Could the next SARS-like virus reach epidemic proportions? Quantifying the likely threat of emerging diseases isn't easy, but evolution is a crucial factor that may tip the balance in favour of such human parasites.
One of the oldest tenets of evolutionary biology is that it is easier to change a little than a lot. We also know that evolutionary change is more easily selected for in a large population than in a small one. On page 658 of this issue, Antia et al.1 combine these facts to reach a previously unappreciated conclusion about emerging infectious diseases: some types of infectious parasites that attack the human population may pose a serious threat even if they are not initially able to cause epidemics. The reason is that certain parasites are specially poised to evolve so that they can cause epidemics.
In epidemiological models, an infectious agent can be characterized by its basic reproductive rate, R0. This is the average number of new infections caused by the first infected individual in the population. The epidemic threshold is R0=1, above which the disease spreads (neglecting random effects) and below which it eventually dies out. R0 is not, however, a measure of virulence or of the harm inflicted on the host by infection. In humans, for example, R0 is virtually zero for some diseases with a high mortality rate (such as rabies, and hantavirus respiratory infections); but R0 is well above unity for other diseases that also have high mortality rates (such as measles in undernourished populations, AIDS and smallpox).
Antia et al.1 point out that, because of evolution, parasites with an R0 value that is hovering just below one can be epidemics-in-waiting. An obvious reason for this is that it takes less change to achieve an R0 of more than one if the initial R0 is just below one. A less obvious reason — and the focus of Antia and colleagues' paper — is that the length of time a parasite persists in the population before disappearing increases with R0; parasites with R0 nearly at unity can persist for a considerable time by chance. The longer the parasite persists, the greater will be its opportunity to evolve to a higher R0. This is merely a population size effect: each additional host infected before the parasite dies out provides yet another opportunity for a mutation that might push R0 over the epidemic threshold.
The model is tractable because of its reliance on the single parameter R0, but the biology behind it is potentially complicated. R0 encapsulates the long chain of events involved in the parasite's association with its host — its first contact, entry, growth in the initial tissue infected, infection of secondary tissues, and so on, until the final stage of its dissemination to make contact with other hosts. A mutation that increases R0 may arise at any stage in this chain, provided that it ultimately leads to an increase in the number of infections. For instance, an increase in R0 may evolve through changes in surface molecules that improve parasite infection of new hosts directly, or it may evolve through improved growth within the host (possibly increasing virulence) so that more parasite progeny are disseminated from the host, resulting in higher numbers of secondary infections.
The rabies and hantavirus examples mentioned above are characterized by good within-host growth of the parasite but poor dissemination to new hosts. The 1976 outbreak of Ebola virus in southern Sudan (which spread from an initial infection of a cotton factory worker to the owner of the local jazz club, then to others at the club, giving at least eight generations of transmission2) is a case in which within-host growth was good and the dissemination–infection stage brought the parasite perilously close to the epidemic threshold. However, a parasite could also begin its foray into humans by being fairly infectious but poor at surviving the onslaught of our immune system (as seems to have been the case with the Ebola virus that, in 1989, destroyed an imported Philippine monkey colony in Reston, Virginia3). Mutations that increase R0 might then improve within-human growth. This type of emerging pathogen is the most easily missed and potentially the most dangerous. Efforts to understand the relationship between parasite adaptation to hosts, virulence and transmission have developed into a small industry in evolutionary biology. The relationship is complicated because it involves group versus individual selection, population bottlenecks and trade-offs4,5.
Antia and colleagues' result1 contributes to a growing awareness of the evolutionary and ecological factors surrounding the emergence of new diseases. We have already had warnings that virulence itself may evolve in response to changes in cultural practices6, and that immunocompromised patients may act as stepping stones to foster evolution of new pathogens capable of attacking people with healthy immune systems7. With the possibility of using grafts from non-human species to replace tissues in humans, we also need to be aware of the potential for activation and genetic recombination of otherwise dormant retroviruses in the human genome or in the graft.
Antia et al. do not, however, emphasize cultural factors in disease emergence. Instead, they provide a way of identifying which agents are most worthy of attention — those closest to the epidemic threshold. An example suggested by Antia et al. is the threat of monkeypox in a world with little resistance to its relative smallpox, because of a lack of either vaccination or exposure. Furthermore, the result draws attention to the neglected topic of parasite dynamics in the pre-epidemic stages. This was brought into focus earlier this year when it was realized that the initial spread of severe acute respiratory syndrome (SARS; Fig. 1) depended heavily on the social connectivity of the first (index) case in a community. Such results and realizations give us a better understanding of how to contain infectious diseases, through early prevention rather than cure. Ultimately, we should learn where and when to apply our efforts to block transmission and so prevent an epidemic.
Currently, the resources and public attention devoted to an infectious disease depend on a combination of social, biological, economic and political factors specific to that disease. Disease virulence, transmissibility and incidence are included in such considerations. The complacency of the pre-HIV and pre-bioterrorism eras has yielded to a growing acceptance of the need to monitor pathogens and even pre-pathogens in our environment. It is not beyond imagination that, even with existing technology, methods could be developed for monitoring emerging pathogens, potentially distinguishing between strains with differing R0 values. The means, provided by Antia et al.1, of identifying these epidemics-in-waiting could become a critical tool in a global defence strategy against emerging pathogens.
References
Antia, R., Regoes, R. R., Koella, J. C. & Bergstrom, C. T. Nature 426, 658–661 (2003).
Garrett, L. The Coming Plague (Farrar, Straus & Giroux, New York, 1994).
Preston, R. The Hot Zone (Anchor Books, New York, 1994).
Frank, S. A. Q. Rev. Biol. 71, 37–78 (1996).
Frank, S. A. Immunology and Evolution of Infectious Disease (Princeton Univ. Press, 2002).
Ewald, P. W. Evolution of Infectious Disease (Oxford Univ. Press, 1994).
Wallace, B. Am. Nat. 133, 578–579 (1989).
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Bull, J., Dykhuizen, D. Epidemics-in-waiting. Nature 426, 609–610 (2003). https://doi.org/10.1038/426609a
Issue Date:
DOI: https://doi.org/10.1038/426609a
This article is cited by
-
Extinction times in the subcritical stochastic SIS logistic epidemic
Journal of Mathematical Biology (2018)
-
Near-critical SIR epidemic on a random graph with given degrees
Journal of Mathematical Biology (2017)