Isolation of Salmonella strains

Two milliliters of venous blood sample taken from patients was inoculated into 20 ml brain-heart infusion broth (BHI) (CM1135; Oxoid Ltd., England, UK). The culture bottles were incubated at 37 °C for 7 days and examined daily for evidence of bacterial growth, including turbidity and hemolysis. If bacterial growth was observed, a subculture was made after 24 h on blood agar (CM0271, Oxoid Ltd., England UK), Salmonella Shigella agar (SS) (CM0099, Oxoid Ltd., England, UK), and MacConkey agar (CM0007, Oxoid Ltd., England, UK).

At the hospital, a swab of stool was inoculated into 45 ml buffered peptone water (CM0509 Oxoid Ltd., England, UK) and incubated at 37 °C for 18 h. In the laboratory, about 10 g fecal sludge sample from the public toilets was inoculated with a sterile plastic spoon into 90 ml buffered peptone water and incubated at 37 °C for 18 h. Then, 0.1 ml of the overnight culture was transferred into 10 ml of Selenite F broth that was homogenized and incubated at 37 °C for 18 h. Following incubation, the sample was streaked onto SS agar and SSI enteric media (Statens Serum Institute, Denmark) and incubated for 18–24 h at 35–37 °C. Transparent colonies with black centers on SS agar and cream colonies with metallic sheen and a black center due to H₂S production on SSI enteric media were identified as presumptive Salmonella.

All presumptive positive Salmonella isolates were subcultured onto nutrient agar plates (CM0309; Oxoid Ltd., England, UK) and incubated at 37 °C for 16–18 h. Isolates were first confirmed by their reaction in the Minibact-E biochemical tests (SSI, Denmark) and then by slide agglutination using polyvalent antisera (Poly A-E+Vi from SSI, Denmark). Specific serovars were established by serotyping at the WHO National Salmonella and Shigella Center, Institute of Health, Bangkok, Thailand [31].

Phage typing was done for S. Typhimurium (28 isolates) and S. Enteritidis (43 isolates) according to the scheme defined by the PHLS Colindale London at the OIE-National reference laboratory for Salmonella, Istituto Zooprofilattico Sperimentale delle Venezie, Italy [32]. Strains showing a typing pattern that did not conform to any recognized phage type was designated “reacted but did not conform” (RDNC) while those that did not show any phage type pattern at all were designated “non-typable” (NT).

Antimicrobial susceptibility of Salmonella isolates was determined by the agar disc diffusion method on Mueller-Hinton agar (CM0337; Oxoid Ltd., England, UK) according to the protocol and guidelines of the European Committee on Antibiotic Susceptibility Testing (EUCAST). Test were carried out in Ghana and repeated in Denmark for quality assurance. The strains were screened for their susceptibility to the following antimicrobials: ampicillin (AMP, 10 μg), cefotaxime (CTX, 30 μg), cefoxitin (FOX, 10 μg), gentamicin (GEN, 10 μg), ceftazidime (CAZ, 30 μg), amoxicillin clavulanic acid (AMC, 30 + 10 μg), tetracycline (TET, 30 μg), chloramphenicol (CHL, 30 μg), trimethoprim (TRI, 5 μg), sulfamethazole (SUL, 240 μg), nalidixic acid (NAL, 30 μg), and ciprofloxacin (CIP, 5 μg) (ROSCO Diagnostic Neosensitabs, Denmark). E. coli ATCC 25922 and Pseudomonas aeroginosa ATCC 27853 were used for quality control. The EUCAST breakpoints [33] were used to interpret zone diameters.

Isolates that showed reduced susceptibility or resistance to ceftazidime were further investigated for ESBL phenotypes by a double disc diffusion test using the cephalosporins CAZ, CTX, and FOX alone and in combination with clavulanic acid. Interpretations were made according to the EUCAST breakpoints [33].

PFGE genotyping was done for the dominant Salmonella serovars (52 strains of S. Enteritidis and 31 strains of S. Typhimurium) of human origin to establish genetic heterogeneity within serovar. PFGE patterns were compared with patterns previously described for S. Typhimurium (N = 1) serovar isolated from poultry [32]. Overnight culture of Salmonella spp. grown in Luria-Bertani (LB) broth (240230; Difco, MD, USA) was used to prepare genomic DNA according to the CDC PulseNet protocol [34] using 1% agarose (SeaKem ® gold agarose, Lonza, Rockland, ME, USA). DNA embedded in the agarose was digested with the restriction endonuclease XbaI (R0145; New England BioLabs, Inc). The DNA fragments were isolated by electrophoresis in 0.5× TBE buffer using CHEF DR III (Bio-Rad Laboratories, Hercules, CA, USA) system at 14 °C with initial switch time of 2.2 s, final switch time of 54.4 s, current 6 V/cm, included angle 120, and run time of 19 h. Salmonella Braenderup was used as the reference strain and a low range marker (NO350S; New England BioLabs, Inc.) used as the size marker. The gel was stained with 1% ethidium bromide solution for 30 min and de-stained in deionized water for 30 min. The gel image was captured by GelDoc EQ system with Quantity One® software (Version 4.2.1; Bio-Rad Laboratories, Hercules, CA, USA).

All data were entered into spreadsheet of Microsoft Excel 2010 and transferred to the statistical product for service solution (SPSS) (version16, 2008) for Windows which was used for descriptive analysis of data. Confidence intervals (95% CI) for prevalence of Salmonella strains found to be resistant to three or more antimicrobials were calculated as ρ ± z*square root of ((ρ*(1 − ρ)/n) where ρ is the estimated prevalence, n is the population size, and z = 1.96. Chi-square was used to test for differences in multi-resistant Salmonella proportion between blood and stool samples using Minitab software. Phylogenetic analysis of PFGE patterns was done using GelCompar® software (Version 4.6). The TIFF images were normalized by aligning the peaks of the size standard strain (S. Braenderup strain H9812) with the database global standard. Cluster analysis was performed by unweighted pair group method of the PFGE patterns using the Dice coefficient of 0.5% optimization with a 2.0% tolerance.