Veterans have higher rate of non-AIDS-related cancers

Clinicians should screen for certain cancers

New research finds that HIV-infected veterans are significantly more likely to have a non-AIDS-defining malignancy (non-ADM) than are HIV-negative veterans.1

The study found that incidence rates were highest for anal cancer, Hodgkin's, liver, and lung cancer. The research was presented at the American Society for Microbiology's 47th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), held Sept. 17-20, 2007, in Chicago, IL.1

About five years ago, some researchers began to note that HIV providers were diagnosing a number of cancers unrelated to AIDS in their patients, says Roger Bedimo, MD, MS, FACP, an assistant professor of medicine and chief of the infectious disease section of the VA North Texas Health Care System in Dallas, TX. Bedimo also is the director of the infectious diseases fellowship training program at the University of Texas Southwestern Medical Center in Dallas.

"We did one study on this in 2004, and then we went back to our database and looked at all the cancer diagnoses we had in patients," Bedimo says. "We divided them into AIDS-defining cancers, and the rest were called non-AIDS-defining cancers."

Investigators examined the incidence rates in two eras, including 1997 to the present and 1996 and earlier, he says.

"We have noticed a decline in AIDS-defining malignancies, but we saw an increase in non-AIDS-defining malignancies, and it was a clear increase," Bedimo says. "But our dataset was so small that we couldn't make a lot of inferences from it."

The problem was that investigators didn't have a population with which to compare their findings.

"It's one thing to say you have an increasing rate of cancers, but if you do not know what the denominator is, then the next step is to try to find a comparative group and see if the rates of cancer are different," Bedimo explains.

Bedimo and co-authors discovered that Veterans Administration's electronic medical records could identify HIV positive and HIV negative veterans who had malignancies.1

"Having an HIV population and a non-HIV population is what makes these findings so valuable," Bedimo says. "We can assert that the risk of many cancers, including lung cancer, liver cancer, skin cancer, and others were significantly higher in the HIV group."

The 2004 study was criticized for its findings, since it contradicted other observational research, Bedimo notes.

But it was the study's timeframe that likely made the difference, he says.

"Our rationale of why other people have not seen those increases before was because if you follow patients long enough, you'll find an increased rate of cancer," Bedimo says.

"One year is not enough time to develop cancers," Bedimo says. "We followed HIV positive patients for five years and 6.4 years for non-HIV patients."

It's taken a long time for the issue to be accepted, but it's important for research to continue and move on, he notes.

"I do have some hypotheses about why the cancer rates are higher among HIV patients," he says. "Some studies are underway to analyze and explore those hypotheses."

The first hypothesis is that many of the cancers are caused by viral infections, Bedimo says.

For example, anal cancer can be caused by HPV infection, and hepatitis C infection can lead to liver cancer.

"Even for those cancers that are not known to have a virus behind them, that's still a possibility," Bedimo says.

A second hypothesis is that HIV patients receiving antiretroviral therapy are living a lot longer than they used to, but there still are questions about whether their immune systems function completely normally under the chronic condition, he says.

"So we might be improving their immune systems, but they're still sort of vulnerable and more predisposed to illness," Bedimo says.

"It is hypothesized, and some studies have shown, that the immune system helps us deal with cells that have become abnormal, either through DNA damage or those that progress to cause cancer," he explains. "So they're either repaired or killed, and if an HIV patient has a subnormal immune system, then the immune system is not adequate enough to reign in or kill that cell that's been tossed out by the virus or an ionizing agent."

This means the HIV patients' bodies are not able to suppress or control cancer as it develops.

An example of this phenomenon can be found in transplant patients who are kept on immunosuppressant medication. Within five years, most will develop cancer — mostly skin cancers, Bedimo says.

"So it's clear that by suppressing the immune system, you can promote cancer, and HIV is suppressing the immune system," he says.

The key factor is time.

"It may take years for an HIV patient who survived to develop cancer," Bedimo adds. "The only reason we're seeing this now is because we're starting to see a large number of patients who've had HIV and are on antiretroviral drugs."

The third hypothesis is that HIV treatment is causing cancer, he notes.

"That is unlikely, but possible," Bedimo says. "If drugs are present in the body for 10-plus years and the cells have been damaged, then you might have a possibility, but we don't have evidence of this."

Bedimo and colleagues will continue to explore this connection, next looking at the cancers involved to see if they carry evidence of another type of infection, or have damage that can be attributed to something else.

"We think it's important to raise questions and look for answers," Bedimo says.

Also, clinicians need to be more suspicious of potential cancer in their HIV patients and maybe screen for certain common cancers, he suggests.

"People should be aware of the fact that these cancers are significantly higher in HIV-positive patients," Bedimo says.


  1. Bedimo RJ, et al. Incidence of non-AIDS-defining malignancies in HIV-infected vs. non-infected veterans in the HAART era: impact of immunosuppression. Abstract presented at the 47th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy, held Sept. 17-20, 2007, in Chicago, IL.