High rifampicin peak plasma concentrations accelerate the slow phase of bacterial decline in tuberculosis patients: Evidence for heteroresistance.

BACKGROUND: Antibiotic treatments are often associated with a late slowdown in bacterial killing. This separates the killing of bacteria into at least two distinct phases: a quick phase followed by a slower phase, the latter of which is linked to treatment success. Current mechanistic explanations for the in vitro slowdown are either antibiotic persistence or heteroresistance. Persistence is defined as the switching back and forth between susceptible and non-susceptible states, while heteroresistance is defined as the coexistence of bacteria with heterogeneous susceptibilities. Both are also thought to cause a slowdown in the decline of bacterial populations in patients and therefore complicate and prolong antibiotic treatments. Reduced bacterial death rates over time are also observed within tuberculosis patients, yet the mechanistic reasons for this are unknown and therefore the strategies to mitigate them are also unknown. METHODS AND FINDINGS: We analyse a dose ranging trial for rifampicin in tuberculosis patients and show that there is a slowdown in the decline of bacteria. We show that the late phase of bacterial killing depends more on the peak drug concentrations than the total drug exposure. We compare these to pharmacokinetic-pharmacodynamic models of rifampicin heteroresistance and persistence. We find that the observation on the slow phase's dependence on pharmacokinetic measures, specifically peak concentrations are only compatible with models of heteroresistance and incompatible with models of persistence. The quantitative agreement between heteroresistance models and observations is very good ([Formula: see text]). To corroborate the importance of the slowdown, we validate our results by estimating the time to sputum culture conversion and compare the results to a different dose ranging trial. CONCLUSIONS: Our findings indicate that higher doses, specifically higher peak concentrations may be used to optimize rifampicin treatments by accelerating bacterial killing in the slow phase. It adds to the growing body of literature supporting higher rifampicin doses for shortening tuberculosis treatments.

vCOMBAT: a novel tool to create and visualize a computational model of bacterial antibiotic target-binding.

BACKGROUND: As antibiotic resistance creates a significant global health threat, we need not only to accelerate the development of novel antibiotics but also to develop better treatment strategies using existing drugs to improve their efficacy and prevent the selection of further resistance. We require new tools to rationally design dosing regimens from data collected in early phases of antibiotic and dosing development. Mathematical models such as mechanistic pharmacodynamic drug-target binding explain mechanistic details of how the given drug concentration affects its targeted bacteria. However, there are no available tools in the literature that allow non-quantitative scientists to develop computational models to simulate antibiotic-target binding and its effects on bacteria. RESULTS: In this work, we have devised an extension of a mechanistic binding-kinetic model to incorporate clinical drug concentration data. Based on the extended model, we develop a novel and interactive web-based tool that allows non-quantitative scientists to create and visualize their own computational models of bacterial antibiotic target-binding based on their considered drugs and bacteria. We also demonstrate how Rifampicin affects bacterial populations of Tuberculosis bacteria using our vCOMBAT tool. CONCLUSIONS: The vCOMBAT online tool is publicly available at https://combat-bacteria.org/ .

Identifying the drivers of multidrug-resistant Klebsiella pneumoniae at a European level.

Beta-lactam- and in particular carbapenem-resistant Enterobacteriaceae represent a major public health threat. Despite strong variation of resistance across geographical settings, there is limited understanding of the underlying drivers. To assess these drivers, we developed a transmission model of cephalosporin- and carbapenem-resistant Klebsiella pneumoniae. The model is parameterized using antibiotic consumption and demographic data from eleven European countries and fitted to the resistance rates for Klebsiella pneumoniae for these settings. The impact of potential drivers of resistance is then assessed in counterfactual analyses. Based on reported consumption data, the model could simultaneously fit the prevalence of extended-spectrum beta-lactamase-producing and carbapenem-resistant Klebsiella pneumoniae (ESBL and CRK) across eleven European countries over eleven years. The fit could explain the large between-country variability of resistance in terms of consumption patterns and fitted differences in hospital transmission rates. Based on this fit, a counterfactual analysis found that reducing nosocomial transmission and antibiotic consumption in the hospital had the strongest impact on ESBL and CRK prevalence. Antibiotic consumption in the community also affected ESBL prevalence but its relative impact was weaker than inpatient consumption. Finally, we used the model to estimate a moderate fitness cost of CRK and ESBL at the population level. This work highlights the disproportionate role of antibiotic consumption in the hospital and of nosocomial transmission for resistance in gram-negative bacteria at a European level. This indicates that infection control and antibiotic stewardship measures should play a major role in limiting resistance even at the national or regional level.

