Long‐term follow‐up of patients with HIV‐related diffuse large B‐cell lymphomas treated in a phase II study with rituximab and CHOP

In the highly active antiretroviral treatment (HAART) era, the addition of rituximab (R) to standard CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) chemotherapy (Boue et al, 2006; Ribera et al, 2008), CHOP-like regimens (Sparano et al, 2010) or infusional therapies (Spina et al, 2005) has proven to be feasible and effective in Phase II trials and in cohort studies (Wyen et al, 2011) in patients with human immunodeficiency virus (HIV)-related B-cell lymphomas. On the other hand, the incidence of non-acquired immunodeficiency syndrome (AIDS)-related malignancies has increased among patients with HIV infection under HAART (Patel et al, 2008; Crum-Cianflone et al, 2009). However, there is scarce information regarding second malignancies or other AIDS-related events occurring in HIV-infected patients who have responded to treatment for lymphoma. We analysed the long-term follow-up of patients with HIVrelated diffuse large B-cell lymphoma (DLBCL) in complete response (CR) included in a Phase II clinical trial of R-CHOP conducted by the Spanish PETHEMA (Programa para el Estudio y Tratamiento de las Hemopatias Maligna), GESIDA (Grupo de Estudio de SIDA), GELTAMO (Grupo Español de Linfoma y Trasplante de Médula Ósea) and GELCAB (Grupo para el Estudio de los Linfomas Catalano-Balear) groups (Ribera et al, 2008). Eighty-one patients with HIV-related DLBCL were included in the Phase II trial of rituximab and CHOP (R-CHOP) therapy. With a median 2-year follow-up (range: 0·1–5·4) of patients alive at the time of publication, the 5-year overall survival (OS) [95% confidence interval (CI)] probability was 56% (43–69%). In the present study the following events were recorded in the 55 patients in CR: relapse of the lymphoma, opportunistic infections (OI) and other cancers. HIV viral load and CD4 lymphocyte count at the nearest time-point to the event were also recorded. Comparison of these counts between patients with and without events was performed by the Mann–Whitney U-test for CD4 lymphocyte count and the chi-square test for viral load (categorized as negative or positive [>50 copies/ml]. Cumulative probabilities of OI and second cancers, as well as OS and event-free survival (EFS) probabilities were calculated by the Kaplan and Meier method. In patients with more than one event the date of the first event was considered for analyses. This follow-up analysis was performed in October 2011. At the time of the analysis the median follow-up of living patients was 6·2 years (range: 4·6–9·5). One patient in CR was lost to follow-up, eight had presented with lymphoma relapse and five had OI (meningoencephalitis [2], Pneumocystis jiroveci pneumonia [1], varicella pneumonia [1], pneumoccal pneumonia [1], oesophageal candidiasis [1], and cytomegalovirus cholitis [1]. Two patients each had two OI. Five patients developed a second cancer (invasive carcinoma of the cervix [1], squamous lung cancer [1], lung adenocarcinoma [1], pancreatic adenocarcinoma [1], and metastatic sarcoma of unknown origin [1]). One patient presented an OI (varicella pneumonia) followed by pancreatic adenocarcinoma 5 years later. The median time of OI appearance was 0·18 years (range 0·02–6·98) and the cumulative probability of OI at 8 years was 15% (95%CI: 7–23%). For second cancers, these values were 2·77 years (range 2·09–4·71) and 12% (95%CI: 2–22%), respectively. Fifteen patients died: six due to lymphoma relapse, three to OI, four to second cancer and two for other reasons (sudden death in one and violent death in the other). The 8-year OS probability for the cohort of 55 patients in first CR of the lymphoma was 64% (95%CI: 45–83%) (Fig 1A) and the EFS probability was 59% (95%CI: 42–76%) (Fig 1B). The 8-year OS probability for the whole series of 81 patients was 46% (95%CI: 31–61%), representing a 10% reduction in survival with respect to the original publication. No differences in the HIV viral load were observed between patients with and without events (Table I). However, patients with events had a significantly lower CD4 lymphocyte count compared to those without events. As expected, this difference was especially evident for OI. This long-term follow-up study of HIV-infected patients with DLBCL treated with R-CHOP shows a remarkable frequency of OI and second cancers, having a significant impact on the survival probability for this group of patients. Lower CD4 lymphocyte counts were observed in the group of patients with these events than in those patients without these complications. The use of HAART has improved the prognosis of HIVrelated lymphomas, leading to an increase in survival, similar to that of the non-immunosuppressed patients in some matched cohort studies (Navarro et al, 2005), and the International Prognostic Index has proven to be useful for prognosis assessment (Bower et al, 2005). The incidence of non-AIDS-related malignancies has increased in the HAART era (Patel et al, 2008; Crum-Cianflone et al, 2009). However, there is scarce information regarding second cancers occurring in patients that have been successfully treated for lymcorrespondence