Background

Ensuring safe food supply has been one of the major challenges and concerns for producers, consumers and public health officials in both developing and developed countries. This is because foods excessively contaminated with pathogenic and spoilage micro-organism are undesirable and can cause food borne illnesses [1–3]. Such illnesses cost billions of dollars in medical care and sometimes even result to death [4]. Several epidemiological reports have implicated foods from animal origin as major vehicles associated with illnesses caused by food borne pathogens [5–7]. Coliform bacteria, especially fecal coliforms, are good microbial indicators of the potential presence of disease causing bacteria and also show the general sanitary quality of the food. Food contamination by Escherichia coli is closely associated with fecal contamination. This is because E. coli are the most prevalent commensal enteric bacteria in animals and humans and are also important zoonotic agents that can be implicated in animal and human infectious diseases [8].

Raw or undercooked chicken meat is particularly prone to contamination. The microbiological quality of chicken meat as purchased by consumers depends mostly on; the slaughter process, sanitation during processing and packaging, maintenance of adequate cold chain storage from the processing to the retail level and to the consumer and finally sanitation during handling at the retail end [9–12]. Chicken meat can also act as a reservoir of drug resistant bacteria. Antimicrobial resistance among E. coli in food animals such as chicken is of increasing concern due to the potential for transfer of these resistant pathogens to the human population [13–15].

In urban areas such as Nairobi, marketing of chicken products is generally undertaken in retail outlets such as supermarkets, local butcheries located in different geo-socio-economic status and even from street vendors in some low income settings. Public health research in countries such as the United States of America focusing on food qualities demonstrated that stores in low socio-economic status populations, because of a higher prevalence of food safety violations, were shown to be consistently exposed to food that is of lower microbial quality. Because of this access pattern, such populations are placed at increased risk of food-borne illnesses [16, 17]. In Kenya, there is a paucity of data on coliform contamination and antimicrobial resistant E. coli in raw retail chicken supplied to retail outlets in the city. Studies related to antimicrobial resistant E. coli have been done on isolates from farm animals and chicken carcass samples from slaughter [18] but not from raw retail chicken which is made available to consumers. Furthermore, no studies are available on the contamination levels in chicken meat available to populations living in different socio-economic status yet, these data are essential for performing risk assessment and risk management for food safety. This study reports on the microbiological quality and safety of chicken meat available to populations who purchase from different retail outlets (supermarkets, retail outlets in high income areas, retail outlets in middle income areas and retail outlets in low income areas) in the city of Nairobi.

Study area and sampling

The study was conducted in Nairobi County. Nairobi is the administrative and commercial capital in Kenya and home to thousands of businesses including the retail chicken business.. There is a huge disparity in income levels and population densities in Nairobi. The people living in the western suburbs are generally the more affluent while the lower and middle-income populations dominate the eastern suburbs [19].

In a cross-sectional study, chicken samples were randomly purchased from August 2011 to February 2012 from different retail outlets spread over 28 locations in Nairobi. In order to take into account compounding factors of socio-economic status within Nairobi, the retail outlets where samples were purchased were classified into supermarkets, shops from high end areas (low densely populated, up-market residential suburbs), shops from middle end areas (middle densely populated areas further classified into high middle and low middle income areas) and shops from low end areas (densely populated slums and informal settings). Classification of locations into these groups was done based on a study on residential segregation in Nairobi [20]. A total of two hundred chicken samples (n = 39, supermarkets; n = 39 high end area retail outlets; n = 84, high middle end area retail outlets; n = 20, low middle end area retail outlets and n = 18, low end area retail outlets) were purchased. Both freshly slaughtered as well as samples that have been in storage in freezers were included in the study. Packaged chicken samples were transported in cool boxes to the laboratory for bacterial isolation within 1 hour from time of collection. Samples were processed at the Kenya Medical Research Institute, Centre for Microbiology Research laboratories, Kenyatta hospital compound. Ethical approval to perform this study was obtained from the Ethics Review Committee at Kenya Medical Research Institute, Reference SSC No. 2036.

Bacteriological analysis

Enumeration of E. coli and coliform bacteria from the chicken samples was performed as described in the Association of Analytical Communities International, official methods of analysis using 3 M petrifilm E. coli/Coliform count plates [21] with slight modification. The samples, which included the skin of the chicken along with the meat itself, were aseptically removed from the package and 100 gram piece of chicken was weighed and placed in 100 ml sterile distilled water. One ml of the rinse water from the sample was placed onto the center of the bottom film and covered carefully avoiding entrapment of air bubbles and the plates from all samples incubated for 24 ± 2 hrs at 35°C. Colony counts for each petrifilm were done the following day and the microbial count results converted to the base 10 logarithm of the number of colony forming units per ml (cfu/ml) rinse water obtained from the samples. Typical E. coli colonies growing on the petrifilm were then selected and sub-cultured in MacConkey agar to obtain pure colonies. After a series of biochemical tests for confirmative identification [22], positive E. coli isolates were stored at -80°C in trypticase soy broth for further antimicrobial susceptibility tests and PCR to determine the presence of virulence genes.

Antimicrobial susceptibility testing

Using the Kirby-Bauer disc diffusion method [23], E. coli were tested for their susceptibility to 12 commonly used antimicrobials on disks containing; Ampicillin (AMP 10 μg), Amoxycillin-clavulanic acid (AMC, 30 μg), Tetracycline (TE 30 μg), sulphamethoxazole-trimethoprin (SXT 30 μg), Ciprofloxacin (CIP 5 μg), Ceftriaxone (CRO 30 μg), Ceftazidime (CAZ 30 μg), Kanamycin (K 30 μg), Streptomycin (S 10 μg), Gentamicin (CN 10 μg), Nalidixic acid (NA 30 μg) and Chloramphenicol (C 30 μg). The concentrations of the antimicrobial disks were selected based on the internationally recognized standards and guidelines on antimicrobial routine testing and reporting on enterobacteriaceae provided by the Clinical and Laboratory Standard Institute. E. coli strain 25922 was used to control for bacterial growth and potency of antibiotic disks. Inoculated agar plates were incubated at 37°C for 24 h. The susceptibility zones were also measured and interpreted according to criteria set by the Clinical and Laboratory Standards Institute [24].

