Clinical and Bacteriological Profiles of Neonatal Sepsis in a Tertiary Hospital, South-Western Nigeria

Abstract

Introduction Antibiotic-resistant bacteria complicate treatment options in neonatal sepsis, especially in developing countries. This study determined the epidemiology and bacteriological characteristics of neonatal sepsis at a tertiary hospital, in southwest Nigeria. Methods This was a cross-sectional study from December 2017 to April 2019 among admitted babies with clinical neonatal sepsis. Blood culture was performed by semi-automated system, sepsis biomarker assay (serum procalcitonin) by a semi-quantitative kit while proforma was used to capture clinico-demographic data. Bacterial identification, antibiotic susceptibility patterns, determination of genetic elements mediating resistance, were performed by standard methods and polymerase chain reaction protocols, respectively. Quantitative data were expressed as frequencies, mean; bivariate and multivariate analyses were performed by Chi-square or Fishers’ exact test and logistic regression. Results Of the 192 cases of neonatal sepsis enrolled, 42.7% (82/192) were blood culture positive. Factors associated with blood culture positivity included respiratory rate ≥60 bpm (60/82; p < 0.03), lethargy/unconsciousness (59/82; FE=7.76; p < 0.001), grunting respiration (54/82; p = 0.04), meconium passage before birth (17/82; p = 0.03) and prolonged rupture of membranes ≥24 h (50/82; FE = 6.90; p = 0.01). On the other hand, mortality in the neonates was associated with elevated serum procalcitonin assay (>0.5 ng/mL) χ² = 13.58; p = 0.03] and Gram-negative bacteremia (χ² = 24.64; p < 0.001). The most common bacterial isolates were Staphylococcus aureus (42/82), coagulase-negative Staphylococcus spp. (17/82), Enterobacter spp. (8/82), and Acinetobacter spp. (6/82). Methicillin resistance was present in 85.7% (36/42) of Staphylococcus aureus and 52.9% (9/17) of coagulase-negative Staphylococcus, while extended-spectrum beta-lactamase (ESBL) and AmpC enzymes were present in (21.1%; 4/19) of the Gram-negative bacilli. Conclusions Almost half of the cases of clinically diagnosed neonatal sepsis have bacterial etiologic confirmation of sepsis. Gram-negative bacteremia and high serum procalcitonin predict mortality in neonatal sepsis. There was high resistance to common antibiotics for the treatment of neonatal sepsis in our settings.

Introduction

While neonatal sepsis is a major public health challenge globally with massive economic, and morbidity consequences, developing countries bear the brunt of mortality because of delayed diagnosis and inappropriate antibiotic therapy [1]. In 2019, a quarter of all 2.4 million neonatal mortalities were attributable to neonatal infections, with the highest burden (27/1000 live births) in Sub-Saharan Africa and Nigeria accounting for the highest rate in Africa [2].

Bacterial agents of neonatal sepsis (NNS) vary widely from one location to another, also depending on the onset of illness, economics of the region, and sophistication of neonatal medical intervention, but Sub-Saharan Africa, with more deaths from NNS, has poorly defined epidemiology and bacterial etiology [1]. Moreover, the common overuse of broad-spectrum empiric antibiotics without recourse to microbiological diagnosis in managing sick neonates drives the emergence of multidrug resistant (MDR) bacterial pathogens including “those that are resistant to last-resort antibiotics” yet the pattern of MDR is poorly characterized in the region [3]. Inadequacy of blood culture as “gold standard” in the investigation of NNS, because negativity does not exclude bacterial infections, necessitates investigation of NNS with the use of serum biomarkers including serum procalcitonin in addition to other clinical parameters, for early assessment of disease severity and prognostication [4].

In neonates, incomplete development of innate and adaptive immunity, protean clinical manifestation, lack of consensus of definition, and rapid fatal progression of illness make empiric antibiotic therapy essential for reducing morbidity and mortality for NNS, which is dependent largely on the knowledge of implicating bacterial agents prevalent in a particular location and their antibiotic susceptibility patterns [5]. While the World Health Organization (WHO) recommended intramuscular/intravenous benzylpenicillin or cloxacillin (when Staphylococcus aureus is the most likely pathogen) and gentamicin or intramuscular/intravenous ampicillin in hospitalized neonates with serious bacterial infection [6], there are concerns because of the emerging antibiotic-resistant pathogens in the etiology of NNS. Against this background, third-generation cephalosporin (ceftriaxone) and gentamicin are the empiric antibiotics for suspected early-onset [EOS] (<72 hours of birth) NNS, while in late-onset [LOS] (>72 hours of birth) NNS, ceftriaxone/ceftazidime and gentamicin are the empiric antibiotic used in our settings. The role of antibiotic resistance in NNS in resource-poor settings is evolving and poorly characterized owing to the paucity of studies. Thus, our study determined the epidemiology and bacteriological characteristics of neonatal sepsis at a tertiary hospital, in Southwestern Nigeria.

