Deciphering the Recent News about Bird Flu Outbreaks in Turkey
By Eric Toner, January 13, 2006
This week’s news of sudden and widespread outbreaks of both human and avian H5N1 influenza in Turkey has been worrisome, not just because of the numbers, but also because of the questions these outbreaks raise about the possible evolution of the virus, our understanding of human infection with H5N1 influenza virus and the effectiveness of existing systems for detection and reporting.
How to Explain the Surprising Numbers, Mild Cases, and Seemingly Rapid Spread?
The outbreaks in Turkey are puzzling because the reported number of human cases is much larger relative to the size of the bird outbreak than has previously been seen in east Asia or Europe. (In Europe, there have been no human cases despite fairly large outbreaks of H5N1 in birds.) And although 3 people have died, most of the 18 confirmed human cases reportedly have been mild. It also seems as if the human and avian cases appeared very suddenly, with more than 23 avian outbreaks and 18 confirmed human cases being reported from across the country in just a 10 day period. The characteristics of the outbreaks in Turkey may suggest the possibility that they are in some fundamental way different from prior outbreaks. However, there are too many questions left unanswered to make such a determination. Did early detection fail? Were all human cases identified? Has the virus mutated? Has the clinical spectrum of this disease been identified? Those questions are important themselves as the world considers the specter of pandemic flu.
Did Early Detection Fail?
It is now apparent that the bird outbreak is much more extensive and long-standing than was previously reported. However, reports coming out now indicate that outbreaks in birds may actually date back several months, but bird owners have been reluctant to report them, and government officials may have been slow to acknowledge results of their testing.
To date, of Turkey’s 81 provinces, 11 have reported confirmed outbreaks this week, and 14 more have suspected outbreaks. Since Turkey has been on the lookout for H5N1 since its first detected in the country last spring, it is worrisome that the virus had become so widespread before it was recognized. It stands to reason that effective detection should have provided evidence of the outbreaks in birds and humans long before now. Early detection is the keystone of WHO and U.S. pandemic influenza strategies. It is important to determine the factors that may have thwarted early detection in Turkey and to take steps to correct any problems that are uncovered.
Were All Human Cases Identified?
Given the likelihood that, without the worldwide media attention focused on the current outbreaks, many of the current human cases, particularly the mild ones, would not have come to light, it is probable that human cases have been missed.
Now, reports indicate that as many as 100 people were hospitalized this week and are awaiting test results. Of the 18 confirmed cases, most had mild symptoms, and 2 children who tested positive for H5N1 flu initially had no symptoms at all. Therefore, it seems likely that mild cases among humans did exist prior to this week, but were not recognized, which raises the question of whether mild cases have been missed in other counties as well. For example, while nearby eastern Romania and the Crimea have experienced large-scale avian outbreaks for many months, no human cases have been reported. Is that because no humans have been infected, or because cases have been mild and therefore unrecognized? Is this a failure of surveillance systems, or the result of a change in the virus, or perhaps the result of a lack of knowledge about the clinical spectrum of H5N1 disease?
Has the H5N1 Virus Mutated?
The WHO reported yesterday that initial sequencing of the hemagglutinin gene from one Turkish specimen reveals the presence of a specific mutation that is thought to enhance bird-to-human transmission. Further analysis of this finding is expected in a few days. So far there is no evidence of person-to-person transmission. The significance of this observed mutation must await further analysis and correlation with clinical and epidemiological data.
Is the Full Clinical Spectrum of H5N1 Disease Known?
The clinical spectrum of human H5N1 disease is unclear. Most of the Turkish cases have been mild, while in east Asia almost all the cases have been severe. But it is impossible to account for these differences—are they due to a change in the virus? Better surveillance in Turkey? Or perhaps to early use of antivirals in Turkey? An article published this week in the Archives of Internal Medicine, and widely quoted in the news, suggests that mild human cases of H5N1 in Vietnam may be much more common than recognized. This conclusion was based on the finding of a higher incidence of mild flu- like illnesses in people who have had contact with sick poultry. It is important to recognize that this study did not involve any diagnostic testing; therefore, any conclusions about the incidence of H5N1 infection are entirely speculative.
This question could be easily resolved through seroprevalance studies. As yet, though, no seroprevalence studies have been published despite that fact that such studies are known to have been conducted in Thailand, China, Indonesia and Vietnam. Dr. Kumnuan Ungchusak, Director of the Bureau of Epidemiology of the Thai Ministry of Health has reported (personal communication) that a study of 800 contacts of H5N1 victims has shown a seropositivity rate of less than 1%. Nature quoted Yi Guan of Hong Kong as saying that that his team tested 4,000 people in southern China in 2001 and found no H5 antibodies. The true incidence of H5N1 infection and its full clinical spectrum is critical information that must be made available.
Hopefully answers to many of these question will become clear in the near future through serological studies, continued collection and sequencing of viral specimens and detailed analysis of clinical cases. A critical review of the apparent failure of early detection systems is also vitally important.