Bacterial DNA extraction and detection of virulence genes in E. coli

DNA from the isolated E. coli as well as from 5 control strains was extracted following the procedures described by Ehrt and Schnappinger [25], with slight modifications. A loop full of overnight bacterial culture was suspended in 1 ml of sterile distilled water and then boiled for 10 minutes at 95°C. The cell mixture was centrifuged for 5 minutes at 14,000 rpm and supernatant was used as the DNA template for PCR amplification. Ten PCR primers were used to detect the target genes enumerating toxins in pathogenic E. coli[26]. A multiplex PCR system was optimized by the progressive incorporation of primers corresponding to the different genes and several combinations of melting temperatures, primer concentrations and DNA template concentration. PuReTaq ready to go PCR beads (GE Health care) were used for PCR. The total reaction volume in each PCR tube was 25 μl containing 3 μl template DNA, 2 μl primer (0.2 μl of each primer) and 20 μ sterile PCR water. PCR was done under the following conditions; initial denaturation at 94°C for 5 minutes. This was followed by 35 cycles of denaturation at 94°C for 1 minute, annealing at 56°C for 30 seconds and extension at 72°C for 1 minute. The final step was amplification at 72°C for 10 minutes. Amplicons were then subjected to agarose gel electrophoresis and viewed under UV light.

Statistical analysis

Statistical analysis including means, medians and standard deviations, was conducted by means of SPSS version 17.0 software. To determine differences in microbial counts in chicken samples purchased from among the various retail outlets, ANOVA was used after converting the counts to base 10 logarithms. The chi-square test was used to assess any statistical significant association between antimicrobial resistances in the E. coli isolated with regard to retail outlet classification from where the raw chicken was purchased.

Contamination rate and microbial count in raw retail chicken

Contamination rate of chicken samples by total coliform bacteria was found to be 97% while contamination by E. coli was 78%. The average E. coli and coliform counts for all samples in general were observed to be above the acceptable range for E. coli counts as set by the Hazard Analysis and Critical Control Point system (HACCP), developed by Food and Agriculture Organization (FAO) and adopted by the Codex Alimentarius Commission (3.911 and 5.0261Log₁₀cfu/ml respectively). According to this system, the acceptable food safety range is 100 cfu/ml or less (2.000 Log₁₀cfu/ml or less), the marginal or intermediate range is over 100 cfu/ml but not above 1000 cfu/ml (over 2.000 but not above 3.000 Log₁₀cfu/ml) and the unacceptable range is above 1000 cfu/ml (above 3.000 Log₁₀cfu/ml) [27]. Only 80 (40%) out of a total 200 retail chicken samples that were purchased fell under the acceptable range for E. coli counts. 76% of samples fell under the unacceptable range for total coliforms. As shown in Tables 1 and 2 supermarkets had the highest percentage of samples within acceptable food safety ranges while average counts of samples from the rest of the retail outlets were within marginal or unacceptable counts of above 2.000 Log₁₀cfu/ml. There was a statistical significant difference in the E. coli count (F (4,175) =16.676; MSE = 106.576; p < 0.001) and coliform counts (F (4,179) =18.37; MSE = 105.097; p < 0.001) among the outlets with supermarkets having the lowest E. coli and coliform count compared to the rest. Graphical evidence of the differences of E. coli and coliform counts observed in the different retail outlet classifications are also shown in Figures 1 and 2.

		Total count log10CFU/ml of carcass rinse
Bacteria	Summary statistics	Supermarkets	H.I.A butcheries	H.M.I butcheries	L.M.I butcheries	L.I.A butcheries
E.coli	Average log10	0.9376	2.6510	5.0007	4.8084	4.9882
	Median	0.9031	1.7243	6.0000	5.9031	6.0792
	SD	0.6794	2.3638	2.8155	2.5910	3.1096
	Minimum	0.0000	0.0000	0.0000	1.0410	0.0000
	Maximum	2.1460	8.0000	8.0790	8.0000	8.0000
Coliforms	Average log10	1.7239	5.0769	6.1855	5.2953	6.0111
	Median	1.5315	6.0000	7.5374	6.0000	7.0731
	SD	1.3324	2.9304	2.3052	2.5475	2.6770
	Minimum	0.6990	0.9030	0.3420	1.1760	0.4770
	Maximum	8.0000	8.0790	8.1610	8.0000	8.1760

H.I.A butcheries- High income area butcheries, H.M.I butcheries- High-middle income area butcheries.

L.M.I- Low-middle income area butcheries, L.I.A- Low income area butcheries.

Assessment of microbial counts Log10cfu/ml
		% acceptable samples		% marginal samples		% unacceptable samples
Retail outlet	E.coli	Coliforms	E.coli	Coliforms	E.coli	Coliforms
Supermarket	97%	84%	3%	13%	0%	3%
H.I.A butcheries	69%	26%	10%	15%	21%	59%
H.M.I butcheries	29%	7%	12%	14%	60%	79%
L.M.I butcheries	25%	10%	10%	20%	65%	70%
L.I.A butcheries	16%	33%	6%	6%	78%	61%

Prevalence of antimicrobial resistant E. coli isolated from raw retail chicken

Prevalence of E. coli isolates harboring virulence markers