Quantifying the impact of treatment history on plasmid-mediated resistance evolution in human gut microbiota.

To understand how antibiotic use affects the risk of a resistant infection, we present a computational model of the population dynamics of gut microbiota including antibiotic resistance-conferring plasmids. We then describe how this model is parameterized based on published microbiota data. Finally, we investigate how treatment history affects the prevalence of resistance among opportunistic enterobacterial pathogens. We simulate treatment histories and identify which properties of prior antibiotic exposure are most influential in determining the prevalence of resistance. We find that resistance prevalence can be predicted by 3 properties, namely the total days of drug exposure, the duration of the drug-free period after last treatment, and the center of mass of the treatment pattern. Overall this work provides a framework for capturing the role of the microbiome in the selection of antibiotic resistance and highlights the role of treatment history for the prevalence of resistance.

Transposon-insertion sequencing screens unveil requirements for EHEC growth and intestinal colonization.

Enterohemorrhagic Escherichia coli O157:H7 (EHEC) is an important food-borne pathogen that colonizes the colon. Transposon-insertion sequencing (TIS) was used to identify genes required for EHEC and E. coli K-12 growth in vitro and for EHEC growth in vivo in the infant rabbit colon. Surprisingly, many conserved loci contribute to EHEC's but not to K-12's growth in vitro. There was a restrictive bottleneck for EHEC colonization of the rabbit colon, which complicated identification of EHEC genes facilitating growth in vivo. Both a refined version of an existing analytic framework as well as PCA-based analysis were used to compensate for the effects of the infection bottleneck. These analyses confirmed that the EHEC LEE-encoded type III secretion apparatus is required for growth in vivo and revealed that only a few effectors are critical for in vivo fitness. Over 200 mutants not previously associated with EHEC survival/growth in vivo also appeared attenuated in vivo, and a subset of these putative in vivo fitness factors were validated. Some were found to contribute to efficient type-three secretion while others, including tatABC, oxyR, envC, acrAB, and cvpA, promote EHEC resistance to host-derived stresses. cvpA is also required for intestinal growth of several other enteric pathogens, and proved to be required for EHEC, Vibrio cholerae and Vibrio parahaemolyticus resistance to the bile salt deoxycholate, highlighting the important role of this previously uncharacterized protein in pathogen survival. Collectively, our findings provide a comprehensive framework for understanding EHEC growth in the intestine.

Drug-target binding quantitatively predicts optimal antibiotic dose levels in quinolones.

Antibiotic resistance is rising and we urgently need to gain a better quantitative understanding of how antibiotics act, which in turn would also speed up the development of new antibiotics. Here, we describe a computational model (COMBAT-COmputational Model of Bacterial Antibiotic Target-binding) that can quantitatively predict antibiotic dose-response relationships. Our goal is dual: We address a fundamental biological question and investigate how drug-target binding shapes antibiotic action. We also create a tool that can predict antibiotic efficacy a priori. COMBAT requires measurable biochemical parameters of drug-target interaction and can be directly fitted to time-kill curves. As a proof-of-concept, we first investigate the utility of COMBAT with antibiotics belonging to the widely used quinolone class. COMBAT can predict antibiotic efficacy in clinical isolates for quinolones from drug affinity (R2>0.9). To further challenge our approach, we also do the reverse: estimate the magnitude of changes in drug-target binding based on antibiotic dose-response curves. We overexpress target molecules to infer changes in antibiotic-target binding from changes in antimicrobial efficacy of ciprofloxacin with 92-94% accuracy. To test the generality of our approach, we use the beta-lactam ampicillin to predict target molecule occupancy at MIC from antimicrobial action with 90% accuracy. Finally, we apply COMBAT to predict antibiotic concentrations that can select for resistance due to novel resistance mutations. Using ciprofloxacin and ampicillin as well defined test cases, our work demonstrates that drug-target binding is a major predictor of bacterial responses to antibiotics. This is surprising because antibiotic action involves many additional effects downstream of drug-target binding. In addition, COMBAT provides a framework to inform optimal antibiotic dose levels that maximize efficacy and minimize the rise of resistant mutants.

Multi-scale modeling of drug binding kinetics to predict drug efficacy.

