Did Yersinia pestis really cause Black Plague? Part 2: Examination of the criticisms

Yesterday I introduced criticisms that have been raised against Y. pestis causation of the Black Death and subsequent plague outbreaks. Today I'll discuss what I see as weaknesses in these criticisms, after the jump.

Selective quoting and interpretation of evidence

First and foremost, a big issue I have with the claims by Duncan and Scott are that they are rather selective in what evidence they choose. For example, they selected several quotes from medieval manuscripts and diaries that suggest that people at the time "knew" it was directly contagious from person to person, and conveniently ignore the other people who thought it was miasmatic, or from drinking polluted water, or simply a divine punishment or result of the constellation of the stars (because, well, they were wrong, after all). They use similar documentation to argue that individuals were well-versed in symptoms of the plague and recognized it immediately when it showed up in a village, but they then discount this recognition when it comes to the actual discovery of the putative causative agent of the plague, Y. pestis, in 1894, or even to the 1720 Marseille outbreak I mentioned in the previous post. Of this outbreak (which they acknowledge was "a genuine outbreak of bubonic plague" rather than their hypothesized hemorrhagic plague), they note:

Even though the health authorities had not seen a major plague epidemic for 50 years, they knew what they must do. They put the old, well-established 40-day quarantine measures and cordons sanitaires into practice...Of course, these precautions were completely ineffective because they were dealing with a disease that was new to them...The people of Marseille were doomed from the start; everything that had been learnt during 300 years of suffering from haemorrhagic plague was worthless in the face of this new enemy. (Return of the Black Death, page 187)

Therefore, they put their trust in descriptions found in remaining documents from the townspeople of direct person-to-person transmission and recognition of the plague when it suits them, but when an outbreak was thought by villagers to be the same illness that had repeatedly struck their city over hundreds of years (and then acknowledged by Scott and Duncan to be bubonic), the villagers were simply mistaken regarding their recognition of the epidemic, and their first-hand knowledge is dismissed.

Additionally, of the quarantine and other preventative measures, they state that the 40-day quarantine was "completely successful," but by definition they eliminate other isolation periods less than 40 days--which evidence shows were rarely successful. (And they give no reference for the 40-day quarantine being so successful; I'm not aware of anyone else making that claim and so can't verify). Indeed, many port cities were still decimated by plague even with a quarantine in place, but I'm not sure in all cases whether the quarantine was for the actual 40 days or fewer. Some support for their assertion would be nice.

The extended incubation period

Their analysis of incubation period is also based on these same historical records. They argue that they can reconstruct the chain of transmission by carefully analyzing diaries, wills, church and burial records, etc., and watching the infection jump from person to person within a family or a town. However, even modern-day contact tracing isn't always easy and straightforward to do, and I have doubts that even in the best-documented cases that they're finding wholly accurate recreations of the epidemic. Even they note, when emphasizing the need to focus on the very start of an epidemic in order to tease out transmission data:

The start of an outbreak is is of critical importance in this work. Events develop slowly at this point and analysis is possible; at the height of the epidemic multiple burials on any one day and the total confusion reigning make it very difficult, if not impossible, to sort out the many lines of infection. (Return of the Black Death, p. 159).

In the village of Penrith (a focus of their book), they base their measure of incubation period on these records, going months into the epidemic. At this point, deaths are occurring every few days (or more), yet they present only a single interpretation of the pattern of transmission through the town (and additionally, present it as definitive). Though they note that the time from appearance of symptoms to death averaged around 5 days (which fits in quite well with Y. pestis), they argue that victims only became symptomatic after a long incubation period (~32 days) during which they were infectious but asymptomatic. This extended incubation period is critical for their thesis, because in many of the records there are gaps of weeks or more between cases--something which can be explained with an epidemic transmitted at least in part by fleas (which could feed from other animals in the town, infect them, and amplify the epidemic that way), but which is more difficult to explain with a directly transmitted disease.

