INTRODUCTION
Antimicrobial resistance is a major public health problem that is projected to result in 10 million deaths worldwide in the year 2050.1 Antibiotic-resistant Gram-negative bacteria are associated with significant morbidity and mortality and are among the most urgent threats.2 Infectious diseases remain the leading cause of death in children under the age of five in low-resource settings,3 and several studies have reported striking mortality rates in neonates and children with multidrug-resistant (MDR) Gram-negative bacterial infections, including Escherichia coli.4–7 Escherichia coli can also serve as a genetic reservoir for plasmid-mediated antibiotic resistance that can be transferred to other pathogens and, thus, represents an important index pathogen for understanding the epidemiology of antibiotic resistance.8,9
Although most surveillance of antibiotic resistance is performed using bloodstream isolates, the gastrointestinal tract is the primary reservoir as well as the source for most invasive E. coli infections, yet little is known about the dynamics of and risk factors for gastrointestinal colonization with resistant E. coli in these populations.10,11 Using archived E. coli isolates and stools from a birth cohort study in Haydom, Tanzania, we, therefore, sought to characterize the antibiotic susceptibility profile of E. coli in these settings and to identify risk factors for the acquisition and carriage of drug resistance.
MATERIALS AND METHODS
Clinical data.
Methods of the multisite Malnutrition and the Consequences for Child Health and Development Project (MAL-ED) birth cohort study have been reported previously.12,13 Briefly, healthy infants were enrolled from November 2009 to February 2012 and followed intensively through 2 years of age, including twice weekly home visits for surveillance of childhood illnesses, antibiotic administration, and breastfeeding practices, as well as monthly anthropometry and a biannual sociodemographic survey. Stools were obtained monthly until 24 months of age and every 6 months thereafter until the age of 5 years. This study was conducted at the Haydom site, Tanzania.14 We selected 100 children from those who had complete follow-up to the age of 2 years, defined as E. coli isolates archived from stool samples collected at 6, 12, 18, and 24 months of age. From these children, we tested pooled E. coli isolates from up to 10 stools per child, collected every 6 months starting at 6 months of age and continuing until the age of 5 years.
Antibiotic susceptibility testing.
For stools collected through 24 months of age, E. coli was cultured from fresh stool samples and archived in pools of five colonies. As per the original MAL-ED study protocol, all of the archived pooled E. coli colonies collected through 24 months of age were confirmed with indole testing. In stool samples obtained after the age of 24 months, whole stool was frozen and stored at −80°C, and lactose-fermenting colonies were selected, with indole testing performed to confirm E. coli on a random subset. For all stools, pooled E. coli colonies underwent antibiotic susceptibility testing for 18 antibiotics, including disc diffusion testing for cefotaxime/clavulanate, cefotaxime, ceftazidime/clavulanate, ceftazidime, aztreonam, cefepime, ceftriaxone, ciprofloxacin, trimethoprim/sulfamethoxazole, cefoxitin, ertapenem, ampicillin, ampicillin/sulbactam, amoxicillin/clavulanate, nalidixic acid, gentamicin, cefazolin (BD BBL Sensi-Disc, Franklin Lakes, NJ) and an Etest for azithromycin (Etest; bioMérieux, Marcy-l’Étoile, France). Specifically, a swab inoculated with pooled E. coli isolates or frozen stool was streaked on MacConkey agar and incubated at 37°C for 24–72 hours. Five lactose-fermenting colonies were inoculated into trypticase soy broth and incubated at 37°C for 2–6 hours. The suspension was adjusted with sterile water to achieve turbidity equivalent to a 0.5 McFarland standard using visual interpretation. Mueller–Hinton agar was then inoculated and antimicrobial discs placed on the plates. The Mueller–Hinton agar plates were incubated at 37°C for 16–18 hours. Zone diameters were measured in millimeters. Azithromycin minimum inhibitory concentration (MIC) was determined using an Etest. Extended-spectrum beta-lactamase (ESBL) testing was completed using two combination disc tests (cefotaxime and cefotaxime/clavulanate, as well as ceftazidime and ceftazidime/clavulanate) defined as a ≥ 5 mm increase in the zone size for either cephalosporin in the presence of clavulanate. Escherichia coli American Type Culture Collection strain 25922 was used as a quality control strain. Zone diameter interpretative criteria were referenced from Clinical and Laboratory Standards Institute documents15 and categorized into susceptible and resistant, with intermediate reclassified as resistant. The Clinical and Laboratory Standards Institute breakpoints to define azithromycin resistance have not been established for E. coli; we used an MIC cutoff of ≥ 32 μg/mL to be consistent with the published literature.16
To identify mechanisms associated with ESBL E. coli and ampicillin resistance, a limited set of isolates underwent molecular testing. Microbial DNA were extracted by suspending bacterial colonies in 200 µL Tris-EDTA buffer and incubating at 95°C for 15 minutes followed by centrifugation, with the supernatant used as the DNA template. Antimicrobial-resistant genes were detected using the Microbial DNA quantitative polymerase chain reaction (qPCR) Array (Qiagen, Inc., Valencia, CA) according to the manufacturer’s protocol. In addition, a qPCR assay for the TEM gene was performed to identify the most common mechanism of aminopenicillin resistance.17
Data analysis.