Methods

Ethical consideration

The Ethics and Research Committee of the Obafemi Awolowo University Teaching Hospitals Complex (OAUTHC) Ile-Ife approved the study with protocol number ERC/2017/08/01.

Study location and study design

This study was a hospital-based, descriptive, prospective cross-sectional study conducted at the neonatal unit of the OAUTHC, a 50 bedded neonatal ward that is a tertiary referral center for the southern region of Nigeria. It is one of the centers for the Kangaroo Care Initiative of the WHO (early, prolonged, and continuous skin-to-skin contact between the mother and her low birthweight infant until at least 40 weeks of age) in preterm neonates.

Study population

This prospective, cross-sectional, hospital-based study was conducted from December 2017 to April 2019, among neonates admitted with the clinical diagnosis of sepsis meeting European Medicines Agency (EMA) Expert Meeting on Neonatal and Paediatric Sepsis criteria [7]. Neonates were consecutively recruited and followed up until discharged or deceased. The neonates with the onset of sepsis occurring within the 72 hours of birth are categorized as early-onset neonatal sepsis (EOS) and those with the onset of sepsis after 72 hours are categorized as late-onset neonatal sepsis.

Sample collection

One to three milliliters of venous blood were obtained by the neonatologist into a set (2 bottles) of commercially produced pediatrics aerobic broth bottles (BACTEC 9050 Peds PlusTM/F culture vials, Becton Dickinson, Belgium) after thorough skin disinfection. The samples were immediately transported to the laboratory [8].

Cultures

Blood culture

Blood in broths was incubated inside a semi-automated BACTEC 9050 blood culture machine (BACTEC ™ 9050, Becton Dickinson, Belgium) for 24 hours to 7 days. Upon positive signaling by the blood culture incubator system, direct Gram of the blood culture broth and subculture was done on 5% sheep blood agar, chocolate agar and MacConkey agar (Oxoid, United Kingdom), which were appropriately incubated under aerobic and capnophilic (5-10% CO₂) conditions at 35-37°C for 18-24 hours. Neonates with clinical signs of meningitis and those with positive blood culture underwent lumbar puncture to rule out accompanying neonatal meningitis [8]. (Group B Streptococcus is not routinely tested in pregnant mothers at 35-37 weeks at OAUTHC).

Interpretation of culture

The result of the cultures was used to categorize the neonatal sepsis into culture-negative sepsis and culture-positive sepsis according to the National Neonatal Forum of India categorization [9].

Culture-negative sepsis assessment was made when neonates had the following, without positive culture results:

• Maternal predisposing factors like fever or foul-smelling liquor or prolonged rupture of membranes (>24 h) or gastric polymorphs (>5 per high power field).
• Presence of at least two of these four parameters:
  • total leukocyte count (<5000/cmm),
  • band to total polymorphonuclear cells ratio of >0.2,
  • absolute neutrophil count <1800/cmm,
  • C-reactive protein (CRP) >1mg/dL and ESR >10 mm-first hour.
• Radiological evidence of pneumonia.

Culture-positive sepsis assessment was made when neonates had clinical signs and symptoms suggestive of systemic infection, lower respiratory tract infections and meningeal infections, in addition to the following:

• Isolation of pathogens from blood or CSF or urine or abscess(es).

Pathological evidence of sepsis on autopsy.

Serum biomarker

StrongStep® Procalcitonin (PCT) Rapid Test (Liming Bio, China), a rapid immune-chromatographic assay for semi-quantitative detection of serum procalcitonin, was used to measure procalcitonin concentration according to the manufacturer’s instruction. The PCT value was interpreted by the age-adjusted B·R·A·H·M·S™ PCT cut-off values in newborns. Serum PCT value of >0.5 ng/mL was considered positive in LOS and >10 ng/mL was considered positive in EOS [10].

