Introduction
Although primary central nervous system (CNS) lymphoma was included among the first acquired immunodeficiency syndrome (AIDS)-defining illnesses,[1] it was not until 1985 that systemic lymphoma, of high or intermediate grade, was recognized as AIDS-defining.[2] The later recognition of lymphoma as part of the spectrum of AIDS is consistent with the fact that lymphoma usually occurs rather late in the course of HIV infection.[3] Thus, while serving as the initial AIDS-defining diagnosis in approximately 3% of patients, lymphoma is the cause of death in approximately 16% to 20% of HIV-infected individuals.[4,5] Moreover, the relative risk of developing lymphoma increases substantially in patients who have already been diagnosed with full-blown AIDS.[6,7] Hence, subsequent to an earlier diagnosis of AIDS, the risk of immunoblastic lymphoma increases approximately 627-fold, the risk of Burkitt’s lymphoma approximately 220-fold, and the risk of diffuse large cell lymphoma 145-fold over what is expected in the general population.[6,7] Notably, when linking cancer and AIDS registries, Goedert and colleagues found that even the risk for low-grade lymphoma increased 14-fold in individuals who had already been diagnosed with an AIDS-defining condition.[6,7] As these data suggest, lymphoma is usually diagnosed in patients with markedly decreased CD4 cell counts, consistent with prolonged periods of HIV infection and subsequent immunosuppression.[3]
Risk Factors for AIDS-Related Lymphoma
Unlike Kaposi’s sarcoma, which occurs primarily in men who have sex with men, lymphoma is seen in all population groups at risk for HIV,[3] and like de novo lymphoma, which occurs in HIV-negative individuals, AIDS-related lymphoma is more common in men than in women.[8] All age groups are affected, and lymphoma is the most common malignancy in HIV-infected children.[9] Epidemiologic studies have failed to identify major environmental factors associated with lymphoma among HIV-infected individuals.[10-12] However, host factors may be operative. Thus, HIV-infected patients who are heterozygotes for the CCR5-delta-32 deletion are statistically less likely to develop lymphoma,[13] whereas those with stromal cell-derived factor 1 (SDF-1) mutations (eg, the 3¢A variant) are statistically more likely to develop lymphoma.[14]
An interplay between genetic and other host factors may also play a role in the development of AIDS-related lymphoma. Grulich and colleagues[15] recently reported the results of a case-control study of 219 patients with AIDS-related lymphoma, who were compared with 219 HIV-infected controls, matched for CD4 cell counts and date of specimen collection. On multivariate analysis, an increased risk of lymphoma was found with longer duration of immunodeficiency (as determined by the time from seroconversion) and lower CD4 cell counts 1 year before the diagnosis of lymphoma. In addition, chronic B-cell stimulation was also predictive of lymphoma, using a higher serum globulin level (as a surrogate for higher immunoglobulins) as a marker of B-cell stimulation. Additionally, although controversial and clearly requiring validation in other studies, the use of acyclovir (Zovirax) has been associated with a decreased risk for AIDS-related lymphoma in one small case-control study of 29 patients,[16] while another, larger study yielded the opposite results.[15]
In trying to predict the incidence of AIDS-related lymphoma in the years ahead, it will be necessary to take all of these factors into account.
Impact of Highly Active Antiretroviral Therapy on Incidence
Role of Immunosuppression
Recent data from several large series[15,17] have demonstrated a substantial decline in the median CD4 cell count among patients with newly diagnosed AIDS-related lymphoma. In the early years of the epidemic, the median CD4 cell count of the patients described by Levine and colleagues was 177/dL. In patients diagnosed recently, however, the median CD4 count was only 55/dL.[17] This recent drop in the median CD4 count in patients with newly diagnosed AIDS lymphoma may also be seen in several large, prospective national trials conducted through the AIDS Clinical Trials Group (ACTG). Patients with newly diagnosed AIDS-related lymphoma accrued to a phase II study of low-dose m-BACOD (methotrexate, bleomycin(Drug information on bleomycin) [Blenoxane], doxorubicin(Drug information on doxorubicin) [Adriamycin], cyclophosphamide(Drug information on cyclophosphamide) [Cytoxan], vincristine [Oncovin], dexamethasone(Drug information on dexamethasone)) between June 1987 and November 1988 had a median CD4 cell count of 150/dL (range: 16 to 1,079/dL).[18] In a subsequent ACTG study comparing low-dose with standard-dose m-BACOD,[19] in which patients were accrued from February 1991 to October 1994 and eligibility requirements were similar, a median CD4 cell count of approximately 100/dL at study entry was reported.[19] The drop in the median CD4 cell count in these recently diagnosed patients is of interest, especially because the widespread use of highly active antiretroviral therapy (HAART) occurred during the same time frame, and would be expected to be associated with an increase in CD4 cells.[20]
The paradoxical decrease in median CD4 count in recent patients with newly diagnosed AIDS-related lymphoma may reflect the lack of HAART use, or the lack of long-term efficacy of HAART in these individuals. Alternatively, the use of HAART may simply have allowed these patients to live long enough to eventually develop lymphoma. If the latter explanation is validated, the incidence of lymphoma would clearly be expected to increase over time, despite the use of HAART.
