A chicken being vaccinated against bird flu in Minsk.

A chicken being vaccinated against bird flu in Minsk.

Vasily Fedosenko/Reuters

Could some vaccines make diseases more deadly?

Vaccines save millions of lives every year by teaching our immune systems how to combat certain viruses or bacteria. But a new study suggests that, paradoxically, they could sometimes teach pathogens to become more dangerous as well.

The study is controversial. It was done in chickens, and some scientists say it has little relevance for human vaccination; they worry it will reinforce doubts about the merits or safety of vaccines. It shouldn't, says lead author Andrew Read, a biologist at Pennsylvania State University, University Park: The study provides no support whatsoever for the antivaccine movement. But it does suggest that some vaccines may have to be monitored more closely, he argues, or supported with extra measures to prevent unintended consequences.

Evolutionary science suggests that many pathogens aren't deadly, or not even very virulent, because if they kill their host too quickly they can’t spread to other victims. Now enter vaccination. Some vaccines don't prevent infection, but they do reduce how sick patients become. As Read first argued in a Nature paper 14 years ago, by keeping their hosts alive, such "imperfect" or "leaky" vaccines could give deadlier pathogens an edge, allowing them to spread when they would normally burn out quickly.

Now, Read has published a paper showing that this seems to have happened with Marek’s disease, a viral infection in chickens. Marek's disease spreads when infected birds shed the virus from their feather follicles, which is then inhaled with dust by other chickens. Poultry farmers routinely vaccinate against the disease, which keeps their flocks healthy but does not stop chickens from becoming infected and spreading the virus. Over the past few decades, Marek's disease has become much more virulent—which some researchers believe is the result of vaccination.

Read and researchers at the Pirbright Institute in Compton, U.K., infected chickens with Marek’s disease virus of different strains known to span the spectrum from low to high virulence. When the birds weren't vaccinated, infection with highly virulent strains killed them so fast that they shed very little virus—orders of magnitude less than when they were infected with less virulent strains. But in vaccinated birds, the opposite was true: Those infected with the most virulent strains shed more virus than birds infected with the least virulent strain.

In one experiment, unvaccinated birds infected with the most virulent strains were housed together with healthy birds. Again, the infected chickens were dead in no time, leaving them no chance to spread the disease to their healthy cagemates. But when vaccinated birds were infected with the highly virulent strain, they lived longer and all the healthy birds housed with them became infected and died. Thus, "vaccination enabled the onward transmission of viruses otherwise too lethal to transmit, putting unvaccinated individuals at great risk of severe disease and death,” the authors write online today in PLOS Biology.

The study is convincing, says Michael Lässig, a physicist at the University of Cologne in Germany who studies the evolution of influenza. "But it’s a very special set of circumstances," he cautions. "I would be careful about drawing general conclusions.”

Adrian Hill, a vaccine researcher at the University of Oxford in the United Kingdom, says the experiments support the idea that vaccines helped make Marek’s disease deadlier, but don't prove it. Many other things have changed in the poultry industry in the last decades; flocks have become much bigger, for instance, which could also favor more virulent strains. But Read says those "hot strains" would die out very quickly if the vaccines were taken away.

Hill doesn't doubt that some vaccines could lead to enhanced virulence; the real question is how likely this is to happen. His answer: It's highly unlikely, and not something we should be worried about. "They have taken 15 years to do an experiment on the only example of this happening.”

Read counters that there may well be other examples. Feline calicivirus, which causes a respiratory infection in cats, is a strong candidate, he says; “there have been outbreaks of “superhot” strains in vaccinated populations.” Read is particularly worried about avian influenza. In Europe and the United States, entire poultry flocks are usually culled to stop an outbreak; Asian farmers often use bird flu vaccines. "You could have the emergence of superhot strains,” as a result, he says. Ab Osterhaus, a virologist at Erasmus MC in Rotterdam, the Netherlands, says this is “very unlikely, but a scenario that cannot be excluded.”

But what about human diseases? Most human vaccines in use today aren't "leaky"; they are very good at stopping disease transmission. But as researchers turn to diseases that are more difficult to protect against, such as malaria or HIV, they are setting their sights lower, aiming for vaccines that prevent severe disease but not infection. "We are entering the era of leaky vaccines in humans,” Read says. Candidate vaccines against Ebola or malaria—one of which recently received an important stamp of approval in Europe—should definitely be used if they are safe and effective, he says, but they could lead to more virulent pathogens. "We need to have a responsible discussion about this.”

But to Hill, these comments themselves are irresponsible. Read "has no more evidence this will happen with an Ebola vaccine than that it’ll happen with any other vaccine in humans,” he says. "He should stop scaremongering.” The whole distinction between leaky and nonleaky vaccines is flawed, Hill argues: "Every vaccine is leaky, in that some people don’t get protected by it, some people are partially protected, some people have prevention of disease, and others prevention of infection.” Millions of people around the world receive a shots every month, and there is no evidence that that has ever led to any disease becoming deadlier, Hill says.

What's more, natural immunity should have the same effect, he adds: After we recover from a disease, we usually end up with a limited, “leaky” protection against a pathogen that is not very different from what vaccines achieve, Hill says. "For malaria, whatever today’s vaccine does is a drop in the ocean of all the immunity that is happening in Africa from all the infections in all the people.”

Hill worries that Read's work will play into the hands of antivaxxers. But Read says that even if a human vaccine is ever shown to cause dangerous evolution of the pathogen, that wouldn't be a reason not to vaccinate. The most important thing would be to support vaccination with other measures that curb transmission, such as bed nets for malaria.

Ironically, increased virulence would make it even more important to vaccinate everyone, he says, because universal vaccination would prevent the more dangerous strains from harming anyone. This is actually what has happened in Marek’s disease, Read says. "I believe because of these vaccines the industry has created superhot strains, but the vaccine still works fantastically well, because it can be delivered to every single vulnerable bird."

*Correction, 28 July, 4:07 p.m.: A quote by Adrian Hill in this story has been corrected.

Follow News from Science

A 3D plot from a model of the Ebola risk faced at different West African regions over time.
dancing shoes