TORONTO — A diagnostic test designed by Canadian researchers and used in Vietnam to detect H5N1 avian flu is out of date, scientists from the National Microbiology Laboratory admit -- raising the possibility some human cases may have been dismissed in error.

The test was designed at the Winnipeg lab using genetic sequencing information from samples of the virus that circulated in the first quarter of 2004. But the virus has changed enough since then that questions have surfaced about the test´s sensitivity.

Tracking the virus´s forays into and among humans is critical, given fears that H5N1 may acquire the ability to easily transmit from person to person, sparking an influenza pandemic.

"Well, you have to be concerned," says Dr. Earl Brown, a virologist at the University of Ottawa who specializes in influenza evolution. Brown was not involved in the design of the test.

"You want to have a good idea of how much infection is going on out there. And if you´re starting calling certain people with pneumonia or whatever infections negative when they´re not, then you´re not recognizing the full situation."

This is not the first suggestion that labs in Vietnam may have underestimated the number of positive cases there.

Earlier this year, Japanese scientists retested a number of specimens that technicians at a lab in Ho Chi Minh had determined were negative. The re-analysis found several of the rejected cases were actually positive.

A scientist from the Winnipeg lab says at least part of the problem behind the out-of-date test stems from the fact that Vietnamese laboratories have had limited success in isolating and growing stocks of the circulating viruses this year.

For reasons that are not clear, the virus is not growing well in the cell culture medium that was used in the past, says Darwyn Kobasa, a respiratory-viruses researcher who recently returned from Vietnam.

"That has largely failed this year. They haven´t been able to grow isolates. So that´s been impeding the collection of new sequence information," Kobasa says.

Genetic sequences from more recent versions of H5N1 are needed to produce more up-to-date test components, called primers.

Primers are tiny strands of synthetic nucleic acid used in PCR or polymerase chain reaction testing. If they are a perfect match for the influenza strain, primers used in flu tests should bind to the RNA of the virus.

Kobasa and lab technician Laura Hart spent several weeks at the National Institute for Hygiene and Epidemiology in Hanoi, sharing diagnostic expertise and helping Vietnamese scientists assess their testing proficiency.

That is when the problem with the primers came to light.

"There´ve been enough changes in the viruses between last year and this year that we found some of our PCR primers did not work that well," says Kobasa, a researcher in the division of respiratory viruses.

"With our own testing we found there were some issues with sensitivity. You can appreciate from a clinical sample that the amount of virus that might be in there is very small. So it can be quite difficult to detect."

The National Institute for Hygiene and Epidemiology has been using the Canadian test as a screening tool.

All specimens from suspected H5N1 cases were tested using the Canadian assay. Samples that tested positive were then retested using primers designed at the U.S. Centers for Disease Control in Atlanta. If the second test came back positive as well, the case was considered confirmed.

But samples that tested negative during the screening phase were not subjected to the second test.

"I think that when they report a positive result that it is positive. I can´t really say anything about whether they´re missing other positives," Kobasa says.

The primers designed in Winnipeg are aimed at three different genetic targets on the H5N1 virus, including the HA or hemagglutinin gene, explains Dr. Frank Plummer, scientific director of the lab, which is part of the Public Health Agency of Canada.

Plummer believes the primers may still be sensitive enough to catch most or all H5N1 cases, but that needs to be confirmed.

"It seemed to be working OK for the HA gene," he says.

"But we have limited information. And the fact is, they (Vietnamese authorities) have limited information too. So until we know more about the whole sequence, we won´t know really they´ve been missing a lot (of cases) or not."

Still, Plummer says it´s possible the out-of-date primers have generated false negative tests.

"Oh, yeah, sure, it´s conceivable. I don´t think it´s that likely. If it can detect one it should be able to detect others that are closely related," he says.

Whether that´s the case would depend on where mutations have occurred on the HA gene.

"These tests don´t span the whole gene, they only take little portions of it. So if the mutations are in that little portion, they may not work as well," he admits.

Kobasa says the Winnipeg lab will now try to design more sensitive primers. He and Hart returned with some genetic material from two recent virus samples, which the lab will try to amplify, isolate and sequence for that purpose.

As well, the lab is trying to get import permits to acquire two new full H5N1 isolates from Vietnam, Plummer says.