Detection of CTX-M-15 beta-lactamases in Enterobacteriaceae causing hospital- and community-acquired urinary tract infections as early as 2004, in Dar es Salaam, Tanzania

Background The spread of Extended Spectrum β-lactamases (ESBLs) among Enterobacteriaceae and other Gram-Negative pathogens in the community and hospitals represents a major challenge to combat infections. We conducted a study to assess the prevalence and genetic makeup of ESBL-type resistance in bacterial isolates causing community- and hospital-acquired urinary tract infections. Methods A total of 172 isolates of Enterobacteriaceae were collected in Dar es Salaam, Tanzania, from patients who met criteria of community and hospital-acquired urinary tract infections. We used E-test ESBL strips to test for ESBL-phenotype and PCR and sequencing for detection of ESBL genes. Results Overall 23.8% (41/172) of all isolates were ESBL-producers. ESBL-producers were more frequently isolated from hospital-acquired infections (32%, 27/84 than from community-acquired infections (16%, 14/88, p < 0.05). ESBL-producers showed high rate of resistance to ciprofloxacin (85.5%), doxycycline (90.2%), gentamicin (80.5%), nalidixic acid (84.5%), and trimethoprim-sulfamethoxazole (85.4%). Furthermore, 95% of ESBL-producers were multi-drug resistant compared to 69% of non-ESBL-producers (p < 0.05). The distribution of ESBL genes were as follows: 29/32 (90.6%) bla CTX-M-15, two bla SHV-12, and one had both bla CTX-M-15 and bla SHV-12. Of 29 isolates carrying bla CTX-M-15, 69% (20/29) and 31% (9/29) were hospital and community, respectively. Bla SHV-12 genotypes were only detected in hospital-acquired infections. Conclusion bla CTX-M-15 is a predominant gene conferring ESBL-production in Enterobacteriaceae causing both hospital- and community-acquired infections in Tanzania. Electronic supplementary material The online version of this article (doi:10.1186/s12879-017-2395-8) contains supplementary material, which is available to authorized users.


Background
Extended Spectrum β-lactamases (ESBLs) have been observed in virtually all species of the family Enterobacteriaceae. Spread of ESBL-producing strains from general wards to intensive care units (ICU) and into the community can contribute to the further propagation of these resistant strains [1].
ESBLs are responsible for resistance against betalactam antibiotics such as penicillins, cephalosporins, monobactams and sometimes also carbapenems [2]. Organisms carrying ESBL enzymes often display coresistance to other antibiotics including aminoglycosides, quinolones, trimethoprim-sulfamethoxazole and tetracycline [3,4]. Spread of ESBL-producing bacterial isolates in the community has made empirical treatment of infections more difficult, and narrows the treatment options to expensive antibiotics like colistin and carbapenems.
Recent studies in Africa and Europe have found substantial increase in ESBL-producing Gram-negative bacteria causing community urinary tract infections, particularly harboring the bla CTX-M-15 allele [5][6][7]. Previously, studies in Tanzania have detected a substantial amount of ESBL-producing bacteria among the inpatients in intensive care and pediatric units in Tanzania [8,9]. However, little is known regarding the frequency of ESBL-producers and ESBL genes in communityacquired (CA) and hospital-acquired (HA) urinary tract infections in Tanzania. A more recent study found a predominance of the bla CTX-M-15 genotype from human feces in a community setting [10].
Previous studies performed at Muhimbili National Hospital documented the presence of ESBL-producers, predominantly of the bla CTX-M-15 genotype, in hospitalacquired infections [8,9]. Therefore, we decided to examine bacterial isolates collected prospectively for a period of six months in 2004 at Muhimbili National Hospital in Dar es Salaam, aiming at investigating whether ESBL-producers were present in the community setting, when hospital-acquired bla CTX-M-15 -producers were first reported in Tanzania. We also aimed to compare the ESBL genotypes circulating in the community to those found in the hospital setting.