Optimizing drug therapies for any disease requires a solid understanding of pharmacokinetics (the drug concentration at a given time point in different body compartments) and pharmacodynamics (the effect a drug has at a given concentration). Mathematical models are frequently used to infer drug concentrations over time based on infrequent sampling and/or in inaccessible body compartments. Models are also used to translate drug action from in vitro to in vivo conditions or from animal models to human patients. Recently, mathematical models that incorporate drug-target binding and subsequent downstream responses have been shown to advance our understanding and increase predictive power of drug efficacy predictions. We here discuss current approaches of modeling drug binding kinetics that aim at improving model-based drug development in the future. This in turn might aid in reducing the large number of failed clinical trials.

Deciphering the landscape of host barriers to Listeria monocytogenes infection.

Listeria monocytogenes is a common food-borne pathogen that can disseminate from the intestine and infect multiple organs. Here, we used sequence tag-based analysis of microbial populations (STAMP) to investigate Lmonocytogenes population dynamics during infection. We created a genetically barcoded library of murinized Lmonocytogenes and then used deep sequencing to track the pathogen's dissemination routes and quantify its founding population (Nb) sizes in different organs. We found that the pathogen disseminates from the gastrointestinal tract to distal sites through multiple independent routes and that Nb sizes vary greatly among tissues, indicative of diverse host barriers to infection. Unexpectedly, comparative analyses of sequence tags revealed that fecally excreted organisms are largely derived from the very small number of L. monocytogenes cells that colonize the gallbladder. Immune depletion studies suggest that distinct innate immune cells restrict the pathogen's capacity to establish replicative niches in the spleen and liver. Finally, studies in germ-free mice suggest that the microbiota plays a critical role in the development of the splenic, but not the hepatic, barriers that prevent L. monocytogenes from seeding these organs. Collectively, these observations illustrate the potency of the STAMP approach to decipher the impact of host factors on population dynamics of pathogens during infection.

Genetic analysis of Vibrio parahaemolyticus intestinal colonization.

Vibrio parahaemolyticus is the most common cause of seafood-borne gastroenteritis worldwide and a blight on global aquaculture. This organism requires a horizontally acquired type III secretion system (T3SS2) to infect the small intestine, but knowledge of additional factors that underlie V. parahaemolyticus pathogenicity is limited. We used transposon-insertion sequencing to screen for genes that contribute to viability of V. parahaemolyticus in vitro and in the mammalian intestine. Our analysis enumerated and controlled for the host infection bottleneck, enabling robust assessment of genetic contributions to in vivo fitness. We identified genes that contribute to V. parahaemolyticus colonization of the intestine independent of known virulence mechanisms in addition to uncharacterized components of T3SS2. Our study revealed that toxR, an ancestral locus in Vibrio species, is required for V. parahaemolyticus fitness in vivo and for induction of T3SS2 gene expression. The regulatory mechanism by which V. parahaemolyticus ToxR activates expression of T3SS2 resembles Vibrio cholerae ToxR regulation of distinct virulence elements acquired via lateral gene transfer. Thus, disparate horizontally acquired virulence systems have been placed under the control of this ancestral transcription factor across independently evolved human pathogens.

Reaction Kinetic Models of Antibiotic Heteroresistance.

Bacterial heteroresistance (i.e., the co-existence of several subpopulations with different antibiotic susceptibilities) can delay the clearance of bacteria even with long antibiotic exposure. Some proposed mechanisms have been successfully described with mathematical models of drug-target binding where the mechanism's downstream of drug-target binding are not explicitly modeled and subsumed in an empirical function, connecting target occupancy to antibiotic action. However, with current approaches it is difficult to model mechanisms that involve multi-step reactions that lead to bacterial killing. Here, we have a dual aim: first, to establish pharmacodynamic models that include multi-step reaction pathways, and second, to model heteroresistance and investigate which molecular heterogeneities can lead to delayed bacterial killing. We show that simulations based on Gillespie algorithms, which have been employed to model reaction kinetics for decades, can be useful tools to model antibiotic action via multi-step reactions. We highlight the strengths and weaknesses of current models and Gillespie simulations. Finally, we show that in our models, slight normally distributed variances in the rates of any event leading to bacterial death can (depending on parameter choices) lead to delayed bacterial killing (i.e., heteroresistance). This means that a slowly declining residual bacterial population due to heteroresistance is most likely the default scenario and should be taken into account when planning treatment length.

Sequence tag-based analysis of microbial population dynamics.

