The Man-made H5N1 Controversy Heats Up: What Next? (Part Three B)
We pick up here with Part B of today's blog on H5N1:
Whether H5N1 is ultimately 60 percent lethal in humanly transmissible form, or an order of magnitude less deadly, even with a 6 percent mortality rate it would be far more dangerous than any previous influenza pandemic. The 1918 strain was lethal in just 2 percent of infections. The far less lethal H1N1 2009 swine flu nevertheless exacted a significant price globally.
But Palese insists that concerns about H5N1 are hyperbolic. There are literally tens of thousands of people in Asia that have been infected with the virus, he argues, but never developed symptoms. During the February 2 NYAS debate, tempers flared between Palese and University of Minnesota epidemiologist Michael Osterholm over this point. Osterholm insisted that very few asymptomatically infected human beings have ever been found, even within households where a family member died of H5N1 flu. Palese claimed the opposite, insisting that bird handlers in Asia carry the virus harmlessly.
The men were arguing over the denominator. "Only hospital cases are counted. People who are asymptomatically infected are not being counted," Palese said. "The unfortunate WHO estimate is very high."
It is commonly said that about 60 percent of human beings that have been infected with H5N1 since 2005 have been killed by the virus. It is also commonly said that no cases, or very few cases, of human-to-human transmission of H5N1 have occurred. The Palese/Osterholm spat was over the size of the world’s asymptomatically infected human population – a number Palese believes is very large, but Osterholm insists approaches zero. Palese claims that in ten studies involving 500 or more individuals the infection rates ranged from 0.2 percent to 5.6 percent. Osterholm, who is a member of the NSABB, countered that the Board had reviewed all available data on asymptomatic infections in people, and in the thirteen studies that met the WHO antibody testing criteria only twenty-six of 5,533 subjects were positive for H5N1 antibodies, for a rate of 0.469 percent.
A review of the literature shows very few cases of asymptomatic human infection have been seen anywhere in the world. None of the ninety-seven res responders to a German H5N1 animal outbreak tested positive; none of the 198 Thai health professionals that treated dying H5N1 patients proved to be infected; eighteen of 700 Cambodians living in a village that had a large outbreak of the disease tested positive (5.6 percent); in a Hong Kong household outbreak, 12 percent of contacts tested positive; another Hong Kong survey found that 10 percent of poultry workers and 3 percent of government investigators in a local H5N1 outbreak tested antibody-positive to the virus.
An international team of scientists surveyed 177 Indonesian H5N1 cases that were identified between July 2005 and July 2009, and 496 of those individual’s close contacts (co-workers, family members, and the like). They found clusters in which it appeared one person had passed the virus to another, but nearly all transmissions were symptomatic, there was no vast reservoir of well carriers, and transmission was extremely rare. “Overall, the study found that AI H5N1 human infection resulting from human transmission of the virus was very limited,” the researchers concluded, “and that the reproduction number was well below the threshold for sustained transmission.”
A survey of poultry markets in Bangladesh found very significant levels of poultry infection for other flu strains among employees of the live animal markets, but near-zero evidence of H5N1. Bangladesh has had many outbreaks of H5N1, but between such episodes of animal and/or human infection the virus seems to entirely disappear from domestic birds and humans.
Egypt is now in the midst of a significant outbreak of H5N1, both in birds and people. Last year a joint U.S./Egyptian team of investigators published an investigation of 119 known human cases; forty of which were lethal. Though the fatality rates in Egypt are far lower than those seen elsewhere in the world, Egypt’s H5N1 episodes also seem to appear, and then disappear, leaving few humans asymptomatically infected. The researchers found fatality rates varied according to seasons, and to the types of domestic birds associated with local outbreaks.
In recent weeks, Vietnamese health authorities have surveyed poultry farmers living and working in an area where two people have died of the disease so far this year: just 0.4 percent were antibody-positive for H5N1, according to a February 6 statement from the National Institute of Hygiene and Epidemiology’s Nguyen Tran Hien.
The second public health utility cited by proponents of the research is for disease surveillance. As the argument goes, knowing which genetic changes are necessary to turn H5N1 into a human contagion allows routine flu investigators to narrow their search to forms of the virus that display one or more of those mutations. Presumably such genetic epidemiology could also be used to minimize the financial catastrophic culling exercises now routinely used to destroy contaminated flocks of poultry in poor countries. Farmers are routinely driven to bankruptcy in these efforts – perhaps, it is suggested, less drastic measures can be deployed when the culprit H5N1 strain lacks the five deadly mutations.
Sadly, genetic epidemiology is not the norm anywhere in the world today. There is no tool kit for affordable, rapid screening of flu strains in the field. Even where careful field work has been done, genetic sequencing of relevant samples of H5N1 is rarely performed. Most sequence information stored in public archives is fragmentary. Moreover, if a strain with pandemic potential were to emerge in Cairo, Jakarta, Guangzhou, or almost anywhere else in the world today it would likely be spreading widely and killing thousands of people long before a laboratory finished sequencing the strain.
The final public health rationale is vaccine production. Knowing what types of viruses might be responsible for a global pandemic could spawn vaccine design and stockpiling, targeting those strains. There are reasons to be skeptical about vaccine stockpiling and production. Total global capacity to produce specific flu vaccines is now 300 million doses in six months, according to U.S. government sources. Based on current technology there is no reason to assume that scale, either in time or volume, will change appreciably in coming years. Based on experiences since 2005 with widespread H5N1 vaccination in poultry, there are limits to cross-reactivity from strain to strain/season to season. Moreover, vaccine-resistant strains of H5N1 have emerged with regularity. Modeling indicates that vaccination that fails to achieve over 95 percent protection against the virus will lead to resistant strains, and may actually worsen epidemic outcomes.
In 2007, the Gates Foundation’s Tachi Yamada was convinced that advances in adjuvant use for vaccination indicated H5N1 prepandemic stockpiling was viable. At the Pacific Health Summit that year in Seattle, Yamada argued that even weakly cross-reactive vaccines could be boosted to confer some level of protection, using new adjuvants to goose up the response. In Yamada’s scheme, delineated at the Pandemics: Working Together for an Effective and Equitable Response 2007 Summit, a currently circulating strain of H5N1 would be selected as the basis for vaccine production. Some 7 billion doses would be manufactured, and stockpiled. At the first sign of sustained human-to-human H5N1 transmission, the WHO would declare a pandemic threat, and the entire world population would be immunized.
The Yamada scheme never got off the ground. Vaccine manufacturers were, predictably, disinterested in rev’ing up production without guarantee that somebody was going to buy $12 to 15 billion worth of product. Public health experts mocked the logistics side of the equation, insisting that immunizing 7 billion human beings would constitute an unparalleled feat, requiring billions of dollars’ worth of logistics, supplies, cold chain capacity, training, and transport.
Tomorrow’s Blog: The Danger of an Overly Narrow Debate