uropathogens, cultured from the urine of patients categorized into one of four infection types (i.e. 50 isolates from each category), to the central testing laboratory (GR Micro Ltd, London). Repeat isolates from the same patient, cultured within 2 weeks of the first isolation, were excluded from the collection. Each isolate was designated as a `pathogen' at the discretion of the centre performing urinalysis. Isolates were stored at 70 C at the centre until a batch of sufficient size was ready for transportation to the Central Testing Laboratory. Identification and susceptibility testing Bacterial isolates were identified by biochemical profiling using API systems (bioMeÂrieux, Basingstoke, UK). MICs were determined using the National Committee for Clinical Laboratory Standards (NCCLS) agar dilution method [10]. Breakpoints used for interpretation of MIC results were those specified by the NCCLS [11].
The overall female to male ratio was 4.1:1, ranging from 2.6:1 in hospitalized patients to 6.8:1 in patients less than 65 years old with community-acquired UTI (Table I). The mean age was 47 years for patients with pyelonephritis compared to 27 years for patients less than 65 years old with community-acquired UTI (Table I). In total, 1291 bacterial isolates were examined in the study. E. coli was the most common urinary pathogen identified. The relative frequency of E. coli was lowest in non-pyelonephritis hospital patients (56.3%), presumably due to the increased frequency of a wide range of other nosocomial pathogens, and highest in patients less than 65 years old with community-acquired UTI (Table II). Enterococcus faecalis, Klebsiella pneumoniae, Proteus mirabilis and Pseudomonas aeruginosa were the next most common species isolated with the rank order varying between infection categories (Table II). Other isolates of importance included Staphylococcus saprophyticus (responsible for 2.0% of infections in patients less than 65 years old with community-acquired UTI) and Klebsiella oxytoca (2.8% of patients with pyelonephritis). The activity of amoxycillin against E. coli (51.3% susceptible) was greatly reduced as a result of blactamase production and only partially restored by the addition of clavulanic acid (78.8% susceptible) (Table III). As expected, K. pneumoniae isolates were almost fully resistant in vitro to amoxycillin. Cefuroxime was very active against the E. coli using parenteral form breakpoints (97.1% susceptible) (Table IV) but not using oral form breakpoints (68.6% susceptible). As expected, cefuroxime was inactive against Enterococcus faecalis or Pseudomonas aeruginosa. (Table III). The activities of cephradine, tetracycline, sulphamethoxazole and
Table I. Gender and age distribution by category. Infection category
Community-acquired UTI (patients less than 65 years old) UTI in hospital patients other than those admitted with pyelonephritis Pyelonephritis Community-acquired UTI (patients greater than 65 years old) All categories combined
Number of isolates
Female/male (ratio)
Age range (years)
Mean age (years)
397
345/51* (6.8:1)
<1±64
27
394
286/108 (2.6:1)
<1±99
33
108 392
87/21 (4.1:1) 321/71 (4.5:1)
<1±92 65±97
47 78
1291
1039/251 (4.1:1)
<1±99
57
*One patient of unknown gender is omitted from the analysis.
96
D. J. Farrell et al. Table II. Relative frequency (% of total) of isolation of species causing various types of UTI. Community-acquired UTI (patients less than 65 years) (n 397)
Hospital patient UTI (excluding pyelonephritis) (n 394)
Pyelonephritis (n 108)
Organism
%
Organism
%
Organism
%
Organism
%
E. coli P. mirabilis E. faecalis K. pneumoniae S. saprophyticus P. aeruginosa S. agalactiae Other species
77.3 4.3 3.8 3.5 2.0 1.8 1.5 5.8
E. coli E. faecalis K. pneumoniae P. mirabilis P. aeruginosa E. cloacae S. aureus E. faecium K. oxytoca Other species
56.3 8.4 6.9 6.3 3.8 2.5 2.5 2.0 2.0 9.1
E. coli K. pneumoniae E. faecalis P. mirabilis K. oxytoca Other species
68.5 6.5 5.6 2.8 2.8 13.9
E. coli E. faecalis K. pneumoniae P. mirabilis P. aeruginosa E. cloacae K. oxytoca Other species
66.6 5.9 5.9 4.3 3.1 2.8 1.8 9.7
