Smear-negative study questions assumptions
May account for big chunk of transmission
In a recent study of case clusters in San Francisco, at least 17% of smear-negative TB patients were shown to be infectious, despite the widespread assumption that such patients are much less infectious (or not infectious at all) compared to their smear-positive counterparts, says Marcel Behr, MD, assistant professor in the divisions of infectious diseases and microbiology at McGill University Health Sciences Center in Montreal.
Because Behr was purposefully conservative in how he categorized and made use of available data, it’s possible that as many as a third of the smear-negative patients were infectious, he adds.
"This simply means you shouldn’t use a [negative] smear as a brain-dead tool," Behr adds. "It’s a century-old test that can detect 10,000 bacteria; and we have a disease that can be caused by 10 bacteria. So it would be very naive to think this should be the sole test used as a screen for infectivity."
When Behr compared RFLP (restriction fragment length polymorphism) results with smear cultures to analyze 71 case clusters in San Francisco, he found 15 of the clusters consisted entirely of smear-negative patients. Of course, he adds, it’s not inconceivable that a smear-positive patient started the chain of infection and then disappeared; but it’s not likely that such an event took place 15 out of 71 times.
Preliminary results from a study of contact investigations at the CDC appear to bolster Behr’s conclusions.
"A significant proportion" of both household contacts and non-household contacts appears to have been infected by smear-negative cases, says Mary Reichler, MD, medical epidemiologist in the Division of Tuberculosis Elimination at the CDC.
Some take-home messages
Practically speaking, Behr says, his study suggests health care providers should exercise caution — not always allowing smear-negative patients to roam the halls of the hospital at will, for example, or watching to make sure patients with three negative smears are responding promptly to whatever their provider is using to treat them.
"A second issue is whether you can use a smear as a guide for when someone can start going out in public," Behr adds. That question was not something the study attempted to address, he adds.
Traditionally, assumptions about the infectiousness of smear-negative patients have been shaped by conventional epidemiology, says Behr — a fact that explains some of the fallacies that trouble the issue.
"In the past, the way to assess how infectious people were was to look at their contacts," Behr explains. "We’d compare the tuberculin skin testing reactivity rates of the contacts of smear-positive people, smear-negative people, and what we called a control’ group," he says. But there were several difficulties inherent in this approach, he adds.
For one thing, contacts’ reactivity isn’t always a reliable gauge of an index case’s infectiousness, Behr says. That’s because a person’s contacts tend to mirror the person himself, he explains: "We hang around with people who are like us."
Second, reactivity in a contact doesn’t tell anything about prior reactor status, and there often is no record of previous tuberculin skin testing status for comparison.
What’s more, the farther the contact investigation extends, the more reactors the investigator may find. Yet the wider the net is cast, the less likely there is a connection between index case and contact. (One-third of the time, in fact, there is no epidemiological connection, merely coincidence, Behr found.)
The upshot isn’t to say that show-leather epidemiology is wrong, adds Behr; "but it does have limits. What it told us was that smear-negatives weren’t too worrisome, and that led us to stop looking. But if you don’t look, you don’t find. It becomes a self-fulfilling prophecy."