As of Friday, Turkey had confirmed 18 cases of human infection with the avian H5N1 flu virus since Jan. 5. In the Jan. 9, 2006, issue of the Archives of Internal Medicine, a report suggested that in Southeast Asia, the true number of H5N1 cases might be many times the 147 since 2003 that have been officially confirmed by the World Health Organization. In a population survey covering nearly 12,000 households and 46,000 inhabitants in Vietnam, "flulike illness attributed to direct contact with sick or dead poultry was estimated to be 650 to 750 cases" over the course of six months.

So is the glass half-full or half-empty?

The fact that avian flu has expanded to the rim of Europe cannot be considered good news. But the cases in Turkey, as well as the large number of possible cases in Vietnam, also suggest that H5N1's reputation as a fearsome killer may be due to the fact that in rural Southeast Asia, with its combination of overtaxed medical systems and cultural barriers to the acceptance of Western-style medical care, only the worst cases come to the attention of doctors or hospitals in the first place.

When the net is cast wider, it seems that many cases of human H5N1 infection are comparatively mild. In Ankara, Turkey, in two preschool-age brothers, blood tests show they are clearly infected with avian flu, but you wouldn't know it from their symptoms: They have none. Several other patients have only mild symptoms. And the authors of the Archives of Internal Medicine study concluded from their data that probably "symptoms most often are relatively mild" in H5N1 infections.

Several factors might explain the cases seen in Turkey. The apparent milder course may be indicative of nothing more than earlier detection and intervention; Guenael Rodier, special adviser on communicable diseases at the World Health Organization's European office, told reporters at a press conference that "most people, if not all in Turkey, are aware of this problem, and we do see parents bringing children into hospitals at a very early stage in the disease, simply on the notion of exposure to sick or dead chickens." Since treatment with Tamiflu began immediately in each instance, early intervention may have headed off a deadly course of the disease in the Turkish cases.

But that explanation cannot account for the findings in the Archives of Internal Medicine article. In that paper, scientists reported on an epidemiological study conducted precisely to identify possible cases of H5N1 that had not come to the attention of medical professionals at the time of infection.

The authors, from Karolinska Institutet in Stockholm, Sweden; the Nordic School of Public Health in Göteborg, Sweden; the Hanoi Medical University in Vietnam; the European Centre for Disease Prevention and Control in Solna, Sweden; and Stockholm Group of Epidemic Modeling used a demographic survey to estimate the true prevalence of H5N1 infections in one rural Vietnamese province. In the survey, two questions were asked about bird flu - whether there had been any contact with poultry (categorized as "well, sick or dead") in the previous six months, and whether anyone in the household had a flulike illness in the same time period. A yes to both questions led to a more detailed follow-up interview.

From their data, the authors estimated that between 650 and 750 cases of a flulike illness occurred as a result of direct contact with sick or dead poultry in their sample. While the authors did not have serum samples and so could not confirm that the flulike illnesses were H5N1, they argued that other possible culprits such as psittacosis or Newcastle disease are less likely causes, and that, in fact, their estimate is conservative, since neither cases resulting from indirect contact nor asymptomatic individuals were included.

One worrisome fact is that while there are no confirmed cases of human-to-human transmission - the worst-case scenario that keeps pandemic planners awake at night - Turkey has had a comparatively large number of cases quite rapidly. The WHO's Rodier said that part of that might have been due to data collection. "The lab work was by batch," he said, meaning that cases reported in bulk on one day may have arisen over several days.

But given that Turkey has seen 18 cases in a week, whereas all of Southeast Asia has had only about 150 confirmed cases from 2003 to the present, he acknowledged that "there does seem to be a more efficient transmission from animals to humans" in Turkey. A more efficient transmission needs to be confirmed scientifically, but if confirmed, it certainly would not be good news. Just how bad depends on whether what's being transmitted more efficiently is an infection that is deadly to half of those who contract it, or most often a case of the sniffles.