High Prevalence of Fecal Carriage of Extended Spectrum β-Lactamase-Producing Enterobacteriaceae Among Newly HIV-Diagnosed Adults in a Community Setting in Tanzania

Colonization in HIV-infected populations with extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBL-PE) is particularly worrisome in low-income settings. This study describes the prevalence of ESBL-PE carriage and associated risk factors among newly HIV-diagnosed adults in a community setting in Tanzania. A total of 595 newly diagnosed HIV adults with a median age of 35 years with interquartile range (IQR) 29–42 years and a median CD4 count of 492 cells/μL (IQR 390–666 cells/μL) were recruited. Among these, 194/595 (32.6%, 95% confidence interval [CI] 28.9–36.6) were ESBL-PE carriers. Participants with low CD4 count (<350 cells/μL) had significantly higher prevalence of ESBL-PE carriage compared with those with CD4 count ≥350 cells/μL (26/58, 44.8%, vs. 168/537, 31.3%, p = 0.04). Antibiotic use in last 4 weeks (odds ratio [OR] 1.55, 95% CI 1.08–2.22, p = 0.02) and CD4 count ≥350 cells/μL (OR 1.78, 95% CI 1.03–3.09, p = 0.04) were independent risk factors for fecal carriage of ESBL-PE. In total, 244 isolates of ESBL-PE were isolated from 194 participants. Of these, 238/244 (97.5%) harbored blaCTX-M genes, with blaCTX-M-15 being predominant (219/238 (92%), followed by blaCTX-M-27 (9/238 (3.8%), blaCTX-M-14 (8/238 (3.4%), blaCTX-M-55 (1/238), and blaCTX-M 211/3 (1/238). blaSHV-2a genes were detected in four isolates, whereas the blaSHV-12 gene was detected in one isolate. Phenotypic carbapenemase-producing Enterobacteriaceae was detected in one HIV-positive person with CD4 count 132 cells/μL. In conclusion prevalence of ESBL-PE carriage is high among newly diagnosed HIV adults in Dar es Salaam, and is significantly associated antibiotic use and low CD4 count.


Introduction
I ndividuals living with HIV are at risk not only of classical HIV-related opportunistic infections such as pneumocystis pneumonia and tuberculosis, but also severe infections caused by common bacterial pathogens such as Escherichia coli, salmonella, pneumococci, and staphylococci. [1][2][3] The World Health Organization (WHO) recognizes antimicrobial resistance (AMR) in such bacterial infections as a major threat to global health. Extended-spectrum blactamase-producing Enterobacteriaceae (ESBL-PE) constitute a particular challenge, as almost all are multidrug resistant. Previously, infections due to ESBL producers were mainly a health care-associated problem. 4,5 However, ESBL-PE infections are now increasingly acquired in the community. 6,7 In resource-constrained settings, ESBL-PE infections are spread more easily due to deficient infection control, whereas expensive reserve antibiotics such as carbapenems and colistin are inaccessible. Consequently, infections caused by ESBL-PE carries high mortality exceeding mortality outcome of bacterial infections in the preantibiotic era. 5,8 Colonization with ESBL-PE frequently precedes invasive diseases. 9,10 Fecal carriage of ESBL-PE could pose greater risk in people living with HIV due to the increased risk of severe bacterial infections. 1,2 In Tanzania, studies have documented high frequency of infections caused by ESBL-PE in hospitalized patients. 5,[11][12][13] ESBL-PE carriage is also common in hospital settings, and carriage rates of 50% have been documented in hospitalized children in Tanzania. 14 Even in community settings, studies have documented ESBL-PE carriage rates of 11-32% in healthy children 15 and 17% in healthy adults 16 in the same country. Although high rates (89.7%) of ESBL-PE carriage have been reported in hospitalized HIV positive children in Tanzania, 14 others documented lower rates (14%) in HIV-infected children in the community in Zimbabwe. 17 To date, no community-based studies from Africa have reported frequencies and risk factors for ESBL-PE carriage among newly diagnosed HIV adults.
HIV is pandemic in sub-Saharan Africa, and Tanzania has an estimated 1.4 million people aged between 15 and 64 years living with HIV. 18 We designed this study aimed to determine prevalence of fecal carriage of ESBL producers in HIV-infected people in the community; and assess any influence of antibiotic use, hospitalization, and CD4 count on fecal carriage of ESBL producers. The findings may help develop strategies for reducing the burden of AMR in a population at high risk of severe infections.

Study sites and study participants
Newly diagnosed HIV-infected adults with CD4 count of ‡350 cells/mL who had not yet started antiretroviral treatment (ART) were recruited consecutively as part of a double-blinded randomized clinical trial (RCT), CoTrimResist (Clinical-Trials.gov identifier: NCT03087890) between April 2017 and May 2018. The study was performed in Dar es Salaam, Tanzania, the largest city and economic capital of Tanzania. Participants were residents of the main municipals of Dar es Salaam city, namely Kinondoni, Temeke, Ilala, Ubungo, and Kigamboni, and were recruited through six HIV care and treatment clinics at Mwananyamala Hospital, Amana Hospital, Temeke Hospital, Pastoral Activities and Services for People with AIDS Dar es Salaam Archdiocese (PASADA), Mbagala Hospital, and Mnazimmoja Hospital. In addition, newly diagnosed HIV-infected adults with CD4 counts of <350 cells/mL (n = 58) were recruited from the same Hospitals during screening for eligibility to the RCT. This article reports on the baseline data before interventions and treatment.