Bacterial identification

The isolated bacterial colonies were identified by Gram staining and biochemical tests. Gram-positive cocci were identified as Staphylococcus spp. by positive tests with catalase, coagulase, growth on DNAse agar and mannitol salt agar, while organisms were identified as Streptococcus spp. using RapID™ STR (Thermo Fisher Scientific, Remel Products, United States) [8]. Coagulase-negative Staphylococcus spp. were categorized as true pathogens when pure cultures of the same isolate were recovered from both blood cultures, as well as other associated evidence of bacterial infections as stated by EMA Expert Meeting on Neonatal and Paediatric Sepsis criteria [7]. Facultative anaerobic and aerobic Gram-negative bacilli (GNB) were identified after Gram staining and oxidase test with Microbact™ GNB 24E (Oxoid, United Kingdom) [8].

Antibiotic susceptibility test

Antibiotic susceptibility testing was done by modified Kirby-Bauer disk diffusion method and interpreted following guidelines by the Clinical and Laboratory Standards Institute (29^th edition, 2019). Control strains for testing were Escherichia coli ATCC® 25922, Escherichia coli ATCC® 35218 (for β-lactam/β-lactamase inhibitor combinations), Pseudomonas aeruginosa ATCC® 27853 and Staphylococcus aureus ATCC® 25923.

Susceptibility of S. aureus to vancomycin was determined with vancomycin E-test strips (256–0.015 µg/mL) in minimum inhibitory concentration (MIC) (Oxoid, United Kingdom) and interpreted according to CLSI guidelines (29^th edition, 2019) [11].

Phenotypic resistance testing of extended-spectrum β-lactamase (ESBL) production, AmpC beta-lactamase and carbapenemase production

A combination disc diffusion test (CDDT) was done for phenotypic confirmation of extended-spectrum β-lactamase production (ESBL) production in Klebsiella pneumoniae, Klebsiella oxytoca, Escherichia coli, and Proteus mirabilis as stated by CLSI guidelines (29^th edition, 2019). Double disk synergy test (DDST) using cefpodoxime (30 μg), cefepime (30 µg) and amoxicillin-clavulanic (20/10 μg) disks was done for group 2 Enterobacterales organisms (Enterobacter spp., Citrobacter freundii, Morganella morganii, Providencia stuartii, Serratia spp. and Hafnia alvei) as described in EUCAST guidelines (Version 2.0, 2017). AmpC beta-lactamase production was confirmed by the AmpC disk test as described by EUCAST guidelines (Version 2.0, 2017). Phenotypic confirmation of carbapenemase production was detected by modified carbapenem inactivation methods (mCIM) according to CLSI guidelines (29^th edition, 2019) [11,12].

Determination of genetic determinants of phenotypic resistance

All multidrug-resistant phenotypes tested (ESBL, carbapenem resistance and methicillin resistance) were further investigated for a genetic determinant of resistance by PCR-based protocols with unique oligonucleotide primers specific for the expected genes as detailed in the supplementary material (Supplementary material 1) [13,14,15,16,17,18].

Primer sequences for selected ESBL genes (blaTEM, blaSHV, blaCTX-M, blaKPC, blaNDM, blaVIM, bla_OXA, mecA) and amplification reactions

DNA templates were prepared from purified bacterial isolates by boiling method, and amplified PCR products were electrophoresed in 1.5% (weight for volume) agarose gel (Biomatik, Canada) in 1X Tris-Acetate EDTA (TAE) buffer for 45 minutes. Standard DNA molecular size markers (100-1500 bp) were used to estimate the size of the PCR product and viewed under a 302 nm UV trans-illuminator [8].

Data and statistical analysis

Statistical Package for Social Sciences (SPSS) version 22 (SPSS Inc., USA) was used for data analysis. Data were summarized in tables with categorical variables compared using Chi-squared or Fisher’s exact test. Bivariate logistic regression analysis was used to identify the risk factors for NNS and the risk of mortality from NNS. A p-value of <0.05 (set at 95% confidence interval) was considered statistically significant.

Results

Incidence rates

There were 2954 live births during the duration of the study of which 763 neonates were admitted for various clinical conditions into the neonatal ward. Out of the 192 neonates with clinical sepsis who were sampled, 86 were in-born, thus the incidence rate of neonatal sepsis per live birth was 29.1 per 1000 (86/2954), as well as 82 of the samples were considered culture-positive neonatal sepsis (Figure 1).

