The Agro-Industrial Roots of Swine Flu H1N1

Mexico appears ground zero for an outbreak of deadly human-specific H1N1.

Of the over 1400 people that have been reportedly infected there so far, 86 have died. Short chains of transmission of the virus have also been reported in California, Texas, Kansas, Ohio, New York City, Canada and New Zealand. The virus has been identified as a new recombinant of influenza A (H1N1).

The World Health Organization has labeled the new strain potentially pandemic and the US has declared a public health emergency. Of great concern, and perhaps a marker of the seriousness of the outbreak, the deaths in Mexico, as in pandemics of eras past, appear concentrated among the young and healthy. In contrast, the mortality associated with most seasonal influenzas falls heaviest upon infants and the elderly.

Researchers have over the past several years hypothesized that a healthy and responsive immune system may explain the greater mortality among patients 20-40 years infected with highly pathogenic influenza.

Once infected with a pathogenic influenza, the blood vessels in a patient’s lungs become porous and fibrinogen—a protein involved in blood clotting—leaks into the lungs. The protein clogs the lungs’ alveolar sacs, where gas exchange takes place, and an acute respiratory disease syndrome results. In a desperate effort to save its charge, the immune system recruits such a storm of immune cells that the lungs suffer oedema. In effect, patients drown in their own fluid only days after infection. Patients with the most responsive immune system produce the greatest immune storms. It remains to be seen, however, whether the same mechanism explains the age distribution of the current outbreak.

The identity of the new virus speaks to a number of long-standing issues about influenza.

The H1N1 subtype was responsible for the 1918 pandemic that killed 50-100 million people worldwide. The genealogical descendents of that virus remain in circulation as a seasonal flu at greatly reduced virulence. The Mexico virus appears a new version of the strain, one U.S. health officials have identified as a multiple recombinant. The new virus contains genes from human, bird and swine hosts.

First, then, the ‘license plate number’ of the virus—here, H1N1—is not enough to define whether a strain will be virulent. A large modeling literature instead hypothesizes a relationship between the rate of transmission and the evolution of virulence, the amount of damage a strain causes its host.

Simply put, to start, there is a cap on pathogen virulence. Pathogens must avoid evolving the capacity to incur such damage to their hosts that they are unable to transmit themselves. If a pathogen kills its host before it infects the next host it destroys its own chain of transmission. But what happens when the pathogen ‘knows’ that the next host is coming along much sooner? The pathogen can get away with being virulent because it can successfully infect the next susceptible in the chain before it kills its host. The faster the transmission rate, the lower the cost of virulence.

That’s the mechanism in the abstract. What of specifics? What objects or processes are responsible for increasing transmission rates in such a way as to ramp up a variety of influenzas to breathtaking virulence? Growing circumstantial evidence points to intensive livestock production or, in the more critical lexicon, factory farming.

Capua and Alexander (2004), reviewing recent bird flu outbreaks worldwide, found no endemic highly pathogenic strains in wild bird populations, the ultimate source reservoir of nearly all influenza subtypes. Instead, multiple low pathogenic influenza subtypes in such populations developed greater virulence only once they entered populations of domestic birds.

While domestic populations can be divided into backyard and industrial, the former have been raised in one form or another for centuries without the now unprecedented outburst of newly pathogenic influenzas. On the other hand, the conditions for supporting such strains appear best represented in industrial animals. Otte et al. (2007) tabulated outbreaks on industrial and smaller farms for two recent outbreaks of highly pathogenic influenza. In British Columbia, 5% of the province’s large farms hosted H7N3 infections in 2004, while 2% of its small farms hosted outbreaks. In the Netherlands, 17% of industrial farms hosted H7N7 outbreaks in 2003, while 0.1% of backyard farms hosted clusters.