trimethoprim were low against E. coli and these agents provided no, or limited, coverage against some of the other prevalent pathogens (Table III). Nitrofurantoin was very active against isolates E. coli (96.3% susceptible) and E. faecalis (100%) but not against K. pneumoniae, P. mirabilis or P. aeruginosa. Ciprofloxacin was highly active against the UTI pathogens examined in this study with susceptibilities of between 87.1% and 97.7% for the most prevalent pathogens (E. coli, n 864, 97.7% susceptible). For all pathogens, regardless of species, ciprofloxacin had the lowest MIC50 (0.015 mg/L) and MlC90 (1 mg/L) for all of the orally available antibiotics (Table III). Ciprofloxacin was the only oral agent tested with activity against P. aeruginosa (88.6% susceptible). All of the parenterally available antibiotics had excellent activity (87.3±100%) against the most prevalent Enterobacteriaceae (Table IV). Escherichia coli was the only species to have sufficient numbers to obtain valid percent susceptible data by infection category (Table V). Susceptibility patterns were similar between the categories, however susceptibilities for all oral agents were lower in isolates from hospital patients (category 2) than outpatients. Also, the susceptibility of E. coli to trimethoprim was considerably lower in patients with pyelonephritis compared to the other categories of infection (64.9% susceptible compared to 72.8±76.5% for the other 3 categories). Within the whole population, 19 multiresistant isolates were found: 12 Enterobacteriaceae extendedspectrum beta-lactamases (ESBLs), two vancomycinresistant enterococci (VRE) (both vanA phenotype) and five methicillin-resistant S. aureus (MRSA). All but one isolate were obtained from hospitalized patients, however one MRSA isolate was from an elderly patient in the community. The multidrug-resistant Enterobacteriaceae
Community-acquired UTI (patients greater than 65 years) (n 392)
(K. pneumoniae 6, E. coli 4, K. oxytoca 1, E. cloacae 1) were all extremely resistant to most oral antibiotics, except 4 isolates which were susceptible to ciprofloxacin and 2 isolates which were susceptible to amoxycillin/ clavulanate. All 12 were susceptible to imipenem and amikacin but not ceftazidime, gentamicin or cefuroxime when using cefuroxime sodium breakpoints. One VRE was susceptible to amoxycillin and imipenem, the other resistant to both agents. Both VRE were resistant to all other agents tested. All 4 MRSA were susceptible to gentamicin, amikacin, trimethoprim, tetracycline and sulphamethoxazole but resistant to all other agents. A total of 37/40 (92.5%) of S. saprophyticus isolates were found in women, with 31/37 (83.8%) in patients less than 65 years of age with uncomplicated UTI or pyelonephritis, and 24/37 (64.9%) of the patients aged between 16 and 35 years old.
Discussion This study shows the distribution of microbial species causing UTIs in the UK during 1999±2000 and their susceptibility patterns to the most commonly used oral and parenteral antimicrobial agents. To reduce bias in the data, UTIs were categorized into 4 categories of infection based on simple or complicated UTI, hospital status, patient age and a clinical diagnosis of pyelonephritis. To reduce technical variation, all isolates were re-identified and had susceptibility tests performed at a central reference laboratory using NCCLS methodology and interpretation [10,11]. Overall, 74 different species of bacteria were identified in the 1291 patients with UTI with between 29 and 49 species being found in each category. This variation in
2
1
NA 128 128 16±>64 NA 8 32 0.015±>64
NA 16 >64 0.015±>64
96.8 1 8 0.5±32
72.6 1 128 0.5±64
NA >64 >64 16±>64
84.5 2 16 1±32
100 0.25 0.5 0.25±1
100 0.25 0.5 0.25±1
1.4 >64 >64 1±128
78.8 8 16 1±64
51.3 8 >64 1±>64
NA 8 128 0.015±>128
NA >128 >128 All > 128
4.8 16 32 8±128
NA 4 32 0.015±>128
NA >128 >128 All > 128
95.2 2 4 0.5±>128
81.7 4 32 1±128
NA 16 >128 2±>128
NA2 16 128 16±128 71.8 8 32 4±>128
68.6 4 8 0.25±>128
Cefuroxime1
65.3 8 16 2±>128
Amoxycillin Amoxycillin/ Cephradine clavulanate
Using NCCLS breakpoints for the oral form of cefuroxime (cefuroxime axetil). NA not applicable. No NCCLS breakpoint available. n Number of isolates.
