Zombies are unlikely creatures to be fashionable. But they are in vogue, with at least three recent blockbuster movies to their credit in the last few years: I Am Legend, Outbreak and most recently World War Z. Despite this, a review of the published scientific literature reveals a paucity of evidence regarding the microbiological characterization of any zombie virus or other pathogen and the epidemiological evolution of any zombie epidemic.
Taking the depiction of zombie epidemics in film as a starting point, this blog post explores the likely microbiological and epidemiological characteristics of any zombie epidemic and its likely maximum cycle time from inception to effective viral eradication. With these estimates in hand, governments and public health agencies will be better prepared to effectively retard or even prevent any zombie epidemic that would otherwise have the potential to destroy our species and human civilization as we know it.
Natural History and Symptomatic Presentation
All three movies have a microbiological theme, presuming that a virus could somehow zombify the host. Here are the assumptions that are typically made in modern zombie movies:
- Zombification is caused by a microbial pathogen, typically some kind of wild or engineered virus.
- Route of transmission of the microbe is through being bitten by an infected person.
- There is an asymptomatic period after being infected.
- Symptoms of infection include photophobia, craving and consumption of animal (ideally, uninfected human) flesh, tachycardia, increased metabolism, hair loss, gait modification, severe cognitive impairment and greying of the skin.
Some of these features, particularly the symptomatic presentation of zombification, are inconsistently reported in the media. For example, in some movies the gait modification may be absent and instead have superhuman speed and strength triggered by the prospect of raw human flesh.
Features of zombification that are often omitted from zombie movies are also numerous and critical to determining the microbiological or epidemiological plausibility of zombification. The most important of these features warranting further research include the reproductive capabilities of zombies, th food preferences and the longer term prognosis of zombification. In this discussion it will be assumed that zombies do not reproduce (although viviparous zombie birth has been reported, it was a consequence of intrauterine infection), and that in extreme cases zombies will eat non-human flesh, including other zombie flesh.
Spontaneous eradication of the pathogen by the host's immune system has not been documented in any zombie movie identified in this review, although genetic immunity or asymptomatic carrier status has been reported in multiple movies. Nor has any movie adequately documented the long-term outcomes of zombification, particularly when in the absence of human flesh.
Even in the absence of information on long-term prognosis, however, it is still possible to discuss the plausibility of zombification during the acute phase of any epidemic, both from a microbiological and an epidemiological perspective.
Evolutionary and Microbiological Speculations
First from the microbiological perspective, we in fact already have an obvious candidate pathogen from which the plausibility of a zombie virus (Z-virus) can be gauged: the Rabies virus. Rabies shares many features with Z-virus, including cognitive modification, transmission through biting and violent or aggressive behaviours that promote such transmission. Like Z-virus, Rabies also has a latency period. However, in the case of Rabies this may be prolonged. Indeed, in humans cases have been documented of many years between infection and onset of behavioural symptoms. For Z-virus the asymptomatic period seems to not exceed 24 hours.
Nonetheless, Rabies remains an obvious candidate for the initial tweaking of microbiological parameters and symptomatic presentation to conceptualize a Z-virus. Much of the behavioral symptoms are already present, for example. Photophobia is also often a symptom of encephalitis, and particularly meningitis for example, and so could easily be presumed to be present with infection of a neurotrophic virus such as Z-virus. Tachycardia and metabolic changes are also frequently associated with infections generally, not just those that are neurotrophic.
But, what of the craving and consumption of animal flesh? To begin, dietary preferences can be strongly influenced by the human microbiome, which in turn can be modified microbiologically (for a recently documented example: through bacteriophage viruses that regulate the prevalence of oral bacteria). So it is not much of a stretch of the imagination to conceive a virus that is able to selectively destroy those gut microbes that aid in digestion of vegetable matter in preference to digestion of animal flesh. The Z-virus may for example result in excessive flatulence or irritable bowel symptoms if zombies tried to eat a banana or bowl of porridge.
