Limits...
How competition governs whether moderate or aggressive treatment minimizes antibiotic resistance.

Colijn C, Cohen T - Elife (2015)

Bottom Line: In this study, we demonstrate how one can understand and resolve these apparently contradictory conclusions.We show that a key determinant of which treatment strategy will perform best at the individual level is the extent of effective competition between resistant and sensitive pathogens within a host.We extend our analysis to the community level, exploring the spectrum between strict inter-strain competition and strain independence.

View Article: PubMed Central - PubMed

Affiliation: Department of Mathematics, Imperial College London, London, United Kingdom.

ABSTRACT
Understanding how our use of antimicrobial drugs shapes future levels of drug resistance is crucial. Recently, there has been debate over whether an aggressive (i.e., high dose) or more moderate (i.e., lower dose) treatment of individuals will most limit the emergence and spread of resistant bacteria. In this study, we demonstrate how one can understand and resolve these apparently contradictory conclusions. We show that a key determinant of which treatment strategy will perform best at the individual level is the extent of effective competition between resistant and sensitive pathogens within a host. We extend our analysis to the community level, exploring the spectrum between strict inter-strain competition and strain independence. From this perspective as well, we find that the magnitude of effective competition between resistant and sensitive strains determines whether an aggressive approach or moderate approach minimizes the burden of resistance in the population.

No MeSH data available.


Related in: MedlinePlus

Trajectories of the between-host model under varying treatment.Treatment is introduced at 5 years. (Left) Parameters are βx = 1.5, βy = 1.04, c = 0.05, r = 0, μ = 0.001. (Middle) Parameters are βx = 2, βy = 1.1, c = 0.3, r = 0.05, μ = 0.0001. (Right) Invasion analysis (bifurcation) plot. The plot shows regions of stability of the disease-free equilibrium (both R0 values less than one), the DR-only equilibrium (top left region), and the equilibrium with both (primarily DS, with low-level DR due to acquisition). The diagonal lines show the boundary at which the DR-only equilibrium loses stability. Lines move to the right as the similarity coefficient increases from 0 (light blue vertical line) to 1 (pink). When it reaches 1, the line is R01 = R02.DOI:http://dx.doi.org/10.7554/eLife.10559.006
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4641510&req=5

fig5: Trajectories of the between-host model under varying treatment.Treatment is introduced at 5 years. (Left) Parameters are βx = 1.5, βy = 1.04, c = 0.05, r = 0, μ = 0.001. (Middle) Parameters are βx = 2, βy = 1.1, c = 0.3, r = 0.05, μ = 0.0001. (Right) Invasion analysis (bifurcation) plot. The plot shows regions of stability of the disease-free equilibrium (both R0 values less than one), the DR-only equilibrium (top left region), and the equilibrium with both (primarily DS, with low-level DR due to acquisition). The diagonal lines show the boundary at which the DR-only equilibrium loses stability. Lines move to the right as the similarity coefficient increases from 0 (light blue vertical line) to 1 (pink). When it reaches 1, the line is R01 = R02.DOI:http://dx.doi.org/10.7554/eLife.10559.006

Mentions: We took several approaches to understand how the parameters of each model relate to whether aggressive or moderate treatment minimizes resistance. The most direct approach is simply to choose a set of parameters, vary the dosage, and examine how resistance changes (Figure 2). Naturally, the result depends strongly on the parameter choice. We also vary one parameter at a time, keeping others fixed, and examine the trajectories (Appendix figures 2, 3). The next approach is to examine, over all simulations simultaneously, how the outcome depends on each parameter by stratifying the outcomes (Figure 3). Using heatmaps or scatter plots, it is also possible to explore how pairs of parameters determine an outcome (Figure 4). We take the same approach in the between-host model, with Figure 5 showing demonstrative trajectories under varying treatment strength, Appendix figure 4 showing a sensitivity analysis varying one parameter at a time, and Figures 6, 7 showing the stratified dependence of the outcome on single and paired parameters while other parameters are allowed to vary.10.7554/eLife.10559.003Figure 2.How treatment changes the trajectory of the in-host model.