We describe sequence tag-based analysis of microbial populations (STAMP) for characterization of pathogen population dynamics during infection. STAMP analyzes the frequency changes of genetically 'barcoded' organisms to quantify population bottlenecks and infer the founding population size. Analyses of intraintestinal Vibrio cholerae revealed infection-stage and region-specific host barriers to infection and showed unexpected V. cholerae migration counter to intestinal flow. STAMP provides a robust, widely applicable analytical framework for high-confidence characterization of in vivo microbial dissemination.

Using Chemical Reaction Kinetics to Predict Optimal Antibiotic Treatment Strategies.

Identifying optimal dosing of antibiotics has proven challenging-some antibiotics are most effective when they are administered periodically at high doses, while others work best when minimizing concentration fluctuations. Mechanistic explanations for why antibiotics differ in their optimal dosing are lacking, limiting our ability to predict optimal therapy and leading to long and costly experiments. We use mathematical models that describe both bacterial growth and intracellular antibiotic-target binding to investigate the effects of fluctuating antibiotic concentrations on individual bacterial cells and bacterial populations. We show that physicochemical parameters, e.g. the rate of drug transmembrane diffusion and the antibiotic-target complex half-life are sufficient to explain which treatment strategy is most effective. If the drug-target complex dissociates rapidly, the antibiotic must be kept constantly at a concentration that prevents bacterial replication. If antibiotics cross bacterial cell envelopes slowly to reach their target, there is a delay in the onset of action that may be reduced by increasing initial antibiotic concentration. Finally, slow drug-target dissociation and slow diffusion out of cells act to prolong antibiotic effects, thereby allowing for less frequent dosing. Our model can be used as a tool in the rational design of treatment for bacterial infections. It is easily adaptable to other biological systems, e.g. HIV, malaria and cancer, where the effects of physiological fluctuations of drug concentration are also poorly understood.

The Role of Adherence and Retreatment in De Novo Emergence of MDR-TB.

Treatment failure after therapy of pulmonary tuberculosis (TB) infections is an important challenge, especially when it coincides with de novo emergence of multi-drug-resistant TB (MDR-TB). We seek to explore possible causes why MDR-TB has been found to occur much more often in patients with a history of previous treatment. We develop a mathematical model of the replication of Mycobacterium tuberculosis within a patient reflecting the compartments of macrophages, granulomas, and open cavities as well as parameterizing the effects of drugs on the pathogen dynamics in these compartments. We use this model to study the influence of patient adherence to therapy and of common retreatment regimens on treatment outcome. As expected, the simulations show that treatment success increases with increasing adherence. However, treatment occasionally fails even under perfect adherence due to interpatient variability in pharmacological parameters. The risk of generating MDR de novo is highest between 40% and 80% adherence. Importantly, our simulations highlight the double-edged effect of retreatment: On the one hand, the recommended retreatment regimen increases the overall success rate compared to re-treating with the initial regimen. On the other hand, it increases the probability to accumulate more resistant genotypes. We conclude that treatment adherence is a key factor for a positive outcome, and that screening for resistant strains is advisable after treatment failure or relapse.

Fitness Costs of Drug Resistance Mutations in Multidrug-Resistant Mycobacterium tuberculosis: A Household-Based Case-Control Study.

BACKGROUND: The projected long-term prevalence of multidrug-resistant (MDR) tuberculosis depends upon the relative fitness of MDR Mycobacterium tuberculosis strains, compared with non-MDR strains. While many experimental models have tested the in vitro or in vivo fitness costs of various drug resistance mutations, fewer epidemiologic studies have attempted to validate these experimental findings. METHODS: We performed a case-control study comparing drug resistance-associated mutations from MDR M. tuberculosis strains causing multiple cases in a household to matched MDR strains without evidence of secondary household cases. RESULTS: Eighty-eight multiple-case and 88 single-case household MDR strains were analyzed for 10 specific drug resistance-associated polymorphisms previously associated with fitness effects. We found that the isoniazid-resistant katG Ser315Thr mutation occurred more than twice as frequently in multiple-case households than in single-case households (odds ratio [OR], 2.39; 95% confidence interval [CI], 1.21-4.70), corroborating previous experimental findings. However, strains carrying both the katG Ser315Thr mutation and the rpsL Lys43Arg mutation were less likely to be found in multiple-case households (OR, 0.09; 95% CI, .01-.73), suggesting a negative epistatic interaction which contrasts previous findings. CONCLUSIONS: The case-control design presents a useful approach for assessing in vivo fitness effects of drug resistance mutations.