But, this does not concord well with the behavioural natural history of the Z-virus as popularly depicted. In particular, the preference for animal flesh appears to develop spontaneously, rather than as the result of behavioural modification following zombie flatulence or explosive diarrhea. Thus, such changes in the microbiome must occur alongside very specific hormonal changes to modify dietary preference. Moreover, zombies have never been documented consuming Z-virus infected animal flesh. Thus, such discrimination must be the result of pre-existing ? that is, common to both zombies and normal humans ? or further Z-virus induced changes to food preference. Perhaps this is easily made plausible, if it is assumed that zombie flesh emits an odour that humans have evolved to associate with disgust (e.g. the odour of rotting meat). In this case, although there would be no a priori reason for zombie flesh to be pathological to other zombies if consumed, there would be a rationale for zombies to have a strong preference for normal human flesh, just as normal humans have a strong preference not to eat rotting meat.
The zombies strong preference for normal human flesh also has sound evolutionary grounds, at least in the acute phase of any epidemic. This is because it would be in the interest of the Z-virus to make its hosts prefer uninfected flesh rather than infected flesh, particularly when the unsuccessful attempt at eating uninfected humans is the primary mode of transmission.
This evolutionary line of reasoning naturally leads on to the discussion of the epidemiology of the Z-virus. Specifically, transmission is not by zombie bites per se but specifically by zombie bites that do not lead to the fatal consumption of the new host by zombies. That is, the newly infected human must escape the zombie attack that resulted in a non-fatal zombie bite, before they can themselves be successfully develop symptomatic zombification and transmit the virus. Therefore, given the necessity of at least some bites leading to viral transmission, it follows that in an evolutionary system (this is a creationist-free thesis!) the Z-virus could only have survived if it produces behavior that trades-off the need for the host to survive through the consumption of human flesh against the need (of the virus) to reproduce.
It is this trade-off that can be informed by theoretical epidemiology. Without known exception, every virus must be partially characterized by a value R, relative to a population P. R is the reproductive co-efficient. Put simply, it is the number of successful viral transmission events per index host. But, this number varies depending on the characteristics of the population (e.g. geographical dispersion, prevalence of immunity, etc.). So, if on average one zombie ? throughout their zombified lifetime ? infects two humans with the Z-virus (without subsequently eating them), then R = 2. So long as R > 1, an epidemic ensues and the virus will eventually infect every susceptible member of P. If R < 1, the virus will eventually be eradicated within P.
What is a plausible value of R for the Z-virus in a naïve human population P? Here again, the problem of the long-term prognosis arises. Based on the limited amount of reports from the movies, within 28 days the Z-virus has resulted in the near extinction of the human race, with perhaps only a few high-security facilities or isolated island populations remaining uninfected. The value of R for these first zombies must be very high to result in the near eradication of uninfected humans within 28 days.
A quick calculation of the minimum value of the related measure of the reproductive rate (r) of the Z-virus illustrates this, where r is the number of new infections per index case per unit time. Let t2 be the length of time the prevalence of infection takes to double within P, i.e. the viral doubling time. Given the rough estimates of the total human population (P) of 8 billion, 28 days for P to develop herd infection with the Z-virus at time tz, and a single index infection of the Z-virus at time t0, then it follows that t2 < log2(8x109)/28. That is, t2 < 1.17, and therefore r > 2/1.17 per day.
So it takes a maximum of 28 hours (1.17 days) during the first 28 days for the Z-virus to effect herd infection within a population of 8 billion susceptible humans. This is far faster than any documented epidemic, but there remains no mathematical reason for its impossibility. Note however that this figure is only an approximation, since in reality humans act as disease vectors or as a food source, but not both. So, the actual population of humans at risk of infection at any moment should exclude those that have or are in the process of being eaten. This proportion has never been explicity reported in the media.