Data collection
Health workers recorded participants' data such as demographics and clinical parameters on tablet computers using standardized electronic case report forms employing the RedCAP system (Research Electronic Data Capture, Vanderbilt University, Nashville, TN), which automatically synchronizes study data to a central data server.

Screening for ESBLand carbapenemase-producing Enterobacteriaceae
From each participant, a rectal swab was collected and transported in liquid Cary-Blair medium (Fecal Transwab, MWE Co Bath Ltd., Corsham, United Kingdom) in a cool box and maintained at 4°C with icepack. Rectal swabs were inoculated into brain heart infusion (BHI) broth and incubated overnight at 37°C. Two drops (0.1 mL) from BHI broth were subcultured on CHROMID ESBL (BioMérieux, Marcy l'Etoile, France) for screening of ESBL-PE, 19 and CHROMID CARBA SMART (BioMérieux) for combined screening for carbapenemase-producing Enterobacteriaceae (CPE) and OXA-48-producing Enterobacteriaceae. The plates were incubated at 37°C for 24 hours and bacterial growth interpreted according to manufacturer's instructions. Molecular characterized strains harboring different combinations of b-lactamases were used for quality control as previously described. 14 Bacterial isolates were identified by MALDI-TOF MS using the Microflex LT instrument and MALDI Biotyper 3.1 software (Bruker Daltonics, Bremen, Germany).
PCR and sequencing for detection and identification of ESBL genes DNA was extracted by a rapid boiling procedure and stored at -20°C. Real-time PCR was used for detection of cefotaxime-Munich (CTX-M) encoding genes using a LightCycler 480 Instrument II (Roche Diagnostics, Mannheim, Germany). Forward primer CTXM-F 5¢-ATGTGCAGYACCAGTAAR GT-3¢, and reverse primers CTXM-R1 5¢-TGGGTGAAGTAA GTGACCAGA-3¢ and CTXM-R2 5¢-TGGGTAAARTAGG TCACCAGA-3¢ (TIB Molbiol, Berlin, Germany), which target a 595 bp internal region present in all the five different CTX-M groups, were used as previously described. 14 Each run included positive and negative controls in duplicates. The PCR products were sequenced using the same reverse primers as for PCR. Sequencing was done using BigDye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA). Sequences were analyzed using the Basic Local Alignment Search Tool program available at the website of the National Center for Biotechnology Information.
For confirmation of ESBL producers, isolates negative for CTX-M PCR were tested for the presence of bla SHV genes as previously described by Kommedal et al. 20

Statistical analysis
We used chi-square test to assess the proportions of fecal carriage of ESBL producers by patient characteristics. We analyzed the association between risk factors of interest (antibiotic use, prior hospitalization, and low CD4 counts) and fecal carriage of ESBL producers, adjusting for age, gender, education status, district of residence, and study sites (individual study sites). For each main exposure variable, we assessed all potential confounders using both logistic regression and Mantel-Haenszel method, defining a confounder as a factor that changes the effect size by ‡10%. Effect modification of factor between different groups were examined by Mantel-Haenszel method. A significance level of 0.05 was used and all p-values refer to two-sided tests. Statistical analysis was performed using STATA version 16 (College Station, TX).

Ethical approval and consent to participate
Ethical approval to conduct the study in Tanzania  Community development, Gender, Elderly, and Children. The clinical trial was also registered by Tanzania Food and Drugs Authority (TZ16CT007) and Clinical trial.org (NCT03087890). In Norway, the study was approved by the Regional Committee for Medical and Health Research Ethics of Western Norway (REK2015/540). Informed written consent was obtained from each of the study participants.

Description of study participants
We recruited a total of 595 newly HIV-diagnosed adults with a median age of 35 years with interquartile range (

Predictors of ESBL producers
Among the main risk factors of interest, only antibiotic use during the last 4 weeks (odds ratio [OR] 1.55, 95% CI 1.08-2.22, p = 0.02) and CD4 count <350 cells/mL (OR 1.78, 95% CI 1.03-3.09, p = 0.04) were independently associated with fecal carriage of ESBL producers (Table 3). Prior hospitalization during the last 1 year was not a significant risk factor (OR 0.79, 95% CI 0.44-1.40, p = 0.41). The effect size of each risk factors did not change significantly (>10%), whereas adjusting for the other main risk factors and potential confounders and effect modifiers, that is, age, gender, education status, district of residence, and study sites (individual study site). Thus, we did not find any significant confounders or effect modifiers.

