A pooled analysis of 1,144 patients with HIV-associated lymphoma: Assessment of lymphoma-, HIV-, and treatment-specific factors on clinical outcomes

Reporter: J Taylor Whaley
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: June 2, 2012

Presenter: Stefan K. Barta, MD
Presenter's Affiliation: Albert Einstein College of Medicine, Bronx, NY


  • Patients suffering from HIV are at an increased risk of numerous complications, including several types of cancers. Three AIDS-defining malignancies include cervical cancer, Kaposi's sarcoma, and non-Hodgkin's lymphoma; however, individuals with HIV are also at increased risk of several others cancers, including lung cancer, anal cancer, Hodgkin's lymphoma, and liver cancer.
  • Although the introduction of highly active anti-retroviral therapy (HAART) has caused a substantial reduction in the incidence of non-Hodgkin lymphoma, it remains the most common malignancy in patients with HIV. The majority of lymphomas encountered in HIV-infected patients are diffuse large B-cell lymphomas or highly aggressive Burkitt-like lymphomas.
  • Current NCCN Guidelines suggest one of several treatment options:
    • Dose-adjusted EPOCH + rituximab
    • CDE + rituximab
    • CHOP + rituximab
    • CDOP + rituximab
  • Outcomes have significantly improved over the last decade, but there is no accepted consensus regarding the optimal initial therapeutic approach. For non-Hodgkin lymphoma in HIV patients. As such, several questions remain:
    • Which histologic types of lymphoma and prognostic factors are significant in the HIV-associated lymphomas?
    • Is rituximab beneficial in all HIV-associated lymphomas?
    • Are infusional chemotherapies more effective than CHOP?
  • The objective of this study was to assess the effects of clinical factors on response and survival in HIV-associated non-Hodgkin lymphomas.

Materials and Methods

  • The authors conducted a systematic review to search for prospective phase II or III clinical trials that assessed therapeutic interventions for HIV-associated NHL.
  • Inclusion criteria for trials were the following:
    • Trials accruing patients between 1990 and 2011 and published from 2001-2012. Trials must have been performed in U.S, Europe, or Canada and were excluded if they evaluated primary CNS lymphoma or radiation.
    • Patients were required to be over the age of 18 years old and HIV positive
  • The authors identified 41 trials that were eligible for evaluation. Primary data was available from principle investigators with utilization of current chemotherapy regimens in 16 trials, including 1144 patients.
  • Lymphoma-associated factors included in the analysis were age-adjusted International Prognostic Index (IPI) and histologic subtype. The HIV-associated factor included was CD4 count and concurrent use of HAART. Treatment factors included type of chemotherapy [CHOP (642 patients); infusional EPOCH (178 patients); infusional CDE (191 patients); intensive regimens, (155 patients)] and rituximab use (542 patients).
  • Endpoints included complete response (CR), progression-free survival (PFS), and overall survival (OS).
  • The authors used logistic regression and Cox proportional hazard models adjusted for age, sex, age-adjusted International Prognostic Index (IPI), type of lymphoma, baseline CD4 count, baseline HIV viral load, use of concurrent antiretroviral therapy, and histology.
  • Odds ratios (OR) > 1 for CR denote improved CR, and hazard ratios (HR) < 1 indicate reduced risk of progression or death.


  • Mean age of patients was 42 years old, range 18-76 y/o. 87% of patients were enrolled during HAART era. 46% of patients were enrolled in 1996-2000.
  • Most common histology was diffuse large B cell lymphoma (67%); 30% of patients had either Burkitt or Burkitt-like lymphoma; 60% had age-adjusted IPI of 2 or 3.
  • Median baseline CD4 count was 222 cells/µL. 61% of patients took HAART concurrent with chemotherapy.
  • Most common chemotherapy was CHOP (55%); 47% of patients received rituximab, and this use increased over time. Patients who received rituximab were more likely to be older, to receive concurrent HAART, to have lower baseline CD4 count, and were more likely to be treated with EPOCH.
  • Compared to patients with diffuse large B cell lymphoma, patients with Burkitt or Burkitt-like lymphoma had a lower CR rate, PFS, and OS.
  • Among the lymphoma- and HIV-specific covariates evaluated, only a higher age-adjusted IPI score was associated with inferior complete response rate, progression-free survival and overall survival. For each 1-point increase in age adjusted IPI, the odds of achieving a CR were decreased by 50% and the risk of disease progression was increased by 57%.
  • Rituximab was associated with a statistically significant higher complete response rate, better PFS, and OS.
  • Patients with a higher baseline CD4 count (?100 cells/µL) benefited more from rituximab with a 37% increase rate of CR and 9% lower risk of disease progression. For each 50 cells/µL increase in CD4 count, a 1% decrease in death was noted.
  • For all histologies, initial therapy with the EPOCH regimen resulted in a statistically better CR rate, PFS, and OS when compared to CHOP.
  • On multivariate analysis, the use of rituximab was associated with 75% increase chance of CR. When using CHOP therapy as a reference, treatment with intensive regimens was associated with similar CR rate but 79% increase rate of PFS and 67% increase in rate of OS.
  • Initial therapy with EPOCH showed 79% increase in CR rate. Infusional CDE was associated with a significant lower CR rate but improved PFS and OS, which were statistically significant. The lower CR rate in this case may be related to different definitions for CR on CDE trials.

Author's Conclusions

  • In this pooled analysis including 1144 patients with HIV-associated NHL, the main prognostic indicator for HIV-associated lymphoma remains age-adjusted IPI. This is similar to non-HIV patients.
  • For each 50 cells/µL increase in baseline CD4 count, a 1% decrease in death was noted.
  • The addition of rituximab to chemotherapy was associated with significantly improved CR rate, PFS, and OS, specifically for patients with a baseline CD4 count ?100 cells/µL.
  • Treatment with infusional EPOCH was also more effective than CHOP with 40-50% decrease in progression or death. This finding supports the current on-going CALGB trial comparing R-CHOP with R-EPOCH.

Clinical Implications

  • This meta-analysis of patients with HIV-associated lymphomas treated on phase II and III clinical trials provides important information and adds to a growing body of literature on the proper treatment of HIV-associated malignancies.
  • This study again indicates the critical nature of HAART in the treatment of HIV-associated malignancies with statistically significant improved in overall survival with a rise in baseline CD4 counts.
  • This particular study is a meta-analysis and utilized pooled data for analysis. Although the evaluation included trial level data, the results must be assessed carefully.
  • Additionally, as this consisted of pooled data, it covered a very large time period and crossed several treatment periods. The model could not account for time of enrollment and the use of rituximab was a confounding factor since its introduction in 2000.
  • Having said this, patients with HIV-associated non-Hodgkin's lymphomas are at higher risk for lower complete responses and increased disease progression. The optimal chemotherapy regimen remains uncertain, but patients with a CD4 count > 100 cells/µL certainly appear to benefit from the use of rituximab.
  • We look forward to the results of the current on-going CALGV trial evaluating the use of R-CHOP vs. R-EPOCH in immunocompetent patients with non-Hodgkin's lymphoma.

Frequently Asked Questions