Slow flu testing causes concern

Because detecting avian flu with standard tests is so difficult and time-consuming, waiting for laboratory confirmation of an outbreak would cause dangerous treatment delays, according to new studies of two flu outbreaks.

The studies, published Thursday in The New England Journal of Medicine, were of family clusters of flu cases in Turkey and Indonesia.


Rapid tests on nose and throat swabs failed every time, and in Turkey, so did all follow-up tests known as Elisas. The only tests that consistently worked were polymerase chain reaction tests, or PCRs, which can be performed only in advanced laboratories and which take several hours.

"It'll be a disaster if we have to use PCRs for everybody," said Dr. Anne Moscona, a professor of pediatrics and immunology at Weill Cornell Medical College. "It just isn't available at a whole lot of places."


If the A(H5N1) flu mutates into a pandemic strain, rapid tests "will be really key," she said.

The studies followed clusters in three families in Indonesia in 2005 and what appears to have been one extended family near Dogubayazit, in eastern Turkey, in January. Case clusters particularly worry public health authorities because they raise the possibility that flu is mutating to spread faster between people.

In the Indonesian cases, the authors, from Indonesia, the World Health Organization and the Centers for Disease Control and Prevention in Atlanta, concluded that human-to-human transmission had probably taken place in two of the three family clusters. In one case, a 38-year-old government auditor appeared to have caught the flu from his 8-year-old daughter or her 1-year-old sister. All three died; his wife and two sons did not get sick. No one in the family had any known contact with poultry, wild birds, animals or sick people, so the source was a mystery.

"But you can't always tell what a young child has done," said Dr. Tim Uyeki, a Centers for Disease Control flu specialist and an author of the study. "There's no magical test, and you don't always get a perfect explanation."

The Dogubayazit cluster was a cause celebre for some Internet flu-watchers following Turkish news reports in January. They contended that widespread human-to-human transmission seemed to be taking place, and that it may have begun at a banquet attended in late December by members of two related families. The Turkish government and the World Health Organization did not link the cases or families and tentatively blamed birds for all transmission.

The studies showed how wide a net was cast: 290 people were tested at one hospital because they had flu symptoms or contact with dying birds or both. All were given the antiviral drug oseltamivir, which is also sold as Tamiflu, and about half were hospitalized.

Only 10 came up positive on PCR tests, and eight of those were confirmed by a World Health Organization laboratory. All were children; four died. The studies confirmed suspicions that the families were linked; seven of the eight children were related or lived near one another.