Modélisa Modélisation en tion en micr microbiolog obiologie ie Lulla Opatowski Pharmaco-Epidemiology and Infectious Diseases unit UMR1181 Institut Pasteur / Inserm / Université de Versailles St Quentin RICAI – Décembre 2019
Emergence, selection, spread of bacteria – a complex multi-scale process Between-host transmission (heterogeneous) Population Public health policies Colonization of ecosystems (gut, skin, naso-, oro-pharynx) Individuals Infection Immunity Drug exposure (vaccine, antibiotics) Genetic evolution : Microorganism Mutations, genes transfer Ecology of pathogens: Colonies Interactions , competition and synergy 2
Nosocomial outbreaks: a public health challenge • Hospitals: frequent outbreaks of different species – High density of individuals and contacts – High antibiotic exposure => high frequency of antibiotic resistant bacteria (ARB) – At-risk patients, eg. intensive care or post surgery – Specificities • Most cases are asymptomatic carriers ( low burden ) • >1 infected patient => critical – Efficient control may require the set up of heavy procedure • Closure of beds, wards etc. • Disorganization and important costs Controlling this dissemination in these • Assessment of benefit is complex settings requires understanding bacterial spread 3
Challenge • Knowledge gaps : Acquisition, transmission and persistence of bacteria, and in particular ARB Pathogen’s intrinsic epidemicity? Routes of acquisition? Factors of clearance? Microbiota, within-host pathogens interactions? Impact of antibiotic exposure? • Analyze human behaviors and microbiological data in an integrative manner => Set up ad-hoc epidemiological studies => development of biostatistics, mathematical modelling tools 4
Reference model for bacterial transmission Natural clearance 1 1 Colonized with a Force of infection S bacteria on day t : 0 Emergence ( p ) 𝜇 𝑢 = 𝛾(𝐷 𝑇 + 𝐷 𝑆 ) Contacts Susceptible 1 Colonized with a R bacteria ds dt = 0 − s ( c S + c R ) + ( c S + c R ) − 1 s dc S dt = + sc S − c S − pc S − 1 c S 1 homogeneous population dc R dt = + sc R − pc R − c R − 1 c R 5 1 drug, 1 simple mechanism => Useful but poorly representative => more specificity s + c S + c R = 1 is required
F ORMALIZING HETEROGENEITY - U SE OF CONTACT NETWORK IN HOSPITAL MODELS • Increasing number of models integrating networks • Different levels of contact networks: between-hospitals, between-wards transfer data from hospital database records, and inter-individual proximities from log sensors Ward 8 Ward 1 Ward 2 Ward 7 Ward 5 [Assab, Nekkab et al, Current Op Inf Dis 2017] 6
2 illustrations of modelling studies • Spread of bacteria in a hospital populations • Impact of host flora in bacterial acquisition 7
STORY 1 - Between human transmission of bacteria over a contact network
T HE I - BIRD STUDY I-BIRD study : « Individual-based Berck sur Mer Investigation of Resistance Long term care hospital Dissemination » 329 patients → Prospective longitudinal study 263 hospital staff → Principal investigator Didier Guillemot 5 months → European project MOSAR → June-October 2009 Daily Weekly Contact swab data Antibiotics N~6620 oropharyngeal rectal N~3500 Detection of all Close proximity interactions (CPI) every 30sec with a distance less than 1m50 Staphylococcus aureus Enterobacteriaceae 2 740 728 contacts 9
C LOSE PROXIMITY INTERACTIONS Frequency of contacts over a day Duration of contacts 10 [Duval A, et al. Scientific report 2018]
C LOSE PROXIMITY INTERACTIONS Frequency of contacts over a day Duration of contacts 11 [Duval A, et al. Scientific report 2018]
[Obadia et al, Plos Comp Biol 2015] [Duval et al, Plos Comp Biol 2019] 1. Gather network & swabs data: Newly colonized Assess the role of contact network Candidate Uncolonized on transmission Real data Randomly permuted data (null hyp) Potential Randomly pick up transmitter New acquisition New acquisition transmitter on contact network Comparaison ( Wilcoxon test) 12
