Abstract

Objectives: To study the clonality of gentamicin-resistant, extended-spectrum β-lactamase (ESBL)-negative and ESBL-producing Escherichia coli isolated from community-onset urinary tract infections (UTIs) in Cornwall.

Methods: Isolates were identified by API, susceptibilities were determined by local disc testing, and MICs were determined at the reference laboratory, both interpreted using BSAC guidelines. blaCTX-M genes were sought by PCR, and isolates were compared by PFGE.

Results: In the years 2004 and 2005, 69 E. coli were submitted by Truro (Cornwall) laboratory for reference laboratory testing: these included 14 gentamicin-resistant, ESBL-negative isolates; 45 with group 1 CTX-M enzymes; seven with group 9 CTX-M enzymes; and three with non-CTX-M ESBLs. By PFGE, nine gentamicin-resistant, ESBL-negative E. coli were distinct (<85% similarity) from all the ESBL producers, but three were related to producers of group 1 CTX-M enzymes, and two isolates were related to a non-CTX-M ESBL producer. An outbreak strain was identified, represented by 11 gentamicin-resistant and one gentamicin-susceptible isolates, all with group 1 CTX-M enzymes, and two gentamicin-resistant, ESBL-negative isolates. This was distinct by PFGE from nationally distributed CTX-M-producing strains. Five of nine patients infected with this strain had been on the same ward in a local hospital; four presented with community-onset UTIs; one inpatient developed a hospital-acquired bacteraemia. Of the other four patients presenting with community-onset UTIs, three were admitted to different hospitals and the fourth had only attended an outpatient clinic.

Conclusions: Community-onset, ESBL-producing and non-producing E. coli were diverse. Two ESBL-negative isolates were closely related to a local CTX-M-producing outbreak strain, suggesting gain or loss of a blaCTX-M-carrying plasmid. An outbreak strain was linked with prior hospital admission and appeared not to represent genuine community acquisition.

Introduction

Multiresistant Escherichia coli strains from community-onset infections are an increasing public health concern, especially those with CTX-M extended-spectrum β-lactamases (ESBLs).1 In the UK, ∼90% of CTX-M-producing E. coli produce group 1 enzymes (mainly CTX-M-15) and PFGE analysis of a national collection has defined five major strains.1 The local epidemiology of E. coli producing CTX-M enzymes may, however, differ from this national picture with, for example, dominance of different strains, or higher prevalence of those with group 9 CTX-M enzymes.

In Cornwall (the south-west peninsula of England), concern was raised over the emergence of gentamicin-resistant E. coli isolates recovered from urines of patients attending general practice. Gentamicin is rarely prescribed in such settings. Several, but not all of these resistant isolates were ESBL producers. We therefore studied the clonality of gentamicin-resistant, ESBL-negative isolates from urines of patients attending general practitioners (GPs) in Cornwall and compared them with ESBL-producing E. coli, many also gentamicin-resistant and also isolated from GP urines.

Materials and methods

Bacterial isolates and initial susceptibility testing

Most E. coli isolates included in this study were associated with community-onset urinary tract infections (UTIs) in patients attending GPs' surgeries in Cornwall during 2004 and 2005. One isolate was from a case of hospital-acquired bacteraemia (onset >48 h after admission).

The isolates were identified as E. coli (API system; bioMerieux, Basingstoke, UK), and their antibiotic susceptibilities were determined by disc testing using BSAC criteria2 in the Department of Clinical Microbiology, Royal Cornwall Hospital. Fourteen gentamicin-resistant, ESBL-negative isolates and 55 ESBL-producing isolates, variably gentamicin-resistant, were referred to the Health Protection Agency's Centre for Infections for further investigation. Multiple isolates were referred from some patients.

Investigation in the reference laboratories

MICs were determined and interpreted using BSAC methodology,2 and blaCTX-M genes were sought by multiplex PCR.3 Selected isolates were compared by PFGE of XbaI-digested genomic DNA, and banding patterns were analysed using BioNumerics software (Applied Maths, Sint-Martens-Latem, Belgium).

Investigation of an outbreak strain

The clinical notes for selected patients were analysed to determine any obvious common factors associated with acquisition of a PFGE-defined outbreak strain.

Results and discussion

Reference investigations and recognition of a ‘Cornish outbreak’ strain

During 2004–2005, the reference laboratories received 69 E. coli isolates from Truro (Cornwall) laboratory for inclusion in this study. These included: 14 gentamicin-resistant, ESBL-negative isolates; 45 with group 1 CTX-M enzymes; seven with group 9 CTX-M enzymes and three with non-CTX-M ESBLs.

All 14 ESBL-negative isolates, and 16 ESBL-producing isolates dating from 2004, were compared by PFGE (Figure 1). Nine gentamicin-resistant, ESBL-negative isolates were distinct (<85% similarity) from the ESBL producers, but three were related to isolates with CTX-M group 1 enzymes, and two isolates from one patient were related to an isolate with a non-CTX-M ESBL.

