The clinical bird flu picture so far as reviewed in the New England Journal

In 2005 the world's bird flu doctors got together and pooled their meager knowledge about the epidemiology and clinical features of this zoonotic disase that has so far infected 350 people and killed 217 of them (latest "official" figures via WHO). In March of 2007 they got together again in Turkey and the New England Journal of Medicine (January 17, 2008 issue) has just published a joint report summarizing their discussion. Helen Branswell sets the stage:

The article, a review of data compiled on human cases to date, answers some questions about how the virus affects people. But it also makes clear that four years after the current wave of H5N1 infections began, many unsolved mysteries remain.

"I think in every area there are unanswered questions and one wants to see progress in every one of those areas, whether it's better understanding of transmission and what are the risk factors there and how often human-to-human transmission actually occurs," said senior author Dr. Frederick Hayden, a scientist with the World Health Organization's global influenza program.

The review article is based on data supplied by clinicians who have treated H5N1 and other research presented at a special WHO meeting held last March in Turkey. Hayden oversaw the project and led an international committee of experts who wrote the article, which appears in this week's New England Journal of Medicine. (Helen Branswell, Canadian Press)

The paper itself has no surprises but it's a worthwhile read, clear and concise. Here are some salient points that sum up the situation as of now:

• Influenza A/H5N1 is now thoroughly entrenched in the poultry of large parts of Asia, Africa and the Middle East
• It has diversified genetically and evolved into distinct lineages or clades, some with subclades (so much for the "extremely stable" virus that M. Vallat of the World Organization for Animal Health (OIE) was promoting just last week)
• Human infections are from avian sources
• Human effected movements of poultry and poultry products are the likely primary means of geographic spread, although migratory birds also play a part. The relative proportion of poultry versus wild birds remains uncertain, as does the importance of wild birds as a reservoir
• Human cases are dramatically skewed toward the young (median age 18). 90% of cases are below the age of 40. Older cases have had better prognosis. Highest case fatality is in the 10 to 19 year age bracketWhether the lack of older cases is a feature of susceptibility or reduced exposure is unknown
• Most cases were previously healthy. This is not a disease of those already at risk for other disease. Of the six pregnant women, four have died and all fetuses were lost.
• H5N1 infections in poultry and people follow the same temporal pattern as "seasonal" flu, increasing in the cooler months
• No "traveler" cases have yet been identified. All cases are among residents of the region where the case was diagnosed
• Some identified risk factors for poultry handling: slaughtering, defeathering, preparing sick poultry for cooking, playing with sick birds, handling fighting cocks or ducks even when they appear to be well, consuming raw or undercoooked poultry or poultry products
• No human cases have been described where transmission has been from a mammalian host other than another person, but there is no obvious reason why this cannot occur
• Clusters of cases have occurred in ten countries, although the discussants judged most to be from common sources. However several highly likely cases of human to human transmission were believed to have occurred when there was very close contact between a sick individual and a caregiver or relative
• However in fully a quarter of cases there is no clearly identified source of infection
• In these and other cases there remains a strong possibility that infection occurred from contact with virus contaminated materials (aerosolized or self contamination with poultry feces, dust, feathers) in the environment
• We still do not understand the factors that determine host range
• Changes in the virus needed to make it easily transmissible between people likely involve several genes. We don't know what those changes are as yet
• Cause of death in most cases is an overwhelming primary viral pneumonia. Exactly what tissues are infected first and how is still unknown with certainty. The effects on the immune system and reported "cytokine storm" in some patients is still not understood. No therapy for this is recommended.
• Cases are seen first by the health care system a median of four days from onset of symptoms. Median time to death is 9 to 10 days. Only one in eight are diagnosed correctly on initial presentation, with "pneumonia" and dengue fever being among the chief incorrect diagnoses. False negatives are common. Multiple specimens are recommended when clinical and history strongly suggest a diagnosis of bird flu
• Rapid diagnostic tests to see if a patient has "influenza" (any subtype) are insensitive and produce many false negatives.

These are some of the items that drew my attention but there is much else in this paper. I recommend it for further details (.pdf here). One of the more worrisome findings is the continued failure to find mild cases. Again, Helen Branswell:

The same test results were obtained for nearly 1,100 people in studies designed to look for missed cases of H5N1 infection. Seven new studies from Cambodia, China, Thailand and Nigeria that looked for H5N1 antibodies in the blood of poultry cullers, live animal market workers, health-care workers and others who were in contact with H5N1 cases turned up only a single positive, in a live animal market worker in China.

Most diseases have a range of severity and experts had hoped the confirmed cases of H5N1 infection represented only the tip of a much larger iceberg. If that were true - if, say, there were 10 mild or virtually symptomless cases for every one found - that would dramatically lower the death rate.

But these studies, when combined with earlier small studies that failed to find any real evidence mild infection was escaping detection, are compelling proof that this isn't the case at present with H5N1, Hayden said.

"I think it's persuasive that there was not (mild disease). Now it doesn't tell you what's going to happen in the future, of course. And that's why you need to keep looking," he said.

Thus the truly horrendous case fatality ratio of over 60% still stands as the best estimate. So this remains a bad disease. A really, really bad disease.