Prevalence of ESBL-producing Enterobacteriaceae isolated from blood cultures in Mali

Introduction: The increasing frequency of extended-spectrum beta-lactamase (ESBL)-producing Enterobacteriaceae is becoming a serious public health concern. This study sought to determine ESBL frequency in Enterobacteriaceae isolated from patients’ blood cultures in two university teaching hospitals of Bamako, Mali. Methodology: During a three-month period, the presence of Enterobacteriaceae from blood cultures of patients admitted to the university teaching hospitals of Bamako was evaluated. The microbial identifications were initially performed with an API 20E gallery and VITEK2 locally in Mali, and then confirmation in France was performed with a mass spectrometry MALDI-TOF in the bacteriology laboratory of the university teaching hospital of Bichat. Antibiotic susceptibility profiles were determined by the diffusion method as recommended by the European Committee on Antimicrobial Susceptibility Testing (EUCAST). Results: The isolated species were K. pneumoniae (14/40; 35.0%), E. coli (11/40; 27.5%), and E. cloacae (9/40; 22.5%). Of the strains isolated, 21/34 (61.8%) had an ESBL phenotype, including 10/14 (71.4%) K. pneumoniae, 8/11 (72.7%) E. coli, and 3/9 (33.3%) E. cloacae. Resistances associated with ESBL strains of K. pneumoniae, E. coli, and E. cloacae were as follows: gentamicin (10/10, 100%; 6/8, 75%; 2/3, 67%, respectively), amikacin (2/10, 20%; 0/8, 0%; 0/3, 0%, respectively), ofloxacin (8/10, 80%; 7/8, 87%; 3/3, 100%, respectively), and cotrimoxazole (10/10, 100%; 6/8, 75%; 3/3, 100%, respectively). Conclusion: Almost two-thirds (61.8%) of Enterobacteriaceae isolated from our blood cultures were ESBL producers. Only susceptibilities to carbapenems and to amikacin were fully conserved within the strains.

In Mali, the rare data available on the prevalence of ESBL-producing Enterobacteriaceae isolated from blood cultures are frightening.Access to antibiotic susceptibility testing currently remains very limited in the country.Most people in the capital city, in most cases, get care in the primary health centers such as the referral health centers or in the community health centers, which are more accessible to them, but drug susceptibility testing is not available.Likewise, outside the capital city in the rest of the country, none of the eight regions of the country has continuously maintained drug susceptibility testing capacity.It has therefore become important to determine the prevalence of ESBL-producing strains among patients who require urgent appropriate treatment such as patients with bacteremia.Indeed, empirical treatments usually administered to these kinds of patients may not be effective if the patients happen to have an ESBLproducing Enterobacteriaceae.In this study, we sought to investigate the prevalence of ESBL phenotype among Enterobacteriaceae isolated from patients with positive blood cultures at the university teaching hospitals of Bamako in Mali.

Type and place of the study
This was a prospective study conducted at the university teaching hospitals of Bamako (CHU Point G and CHU Gabriel Touré) from January to March 2014.The university teaching hospitals of Bamako are the structures of third reference level, representing the top of the pyramid of health care in Mali.In general, these are centers that receive patients with complicated diseases from other health care centers.

Patient recruitment
The study was open to any patients referred from another health center where they had been previously hospitalized, who had a body temperature ≥ 39°C with suspected invasive bacterial infection.

Sample collection
For each blood culture, venous blood (8 to 10 mL from adults and 1 to 5 mL from infants) was collected and injected directly into a BD Bactec Plus Aérobie/F blood culture bottle (Becton Dickinson, Franklin Lakes, USA) or BD Bactec Peds Plus/F blood culture bottle (Becton Dickinson, Franklin Lakes, USA), which was then introduced into a Bactec 9050 (Becton Dickinson, Franklin Lakes, USA).Only one blood culture per patient was performed.The aseptic blood sample collection was performed following the manufacturer's instructions.

Bacterial identification
In Bamako, Mali, Drigalski lactose agar (bioMérieux, Marcy l'Etoile, France) was used for the selective isolation of Enterobacteriaceae.The primary identification was done by the API 20E system (bioMérieux, Marcy l'Etoile, France), with the galleries inoculated with calibrated bacterial suspensions.The confirmation was done using the automated phenotypic identification system VITEK2 (bioMérieux, Marcy l'Etoile, France).
Enterobacteriaceae strains were kept in cryovials in 10% of glycerol and stored in a -80°C freezer before getting shipped to Paris for further testing.