Clinical characteristics of neonates

One hundred and ninety-two sampled neonates distributed a 1.3:1 (107/85) male-to-female ratio. The mean gestation age at birth was 35.4 weeks ± 4.30 weeks, while the mean age at the onset of admission was 6.2 days ± 7.4 days. In-born neonates represented 44.8% (86/192) and the number of neonates delivered at the tertiary hospitals accounted for 37.5% (72/192) –

Table 1. Characteristics of the neonates with clinical neonatal sepsis sampled at Obafemi Awolowo University Teaching Hospitals Complex, Ile-Ife, South-west Nigeria, December 2017 to April 2019. 

.

Signs and symptoms of sepsis associated with culture-positive neonatal sepsis

The signs of neonatal sepsis and risk factors significantly associated with culture-positive neonatal sepsis were tachypnea [crude odds ratio 0.43; p=0.01], grunting (crude odds ratio 4.70; p=0.04), lethargy/unconsciousness (crude odds ratio 6.20; p=0.01) and prolonged rupture of membranes (>24 hours) (crude odds ratio 7.56; p=0.06). However, on logistic regression, the independent risk factors for positive culture neonatal sepsis were tachypnea (adjusted odds ratio 0.49; p=0.05] and lethargy/unconsciousness (adjusted odds ratio 0.54; p=0.01) –

Table 2. Association between clinical variables of neonates and neonatal sepsis at Obafemi Awolowo University Teaching Hospitals, Complex, Ile-Ife, South-west Nigeria, December 2017 to April 2019. 

.

Factors associated with neonatal deaths

Positive serum procalcitonin assay (χ²=10.224; p=0.001] and Gram-negative bacilli bacteremia (χ²=12.948; p<0.001) were significantly associated with neonatal deaths (

Table 3. Association between neonatal sepsis determinant and treatment outcome at Obafemi Awolowo University Teaching Hospitals Complex, Ile-Ife, South-west Nigeria, December 2017 to April 2019. 

).

Bacterial profile of neonatal sepsis

Eighty-two isolates were retrieved from 192 blood culture samples, categorized into Gram-positive cocci 61 (74.3%), Gram-negative bacilli 19 (23.2%) and Candida spp. 2 (2.4%). The distributions of the bacteria are shown in

Table 4. Antibiotic resistance pattern of isolated of bacteria associated with neonatal sepsis at Obafemi Awolowo University Teaching Hospitals Complex, Ile-Ife, South-west Nigeria, December 2017 to April 2019. 

. Staphylococcus aureus was the most isolated pathogen 51.2% (42/82) in NNS while the most common Gram-negative bacteria isolated was Enterobacter agglomerans 7.3% (6/82). Staphylococcus aureus was the most common bacteria isolated in EOS 22.5% (33/147) and LOS 20.0% (9/45). Also, the most common bacteria isolated in term neonates with sepsis was Staphylococcus aureus 23.6% (26/110) as well as in preterm neonates 19.5% (16/82). There were no bacterial isolates significantly associated with the onset of NNS and maturity of the neonates at birth (

Table 5. Association between the isolated bacteria, the onset of neonatal sepsis and maturity of neonates at birth at Obafemi Awolowo University Teaching Hospitals Complex, Ile-Ife, South-west Nigeria, December 2017 to April 2019. 

).

Pattern of antibiotic resistance in isolated bacteria in neonatal sepsis

The pooled Gram-negative bacilli (GNB) organisms showed high to moderate resistance rates to the tested antibiotics: ampicillin 88.9% (8/9), cefuroxime 75.0% (12/16), ceftriaxone 75.0% (12/16), ciprofloxacin 52.6% (10/19), gentamicin 40.0% (6/15) and meropenem 31.3% (5/16). Four of the 18 (22.2%) Gram-negative bacilli were phenotypically confirmed as ESBL- and AmpC beta-lactamase producers respectively.

Among Gram-positive cocci, forty-one (97.6%) Staphylococcus aureus isolates were resistant to penicillin and 76.2 % (32/42) resistant to each gentamicin and ciprofloxacin. The resistance to ceftaroline was 42.9% (18/42) while resistance to quinupristin-dalfopristin was low at 7.2%. S. aureus was susceptible to linezolid and vancomycin. Coagulase-negative Staphylococcus spp. showed considerable resistances to penicillin (100%), gentamicin (47.1%) and ciprofloxacin (70.6%). A total of 36 (85.7%) Staphylococcus aureus and 9 (52.9%) CoNS were methicillin-resistant (Table 4).