Study setting and patient population
The study was conducted at Muhimbili National Hospital (MNH, Dar es Salaam, Tanzania. MNH is a tertiary health care facility that serves a population of about 4 million residents. The Department of Microbiology, MNH, receives samples from inpatients in the hospital wards, from the outpatient clinics and private health facilities in the city. Included in the study were samples from inpatients and outpatients with urinary tract infections seen at MNH between June 2004 and January 2005. Hospital-acquired infections were defined as those occurring in inpatients admitted at MNH for at least 72 h. Community-acquired infections were defined as those occurring in patients attending outpatient clinics at MNH. A UTI was defined as a positive urine culture of ≥10 5 CFU/ml of pure bacterial growth.

Clinical isolates
A total of 172 isolates of Enterobacteriaceae were collected, 84 and 88 isolates from hospital and community patients, respectively. All isolates were identified to the species level using established conventional procedures, the API 20E system (bioMérieux SA, Marcy l‵Etoile, France) or the Vitek 2 system (BioMérieux, Inc., Durham, N.C).

ESBL detection
All isolates with reduced susceptibilities to ceftazidime (zone of inhibition < 22mm), ceftriaxone (zone inhibition <25 mm) and cefotaxime (zone inhibition <27 mm) disks according to CLSI guidelines [11], were tested for an ESBL using E-test ESBL strips as previously described [8] and PCR. Isolates with reduced susceptibilities to cephalosporin were confirmed ESBL by either using Etests ESBL strips or PCR.

Detection and identification of ESBL genotypes
All strain with reduced susceptibility to cephalosporin were examined for the presence of the bla TEM , bla SHV and bla CTX-M genes by PCR, using genomic DNA isolated by boiling. For bla TEM amplification the primers described by Dubois et al. were used [13]. The cycling conditions were 95°C for 15 min, followed by 30 cycles of denaturation at 95°C for 1 min, annealing at 55°C for 1 min, elongation at 72°C for 1 min, followed by a final extension of 72°C for 10 min.
For bla SHV amplification the primers SHV-1F (5′ -CGG CCT TCA CTC AAG GAT G -3′) and SHV-1R (5′ -CGG STT AGC GTT GCC AGT -3′) were used. The cycling conditions were the same as for those of blaTEM, but with an annealing temperature of 60°C. In PCR amplification targeting the blaCTX-M gene the primer pairs described by Pagani et al. [14] were used. The cycling conditions for these two primer pairs were initial activation at 95°C for 15 min, followed by 30 cycles of denaturation at 95°C for 30 s, annealing at 50°C for 40 s, elongation at 72°C for 1 min, followed by a final extension at 72°C for 10 min, and initial activation at 95°C for 15 min, followed by 30 cycles of denaturation at 95°C for 50 s, annealing at 50°C for 40 s, elongation at 72°C for 1 min, followed by a final extension at 72°C for 10 min. HotStarTaq Master Mix Kit (Qiagen, Hilden, Germany) and 1 μM of each primer were used for all PCR amplifications.
The PCR products were purified using either QIAquick PCR Purification Kit (Qiagen, Hilden, Germany) or ExoSAP-IT (GE Healthcare). Both strands were sequenced using the same primers as for PCRs, and sometimes internal sequencing primers were added as described by Arpin et al., Bermudes et al. and Rasheed et al. [15][16][17]. The BigDye® Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) were used for sequencing, followed by analysis by capillary electrophoresis with an ABI Prism 3700 DNA Analyzer (Applied Biosystems). Point mutations were accepted if present in both the forward and reverse sequences.