The problem is that their logic here is circular, and they again dismiss any evidence that is counter to their ideas. Due to their focus on some of the quotes claiming a direct person-to-person transmission, they've already ruled out anything vector-borne as a pathogen. So regardless of the gaps in burials or in reported deaths from Black Plague in a town, they must assume that every new case was directly infected from an old one, leading to their formulation of a lengthy incubation period (during which they claim the victim must have been actively shedding virus). Now, as they note, this is an attractive hypothesis in one way: if true, it could explain the rapid spread of the plague, as individuals would remain asymptomatic (but spreading the infection) for a full month--but it doesn't seem to be supported by the totality of the data.

Indeed, on page 118-120 of Return they summarize "an emerging profile of the serial killer." They suggested there were 2 different types of epidemics: a rural one which typically hit one or two families and then fizzled, and urban outbreaks that really caught hold, and then "exploded" in the summer months before dying out in the colder periods. They also note:

The disease struggled through the winter with difficulty. An epidemic that started in the autumn limped along and often fizzled out completely in the winter, doubtless to the relief of the population.

Once the epidemic had established itself, the citizens put the plague orders in force, although this made little difference in the outcome.

The time between the appearance of the dreaded symptoms and death was quite short, perhaps an average of five days.

Though Scott and Duncan are dismissive, what they describe is consistent with bubonic plague.

I could go on and on with additional criticisms of Scott and Duncan's work, but I'll just add one more example that demonstrates some of their scholarship. This is from a passage where they again emphasize the importance of historical observers of the plague outbreaks, whom they claim recognized that the infection was transmitted directly person-to-person. They include the following quote to emphasize this:

...because of its infectious nature, the disease may be spread by apparently healthy people who harbour the disease but have not yet exhibited the symptoms. Such a person was in fact a poisoner, a walking destroyer perhaps for a week or a fortnight before his death, who might have ruined those that he would have hazarded his life to save...(Return of the Black Death, p. 166.

This quote was referring to the Great Plague of London in 1665, and Scott and Duncan muse what a pity it was that "little attention has been paid" to this author's "observation."

The author, however, was Daniel Defoe (1660-1731), from his "A Journal of The Plague Year"--written in 1722. While this was indeed an account of the plague based on history, surely Scott and Duncan must be aware that Defoe didn't really "observe" this outbreak as they suggest. If this was the best evidence they could come up with, that really raises doubts about the rest of their scholarship.

Discounting combination of bubonic and pneumonic plague

Scott and Duncan also discount the complex epidemiology of Y. pestis, and the probability that the Black Death (and other large outbreaks of plague) consisted of both the bubonic and the pneumonic forms of disease. They address this briefly, but dismiss it because they claim an outbreak, to persist, must contain both bubonic and pneumonic forms of the disease: that for the pneumonic form to exist, the bubonic form must have come first. Since the objections they've raised stand against the presence of bubonic plague, therefore pneumonic plague caused by Y. pestis could not have been present either.

They also suggest that the presence of buboes (the black lumps characteristic of the bubonic form of the disease) during the Black Plague rules out pneumonic plague, because these don't appear in the pneumonic form of the disease (apparently ignoring that not every patient had buboes, as would be expected in a mixed infection). They also contradict themselves again by quoting from medieval accounts of the plague in which one observer notes there are several forms of the plague, including one in which "firstly, men suffer in their lungs and breathing, and these victims, even if they are slightly attacked, cannot by any means escape, nor live beyond two days." Nowhere in that description of disease are buboes mentioned, though another form of the sickness "running its course concurrently with the first" is noted, in which victims develop "certain swellings under both arms." But this description is early in their book, "Return of the Black Death;" apparently by the middle of the book, they'd forgotten or dismissed it in favor of other contemporary writings that suit their hypothesis better.

No rats, no plague?

