Summary
Pathogens that can replicate and spread in the environment and transmit to humans are likely unique among biological threats in their potential to directly cause human extinction. These environment-to-human (‘E2H’) pathogens may be able to infect humans through multiple routes, including food, water, surfaces, and air. Airborne transmission is probably the most dangerous and hardest to combat pathway.
Other classes of biological threats, like pathogens that spread person-to-person or that can be released in the environment but with limited potential to replicate, can still cause massive amounts of harm, but are very unlikely to kill everyone.
Given their potential to cause human extinction, we believe the community of people working to mitigate existential biological risks should consider preventing and preparing to defend against environment-to-human (‘E2H’) pathogens a top priority.
Environment-to-human biothreats
Most of the biological weapons pursued by historical bioweapons programs fall into one of two categories: localized attacks, or transmissible pathogens. The first are things like anthrax, where the biological agent is released into the environment and victims are exposed by airborne particles or surface contaminants. The second category are adversarially released epidemic- or pandemic-causing agents, with the fastest spreading examples generally being respiratory viruses. These transmit person-to-person, allowing for a much larger reach than the site of their initial introduction.
While extremely dangerous, both of these types of attacks have some limits on their potency. Localized attacks using non-transmissible agents such as anthrax don’t replicate in the environment, so they don’t spread beyond the initial attack. Highly transmissible pathogens pose a risk of much wider spread, but it is still possible for most people to avoid infection through isolation. For these, the central challenge is protecting people when they have to share indoor spaces, such as essential workers who need to do in-person work to keep society functioning.
A pathogen that could replicate outside of a host, quickly spread in the environment, and transmit to humans through environmental aerosols, surface particles, and contaminated food or water would face neither of these limits. People who remained totally isolated could still get infected. Society would need to protect not just essential workers at their job, but everyone all the time. Even if you have great PPE, infection could be possible when one inevitably has to remove it to eat, drink, or sleep.
An example of such a pathogen could be mirror bacteria—a bacteria whose biomolecules have the opposite chirality as normal life, potentially giving it the ability to evade host immune defenses and spread across ecosystems without being checked by predation. This is why we and the scientific community now believe the creation of mirror life should not be pursued.
Surviving the release and spread of an E2H pathogen, especially if it got to high concentrations in outdoor air and other environmental mediums, would be extremely challenging. People would potentially have to spend most of their time in either extremely robust bioshelters or wearing high-grade PPE. While there are cleanrooms for advanced manufacturing, laboratories, and hospital isolation rooms that could provide protection, there are nowhere near enough to house a large share of society. Similarly, there is not enough high grade PPE for everyone.
Aerosols as the weak link
A mirror bacterium or other severe E2H pathogen could potentially transmit to humans through food, water, surface particles we accidentally ingest, wounds, or airborne particles. We have reliable methods for closing off most of these routes. However, we think airborne transmission is probably far and away the hardest exposure route to close off.
Even in developed countries with high water and food safety standards, the typical person might drink millions of live bacteria every day and eat at least as many[1]. Normally, almost none of these pose a risk of serious infection, so this doesn’t matter. But if a lethal pathogen came to occupy even a small share of the current bacterial ecosystem, this could be a major exposure route. Thankfully, standard practices like boiling water and thoroughly cooking food can reduce bacterial content by several orders of magnitude, likely making food and water safe even in a severe E2H outbreak[2]. These basic approaches are most effective if they are conducted in an otherwise low-contamination environment, i.e. where bacteria won’t be easily re-introduced from contaminated surfaces or airborne particles[3].
Even when instructed not to, people frequently touch their face[4], and since people’s hands can transfer a large share of bacteria on a surface they touch, it may be important to disinfect surfaces people regularly come in contact with. Most household surfaces usually have on the order of tens of bacteria per cm2, so getting a several order of magnitude reduction in surface concentrations may be important, but we expect this is technically feasible using widely available materials. Cleaning surfaces with salt water can reduce pathogen loads by two orders of magnitude or more, and people can make a more powerful disinfectant, hypochlorous acid (HOCl), using salt, batteries, and an electrode like the graphite in pencils[5]. It may not be possible to fully eliminate an E2H pathogen from a bioshelter, potentially allowing the population of pathogens to replicate on surfaces in the shelter. However, we expect that this is unlikely to be a major risk; most bacteria only survive but do not replicate on dry, nutrient-poor surfaces, so by default there would likely not be a large build-up of pathogens.
