Tim Cook Profile picture
Bath-based anaesthetist/intensivist. I study medical complications to improve safety. Own views except where I knick others’. Insta @timcooksnaps. Airway goblin

May 3, 2021, 17 tweets

A thread on what studying "AGPs" tells us about 'aerosols and droplets'

Focus on this paper
…-publications.onlinelibrary.wiley.com/doi/10.1111/an…

Done in Sydney
Hats off to @NWilson247 @EuanTovey
My role was microscopic

First: discussion should be 'what is the proportion of each in disease risk: not A vs B

The study builds on work by others esp Bristol AERATOR study which has examined designated 'aerosol generating procedures' in a laminar flow theatre

-intubation/extubation
-NIV/HFNO
-tracheostomy

& largely identified a LACK of aerosol

More to come on other designated AGPs

For balance here are some studies with slightly different conclusions
ccr.cicm.org.au/config/cicm-cc…

…-publications.onlinelibrary.wiley.com/doi/full/10.11…

Another important paper by @nate_gaeckle
Studied NIV/HFNO in a clean isolation chamber

showed
-no more aerosol than other oxygen delivery methods
-massive variation between individuals

atsjournals.org/doi/full/10.11…

Key in all these studies is conduct in ULTRAclean (HEPA filtered) environment

...needed to detect respiratory aerosols (signal) over non-respiratory (noise)

Note
-noise amplitude >> signal
-size distribution differs

(studies in non-clean settings can be ignored!)

...back to the paper at hand
…-publications.onlinelibrary.wiley.com/doi/10.1111/an…

Done in an ultraclean chamber

Key difference from other papers
-massive head sized cone
-high sampling rates
to try to capture ALL exhaled aerosol

Expired particles sampled

During respiratory ACTIVITIES
-breathing
-shouting
-exercise
-force expiration
-cough

With respiratory PROCEDURES
-nothing
-NIV or CPAP
-HFNO
-surgical mask

First key finding.
A modest increase in aerosols with respiratory THERAPIES during quiet breathing

BUT (big but) this disappeared & emissions fell with respiratory therapy during exertional respiratory activities

This fall in exhaled aerosols with NIV/CPAP/HFNO compared to without their use shows they are NOT AGPs

The fall likely due to
-PEEP reducing airway collapse and open/close cycling
-positive inward flow reducing egress

You may have picked it up already....
Yes coughing, exercise & forced expiration (each mimicking breathing patterns in illness) produce many-fold more aerosols than quiet breathing (log scale)

Respiratory ACTIVITY should be focus of concern
NOT respiratory PROCEDURES

For those who want to dive deep here are the particle size distributions (generally >95% <5micron) & the % of all particle volume in these small aerosols
…-publications.onlinelibrary.wiley.com/action/downloa…

Again - as per Gaeckle and others - there was a massive inter-individual variation

So what do we learn

1 NIV, CPAP, HFNO are not AGPs
2 Exertional activities (cough, fast deep breathing) generate up to 370-fold more aerosol than quiet breathing
3 This aerosol is respirable
4 It constitutes a significant volume of all expired volume
5 We need to rethink risk

We know some superspreader events create massive infection & can only be explained by aerosols

These likely combination of
-high viral load (disease time course)
-high risk person (high emitter/spreader)
-high risk activity (sing, shout etc)
-poor ventilation
-prolonged exposure

While droplets may be a high risk at short distance...

...the high production of respirable aerosols by exertional respiratory activities (common in illness) point to a prominent role of aerosol transmission at short distances

(& in some settings over long distances)

The other point I should’ve made. Even more important than an ultraclean setting..

....collaboration between aerosol/environment scientists & clinicians

Key to all the papers in the thread (& missing from many others)

Share this Scrolly Tale with your friends.

A Scrolly Tale is a new way to read Twitter threads with a more visually immersive experience.
Discover more beautiful Scrolly Tales like this.

Keep scrolling