Regarding the absence of rats, that's more difficult to judge. There is published work suggesting that, indeed, Iceland was rat-free during the time that plague was reportedly decimating the country. (They cite unpublished data to show that rural England was free of rats, with no means to double-check). However, even if rats were absent, this doesn't necessarily lead to the conclusion that the Black Plague wasn't caused by Y. pestis--and indeed, that's not the position the author of the Iceland paper cited above takes either. Rather, it suggests that we have a lot to learn about the mode of transmission of the bacterium--not all that surprising, since we know relatively little about transmission of many well-studied pathogens. We know that many different species of flea (including the human flea, Pulex irritans), for example, can carry plague. These all have different host preferences, and the hosts have varying susceptibility to Y. pestis (while the fleas have varying transmission efficiency). It's quite possible that animals besides rats became infected with Y. pestis and assisted in its spread over the terrain, and that this combination of human and animal (including, but not limited to, rat) dispersion resulted in the widespread epidemic that we know as the Black Death.

Scott and Duncan counter on page 178 of Return that zero animals in Europe (excluding rats) have ever been found to be infected with Yersinia pestis, in contrast to 51 in North America--but again, this is an unreferenced claim, which is another big shortcoming with Scott and Duncan's work.

What about CCR5?

Scott and Duncan argued that their proposed hemorrhagic plague virus would be a better fit as a pathogen that could increase the frequency of the CCR5Î32 allele, which they suggested provided resistance to the Black Death (as it currently does against HIV infection). However, further research has poked a lot of holes in the whole CCR5 connection. Even the mouse data looking at Y. pestis and CCR5Î32 isn't certain. First, the mouse model of infection. Though the study I mentioned yesterday found no difference in mortality from infection with Y. pestis among mice who had the CCR5 mutation and those who didn't, another study pointed out that, although the mortality rates weren't different, there were differences in how Y. pestis was taken up by macrophages harvested from either wild-type or CCR5-deficient mice. Therefore, the CCR5Î32 mutation could play a role in the outcome of Y. pestis infection after all.

Even more important, there is disagreement over the very basic premise: has the CCR5Î32 allele been selected for at all? Did it increase in frequency in the population around the time of the Black Plague? A 2005 paper suggests that CCR5Î32 arose thousands of years ago, and hasn't experienced much in the way of positive selection--in other words, that the mutation is old, and has increased due to chance as much as any selection pressures (such as that postulated from the Black Plague). Additionally, analysis of skeletons from 14th century Italy and Germany and from remains dating back almost 3000 years show that CCR5Î32 was already prevalent in ancient European populations, and that there was no difference in prevalence between 14th century plague victims and a historic control group. Therefore, that line of evidence either for or against Y. pestis causation seems to be a non-starter at this point in time.

So, while criticisms of Y. pestis causation of Black Death and subsequent plague outbreaks in Europe may seem to have some meat on the surface, with a bit of investigation it can be seen that they're really not as substantial as media reports have made them out to be. Are there questions remaining? Of course, but I think that taking the evidence as a whole supports Y. pestis as a causative agent much more than an as-yet-undiscovered hemorrhagic fever virus with an unusually long incubation period.

Additionally, some readers will be familiar with paleomicrobiological studies, examining "plague pits" and looking for evidence that the victims had been infected with Y. pestis. Don't those prove that this bacterium was the cause of the Black Death? I'll discuss the pros and cons of that research tomorrow.

References and further reading

Sabeti PC et al. 2005. The Case for Selection at CCR5-Î32. PLoS Biology. 3(11): e378. Link.

Elvin SJ et al.. 2004. Evolutionary genetics: ambiguous role of CCR5 in Y. pestis infection. Nature. 430: 418. Link.

Scott and Duncan. 2004. Return of the Black Death. Wiley and Sons.

Scott and Duncan. 2001. Biology of Plagues. Cambridge University Press.

Duncan and Scott. 2005. What caused the Black Death? Postgrad. Med. J. 81:315-320.

Karlsson G. 1996. Plague without rats: the case of fifteenth-century Iceland. Journal of Medieval History. 22:263-284.

Image from http://www.uwmc.uwc.edu/csepa/mhall/IGS/Plagues/PIA/Images/bocklinplagu… Arnold Bocklin, "The Plague," 1898