Advertisement

Global forecast of antimicrobial resistance in invasive isolates of Escherichia coli and Klebsiella pneumoniae

Open AccessPublished:February 02, 2018DOI:https://doi.org/10.1016/j.ijid.2018.01.011

      Highlights

      • We aimed to estimate global trends of antimicrobial resistance of E. coli and K. pneumoniae invasive isolates.
      • By 2030, over half of E. coli and K. pneumoniae invasive isolates could become resistant to third generation cephalosporins.
      • Carbapenem resistance is increasing rapidly among K. pneumoniae invasive isolates.

      Abstract

      Objectives

      To project future antimicrobial resistance (AMR) in Escherichia coli and Klebsiella pneumoniae.

      Methods

      Mixed linear models were constructed from a sample of countries with AMR data in the ResistanceMap database. Inverse probability weighting methods were used to account for countries without AMR data.

      Results

      The estimated prevalence of AMR in 2015 was 64.5% (95% confidence interval (CI) 42–87%) for third-generation cephalosporin-resistant (3GCR) Escherichia coli, 5.8% (95% CI 1.8–9.7%) for carbapenem-resistant (CR) E. coli, 66.9% (95% CI 47.1–86.8%) for 3GCR Klebsiella pneumoniae, and 23.4% (95% CI 7.4–39.4%) for CR K. pneumoniae. The projected AMR prevalence in 2030 was 77% (95% CI 55–99.1%) for 3GCR E. coli, 11.8% (95% CI 3.7–19.9%) for CR E. coli, 58.2% (95% CI 50.2–66.1%) for 3GCR K. pneumoniae, and 52.8% (95% CI 16.3–89.3%) for CR K. pneumoniae.

      Conclusions

      The models suggest that third-generation cephalosporins and carbapenems could be ineffective against a sizeable proportion of infections by E. coli and K. pneumoniae in most parts of the world by 2030, supporting both the need to enhance stewardship efforts and to prioritize research and development of new antibiotics for resistant Enterobacteriaceae.

      Keywords

      Introduction

      Antimicrobial resistance (AMR) is now a global problem, and resistance in Enterobacteriaceae, specifically Escherichia coli and Klebsiella pneumoniae, is a critical threat to human health (
      • World Health Organization
      Antimicrobial resistance: global report on surveillance.
      ,

      World Health Organization. Global priority list of antibiotic-resistant bacteria to guide research, discovery, and development of new antibiotics. 2017. Available at: http://www.who.int/medicines/publications/global-priority-list-antibiotic-resistant-bacteria/en/. [Accessed 22 August 2017, n.d.].

      ). Infections caused by third-generation cephalosporin-resistant (3GCR) Enterobacteriaceae are associated with increased mortality, length of stay, and costs compared with drug-sensitive strains (
      • Stewardson Andrew J.
      • Allignol Arthur
      • Beyersmann Jan
      • Graves Nicholas
      • Schumacher Martin
      • Meyer Rodolphe
      • et al.
      The health and economic burden of bloodstream infections caused by antimicrobial-susceptible and non-susceptible Enterobacteriaceae and Staphylococcus aureus in European hospitals, 2010 and 2011: a multicentre retrospective cohort study.
      ). Carbapenems are less reliable as last-resort antibiotics because of increasing resistance (
      • Gelband Hellen
      • Miller-Petrie Molly
      • Pant Suraj
      • Gandra Sumanth
      • Levinson Jordan
      • Barter Devra
      • et al.
      The state of the world’s antibiotics, 2015.
      ). AMR already imposes a heavy economic burden on health systems (
      • Stewardson Andrew J.
      • Allignol Arthur
      • Beyersmann Jan
      • Graves Nicholas
      • Schumacher Martin
      • Meyer Rodolphe
      • et al.
      The health and economic burden of bloodstream infections caused by antimicrobial-susceptible and non-susceptible Enterobacteriaceae and Staphylococcus aureus in European hospitals, 2010 and 2011: a multicentre retrospective cohort study.
      ). Projecting future prevalence of AMR may help prioritize research projects and interventions. The aims of this study were to estimate global trends in AMR in E. coli and K. pneumoniae and to project future AMR prevalence to 2030.

      Methods

      Data on population and gross national income per capita (GNIPC) from the World Bank and on AMR from ResistanceMap, a global repository of AMR data from quality-assured and accredited hospitals and laboratory networks, were used (

      ResistanceMap. Center for Disease Dynamics, Economics & Policy (CDDEP). Available at: http://resistancemap.cddep.org/. [Accessed 11 July 2016, n.d.].