Discussion
This first community-based study reports prevalence of fecal carriage of ESBL-PE in almost one third (194/595,

32.6%) of individuals newly diagnosed with HIV in Dar es
Salaam. This rate of ESBL-PE fecal carriage is much higher than found among adults with unknown HIV status in community-based studies in Northern Tanzania (55/334, 16.5%, p < 0.001), 16 Burkina Faso (22/101, 21.8%, p = 0.03), 21 Gambia (28/565, 5.0%, p < 0.001), 22 and Morocco (4/93, 4.3%, p < 0.001). 23 Although we are not aware of other communitybased studies from HIV-infected adults in Africa, there was significantly lower ESBL-PE carriage rate (24/175 13.7%, p < 0.001) in a community-based studies of HIV-infected children in Zimbabwe. 17 The high carriage rate observed in our study could be attributed to several factors, including the choice of screening technique employed. Our use of broth enrichment of rectal swabs may have increased the yield in detecting ESBL-PE colonization. 24,25 In contrast, our use of rectal swabs may have inferior sensitivity compared with stool culture used in other studies. 19 We observed that antibiotic use during the last 4 weeks was a strong risk factor for fecal carriage of ESBL producers in this cohort. In Tanzania, almost all antibiotics can be accessed over the counter. Overuse of antibiotics and misuse of broadspectrum antibiotics, such as cephalosporins for trivial infections, are known to contribute to increasing AMR, including ESBL-PE. In this setting, people with an undiagnosed HIV infection are likely to opt for self-medication with antibiotics for intercurrent infections and symptoms related to an, as yet, unrecognized immunodeficiency. Such overuse of antibiotics may partially explain the high prevalence of ESBL-PE carriage in our study. This study confirms our hypothesis that antibiotic use affects carriage of resistant bacteria in HIV populations, and our finding is consistent with earlier observations from community and hospital settings in Africa. 14,16,21,22 Previous hospitalization is a well-known risk factor for fecal carriage of ESBL producers, 12,21 contributing both to increased likelihood of antibiotics uses and acquisition of ESBL producers in gastrointestinal tract. However, this study could not add support this observation. Tanzania's current implementation of the ''test and treat'' strategy for HIV is expected to result in more people getting diagnosed before actually experiencing symptomatic disease. Despite this, it is  EXTENDED SPECTRUM b-LACTAMASE HIV ADULTS TANZANIA plausible that many patients will have consulted health care services, without actually having been admitted, for unknown HIV-related ailments before getting diagnosed with HIV, and that this may have led to increased risks of gastrointestinal colonization with ESBL-PE. This study found low CD4 count (<350 cells/mL) was an independent risk factor for fecal carriage of ESBL. Low immunity predisposes to risk of opportunistic infections increasing the likelihood of antimicrobial use and hospitalization. However, adjustment for both antibiotic use and hospitalization did not alter the independent association between low CD count and fecal carriage of ESBL producers. Furthermore, Wilmore et al. 17 found no association between low CD count and fecal carriage of ESBL producers in children in Zimbabwe. 17 The differences could be explained by inclusion criteria, this study recruited newly diagnosed HIV-infected adults not yet on ART, whereas Wilmore et al. enrolled children who were stable on ART. 17 Our study confirmed bla CTX-M 15 as the predominant ESBL gene type accounting 92% of all isolates, supporting other reports on the extraordinary dissemination of the bla CTX-M-15 genotype worldwide. In the first ever report of ESBL-PE from Tanzania, 5,13 we showed that bla CTX-M-15 , alongside TEM-63, was the dominant cause of ESBL-PE bloodstream infections in Tanzania as early as 2001-2002. Later on, we found predominance of bla CTX-M-15 in clinical isolates from community-acquired urinary tract infections from 2004 (Ref. 7 ). Reports from elsewhere 26,27 indicate that when bla CTX-M-15 genotype penetrates in a new landscape where other ESBL genes persist, bla CTX-M-15 tend to displace other genes and becomes predominant. All recent studies from community and hospital settings in Tanzania have supported the predominance of bla CTX-M 15 genes among Enterobacteriaceae isolated from feces. 12,15,16 The CTX-M group 9, bla CTX-M-27 , and bla CTX-14 were also documented in this study and were only isolated from E. coli. This correlates with a recent study among HIV-infected children in Zimbabwe. 17 Our previous study on uropathogenic isolates did not detect any bla CTX-M group 9 genes, 7 whereas a pediatric study reported only two E. coli isolates carrying bla CTX-M-14 genes and none with bla CTX-M-27 . 14 In conclusion, the study found a high prevalence of fecal ESBL-PE carriage among people newly diagnosed with HIV. Antibiotic use during the last 4 weeks and low CD4 count were risk factors for fecal carriage of ESBL producers. Further studies should investigate transmission dynamics of ESBL-PE in the community, complex factors driving the emergence of multidrug-resistant bacteria, and the relationship between ESBL-PE carriage and invasive disease. Social behavior and policies on AMR need to be crafted and regulation on antimicrobial use must be enforced.