Quality control
All identifications were confirmed in France by mass spectrometry MALDI-TOF (Bruker, Wissenbourg, France) in the bacteriology laboratory of Bichat University Teaching Hospital.The mass spectrometer allows identification of bacteria by analysis of their total proteins (ribosomal proteins and membrane-associated proteins).

Ethical considerations
This study was approved by the ethics committee of the Faculty of Medicine and Odonto-Stomatology/Faculty of Pharmacy, University of Sciences, Techniques and Technologies of Bamako (USTTB).Written consent was obtained from all patients included in the study.

Patients' characteristics
From a total of 843 blood cultures collected from patients hospitalized at CHU Point G, Enterobacteriaceae were isolated from 14 blood cultures.Similarly, of 1,262 blood cultures taken from children hospitalized in the pediatric department of CHU Gabriel Touré, Enterobacteriaceae were isolated from 25 blood cultures.The patients' ages ranged from 31 to 40 years (42.9%) at CHU Point G, and from 1 to 5 years (48.0%) at CHU Gabriel Touré (Table 1).Of all participants, 57.1% and 64.0% from CHU Point G and CHU Gabriel Touré, respectively, were males; 28.6% and 28.0% of patients, respectively, were living in the urban area of the Commune I of Bamako City (the most populated neighborhood of Bamako) (Table 1).
K. pneumoniae was the predominant species on both sites, representing 42.9% and 30.8% of Enterobacteriaceae isolated at CHU Point G and CHU Gabriel Touré, respectively (Table 2).E. coli was the second-most common at CHU Point G (28.6%).At CHU Gabriel Touré, E. coli and E. cloacae shared the same second position, with 26.9% of Enterobacteriaceae isolated (Table 2).One patient had two strains of E. coli and E. cloacae at CHU Gabriel Touré.
In our study, none of our Salmonella enterica were ESBL producers, while in Nepal, 72.0% of Salmonella isolated from blood cultures were ESBL producers [26]; however, in another study, only 0.5% of S. enterica were found to be ESBL producers [27].Moreover, we noted that the frequency of ESBLproducing Enterobacteriaceae varies randomly between countries.For example, 32.0% of K. pneumoniae were reported in India [28], but only 8.1% of E. coli in the United Kingdom [29].Also, on 44 ESBL-producing Enterobacteriaceae in New Zealand, 40.9% were K. pneumoniae, 36.3% were E. coli, 20.5% were E. cloacae, and 2.3% were E. aerogenes [30].
Another interesting finding on the distribution of species in our study is the predominance of K. pneumoniae in both sites: 42.9% and 30.8% at CHU Point G and CHU Gabriel Touré, respectively.This result confirms the predominance of K. pneumoniae among Enterobacteriaceae from blood cultures in hospitalized patients exposed to nosocomial infections.E. coli and E. cloacae came in second position with, respectively, 28.6% and 14.3% at CHU Point G, and 26.9% and 26.9% at CHU Gabriel Touré.This predominance of K. pneumoniae was also reported in Kabul, Afghanistan; of 212 Enterobacteriaceae isolated from blood cultures, 66 sere K. pneumoniae, 42 (19.8%)were E. cloacae, and 35 (16.5%) were E. coli [19].That study had a bigger sample size and the study period covered two years (November 2009 to November 2011), while ours covered only a period of three months in order find preliminary data for Mali.Our results were different from those of a tertiary care university teaching hospital in South Korea, where 101 Enterobacteriaceae strains were isolated from blood cultures, of which 80 (79.2%) were E. coli and 21 (20.8%) were K. pneumoniae [22].This distribution has been also observed in China in blood cultures performed in children having hematological malignancy and undergoing chemotherapy; of 109 Enterobacteriaceae strains, 58 (53.2%) were E. coli and 51 (46.8%) were K. pneumoniae [20].Therefore, it can easily be concluded that in these South Korean and Chinese studies, E. coli was the predominant bacteria, although those studies were conducted in a hospital setting.

Conclusions
Among our isolates, K. pneumoniae was the predominant species of the Enterobacteriaceae of our study.About two-thirds (61.8%) of the isolates from blood cultures were ESBL-producing Enterobacteriaceae.Of the panel of antimicrobials we tested, only carbapenems and amikacin were fully effective on all the strains.

Table 1 .
Distribution of patients according to sex, age range, and location

Table 2 .
Distribution of patients based on isolated germs and identified extended-spectrum beta-lactamases (ESBLs) * 1 Escherichia coli and 1 Enterobacter cloacae isolated from the same patient; ** Morganella morganii, Proteus mirabilis and Salmonella enteritidis were not ESBL-producing and therefore not counted.

Table 3 .
Resistant to antibiotics at both sites