To identify risk factors for acquisition and carriage of drug-resistant E. coli, we created 10 antibiotic classes: ampicillin, cefazolin, aztreonam, ertapenem, azithromycin, gentamicin, cotrimoxazole, fluoroquinolones (combining nalidixic acid and ciprofloxacin), extended-spectrum cephalosporins (ceftriaxone, ceftazidime, cefotaxime, cefepime, and cefoxitin), and beta-lactam/beta-lactamase inhibitors (ampicillin/sulbactam and amoxicillin/clavulanate). We then calculated the number of antibiotic classes to which each E. coli pool was resistant. Multidrug resistance was defined as resistance to at least one drug in three or more antimicrobial categories.18
To assess for risk factors for carriage of drug-resistant E. coli, we used generalized estimating equations to fit a multivariate linear regression model. In addition to age and gender, risk factors were chosen based on biological plausibility and sufficient heterogeneity for the children included in the analysis. All analyses were performed using R 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria).
RESULTS
Of 262 enrolled in the cohort, 209 had complete follow-up to 24 months of age; of these, 137 had archived E. coli from 6, 12, 18, and 24 months of age. From these, 100 children were randomly selected for this study, from whom 928 stools were originally collected and 837 had antibiotic susceptibility testing for at least one antibiotic from each of the pre-defined antibiotic resistance classes. Most households did not have improved sanitation, while antibiotic exposure, diarrheal incidence, and malnutrition were common (Table 1).
Study population characteristics (N = 100)
| Risk factors | N |
|---|---|
| Male gender | 50 |
| Hospital birth | 48 |
| Crowding (greater than two people for each room) | 31 |
| High income (> 20 USD per month) | 47 |
| Improved source of drinking water | 41 |
| Improved sanitation | 7 |
| Dirt floor in home | 97 |
| Household owns chickens | 97 |
| Household owns cattle | 74 |
| Household owns agricultural land | 100 |
| Length-for-age Z score at enrollment [median (IQR)] | −0.96 (−1.69 to 0.24) |
| Duration of exclusive breastfeeding (months) [median (IQR)] | 1.8 (1.2–2.7) |
| Number of episodes of diarrhea in 24 months [median (IQR)] | 2 (1–3) |
| Number of episodes of acute lower respiratory tract infection in 24 months [median (IQR)] | 0 (0–1) |
| Number of episodes of fever in 24 months [median (IQR)] | 4 (3–7) |
| Number of antibiotic courses in 24 months [median (IQR)] | 6 (4–9) |
IQR = interquartile range; USD = United States dollars.
* Socioeconomic status derived from the Malnutrition and the Consequences for Child Health and Development study populations.
We first aimed to describe the proportion of resistant samples from all children at all of the time points tested (Figure 1). Most E. coli was resistant to ampicillin (749/837; 89.5%), cefazolin (742/837; 88.6%), and cotrimoxazole (721/837; 86.1%). Less than half was resistant to amoxicillin/clavulanate (361/836; 43.2%). Resistance to ampicillin/sulbactam (178/819; 21.7%), nalidixic acid (131/831; 15.8%), and azithromycin (115/837; 13.7%) was less common. The resistance was rare for the extended-spectrum cephalosporins (ceftriaxone: 22/836; 2.6%, ceftazidime: 10/836; 1.2%, cefotaxime: 18/835; 2.2%, cefepime: 15/837; 1.8%, and cefoxitin: 17/834; 2.0%), aztreonam (11/837; 1.3%), gentamicin (14/837; 1.7%), and ciprofloxacin (19/836; 2.3%). No ertapenem resistance was identified. Only 1.8% (15/837) of the pooled E. coli isolates met the criteria for ESBL production, whereas 82.6% (691/837) met criteria for MDR. All children had at least one E. coli isolate by 24 months of age that met the criteria for MDR. Ampicillin resistance was confirmed by the presence of TEM in the majority of tested isolates (27/30). Of the three ESBL E. coli isolates screened for resistance genes, a CTX gene was detected in two and a SHV gene in one.
