DNA solves AIDS epidemiological whodunit.

When Florida dentist David Acer was accused in 1990 of passing on the AIDS-causing virus (HIV) to seven of his patients, the case raised many nagging questions. People wondered whether a visit to the doctor or dentist was still safe. Policymakers debated whether HIV-infected health care workers should continue to see patients.

But epidemiologists, and later lawyers, faced probably the most vexing and fundamental question: Had the dentist really infected those seven people?

When the Centers for Disease Control (CDC) in Atlanta conducted its initial investigation, the results were suggestive but far from conclusive. Two of the seven infected patients had engaged in behavior that put them at risk for HIV, whereas the other five had not. Investigators knew they'd have to dig deeper for answers.

"Without being able to look at the virus," says CDC epidemiologist Harold W. Jaffe, "it would be very hard to say for sure what the source of infection was."

Now, Jaffe and other CDC researchers have done just that. In some novel genetic detective work, the group has studied the virus' DNA to determine whether the Florida dentist actually infected his patients with HIV. The group's findings, reported in the May 22 SCIENCE and the May 15 ANNALS OF INTERNAL MEDICINE, confirm their initial suspicions: The five patients not at risk for HIV infection contracted the virus from Acer; the other two patients did not.

"This is a pioneering study," says physicist Temple F. Smith of the Biomolecular Engineering Research Center at Boston University. "It's an attempt to look at the DNA and say: Can you actually tell who infected whom?"

The epidemiologists' conclusions hinge on two previous observations about HIV. First, because its genetic data are constantly shuffling, HIV comes in many varieties. Second, HIV strains in people who directly infect each other--for example, sex partners, mothers and their infants, blood donors and recipients--look more genetically alike than HIV strains in people with no contact. Thus, the researchers reasoned that persons whose HIV looked genetically similar to the dentist's had been infected by him.

To find out, the CDC team isolated three short genetic segments from HIV in Acer, his seven patients and 35 HIV-infected Floridians who had had no contact with Acer.

Using a technique called polymerase chain reaction (PCR), the scientists then multiplied and sequenced all these segments. The sequences were analyzed statistically and compared to one another. The only sequences matching Acer's were the ones belonging to the five patients with no prior risk for HIV.

Although both CDC studies taken together seem to resolve the "who infected whom" question, several mysteries remain. One of the most frustrating, says Jaffe, is that CDC researchers still don't know exactly when or how the dentist infected his patients. Investigators conclude, however, that neither deliberate infection nor sexual contact caused the HIV transmission.

Furthermore, although the DNA analysis techniques used in the study look promising, some scientists believe they're not yet accurate enough to use as evidence in court or as a basis for public policy. "If I was on trial," says Smith, "I would not want data used that has some of the potential problems this may have." But someday, he adds, these techniques will become the "ultimate method for identification."

As for the safety question raised by the Acer case, CDC researchers point out that this is the first known case of a health care worker infecting a patient. In fact, they calculate that an HIV-infected surgeon would infect a patient just once every 42,000 to 420,000 procedures.