Socio-demographic characteristics of study participants colonized by Enterobacteriaceae
In the current study, females were the more infected with Enterobacteriaceae at 86.36% implicated in Urinary Tract Infections. This can be attributed to physiological and anatomical differences [33] compared to males. In addition males have poor health seeking behaviour compared to female counterparts and as such they are less likely to visit health facilities even when they are sick. This percentage is higher compared to 58% previously reported in Sri Lanka [34] and another related study conducted in Spain which reported bacterial isolation rates of 57% among females.
A related study conducted in Pakistan revealed a lower prevalence of Urinary Tract Infections among females at 72% [35] The most affected age category was 15-29 followed by 30-44 years at 37.50% and 30.68% respectively . This can be attributed to the fact that it is the most sexually active age hence higher chances of contracting Urinary Tract Infections[35]. However the association between age and infection was not statistically significant in the current study.
It was reported that most participants infected with Enterobacteriaceae possessed livestock at 70.45%. This is because Mbarara is found in south-western region of Uganda where most occupants are agro-pastoralists.
Mbarara Regional Referral Hospital yielded the highest number of participants infected with Enterobacteriaceae at 46.59% followed by Bwizibwera HCIV and then Rubaya HCIII. This is because MRRH is a high volume facility compared to Bwizibwera HCIV and Rubaya HCIII and the fact that it is a regional referral hospital makes I attract more patients seeking for quality services in addition to having complicated cases.
Bacterial isolation trends
The most isolated organism from urine was Escherichia coli at 70.45% followed by Klebsiella spp, proteus mirabilis and Enterobacter aerogenes. The findings are consistent with other studies indicating that Escherichia coli is the commonest bacteria implicated in community acquired UTIs [36] and other bacterial infections [37], [38]. However the isolation rate of Escherichia coli from this study was slightly lower compared to the common isolation rates which are reported between 75-90% [39]
A related study in Uganda demonstrated that Escherichia coli was the most isolated organism among outpatients at 89% which is higher than the isolation rate reported in our study. [7].
It has been reported previously that Escherichia coli and Klebsiella spp are the major ESBL producing Enterobacteriaceae isolated from both community and hospital settings [40] [41, 42] and [15] which is in agreement with results from this study. A related study in Pakistan demonstrated 33.5% prevalence of ESBL among Escherichia coli and 15.25% among Klebsiella spp.
Proportion of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae
The current study reported the prevalence of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae at 23.86% demonstrating an increase in community spread of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae in Uganda as compared to a previous study by [7].
The results from this study demonstrated higher prevalence of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae from the community compared to 8.5% prevalence reported in Nigeria [43].
A related study in Morocco demonstrated a lower prevalence of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae at 7.5% [35] compared to our study. This can be attributed to difference in climatic conditions, socio-economic factors and regulations regarding antibiotic use which act as major sources of selection pressure for production of ESBLs.
Related studies have reported higher prevalence of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae compared to this study. For instance a study carried out in Iran demonstrated 52% prevalence of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae from community acquired Urinary Tract Infections [44]
A related study in Canada demonstrated prevalence of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae from the community at 71% [15] while another study in Nepal reported 35.9% prevalence [38].
A related study in Nigeria demonstrated prevalence of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae from community settings at 35% [45]. Related studied in India and Turkey demonstrated higher prevalence compared to this study at 40% [46] and 37.1% [40] respectively. The differences in prevalence can be attributed to differences in climatic conditions, health seeking behaviour, socio-economic factors, antibiotic availability and use which either singly or in combination contributes to selection pressure for ESBLs production. In addition, countries with highest antibiotic consumption rates have been found to have high prevalence of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae for instance Turkey, Tunisia, Algeria, Greece and Romania [47], [48], and [49].
It has also been noted that the spread of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae is attributed to the existence of reservoirs in the community especially by recently discharged patients [50]. Other potential reservoirs have been identified for instance a study in Tanzania demonstrated existence of ESBL producing Enterobacteriaceae in public latrines [51]. Other studies have demonstrated faecal carriage of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae among humans which potentiates the spread of such organisms [29, 52] [8] [53] and [54].
In addition it has been of great concern that ESBL producing Enterobacteriaceae are transmitted from animals to humans leading to cross infestation [55] and [56], [57] and [58].