Even if these and other such influenza strains first developed on small holdings, industrial livestock appear ideal populations for supporting virulent pathogens. Growing genetic monocultures removes whatever immune firebreaks may be available to slow down transmission. Larger population sizes and densities facilitate greater rates of transmission. Such crowded conditions depress immune response. High turnover, a part of any industrial production, provides a continually renewed supply of susceptibles, the fuel for the evolution of virulence.

There are additional pressures on influenza virulence on such farms. As soon as industrial livestock reach the right bulk they are killed. Resident influenza infections must reach their transmission threshold quickly in any given animal, before it is sacrificed. The quicker viruses are produced, the greater the damage to the animal. Increasing age-specific mortality in industrial livestock should select for greater virulence. With innovations in production, the age at which chickens, for one, are processed has been reduced from 60 days to 40 days, increasing pressure on viruses to reach their transmission threshold—and virulence load— that much faster.

For more on the industrial roots of the evolution of influenza A (H5N1), until last week influenza’s star serotype, see this previous post.

If swine are indeed a source of any of the genomic segments of the new H1N1, industrial agriculture is by definition involved. Swine have long been identified as a mixing vessel in which avian and human strains reassort into new human-specific strains. And industrial pig production has squeezed out small farming, including across Mexico. On the Mexican hog industry, the USDA wrote in 2005,

[T]he industry has been undergoing further consolidation and concentration, due in part to the wide range of technological and commercial developments, as well as increased competition among producers. The onus of consolidation is falling primarily on Mexico’s small commercial producers (200-500 sows). These producers account for about 20 percent of Mexican hog production, while larger, more technically advanced producers (500 sows or more) account for about 50 percent.

Speculation as to the specific source of the outbreak has already begun. Blogger Tom Philpott identified a Smithfield Foods subsidiary as a possible culprit. Granjes Carrol operates large-scale hog operations for Smithfield in Perote, Veracruz, where the outbreak may have begun.

According to Biosurveilliance, a disease-tracking blog produced by an employee at Veratect, a corporate infectious disease surveillance service that is “empowering a world at risk”, health officials in Perote reported 400 cases of a flu-like illness in a local town of 3000:

Residents believed the outbreak had been caused by contamination from pig breeding farms located in the area. They believed that the farms, operated by Granjas Carroll, polluted the atmosphere and local water bodies, which in turn led to the disease outbreak. According to residents, the company denied responsibility for the outbreak and attributed the cases to “flu.” However, a municipal health official stated that preliminary investigations indicated that the disease vector was a type of fly that reproduces in pig waste and that the outbreak was linked to the pig farms. It was unclear whether health officials had identified a suspected pathogen responsible for this outbreak.

The first human H1N1 death was reported about a week later in nearby Oaxaca.

Whether Philpott’s charge sticks remains to be seen. The fly part of the putative epidemiology appears a bit strange. On its face, the recombinant nature of the new virus casts additional suspicions on the factory farm angle. How does a virus produced from multiple hosts arise from isolated large-scale livestock?

There is a way. Large pig and poultry operations are now concentrated in the same areas. As Otto et al. (2007) describe it,

Over the past 60 years, the geographic distribution of both pig and poultry production in the US, for example, has become more clustered, with poultry production now being highly concentrated in the southeastern states and pig production concentrated in some of these same states, as well as in the Midwest. Similar trends have occurred worldwide with pig and poultry populations increasingly concentrated in particular locations which are often geographically coincident.

International health officials have waved their fingers at small farmers who raise poultry atop backyard pigs, increasing the chance influenza reassorts across host species. But agribusiness has done the same at a broader geographic scale and at a much greater magnitude.

We may be caught in the ‘fog of pandemic’. It may be years before a source is identified beyond a shadow of a doubt. And it remains to be seen whether this influenza, a new addition to a veritable killers’ row of influenzas capable of infecting humans—H5N1, H7N1, H7N3, H7N7, H9N2, in all likelihood H5N2, and perhaps even some of the H6 serotypes—becomes the world’s next pandemic.

That said, the role industrial livestock production plays in the evolution of deadly influenzas must now move front and center.