Cycling empirical antibiotic therapy in hospitals: meta-analysis and models.

The rise of resistance together with the shortage of new broad-spectrum antibiotics underlines the urgency of optimizing the use of available drugs to minimize disease burden. Theoretical studies suggest that coordinating empirical usage of antibiotics in a hospital ward can contain the spread of resistance. However, theoretical and clinical studies came to different conclusions regarding the usefulness of rotating first-line therapy (cycling). Here, we performed a quantitative pathogen-specific meta-analysis of clinical studies comparing cycling to standard practice. We searched PubMed and Google Scholar and identified 46 clinical studies addressing the effect of cycling on nosocomial infections, of which 11 met our selection criteria. We employed a method for multivariate meta-analysis using incidence rates as endpoints and find that cycling reduced the incidence rate/1000 patient days of both total infections by 4.95 [9.43-0.48] and resistant infections by 7.2 [14.00-0.44]. This positive effect was observed in most pathogens despite a large variance between individual species. Our findings remain robust in uni- and multivariate metaregressions. We used theoretical models that reflect various infections and hospital settings to compare cycling to random assignment to different drugs (mixing). We make the realistic assumption that therapy is changed when first line treatment is ineffective, which we call "adjustable cycling/mixing". In concordance with earlier theoretical studies, we find that in strict regimens, cycling is detrimental. However, in adjustable regimens single resistance is suppressed and cycling is successful in most settings. Both a meta-regression and our theoretical model indicate that "adjustable cycling" is especially useful to suppress emergence of multiple resistance. While our model predicts that cycling periods of one month perform well, we expect that too long cycling periods are detrimental. Our results suggest that "adjustable cycling" suppresses multiple resistance and warrants further investigations that allow comparing various diseases and hospital settings.

Smear positivity in paediatric and adult tuberculosis: systematic review and meta-analysis.

BACKGROUND: Tuberculosis (TB) diagnosis continues to rely on sputum smear microscopy in many settings. We conducted a meta-analysis to estimate the percentage of children and adults with tuberculosis that are sputum smear positive. METHODS: We searched PubMed, MEDLINE, Embase, and Global Health databases for studies that included both children and adults with all forms of active TB. The pooled percentages of children and adults with smear positive TB were estimated using the inverse variance heterogeneity model. This review was registered in the PROSPERO database under registration number CRD42015015331. RESULTS: We identified 20 studies meeting our inclusion criteria that reported smear positivity for a total of 18,316 children and 162,574 adults from 14 countries. The pooled percentage of paediatric TB cases that were sputum smear positive was 6.8 % (95 % Confidence Interval (CI) 2.2-12.2 %), compared with 52.0 % (95 % CI 40.0-64.0 %) among adult cases. Eight studies reported data separately for children aged 0-4 and 5-14. The percentage of children aged 0-4 that were smear positive was 0.5 % (95 % CI 0.0-1.9 %), compared with 14.0 % (95 % CI 8.9-19.4 %) among children aged 5-14. CONCLUSIONS: Children, especially those aged 0-4, are much less likely to be sputum smear positive than adults. National TB programs relying on sputum smear for diagnosis are at risk of under-diagnosing and underestimating the burden of TB in children.

Bi-modal distribution of the second messenger c-di-GMP controls cell fate and asymmetry during the caulobacter cell cycle.

Many bacteria mediate important life-style decisions by varying levels of the second messenger c-di-GMP. Behavioral transitions result from the coordination of complex cellular processes such as motility, surface adherence or the production of virulence factors and toxins. While the regulatory mechanisms responsible for these processes have been elucidated in some cases, the global pleiotropic effects of c-di-GMP are poorly understood, primarily because c-di-GMP networks are inherently complex in most bacteria. Moreover, the quantitative relationships between cellular c-di-GMP levels and c-di-GMP dependent phenotypes are largely unknown. Here, we dissect the c-di-GMP network of Caulobacter crescentus to establish a global and quantitative view of c-di-GMP dependent processes in this organism. A genetic approach that gradually reduced the number of diguanylate cyclases identified novel c-di-GMP dependent cellular processes and unraveled c-di-GMP as an essential component of C. crescentus cell polarity and its bimodal life cycle. By varying cellular c-di-GMP concentrations, we determined dose response curves for individual c-di-GMP-dependent processes. Relating these values to c-di-GMP levels modeled for single cells progressing through the cell cycle sets a quantitative frame for the successive activation of c-di-GMP dependent processes during the C. crescentus life cycle. By reconstructing a simplified c-di-GMP network in a strain devoid of c-di-GMP we defined the minimal requirements for the oscillation of c-di-GMP levels during the C. crescentus cell cycle. Finally, we show that although all c-di-GMP dependent cellular processes were qualitatively restored by artificially adjusting c-di-GMP levels with a heterologous diguanylate cyclase, much higher levels of the second messenger are required under these conditions as compared to the contribution of homologous c-di-GMP metabolizing enzymes. These experiments suggest that a common c-di-GMP pool cannot fully explain spatiotemporal regulation by c-di-GMP in C. crescentus and that individual enzymes preferentially regulate specific phenotypes during the cell cycle.

