Survival and predictors of death in people with HIV-associated lymphoma compared to those with a diagnosis of lymphoma in general population

OBJECTIVES: to compare overall survival in HIV-associated lymphoma (HIV-L) and lymphoma raising in HIV-negative population (nHIV-L) and to identify predictors of increased risk of death. METHODS: All HIV+ patients with HIV-associated lymphoma (Hodgkin lymphoma, HL; non-Hodgkin Lymphoma, NHL) observed between 1.2000 and 12.2013 in the ICONA Foundation Study cohort or in three collaborating centres, and, as control group, nHIV-L individuals followed in one of the four collaborating centres over the same time period, were included. Survival estimates were calculated by use of Kaplan-Meier (KM) and multivariable Cox regression models. RESULTS: 1,331 pts were included (465 HIV-L, 866 nHIV-L): 909 (68%) NHL, 422 (32%) HL. 3 years-cumulative probability (95% confidence interval, CI) of death was higher in HIV-L compared to nHIV-L in NHL (38% (33-44) vs. 22% (19-26); p<0.001), and HL (22% [15-29] vs. 10% (6-13), p<0.001). Among HL, HIV was associated with an increased risk of death (hazard ratio [HR] = 2.37 [95% CI: 1.24-4.55], p = 0.009) independently of calendar year, age, gender, type of chemotherapy and stage; in NHL, HIV was no longer an independent predictor of death after controlling for rituximab use and IPI (HR = 1.26 (0.97-1.63), p = 0.08). CONCLUSIONS: Our analysis shows a reduced overall survival in HIV+ patients diagnosed with lymphoma compared to HIV-negative controls. Whereas in HIV people with HL, the increased risk of death was confirmed even after adjustment for main confounders, the association between HIV status and survival in NHL appears to be somewhat attenuated after controlling for more aggressive presentation and lower frequency of rituximab use in HIV-+ people.