Cancers in people with HIV: who gets them, and which ones?

A very low CD4 count in the past, and a history of smoking, are the most consistent risk factors to emerge from large studies of risks for developing both AIDS-defining and non-AIDS-defining cancers, researchers reported at the 19th Conference on Retroviruses and Opportunistic Infections in Seattle this week.

In particular, studies found that cancers were concentrated in older people, those who had experienced very low CD4 cell counts before starting highly active antiretroviral therapy, and those who had very high viral loads before starting treatment – but risk factors differed according to the type of cancer, indicating the complex interactions between HIV disease, ageing and the era in which individuals diagnosed with cancer first began HIV therapy.

Fears that people with HIV might suffer a disproportionate burden of cancers as they age have been expressed as a result of epidemiological evidence from a number of large cohort studies indicating a higher rate of some types of cancers, particularly those with an infectious cause, in people living with HIV.

These cancers are known as non-AIDS-defining malignancies. They are not AIDS-defining illnesses and can occur even at high CD4 cell counts, just as in the rest of the population. They include organ cancers such as liver and lung cancer, tissue cancers such as skin cancers, oral and pharyngeal cancer, and anal cancer caused by HPV. (Invasive cervical cancer is an AIDS-defining cancer.)

A number of analyses of large patient populations were presented last week at the 19th Conference on Retroviruses and Opportunistic Infections in Seattle, providing reassurance regarding some aspects of cancer risk in people living with HIV, but also underlining the importance of vigilance for potential cancers in people with a past history of severe immunosuppression.

Oral and pharyngeal cancers

A review of the North American AIDS Cohort Collaboration, which collates data on over 110,000 people receiving HIV care in the United States and Canada, looked at the incidence of oral and pharyngeal cancers in 51,151 HIV-positive individuals.

These cancers are relatively uncommon in the general population, despite the level of media attention given to HPV-related oral cancers in recent years, and are also uncommon in people living with HIV. The researchers found just 41 new cases of such cancers in their cohorts, and combined with cases already present at the time of HIV diagnosis (prevalent cases), this represented a rate of 21.7 cases per 100,000 person-years. This means that for every 100,000 people with HIV receiving care for a full year, 21 people might be expected to develop one of these cancers.

Only 28% of the incident cases were taking antiretroviral therapy when they were diagnosed with one of these cancers, at a median CD4 cell count of 324 cells/mm³. All cases occurred in people who had a history of smoking.

The researchers also examined whether the risk of oral and pharyngeal cancers has increased over time. They found that the age-standardised incidence had increased significantly from 10.6 cases per 100,000 patient-years  in the period 1996-2000 to 30.1 cases per 100,000 patient-years in the period 2006-10 (p = 0.03) whereas in the general population the incidence of these cancers has shown a decreasing trend over the same period.

However this big difference appears to have been driven chiefly by cancers in people over 50 with CD4 counts below 350; cancer rates in other groups remained stable over time.

Examination by cancer type showed that it was non-oropharyngeal cancers that had increased over time, whereas the rate of oropharyngeal cancers had remained stable since 1996. Non-oropharyngeal cancers include cancers of the anterior epiglottis, the vallecula and the mouth other than the base of the tongue. Once again, the increase in incidence occurred in the over-50s rather than in those with low baseline CD4 cell counts.

Although all cancers were diagnosed in people with a history of smoking, presenter Alison Abrahams of Johns Hopkins University Bloomberg School of Public Health noted that the lack of information on smoking intensity and alcohol consumption was a limitation of this analysis.

Non-Hodgkin lymphoma

Although rates of the AIDS-defining cancer non-Hodgkin lymphoma (NHL) have declined since the introduction of HAART in 1996, rates reported in cohort studies continue to be between four and 23 times higher than in the general population. This variation has been variously attributed to calendar time, immunosuppression, age and viral load, as well as location and demographics.

Chris Achenbach presented data on NHL incidence in the CFAR network of Integrated Clinical Systems representing eight large HIV clinics in the United States serving diverse populations and comprising just over 10,000 patients in care. The cohort was largely diagnosed in the modern era of HAART, with a median year of viral load suppression of 2004. They had a median nadir CD4 count of 180 cells/mm³ and peak viral load of 4.6 log₁₀ copies/ml. Eighty-five per cent achieved viral load suppression below 50 copies/ml but 54% also experienced at least one viral load rebound above 500 copies/ml.