      ). Countries for which samples were obtained from a single hospital were excluded from this study. Annual AMR data that had fewer than 30 isolates were also excluded (
      • Agresti Alan
      • Caffo Brian
      Simple and effective confidence intervals for proportions and differences of proportions result from adding two successes and two failures.
      ).
      With few exceptions, low- and middle-income countries are less likely to monitor AMR; therefore, high-income countries are overrepresented in the ResistanceMap database. Not taking this into account may lead to selection bias and an underestimation of the prevalence of AMR because of the strong negative association between GNIPC and the prevalence of AMR (
      • Alvarez-Uria Gerardo
      • Gandra Sumanth
      • Laxminarayan Ramanan
      Poverty and prevalence of antimicrobial resistance in invasive isolates.
      ). To overcome this problem, inverse probability of inclusion (IPI) weighting, a method analogous to the use of inverse probability weights, was used to account for non-responders in surveys (
      • Dugoff Eva H.
      • Schuler Megan
      • Stuart Elizabeth A.
      Generalizing observational study results: applying propensity score methods to complex surveys.
      ). IPI weights were calculated based on the inverse probability of being included in the study, using a logistic regression model that included data from countries in the world for which GNIPC data were available (countries with no GNIPC data comprised 1.3% of the world population) (

      Open Data Catalog, The World Bank. Available at: http://datacatalog.worldbank.org/. [Accessed 28 November 2016, n.d.].

      ). In this logistic regression model, the availability of national AMR data (thus being included in the study) was the dependent variable, and orthogonal cubic spline transformations of 2014 GNIPC and 2014 country populations were the independent covariates (
      • Dugoff Eva H.
      • Schuler Megan
      • Stuart Elizabeth A.
      Generalizing observational study results: applying propensity score methods to complex surveys.
      ). IPI weights gave more ‘weight’ to countries that were less likely to have AMR data in the ResistanceMap database, based on their GNIPC and population. This method helps generalize the results of the study to the world population. IPI weights were multiplied by population weights, which gives more weight to countries with larger populations, and the results were used as probability or sample weights in the final mixed model with random intercept and slopes (
      • Dugoff Eva H.
      • Schuler Megan
      • Stuart Elizabeth A.
      Generalizing observational study results: applying propensity score methods to complex surveys.
      ). The mixed models were used to project AMR up to 2030.

      Results

      The study included 45 countries with AMR data for E. coli and 43 countries with AMR data for K. pneumoniae. In countries with E. coli AMR data, the median number of AMR point estimates was 14 (interquartile range 1–15), and 31 were high-income countries. In countries with K. pneumoniae AMR data, the median number of AMR point estimates was 10 (interquartile range 2–14), and 28 were high-income countries. No country had AMR data beyond 2015.
      Forecast estimates of global AMR are presented in Figure 1. The estimated prevalence of AMR in 2015 was 64.5% (95% confidence interval (CI) 42–87%) for 3GCR E. coli, 5.8% (95% CI 1.8–9.7%) for carbapenem-resistant (CR) E. coli, 66.9% (95% CI 47.1–86.8%) for 3GCR K. pneumoniae, and 23.4% (95% CI 7.4–39.4%) for CR K. pneumoniae. The projected annual variation (slope) of AMR was 0.83% (95% CI 0.73–0.93%) for 3GCR E. coli, 0.4% (95% CI 0.12–0.68%) for CR E. coli, −0.58% (95% CI −1.46% to 0.3%) for 3GCR K. pneumoniae, and 1.96% (95% CI 0.59–3.33%) for CR K. pneumoniae. The projected AMR prevalence in 2030 was 77% (95% CI 55–99.1%) for 3GCR E. coli, 11.8% (95% CI 3.7–19.9%) for CR E. coli, 58.2% (95% CI 50.2–66.1%) for 3GCR K. pneumoniae, and 52.8% (95% CI 16.3–89.3%) for CR K. pneumoniae. Projections for individual countries with at least four AMR point estimates using simple linear regression are presented in the Supplementary material.
      Figure 1
      Figure 1Forecast estimates with 95% confidence intervals of global resistance of Escherichia coli (A) and Klebsiella pneumoniae (B) to third-generation (3G) cephalosporins and carbapenems based on population weighted mixed models with random slopes and intercepts.

