Best Methods for Studying Droplet Spray Transmission

							        Breakout Group:
Best Methods for Studying Droplet
      Spray Transmission




       “Understanding the Modes of Influenza Transmission” Workshop
       November 4 , 2010 and November 5, 2010 – Atlanta, GA
What are the key questions/ gaps
 remaining in understanding the
      contribution of droplet
  transmission to the spread of
    influenza among humans?
         Droplet Transmission Breakout Group:
            Remaining Key Questions/Gaps
   There are no data discriminating relative contribution of spray
    compared to other potential modes of transmission
        Use of masks versus respirators cannot distinguish between these modes because
           •   The conjunctiva are exposed which may or may not be a route of infection
           •   Masks differ widely by their ability to filter some smaller particles
        People touch their faces many times in an hour and so difficult to exclude indirect
         contact transmission in non-experimental model
   The experimental models that mimic human generation of influenza
    droplets and aerosols may not be valid in
        Methods that generation propulsion of fluids/aerosols may
           •   Alter viability of virus
           •   Not be designed to capture particils of all size, especially large sizes
           •   Use an artificial matrix or have missing cofactors that alter viability and/or infectivity of the
               viruses or the particle size distribution

   Size distribution of particles in natural infection not well characterized
        Needs to be determined for different states (e.g. tidal breathing, talking, coughing,
         singing)
        More information is needed to understand person to person variability in particle size
         distribution
         Droplet Transmission Breakout Group:
            Remaining Key Questions/Gaps
   Need to characterize deposition of droplet spray on different sites of
    infection, and infectious dose by site of deposition
        What sites of deposition of droplets (eyes, nose, etc) will most likely results in
         infection (efficiency of infection by site of deposition)
        During droplet spray, what proportion of droplets is distributed at different sites of
         the body that might lead to human infection
        These data could be applied to a model such as that developed by Nicas, et al.
   Better understanding of correlation between viral copies as measured
    by rt-PCR compared to live virus and how this relates to potential
    infectivity
   Better understanding of the virus dose needed by route of infection
        Human experimental models did not look at more than one route of transmission in
         the same experiment
        Relatively low secondary attack rate and Ro for influenza inconsistent with report of
         very small inoculum requirement reported for aerosol transmission in studies such as
         Alford, et al
   Variability in viral shedding among individuals and pre-existing cross-
    reactive immunity within households/populations difficult to assess and
    control for in population-based studies
What are the best study designs
     and their pros/cons?

What study designs would be best
for understanding the contribution
  of droplet transmission to the
    other transmission routes?
    Droplet Transmission Breakout Group: Best
           Study Design and Pros/Cons
   Considerations for studying the relative contribution of droplet spray
        Because spray implies a ballistic event, issues of temperature and humidity in the
         environment should be less of an issue
           •   However, humidity and temperature effects on the host may also affect susceptibility by this
               route as for other transmission routes
        Study may be less complicated compared to aerosol or contact since the virus from
         droplets is expected to be outside of the infected person for much less than a second
         before it reaches the new host
   Studies to assess the distribution of influenza virus from droplet spray
        Persons infected with influenza (preferentially persons naturally infected ) would
           •   Cough, sneeze and/or speak toward an artificial target (e.g. mannequin, splash plates, growth
               media)
                 o  Assess distribution of amount of virus hitting mannequin targets that may lead to
                    infection (e.g. lips, conjunctiva, mouth, nasal mucosa)
        This would allow for better understanding of the distribution of large droplets onto a
         target
   Human experimental studies that would eliminate different modes of
    transmission
        Assess risk of human to human transmission using methods to eliminate droplet
         spray, e.g.
           •   Face shield protecting eyes, nose and mouth on potential recipient which prevents only spray,
               but not aerosol
           •   Using UV light in the room when exposure occurs ( to eliminate aerosol mode)
           •   Using a gentle crosswind between the infected and exposed persons to eliminate aerosol
    Droplet Transmission Breakout Group: Best
           Study Design and Pros/Cons
   Community studies of influenza
        Use face shields instead of masks for intervention group and compare to a control
         group
        Prospective design would help to eliminate delay in initiating intervention
           •   E.g. this is has been a difficulty with community studies that enrolled an index case after
               influenza diagnosis and then sought to establish different control measures among house hold
               members, may of whom were already exposed to influenza from the index case prior to
               enrollment
   Comparison of the relative efficiency of transmission by deposition on different
    parts of face
        One of the gaps in knowledge is where influenza viruses may deposit on the face and
         what proportion of viruses expelled from an infected person land on a mucosal
         surface where infection may then result
        Concern expressed was that artificially made sprays or sprays onto a mannequin
         may not provide good comparison to natural infection
   Prospective surveillance for and identification of influenza-infected hospitalized
    patients and conduct follow-up of exposed persons to identify secondary cases
    and the types of exposures they had to the index cases
        Would require intensive monitoring of staff and their activities, use of PPE, amount of time spent
         with the index patient, particularly face-to-face time that may have allowed for droplet spray
        Conduct full genome sequencing of viruses from index case to any contacts that become infected
         to determine if viruses likely from the same source or from another exposure
   Observational studies of healthcare workers to determine the amount of time
    they are face to face with a patient and may be likely exposed to droplet sprays
    through patient coughing, talking or sneezing
          Pros and Cons of Different Methods

   Series of studies to estimate probability of exposure to droplets
    and types of exposures needed to cause infection
   Cost high especially for human experimental infection studies
   Ethical concerns for human inoculation/exposure studies
   Variability in results due to differences between viruses, host
    variability, pre-existing immunity among exposed persons, and
    routes of infection
   Variability in models and samplers used in terms of
      Ability to capture different sizes of particles
      Ability to detect live virus
   Sample sizes for human studies would be high
   Generalizability of findings
   Compliance with interventions and control for multiple exposures in
    community studies