Given the ready availability of ways to reduce bacterial counts in food, water, and surfaces, the dominant exposure route would likely be from small particles that can remain suspended in air for hours or days (aerosols). While the variation between different environments is large, outdoor air often contains tens of thousands to millions of bacterial cells per m3, and indoor air can have similar or slightly lower concentrations. Since people breathe in close to 1 m3 per hour, our total inhalation of bacteria is very high, such that if even a small share of the normal bacterial population was able to evade our immune defenses and cause a serious infection, people would need high grade and very consistent protection to survive.
We cannot practically wear PPE 24/7, and buildings need to constantly bring in fresh outdoor air for us to breathe. This means that keeping a low airborne concentration of outdoor-originating pathogens is likely the biggest challenge in protecting humanity against a severe E2H outbreak.
Conclusion
While the top priority is to prevent the development of an E2H pathogen like mirror bacteria, we should aim for defense in depth and develop robust response plans if an E2H pathogen is released. Increasing the deployment of air cleaning technologies like air purifiers, glycol vapors, and (far) UVC. However, these may be insufficient for creating well-protected spaces, and deployment speeds may not be fast enough to counter a fast-spreading E2H pathogen if it is released soon[6]. In the next blog post, I outline how it may be possible for humanity to quickly scale up defenses against outdoor-originating aerosols with improvised bioshelters and PPE. However, making it likely that humanity implements these defenses will require a lot of work, and we’re excited for more people to build and test different protocols.
The number of bacteria one eats ranges by several orders of magnitude depending on diet, but is usually at least around one million colony forming units (CFUs) per day. US tap water is supposed to have <500 CFU/mL of culturable bacteria (typically having tens to low hundreds per mL), though limiting to culturable bacteria probably undercounts the total number, potentially by an order of magnitude. Since the typical person drinks a few liters per day, 1000 bacteria/mL amounts to about a million bacteria per day. ↩︎
In most environments where you find bacteria, a given species only occupies a very small share of the total population, often less than a few percent. We expect this would hold even for a pervasive environmental pathogen like mirror bacteria, since such a pathogen would still have to compete with other bacteria for resources, and would likely not be the best at exploiting most niches. So even if the default exposure to all environmental bacteria through a given exposure pathway is tens of thousands or millions per day, reducing overall exposure by a few orders of magnitude likely lowers the risk of exposure to a given bacterial species, including a lethal one, to a survivably low level. ↩︎
In studies on fecal-oral transmission, researchers have found that interventions are effective if the background prevalence of contaminants is low, but much less so if there is poor overall hygiene. ↩︎
For example, medical students have been observed to touch their face (even when instructed not to) an average of >20 times per hour, almost half the time touching their mouth or nose. Another study, looking at untrained volunteers, found similar results. ↩︎
The average US home has about 170 m2 of floor area. A study on HOCl disinfection found that ~0.3 g of HOCl per m2 was enough to achieve a >3-log reduction in viral contamination on surfaces, so covering all the floors in a typical home takes about 50 g of HOCl per application. Making 1 g of HOCl through electrolysis takes about 2.1 g of table salt, so that's ~105 g of salt per home per cleaning. This is a small enough amount of salt that all US households could regularly disinfect floors and other surfaces, especially because, given how little if at all an E2H pathogen might grow in most household environments, cleanings wouldn’t have to be that frequent to keep contamination very low. If all ~120 million US housing units used 105 g of salt once a week, total demand would be about 650 kt of salt per year, or <2% of current annual US salt production. The electricity needed is similarly low, since about 3 Wh are needed to make one gram of HOCl, which even at daily whole-house application comes to well under 1% of US residential electricity consumption. ↩︎
These technologies work by increasing the rate at which airborne pathogens (or aerosols more generally) are removed from the air or inactivated (rendered uninfectious). Without them, mechanisms like ventilation, particle settling, and normal pathogen decay typically remove infectious pathogens from a building's air at a combined rate of 1–10 h-1. Air cleaning technologies like those listed in the text typically increase the effective removal rate by a factor of 2–10× (the higher end is usually only achieved in buildings with low baseline removal rates). Holding the rate of pathogen introduction constant, this means these technologies reduce pathogen concentrations by roughly 2–10×. However, we think it may be necessary to reduce indoor aerosol exposure by several orders of magnitude relative to outdoor levels, so these technologies, on their own, are likely insufficient—though still very helpful. Several air cleaning technologies face limits on how widely they can be deployed, at least outside emergency contexts: glycol vapors are not currently legal to use for health-protective purposes in the US; portable air cleaners can cause noise issues; and far-UVC comes with various safety and air-quality uncertainties. We're optimistic and excited about efforts to overcome these limits and push for broader adoption, but don't want to bet too heavily on these longer-term approaches when more readily scalable options may be available. ↩︎