P. aeruginosa (35, 2.7%) % susceptible MIC50 MIC90 Range All isolates combined % susceptbile MIC50 MIC90 Range
K. pneumoniae (71, 5.5%) % susceptible MIC50 MIC90 Range P. mirabilis (62, 4.8%) % susceptible MIC50 MIC90 Range
E. coli (864, 66.9%) % susceptible MIC50 MIC90 Range E. faecalis (77, 6.0%) % susceptible MIC50 MIC90 Range
Organism (n, % of UTI)
NA 2 >64 0.12±>64
0 32 32 16±>64
0 64 >64 32±>64
81.7 2 >64 1±>64
29.9 64 >64 0.25±>64
74.4 2 >64 0.5±>64
NA 32 >512 0.5±>512
57.1 256 512 64±>512
87.1 16 >512 4±>512
84.5 32 >512 8±>512
NA >512 >512 All > 512
62.3 32 >512 1±>512
NA 0.5 >64 0.12±>64
NA >64 >64 32±>64
58.1 4 >64 1±>64
84.5 1 >64 0.25±>64
NA 0.5 >64 0.12±>64
73.3 0.5 >64 0.12±>64
NA 32 128 2±>512
NA >512 >512 All > 512
0 128 128 64±512
15.5 128 256 16±512
100 8 8 4±16
96.3 32 32 2±256
NA 4 >64 1±>64
NA >64 >64 All > 64
83.9 4 >64 2±>64
81.7 4 >64 2±>64
NA >64 >64 All > 64
94.7 2 4 1±>64
NA 0.015 1 0.002±>64
88.6 0.25 2 0.03±16
87.1 0.03 2 0.015±4
94.4 0.03 0.5 0.008±32
NA 1 64 0.25±64
97.7 0.015 0.015 0.002±64
Tetracycline Sulphamethoxazole Trimethoprim Nitrofurantoin Nalidixic acid Ciprofloxacin
Table III. Percent susceptible, MIC50, MIC90 and range (in mg/L) for orally administered antimicrobial agents against the most prevalent uropathogens.
UK Surveillance of UTI Pathogens 97
98
D. J. Farrell et al.
Table IV. Percent susceptible, MIC50, MIC90 and range (in mg/L) for parenterally-administered antimicrobial agents against the most prevalent uropathogens. Organism (n, % of UTI)
Ceftazidime
Imipenem
Gentamicin
Amikacin
Cefuroxime1
E. coli (864, 66.9%) % susceptible MIC50 MIC90 Range
99.3 0.12 0.25 0.06±>128
100 0.12 0.25 0.06±1
98.8 0.25 0.5 0.12±>64
100 1 2 0.25±16
97.1 4 8 0.25±>128
E. faecalis (77, 6.0%) % susceptible MIC50 MIC90 Range
NA2 64 >128 8±>128
NA 1 2 0.25±4
NA 8 >64 1±>64
NA >64 >64 8±>64
NA 16 >128 2±>128
K. pneumoniae (71, 5.5%) % susceptible MIC50 MIC90 Range
91.6 0.12 1 0.06±>128
100 0.12 0.25 0.06±0.5
91.6 0.25 0.5 0.12±64
100 1 1 0.5±16
87.3 4 32 1±128
P. mirabilis (62, 4.8%) % susceptible MIC50 MIC90 Range
100 0.06 0.12 0.06±0.25
100 0.5 2 0.06±4
100 0.5 1 0.12±1
100 2 4 0.5±4
98.4 2 4 0.5±>128
P. aeruginosa (35, 2.7%) % susceptible MIC50 MIC90 Range
60.0 8 32 1±>128
97.1 2 2 0.12±16
97.1 2 4 0.25±32
100 4 8 0.5±8
NA >128 >128 All > 128
All isolates combined % susceptible MIC50 MIC90 Range
NA 0.25 16 0.06±>128
NA 0.12 1 0.008±>64
NA 0.25 4 0.03±>64
NA 1 4 0.12±>64
NA 4 32 0.015±>128
1 2
Using NCCLS breakpoints for the parenteral form of cefuroxime (cefuroxime sodium). NA not applicable. No NCCLS breakpoint available.