      Discussion

      The projections of AMR in this study signal a potentially serious shortage of effective antimicrobials for common causes of infection by 2030. Under current trends, over three-fourths of E. coli globally will be 3GCR, and over half of K. pneumoniae invasive isolates will be CR. The consequences of the high prevalence of AMR could be devastating for health systems (
      • World Health Organization
      Antimicrobial resistance: global report on surveillance.
      ,
      • de Kraker Marlieke E.A.
      • Davey Peter G.
      • Grundmann Hajo
      • on behalf of the BURDEN study group
      Mortality and hospital stay associated with resistant Staphylococcus aureus and Escherichia coli bacteremia: estimating the burden of antibiotic resistance in Europe.
      ).
      The models showed that the annual variation in the prevalence of 3GCR K. pneumoniae was not significantly different from zero, with narrowing of the confidence interval over time. This can be explained by the fact that countries with initial low prevalence of 3GCR K. pneumoniae showed a rising trend over time, while the trends were stable or mildly decreasing in countries with initial high prevalence, such as India and South Africa (Supplementary material, Figure S3). CR K. pneumoniae had the highest annual increase of AMR, which could reach 53% by 2030, but the confidence intervals were wide, indicating uncertainty of the projections. The projected increase in the prevalence of CR E. coli was more modest. However, empirical treatment of infections will need to cover 3GCR, leading to an increased use of carbapenems, and this, in turn, may accelerate the pace of CR.
      Enterobacteriaceae are part of the human gut microbiota, and the spread of AMR is facilitated by conditions that are more common in resource-poor settings, such as suboptimal sewage systems and a lack of access to clean water (
      • Holmes Alison H.
      • Moore Luke S.P.
      • Sundsfjord Arnfinn
      • Steinbakk Martin
      • Regmi Sadie
      • Karkey Abhilasha
      • et al.
      Understanding the mechanisms and drivers of antimicrobial resistance.
      ). Previous studies have shown that resistance in Enterobacteriaceae can emerge anywhere and spread around the globe (
      • Nordmann Patrice
      • Poirel Laurent
      • Walsh Timothy R.
      • Livermore David M.
      The emerging NDM carbapenemases.
      ). Isolated interventions in high-income countries alone, without intervention efforts in low- and middle-income countries, may be ineffective in a globalized world (
      • Nordmann Patrice
      • Poirel Laurent
      • Walsh Timothy R.
      • Livermore David M.
      The emerging NDM carbapenemases.
      ).
      This study has important limitations. The total population of all countries included in the study was approximately a third of the world population and was biased towards high-income countries. While IPI models were used to attempt to correct for underrepresentation of low- and middle-income countries, more surveillance data are urgently needed to improve current and future estimates of AMR. The projections for future levels of AMR were based on linear models, which assumed no changes in the growth rate of resistance. They also did not account for saturation or stabilization of AMR levels, as was observed with 3GCR K. pneumoniae. In addition, it was not possible to distinguish the case mix of community- and hospital-acquired infections among the countries included in the study, and the high prevalence of AMR in some countries could be influenced by a higher proportion of hospital-acquired infections (
      • Dat Vu Quoc
      • Vu Hieu Ngoc
      • Nguyen The Hung
      • Nguyen Hoa Thi
      • Hoang Long Bao
      • Vu Tien Viet Dung
      • et al.
      Bacterial bloodstream infections in a tertiary infectious diseases hospital in Northern Vietnam: aetiology, drug resistance, and treatment outcome.
      ,
      • Thaden Joshua T.
      • Li Yanhong
      • Ruffin Felicia
      • Maskarinec Stacey A.
      • Hill-Rorie Jonathan M.
      • Wanda Lisa C.
      • et al.
      Increased costs associated with bloodstream infections caused by multidrug-resistant gram-negative bacteria are due primarily to patients with hospital-acquired infections.
      ).
      These results suggest that if current trends were to continue, third-generation cephalosporins and carbapenems could become ineffective against E. coli and K. pneumoniae in most parts of the world in the not-too-distant future. Empirical antimicrobial therapy for sepsis or for urinary tract or abdominal infections might shift to non-beta-lactam antibiotics, which, in turn, may lead to an increase in AMR in other antibiotic groups. These results underscore the need to improve the judicious use of antimicrobials and support recent World Health Organization recommendations to prioritize the research, discovery, and development of new and effective antibiotic treatments for beta-lactam-resistant Enterobacteriaceae (

      World Health Organization. Global priority list of antibiotic-resistant bacteria to guide research, discovery, and development of new antibiotics. 2017. Available at: http://www.who.int/medicines/publications/global-priority-list-antibiotic-resistant-bacteria/en/. [Accessed 22 August 2017, n.d.].

      ).

      Funding

      The funders had no role in the study design, in the collection, analysis, and interpretation of the data, in the writing of the report, or in the decision to submit the article for publication.

      Ethics approval

      This study used data available in the public domain and thus did not require ethics approval.

      Conflict of interest

      There are no conflicts of interest to disclose.

      Acknowledgments

      The research leading to these results received support from the Bill & Melinda Gates Foundation for the ResistanceMap project (Investment ID OPP1112355 ) and from the Innovative Medicines Initiative Joint Undertaking under grant agreement number 115618 (Driving re-investment in R&D and responsible antibiotic use, DRIVE-AB, www.drive-ab.eu) (for SG, RL), resources of which are composed of financial contributions from the European Union Seventh Framework Programme ( FP7/2007-2013 ) and European Federation of Pharmaceutical Industries and Associations ( EFPIA ) companies in kind contribution.

