Accuracy of HIV test prompts concern

San Francisco medical officials are concerned about a federally approved oral HIV test that they say produces a higher number of false positives in their public health clinics than they expected.

At issue is the OraQuick Advance Rapid HIV Antibody Test, which the U.S. Food and Drug Administration approved for marketing in June 2004. Using fluid swabbed from the mouth, the 20-second test detects antibodies to the virus that causes AIDS.

The FDA is considering to allow over-the-counter sales of the product, making it the first such HIV test that would not require mailing a sample to a lab to get results.

Officials of the San Francisco Department of Public Health said that, in analyzing more than 6,000 oral HIV tests conducted since last spring, they discovered at least 47 instances in which initially positive test results turned out to be negative on follow-up examinations.

In a telephone interview yesterday, Shelley Facente, the city’s rapid-testing coordinator, confirmed a report in the San Francisco Chronicle that revealed some positives detected by the OraQuick Advance test were wrong.

“In the last three months, we’ve seen a rate of eight false positives per 1,000 tests,” twice the usual rate of four per 1,000, Ms. Facente said.

Despite these numbers, the OraQuick Advance test is very accurate, according to Teri Dowling, manager for HIV counseling and testing for the San Francisco health department.

The test’s manufacturer, OraSure Technologies of Bethlehem, Pa., says OraQuick Advance, has an accuracy or specificity rate of 96.6 percent. “The specificity here (in San Francisco) has been 96.2 percent,” said Ms. Dowling.

She said city health officials already recognized that OraQuick Advance has a “shelf life of six months,” and beyond that time false positives tend to increase. “But we were finding our rate of false positives were a little higher than expected,” Ms. Dowling said.

She said the San Francisco health department began using the OraQuick Advance test in February and has consistently found that “finger-stick testing is a little more sensitive.”

One bright spot from the San Francisco situation is that there have been no situations, where the OraQuick Advance failed to detect any real HIV infection.

Douglas A. Michels, OraSure’s president and chief executive officer, said that between Jan. 1 and Nov. 1 of this year 20 customers using the test, including clinics, hospitals, and doctors’ offices, reported a total of 107 “confirmed complaints of false positives” nationwide.

That number was out of 28,436 test results, Mr. Michels said, for a specificity of 99.6 percent. “That is within the FDA’s range (of acceptability). No product or diagnostic test is 100 percent sensitive or specific.”

The FDA did not return phone calls yesterday from a reporter seeking information about whether it has received any complaints about OraQuick Advance. Complaints made about products already approved by the FDA are voluntary.