Sunday, 17 August 2014

Ebola, pigs, primates and people

This is a companion piece to my collaborative article, Ebola virus may be spread by droplets, but not by an airborne route: what that means, posted a couple of days ago. I suggest you read the both together.

In this post, I'd like to make sure we all understand that an airborne route of Ebola virus infection has been used to deliberately infect non-human primates (NHPs). It is possible and it can be done. Okay? I'm not covering up any secret knowledge or trying to conceal facts that only we few evil-society-of-science types know. I don't secretly work for an agency aiming to delude you dear readers into feeling falsely safe about the risks associated with being near an Ebola virus infected person (which most reading this will likely never be). Frankly, I'm learning this as I go.

Don't expect perfection from risk mitigation advice.

Like all things that involve biology, there are hardly ever clear-cut lines and yes or  no definitions and explanations. Sometimes that's because things vary...because biology! Sometimes that's because we haven't yet done enough science to know those answers. I'm not an expert on ebolaviruses nor on Ebola virus disease (EVD) - but in my time learning about the viruses and the disease, its clear that this is (yet another) area that is lacking in all sorts of information. So risks are judged using what we do know and can support and verify, with softer language used when decision makers don't know for certain; less so when they think they do. 

When that message of risk gets passed to the public, it is important to be accurate, clear, concise but not to over-simplify things because that may degrade trust in the body(s) sending the message if things change later. That's a very tough balance when dealing with biological risks.

So having said that, let's talk pigs.

Pigs are not primates.

In ebolaville - the virtual world created by social and mainstream media stories and discussion about ebolaviruses - a lot of people have been throwing the 2012 pig to macaque study (8) around as an argument for why we should admit that ebolaviruses spread by an airborne route and run for the hills. This is why that is not a good comparison:
  • Pigs have a different disease and replication process to humans. 
    • Pigs tend to have much more virus growing in their lungs.(12) 
    • Pigs tend to cough and sneeze and generally propel more of said pathogen from their lungs.(11) 
  • Pigs may eject more infectious viruses in their droplets than do primates
If we look at the study of disease occurrence and spread in previous outbreaks, that epidemiology does not suggest an airborne spread - the numbers and nature of human-to-human spread don;t show it as any sort of major contributor to spread. Might it be a minor contributor? Possibly. We don;t know either way with 100% surety. But we do  know some other things. One of these is that it takes very little virus to infect pigs and NHPs. If there are not obvious signs of an airborne spread in humans, we just not have detected it yet or, it may have a biological basis. It is possible that the infectious dose (amount of virus needed to get a foothold and start an infection) may be much higher for humans; infected and severely ill human cases may not breathe out infectious virus or ebolaviruses may not survive for long in the aerosols expired by humans,(19) even if they can survive on hard surfaces or in generated aerosols under laboratory conditions.(20)

Non-human primates can be infected with ebolaviruses via a lab-made aerosol with lots of lab virus at lab temperatures and lab humidity and other lab conditions in a lab.

You get that this is done in a lab? Cool.

It apparently does not take much virus to infect a human via an aerosol according to the Public Health Agency of Canada's (PHAC) Pathogen Safety Data Sheet (PSDS) on ebolavirus.(1) Only 1-10 infectious organisms (see above). But one problem with that PSDS is that it cites only 1 paper to support that range. Ref 21 from the PSDS is entitled Clinical recognition and management of patients exposed to biological warfare agents.(2) It is 1997 review that does not specify if this range is specific to any 1 or more of the ebolaviruses, just "viral hemorrhagic fevers". The PSDS seems to rely on that 1 line. In that reference, there are no further links to studies that define this range for humans, ebolaviruses or an Ebola virus (EBOV) of the species Zaire ebolavirus. I've sent a couple of emails in the past week, seeking further clarification from ebolavirus experts, but have yet to hear anything back.

In a laboratory experiment reported in 1995, transmission of an EBOV from one set NHPs infected by injection, to another set  resulted in 2 of 3 NHPs (1 with a heavy load of virus in the lung) becoming infected and that seemed to have occurred through some kind of airborne route as the 2 groups of animals were separated by 3m and care was taken to avoid creating bigger droplets and splashes during cage cleaning.(15) While the authors noted that fomites (contaminated objects and surfaces) or contact droplet transmission of virus was unlikely, the exact mode of transmission to the second group of NHPs could not be determined. In a follow-up study, the authors were able to prove that conjunctival and oral exposure to an EBOV could indeed result in infection in NHPs.(18) Thus we have plenty of reason for the use of masks, goggles and face shields that are already part of the recommended personal protective equipment (PPE) items for dealing with infected humans.