Prevalence and types of resistance genes among the bacterial isolates in neonatal sepsis

Of the 36 phenotypically confirmed methicillin-resistant Staphylococcus spp., the mecA gene was present in 21 (58.3%) S. aureus and 9 (44.4%) coagulase-negative Staphylococcus. Four of the 19 (21.1%) Gram-negative bacteria were phenotypically confirmed ESBL-producers harboring one or more resistance-determining genes. Of the total 11 genes, 4 were bla_CTX-M, 3 bla_SHV and 4 bla_TEM. There was no carbapenemase determining gene detected in any of the 5 phenotypically confirmed carbapenem-resistant isolates.

Discussion

This study revealed that the incidence rate of neonatal sepsis in our settings is markedly high compared to the pooled incidence reported by Hibberd et al. (7.7/1000 live births) from a multicenter prospective cohort study involving 7 countries from Sub-Saharan Africa, South Asia and Latin America. This lower incidence could be due to the fact that this was a community study instead of a hospital-based study as our was [19]. Also, our incidence was high compared to the global estimated incidence rate of neonatal sepsis in the hospital setting (19.9/1000 live births) by a systematic review by Fleischmann et al. This is not surprising because of poor obstetric care, unhygienic delivery environment, insufficient standard hospitals and delayed response to both neonatal and maternal care in our settings. The incidence of 29.1/1000 live births was also higher than the 22.9/1000 live births which was documented by Adejuyigbe et al. about two decades ago in the same hospital. While the incidence of neonatal sepsis fell steadily over this time globally, however, this increased incidence could be explained by improvement in neonatal care and facility at OAUTHC which could be driving patronage [20].

The clinical signs and symptoms associated with neonatal sepsis in this study were tachypnea and lethargy/unconsciousness which were consistent with the various studies on clinical signs that best identify severe infections in young infants [21]. Early detection of these signs can promote early decisions on referral to hospital for diagnosis and treatment to ensure better neonatal outcomes, especially in low- and middle-income countries (LMIC) where the sophisticated algorithm for diagnosis might not be readily available [21].

Serum procalcitonin has been used to differentiate between bacterial and non-bacterial infections. Furthermore, persistence in the serum procalcitonin elevation indicates continuous infection which can lead to a fatal outcome. Elevated serum procalcitonin (>0.5 ng/mL) was associated with mortality in this study and was congruent with findings from Japan where elevated serum procalcitonin correlated with mortality (p<0.001) [22]. Serum procalcitonin can be used to measure the severity of the bacterial infection and its utility in the routine management of neonatal sepsis in LMICs with limited access to standard blood culture may help in a timely escalation of treatment thereby ameliorating morbidity and reducing mortality.

Studies have highlighted the diversity of bacterial agents of neonatal sepsis in hospital settings. The most common pathogen isolated in our study was Staphylococcus aureus, which accounted for slightly more than half of the isolates (51.2%), which was congruent with a study from Calabar (53.0%) in Southern Eastern, Nigeria. Moreover, studies from Ilorin (29.3%), Maiduguri (46.2), Abakaliki (45%), also documented Staphylococcus aureus as the most common pathogens of neonatal sepsis, different from Dawodu et al. from Lagos, South-western Nigeria, which reported Klebsiella spp. (46.6%) as the most common pathogen in their study [24]. The etiology of neonatal sepsis is dynamic and varies widely based on geographical region, level of technological advancement available for neonatal sepsis management, adherence to infection prevention control, prior antibiotic exposure and level of hygiene at the delivery centers. The systematic review of the burden of neonatal sepsis in South East Asia by Chaurasia et al. reported Klebsiella spp. as the most common pathogen responsible for neonatal sepsis and Staphylococcus aureus was ranked third after Escherichia coli [25]. The role of Streptococcus agalactiae in neonatal sepsis in our facility was not evident, and this is similar to other studies from low-and-middle-income countries, as well as in Nigeria where the role of Group B Streptococcus as an agent of neonatal sepsis is minimal as opposed to high-income-countries [24]. Gram-positive bacteremia was more pronounced in this study, congruent with findings studies from LMIC [26].

