Polymerase chain reaction (PCR) testing, which is much more expensive than AIDS antibody tests, should be used routinely to detect HIV infection in infants but not in adults, according to two new studies from Veterans Affairs and Stanford researchers.
"We found that PCR is not as accurate as antibody tests in adults, but in infants it's one of the best tests available," said Dr. Douglas Owens, health services senior research associate at the Veterans Affairs Palo Alto Health Care System and assistant professor of medicine and of health research and policy at Stanford University School of Medicine.
Researchers led by Owens and Dr. Mark Holodniy reported their findings in the May 1 issues of the Annals of Internal Medicine and the Journal of the American Medical Association. They based their conclusions on analyses of 96 studies that evaluated PCR screening for HIV in adults and 32 studies looking at the technique in infants.
Unlike antibody tests, which detect molecules formed by the immune system in response to HIV, PCR is a newer method that detects viral genes in a blood sample. Within hours, this technique can turn a tiny shred of genetic material into millions of easily analyzed copies.
Method's Worth Has Been Debated
When PCR became available about 10 years ago, Owens said, many physicians hoped it could be a useful test for HIV infection in both infants and adults--despite its expense. Since then, the method has become controversial, as numerous studies have failed to agree on its value in HIV testing. "Some studies indicated PCR testing was perfect, while others indicated it was inaccurate," Owens noted.
To help settle the issue, Owens and his colleagues performed a meta-analysis on data from a number of smaller studies. The researchers concluded that antibody tests are more accurate than PCR in adults and recommended adult PCR testing in only a few situations. For example, PCR may help detect HIV in people exposed so recently that their bodies haven't yet generated antibodies against the virus.
For infants, however, the researchers concluded that PCR is very useful in diagnosing HIV infection. All babies born to mothers with HIV have maternal antibodies against the virus in their blood for about 15 to 18 months, and therefore, antibody tests cannot tell physicians whether a newborn is actually infected, Owens noted.
Although PCR is one of the most accurate tests for use in infants, he said, a single test result is not sufficient to make or exclude the diagnosis. Rather, doctors should follow the infant closely for signs of HIV infection.
PCR Accuracy Depends Partly on Age
The accuracy of PCR depends, in part, on the child's age, Owens explained. "Current thinking is that a baby often is infected during birth, and it takes a while for the numbers of virus to build," he said. The researchers recommend using PCR after a child is 30 days old or repeating the test if it is used in newborns.
In addition to Owens and Holodniy, who directs the AIDS research center at the Palo Alto VA, coauthors of the JAMA report on adult studies were Dr. Thomas McDonald, John Scott, and Seema Sonnad. The Annals coauthors, along with Owens, Holodniy, Scott, and Sonnad, were Dr. Alan Garber, Lincoln Moses, Dr. Bruce Kinosian, and Dr. J. Sanford Schwartz.
The research was supported by the U.S. Department of Veterans Affairs and the National Institutes of Health.
A 70-year-old man with a history of localized prostate cancer treated with whole-pelvis radiation therapy with a boost to the prostate, in conjunction with androgen deprivation therapy 7 years prior, presented with lower back pain. A bone scan revealed an area of activity in the sacrum. What is the most likely diagnosis?
