CROI: Risk of lymphomas depends on cumulative viral load and latest CD4 counts

This article is more than 16 years old. Click here for more recent articles on this topic

The cumulative effect of having detectable viral load over time in individuals on non-suppressive highly active antiretroviral therapy (HAART) doubles the risk of Burkitt or Burkitt-like lymphoma compared with other high-grade non-Hodgkin's lymphomas, according to a German study presented to the Fifteenth Conference on Retroviruses and Opportunistic Infections in Boston earlier this week.

However, it is CD4 count, rather than viral load, that appears to primarily influence the risk of non-Hodgkin's lymphomas other than Burkitt or Burkitt-like lymphoma, the study found..

Following the widespread use of highly active antiretroviral therapy (HAART) in wealthier nations more than a decade ago, a decrease in the incidence of AIDS-related illnesses was observed, including that of non-Hodgkin’s lymphoma. However, there are few data on the incidence of non-Hodgkin’s lymphoma in the HAART era and the effects of HAART on incidence over time.



A type of cancer that starts in the tissues of the lymphatic system, including the lymph nodes, spleen, and bone marrow. In people who have HIV, certain lymphomas, such as Burkitt lymphoma, are AIDS-defining conditions.

non-Hodgkin lymphoma

A group of lymphomas (cancers of the lymphatic system). The many types of non-Hodgkin lymphoma (NHL) are classified according to how fast the cancer spreads. Although the symptoms of NHLs vary, they often include swollen lymph nodes, fever, and weight loss. Certain types of NHLs, such as Burkitt lymphoma and immunoblastic lymphoma, are AIDS-defining cancers in people with HIV.


The presence of virus in the blood.


statistical significance

Statistical tests are used to judge whether the results of a study could be due to chance and would not be confirmed if the study was repeated. If result is probably not due to chance, the results are ‘statistically significant’. 


Involving the brain.

On Monday, Dr Alexander Zoufaly of the University Medical Centre Hamburg-Eppendorf presented data on behalf of the CLINSURV Study group. Their analysis aimed to determine risk factors for the development of AIDS-related lymphoma in individuals taking HAART, which in this cohort was defined as three or more drugs of any class.

The CLINSURV cohort has enrolled 18,451 HIV-positive individuals from 15 centres throughout Germany since 1999, in order to collect retrospective and prospective data. For this analysis, the 6,458 individuals on HAART at the end of 2006 were included, resulting in a total of 28,125 person-years of follow-up.

A total of 94 individuals were found to have incident non-Hodgkin's lymphoma (78 AIDS-related non-Hodgkin's lymphomas, and 16 primary central nervous system lymphomas). Incident non-Hodgkin's lymphoma was defined as being diagnosed for the first time at least 30 days after inclusion in the cohort, and after at least 30 days on HAART. This resulted in an incidence rate of 3.3 lymphomas per 1000 person-years.

The investigators were particularly interested in the role of HIV viral load, and used two different viraemia models for their risk factor analyses. In their categorical model, if an individual’s viral load was above 500 copies/ml for 75% of the time or longer, this was deemed to be persistent viraemia. A second model, termed the online model, was used for their final analysis. Here, they calculated the cumulative area under the curve (AUC) of an individual’s viral load over time.

In their first multivariate analysis, using the categorical model, the following risk factors were identified (all p

  • an AIDS diagnosis prior to commencing HAART (Hazard Ratio 5.29; 95% CI, 3.48-8.06)
  • at least 75% of viral load measurements >500 copies/mL (HR 3.41; 95% CI, 2.17-5.35)
  • latest CD4 count
  • HIV acquired via sex between men (HR 1.86; 95% CI, 1.19-2.90) and
  • age, per ten years older (HR 1.40, 95% CI, 1.17 to 1.68).

Before performing their final analysis, the investigators confirmed the non-Hodgkin's lymphoma genotypes diagnosed in the cohort. They found 40 cases of non-Burkitt, high-grade non-Hodgkin's lymphoma; 15 cases of AIDS-associated non-Hodgkin's lymphoma; twelve cases of Burkitt or Burkitt-like lymphoma; nine cases of primary cerebral non-Hodgkin's lymphoma; and nine cases of ‘other’ lymphomas. A further nine cases could not be confirmed and were dropped from the final analysis.

In the final analysis, using the online model for viral load, investigators found that three risk factors remained statistically significant for the development of all lymphomas (all p

  • age, per ten years older (HR 1.4; 95% CI, 1.1-1.7)
  • cumulative viraemia (HR 1.4; 95% CI, 1.1-1.7) and
  • latest CD4 count

However, when they analysed the risk factors for different types of lymphomas, they found that cumulative viraemia doubled the risk for Burkitt or Burkitt-like lymphoma (HR 3.0; 95% CI, 1.5-5.3; p

Conversely, they found that latest CD4 count (either between 200-350, or below 200 cells/mm3) was not statistically significant for Burkitt or Burkitt-like lymphoma, but was highly associated with non-Burkitt high-grade non-Hodgkin's lymphoma (latest CD4 count

Only latest CD4 count below 200 cells/mm3 remained a significant risk factor for primary cerebral non-Hodgkin's lymphoma (HR 15.2; 95% CI, 1.6-141.2; p=0.02).

A previous study, from France, published in Clinical Infectious Diseases in 2006, had found that prolonged high viral loads were associated with an increased risk of non-Hodgkin's lymphoma independent of CD4 cell counts in individuals not on HAART, although this study did not analyse the relative risks for different lymphoma genotypes.

Discussing the role of cumulative viraemia, Dr Zoufaly said that he and his colleagues hypothesed that ongoing HIV viraemia might trigger ongoing immune activation which could then lead to increased B-cell turnover. This, in turn, might eventually lead to the development of lymphoma in the presence of immunodeficiencya and co-infections, such as Epstein-Barr Virus (EBV) and Human Herpes Virus-8 (HHV-8), both of which have been associated with certain lymphomas.

During the question and answer session that followed, Dr Zoufaly said that although HIV acquired via sex between men had been a risk factor in the first analysis, it may be have been an artefact of their second analysis – which excluded nine cases – that resulting in this being no longer a statistically significant risk factor.

He concluded by reiterating that in this study, age, latest CD4 count and cumulative viraemia were strong predictors for the development of AIDS-related lymphoma, and that the impact of viral load may be higher for Burkitt lymphoma than non-Burkitt lymphomas.

Since the only potentially modifiable risk factor identified is ongoing viraemia, he recommended that a clinical strategy to optimise HAART at any time-point with respect to viral suppression could help to minimise the incidence of AIDS-related lymphoma.

Last month, a Swiss study published in AIDS found that use of suppressive HAART reduced the risk of non-Hodgkin’s lymphoma by half within the first five months of use, and the risk remained extremely low after ten years of suppressive HAART.


Zoufaly A et al. Insufficient virus suppression during HAART is a strong predictor for the development of AIDS-related lymphoma: German CLINSURV Cohort. Fifteenth Conference on Retroviruses and Opportunistic Infections, Boston, abstract 16, 2008.