In addition, Extended Spectrum Beta-Lactamase producing Enterobacteriaceae have recently been isolated in food products which is a potential source to humans especially when these products are consumed in raw form [59-61], [62] and [24]
Risk factors to acquisition of infection by Extended Spectrum Beta-Lactamase producing Enterobacteriaceae
Among the risk factors assessed, female gender was associated to increased risk of infection by Extended Spectrum Beta-Lactamase producing Enterobacteriaceae with P-value <0.005. This observation is similar to a study conducted in Canada which demonstrated that female gender was associated with infection by Extended Spectrum Beta-Lactamase producing Enterobacteriaceae [11]. Other related studies have demonstrated an association between female gender with infection with Extended Spectrum Beta-Lactamase producing Enterobacteriaceae [63], [27, 64]
Contrary to this study, related studies in Uganda [7] and India [46] demonstrated that there was no single risk factor associated with Extended Spectrum Beta-Lactamase production by Enterobacteriaceae.
Other factors assessed were not statistically significant to infection with Extended Spectrum Beta-Lactamase producing Enterobacteriaceae which included age category, livestock possession and health facility with P-values 0.680, 0.663, 0.620 respectively.
Other related studies worldwide have reported a number of factors related to infection by Extended Spectrum Beta-Lactamase producing Enterobacteriaceae for instance a study in United States of America demonstrated that prior exposure to broad spectrum antibiotics, recent hospitalization and surgery were potential risk factors to infection with Extended Spectrum Beta-Lactamase Producing Enterobacteriaceae [65].
A similar study conducted in Nigeria demonstrated that recent surgery, hospitalization, recent antibiotic use were associated with colonization and infection with Extended Spectrum Beta-Lactamase producing Enterobacteriaceae [43]. A related study in Australia reported that recent fluoroquinolone use was related to production of ESBLs among Enterobacteriaceae due to accidental expression of genes because ESBL genes are located on the same plasmid with those genes responsible for quinolone resistance [14] and a related study demonstrated that ESBL production was directly related to fluoroquinolone use [66].
Antibiotic susceptibility patterns of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae
A total of 13 antibiotics selected from 7 different classes were tested on ESBL producing Enterobacteriaceae. Extended Spectrum Beta-Lactamase producing Enterobacteriaceae were most resistant to Ampicillin, Cefepime, Aztreonam, Nalidixic acid, Ciprofloxacin and Amoxicillin/clavulanic acid. Several studies worldwide have demonstrated comparable findings. For instance a review in United States of America pointed out that fluoroquinolones were no longer consistently reliable in treatment of bacterial infections [67].
A related study conducted in Nigeria demonstrated consistent results to this study where it was shown that Extended Spectrum Beta-Lactamase producing Enterobacteriaceae were resistant to quinolones, penicillins and beta-lactam combinations [45], [15] [68], [34]. A related study in India demonstrated high resistance rates to flouroquinolones and beta-lactam agents [46]. A related study in Japan demonstrated similar findings to this study where it was revealed that Cefepime was ineffective at treatment of infections caused by ESBL producing Enterobacteriaceae [69]
The findings from this study differ from other studies done worldwide for instance studies which have recommended the use of Cefepime to treat complicated bacterial infections caused by ESBL producing Enterobacteriaceae [70], [71], [72] and [73].
This study revealed high sensitivity rates of ESBL producing Enterobacteriaceae to Imipenem, Meropenem, Amikacin, Gentamycin, Piperacillin/tazobactam, Cefoxitin and Nitrofurantoin. The findings from this study are in agreement with several studies conducted worldwide for instance a study conducted in Nepal revealed that all isolated tested were sensitive to Imipenem [38] and a similar study conducted in Nigeria demonstrated 97.14% sensitivity to Imipenem [43] and thus carbapenems have been recommended as drug of choice to treat infections caused by ESBL producing Enterobacteriaceae [74] and [75] although carbapenems have been linked to increased risk of fungal infections [76]. In addition there have been concerns of emergence of carbapenem resistance among Enterobacteriaceae which makes treatment of complicated bacterial infections more difficult [11], [36] and [77]
A related study in Iran demonstrated similar findings to this study where high sensitivity rates were reported to Gentamycin [78]. A related study in India demonstrated similar findings to this study where high sensitivity rates were reported to Imipenem, Amikacin and Nitrofurantoin [79].
