Epidemiology

Various studies have compared rates of cancer in people with and without HIV, leading to an identification of those cancers that are more common among people with HIV.

In the United States, the National Cancer Institute collaborated with population-based cancer registries and AIDS registries and analysed data on nearly 100,000 people with AIDS and over one million people with cancer. This study found that people with AIDS have a greater risk of particular cancers: a 310-fold risk of Kaposi’s sarcoma (KS); a 113-fold risk for non-Hodgkin lymphoma's (NHL); a 36.7-fold risk of blood vessel cancer; a 7.6-fold risk of Hodgkin's disease; a 4.5-fold risk of multiple myeloma; a 3.5-risk of brain cancer and a 2.9-fold risk of testicular cancer.¹

In 2001, research conducted by the Medical Research Council in the United Kingdom found that people with HIV were at greater risk of non-AIDS-defining cancers. Data from over 2000 HIV-infected people in the United Kingdom was compared with the Thames Cancer Registry for southeast England. The risk of anal cancer was greater for HIV-infected, non-African men and women. The incidence of Hodgkin's disease was also higher among people with HIV.²

Two more recent studies also found that people with HIV have a greater risk of non-AIDS-defining cancers than do non-infected people. An Italian study of over 5000 HIV-infected men and 2000 HIV-infected women found that HIV-infected individuals had a greater risk of liver cancer (19-fold risk), Hodgkin's disease (eleven-fold risk), and lung cancer (four-fold risk).³ A study comparing the incidence of cancers among several cohorts of HIV-infected men in southern Europe to that of uninfected men found that Hodgkin's disease, liver cancer, and cancer of the salivary glands were more common in HIV-infected men. HIV-infected women were more likely to have cervical cancer (17-fold risk). When considering all non-AIDS defining cancers, those who were HIV-infected had twice the risk of cancer as did uninfected individuals.³

An American study compared cancers in 871 HIV-infected women and 439 uninfected women. Among the HIV-infected women there were eight cases of non-Hodgkin's lymphoma, five cases of invasive cervical cancer, one case of KS, and twelve non-AIDS defining cancers, including four cases of lung cancer. In comparison, there were only four cases of cancer among the HIV-uninfected women, including one case of lung cancer. Statistical analysis showed that women with HIV were at a significantly greater risk of cancer than were HIV-negative women.⁴

A 2000 analysis of causes of death in HIV-positive persons in France in 2000 found that cancer accounted for 28% of the 964 deaths recorded. AIDS-related cancers represented 17% of total causes of death, most of which were non-Hodgkin's lymphoma (11%), Kaposi’s sarcoma (5%), and cervical cancer (1%). Malignancies that were not AIDS-related represented 13% of total causes of death, most of which were tumours in the lungs, liver, stomach, and anus. Systematic cancers accounted for just 17 cases, the majority of which were Hodgkin's lymphoma and myeloid leukaemias. Those who died from non-AIDS tumours were significantly more often men, smokers, and had a higher CD4 count than those dying from others causes.⁵

A 2005 update of this analysis revelead that the proportion of deaths in people with HIV receiving care in France had risen to 34%, a significant increase compared to the 2000 survey, and half of these deaths were due to non-AIDS-defining cancers, also a significant increase compared to 2000.⁶

Can the excess of particular cancers among HIV-infected individuals be directly attributed to HIV, or are other factors at play? Can HAART modify this risk?

An Australian case-control study of 438 patients published in 2000 showed that B-cell stimulation and prolonged immune deficiency are risk factors for non-Hodgkin's lymphoma (NHL) in people with AIDS. The authors concluded that unless immune deficiency in HIV-infected persons can be reversed, NHL would be an increasing cause of morbidity and mortality.⁷

A 2002 study linking data from cancer, HIV, and AIDS registries in Australia showed a decreasing incidence of AIDS-related KS, and to a lesser extent, non-Hodgkin lymphoma, since the widespread use of antiretroviral therapy. NHL was the most common AIDS-associated cancer in Australia.⁸

Researchers also looked for an association between immune deficiency and cancer risk and at the incidence of non-AIDS-defining cancers before and after a diagnosis of AIDS. There were significantly increased rates for cancer of the lip (IR 2.26), anus (IR 37.1), connective tissue (IR 9.7), Hodgkin's disease (IR 7.85), myeloma (IR 4.15), and leukaemia (IR 3.4) than might be expected in the general population. A significant trend of increasing relative risk of cancer with increasing time since HIV diagnosis was found for Hodgkin's disease and multiple myeloma.⁸

In both the HIV/AIDS and solid organ transplant recipient cohorts, there was a pattern of increased cancer risk related to immune deficiency. Cancers of an infectious origin were the most common; epithelial cancers did not show an increased incidence.

A meta-analysis presented in 2007 identified seven studies of cancer incidence in people with HIV from the United States, United Kingdom, Australia, Switzerland, and Italy, and five studies of cancer incidence in people immunosuppressed after organ transplants from Scandinavia, Canada, and Australia. The data were derived from cohort studies in which cancer incidence was confirmed by reference to national cancer registries, ensuring the reliability of the data captured. All studies were published in peer-reviewed medical journals.⁹

The cohorts included 444,172 people with HIV and 31,977 transplant recipients. There were 42,000 cancers in the HIV-positive group and 3000 in the transplant group. The researchers wanted to compare cancer rates in people with HIV and in immunosuppressed people in order to see if there were any differences with expected rates of cancer in the general population; additionally, to see whether any common patterns emerged that might be linked to long-term immunosuppression.

