Conceived and designed the experiments: AJS AJ FD. Performed the experiments: AJS FD. Analyzed the data: AJS AJ EB dkE fT TL MAF PR PN DM FD. Wrote the paper: AJS AJ EB dkE fT TL MAF PR PN DM FD.
The authors have declared that no competing interests exist.
With the lengthening of life expectancy among HIV-positive subjects related to the use of highly active antiretroviral treatments, an increased risk of cancer has been described in industrialized countries. The question is to determine what occurs now and will happen in the future in the low income countries and particularly in sub-Saharan Africa where more than two-thirds of all HIV-positive people live in the world. The objective of our paper is to review the link between HIV and cancer in sub-Saharan Africa, putting it in perspective with what is already known in Western countries.
Studies for this review were identified from several bibliographical databases including
Studies conducted in sub-Saharan Africa show that HIV infection is not only strongly associated with AIDS-classifying cancers but also provided some evidence of association for other neoplasia. African countries need now to implement well designed population-based studies in order to better describe the spectrum of AIDS-associated malignancies and the most effective strategies for their prevention, screening and treatment.
Infection with the human immunodeficiency virus (HIV) entails an increased risk of developing cancer
The objective of our paper is to review the evidence on the link between HIV and cancer in sub-Saharan Africa, putting it in perspective with what is already known in Western countries. A better understanding of the link between HIV and cancer is indeed necessary for the improvement of care of HIV-infected subjects as well as may shed understanding on the mechanisms of cancer occurrence in the general population.
Studies from Africa were identified from several bibliographical databases (
Similarly, references from Western countries were identified from
Measures of association between HIV and cancer varied in the published materials, depending on study design. All measures, be they Standardized Incidence Ratio (SIR) or Hazard Ratio (HR) in cohort studies or Odds Ratios (OR) in case-referent studies can be taken as estimates of relative risk (RR) and express by how much the risk of cancer is increased in HIV-positive
Since the first report by Hymes
In sub-Saharan Africa, KS has been much more frequent than in Western countries even before the HIV epidemic, and the African-endemic form has been described since the 1960s.
NHL is the second most common malignant disorder associated with HIV infection worldwide. After the first description in 1982
In Western countries, the introduction of HAART induced a significant reduction of the NHL incidence, although less marked than for KS incidence. A first international meta-analysis based on 23 prospective studies compared the incidence of NHL between 1992–96 and 1997–99
Several NHL studies have been conducted in sub-Saharan Africa. Data coming from a cancer registry in Uganda showed an overall increase of the NHL annual incidence in the nineties. In Kampala, Uganda, the NHL annual incidence of just more than 3 cases per 100 000 py, stable up to the early nineties, increased to 7.4 per 100 000 py during the 1995–1997 period
In children, malignant lymphoma is mostly represented by endemic Burkitt lymphoma associated with early and persistent Epstein-Barr Virus (EBV) infection. A case-referent study conducted in Uganda by Parkin
According to the World Health Organization (WHO), ICC is the second most common cancer in women worldwide, and is more frequent in low income countries
Cervical cancer is much more frequent in sub-Saharan Africa, where it has been highly prevalent even before the HIV epidemic, than in Western countries. Data coming from different cancer registries showed contradictory trends. While the Ugandan tumour registry finds increased cervical cancer rates on the parallel rise of the HIV epidemic
* squamous intraepithelial lesions † low-grade SIL ‡ high-grade SIL.
However, findings from Mbulateye
Squamous Cell Carcinoma of the Conjunctiva (SCCC) is a very rare tumor, especially in Western countries where its incidence has been estimated at 0.03 cases per 100 000 py
SCCC has been recognized to be associated with HIV in South Africa, Rwanda, Malawi and Uganda since the early 90's, with ORs ranging from 2.6 (95% CI 1.4–4.9) to 13.1 (95% CI 4.7–37.6)
The incidence of HL is clearly increased among HIV-infected patients
The record-linkage study conducted in Uganda showed an association between HL and HIV/AIDS in Africa with a SIR of 5.7 (95% CI 1.2–17)
HIV-infected persons have an elevated risk for lung cancer, varying from two to seven-fold higher than the general population according to several cohort studies
The record linkage study conducted in Kampala, Uganda, showed a higher incidence of lung cancer in HIV-infected people with a SIR of 5.0 (95% CI 1.0–15.0)
The most common risk factors for hepatocellular carcinoma (HCC) are hepatitis B or C virus (HBV, HCV) infection, chronic alcohol intake as well as aflatoxines in the diet in Africa. Although the etiological link between HCC and HBV and HCV is well documented, the influence of HIV infection and immune status on the development of this type of liver cancer is unclear and particularly pertinent for Africa
In some parts of Africa, especially in West Africa where HBV infection is endemic, liver cancer is highly prevalent. In a publication from the Gambia Liver Cancer Study, HCC was the commonest cancer and 15% to 20% of the total population are chronic HBV carriers
Early studies on breast cancer showed that the incidence was lower in HIV-infected than HIV-negative women
The reduced breast cancer incidence in HIV-infected patients initially described in Western countries has also been found in Africa. Data collected between 1968 and 1996 by the Tanzanian cancer registry showed a significant reduction in the incidence of breast cancer in women as well as men with HIV
Anal cancer is closely associated with HIV infection. Bower
Some HPV-related malignancies such as anogenital cancers (other than cervical cancer) have also be shown to be related to HIV infection with ORs of 4.8 (95% CI 1.9–12.2) and 2.2 (95% CI 1.4–3.3) in the case-referent studies conducted in South Africa
Recent findings from a case-referent study conducted by Stein
Publications that aimed to assess the association between AIDS-defining cancers and HIV/AIDS in sub-Saharan Africa showed that KS and NHL, known to be directly linked to the severity of immunosupression, were strongly and significantly associated with HIV/AIDS. This association was much weaker for ICC.
