Researchers Report Conflicting Data on Cervical Cancer in AIDS

July 1, 1997

BETHESDA, Md--A review of 10 years' experience with HIV-infected patients treated at University Hospital, Newark, NJ, revealed significantly elevated levels of several types of cancers, but a surprising dearth of invasive cervical cancers, which prompted the study's lead author to suggest dropping cervical cancer from the list of AIDS-associated malignancies.

BETHESDA, Md--A review of 10 years' experience with HIV-infected patientstreated at University Hospital, Newark, NJ, revealed significantly elevatedlevels of several types of cancers, but a surprising dearth of invasivecervical cancers, which prompted the study's lead author to suggest droppingcervical cancer from the list of AIDS-associated malignancies.

Stanley H. Weiss, MD, and colleagues at the University of Medicine andDentistry of New Jersey-New Jersey Medical School and University Hospital,examined data from the hospital's cancer registry of patients treated from1986 through 1995.

They compared the proportion of specific cancers diagnosed among 118HIV-positive cancer patients (93 male, 25 female) with those occurringin 3,289 individuals not known to carry HIV, and also compared their outcomes.

"Newark is clearly an HIV epicenter," Dr. Weiss told the firstNational AIDS Malignancy Conference held at the National Cancer Institute."There are particularly high rates among women and injection drugusers, providing an interesting community in which to look at issues relatedto cancer."

Excess Kaposi's Sarcoma and NHL

The tumor types among the HIV-positive patients included 45 Kaposi'ssarcoma (38%); 22 lymphoreticular (19%), which included 16 non-Hodgkin'slymphoma (NHL), 3 Hodgkin's, and 3 leukemia; 16 lung (14%); 13 head andneck (11%); 11 gastrointestinal (9%); 3 breast (2.5%); 3 invasive cervical;and 5 miscellaneous (4%).

Not surprisingly, HIV patients showed significant excesses of Kaposi'ssarcoma and non-Hodgkin's lymphomas, with odds ratios of 66 and 8, respectively.Other cancers with statistically excessive ratios were lung (2.7), leukemia(3.4), and Hodgkin's lymphoma (4.4).

Dr. Weiss and his colleagues had previously reported an associationbetween HIV infection and cervical dysplasia, with the highest rates amongthose with the fewest CD4 cells. They had also documented that, in theirregion, historically the majority of HIV-infected women infrequently ornever had cervical cytology screening, despite physician counseling.

Said Dr. Weiss: "We therefore anticipated finding an excess ofinvasive cervical cancer." He noted that the total of 230 cervicalcancers in his series constituted 14% of the 1,654 cancer registry casesamong women. "This proportion is significantly elevated compared tonational data [8% in 1993 for uterine cancers, including invasive cervicalcancers]," he said. "Cervical cancer remains epidemic in ourregion."

To the investigators' surprise, cervical cancer among HIV-infected womenwas not increased in the registry study. The reduced odds ratio was 0.87(not statistically significant), Dr. Weiss said.

A University Hospital clinic run by Dr. Patricia Kloser has cared forseveral thousand HIV-positive women, he said. All of the cervical cancersoccurring among women at this clinic were already in the cancer registryrecords. The last HIV clinic case of cervical cancer during the study periodoccurred in 1992, "so despite much other cervical pathology, it hasnot been translating into invasive cancer here," Dr. Weiss pointedout.

Cervical cancer was added in 1993 as an AIDS-defining condition amongHIV-positive women, "despite there then being only 18 such cases describedin the world," he said.

The strength of the association with cervical dysplasia was persuasiveon a biologic basis, but, he said, "our new data suggest that we shouldreconsider this decision."

More important, Dr. Weiss noted, the findings suggest that a biologicprogression from dysplasia to invasive cancer may involve a step in thecarcinogenesis pathway that is somehow impeded or reduced among HIV-positiveindividuals.

"If so, our data offer a clue that alternative therapies, perhapsimmunologic, may exist that could reduce the risk of cervical cancer inall women," Dr. Weiss commented.

With the exception of Kaposi's sarcoma, the overall outcome among allthe HIV-infected patients in this study was "no worse than expectedfor patients with cancer of the same site and stage without known HIV infection,"Dr. Weiss said. However, whites in both groups had a longer median survivaltime than nonwhites.

The study patients were 69% black, 17% Hispanic, and 14% white. Of the118 HIV patients, 47% smoked cigarettes, 12% never smoked, and the smokingstatus of the others was unknown.

Mixed Reaction to Proposal

Dr. Weiss' proposal to drop invasive cervical cancer as AIDS definingmet mixed reactions at the conference.

"I would advocate that we do not change the definition," saidMitchell Maiman, MD, of the State University of New York Downstate MedicalSchool in Brooklyn. "Cervical cancer is by far our leading gynecologicalmalignancy, unlike the rest of the country."

Dr. Maiman described experience at his institution that contrasted strikinglywith that presented in the New Jersey study. He noted that while womenaccount for 15% of US AIDS cases and 19% of AIDS cases in New York City,they make up 26% of AIDS diagnoses in Brooklyn.

"Thus far, we're running an HIV positivity rate of 18% or 19% inwomen who come in with invasive cervical cancer under the age of 50,"he said. "Those are pretty high rates and certainly indicate thatHIV testing in all patients with cervical cancer in high-risk populationsis a good idea."

Dr. Maiman reported data drawn from an initial pool of 221 women youngerthan age 51 with cervical cancer. Of these, 149 were tested for HIV and28 were positive. HIV-infected women tended to have a more advanced stageof cervical cancer when diagnosed.

"The majority of these patients came in with advanced disease,that is, stage ID2," Dr. Maiman said. In contrast, 43% of HIV-negativepatients had early-stage disease (stage IA or IB1). "Most important,"he said, "most of the HIV-positive patients came in asymptomatic withregard to HIV infection. So if you don't test, you don't know."

The Downstate data indicate that HIV-positive women who develop cervicalcancer are not among the more immunosuppressed AIDS patients and that theiroutcome is worse.

"Sixty-four percent had CD4 counts between 200 and 500, and another16% had counts over 500," Dr. Maiman noted. "About half the patientswho are HIV negative recurred over time, while almost 90% of the patientswho are HIV positive recurred."

Downstate's experience since the study is slightly better, he added."We now have a few patients who have more prolonged courses."

Dr. Maiman also noted a high prevalence of oncogenic strains of humanpapillomavirus in women with AIDS. "We must look at HPV as almosta ubiquitous or opportunistic infection in HIV-positive patients,"he said.

Dr. Weiss strongly concurred that HPV is a serious problem in urbancommunities and is highly prevalent among HIV-infected women. Resourcesfor gynecologic care remain critically important for these regions, hesaid.

Dr. Weiss added that "a research application to explore and explainthe differences in these local findings of cervical cancer incidence isunder current development as a joint endeavor between researchers at Downstateand our group."