Related studies have tested and recommended Beta-lactam combinations such as Piperacillin/tazobactam for the treatment of infections caused by Extended Spectrum Beta-Lactamase producing Enterobacteriaceae [80], [81], [82] and [83].
The findings from this study are contrary to a related study conducted in Iran which reported high resistance rates to Nitrofurantoin and amikacin [78]
Related studies in Spain and South Korea recommended the use of Piperacillin/tazobactam as an alternative to carbapenems for the treatment of complicated infections caused by ESBL producing Enterobacteriaceae [84] and [85] although it should be used in stable conditions and where antibiotic susceptibility profiles have been carried out [86] because it has been associated to high mortality rates in some studies [17].
Multi- drug resistance (MDR) among ESBL producing Enterobacteriaceae from this study was very high and established at 85.71%. This is in agreement to other studies conducted worldwide for instance a study carried out in Morocco established multi-drug resistance among ESBL producing Enterobacteriaceae at 91.1% [35] while a related study in India established multi-drug resistance at 82.6% [79]. These observations are consistent to the previously reported co-resistance to other antibiotics [87]. This has been attributed to the fact that genes responsible for ESBL production are encoded on big plasmids which contain other resistance genes to other antibiotics [75].
The high resistance rates observed in this study can be attributed to overuse and misuse of antibiotics to treat common infections [88]. It can also be attributed to socio-economic differences where patients lack consultation fees on which treatment to consider in addition to prior knowledge of which antibiotics to use hence resorting to self-medication [89]. Thus it has been reported that countries with the most reserved antibiotic prescribing patterns have relatively low rates of resistance [90].
Genotyping of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae
The genes responsible for production of Extended Spectrum Beta-Lactamases among Enterobecteriaceae were determined using Polymerase Chain Reaction (PCR) where 90.48% of the isolates harboured the genes CTM-M, TEM and SHV which is in agreement with what has been reported from other studies that these genes are the most prevalent ESBL genes in both community and hospital settings [17]
This study demonstrated that an organism can harbour more than one ESBL gene which is consistent with other studies carried out previously for instance a related study carried out in Western Uganda demonstrated that 40.8% of the organisms harboured more than one gene [32]. A related study in Morocco also demonstrated that an organism can harbour more than ESBL gene [35].
The gene CTX-M was the most prevalent at 46.7%, followed by TEM at 30% and SHV at 23.3%. The findings from this study are consistent with other studies conducted worldwide. For instance a study conducted in Libya demonstrated that CTX-M was the commonest isolate [91] and similarly higher rates of CTX-M have been found in Europe [92], [17] and in other parts of Africa [93] and [91]. The high prevalence of CTX-M in this study can be attributed to the widespread use of ceftriaxone to treat serious bacterial infections [94]
The prevalence of TEM and SHV was found to be lower than that of CTX-M because TEM and SHV are more prevalent in hospital settings compared to community settings [95]. Different rates of SHV and TEM have been reported worldwide for instance a study conducted in Sri Lanka indicated higher prevalence of TEM compared to SHV [34] and a related study conducted in Western Uganda demonstrated higher prevalence of TEM compared to SHV [32]. Some studies have reported equal prevalence of TEM and SHV for instance a related study in Spain demonstrated 18% proportions for both SHV and TEM genes [17]. A related study conducted in Iran demonstrated contrary findings to this study where higher prevalence of SHV compared to TEM were reported [44].
This study also demonstrated that ESBL genes in community settings are similar to those in hospital care settings irrespective of the proportions which suggests possible importation of ESBL producing Enterobacteriaceae from the community into the hospital and vice versa for instance a related study in Israel involved screening of patients for ESBL at admission and it was discovered that 10.8% of the patients were colonized by ESBL producing Enterobacteriaceae [8]
The findings from this study clarified that there are other types of ESBL genes that were not detected using the primers specific for CTX-M, TEM and SHV hence a discrepancy between phenotypic and genotypic methods for ESBL detection.
