Safety and efficacy of cyclophosphamide, adriamycin, vincristine, prednisone and rituximab in patients with human immunodeficiency virus-associated diffuse large B-cell lymphoma: results of a phase II trial


Josep-María Ribera, Servicio de Hematología Clínica, Institut Català d’Oncologia-Hospital Universitari Germans Trias i Pujol, C/Canyet S/N, 08916 Badalona, Spain. E-mail:


Immunochemotherapy with cyclophosphamide, adriamycin, vincristine, prednisone and rituximab (R-CHOP) is the standard treatment in non-immunosuppressed patients with diffuse large B-cell lymphoma (DLBCL), but its adequacy has not been definitively established in patients with human immunodeficiency virus (HIV)-related lymphoma. This phase II trial aimed to evaluate the safety and efficacy of six cycles of R-CHOP in patients with HIV-related DLBCL and to determine whether response to highly active antiretroviral therapy (HAART) had prognostic impact. Patients were eligible if they had performance status <3 and absence of active opportunistic infections. Eighty-one patients were enrolled, 57 in stages III or IV, International Prognostic Index (IPI) 0 or 1 (n = 26), 2 (n = 19), 3 (n = 20) and 4 or 5 (n = 16), and median CD4 lymphocyte count of 0·158 × 109/l. The main adverse events were neutropenia (48% of cycles) and infections (10% of cycles), which were fatal in seven patients. Complete response was achieved in 55 (69%) patients, with an estimated 3-year disease-free survival of 77% and 3-year overall survival of 56%. IPI score and virological response to HAART were the prognostic parameters for response and survival. In HIV-related DLBCL R-CHOP is feasible, safe and effective. The prognosis depends on lymphoma-related parameters and on the response to HAART.

The generalized use of highly active antiretroviral therapy (HAART) has dramatically improved the survival of human immunodeficiency virus (HIV)-infected patients (Lewden et al, 2005). This has been because of the decrease in the frequency of opportunistic infections and some types of cancer, including acquired immunodeficiency syndrome (AIDS)-related lymphomas (ARL) (Engels et al, 2006). In addition to the reduction in frequency, there is clear evidence that the prognosis of ARL has improved, being similar to that of the lymphomas in non-HIV-infected patients in some studies (Spina et al, 2004; Navarro et al, 2005; Diamond et al, 2006). Nevertheless, ARL still constitute a major cause of death in HIV-infected patients. Most ARL are diffuse large B-cell lymphomas (DLBCL) or Burkitt or Burkitt-like lymphomas, and express CD20 on the surface of lymphoma cells (Navarro & Kaplan, 2006).

In recent years several trials with standard chemotherapy (CHT), such as CHOP (cyclophosphamide, doxorubicin, vincristine and prednisone) (Boue et al, 2006), dose-adjusted EPOCH (96-h infusional regimen of doxorubicin, etoposide and vincristine in conjunction with intravenous bolus of cyclophosphamide plus oral prednisone) (Little et al, 2003) or CDE (infusional cyclophosphamide, doxorubicin and etoposide) (Spina et al, 2005) combined with the anti-CD20 monoclonal antibody rituximab have been published, with encouraging results in terms of efficacy and acceptable toxicity. However, in a randomized trial comparing CHOP with CHOP and rituximab (R-CHOP) (Kaplan et al, 2005), the efficacy of the latter regimen was not statistically superior and there was a significant increase in the frequency of treatment-related infectious deaths, especially in the patients with baseline CD4 counts below 0·05 × 109/μl. Thus, the usefulness of the combination with rituximab and CHT in ARL remains to be definitively established, and additional studies assessing the efficacy and toxicity are needed.

On the other hand, the impact of the response of HIV infection to HAART in the prognosis of ARL treated with standard chemotherapies has been well established (Navarro et al, 2002; Oriol et al, 2005), but there are no data on its influence in patients treated with immunochemotherapy. This multicentre trial from the Spanish PETHEMA (Programa de Estudio y Tratamiento de las Hemopatías Malignas), GELTAMO (Grupo Español de Linfomas-Trasplante de Medula Ósea), GELCAB (Grup d’Estudi dels Limfomes de Catalunya i Balears) and GESIDA (Grupo de Estudio del SIDA de la Sociedad Española de Enfermedades Infecciosas y Microbiología Clínica) groups aimed to evaluate the safety and efficacy of rituximab in combination with CHOP in patients with HIV-related DLBCL and to determine the prognostic impact of response to HAART in these patients.