Figure 1

Dendrogram (UPGMA, Dice) showing the relatedness of PFGE banding patterns for 30 E. coli isolates from Cornwall. Bold lines indicate clusters with >85% similarity. Symbols: +, CTX-M group 1 producer; +a, CTX-M group 9 producer; −b, ESBL (non-CTX-M) producer; −, non-ESBL producer. Isolates of the local outbreak strain and UK epidemic strain A are shaded light grey and dark grey, respectively.

Figure 1

Dendrogram (UPGMA, Dice) showing the relatedness of PFGE banding patterns for 30 E. coli isolates from Cornwall. Bold lines indicate clusters with >85% similarity. Symbols: +, CTX-M group 1 producer; +a, CTX-M group 9 producer; −b, ESBL (non-CTX-M) producer; −, non-ESBL producer. Isolates of the local outbreak strain and UK epidemic strain A are shaded light grey and dark grey, respectively.

One major outbreak strain was identified, and comprised 11 gentamicin-resistant and one gentamicin-susceptible CTX-M group 1 producers and two gentamicin-resistant, ESBL-negative isolates (12 of these are shown in Figure 1); the 14 isolates were from nine patients. This ‘Cornish outbreak’ strain was distinct by PFGE from five nationally distributed CTX-M-producing strains, including the UK epidemic strain A (Figure 1),1 which accounted for only three of the 69 E. coli isolates referred from Truro laboratory.

Geometric mean MICs for the ‘Cornish outbreak’ strain are shown in Table 1. CTX-M ESBL-producing isolates of this strain were resistant to penicillins, all cephalosporins tested (but with potentiation by clavulanate), and also to ciprofloxacin, gentamicin and tobramycin. Isolates of the outbreak strain were susceptible to piperacillin/tazobactam, amikacin and carbapenems.

Table 1

Susceptibilities of the ‘Cornish outbreak’ CTX-M-producing E. coli strain compared with UK epidemic strain A1

 Geometric mean MIC (mg/L) 
 
 
Antibiotic ‘Cornish outbreak’ strain (n = 12)a strain A1 
AMP 64 64 
AMC 17 16 
CTX 71.8 64 
CTX + CLA 0.1 ≤0.06 
CAZ 26.9 
CAZ + CLA 0.4 0.25 
FEP 56.4 
FEP + CLA 0.1 ≤0.06 
FOX 10.7 16 
PIP 64 64 
TZP 4.2 16 
IPM 0.2 0.125 
MEM 0.1 ≤0.06 
ETP 0.1 ≤0.125 
CIP 17.5 
GEN 24 1.1 
AMK 2.6 
TOB 10.7 16 
MIN 3.5 8.2 
 Geometric mean MIC (mg/L) 
 
 
Antibiotic ‘Cornish outbreak’ strain (n = 12)a strain A1 
AMP 64 64 
AMC 17 16 
CTX 71.8 64 
CTX + CLA 0.1 ≤0.06 
CAZ 26.9 
CAZ + CLA 0.4 0.25 
FEP 56.4 
FEP + CLA 0.1 ≤0.06 
FOX 10.7 16 
PIP 64 64 
TZP 4.2 16 
IPM 0.2 0.125 
MEM 0.1 ≤0.06 
ETP 0.1 ≤0.125 
CIP 17.5 
GEN 24 1.1 
AMK 2.6 
TOB 10.7 16 
MIN 3.5 8.2 

AMP, ampicillin; AMC, co-amoxiclav; CTX, cefotaxime; CLA, clavulanate; CAZ, ceftazidime; FEP, cefepime; FOX, cefoxitin; PIP, piperacillin; TZP, piperacillin/tazobactam; IPM, imipenem; MEM, meropenem; ETP, ertapenem; CIP, ciprofloxacin; GEN, gentamicin; AMK, amikacin; TOB, tobramycin; MIN, minocycline.

aTwo non-CTX-M-producing isolates of the outbreak strain were excluded from this analysis.

Two gentamicin-resistant, ESBL-negative isolates (unrelated by PFGE and distinct from the ‘Cornish outbreak’ strain) were highly resistant to the veterinary aminoglycoside apramycin (MICs > 512 mg/L), which is consistent with AAC(3)-IV activity;4 the range of apramycin MICs for all other isolates was 2–32 mg/L.

Analysis of case notes

Recognition of the ‘Cornish outbreak’ strain was unexpected and prompted analysis of clinical notes for the nine patients (5 male and 4 female) affected to determine any obvious common factors that might be associated with its acquisition. The average age of these patients was 80 years (range, 57–93 years).