Proportion of Escherichia coli isolates resistant to tested antibiotics from 100 children aged 6–60 months.
Citation: The American Journal of Tropical Medicine and Hygiene 100, 5; 10.4269/ajtmh.18-0789
Proportion of Escherichia coli isolates resistant to tested antibiotics from 100 children aged 6–60 months.
Citation: The American Journal of Tropical Medicine and Hygiene 100, 5; 10.4269/ajtmh.18-0789
Proportion of Escherichia coli isolates resistant to tested antibiotics from 100 children aged 6–60 months.
Citation: The American Journal of Tropical Medicine and Hygiene 100, 5; 10.4269/ajtmh.18-0789
Then, using antibiotic susceptibility testing results for samples through 24 months of age, up to which point active surveillance was performed for childhood illness and antibiotic use, we examined risk factors for the presence of drug-resistant E. coli (Table 2). A total of 387 samples of 400 possible underwent testing and were available for this analysis from 6, 12, 18, and 24 months of age in 100 children. Child age was not statistically significantly associated with the number of resistant antibiotic classes (Figure 2). High antibiotic use (0.26 additional resistant antibiotic classes; 95% CI: 0.05, 0.47) and high income (0.28 additional resistant antibiotic classes; 95% CI: 0.06, 0.50) were associated with the carriage of drug-resistant E. coli. All other evaluated putative risk factors, including hospital birth, crowding in the home, improved drinking water and sanitation, and common childhood illness were not associated with the carriage of antibiotic-resistant E. coli. Limiting exposures to a 30-day period before the collection of the stool sample did not change these findings, and exposure to individual antibiotics was not associated with subsequent resistance to the associated antibiotic/antibiotic class. Finally, exposures most relevant for the first 6 months, specifically duration of exclusive breastfeeding and hospital birth, were not associated with the number of resistant antibiotic classes at 6 months of age.
Risk factors for carriage of antibiotic-resistant Escherichia coli
| Risk factor | Change in number of antibiotic-resistant classes (95% CI) |
|---|---|
| Sociodemographic | |
| Male gender | 0.00 (−0.22, 0.22) |
| Hospital birth | 0.09 (−0.15, 0.33) |
| Crowding (greater than two people per room) | 0.02 (−0.22, 0.25) |
| High income (> 20 USD per month) | 0.28 (0.06, 0.50) |
| Improved source of drinking water | −0.06 (−0.29, 0.17) |
| Improved sanitation | −0.26 (−0.90, 0.38) |
| Household owns cattle | −0.20 (−0.46, 0.05) |
| Illness and antibiotic use | |
| Frequent diarrhea (greater than one episode in last 6 months) | −0.33 (−0.69, 0.03) |
| Frequent fever (greater than two episodes in last 6 months) | −0.20 (−0.49, 0.10) |
| Acute lower respiratory tract infection in last 6 months | −0.07 (−0.41, 0.27) |
| High antibiotic use (greater than two courses in last 6 months) | 0.26 (0.05, 0.47) |
USD = United States dollars.
Number of classes of antibiotic resistance in Escherichia coli isolates from 100 children aged 6–24 months.
Citation: The American Journal of Tropical Medicine and Hygiene 100, 5; 10.4269/ajtmh.18-0789
Number of classes of antibiotic resistance in Escherichia coli isolates from 100 children aged 6–24 months.
Citation: The American Journal of Tropical Medicine and Hygiene 100, 5; 10.4269/ajtmh.18-0789
Number of classes of antibiotic resistance in Escherichia coli isolates from 100 children aged 6–24 months.
