| dc.description.abstract | Background: Extended spectrum β-lactamase producing Enterobacteriaceae (ESBL-PE) have been identified as a major cause of both community and facility associated infections and are closely linked to faecal carriage in the infected individuals. Faecal carriage of ESBL-PE has been reported from several countries with varied prevalence and related characteristics. This study sought to estimate the prevalence of carriage of ESBL-PE, to determine the factors associated with such carriage and to identify the type of ESBLs, their nature of spread and co-linked non-βlactam resistance in an urban district, Kampala and two rural districts of Kayunga and Mpigi. The objectives of this study were: 1. To estimate the prevalence of ESBL-producing Enterobacteriaceae in the flora of clients attending OPD clinics in Kampala, Kayunga, and Mpigi 2. To identify the factors associated with carriage of ESBL-producing Enterobacteriaceae in the flora of clients attending OPD clinics in Kampala, Kayunga, and Mpigi 3. To determine the type of ESBLs and the extent to which they have evolved in Uganda 4. To determine the nature of spread of ESBLs among Enterobacteriaceae in Ugandan communities 5. To determine non-β-lactam co-resistant genetic determinants among Enterobacteriaceae in Ugandan communities Methods: Using probability proportion to size cluster sampling, 30 clusters of participants per district (N=484 /district) were recruited. Informed consent was obtained, and a standardized questionnaire administered. Participants then provided a convenience sample of stool or urine. E.coli and K.pneumoniae were isolated from samples and were subjected to susceptibility testing using the CLSI Kirby Bauer method. ESBLs were detected using the double disc synergy test and Vitek II and factors associated with ESBL-PE carriage determined using two-by two tables, the chi square test and logistic regression models. The genes encoding ESBLs were detected by PCR followed by sequencing and homologous sequences from the gene bank were identified using the BLAST tool. The nucleotide sequences were translated into amino acids using the EXPASY tool and aligned with known sequences from the gene bank. The ESBL-PE were subjected to PCR based finger printing and multilocus sequencing typing (MLST) to determine
xx
relatedness and circulating clones. Plasmids from trans-conjugants were characterized by replicon typing and the population structure of the E.coli ESBL-PE was determined by phylogenetic typing. Resistance determinants encoded on plasmids from trans-conjugants and cassette arrays from the conserved region of integron class 1 were identified by PCR and sequencing.
Results: The samples collected from the 1448 participants varied by district with one district submitting 469 (92%) of the desired stool samples while clients at the other two sites preferred giving urine. Thus, the stools were 167(35%) and 94(20%). Accordingly, the overall yield of isolates was 985 from 636 (87%) stool samples; 620 E.coli and 16 Klebsiella pneumoniae and 349 urine samples; 310 E.coli and 39 Klebsiella pneumoniae. Susceptibility testing revealed low resistance to ceftriaxone at 3% but higher levels resistance to β lactam/β lactamase inhibitor combinations at 36% for amoxicillin clavulanate and 27% for piperacillin tazobactam, and cephamycins (cefoxitin) at 22%. In addition, carbapenem resistance was not detected but isolates with reduced zones that would necessitate screening for carbapenemase activity were noted at 17%, thus these findings are suggestive of multiple β lactam resistance mechanisms. The prevalence of ESBL-PE was high in the urban district but low and varied in rural districts. There was no significant difference in prevalence of ESBL-PE between the two rural districts, however the difference between both rural districts and the urban one was significant (p<0.001). It was not possible to establish independent risk factors for carriage of ESBLA, however association with health facilities was a significant contributor to carriage of isolates that were AmpC producers or had an MDR phenotype. The most predominant ESBL type was the CTX-M-15 and it was possible to transfer the phenotype by conjugation. It was not possible to complete alternate transfer methods neither was it possible to locate the blaCTX-M or determine the size of plasmids it was located on. It was not possible to get complete sequences in majority of reactions; this made it difficult to rule out other mutants of the CTX-M-3 members from the CTX-M-1 group, like CTX-M 55. The existence of two different ESBLA types occurred in only one isolate, however non ESBL- β lactamases like TEM-1, OXA-1, SHV-like β lactamases, were common. ESBLA coexisted with AmpC in several isolates. The spread of ESBL-PE in the three districts is not clonal in nature, and is mainly associated with plasmid spread, mainly of the incF;
xxi
FIA-FIB replicon types. It was difficult to amplify several of the house keeping genes from both E.coli and Klebsiella pneumoniae. The multidrug resistance phenotype associated with ESBL-PE was transferable in 45% of isolates. Resistance rates of 70% to cefotaxime, 75% to trimethoprim sulphamethoxazole, 45% to gentamicin, 25% to ciprofloxacin and nitrofurantoin and 20% to cefoxitin were observed in transconjugants. Furthermore resistance determinants against several of the commonly used antimicrobials such as (i) Sul1 and Sul2 of sulfamethoxazole resistance together with dhfra1,dhfra5 and dhfra7 of trimethoprim resistance (ii) aac for gentamicin and aac(6’)Ib-cr for amikacin tobramycin and quinolones (iii) qnrA,qnrB of quinolone resistance (iv) cat1 and CmlA of chloramphenicol resistance (v) tetA and TetB of tetracycline resistance and (vi) OXA-1 and TEM-1 of penicillin resistance were co-transferred with the blaCTX-M-15. Much as integron class 1 was common, the cassette arrays carried only two resistance determinants like aad that conferred resistance to streptomycin and dhfr that conferred resistance to trimethoprim. No sul1 or qac delta genes were detected thus the 3Cs‘ end of the conserved region was either not amplified by the primers used, could have experienced sequence rearrangement or may be missing from the integron structures in this setting.
Conclusion: ESBL-PE are present in the gastrointestinal and to a lesser extent in the urinary tracts of clients attending outpatient clinics. Individuals who have visited health facilities in the recent past are likely to carry multidrug resistant bacteria including ESBL-PE. The CTX-M-15 is the most predominant ESBL. Dissemination of ESBLs is mainly through horizontal gene transfer mainly by plasmids, and the blaCTX-M-15 is co transferred with other resistance determinants mainly of commonly used antimicrobial agents. There is epidemiological linkage between CTXM-15 enzymes in the urban and rural districts. | en_US |