      Appendix A. Supplementary data

      The following are Supplementary data to this article:

      References

        • Agresti Alan
        • Caffo Brian
        Simple and effective confidence intervals for proportions and differences of proportions result from adding two successes and two failures.
        Am Stat. 2000; 54: 280-288https://doi.org/10.1080/00031305.2000.10474560
        • Alvarez-Uria Gerardo
        • Gandra Sumanth
        • Laxminarayan Ramanan
        Poverty and prevalence of antimicrobial resistance in invasive isolates.
        Int J Infect Dis. 2016; 52: 59-61https://doi.org/10.1016/j.ijid.2016.09.026
      1. ResistanceMap. Center for Disease Dynamics, Economics & Policy (CDDEP). Available at: http://resistancemap.cddep.org/. [Accessed 11 July 2016, n.d.].

        • Dat Vu Quoc
        • Vu Hieu Ngoc
        • Nguyen The Hung
        • Nguyen Hoa Thi
        • Hoang Long Bao
        • Vu Tien Viet Dung
        • et al.
        Bacterial bloodstream infections in a tertiary infectious diseases hospital in Northern Vietnam: aetiology, drug resistance, and treatment outcome.
        BMC Infect Dis. 2017; 17: 493https://doi.org/10.1186/s12879-017-2582-7
        • Dugoff Eva H.
        • Schuler Megan
        • Stuart Elizabeth A.
        Generalizing observational study results: applying propensity score methods to complex surveys.
        Health Serv Res. 2014; 49: 284-303https://doi.org/10.1111/1475-6773.12090
        • Gelband Hellen
        • Miller-Petrie Molly
        • Pant Suraj
        • Gandra Sumanth
        • Levinson Jordan
        • Barter Devra
        • et al.
        The state of the world’s antibiotics, 2015.
        Center for Disease Dynamics, Economics & Policy (CDDEP), Washington, D.C2015
        • Holmes Alison H.
        • Moore Luke S.P.
        • Sundsfjord Arnfinn
        • Steinbakk Martin
        • Regmi Sadie
        • Karkey Abhilasha
        • et al.
        Understanding the mechanisms and drivers of antimicrobial resistance.
        Lancet. 2016; 387: 176-187https://doi.org/10.1016/S0140-6736(15)00473-0
        • Nordmann Patrice
        • Poirel Laurent
        • Walsh Timothy R.
        • Livermore David M.
        The emerging NDM carbapenemases.
        Trends Microbiol. 2011; 19: 588-595https://doi.org/10.1016/j.tim.2011.09.005
        • Stewardson Andrew J.
        • Allignol Arthur
        • Beyersmann Jan
        • Graves Nicholas
        • Schumacher Martin
        • Meyer Rodolphe
        • et al.
        The health and economic burden of bloodstream infections caused by antimicrobial-susceptible and non-susceptible Enterobacteriaceae and Staphylococcus aureus in European hospitals, 2010 and 2011: a multicentre retrospective cohort study.
        Euro Surveill. 2016; 21https://doi.org/10.2807/1560-7917.ES.2016.21.33.30319
        • Thaden Joshua T.
        • Li Yanhong
        • Ruffin Felicia
        • Maskarinec Stacey A.
        • Hill-Rorie Jonathan M.
        • Wanda Lisa C.
        • et al.
        Increased costs associated with bloodstream infections caused by multidrug-resistant gram-negative bacteria are due primarily to patients with hospital-acquired infections.
        Antimicrob Agents Chemother. 2017; 61: 1709-1716
      2. Open Data Catalog, The World Bank. Available at: http://datacatalog.worldbank.org/. [Accessed 28 November 2016, n.d.].

      3. World Health Organization. Global priority list of antibiotic-resistant bacteria to guide research, discovery, and development of new antibiotics. 2017. Available at: http://www.who.int/medicines/publications/global-priority-list-antibiotic-resistant-bacteria/en/. [Accessed 22 August 2017, n.d.].

        • World Health Organization
        Antimicrobial resistance: global report on surveillance.
        World Health Organization, Geneva, Switzerland2014
        • de Kraker Marlieke E.A.
        • Davey Peter G.
        • Grundmann Hajo
        • on behalf of the BURDEN study group
        Mortality and hospital stay associated with resistant Staphylococcus aureus and Escherichia coli bacteremia: estimating the burden of antibiotic resistance in Europe.
        PLoS Med. 2011; 8e1001104https://doi.org/10.1371/journal.pmed.1001104#pmed.1001104.s004