Patients and methods

Study eligibility

Patients were eligible for enrollment at diagnosis if they fulfilled the following criteria: HIV seropositivity, biopsy confirmed DLBCL according to the World Health Organization (WHO) Classification (Harris et al, 1999) with Ann Arbor stage I–IV, CD20 positivity by immunophenotyping on lymphoma cells, age from 18 to 75 years, Eastern Cooperative Oncology Group (ECOG) score ≤3 and absence of active opportunistic infection at the time of lymphoma diagnosis. Informed consent was given by all the patients included in the study. The study was approved by the Institutional Review Boards of all participating centres and the procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation and with the Helsinki Declaration of 1975, as revised in 2000. The trial was activated in April 2001 and closed for patient follow-up in December 2006.


Staging procedures included a detailed clinical history (including the date of the first known HIV infection and the use of antiretroviral therapy), physical examination (baseline height, weight, B symptoms, performance status and dimensions of all palpable lesions), complete blood count, blood chemistry [lactate dehydrogenase (LDH) and beta-2 microglobulin levels], electrocardiogram, chest radiography, computed tomography of the chest and abdomen, bone marrow (BM) biopsy, study of cerebrospinal fluid at baseline, CD4 cell count and HIV viral load. Stage was assessed by the Ann Arbor staging system (Carbone et al, 1971). The International Prognostic Index (IPI) (International Non-Hodgkin’s Lymphoma Prognostic Factors Project, 1993) was also calculated for each patient.


All patients received cyclophosphamide 750 mg/m2 i.v. on day 1, doxorubicin 50 mg/m2 i.v. on day 1, vincristine 1·4 mg/m2 (maximum 2 mg) i.v. on day 1 and prednisone 40 mg/m2/d p.o. for 5 d. Rituximab (375 mg/m2 i.v.) was administered on day 1 prior to the CHOP CHT. Patients were treated every 3 weeks for six cycles. Involved field radiotherapy was given after CHT if there was bulky disease at diagnosis or a residual mass after the completion of CHT. CNS prophylaxis with IT methotrexate 12 mg, cytarabine 40 mg and hydrocortisone 20 mg was given with each cycle of CHT. Administration of granulocyte colony-stimulating factor (G-CSF, 5 μg/kg/d s.c.) was recommended if grade 4 neutropenia was observed in the previous CHT cycle. HAART was required from the beginning of treatment, if patients were not already receiving it, and consisted of one or two protease inhibitors and two nucleoside reverse transcriptase inhibitors. Zidovudine therapy was not recommended. Patients received trimetroprim–sulfametoxazol (160/800 mg) thrice weekly or aerosolized pentamidine (300 mg inhalated) as prophylaxis against Pneumocystis jiroveci pneumonia.

Response assessment and adverse events

Response was assesssed after six cycles of R-CHOP CHT or at the end of prematurely interrupted treatment. The International Workshop criteria were employed to assess response and survival (Cheson et al, 1999). Complete response (CR) was defined as the lack of evidence of all lesions or radiological or biological abnormalities observed at diagnosis at the end of treatment and lasting for at least 1 month after the completion of CHT. Unconfirmed CR (UCR) was defined as CR with the persistence of some radiological abnormalities, which had to have reduced by at least 75% in size. Any other situation (partial response, stable disease or progression) was considered as therapeutic failure. Relapse was defined as the presence of lymphoma in a patient that had been in CR for at least 2 months. Overall survival (OS) was considered as the period between the date of diagnosis and the date of death by any cause or the last follow-up of the patient. Disease-free survival (DFS) was defined as the period between the date of CR and the date of relapse, death by any cause or the last control in CR.

Virological response to HAART was considered when the total HIV-1 RNA loads were below the limit of detection in serum (according to the reference level of each centre) after being on HAART treatment for at least 6 months. Immunological response to HAART was considered if CD4 lymphocyte counts were >0·2 × 109/l or there was an increase in CD4 lymphocyte counts over 0·1 × 109/l after being on HAART treatment for at least 6 months (Antinori et al, 2001).