For KS, the reduced association observed among African HIV-infected populations compared to the one of industrialized countries might be related to the high background risk of KS among HIV-negative individuals in Africa and the higher seroprevalence of HHV8 reported in sub-Saharan Africa or the influence of other co-factors
A weaker association was found between HIV and NHL in sub-Saharan Africa compared to industrialized countries. This result might be partly explained by the under-ascertainment of NHL in middle and low-income settings, particularly marked for HIV positive subjects. Indeed, from a clinical point of view, as this malignancy requires a costly histological diagnosis it is possible that many patients presenting with polyadenopathies might have been classified as tuberculosis and died without any histological exploration. In addition, NHL cerebral localization has to be differentiated from current infectious opportunistic conditions such as cryptoccocosis (which is quite easy to diagnose), cerebral toxoplasmosis or cytomegalovirus (CMV) infection which remain challenging diagnoses in countries with limited facilities. In this respect, Lucas
Although there appears to be a clear association between HIV infection and the occurrence of CIN, ICC has been somewhat less strongly associated with HIV in sub-Saharan Africa than in Western countries. Like NHL, this lower association could be due to the competing risk of mortality from other conditions associated with HIV, particularly in settings where HAART is not widely available
The association between HIV infection and HL in Uganda was consistent with previous findings from other studies conducted in Western countries. Inversely, the association between SCCC and HIV infection was primarily and more frequently documented in sub-Saharan Africa rather than Western countries. Even if this difference is partly explained by the higher exposure to solar UV radiations, a well known risk factor of SCCC, the impact of HIV infection on the occurrence of malignancies found in Southern countries may not be directly extrapolated from the Western one. Indeed, confounding factors such as local climatic conditions and particular innate immunity should be sought. Certain non AIDS-defining cancers like skin cancers are also thought to occur with increased frequency or altered course in patients with HIV. This is the case of squamous carcinoma of the skin which shares close patterns with SCCC
There is a real need to describe the patterns of malignancies in African populations as they may not share the same genetic background and as they are not exposed to the same degree to some carcinogenic factors such as solar UV radiation, endemic HBV and EBV infections as well as food contaminations by aflatoxins among others
Women now constitute almost half of all AIDS cases. With the increasing use of HAART in sub-Saharan Africa, the overall health and survival of HIV individuals is expected to improve in the coming years. Based on epidemiologic data from Western countries and Africa, HIV infection is not considered as permissive for breast cancer
Results from non AIDS-defining cancers available from the few studies that were conducted so far are based on small-sized populations. This lack of power partly explains the wide confidence intervals associated with estimated SIRs and ORs in these studies. As a number of non AIDS-defining cancers occur with a relatively low incidence, available studies are limited to evaluate the association between cancer and HIV. There is a need for wide and prospective data collection in order to monitor cancer events among HIV-infected persons in low-income settings.
The incidence rate ratio is the most appropriate statistic for making comparisons between populations for which incidence rates are computed. For that purpose the main method to quantify cancer occurrence in an HIV-infected cohort relies on linkage studies between cancer registries and HIV/AIDS databases. Such studies are actually conducted in Western countries
So far, the epidemiological studies have focused on the strength and precision of the associations between the different malignancies and HIV infection. Not only the new studies will have to consolidate some of these estimates and sometimes reconciliate controversial findings for specific tumors, they will also need to estimate for the first time the general population attributable risk and more importantly perhaps the attributable risk in the exposed (HIV-infected) population that will may deserve focused interventions.
In Western countries, as HAART has improved the survival of HIV/AIDS patients, the cumulative risk of developing, and dying from, non AIDS-defining cancers is likely to increase. In sub-Saharan Africa, morbidity and mortality are mainly caused by transmissible diseases and complications of HIV-infection are predominantly of infectious origin. With the worldwide scale-up of HAART, chronic conditions such as cancer are likely to represent a growing part in the burden of the HIV-associated morbidity. Some studies in sub-Saharan Africa show that HIV infection is already associated with AIDS-classifying cancers and others provide some evidence of an association between HIV and non-AIDS classifying neoplasia. African researchers need to implement now large and well-designed population-based studies in order to better define the spectrum of HIV-associated malignancies and the most effective strategies for their prevention, screening and treatment.
The authors thank Dr. Elisabeth Poulet, Dr. Mathias Bruyant and Dr. Charlotte Lewden for helpful discussions at the time of the initial preparation of this manuscript which started as a Master student assignment (FT, TLM, MAF) and was part of a Public Health internship (EB) under the direction of AJS at the ECP team of the Inserm U 897 of the Victor Segalen Bordeaux 2 University. The authors are also indebted to the ISPED documentary team for its bibliographical assistance and to Luc Letenneur and Dr. Jean-François Tessier for their dedicated help in the very final stages of manuscript preparation.