TORONTO — A rapid, commercially available test for SARS is 100 per cent effective in detecting the disease in tissue taken during autopsies, some Toronto researchers reported Thursday, leading to hopes a way can be found to use the test on living patients.

"The first thing that we had to show in a really well controlled setting was that the test was specific and sensitive," senior author Dr. Kevin Kain said Thursday.

"Now the next step is: OK, we know this test will work if you get a decent sample. So how do we get an appropriate sample from living people?"

The work done to evaluate the efficacy of the test also turned up a troubling and unexpected finding about the course of the disease. People who contracted SARS and died after weeks or even months were likely still infectious at the time of their deaths.

That isn't true of most respiratory illnesses and points out the need for high-level precautions in the treatment of SARS patients throughout their illness, and even during their autopsies if they die. It will likely also change the thinking about why people who die after a protracted fight with SARS actually succumb.

Earlier this spring, as the medical community tried to piece together knowledge about the new disease, it was assumed an over-active immune response, not the virus itself, was the cause of death in the most severe cases of SARS.

"This would suggest that even weeks into illness, they still have virus - high levels of virus - and that means it's probably very much contributing to their death," said Kain, director of the centre for travel and tropical medicine at Toronto General Hospital.

"We don't have to evoke funny immune responses as the reasons why these patients aren't getting off ventilators," added Dr. Donald Low, another author of the study. "They've got replicating virus in their lungs."

Those findings also highlight the urgency of finding anti-viral drugs that work against the SARS coronavirus, said Low, chief microbiologist at Mount Sinai Hospital.

"They still have disease," Low said of patients who remain sick with the disease weeks after contracting it. "And maybe if we cure the disease we can reduce the mortality."

The study was conducted by researchers from Mount Sinai Hospital and the University Health Network, a hospital network to which Toronto General belongs. It was published Thursday in Emerging Infectious Diseases, a journal published by the U.S. Centers for Disease Control.

Since SARS came to the attention of the world earlier this year, health-care professionals and public health officials have been bemoaning the lack of a rapid and accurate test to weed SARS patients from those suffering respiratory distress due to other causes.

As officials in Toronto learned all too well, failing to quickly identify and isolate SARS patients can set off devastating waves of disease in hospitals and beyond.

This test was developed by Artus, a biotech company from Hamburg, Germany. It detects RNA from the SARS virus using polymerase chain reaction technology. And it does it in under an hour.

Artus provided the research team with the kits used to run the test, but provided no funding and had no input into the study. The study was supported by the Canadian Institute of Health Research, which funds Kain's research.

The team ran the test on multiple lung tissue samples taken from 11 people who died during Toronto's first SARS outbreak. They also tested lung samples from controls - people who died this spring but not from SARS, as well as lung tissue that was banked from five years ago.

The trial was a blinded one, meaning researchers didn't know which samples they were handling until all results were in.

All the SARS tissues tested positive and there were no false positives, Kain said.

That accuracy level is impressive, but can they be replicated in living patients, given that emergency room physicians are not going to be able to subject people suspected of having SARS to lung biopsies?

Kain believes sputum specimens drawn from deep in the lung could provide as accurate a source of testing material as lung tissue.

Inducing suspected SARS patients to cough up such specimens can be done, though that would have to be done under infection controlled settings to ensure health-care workers and other patients didn't get infected, he said.

"Bees go where the honey is. If the virus is in the lung and we know that and it's in the lung for a long time and we want to know if someone's infected, then that's going to be the most rewarding area to sample," Kain said.