Patients were regularly restaged every 3 months in the first year, every 6 months during the second year and once a year thereafter. All adverse events were collected from the case report forms and were graded according to the National Cancer Institute Common Toxicity Criteria.

Statistical analyses

The primary objective of this study was to estimate the response rate (CR and UCR). Secondary objectives were to determine DFS and OS probabilities and the estimation of toxicities higher than grade 2. A sample size of 78 measurable patients was calculated to confirm the non-inferiority of treatment with R-CHOP in relation to our previous results with CHOP with an expected response rate of 70%, 55% CR rate with CHOP (Navarro et al, 2005), a one-sided significance level of 5% and statistical power of 80%. Baseline characteristics were described and their univariate association with CR was analysed using bivariate tests (Student t-test or Mann–Whitney U-test when appropriate) for quantitative variables and the chi-square or Fisher’s exact test for categories. Variables with a significance <0·05 in univariate tests were included in a stepwise logistic regression model to identify statistically independent predictive factors for response achievement. Actuarial curves for event-free survival and OS were plotted according to the Kaplan–Meier method (Kaplan & Meier, 1958) and were compared by the log-rank test (Peto & Pike, 1973). The statistically significant (P < 0·05) or borderline significant (0·1 < P < 0·05) variables identified in univariate studies were incorporated stepwise to a Cox proportional hazard-regression model (Cox, 1972). The virological and the immunological response to HAART were included in these models as time-dependent variables. Odd ratios (OR) were reported with a 95% confidence interval (CI). Statistical analyses were carried out using the spss (Statistical Package for Social Sciences) software, version 12 for Windows.


Pretreatment characteristics

Eighty-six patients were enrolled in 24 centres from the Spanish PETHEMA, GEL-TAMO, GELCAB and GESIDA groups between April 2001 and April 2006. Five patients were excluded after review (Burkitt lymphoma in three cases, Hodgkin lymphoma in one and unclassified B-cell lymphoma in the remaining patient). The main baseline characteristics of the 81 evaluable patients are listed in Table I. The median age was 44 years (range 21–74 years) with male predominance (82%). Sexual activity was the main risk factor for HIV infection (60% of the cases). Non-Hodgkin lymphom (NHL) was diagnosed simultaneously with HIV infection in 40 (49%) patients. The median time interval between the diagnosis of HIV infection and the onset of NHL was 7 years (range 0·2–21 years) for the remaining 41 cases. In patients receiving HAART for more than 3 months prior to NHL diagnosis, the median time on HAART was 1 year (range 0·3–12 years). The median value of the CD4 lymphocyte count was 0·158 × 109/l (range 0·006–0·905 × 109/l) and the median HIV viral load was 7058 copies/ml (range 0–2 × 106 copies/ml). Bulky disease was present in 13 (16%) patients and BM involvement was observed in 10 (12%) patients. In 48 of 80 (60%) patients serum LDH levels were over the normal range and 37 of 73 (51%) patients had serum beta-2 microglobulin levels over the upper normal value.

Table I.   Main clinical and biological characteristics of the 81 patients studied in the series.
  1. *One IV drug user was also homosexual.

  2. †Defined as HAART started <3 months before NHL diagnosis. Data known in 73 patients.

  3. ‡Available in 80 patients.

  4. HIV, human immunodeficiency virus; NHL, non-Hodgkin lymphoma; HAART, highly active antiretroviral therapy; ECOG, Eastern Cooperative Group; IPI, International Prognostic Index.

Risk activity for HIV infection
 IV drug use*2936
Prior opportunistic infections3847
Prior neoplasia34
NHL as first HIV-associated event4049
Prior HAART†
CD4 lymphocyte count‡ (109/l)
HIV viral load <50 copies/ml‡2228
ECOG status‡
B symptoms4353
Extranodal involvement (organ)5467
Ann Arbor stage
IPI score
 0–1 (low)2632
 2 (intermediate-low)1923
 3 (intermediate-high)2025
 4–5 (high)1620

Results of treatment and clinical outcome

Figure 1 shows the main results of the study. One patient withdrew early from the study. Sixty-two (78%) of the 80 evaluable patients completed the six scheduled cycles of R-CHOP CHT. Involved field radiotherapy was presented to two patients for residual masses at diagnosis. A dose reduction or treatment delay was required in 38% of the patients because of haematological toxicity. G-CSF support was administered to 75 (94%) patients, and all patients received HAART during R-CHOP therapy.