Data analysis
Data were analysed using SPSS software version 20.0 (IBM SPSS statistics 20.0, SPSS Inc., Chicago, IL, USA). Chi-square test was used to determine associations between categorical variables, p < 0.05 was considered statistically significant.
When comparing rates of resistance between HA and CA ESBL, we found that hospital-acquired E. coli and K. pneumoniae were more frequently resistant to ciprofloxacin, gentamicin and nalidixic acid than those isolated from community-acquired infections (Table 3).
Non-ESBL-producing E. coli and K. pneumoniae from hospital-and community-acquired infections were less frequently resistant to gentamicin, nalidixic acid and ciprofloxacin. However, non-ESBL-producing E. coli from outpatients showed moderately high rates of resistance to trimethoprim-sulfamethoxazole, nitrofurantoin and doxycycline compared to isolates from inpatients. Higher rates of resistance to amoxicillinclavulanic acid and trimethoprim-sulfonamide were observed in hospital-acquired as compared to community-acquired isolates of non-ESBL-producing K. pneumoniae. (Susceptibility profile for other isolates see Additional file 1: Table S1).

Discussion
In recent years there has been an alarming increase in community acquired infections with ESBL-producing bacteria [5,12,18]. Spread of these strains in the community is a major concern to patient healthcare, since   The overall frequency of ESBL-producing Enterobacteriaceae among urinary tract pathogens in this study was 23.8%. The frequency of ESBL-producing pathogens was significantly higher in hospital-acquired compared to community-acquired uropathogens. Our finding is in agreement with other studies [19][20][21][22] reporting higher frequency of ESBL-producers in hospital-acquired urinary tract infections compared to community-acquired infections. A possible explanation for this could be that hospital-acquired infections were more likely associated with prolonged hospitalization, comorbidities, previous antibiotic use and urinary catheterization which are well-known risk factors for acquisition of ESBLproducing pathogens [19]. However, the finding of ESBL-producing isolates in community urinary infections is worrisome because of the limited treatment options, considering most of these isolates display multidrug resistance.
Similar to other studies in Africa [4,10,22], we found that ESBL-producing isolates from both hospital and community settings displayed high rates of resistance to ciprofloxacin, trimethoprim-sulfamethoxazole, gentamicin, nalidixic acid and doxycycline. Resistance to commonly prescribed oral antimicrobials in these resource-limited settings, specifically to ciprofloxacin and trimethoprim-sulfamethoxazole, limits outpatient therapeutic options. Considering that most of the outpatients present with uncomplicated urinary tract infections, opting to injectable and expensive antimicrobials increases health-care burdens. We also found non-ESBL-producing E. coli from community-acquired urinary tract infections had moderate to high rates of resistance to trimethoprim-sulfamethoxazole, doxycycline and nitrofurantoin. This could be expected, since oral antimicrobials are inexpensive and easily available over the counter, and self-treatment is common in Africa; these are well known factors driving emergence of antimicrobial resistance bacteria.
Among 41 ESBL defined isolates, 78% were found to carry ESBL genes. Our finding of a predominance of CTX-M-15 is in line to previous and recent studies from hospital and community urinary tract infections [6,7,18,23], and our finding concurs with those of studies from the same setting, which found CTX-M-15 as the dominant ESBL genotype [8][9][10]24]. CTX-M types ESBLs, in particular CTX-M-15, are known for their rapid dissemination world-wide among the members of Enterobacteriaceae [6,12,23,25,26]. It has also been suggested that the widespread use of ceftriaxone and cefotaxime could be a reason of emergence and spread of CTX-M enzymes [27].
Our study had some limitations; one our isolates were collected in 2004 and may not imply the current situation. However, our findings shed lights on community spread of ESBL-producers and suggest existence in Tanzania at least since 2004. Second being a laboratorybased study, clinical information was not obtained, and we could not analyze risk factors for ESBL infections. Furthermore, epidemiological typing to assess clonality of the isolates was not performed, and this could have added value to the understanding of the epidemiological spread of ESBL genes.

Conclusion
In conclusion, we report the presence of Enterobacteriaceae harboring CTX-M-15 type ESBL causing communityacquired urinary tract infections in Tanzania as early as 2004. Furthermore, both ESBL and non-ESBLproducing isolates displayed high rates of multidrug resistance. Further investigation needs to be performed to understand the transmission dynamics of CTX-M type of ESBL resistance.