However, there are a number of issues related to forced aerosol infection of NHPs, many of which can be found in a massive and detailed 2008 review by Dr. Jens Kuhn.(3) These include:
  • Often unrealistically high viral loads - the exact amount of infectious virus humans are exposed to during outbreaks has not been defined.
  • Temperature and humidity conditions that were unlikely to reflect conditions during outbreaks in Africa - but may reflect conditions in hospitals.
  • An initially lung-focussed pattern of viral replication (7) results from direct aerosol delivery of virus to NHP airways which seems different to infection of humans via the more frequent natural direct contact route. Systemic spread to multiple organs then follows via infected dendritic cells and macrophages and blood monocytes.
  • Different routes of virus acquisition can lead to different incubation periods.
  • Different virus isolates, sources and preparations may affect the course of infection and disease
  • Because of the small and enclosed space and air throughput in head-only chambers, droplets rather than droplet nuclei may be the vehicle carrying infectious virus. This is important because, as you can read in the companion piece, droplet nuclei are the component of a lingering "airborne route" of acquisition and if NHPs are in fact infected by the droplets, that may be more indicative of direct fluid contact than true airborne travel.
A head-only inhalation chamber of the sort used in NHP
aerosol inoculation studies. Biaera Technologies.
Image from
. See also (17)

Click on image to enlarge.
Some NHP studies that have successfully caused initial respiratory infection using an airborne route to infect NHPs under controlled experimental conditions include the following:
  • 1,000 plaque forming units (PFU; a measure of how much virus is in a preparation using cell culture methods in the lab) of either a Kikwit EBOV isolate or a Boniface isolate of Sudan virus (SUDV; species Sudan ebolavirus) isolate were delivered using a Collison nebulizer (producing small droplets) after intramuscular immunization with a recombinant adenovirus vaccine.(5) 
  • 1,000 PFU of a Kikwit EBOV isolate was delivered with a Collision nebulizer via a head-only aerosol chamber, after intramuscular immunization with a recombinant vesicular stomatitis virus (VSV) vaccine.(6)
  • 743-274,000 PFU of a Kikwit EBOV isolate was delivered to with a Collision nebulizer via a head-only aerosol chamber, to examine aerosol-related pathology.(7)
  • ~50 or ~500 PFU of a Boniface SUDV isolate were delivered to 3 different NHP species using a Collison nebulizer via a head-only chamber to compare species-specific effects.(14) 
  • 0.8-128 PFU of a Kikwit EBOV isolate was delivered to 3 different NHP species using a Collision nebulizer via a head-only aerosol chamber, to examine disease course between species.(9)
  • ~300-50,000 PFU of an EBOV isolate was delivered to with a Collision nebulizer (0.8-1.2um droplets) via a head-only aerosol chamber, to examine aerosol-related pathology.(16)
Are primates humans?

Judging by the effort we put into getting rid of our fur compared to an NHP, I'd say we're not! 

But on the topic of EVD, some NHPs that we infect with an ebolavirus, show very similar disease signs, symptoms and disease progression to those of EVD in humans; especially rhesus macaques [Macaca mulatta] although oneo f the studies above showed that 3 different NHP species were not that different in the way they responded to infection (rhesus macaques as well as cynomolgous macaques [Macaca fascicularis] and African green monkeys [Chlorocebus aethiops]).(9) 

Rhesus macaques become febrile, anorexic, lethargic, viraemic, develop a rash and sometimes develop diarrhoea and melena (gastrointestinal bleeding).(3)

But no animal model seems to completely capture other components of human disease which have historically included conjunctivitis, diarrhoea and vomiting and coughing up blood. Vomiting up blood and having bleeding gums occurs more often in fatal cases than in survivors.(3) 

Bleeding only occurred in 41% of 103 observed human patients during the 1995 Kikwit outbreak of an EBOV.(3)

So the answer is, primates are not humans when it comes to EVD, but they are pretty close. Yet within that "pretty close" lies an immeasurable amount of variation that may mislead when trying to map the course of NHP disease onto that of humans.

Where does that leave us?

I admit to being very uneasy saying that there is no risk at all of an airborne route of ebolavirus infection. Clearly it can be forced to happen, but we have no evidence that it has ever happened in humans in an outbreak. But let's put that into context. An absence of evidence is not evidence of absence. Outbreaks of ebolaviruses are not particularly conducive to large careful research projects measuring infectious droplet nuclei around critically ill people, especially when the occur in exotic locations in someone else's back yard.

So have I deserted by position from yesterday's post stating no airborne role for ebolavirus transmission between humans? No, not at all. What we know is that the overwhelming majority of human EVD cases acquire their infection during the time they are in direct contact with the fluids of a very ill EVD case; be that through physical contact or wet droplet spray impact. Beyond that fact, it may just be a discussion based on academic musings and hand-waving. But it is a discussion we should be having a little more I feel. A back-and-forth rather than messages with guarantees and statements dealing in black and white absolutes. I'm not sure the public believe in or feel safer with such absolutes today. We're all a bit too cynical for that.

If infection can happen between primates via the air, it is a very, very inefficient process as a study of 78 people from 27 households with EVD cases during the 1995 Kikwit  revealed.(10) Those 78 household members had no physical contact with the cases, and they did not get sick. Others who had physical contact, got EVD. 

In a recent study by the authors of the 2012 pig/macaque study we started this post with, infected NHPs did not pass EBOV to uninfected NHPs only 30cm away.[21] Not only was there no disease in the inoculated animals but no antibodies were detectable in the uninfected NHPs 4-weeks later. There had been no infection at all.

While at some point we'll need to be more sure of all this for humans than we are now, we can say that pigs aren't primates and airborne route has not been shown to be a risk for human acquisition of an EBOV.