There are other restrictions on epidemic growth however. Namely, the availability of alternative food sources following the consumption or infection of all humans.
If we make some reasonable assumptions regarding zombie metabolism and dietary restrictions, then given the above lower bound estimate on the reproductive ratio of the Z-virus, very soon the zombie population faces an energy crunch; there's no more human flesh to eat. At this point, conceivably, they may switch to eating mammalian flesh. Making such assumptions on alternative food sources should also inform the long-term prognosis of zombification.
A typical human may take many weeks to die due to starvation. The upper limit reported from hunger strikers is 40 days. However, we should not assume from this that a zombie epidemic will have burnt itself out after a maximum of 40 + 28 days. This would presume both that zombies have a similar metabolism to uninfected humans, as well as that non-human or zombie flesh is toxic to them. This is clearly not the case, since multiple zombie movies have either implicitly or explicitly (e.g. Day of the Dead) indicated a specific zombie metabolism, and zoonotic transmission of the Z-virus is also well documented, as well as consumption of non-human mammals by zombies. This would be consistent with reports of the period taken for a zombie population to starve following herd infection and consumption of any remaining humans as 28 weeks (from the movie 28 Weeks Later).
Assuming for the moment that the Z-virus can infect non-human species with similar values of r as among humans, that zombified flesh is toxic to zombies but that non-human mammalian flesh is an alternative food source to zombies than human flesh, then given reasonable estimates of non-human mammalian biomass it is possible to project population-level prognosis beyond the limit of 40 + 28 days.
A rough approximation of the biomass of mammalian livestock is that it is 3-fold greater than human biomass. But with such a short doubling time, this additional available biomass would provide only an extra day or so of lifespan for the zombie population. Adding wild mammals would increase the biomass substantially, say by a further factor of 2, thus adding an additional 28 hours. Still, this remains a poor prognosis. Biomass of non-mammalian species however is huge, and dwarfs that of mammals. But these species (insects, birds, etc.) are difficult for such cognitively impaired and non-dexterous creatures such as zombies to use as a reliable food source.
Therefore, it seems the zombies (whether human or non-human) must resort at some point to eating each other. This at least makes the calculation simpler since there would be no longer any new infections. Restricting the analysis to human zombies, a typical human needs approximately 1,500 calories per day to stay alive. A 250g steak contains approximately 500 calories. Assuming the mass of an average zombie to be comparable to the edible weight of a normal human (excluding free fluid content and skeletal mass) at 50kg, and comparable nutritional value of zombie flesh and beef steak, then this represents a total of 100,000 calories per human cadaver. That's enough to keep a zombie alive for approximately 67 days. Assuming zombies do not reproduce by viviparous birth, and the cognitive capabilities to farm, and that no zombie eats more than what is necessary to keep them alive and functioning, then this means that every 67 days the population of zombies will halve. Performing the above calculation to derive an estimate of the doubling time in reverse, and assuming that 50 percent of all humans were used as food during the first 28 days, then this provides the zombie population with an additional 27 x 67 = 1,809 days of sustenance. In total therefore, it is reasonable to assume as a rough approximation that any zombie epidemic would have a maximum cycle time of 1,809 + 40 + 28 = 1,877 days or just over 5 years.
Using cultural references and reasonable assumptions as to zombie metabolism, dietary requirements and estimates of the reproductive ratio of any zombie virus, the duration of any zombie epidemic is estimated to be approximately 5 years. Following this time, it should be safe for any remaining human populations to regroup and re-establish humans as the dominant species.
The existence of zombies has not yet been empirically verified. Nonetheless, in the event that a Z-virus is discovered and microbiologically characterized, the results of this discussion should assist public health officials in taking measures to contain the spread of the virus wherever possible, and help governments to better manage the social, economic and ecological consequences of the epidemic.
Michael Hughes, Ph.D., is a principal epidemiologist with Decision Resources Group.