Incidence of AIDS-Related Lymphoma in the Era of HAART
While HAART has been associated with a significant decline in the incidence of various opportunistic infections and Kaposi’s sarcoma,[20-22] such a major and significant decline has not yet been seen in patients with systemic AIDS-related lymphoma. In a cohort of 6,636 HIV-infected individuals from Switzerland, reflecting over 18,000 person-years of follow-up, Lederberger and colleagues[21] compared the incidence of various AIDS-defining conditions in the period from 1992 to 1994 (prior to the widespread use of HAART) with the period from July 1997 to June 1998. No decrease in AIDS-related lymphoma was seen (relative hazard = 0.61, 95% confidence interval = 0.30-1.29). Likewise, although patient numbers were quite small (27 cases), no decrease in the incidence of primary CNS lymphoma was evident from these data.[21]
A recent report of AIDS cases from 17 Western European countries evaluated the incidence of systemic and primary CNS lymphoma in 1988, compared with that in 1997.[23] The number of lymphoma cases among all AIDS patients rose steadily from 1988 to 1996, and then declined somewhat in 1997. As a percentage of all AIDS cases, however, systemic lymphoma actually increased from 3.6% in 1994 to 4.9% in 1997. This would be consistent with the significant decline in opportunistic infections reported with the advent of HAART.[20] A total of 989 cases of primary CNS lymphoma were diagnosed in this cohort, representing approximately 0.5% of all AIDS cases, with no significant change in incidence over time.[23] In contrast, investigators from Italy have documented a significant decrease in primary CNS lymphoma prevalence from 1991-1997 to 1997-1998.[24]
In Australia, national HIV reporting has been ongoing since 1985, with seroconversion data available from 1991 on. Almost the entire population of HIV-infected persons in that country has been tabulated, providing a unique ability to track trends in cancer incidence over time. In this large and comprehensive database, the standardized incidence rates of lymphoma among seroconverters decreased from 67 prior to July 1996 to 33.3 thereafter, indicating a 50% reduction in AIDS-related lymphoma coincident with the widespread use of HAART.[25]
Data from the International Collaboration on HIV and Cancer were presented recently.[26] Investigators representing over 20 cohort or case-control studies conducted around the world were convened after submitting initial raw data from these various studies for re-review in the United Kingdom. Different methods were used to analyze cohort vs case-control studies, but all relative risk estimates were adjusted for age, sex, and race. In the case-control studies, additional adjusting factors included number of sexual partners and parity.
Eleven of the studies allowed assessment of cancer risk in the years before and after widespread use of HAART. In this subgroup, rate ratios of the AIDS-defining cancers were calculated for the period 1992-1996, and were compared with the rate ratios for 1997-1999, when HAART was widely available in resource-rich areas of the world.[26] For non-Hodgkin’s lymphoma (NHL), the rate fell from 6.2 to 3.6 per 1,000 patient-years, indicating a significant decline, with a rate ratio between the two periods of 0.58. No decline in Burkitt’s lymphoma was seen. A significant reduction in the incidence of primary central nervous system lymphoma was also documented.[26]
These studies are thus inconsistent in terms of the incidence of systemic or primary CNS lymphoma in the era of HAART. While most studies do not indicate a statistically significant decline in the incidence of AIDS-related lymphoma, other large studies, such as the International Collaborative Study, indicate that there has been a meaningful decline. However, one fact remains clear: the decrease in the incidence of AIDS-related lymphoma is not as impressive as the decline in the incidence of Kaposi’s sarcoma.[21] Furthermore, while initial controlled clinical trials indicated that approximately 80% of treated subjects will achieve a nondetectable HIV viral load after HAART, only about 40% will achieve this end point in "real world" conditions.[27] The effect of HAART on the incidence of AIDS-related lymphoma will clearly be dependent upon the long-term efficacy of combination antiretroviral therapy when assessed at the population level. Issues of access, compliance, drug resistance, and underlying host and environmental factors will all likely be operative. Thus, further time will be required to elucidate the full impact of HAART on the incidence of AIDS-related systemic and CNS lymphoma.
