High prevalence of ESBL-Producing E. coli in private and shared latrines in an informal urban settlement in Dar es Salaam, Tanzania

Background Data about the burden of extended-spectrum beta-lactamase (ESBL)-producing microorganisms in Africa are limited. Our study aimed to estimate the prevalence of human faecal ESBL carriage in the community of an informal urban settlement in Dar es Salaam (Tanzania, East Africa) by using environmental contamination of household latrines with ESBL as a surrogate marker. Methods Within the context of a large survey in February 2014 assessing 636 randomly selected household latrines for faecal contamination by the detection of growth of E. coli and total faecal coliform bacteria, a randomly selected subset of the samples were screened for ESBL. Results Seventy latrines were screened for ESBL. An average of 11.4 persons (SD ±6.5) were sharing one latrine. Only three (4.3%) latrines had hand-washing facilities and 50 showed faeces on the floor. ESBL-producing Enterobacteriaceae were confirmed in 17 (24.3%) of the 70 latrine samples: 16 E. coli and 1 Klebsiella pneumoniae. Five ESBL E. coli strains were detected on door handles. The most prevalent ESBL type was CTX-M-1 group (76.5%). Pulsed-field gel electrophoresis typing of a subset of ESBL-producing E. coli isolates revealed both diverse singular types and a cluster of 3 identical isolates. There was no significant difference of the latrine and household characteristics between the group with ESBL (n = 17) and the group with non-ESBL E. coli (n = 53) (p > 0.05). Conclusions Almost a quarter of private and shared latrines in an informal urban settlement in Tanzania are contaminated with ESBL-producing microorganisms, suggesting a high prevalence of human ESBL faecal carriage in the community. Shared latrines may serve as a reservoir for transmission in urban community settings in Tanzania.


Background
Extended-spectrum beta-lactamase (ESBL)-producing gramnegative bacteria have become an emerging global health threat and have been associated with high mortality [1,2]. Whereas ESBL infections were initially associated with nosocomial outbreaks, there is now increasing recognition of high rates of faecal carriage and the importance of community-acquired infections due to ESBL-producing Escherichia coli in industrialized countries [3,4].
Data about the prevalence of ESBL-producing microorganisms in Africa are limited. A wide variation from 0.6% up to 77.8% has been reported in hospital-based surveys of clinical isolates [5][6][7][8][9][10][11]. There is little published data on the magnitude of the community carriage of ESBL in African countries like Tanzania.
Our study aimed to estimate the prevalence of human faecal ESBL carriage in the community in an urban setting in Tanzania, East Africa, by using environmental contamination of household latrines with ESBL as a surrogate marker.

Methods
The study was performed in February 2014 in an urban study site in Keko Machungwa, part of the largest unplanned and under-serviced settlement in Temeke district, Dar es Salaam, Tanzania.
Within the context of a large survey assessing the improvement of sanitation facilities [12], 636 randomly selected household latrines were screened for the presence and concentration of faecal contamination (E. coli and total faecal coliform bacteria) [13]. Surface swipe swabs from high frequency contact points (10 cm 2 ) like door handle and footrest were taken and analysed using direct membrane filtration technique (Merck Millipore, Billerica, MA, USA) with commercial medium m-ColiBlue® (HACH, Loveland, CO, USA) [14]. In this survey, 492 latrines showed growth of E. coli or coliform bacteria either at the footrest, the door handle or both (unpublished data, Fig. 1).
Univariable analysis was performed by the chi-square test or Fisher's exact test, where appropriate, for categorical variables and two-tailed Student's t test for continuous variables. Two-tailed p-values of <0.05 were considered statistically significant.
The study was approved by the national ethics committee of the National Institute for Medical Research Tanzania (NIMR/HQ/R.8a/Vol.IX/1632).