the number of species involved in UTI is greater than a recent report of 32 species identified in a study of 1510 hospital patients (43 species were found in hospitalized patients in our study with a population size of 394 patients) [8]. E. coli was the most frequent uropathogen in all categories of infection. The frequency of E. coli isolated (66.9% in total, category range 56.3±77.3%) is lower than a recently reported frequency for the UK of 81.2% [12]. E. faecalis was the second most common isolate causing 6.0% of infections although the frequency of the most prevalent Enterobacteriaceae combined were greater. E. faecalis was more frequently isolated in hospitalized patients than non-hospitalized. Staphylococcus saprophyticus is a known opportunistic pathogen predominantly in young, sexually active females [13]. Recent reports have found this pathogen to be present in 9.7% (n 5739) and 4.0% (n 4342) of women with uncomplicated cystitis [9,13]. In our study, S. saprophyticus infection in women less than
65 years old with community-acquired UTI, therefore, is demographically similar but lower in frequency (2.0%) than previous reports [9,13]. A recent European study reported an ampicillin susceptibility of 58.9% for E. coli from 246 patients in the UK [12]. The rate of susceptibility to amoxycillin in our study (total 51.3%, category range 48.5±55.6%) is slightly lower, but similar to this percentage. The same study reported an amoxycillin/clavulanate susceptibility of 96.3% for UK patients [12]. This rate is vastly different from our rate (78.8%) but similar to recently reported data from Canada (95%, n 694) [14]. Possible reasons for this difference are geographical variation due to a higher prevalence of variants of b-lactamase resistant E. coli (such as hyper-b-lactamase producers with TEMmediated resistance) or methodological discrepancies. For example, the study reporting low levels of amoxycillin/clavulanate resistance in the UK did not perform MICs using NCCLS methodology, as in our study, but
307 222 74 261 864
1 2 3 4 All categories
Ceftazidime
100 97.3 100 100 99.3
100 100 100 100 100
Imipenem
Parenteral agents
Gentamicin 100 97.3 97.3 99.2 98.8
Amikacin 100 100 100 100 100
Cefuroxime1 97.1 93.7 98.6 98.1 96.6
Oral agents
48.5 48.6 55.4 55.6 51.3
Amoxycillin
1
n Number of isolates. Using NCCLS breakpoints for the parenteral form of cefuroxime (cefuroxime sodium). 2 Using NCCLS breakpoints for the oral form of cefuroxime (cefuroxime axetil). 3 NCCLS breakpoints for the parenteral and oral forms of ciprofloxacin are identical.
n
Category
78.8 74.3 86.5 80.5 78.8
Amoxycillin/clavulanate
Table V. Percent susceptibility of E. coli isolates from each category of infection.
Cephradine 68.4 58.1 78.4 64.0 65.3
Cefuroxime2 71.0 60.8 78.4 69.7 68.6
Tetracycline 74.9 73.0 75.7 75.1 74.5
Sulphamethoxazole 65.5 57.7 59.5 63.2 62.3
Trimethoprim 76.5 72.1 64.9 72.8 73.3
Nitrofurantoin 97.7 95.9 95.9 95.0 96.3
Nalidixic acid 96.7 91.9 90.5 95.8 94.7
Ciprofloxacin 3 99.3 91.9 93.2 98.5 95.8
UK Surveillance of UTI Pathogens 99
100
D. J. Farrell et al.
used disc susceptibility testing [12]. The decision to prescribe amoxycillin/clavulanate empirically, therefore, requires consideration of current local susceptibility data. Comparative studies between different amoxycillin/clavulanate susceptibility testing methodologies are needed to increase confidence in the reliability of these results. The two oral cephalosporins (cephradine and cefuroxime), trimethoprim, sulphamethoxazole, and tetracycline did not have good in vitro coverage for many of the uropathogens isolated in this study. Because of this, they are of little use as empirical therapeutic options. In general, the parenteral agents showed excellent in vitro coverage of the uropathogens isolated in this study. This information will be useful in choosing empiric therapy for seriously ill hospitalized patients with suspected urosepsis. Although nitrofurantoin demonstrated excellent in vitro activity against E. coli and E. faecalis (96.3% and 100% susceptible, respectively) it demonstrated little or no activity against other prevalent Enterobacteriaceae and P. aeruginosa. Also, although debateable, nitrofurantoin is not traditionally used in urinary tract infections with any significant tissue involvement (such as pyelonephritis). Nalidixic acid demonstrated excellent in vitro activity against E. coli (94.7% susceptible), good activity against other prevalent Enterobacteriaceae but no activity against Gram-positive organisms or P. aeruginosa. In addition, as with nitrofurantoin, nalidixic acid has poor tissue penetration. Overall, ciprofloxacin demonstrated the greatest in vitro activity against the widest range of uropathogens for all of the oral agents tested with susceptibilities of between 87.1% and 97.7% for the most prevalent pathogens (E. coli, n 864, 97.7% susceptible). This activity was consistent across each of the infection categories. Importantly, ciprofloxacin was the only oral agent tested with activity against Pseudomonas spp. Although no NCCLS breakpoints exist for ciprofloxacin against E. faecalis, ciprofloxacin demonstrated activity against this species (MIC50 1 mg/L). In conclusion, these data provide much needed information on the prevalence of antimicrobial resistance amongst pathogens currently causing UTI in the UK. This data was obtained using standardized (NCCLS) methodology and has been divided into infection type categories to remove bias. Overall, ciprofloxacin demonstrated the most complete in vitro coverage amongst the antimicrobials tested and hence emphasizing its great utility as a choice for empiric therapy of UTI in the UK. As with other agents, increased use results in concern of resistance development. Prudent prescribing, establishment and use of infection control guidelines,
local monitoring of resistance and longitudinal surveillance studies are necessary to minimized and monitor the development of resistance. We hope to repeat this study 2002/2003 to monitor the development of resistance in the UK. Acknowledgments We thank Bayer AG for sponsoring this study and, as always, we are grateful to the collecting centres for their participation in the study.
