Quantiferon looks good, investigators say, but dicey research remains

Don’t throw out tuberculin skin tests just yet

Data from the first round of field trials that pitted newcomer Quantiferon against the tuberculin skin test, or TST, have shown exactly what investigators had hoped: Quantiferon, a one-step assay already in use in veterinary medicine, looks like a much more accurate test.

In an ironic twist, the measures of just how good the test may prove are when and how often it disagrees with the so-called "gold standard," the much-maligned TST.

"If you weed out those things known to cause false positives" — namely, cross-reactions from nontuberculous mycobacteria and reactions to BCG — "what you get is increasing agreement between the tests, up to about 90%," says Jerry Mazurek, MD, the project officer for a five-site study conducted by the Atlanta-based Centers for Disease Control and Prevention’s Division of Tuberculosis Elimination.

Tests agree about 81% of the time

Hospitals and programs shouldn’t start throwing out their vials of skin-testing reagents just yet, however, because the last phase of research will be tough going, Mazurek adds.

Here’s how the two tests have stacked up against each other so far. "Overall, they agree about 81% of the time," says Mazurek. "If subjects had had a BCG vaccination, there was greater discordance," yielding only about 70% agreement, compared with instances in which there was evidence of atypical mycobacterial infection, he says.

"If evidence showed infection with an atypical [mycobacterium] — usually, Mycobacterium avium — the risk ratio is about 3, so if you had evidence of M. avium, you were three times as likely to have a positive TST and a negative Quantiferon than if you had no evidence of an [atypical infection]," he adds.

The only other time that results varied measurably was when test readers seemed to exhibit what Mazurek terms "digit preference" — that is, when they were more apt to call a reaction a 9 mm than the cut-off point of 10 mm, presumably to spare subjects the trouble of having to get preventive therapy.

"So if we had more nines than 10s recorded at a site, then there was a slightly greater degree of discordance," Mazurek adds.

The bovine gold standard’

What makes it tough to interpret results from the 1,500 subjects tested in the first round of the trial is that there isn’t a biological correlate, or "gold standard," to back up investigators’ hunches about what the tests mean in humans. That means the TST has to serve as the de facto gold standard, even though it falls short, in most peoples’ estimates.

Take the instances in which subjects with a history of BCG tested negative with Quantiferon and positive with the TST. As Mazurek puts it, "Is that because people who got BCG were from areas where there was a lot of TB? If so, maybe what the TST was telling us was true. My hunch is that’s not the case."

What bolsters Mazurek’s hunch is the way there is gold-standard-quality evidence in support of Quantiferon in cattle. "There is a gold standard in cattle, and that’s [evidence from] cultures," he explains.

It works like this. Usually by the time animals are brought for evaluation, the disease is well-established. "They can’t complain of a cough, obviously, but they usually come in looking pretty sick and having lost weight," Mazurek explains. The animal is tested with Quantiferon for Mycobacterium bovis; those that test positive typically are slaughtered, and lymph nodes from as many as two dozen sites are dissected and cultured.

That opens the door for side-by-side skin tests, with a culture available to test results. When CSL, the Australian company that produces uantiferon, has done side-by-side comparisons, Quantiferon has shown greater predictive value than the TST, says Mazurek.

Plus, even though the available evidence so far is slight, in instances where cattle also have been dosed experimentally with BCG vaccine, Quantiferon again reportedly performs more accurately than the TST, adds Mazurek.

Along with its promise of greater accuracy, Quantiferon offers the prospect of a much speedier skin test than the TST.

That’s because the assay depends on a single step — a blood draw — instead of a two-step procedure in which subjects have to return for test readings. Instead of getting an intradermal injection, subjects provide a small amount of blood, which is portioned out into wells and mixed with three antigens: one for M. bovis, one for M. tuberculosis, and one for M. avium.

Specimens and antigens incubate for about five hours. Then the amount of gamma interferon produced by each reagent is measured. Gamma interferon, a cytokine that regulates cell-mediated immunity, is produced by lymphocytes that already have been sensitized, Mazurek explains; thus, the infective agent is presumed to be the one that produces the most gamma interferon.

Even though the news on the newcomer assay is encouraging, measuring the predictive value of the new test will be tough, Mazurek warns. "What we don’t know yet is how well [a positive result] from Quantiferon predicts that someone will get TB disease," he says. "Is it one in 10? One in 20? We still need to do that study, and it’s going to be hard."

An ethical Catch-22

For one thing, there’s the built-in ethical Catch-22: Since the TST is the accepted diagnostic test, how can researchers rightly deny access to preventive therapy to someone who tests positive with the TST but negative with Quantiferon? In theory, researchers could offer such people isoniazid and try to follow those who refused treatment. "But generally, those people are the hardest to follow," he notes.

There’s also the sheer scope of the trial. "It would take a tremendous number of people [to do a study], since you’d need to recruit about 4,000 people with a positive tuberculin skin test," he says. "To get that many, you’d have to test about 100 times that many people," given the low incidence of TB infectivity in the United States. Then researchers would have to follow the population for at least several years.

"Some people wouldn’t take their INH," concludes Mazurek. "Your only hope is that you’d get enough of them that you’d be able to draw some conclusions."