Results
From the 492 latrines with growth of E. coli, 70 latrines were randomly selected for ESBL screening. All 70 latrines had either a cement, tile or brick floor and met the WHO/UNICEF Joint Monitoring Program for Water Supply and Sanitation definitions for improved latrines [12]. Only 3 (4.3%) latrines had hand-washing facilities and 50 showed faeces on the floor. An average of 11.4 persons (SD ±6.5) were sharing one latrine.
Twenty-four (34.3%) latrines were private (i.e. each of these latrines were used by only one household), the remaining latrines were shared by ≥2 households. The latrines were not accessible for the public but restricted to individuals living in the respective household (Table 1).
Forty samples showed ESBL suspected colonies on chromID® ESBL culture medium. ESBL-producing Enterobacteriaceae were finally confirmed in 17 (24.3%) of the 70 latrine samples: 16 E. coli and 1 Klebsiella pneu-    Table 3. PFGE typing of a subset of ESBL-producing E. coli isolates revealed both diverse singular types and a cluster of 3 identical isolates (Fig. 2). Ten isolates were nontypable by PFGE as observed in other studies [17]. There was no significant difference of the latrine and household characteristics between the group with ESBL (n = 17) and the group with non-ESBL E. coli (n = 53) (p > 0.05).

Discussion
In this study, we found that a quarter of the E. coli from contaminated latrines in an urban informal settlement in Dar es Salaam express ESBL, suggesting a high prevalence of human ESBL faecal carriage in the community, and that shared latrines may serve as a reservoir for transmission in urban areas of Tanzania.
In a recent meta-analysis, the median proportion of ESBL-producing Enterobacteriaceae in patients of Tanzanian healthcare facilities was 39% (range 14.2-75.9%) [11]. Only little is known about the ESBL burden in a community setting in Tanzania. Tellevik et al. reported an ESBL faecal carriage rate in healthy community children in Dar es Salaam of 11.6% [18]. An elderly survey from 2004 found ESBL in 16% of Enterobacteriaceae causing community- acquired urinary tract infections [19]. Studies from other African countries like Senegal, Niger and Madagascar have reported ESBL carriage rates in the community ranging as high as 10% to 31% [3]. In cases where they have been identified, CTX-M enzymes were predominantly of CTX-M-1 group as in our isolates. Community-onset infections with ESBL-producing pathogens are now increasingly reported. Empirical and targeted antibiotic treatment of such infections is challenging particularly in resource-limited countries. ESBL producing gram-negative bacteria are by definition resistant to extended-spectrum cephalosporins and frequently carry additional resistance genes conferring reduced susceptibility to many other antibiotics like e.g. fluoroquinolones. In many cases the carbapenems remain the only choice for treatment of infections caused by these resistant bacteria. However, the access to antimicrobial agents active against ESBL is limited in many regions of East Africa, thus common infections with multidrug-resistant pathogens cannot be treated adequately.
Intestinal colonization with ESBL-producing microorganisms may last several weeks, months or even years and is a potential source for human-to-human transmission. Precarious hygienic conditions of latrines, which are commonly shared amongst different households and household members, and the presence of ESBL-producing microorganisms on door handles may enable further spread. The cluster of 3 identical E. coli strains from 3 different latrines detected by PFGE typing in our study points to the potential of ESBL transmission in the community.
Our study has limitations: 1) Due to ethical and logistical reasons rectal swabs were not feasible. However, environmental samples of the latrines may give a valuable estimate of the frequency and distribution of ESBL-producing microorganisms. In addition, there is a lower risk of selection bias in the community and the results provide information about potential transmission pathways of multidrug-resistant pathogens in the community, similar to a recent study from airport toilet door handles [20]. 2) Our study was restricted to one informal urban settlement in Dar es Salaam and might differ from other urban and rural settings in Tanzania and East Africa. 3) Due to financial constraints only 70 latrines could be screened for ESBL.

Conclusions
Our study provides evidence of a high prevalence of human ESBL faecal carriage in the community of a resource-limited country such as Tanzania. Further larger surveillance studies are needed in Africa to better describe the epidemiology of ESBL, to raise awareness of the need of strategies to prevent further dissemination, and to improve the access and responsible use of appropriate antimicrobial agents in the empirical treatment of infections caused by otherwise deadly multidrugresistant microorganisms.