Figure 1.

 Main results of the 82 patients in the study.

At the end of treatment 55 (69%) patients attained CR, 16 experienced resistance after re-evaluation (n = 7) or disease progression during therapy (n = 9) and nine patients died during therapy. The main causes of death were infection (seven patients), liver failure (n = 1), and multi-organ failure (n = 1). Five of 55 patients relapsed. Five patients died in CR (P. jiroveci pneumonia in one case, pneumococcal pneumonia in one, AIDS-dementia complex in one, sudden death in one and violent death in the remaining patient). Forty-five patients remain alive and disease free. After a median follow-up of 24 months, the estimated 3-year DFS was 77% (95% CI, 62–92%) (Fig 2) and the estimated 3-year OS was 56% (95% CI, 43–69%) (Fig 3). Virological response was observed in 50 of the 80 patients (62%) and 19 of 38 evaluable patients (50%) showed an immunological response to HAART.

Figure 2.

 Disease-free survival for the 55 patients in complete remission.

Figure 3.

 Overall survival for the 81 patients from the series.


Table II shows the main toxic events observed in 426 evaluated R-CHOP cycles. The most frequent adverse event was grade 3 or 4 neutropenia (208 cycles, 49%), accompanied by infection in 42 cycles (10%). Six patients had infections (Escherichia coli in one, Pseudomonas spp. one, methicillin-resistant Staphylococcus aureus one and unknown origin in three) and died. The next most frequent toxic event was digestive (24 episodes, 6%), being fatal in one case (liver failure). There were few AIDS-related events: P. jiroveci pneumonia (n = 1), tuberculosis (n = 1), cytomegalovirus infection (n = 2), aspergillosis (n = 1), varicella-zoster pneumonia (n = 1) and esophageal candidiasis (n = 1). The actuarial probability of having an opportunistic infection was 10% (95% CI 3–17%) at 3 years (Fig 4). Patients with opportunistic infections had a significantly lower CD lymphocyte count at lymphoma diagnosis than those who did not present this complication [0·137 ± SD: 0·099 × 109/l vs. 0·241 ± 0·22 × 109/l; P = 0·02].

Table II.   Main grade 2–4 toxic events in 426 R-CHOP cycles evaluated.
CycleNo. evaluable cyclesNeutropeniaInfectionDigestiveNeurologicalSkinHeart
Figure 4.

 Actuarial probability of HIV-related infections OI, opportunistic infections.

Prognostic factors

Bone marrow involvement and IPI score >1 were associated with failure to achieve CR on univariate analysis (Table III). Only an IPI score >1 retained statistically independent significance as a prognostic factor for response on logistic regression analysis. No baseline characteristics were found to be associated with a shorter DFS whereas an IPI score >1 had a negative impact on OS (Table IV and Fig 5). Virological response to HAART was included as a time-dependent variable in multiple regression analyses and proved to have a positive impact on both DFS and OS. Low or low-intermediate IPI score retained an independent significance on OS (Table V).

Table III.   Univariate analysis of prognostic factors for attaining complete remission (CR) in the 80 evaluable patients*.
VariableCategorynCR (%)No CR (%)P-value
  1. *One patient was withdrawn early.

  2. †Includes the following variables: age, ECOG score, Ann Arbor stage, extranodal involvement and LDH serum level.

  3. NHL, non-Hodgkin lymphoma; AIDS, acquired immunodeficiency syndrome; HIV, human immunodeficiency virus; VL, viral load; ECOG, Eastern Cooperative Oncology Group; BM, bone marrow; LDH, lactate dehydrogenase; IPI, International Prognostic Index; G-CSF, granulocyte colony stimulating factor; HAART, highly active antiretroviral therapy.