Citation: The American Journal of Tropical Medicine and Hygiene 100, 5; 10.4269/ajtmh.18-0789
DISCUSSION
Although the carriage of ESBL E. coli was rare, we noted a high prevalence of carriage of MDR E. coli in young children in rural Tanzania. In particular, resistance to ampicillin, cotrimoxazole, and cefazolin, a first-generation cephalosporin, was widespread. These data are consistent with a recent systematic review of antimicrobial resistance in Africa, which focused on invasive isolates.19 It is possible that examination of the carriage of E. coli and other Gram-negative bacteria from stool represents an efficient way to survey community levels of antibiotic resistance. A benzyl- or aminopenicillin in combination with gentamicin remains the first-line therapy for possible serious bacterial infection in young children in these settings,20 and these data suggest that inclusion of an aminoglycoside is warranted.
The drug resistance profile of these E. coli isolates was similar across the range of child ages studied, consistent with early-life colonization with E. coli as a component of the developing infant microbiota.21 High antibiotic use was associated with a slight increase in the number of classes of antibiotic resistance, although there was no association observed for other putative risk factors, including diarrhea and other common childhood illnesses, improved drinking water or sanitation, crowding, hospital birth, and duration of exclusive breastfeeding. The absence of identifiable risk factors for the acquisition of drug-resistant E. coli is also consistent with the assumption that drug-resistant E. coli is acquired from the environment early in life and that risk factors traditionally associated with exposure to enteric pathogens do not mediate this acquisition. This suggests that environmental interventions such as improvements in water, sanitation, or hygiene may be unlikely to reduce the carriage of antibiotic-resistant E. coli. Furthermore, although individual exposure to antibiotics may slightly increase the antibiotic resistance in commensal E. coli, population-level interventions may be needed, with reduction in antibiotic use the clearest and most feasible intervention. The association between income and carriage of drug-resistant E. coli was surprising. Although this association was present despite inclusion of antibiotic use in the same model, it is possible that this still serves as a marker of access to and use of health care.
Our study has several limitations. First, our study population of only 100 children was relatively small. Although we tested E. coli isolates at 10 time points, the 6-month intervals in between sample testing may be too long to detect transient resistance pattern changes. Antibiotic use was relatively low in this cohort, for example, in comparison with other sites in the same birth cohort study.22 For amoxicillin/clavulanate, ampicillin/sulbactam, and azithromycin, we saw a marked change in resistance rates in between months 24 and 30. Because frozen E. coli pools were tested 24 months and earlier, whereas fresh cultures were tested 30 months and later, this may suggest that the stored isolates have artifactually lower drug resistance, for instance, because of losing plasmids harboring resistance genes over time.23
In summary, we noted surprisingly high rates of antibiotic-resistant E. coli, including MDR E. coli, in young children in rural Tanzania. The prevalence of resistance was consistent across children aged 6 months to 5 years, suggesting that the carriage of MDR E. coli in children in this setting is acquired from the environment or very early in life. This suggests the importance of population-level interventions to decrease the carriage of drug-resistant E. coli.
Acknowledgments:
The Etiology, Risk Factors and Interactions of Enteric Infections and Malnutrition and the Consequences for Child Health and Development Project (MAL-ED) is a collaborative project supported by the Bill & Melinda Gates Foundation, the Foundation for the NIH, and the NIH, Fogarty International Center. We thank the staff and participants of the MAL-ED Network Project for their important contributions.
REFERENCES
- 1.↑
O’Neill J, 2014. Tackling Drug-Resistant Infections Globally: Final Report and Recommendations. The Review on Antimicrobial Resistance. Available at: https://amr-review.org/Publications.html. Accessed September 28, 2018.
- 2.↑
World Health Organization, 2017. Global Priority List of Antibiotic-Resistant Bacteria to Guide Research, Discovery, and Development of New Antibiotics. Available at: http://www.who.int/medicines/publications/global-priority-list-antibiotic-resistant-bacteria/en/. Accessed September 28, 2018.
- 3.↑
Liu L, Oza S, Hogan D, Perin J, Rudan I, Lawn JE, Cousens S, Mathers C, Black RE, 2015. Global, regional, and national causes of child mortality in 2000–13, with projections to inform post-2015 priorities: an updated systematic analysis. Lancet 385: 430–440.
- 4.↑
Laxminarayan R, Matsoso P, Pant S, Brower C, Rottingen JA, Klugman K, Davies S, 2016. Access to effective antimicrobials: a worldwide challenge. Lancet 387: 168–175.
- 5.
Investigators of the Delhi Neonatal Infection Study (DeNIS) Collaboration, 2016. Characterisation and antimicrobial resistance of sepsis pathogens in neonates born in tertiary care centres in Delhi, India: a cohort study. Lancet Glob Health 4: e752–e760.