The primary analysis of this cohort examined the incidence of NHL after viral suppression. Seventy-six cases of NHL occurred after initial viral suppression, of which 35 occurred while viral suppression was still maintained. The majority of lymphomas were diffuse large B-cell lymphoma cases.

Fifty-four per cent of cases occurred in the first two years after viral suppression. The highest rate of lymphoma was observed in those with a nadir CD4 count below 50 cells/mm³ (20 events per 10,000 person-years compared to 15 events in those with a nadir CD4 cell count above 200), and Kaplan-Meier analysis showed a higher incidence of NHL cases in this stratum during the first five years after viral suppression.

Overall the study found that the rate of NHL in this HIV-positive cohort was at least seven times higher than in the US general population, at 15 cases per 10,000 patient-years of follow-up, compared to rates of between 1.7 and 2 cases per 10,000 patient-years in HIV-negative cohorts. Although male gender and white race were the strongest predictors of NHL, as in the general population, the level of immune suppression was also predictive after controlling for age, sex and race (each 100 cell increment in current CD4 cell count reduced the risk of NHL by 40%). There was also a modestly increased risk in those with detectable viral load after having previously achieved viral suppression, and a more pronounced increase in risk among those with a viral load between 50 and 500 copies/ml in the population diagnosed with NHL with a viral load below 500 copies/ml.

The authors of this study concluded that although earlier detection of lymphoma, together with earlier HIV diagnosis and treatment, may reduce the incidence of NHL, an elevated risk of lymphoma is likely to persist in people living with HIV due to immune dysregulation that cannot be normalised by current antiretroviral therapies, and due to the state of B-cell activation that persists in people with HIV infection. Understanding how these factors contribute to the development of NHL in people with HIV, and how these states can be corrected, is likely to make an important contribution to reducing the burden of cancers in people with HIV.

Non-AIDS defining malignancies in the D:A:D cohort study

The D:A:D (Data collection on Adverse events of anti-HIV Drugs) cohort study collects data on more than 49,000 patients in 33 countries in Europe, North America and Australia. Investigators analysed risk factors for non-AIDS-defining malignancies diagnosed at participating centres between January 2004 and January 2010.

During 176,775 person-years of follow up, 880 patients developed a new non-AIDS-defining malignancy, an incidence of 4.98 cases per 1000 person-years of follow-up. In other words, in a clinic cohort of 1000 patients followed for a full year, one might expect to see five non-AIDS cancers in one year.

The three most common malignancies were: lung cancer (0.79 per 1000 py), Hodgkin’s lymphoma (0.63 per 1000 py) and anal cancer (0.45 per 1000 py). A wide range of other cancers were seen at much lower incidences. The analysis did not present any comparison with rates seen in the general population, presumably because this was a multinational cohort study, making comparison with nationally derived incidence data unsound.

The median age of diagnosis of a malignancy was 50 years, and diagnosis occurred at a median CD4 count of 392 cells/mm³ (interquartile range 245-580). The median nadir CD4 count of diagnosed persons was 127 cells/mm³ (interquartile range 49-245), and the median viral load was 50 copies/ml (undetectable). Hepatitis C co-infection was present in 8.5% and hepatitis B in 6.3%. Just over one-third of those diagnosed was a current smoker (34.4%).

Multivariable analysis which adjusted for demographic variables, prior history of malignancies, co-infection, body mass index and prior AIDS diagnosis found a modest protective effect of latest CD4 count (3% per 50 cells, risk ratio 0.97, 95% confidence interval 0.95-0.98, p = 0.001) and a more pronounced effect of nadir CD4 cell count below 100 cells/mm³ (RR 1.22, 95% CI 1.03-1.44, p = 0.02). Duration of immunosuppression below 200 cells/mm³ also had a modest impact on risk (RR 1.04, 95% CI 1.02-1.05, p = 0.0001), as did the duration of immune recovery above 200 cells/mm³ (RR 0.94 per year, 95% CI 0.92-0.97, p = 0.001).

In the cases of lung cancer, anal cancer and Hodgkin’s lymphoma, which were analysed separately, higher current CD4 count was correlated with somewhat more pronounced reductions in risk for each cancer (lung 0.93, HL 0.89, anal 0.93) per 50 cell increment in CD4 cell count.

Signe Worm, presenting the results on behalf of the D:A:D collaboration, noted that a low nadir CD4 count remained associated with a persistent increase in risk, but that for the subgroup of patients with CD4 cell increases from nadir to above 200 cells/mm³, a low CD4 count nadir ceases to become relevant to current risk of developing a non-AIDS malignancy over time.