SexMale6542 (65)23 (35)0·084
Female1513 (87)2 (13)
Risk activityDrug abuse2818 (64)10 (36)0·506
Homosexual/bisexual2415 (63)9 (37)
Heterosexual2519 (76)6 (24)
Unknown33 (100)0 (0)
NHL as a first AIDS-related eventYes5528 (51)27 (49)0·567
No2511 (44)14 (56)
Undetectable HIV VL (at diagnosis)Yes2617 (65)9 (35)0·779
No5437 (69)17 (31)
B symptomsYes4227 (64)15 (36)0·365
No3828 (74)10 (26)
BM involvementYes104 (40)6 (60)0·045
No7051 (73)19 (27)
Beta-2-microglobulin (mg/l)>3·53725 (68)12 (32)0·863
≤3·53625 (69)11 (31)
IPI score†0–12624 (92)2 (8)0·002
>15431 (57)23 (43)
HAARTBefore NHL5437 (69)17 (31)0·949
From treatment2618 (69)8 (31)
CD4 lymphocyte count (109/l)≥0·24630 (65)16 (35)0·612
<0·23424 (71)10 (29)
Virological responseYes5038 (76)12 (24)0·071
No3017 (57)13 (43)
Table IV.   Univariate analysis of prognostic factors for disease free survival (DFS) and overall survival (OS).
VariableCategorynEvents for DFSDFSnEvents for OSOS
CCR (%) 3 years (95% CI)P-valueAlive (%) at 3 years (95% CI)P-value
  1. CCR, continuous complete remission; NHL, non-Hodgkin lymphoma; AIDS, acquired immunodeficiency syndrome; HIV, human immunodeficiency virus; VL, viral load; ECOG, Eastern Cooperative Oncology Group; BM, bone marrow; LDH, lactate dehydrogenase; IPI, International Prognostic Index; G-CSF, granulocyte colony stimulating factor; HAART, highly active antiretroviral therapy; CHT, chemotherapy.

SexMale42872 (54–100)0·378662851 (37–65)0·144
Female13192 (77–100)15380 (60–100)
Risk activityDrug abuse18466 (34–98)0·350291153 (29–77)0·348
Homosexual15468 (42–94)241342 (40–64)
Heterosexual19194 (83–100)25770 (51–89)
NHL as a firstNo27667 (42–92)0·210412043 (24–62)0·059
AIDS-related eventYes28385 (70–100)401169 (54–84)
Undetectable VL at diagnosisYes17647 (16–78)<0·001261247 (25–69)0·197
No37289 (74–100)541861 (46–76)
B symptomsYes27387 (72–100)0·343431660 (44–76)0·909
No28669 (47–91) 381553 (34–72)
BM involvementYes4167 (14–100)0·80710640 (10–70)0·092
No51877 (61–93)712558 (44–72)
Beta-2-microglobulin (mg/l)≤3·524479 (61–97)0·923351263 (50–76)0·390
>3·525573 (52–94)381749 (31–67)
IPI score0–124376 (50–100)0·53726570 (45–95)0·015
>131677 (60–94)552650 (36–64)
HAARTBefore NHL35674 (55–93)0·577511857 (42–74)0·829
From CHT16285 (66–100)23958 (37–79)
CD4 lymphocyte count (109/l)>0·224573 (53–93)0·371351358 (40–76)0·924
0·1–0·19913367 (33–100)20948 (23–73)
0·05–0·0996150 (0–100)9433 (0–82)
<0·05110100 (-)16568 (45–91)
ImmunologicalYes19288 (72–100)0·85019287 (70–100)0·711
responseNo19275 (41–100)19180 (45–100
Figure 5.

 Overall survival according to International Prognostic Index Score.

Table V.   Multivariate analyses of prognostic factors for tumour response, disease-free survival (DFS) and overall survival (OS).
CategoryVariableOR (95% CI)P-value
  1. IPI, International Prognostic Index.

ResponseLow IPI score8·93 (1·91–41·67)0·005
DFSVirological response11·403 (1·16–112·32)0·037
OSVirological response3·07 (1·16–8·10)0·023
Low IPI score2·72 (1·04–7·10)0·041


The present study showed that, in patients with HIV-related DLBCL, immunochemotherapy with R-CHOP with concomitant administration of HAART is feasible, safe and effective. The main prognostic factors were related to the lymphoma itself. In addition, this is the first study to confirm the favourable impact of virological response to HAART in the survival of patients treated with rituximab-based CHT.

