1. Overseas travel and antibiotic use are more important than community demographics or measures of social and material deprivation in predicting ESBL colonization at the time of hospital admission
Jonathan A. Otter, Alessandra Natale, Rahul Batra, Olga Tosas Auguet, Eleonora Dyakova, Simon D. Goldenberg, Jonathan D. Edgeworth
@jonotter

2. Introduction
Carriage rate of ESBL in the community
ESBL-producing Enterobacteriaceae (ESBL-E) are clinically problematic and drive the use of carbapenems, which can select for carbapenemase-producing Enterobacteriaceae.
A range of enzymes lead to an ‘ESBL’ phenotype, most commonly CTX-M-type ESBLs.
Risk factors associated with travel, especially to the Asian sub-continent, are typically associated with ESBL carriage.1,2
Various indices of social and material deprivaiton are risk factors for other MDROs (e.g. MRSA).3
Woerther et al. Clin Microbiol Rev 2013.
Woodford et al. FEMS Microbiol Rev 2011.
3. Tosas Auguet et al. Plos Med 2016.

3. Methods
Patients
Risk factor questionnaire
Hospital database
Community risk factors
All admissions to a London hospital group were screened for ESBL-E carriage from rectal swabs for four months in 2015.
ESBL-E were cultured on chromogenic media, and ESBLs detected on a microarray (CheckPoint Check-MDR).
Patients completed a risk factor questionnaire, and those with a residential postcode in the catchment area were linked to a database containing community-based risk factor data.
Risk factors for ESBL-E carriage were determined by binary logistic regression.

4. Risk factors for phenotypic ESBL-E carriage for all patients included in the study%+ p OR
Upper 95% CI
Lower 95% CI
Asian or Asian British 98 29
22.8%
.000
3.2
1.9
5.3
Black or Black British
469 59
11.2%
.045
1.4
1.0
2.1
Travel to Asia in the past 12 months
189 41
17.8%
.003
1.3
3.4
Travel to Africa in the past 12 months
128 25
16.3%
.026
1.1
3.5
Travel to Europe in the past 12 months
790 55
6.5%
.020
0.6
0.4
0.9
Overseas overnight hospital stay in the past 12 month 31 10
24.4%
.005
1.6
16.7
2 or more courses of antibiotics in the past 6 months
875
136
13.5%
2.9
Multivariable analysis. 360 (9%) of 4006 patients were ESBL+

5. Risk factors for phenotypic ESBL-E carriage in patients with residential postcodes in the catchment area of the study
Factor n- n+ %+ p OR
Upper 95% CI
Lower 95% CI
Travel to Asia in the past 12 months 56 21
27.3%
.000
5.0
2.5
10.0
Travel to Africa in the past 12 months 44 11
20.0%
.016
2.9
1.2
7.0
Two or more courses of antibiotics in the past 6 months
341 54
13.7%
2.2
1.5
3.4
Multivariable analysis. 157 (10%) of 1633 patients with residential postcodes within the catchment were ESBL+.
Factors not associated with ESBL-E included: the Index of Multiple Deprivation (a composite deprivation score including income, employment, health and disability, education skills and training, barriers to housing and services, crime, and living environment), household by deprivation dimensions (a composite deprivation score including employment, education, health and disability, and housing status), population density, health, household overcrowing, proportion of residents in communal establishments, proportion of residents in various ethnic groups, household space by dwelling type, and length of residence in the UK.
Residence in an area with a high proportion of Arab residents and residence in an area with a low proportion of houses with two or more bedrooms were associated with ESBL-E carriage in univariable but not multivariable analysis.

6. Risk factors for phenotypic E
Risk factors for phenotypic E. coli ESBL-E carriage compared with other species
Factor
Other species
E. coli
%E. coli p OR
Upper 95% CI
Lower 95% CI
Female 45
121
72.9%
.027
0.5
0.3
0.9
Overnight hospital stay in the past 12 months 56
124
68.9%
.001
0.4
0.2
0.8
1 course of antibiotics in the past 6 months 25 67
72.8%
.026
2 or more courses of antibiotics in the past 6 months 40 93
69.9%
.002
0.1
0.6
Multivariable analysis. 274 (76%) of ESBL-E were E. coli.
The following variables were associated with ESBL-E E. coli in univariable but not multivariable analysis:
Overseas travel in the past 12 months.
Travel to Asia in the past 12 months.
Mean number of days spent overseas in the past 12 months.

7. Summary of species / ESBL gene groups

8. Discussion
ESBL-E carriage is associated with travel-related risk factors, certain ethnic groups, and multiple antibiotic courses.1,2
Travel to Europe was protective, perhaps related to sociodemographics.
A very limited evidence base for community-based risk factors for Gram-negative bacteria – some more data for MRSA.3-5
It was surprising that no community-based factors related to social or material deprivation were associated with ESBL-E carriage.
Woerther et al. J Travel Med 2017.
Kantele et al. Clin Infect Dis 2015.
Sun et al. PloS ONE 2014.
Bonelli et al. Drug Res Updates 2014.
Tosas Auguet et al. Plos Med 2016.

9. Which anatomical site to screen?
Paired rectal and perineal swabs from the same individuals yielded ESBL-E in 7.8% of rectal swabs vs. 3.8% of perineal swabs, p<0.001
Dyakova et al. Clin Microbiol Infect 2017.

10. Can I swab your rectum please?
Factors associated with patients declining to provide a rectal swab were:
younger age (odds ratio (OR) 0.99, 95% confidence interval (CI) ) female gender (OR 1.26, CI ),
transfers from other hospitals (OR 1.77, CI ) or an unknown admission route (OR 1.61, CI ),
admission before the change in study description (OR 0.39, CI ).
the staff member who consented the patient (p<0.001);
ethnicity was not a significant factor.
Dyakova et al. Clin Microbiol Infect 2017.

11. Improving screening compliance
Dyakova et al. Clin Microbiol Infect 2017.

12. Conclusion
We linked individual risk factor information with community-based risk factor information, concluding that individual risk factors (including antibiotic use and overseas travel) were more important than community-based risk factors for predicting colonisation with ESBL-E at the time of hospital admission.
This information is useful when identifying risk groups for targeted screening.

13. Overseas travel and antibiotic use are more important than community demographics or measures of social and material deprivation in predicting ESBL colonization at the time of hospital admission
Jonathan A. Otter, Alessandra Natale, Rahul Batra, Olga Tosas Auguet, Eleonora Dyakova, Simon D. Goldenberg, Jonathan D. Edgeworth
@jonotter