Antagonism between bacteriostatic and bactericidal antibiotics is prevalent.

Combination therapy is rarely used to counter the evolution of resistance in bacterial infections. Expansion of the use of combination therapy requires knowledge of how drugs interact at inhibitory concentrations. More than 50 years ago, it was noted that, if bactericidal drugs are most potent with actively dividing cells, then the inhibition of growth induced by a bacteriostatic drug should result in an overall reduction of efficacy when the drug is used in combination with a bactericidal drug. Our goal here was to investigate this hypothesis systematically. We first constructed time-kill curves using five different antibiotics at clinically relevant concentrations, and we observed antagonism between bactericidal and bacteriostatic drugs. We extended our investigation by performing a screen of pairwise combinations of 21 different antibiotics at subinhibitory concentrations, and we found that strong antagonistic interactions were enriched significantly among combinations of bacteriostatic and bactericidal drugs. Finally, since our hypothesis relies on phenotypic effects produced by different drug classes, we recreated these experiments in a microfluidic device and performed time-lapse microscopy to directly observe and quantify the growth and division of individual cells with controlled antibiotic concentrations. While our single-cell observations supported the antagonism between bacteriostatic and bactericidal drugs, they revealed an unexpected variety of cellular responses to antagonistic drug combinations, suggesting that multiple mechanisms underlie the interactions.

The path of least resistance: aggressive or moderate treatment?

The evolution of resistance to antimicrobial chemotherapy is a major and growing cause of human mortality and morbidity. Comparatively little attention has been paid to how different patient treatment strategies shape the evolution of resistance. In particular, it is not clear whether treating individual patients aggressively with high drug dosages and long treatment durations, or moderately with low dosages and short durations can better prevent the evolution and spread of drug resistance. Here, we summarize the very limited available empirical evidence across different pathogens and provide a conceptual framework describing the information required to effectively manage drug pressure to minimize resistance evolution.

On being the right size: the impact of population size and stochastic effects on the evolution of drug resistance in hospitals and the community.

The evolution of drug resistant bacteria is a severe public health problem, both in hospitals and in the community. Currently, some countries aim at concentrating highly specialized services in large hospitals in order to improve patient outcomes. Emergent resistant strains often originate in health care facilities, but it is unknown to what extent hospital size affects resistance evolution and the resulting spillover of hospital-associated pathogens to the community. We used two published datasets from the US and Ireland to investigate the effects of hospital size and controlled for several confounders such as antimicrobial usage, sampling frequency, mortality, disinfection and length of stay. The proportion of patients acquiring both sensitive and resistant infections in a hospital strongly correlated with hospital size. Moreover, we observe the same pattern for both the percentage of resistant infections and the increase of hospital-acquired infections over time. One interpretation of this pattern is that chance effects in small hospitals impede the spread of drug-resistance. To investigate to what extent the size distribution of hospitals can directly affect the prevalence of antibiotic resistance, we use a stochastic epidemiological model describing the spread of drug resistance in a hospital setting as well as the interaction between one or several hospitals and the community. We show that the level of drug resistance typically increases with population size: In small hospitals chance effects cause large fluctuations in pathogen population size or even extinctions, both of which impede the acquisition and spread of drug resistance. Finally, we show that indirect transmission via environmental reservoirs can reduce the effect of hospital size because the slow turnover in the environment can prevent extinction of resistant strains. This implies that reducing environmental transmission is especially important in small hospitals, because such a reduction not only reduces overall transmission but might also facilitate the extinction of resistant strains. Overall, our study shows that the distribution of hospital sizes is a crucial factor for the spread of drug resistance.