There was high methicillin resistance among Staphylococcus aureus (85.7%) and coagulase-negative Staphylococcus spp. (52.9%) in this study, and this is a common experience in developing countries across Africa, South Asia and South America [25,26]. The increasing methicillin resistance among staphylococci makes the efficacy of the WHO-recommended empiric therapy inadequate in managing neonatal sepsis. While evidence from clinical trials among neonates with the clinical diagnosis of sepsis where referral is not possible favored a simplified therapeutic regimen of gentamicin and amoxicillin or ampicillin across poor economic settings, this might not be an objective reflection of antibiotic susceptibility of bacterial agents of neonatal sepsis due to lack of laboratory evidence [27,28]. Therefore, the choice of the empiric therapy for neonatal sepsis should be based on the locally generated evidence to reduce morbidity, mortality and improve antibiotic stewardship in neonatal units [5]. The most common Gram-negative bacteria isolated in this study was Enterobacter agglomerans, followed by Acinetobacter spp. This is different from the pattern observed in the multi-national study by Kirsty et al., which spanned 7 LMIC of Africa (including Nigeria) and Asia which reported Klebsiella spp. as the most common Gram-negative bacterial agent of neonatal sepsis [29]. Emergence of Acinetobacter spp. and Enterobacter spp. as common agents of Gram-negative sepsis was highlighted in various studies with associated resistance to several broad-spectrum antibiotics, coupled with limited access to effective last-resort antibiotics useful against multidrug resistant bacteria, contributing to poor neonatal outcomes [25,29]. Of the phenotypically confirmed MRSA, slightly more than half had the mecA gene as the genetic determinant. The others lacking this gene may have alternate methicillin resistance genes like the mecC gene and other numerous described mec genes types and subtypes [30]. All the Gram-negative isolates harbored at least one genetic determinant for ESBL, which is in agreement with the high level of antibiotic resistance in Gram-negative pathogens of neonatal sepsis. bla_CTX-M is the most successive gene for ESBL strains, and in this study, it is one of the prevalent ESBL genes as also demonstrated by the Burden of Antibiotic Resistance in Neonates from Developing Societies (BARNARDS) study in Nigeria [29].

The empiric therapy for neonatal sepsis at OAUTHC had been switched from the WHO recommended first-line antibiotics (ampicillin and gentamicin) to the second-line antibiotics including ceftriaxone and gentamicin to which the bacteria have shown considerable resistance also. The BARNARDS study further exemplified the enormity of resistance making the use of WHO-recommended first- and second-line antibiotics ineffective for treatment in many cases [29]. In this study, about a third (31.3%) of cumulative Gram-negative isolates were non-susceptible to meropenem. Overuse of carbapenems is an important factor driving the emergence of carbapenem-resistant Enterobacterales, which is more worrisome in our environment, calling for increased evidence-based targeted therapy by prioritizing the use of microbiological laboratory and antibiotic surveillance.

The pathogens isolated and their antibiotic resistance patterns represent the occurrence at the referral center, which may have a bias for severe bacterial infections. Therefore, generalization of the results should be done with caution. Also, prior antibiotic exposure before blood culture sampling in some neonates might also amplify antibiotic-resistant bacteria more in this study. Moreover, further studies using discriminatory tools like whole genome sequencing should be performed on the isolates to investigate their transmission dynamics as this would inform better infection prevention strategies for NNS prevention.

Conclusions

In conclusion, tachycardia and lethargy/unconsciousness were signs associated with culture-positive neonatal sepsis while Gram-negative bacteriemia and high serum procalcitonin assay predicted mortality in neonatal sepsis. There was a high level of antibiotic resistance to the WHO-recommended empiric antibiotics for the treatment of neonatal sepsis in our settings, and this calls for the review of the local antibiotic treatment protocol for neonatal sepsis in this geographical region, which should be based on laboratory surveillance evidence.

Supplementary material

Supplementary table 1. Primer sequences for selected ESBL genes (bla_TEM, bla_SHV, bla_CTX-M, bla_KPC, bla_NDM, bla_VIM, bla_OXA, mecA) and amplification reactions. 

Supplementary table 1. Primer sequences for selected ESBL genes (bla_TEM, bla_SHV, bla_CTX-M, bla_KPC, bla_NDM, bla_VIM, bla_OXA, mecA) and amplification reactions.