A Pooled Analysis of 1,546 Patients with HIV-Associated Lymphoma: Assessment of Lymphoma-, HIV-, and Treatment-Specific Factors On Clinical Outcomes

Non-Hodgkin Lymphoma (NHL) remains the most common malignancy in patients with HIV infection, but there are few randomized phase II or III trials that provide level 1 evidence regarding the optimal treatment. We previously reported a pooled analysis of 16 trials including 1,144 patients with HIV–associated NHL (ASCO 2012, abstract 8005), and herein report the updated final analysis of 1,546 patients enrolled on 19 trials. Our objective was to assess lymphoma-, HIV-, and treatment-specific factors and their influence on response and survival.

We performed a systematic review to search for prospective clinical phase II or III trials that assessed therapeutic interventions for HIV-associated NHL and assembled individual patient data from 19 trials published between 1993 and 2011 including 1,546 patients (median N=61/trial, range 14–467). Treatment factors included rituximab use (N=542) and type of chemotherapy (CHOP, N=632; EPOCH, N=166; CDE, N=191; dose-intense (intensive) regimens, N=155; ABCVP-based regimens, N=158; low-dose CHOP, N=165; VS, N=41; oral regimens, N=38). Endpoints included complete response (CR), progression-free survival (PFS), and overall survival (OS). We used logistic regression and Cox proportional hazard models adjusted for age, sex, histological subtype, age-adjusted international prognostic index (aaIPI), CD4 count at baseline, prior history of AIDS, type of chemotherapy, use of rituximab, concurrent use of GCSF and combination antiretroviral therapy (cART) with chemotherapy, and enrollment period.

The majority of patients were enrolled in the cART era (75%), were male (84%), had DLBLC (69%; BL/BLL 26%; other 6%), had high risk disease (aaIPI≥2; 59%) and a baseline CD4 count ≥50 cells/μl (86%). CHOP was the most commonly prescribed regimen (41%). Among the lymphoma- and HIV-specific covariates evaluated, lower baseline CD4 count and higher aaIPI score were associated with inferior CR rates (CD4: p<0.05; IPI: p<0.001), PFS (CD4: p=0.06; IPI: p<0.001) and OS (p<0.001 for both). Rituximab was associated with a higher CR rate (OR 2.89, p<0.001), better PFS (HR 0.50, p<0.001) and OS (HR 0.51, p<0.0001). For all histologies, initial therapy with more intensive therapy (ACVBP-based or intensive regimens) resulted in better CR rates (ACVBP: OR 1.70, p<0.05), PFS (ACVBP: HR 0.72, p=0.049; intensive: HR 0.35; p<0.001) and OS (intensive: HR 0.54, p<0.001) when compared to CHOP, whereas less intensive therapy resulted in significantly reduced CR rates, PFS and OS (p<0.0001 for all outcomes with low-dose CHOP or VS). In a subgroup analysis of only patients diagnosed with DLBCL treated with either CHOP or EPOCH, we found that the use of EPOCH was associated with significantly better OS (HR 0.33; p=0.031). In the subset of patients with either BL or BLL the outcomes for treatment with intensive regimens versus the infusional regimens EPOCH and CDE were not significantly different except for a better PFS with the intensive regimens (HR for PFS: 0.40, p=0.04). Concurrent use of cART was associated with improved CR rates (OR 1.89; p=0.005), and OS (HR 0.78; p=0.07). When examining causes of death, we found that the type of chemotherapeutic regimen used was associated with an increased risk of dying of lymphoma progression (OR 1.13; p=0.02). Treatment with rituximab-containing regimens was associated with a reduced risk of death secondary to lymphoma (OR 0.31; p<0.0001) and treatment-related mortality (OR 0.59; p=0.02), but not death secondary to HIV-related causes. Lower CD4 counts at the time of lymphoma diagnosis were significantly associated with higher odds for death secondary to HIV-related causes. This association was strongest for patients with baseline CD4 counts <50 cells/μl (p=0.04) and was no longer significant for patients with a CD4 count ≥200 cells/μl (p=0.18). In the adjusted model, there was no association between CD4 count and death due to treatment toxicities or lymphoma.

In this pooled analysis including 1,546 patients with HIV-associated NHL, the addition of rituximab to chemotherapy was associated with significantly improved CR rate, PFS, and OS. Rituximab use did not result in a higher risk of infectious or HIV-related death. Treatment with infusional EPOCH was also more effective than CHOP for patients with DLBCL. Concurrent use of cART was associated with better outcomes.

No relevant conflicts of interest to declare.