References 1 Stamm WE, Hooton TM. The management of urinary tract infection. N Engl J Med 1993; 329: 1328±1334. 2 Warren JW, Abrutyn E, Hebel JR, Johnson JR, Schaeffer AJ, Stamm WE. Guidelines for antimicrobial treatment of uncomplicated acute bacterial cystitis and acute pyelonephritis in women. Clin Infect Dis 1999; 29: 745±758. 3 Gastmeier P. Nosocomial urinary tract infections: many unresolved questions. Clin Micro Infect 2001; 7: 521±522. 4 Nicolle LE, Strausbaugh LJ, Garibaldi RA. Infections and antibiotic resistance in nursing homes. Clin Microbiol Rev 1996; 9: 1±17. 5 GruÈneberg RN. Changes in urinary pathogens and their antibiotic sensitivities. 1971±1992. J Antimicrob Chemother 1994; 33(Suppl. A): 1±8. 6 MacGowan AP, Brown NM, Holt HA, Lovering AM, McCulloch SY, Reeves DS. An eight-year survey of the antimicrobial susceptibility patterns of 85,971 bacteria isolated from patients in a district general hospital and the local community. J Antimicrob Chemother 1993; 31: 543±557. 7 Barrett SP, Savage MA, Rebec MP, Guyot A, Andrews N, Shrimpton SB. Antibiotic sensitivity of bacteria associated with community-acquired urinary tract infection in Britain. J Antimicrob Chemother 1999; 44: 359±365. 8 Mathai D, Jones RN, Pfaller MA. Epidemiology and frequency of resistance among pathogens causing urinary tract infection in 1,510 hospitalized patients: a report from the SENTRY Antimicrobial Surveillance Program (North America). Diag Microbiol Infect Dis 2001; 40: 129±136. 9 Karlowsky JA, Jones ME, Thornsberry C, Critchley I, Kelly L, Sahm D. Prevalence of antimicrobial resistance among urinary tract pathogens isolated from female outpatients across the US in 1999. Int J Antimicrob Agents 2001; 18: 121±127. 10 NCCLS (2000). Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically; Approved StandardÐ5th edn. NCCLS document M7-A5 (ISBN 1-56238-394-9). NCCLS, Wayne, Pennsylvania 19087-1898, USA. 11 NCCLS (2002). Performance standards for antimicrobial susceptibility testing; Twelfth informational supplement. NCCLS document M100-12 (ISBN 1-56238-454-6) NCCLS, Wayne, Pennsylvania 19087-1898, USA. 12 Kahlmeter G. The ECO-SENS project: a prospective, multinational, multicentre epidemiological survey of the prevalence and antimicrobial susceptibility of urinary tract pathogens. J Antimicrob Chemother 2000; 46(Suppl. S1): 15±22. 13 Gupta K, Scholes D, Stamm WE. Increasing prevalence of antimicrobial resistance among uropathogens causing acute uncomplicated cystitis in women. JAMA 1999; 281: 736±738. 14 Blondeau JM, Yaschuk Y, Suter M, Vaughan D, and the Canadian Antimicrobial Study Group. In-vitro susceptibility of 1982 respiratory pathogens and 1921 urinary tract pathogens against 19 antimicrobial agents: a Canadian multicentre study. J Antimicrobial Chemotherapy 1999; 43(Suppl. A): 3±23.