The design of this study took into account the promising results observed with standard CHOP CHT in patients with ARL in the HAART era (Coiffier et al, 2002; Pfreundschuh et al, 2006). As immunochemotherapy with rituximab in combination with CHOP-based or infusional regimens has shown good results in non-immunosuppressed patients with DLBCL, we developed the present phase II trial of R-CHOP in this particular subtype of lymphoma in patients with HIV infection, in whom the safety and efficacy were unknown. In fact, to date, only two phase II studies with immunochemotherapy (Spina et al, 2005; Boue et al, 2006) and one randomized trial (Kaplan et al, 2005) comparing CHOP and R-CHOP schedules have been published.

The baseline characteristics of the patients enrolled in the present trial were similar to those from other trials in the HAART era (Little et al, 2003; Sparano et al, 2004; Kaplan et al, 2005; Spina et al, 2005; Boue et al, 2006). It is of note that an extremely low-CD4 lymphocyte count was not an exclusion criterion for the present study, although patients with active opportunistic infection at the time of DLBCL diagnosis or very poor performance status were not included.

The results of the present study are similar to those from the two phase II trials with CHT and rituximab. Boue et al (2006) treated 61 patients with several types of ARL (most with DLBCL) with R-CHOP, 52 of whom were assessable for response. The CR was 77% and the 2-year OS rate was 75%, similar to rates observed in the present study. Spina et al (2005) reported the results of the infusional CDE regimen together with rituximab in three pooled trials in patients with ARL, with similar results (CR 70%, 2-year OS: 64%). The results observed in these three trials are similar to those observed in non-immunosuppressed patients, a feature that has already been observed in ARL receiving HAART and CHT before the use of rituximab (Spina et al, 2004; Navarro et al, 2005; Diamond et al, 2006).

Infections were the most important adverse event in the present study. The randomized trial comparing R-CHOP and CHOP schedules in patients with ARL conducted by the AIDS-Malignancies Consortium (AMC) Group (Kaplan et al, 2005) identified a statistically significant increase in infectious deaths occurring in the rituximab arm. These treatment-related infectious deaths were more frequent (60%) in patients with baseline CD4 counts below 0·05 × 109/l, representing 25% of the cases from that arm, compared with 20% in our study. In the French study (Boue et al, 2006) with R-CHOP, severe neutropenia occurred in 25% of patients, but only three patients presented severe sepsis, being fatal in one. However, in that trial only five (8%) patients had a CD4 lymphocyte count below 0·05 × 109/l (Spina et al, 2007) versus 16 (20%) in our series. In the CDE and rituximab study (Spina et al, 2005) (in which the number of patients with a CD4 lymphocyte count below 0·05 × 109/l is not provided), grade 3–4 neutropenia occurred in 78% of patients and the frequency of opportunistic and non-opportunistic infections during or within 3 months of the end of R-CDE the treatment was higher than that observed in historical controls with CDE (31% vs. 20%). However, only 2% of patients in the R-CDE arm (vs. 0% in the historical CDE arm) had treatment-related infectious deaths. In the present study, severe neutropenia occurred in 49% of R-CHOP cycles and infections were observed in 42 (10%) cycles, leading to death in seven patients. This supports the concept that in patients under R-CHOP therapy the frequency and severity of infections is considerable, especially in patients with extremely low-CD4 lymphocyte counts at the time of lymphoma diagnosis. The frequency of opportunistic infections in the present study was 13%, similar to 8% in the R-CHOP arm in the AMC Group study (Kaplan et al, 2005) or to 14% in the R-CDE trial (Spina et al, 2005), but superior to the 3% observed in the phase II study or R-CHOP by Boue et al (2006). It is of note that in the latter trial the number of patients with severely immunosuppressed at NHL diagnosis was lower than in the three aforementioned trials (Spina et al, 2007).