The meta-analysis found that for 20 out of 28 cancers, incidence rates in both groups were higher than that those found in the general population. Most of these cancers were from a known infectious cause: human papillomavirus (cervix, vulva, vagina, penis, anus, oral cavity, and pharynx), Epstein-Barr virus (Hodgkin's and non-Hodgkin lymphoma), HHV-8 (Kaposi’s sarcoma), hepatitis C (liver), and helicobacter pylori (stomach). The researchers concluded that immune deficiency, rather than other risk factors for cancer, led to high incidence rates in both of these groups. They also predicted that infection-related cancers would be an increasingly frequent complication in long-term HIV management.

The United States AIDS-Cancer Match Registry Study Group conducted a comprehensive analysis of cancers found in over 300,000 persons with HIV/AIDS in the period from 1978 to 1996 to see which cancers were influenced by immune suppression.² The relative risk (RR) of particular cancers was assessed by dividing the number of cancers among people with HIV by the number of cancer cases in the general population. Cancers were deemed to be influenced by immune deficiency if relative risk rose over three periods: 60 to 27 months before AIDS, four to 27 months before AIDS, and four months before AIDS to 27 months post-AIDS.

Hodgkin's disease seemed to be linked to immune suppression, as were possibly, lip cancer and testicular seminoma. In Hodgkin's disease (particularly mixed cellularity and lymphocytic depletion subtypes), the relative risk (RR) went from 2.6 distant pre-AIDS to 6.7 early post-AIDS. Lip cancer rates rose from 1.6 in the distant pre-AIDS period to 5.8 post AIDS. Testicular seminoma cancer RR rose from 0.7 distant pre-AIDS to 1.8 post AIDS.

A review of the D:A:D (Data Collection on Adverse Events of Anti-HIV Drugs) international study found a higher rate of death from non-AIDS-defining cancers (63%) than from AIDS-defining cancers (37%).¹⁰ 

In the 112 deaths from AIDS-defining malignancies, 73% were due to NHL and 25% to Kaposi's sarcoma. The most common fatal non-AIDS-defining cancers were lung, gastrointestinal, haematological, and anal. In the 193 deaths due to non-AIDS-defining cancers, lung cancer was most frequent (32%), followed by gastrointestinal tract cancer (21%), haematological cancer (10%) and anal cancer (10%). In this study, HIV RNA levels did not influence the risk of fatality from either type of malignancy; however, the risk of dying from AIDS-defining and non-AIDS-defining cancer was increased at lower CD4 cell counts. Other risk factors thought to influence mortality included smoking and chronic hepatitis B infection.

A retrospective analysis of cancers occurring in two large US cohort studies, the HIV Outpatient Study (HOPS) and the Adult/Adolescent Spectrum of Disease study (ASD), between 1991 and 2002 found that participants on HAART experienced a significant decrease in cases of Hodgkin's lymphoma, melanoma, and colon, lung, oropharyngeal, and testicular cancers. Similarly, in AIDS-defining cancers, there was a decrease in the incidence of cervical cancer, Kaposi's sarcoma, and NHL in the HAART era. Antiretroviral therapy was independently associated with a decreased risk in all types of cancers with the exception of melanoma and prostate cancer. Although cancer incidence in the HIV population is higher than in the general population, the gap has narrowed in the HAART era.¹¹ 

In 2008, results from the French ANRS CO3 Aquitaine Cohort of over 4000 patients in the period from 1998 to 2006 were presented. Investigators found that, independent of CD4 cell count, longer exposure to uncontrolled HIV RNA was associated with a higher risk of AIDS-defining cancers.  HIV RNA over 500 copies/mL increased the relative risk (RR) by 1.20 per year of exposure. A longer duration of CD4 counts less than 200 cells/mm³ increased RR by 1.35 per year of exposure. Exposure to combination ART (cART) reduced the incidence of AIDS-defining cancer per year by 0.82 RR.

CD4 cell count (but not viral load or use of cART) was associated with a higher incidence of non-AIDS-defining cancers. A longer duration of CD4 cell counts below 500 cells/mm³ resulted in an increased RR of 1.11 per year of exposure; a longer duration of cell counts less than 200/mm³ carried a RR of 1.16 per year of exposure. Overall, prolonged immunosuppression was associated with a higher incidence of both AIDS-defining and non-AIDS defining cancers.¹²

A larger French analysis of 52,278 patients with 253,353 person years of follow-up published in 2009 found a clear relationship between current CD4 cell count below 500 and an elevated risk of both AIDS-defining cancers and cervical cancer, anal cancer, Hodgkin's lymphoma, lung cancer and liver cancer. Anal cancer was also associated with a greater duration of immunosuppression below 200 and viral load elevation above 100,000 copies/ml.¹³

Most experts now agree that even moderate, but prolonged immunosuppression can increase the risk of non-AIDS defining cancers, particularly those cancers due to infectious causes, as well as non-AIDS-defining cancers. Meta-analysis shows that the risks of non-AIDS defining cancers related to infectious agents or smoking are elevated in people with HIV.¹⁴ However, analysis of patients receiving care in California's Kaiser Permanente managed-care programme shows that although the risk of non-AIDS cancers of an infectious cause is higher in people with HIV, the risk of anal cancer has declined by 6% per year since 1996.¹⁵