How competition governs whether moderate or aggressive treatment minimizes antibiotic resistance.

Colijn C, Cohen T - Elife (2015)

Trajectories of the between-host model under varying treatment.Treatment is introduced at 5 years. (Left) Parameters are βx = 1.5, βy = 1.04, c = 0.05, r = 0, μ = 0.001. (Middle) Parameters are βx = 2, βy = 1.1, c = 0.3, r = 0.05, μ = 0.0001. (Right) Invasion analysis (bifurcation) plot. The plot shows regions of stability of the disease-free equilibrium (both R0 values less than one), the DR-only equilibrium (top left region), and the equilibrium with both (primarily DS, with low-level DR due to acquisition). The diagonal lines show the boundary at which the DR-only equilibrium loses stability. Lines move to the right as the similarity coefficient increases from 0 (light blue vertical line) to 1 (pink). When it reaches 1, the line is R01 = R02.DOI:http://dx.doi.org/10.7554/eLife.10559.006
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4641510&req=5

fig5: Trajectories of the between-host model under varying treatment.Treatment is introduced at 5 years. (Left) Parameters are βx = 1.5, βy = 1.04, c = 0.05, r = 0, μ = 0.001. (Middle) Parameters are βx = 2, βy = 1.1, c = 0.3, r = 0.05, μ = 0.0001. (Right) Invasion analysis (bifurcation) plot. The plot shows regions of stability of the disease-free equilibrium (both R0 values less than one), the DR-only equilibrium (top left region), and the equilibrium with both (primarily DS, with low-level DR due to acquisition). The diagonal lines show the boundary at which the DR-only equilibrium loses stability. Lines move to the right as the similarity coefficient increases from 0 (light blue vertical line) to 1 (pink). When it reaches 1, the line is R01 = R02.DOI:http://dx.doi.org/10.7554/eLife.10559.006
Mentions: We took several approaches to understand how the parameters of each model relate to whether aggressive or moderate treatment minimizes resistance. The most direct approach is simply to choose a set of parameters, vary the dosage, and examine how resistance changes (Figure 2). Naturally, the result depends strongly on the parameter choice. We also vary one parameter at a time, keeping others fixed, and examine the trajectories (Appendix figures 2, 3). The next approach is to examine, over all simulations simultaneously, how the outcome depends on each parameter by stratifying the outcomes (Figure 3). Using heatmaps or scatter plots, it is also possible to explore how pairs of parameters determine an outcome (Figure 4). We take the same approach in the between-host model, with Figure 5 showing demonstrative trajectories under varying treatment strength, Appendix figure 4 showing a sensitivity analysis varying one parameter at a time, and Figures 6, 7 showing the stratified dependence of the outcome on single and paired parameters while other parameters are allowed to vary.10.7554/eLife.10559.003Figure 2.How treatment changes the trajectory of the in-host model.

Bottom Line: In this study, we demonstrate how one can understand and resolve these apparently contradictory conclusions.We show that a key determinant of which treatment strategy will perform best at the individual level is the extent of effective competition between resistant and sensitive pathogens within a host.We extend our analysis to the community level, exploring the spectrum between strict inter-strain competition and strain independence.

View Article: PubMed Central - PubMed

Affiliation: Department of Mathematics, Imperial College London, London, United Kingdom.

ABSTRACT
Understanding how our use of antimicrobial drugs shapes future levels of drug resistance is crucial. Recently, there has been debate over whether an aggressive (i.e., high dose) or more moderate (i.e., lower dose) treatment of individuals will most limit the emergence and spread of resistant bacteria. In this study, we demonstrate how one can understand and resolve these apparently contradictory conclusions. We show that a key determinant of which treatment strategy will perform best at the individual level is the extent of effective competition between resistant and sensitive pathogens within a host. We extend our analysis to the community level, exploring the spectrum between strict inter-strain competition and strain independence. From this perspective as well, we find that the magnitude of effective competition between resistant and sensitive strains determines whether an aggressive approach or moderate approach minimizes the burden of resistance in the population.

No MeSH data available.


Related in: MedlinePlus