As occurred in the two phase II studies (Spina et al, 2005; Boue et al, 2006), from the data of the present study it is difficult to determine the contribution of rituximab to the overall efficacy. The only comparative trial published to date, the AMC-10 trial (Kaplan et al, 2005), showed a non-statistically significant trend for a better response rate (57·6% vs. 47%) in the R-CHOP arm, that did not translate into better survival, at least in part because of the excess number of infectious deaths in that arm (14% vs. 2%). However, even in moderately immunosuppressed patients (i.e. those with CD4 lymphocyte count >0·1 × 109/l) no significant benefits in terms of survival were observed in the R-CHOP arm. The inclusion of a high number of severely immunosuppressed patients in that trial could explain the inferior results with respect to those from the three aforementioned phase II trials, as the CD4 lymphocyte count has been a poor prognostic factor in patients with ARL treated with CHOP-like chemotherapies in the HAART era in some studies (Bower et al, 2005).

As in most studies in the HAART era (Lim et al, 2005; Miralles et al, 2007), the prognostic factors identified in the present study are lymphoma related. As a significant proportion of patients began HAART treatment at lymphoma diagnosis, baseline CD4 lymphocyte counts and viral load did not appear as adverse prognostic factors. The prognostic value of immunological response is difficult to ascertain given that a period of time, at least 6 months, is necessary after CHT for the recovery of the decreased CD4 lymphocyte count (Little et al, 2003). Both the relative lack of events after that period and the fact that the immunological response could only be effectively assessed in 38 patients limit the statistical analysis of that variable in our study. On the other hand, the virological response is, theoretically, not directly influenced by CHT and could be sequentially assessed during and after treatment, presenting a clear association with a better prognosis in a time-dependent manner. The prognostic impact of virological response to HAART in patients receiving immunochemotherapy has not been shown before. In fact, in the R-CDE trial (Spina et al, 2005) the detection of HIV viral load on completion of CHT was associated with borderline significance to increased risk for treatment failure, but this was not translated to an improved survival. The results of the present study clearly suggest that a good control of HIV infection is prognostically important for survival in patients with ARL treated with immunochemotherapy and HAART.

In conclusion, the data from this trial show that immunochemotherapy with R-CHOP is feasible and effective in patients with HIV-related DLBCL. As occurs in non-immunosuppressed patients lymphoma-related parameters, such as the IPI score, together with a good control of HIV viral load, are the most important determinants of prognosis.


This work was supported in part by grants 3690-02 and 36606/06 from FIPSE, and FIJC P-EF/06, from the José Carreras International Leukemia Foundation.


The following institutions and investigators participated in this study: ICO-Hospital Duran y Reynals, L’Hospitalet de Llobregat (E González-Barca, E Domingo), Hospital Gregorio Marañón, Madrid (P Miralles, J Berenguer), ICO-Hospital Germans Trias i Pujol, Badalona (JM Ribera, A Oriol, M Morgades, JT Navarro, B Xicoy), Hospital Clinic, Barcelona (A López-Guillermo), Hospital Vall d’Hebron Barcelona(A López, M Navarrete), Hospital Del Mar, Barcelona (E Abella), ICO-Hospital Josep Trueta, Girona (S Gardella), Consorci Hospitalari, Terrassa (M García), Hospital Son Llàtzer, Palma Mallorca (J Bargay), Hospital de Navarra, Pamplona (A Gorosquieta, K Perez-Equiza), Hospital de Sant Pau, Barcelona (J Briones), Consorci Sanitari, Mataró (Ll Rodríguez, JA Hernández-Rivas), Hospital Puerta de Hierro, Madrid (M Provencio), Hospital Joan XXIII, Tarragona (L Escoda), Hospital Mutua de Terrassa, Terrassa (C Estany), Consorci Hospitalari Parc Taulí, Sabadell (A Soler), Hospital Verge de la Cinta, Tortosa (X Ortín), Hospital La Paz, Madrid (M Canales), Hospital Els Camils, Sant Pere de Ribas (A Asensio), Hospital Virgen de las Nieves, Granada (A Romero), Hospital Dr Peset, Valencia (S Ferrer), Hospital Arnau de Vilanova (E Monzó), Hospital de la Princesa, Madrid (R Arranz), Hospital Gregorio Marañón, Madrid (J Berenguer, P Miralles), Hospital de Figueres, Figueres (P Romero).