It was demonstrated that 2(9.52%) Enterobacteriaceae did not harbour any of the 3 genes under detection. This can be attributed to the fact that new ESBL genes are emerging as a result of mutations. The findings from this study are consistent with other studies conducted worldwide for instance a study conducted in Iran demonstrated that 8% of the organisms that were ESBL positive at phenotypic detection contained VEB gene which would have been missed out if only 3 primers for CTX-M, TEM and SHV were used [44].
A related study in Israel demonstrated phenotypically ESBL positive isolates that lacked common ESBL genes at genotyping [8]. Similarly, a related study conducted in Sudan indicated that 47.7% of phenotypically ESBL positive isolates did not harbour common ESBL genes [91]. A related study in India demonstrated that not all phenotypically positive ESBL isolates harbour the three common genes as 61.6% of the isolates did not contain CTX-M, TEM or SHV genes [96]. Several other studies have demonstrated that there are other ESBL genes that need to be detected as far as genotyping for ESBL is concerned [78] and [32].
This study demonstrated that all genotypically positive Enterobacteriaceae were phenotypically positive which is contrary to other studies for instance a study carried out in Western Uganda reported that 37% of the isolates contained ESBL genes but were negative on phenotypic assessment [97]. This can be attributed to lack of effective selection pressure to trigger gene expression [75].
The high prevalence of ESBL genes in community settings can be attributed to the fact that there is horizontal transfer of ESBL genes between Enterobacteriaceae existing in the environment. This is because ESBL genes are located on mobile genetic elements which favour their dissemination between Enterobacteriaceae. For instance a related study in Turkey demonstrated that ESBL genes were present in Enterobacteriaceae isolated from faecal samples and this potentiates spread of genes to other isolates colonizing humans [52].
A related study in Korea demonstrated the potential spread of ESBL producing Enterobacteriaceae and exchange of ESBL genes from farm animals and farm environment to humans [98]. Since this study was carried out in an agro-pastoral community, such spread cannot be ruled out.
Conclusions
The prevalence of Extended Spectrum Beta-Lactamase producing Enterobacteriaceae from our study is high whileThe most isolated bacterial species are Escherichia coli and Klebsiella spp. Carbapenems are effective against Extended Spectrum Beta-Lactamase producing Enterobacteriaceae while Beta Lactam/ Beta Lactamase inhibitor combination antibiotics are second in effectiveness to Carbapenems. The rate of Multi Drug Resistance (MDR) to common antibiotics used to treat Urinary Tract Infections is high.The most prevalent genes responsible for ESBL production are CTX-M and TEM
Recommendations
Double disc synergy method should be implemented in routine confirmation of ESBL producing Enterobacteriaceae in health facilities in resource scarce settings. This would encourage definitive therapy instead of empirical therapy in management of bacterial infections.
More studies about antimicrobial resistance in community settings should be carried out. This will help to establish action areas and points where to reduce spread of resistance either from the community to the healthcare settings.
Infection control points should be established especially at the entrance of health facilities to reduce influx of ESBL producing Enterobacteriaceae from the community. Likewise personal hygiene like hand washing, cleanliness should be emphasized in health facilities and community to reduce spread of ESBL producing Enterobacteriaceae.
Antimicrobial surveillance should be done both in both community and health care settings to be able to identify action areas to combat spread of antimicrobial resistance. Data obtained should direct policy makers for effective decision making.
The antibiotics that were identified as ineffective for instance Ampicillin, Amoxicillin/clavulanic acid, Cefepime, Nalidixic acid, Ciprofloxacin and Aztreonam should be rested in order to reduce the selection pressure against them by the bacteria. In future, these can be reintroduced as guided by data when they have regained their potency. Piperacillin/tazobactam should be emphasized for treatment of infections caused by ESBL producing Enterobacteriaceae as an alternative to carbapenems to reduce the risk of development of carbapenem resistance among Enterobacteriaceae.
Study limitations
The study was limited to three health facilities in Mbarara district. This may not reflect the real picture in all facilities in South western Uganda which makes generalization of findings difficult. Only Enterobacteriaceae isolated from outpatients were considered hence the results cannot be generalized to hospitalized patients. Data about history of recent antibiotic use was not collected hence making it difficult to determine association as far as resistance development is considered. The study was limited to human beings and no comparative study was carried out in livestock to establish existence of ESBL.