- 6.
Kayange N, Kamugisha E, Mwizamholya DL, Jeremiah S, Mshana SE, 2010. Predictors of positive blood culture and deaths among neonates with suspected neonatal sepsis in a tertiary hospital, Mwanza-Tanzania. BMC Pediatr 10: 39.
- 7.↑
Fox-Lewis A et al. 2018. Antimicrobial resistance in invasive bacterial infections in hospitalized children, Cambodia, 2007–2016. Emerg Infect Dis 24: 841–851.
- 8.↑
Bailey JK, Pinyon JL, Anantham S, Hall RM, 2010. Commensal Escherichia coli of healthy humans: a reservoir for antibiotic-resistance determinants. J Med Microbiol 59: 1331–1339.
- 9.↑
Klemm EJ et al. 2018. Emergence of an extensively drug-resistant Salmonella enterica serovar typhi clone harboring a promiscuous plasmid encoding resistance to fluoroquinolones and third-generation cephalosporins. MBio 9: e00105-18.
- 10.↑
Turner P, Pol S, Soeng S, Sar P, Neou L, Chea P, Day NP, Cooper BS, Turner C, 2016. High prevalence of antimicrobial-resistant gram-negative colonization in hospitalized Cambodian infants. Pediatr Infect Dis J 35: 856–861.
- 11.↑
Das P, Singh AK, Pal T, Dasgupta S, Ramamurthy T, Basu S, 2011. Colonization of the gut with Gram-negative bacilli, its association with neonatal sepsis and its clinical relevance in a developing country. J Med Microbiol 60: 1651–1660.
- 12.↑
Richard SA, Barrett LJ, Guerrant RL, Checkley W & Miller MA MAL-ED Network Investigators, 2014. Disease surveillance methods used in the 8-site MAL-ED cohort study. Clin Infect Dis 59: S220–S224.
- 13.↑
Houpt E et al. MAL-ED Network Investigators, 2014. Microbiologic methods utilized in the MAL-ED cohort study. Clin Infect Dis 59: S225–S232.
- 14.↑
Mduma ER et al. 2014. The etiology, risk factors, and interactions of enteric infections and malnutrition and the consequences for child health and development study (MAL-ED): description of the Tanzanian site. Clin Infect Dis 59: S325–S330.
- 15.↑
Clinical and Laboratory Standards Institute (CLSI), 2016. Performance Standards for Antimicrobial Susceptbility Testing. Twenty-Sixth Informational Supplement. M100-S26. Wayne, PA: CLSI.
- 16.↑
Seidman JC, Coles CL, Silbergeld EK, Levens J, Mkocha H, Johnson LB, Munoz B, West SK, 2014. Increased carriage of macrolide-resistant fecal E. coli following mass distribution of azithromycin for trachoma control. Int J Epidemiol 43: 1105–1113.
- 17.↑
Roschanski N, Fischer J, Guerra B, Roesler U, 2014. Development of a multiplex real-time PCR for the rapid detection of the predominant beta-lactamase genes CTX-M, SHV, TEM and CIT-type AmpCs in Enterobacteriaceae. PLoS One 9: e100956.
- 18.↑
Magiorakos AP et al. 2012. Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance. Clin Microbiol Infect 18: 268–281.
- 19.↑
Tadesse BT, Ashley EA, Ongarello S, Havumaki J, Wijegoonewardena M, Gonzalez IJ, Dittrich S, 2017. Antimicrobial resistance in Africa: a systematic review. BMC Infect Dis 17: 616.
- 20.↑
Mathur S, Fuchs A, Bielicki J, Van Den Anker J, Sharland M, 2018. Antibiotic use for community-acquired pneumonia in neonates and children: WHO evidence review. Paediatr Int Child Health 38(Supp 1): S66–S75.
- 21.↑
Backhed F et al. 2015. Dynamics and stabilization of the human gut microbiome during the first year of life. Cell Host Microbe 17: 690–703.
- 22.↑
Rogawski ET et al. 2017. Use of antibiotics in children younger than two years in eight countries: a prospective cohort study. Bull World Health Organ 95: 49–61.
- 23.↑
Chen S, Larsson M, Robinson RC, Chen SL, 2017. Direct and convenient measurement of plasmid stability in lab and clinical isolates of E. coli. Sci Rep 7: 4788.