One patient developed a hospital-acquired bacteraemia (defined as onset >48 h after admission) with the ‘Cornish outbreak’ strain while an inpatient on ward G of a local hospital (RCH) in March to May 2004 during a cluster of infections on this ward caused by ESBL-producing E. coli. This blood culture isolate was the only saved representative from the ward cluster. The other eight patients infected by the ‘Cornish outbreak’ strain had presented at various GP surgeries, separated by up to 30 miles. The strain was isolated from urine in eight patients, from sputum in three, from a peg site in one, and two patients were widely colonized. Despite the community-onset of their UTIs, seven of these eight patients had prior, and in some instances multiple, hospital inpatient admissions, and multiple transfers between wards. The eighth only attended an outpatient clinic. The average time after hospital discharge before presentation with a UTI at their GP’s surgery was 39 days (range 6–89 days). Four of the eight GP patients had been inpatients on ward G of RCH during March–May 2004. Another had been in other wards at RCH during this period, though not on ward G. Two other patients had been on ward L of a second hospital (CRH), but there was no temporal overlap; one had been an inpatient from December 2003 to March 2004; the other was transferred from ward G of RCH in April 2004 and was discharged in May 2004. Six of the nine affected patients died. Mortality was not directly attributable to the ESBL-producing strain in five cases, but the sixth died with ESBL-producing E. coli septicaemia, implying causality.

These data suggest that the ‘Cornish outbreak’ strain may have caused the cluster of infections on ward G at RCH during March–May 2004, although only one isolate was available for testing. The alarming finding of a CTX-M-producing E. coli strain presenting in the community, but that could be traced to a hospital source led to a rapid review of susceptibility testing of all ‘coliforms’ in the clinical laboratory, as advocated previously,1,5 and methodology for detecting ESBLs was in place within two months of these findings.

In summary, most gentamicin-resistant, ESBL-negative isolates from community-onset UTIs in Cornwall were diverse and were distinct from ESBL-producing E. coli from community-onset UTIs. However, two ESBL-negative, gentamicin-resistant isolates were closely related to a local CTX-M-producing and gentamicin-resistant outbreak strain, suggesting gain or loss of the blaCTX-M gene or a plasmid carrying it. Most of the patients infected by this outbreak strain had hospital admissions in the three months prior to onset of their UTIs. This has been identified previously as a risk factor for infections caused by ESBL producers in non-hospitalized patients.6,7 Old age (>60 years), as seen in patients described here, is also a known risk factor.6

Although manifesting in the community, many of the UTIs of these Cornish patients appeared to have their origins in local hospitals and did not represent genuine community acquisition. This study highlights weaknesses in the concept of a clear distinction between hospital- and community-acquired infections caused by CTX-M-producing and other multiresistant E. coli. Clearly, acquisition in hospital may be followed by a period of colonization, which may be prolonged in some instances, prior to infection. Screening for faecal colonization at discharge, especially during on-going hospital clusters with ESBL-producing strains, would identify those patients most at risk of developing subsequent community-onset infections with multiresistant strains.

Transparency declarations

N. W. has received research grants from Merck and AstraZeneca; N. W. and J. W. H. have received conference fees from Merck.

Part of this work was presented at the Forty-Sixth Interscience Conference on Antimicrobial Agents and Chemotherapy (San Francisco, USA, 2006; Poster C2-0078). No external financial support was received for this study.

References

1
Woodford
N
Ward
ME
Kaufmann
ME
, et al.  . 
Community and hospital spread of Escherichia coli producing CTX-M extended-spectrum β-lactamases in the UK
J Antimicrob Chemother
 , 
2004
, vol. 
54
 (pg. 
735
-
43
)
2
Andrews
J
for the BSAC Working Party on Susceptibility Testing. BSAC standardized disc susceptibility testing method (version 5)
J Antimicrob Chemother
 , 
2006
, vol. 
58
 (pg. 
511
-
29
)
3
Woodford
N
Fagan
EJ
Ellington
MJ
Multiplex PCR for rapid detection of genes encoding CTX-M extended-spectrum β-lactamases
J Antimicrob Chemother
 , 
2005
, vol. 
57
 (pg. 
154
-
5
)
4
Johnson
AP
Burns
L
Woodford
N
, et al.  . 
Gentamicin resistance in clinical isolates of Escherichia coli encoded by genes of veterinary origin
J Med Microbiol
 , 
1994
, vol. 
40
 (pg. 
221
-
6
)
5
Munday
CJ
Whitehead
GM
Todd
NJ
, et al.  . 
Predominance and genetic diversity of community- and hospital-acquired CTX-M extended-spectrum β-lactamases in York, UK
J Antimicrob Chemother
 , 
2004
, vol. 
54
 (pg. 
628
-
33
)
6
Colodner
R
Rock
W
Chazan
B
, et al.  . 
Risk factors for the development of extended-spectrum β-lactamase-producing bacteria in nonhospitalized patients
Eur J Clin Microbiol Infect Dis
 , 
2004
, vol. 
23
 (pg. 
163
-
7
)
7
Calbo
E
Romani
V
Xercavins
M
, et al.  . 
Risk factors for community-onset urinary tract infections due to Escherichia coli harbouring extended-spectrum β-lactamases
J Antimicrob Chemother
 , 
2006
, vol. 
57
 (pg. 
780
-
3
)