S. aureus 1. Assess the role of contact network on transmission p-value < 0.001 Shorter distance found in observed data compare to permutated data Distance to the nearest potential transmitter ESBL K. pneumoniae ESBL Escherichia coli 35 acquisitions 20 acquisitions p-value = 0.25 p-value = 0.025 S. aureus and K. pneumoniae transmission supported by the proximity network Not clear for E. coli , probably other important sources [Obadia et al, Plos Comp Biol 2015] 13 [Duval et al, Plos Comp Biol 2019]
2. Estimating S. aureus strain’s epidemicity over the network Contact data, Likelihood-based parameter swabs, antibiotics estimation … MCMC Transmission rate per contact.day for the different strains 14
Step 1: Identify potential transmitters and acquirers • Transmitters P 0 : colonized individuals having recorded contacts • Potential acquirers: individuals at risk of acquisition • Recorded contacts with P 0 over [t 0 , t n ] • A negative swab over the period or before P0 P0 P0 P0 t n : last positive t 0 : first positive swab (strain s ) swab (strain s ) Negative swab Recorded proximity 15 Positive swab
Step 2: Transmission model for strain s For any non-colonized individual i , we define the force of acquisition on day t : 𝜇 𝑗 𝑢, 𝑡 = 𝛾(𝑦 𝑘 , 𝑦 𝑗 , 𝑢) 𝑘≠𝑗, 𝑡 𝑘 𝑢,𝑡 =1 𝑗∈𝐷 𝑘 (𝑢,𝑡) – x i , x j , denote for individuals i and j , the vectors of individual variables eg. status [HCW versus patient] – 𝛾(𝑦 𝑘 , 𝑦 𝑗 , 𝑢) , the probability of transmission of strain s from individual j to individual i on a given day of contact t – Case transmission network of j 𝑘 𝑢, 𝑡 = ቊ 𝑗, 𝑗, 𝑘 ∈ 𝐹 𝑢 , 𝑡 𝑗 𝑢, 𝑡 = 0 if 𝑢 ∈ [𝑢 1 , 𝑢 𝑜 ] 𝐷 ∅ otherwise 16
2. Estimating S. aureus strain’s epidemicity over the network MRSA MSSA => significantly higher epidemicity for MRSA t777 => More generally, higher epidemicity for MRSA 17 [Opatowski et al, in prep]
BACK TO MICROBIOLOGY → Some species spread over interindividual networks but lack of evidence for other (ESBL E. coli ) → In MRSA: different epidemicities according to spatype and resistance to methicillin • Strains’ characteristics conferring: – differences in modes of transmission? – differences de transmissibility between MRSA and MSSA for t777 ? 18
STORY 2 - Modelling within-host flora and microbial colonisation
1) A T THE HOST LEVEL - MODELLING BETWEEN - BACTERIA INTERACTIONS When longitudinal microbiota data is available … Dynamics of relative abundance of different species/groups in the microbial community Ι N : the number of different species 𝑦 𝑗 (𝑢): the abundance of species i at time t 𝑠 𝑗 : growth rate of species i (can be negative when receiving antibiotic or under immune response) 𝑏 𝑘𝑗 : interaction strength between species j and i (positive or negative) K : carrying capacity of the entire community [Xuefeng Gao et al, Frontiers in Microbiology 2018] 20
1) A T THE HOST LEVEL - MODELLING BETWEEN - BACTERIA INTERACTIONS When longitudinal microbiota data is available … [Xuefeng Gao et al, Frontiers in Microbiology 2018] 21
1) A T THE HOST LEVEL - MODELLING BETWEEN - BACTERIA INTERACTIONS When longitudinal microbiota data is available… Dynamics of relative abundance of different groups in the microbial community Ι Antibiotics Feeding Symptoms Colonization with ESBL (…) [Xuefeng Gao et al, Frontiers in Microbiology 2018] 22
AT THE POPULATION LEVEL Natural clearance 1 1 Colonized with a Force of infection S bacteria on day t : 0 Emergence ( p ) 𝜇 𝑢 = 𝛾(𝐷 𝑇 + 𝐷 𝑆 ) Contacts Susceptible 1 Colonized with a R bacteria ds dt = 0 − s ( c S + c R ) + ( c S + c R ) − 1 s dc S dt = + sc S − c S − pc S − 1 c S 1 homogeneous population dc R dt = + sc R − pc R − c R − 1 c R 23 1 species 1 drug s + c S + c R = 1
A SIMPLE MODEL Hospital FOR MICROBIOME - High-D flora (H) Low-D flora (L) PATHOGEN CO - No pathogen 1 COLONIZATION • Compartmental ODE 2 model • Patient population in healthcare setting 3 4 3 4 • Antibiotic use ( q a ) • Proportion of patients receiving antibiotics Pathogen (P) 1 • Antibiotic resistance ( q r ) • Proportion of 2 antibiotics against which pathogen resistant 1. Flora disruption ( 𝜏 ) 4. Pathogen clearance ( 𝛿 + 𝜏 ) 2. Flora recovery ( 𝜀 ) [Smith et al, in prep] 3. Pathogen acquisition ( 𝛾 + 𝛽 )
Recommend
More recommend
