Since the introduction of highly active antiretroviral therapy (HAART), HIV-associated morbidity and mortality have shown impressive declines [1,2]. Although previously common opportunistic diseases have revealed unusual clinical presentations and manifestations, the real effect of HAART on non-Hodgkin's lymphoma (NHL) temporal trends still needs to be defined [3–8]. In this regard, it is of particular importance to identify the potential role of the recovered immune function as a result of HAART on the clinical presentation and outcome of AIDS NHL.
In the HAART era, data from HIV-infected patients with NHL seem to indicate higher remission and improved survival rates than those observed among patients treated in the pre-HAART period [9,10]. However, no longitudinal data are yet available concerning the effect of HAART on the presentation and outcome of AIDS-related NHL, and on the tolerability and feasibility of combined treatment approaches with chemotherapy and HAART.
On the basis of these considerations, we conducted a 27 month longitudinal study with the following primary objectives: (i) to determine the response rate and survival of patients with AIDS-related NHL who responded to HAART; and (ii) to analyse the feasibility and toxicity of the combined HAART-chemotherapy approach. The secondary end-point was to evaluate the relationship between the response to HAART and the extent of the tumour at diagnosis.
Materials and methods
Patients and pretreatment evaluation
Between June 1997 and December 1999, all consecutive HIV-infected patients with a newly, histologically confirmed, diagnosis of NHL were enrolled in the study. The study was conducted in two major HIV tertiary care centres in Rome, i.e., the Department of Infectious Diseases of the Catholic University and the National Institute for Infectious Diseases Lazzaro Spallanzani, IRCCS. All of the enrolled patients were 18 years of age or older, had measurable or assessable lymphoid disease and gave their informed consent.
At baseline, a complete blood count, CD4 T cell count and plasma HIV-1-RNA load was collected. The HIV-1 load was measured by Quantiplex HIV-1 branched DNA (Bayer Diagnostics Emeryville, CA, USA) with a detection limit of 50 copies/ml, or by nucleic acid sequence-based amplification HIV-1 RNA QT (Organon Teknika, Durham, NC, USA) with a detection limit of 80 copies/ml; values were adjusted for technique cut-off and the latter was assumed to be comparable to the Amplicor HIV-1 Monitor assay (Roche Molecular System, Sommerville, NJ, USA) .
According to the Ann Arbor system, clinical evaluation of the stage of disease included: chest radiography, computed tomographic or magnetic resonance scanning of the chest, abdomen, pelvis and brain; iliac-crest bone marrow biopsy; lumbar puncture with routine studies and cytological examination. Patients were stratified into three risk group (i.e. at low, intermediate, or high risk), on the basis of the following prognostic factors: Eastern Cooperative Oncology Group (ECOG) performance status greater than 1; previous AIDS diagnosis; and CD4 cell count less than 100 cells/mm3[12,13]. The Revised European American Classification for Lymphoid Neoplasms classification was used .
All patients were scheduled to receive the concurrent administration of chemotherapy and HAART. A cyclophosphamide–hydroxydaunomycin–Oncovin– prednisone (CHOP) regimen was administered for NHL treatment in 34 cases, whereas 10 cases underwent a different chemotherapeutic regimen [i.e. bleomycin–doxorubicin–cyclophosphamide–vincristine–methotrexate–dexamethasone (m-BACOD); modified M-BACOD; methotrexate–doxorubicin–cyclophosphamide–vincristine–prednisone–bleomycin (MACOP-B); Magrath and adriamycin–cyclophospha-mide–vinblastine–bleomycin (ACVB)][13,15–19]. All patients underwent intrathecal prophylaxis for central nervous system localization.
Patients naive to antiretroviral therapies were planned to receive, at NHL diagnosis, at least one protease inhibitor (indinavir, ritonavir, or nelfinavir) or a non-nucleoside reverse transcriptase inhibitor (nevirapine or efavirenz), plus at least two nucleoside transcriptase inhibitors (zidovudine, didanosine, stavudine, lamivudine, abacavir). At NHL diagnosis, patients with a virological or immunological response to HAART (see section on response to HAART) continued the already prescribed antiretroviral regimens. In contrast, in patients who had failed previous antiretroviral treatment, a HAART regimen was implemented according to international guidelines .
Supportive care included Pneumocystis carinii pneumonia prophylaxis (trimethoprim plus sulphamethoxazole once a day, or inhaled pentamidine in sulpha-intolerant patients), anti-bacterial prophylaxis (ciprofloxacin), anti-fungal (fluconazol) and anti-herpetic (acyclovir) prophylaxis. Granulocyte colony-stimulating factor was given as the primary prophylaxis for neutropenia during chemotherapy.
A complete blood count was obtained twice a week during chemotherapy. Plasma HIV-RNA and CD4 T cell counts were performed every 3 months up to 9 months after enrolment.
Response to chemotherapy and toxicity evaluation
An evaluation of response was performed after the first two cycles of chemotherapy, at the end of chemotherapy, and every 3 months thereafter. Complete response (CR) was defined as the absence of clinically detectable disease and negative radiographic findings at any previously abnormal site for at least 4 weeks. The criteria of the ECOG modified by Sparano et al. were adopted for the definition of CR. Partial response (PR) was defined as a decrease of at least 50% in the measurable tumour burden, lasting for at least 4 weeks. Progressive disease (PD) was defined as a reduction of less than 50%, or any increase in measurable disease, or the appearance of the disease at a new site . Toxicity was graded according to WHO criteria.
In a standard CHOP regimen, the method of Hryniuk and Bush  was used to calculate the dose intensity of cyclophosphamide and doxorubicin. The relative dose intensity (RDI), for regimens including cyclophosphamide and doxorubicin other than CHOP, was calculated dividing the dose intensity of cyclophosphamide and doxorubicin of each regimen by the dose intensity of these agents in the CHOP regimen .
Response to highly active antiretroviral therapy
Virological and immunological response criteria were used in this aim. Virological response was defined as a reduction of plasma HIV RNA of at least 2 log10 after 3 months from the beginning of treatment, or an undetectable HIV-RNA load at 6 months in the absence of viral rebound after suppression. Immunological response was defined as an increase, at the last measurement, of the CD4 cell count of more than 30% compared with the baseline value.
Primary end-points of the study included: the rate of CR to chemotherapy; survival from the time of the initiation of chemotherapy; and tolerability of the combined HAART-chemotherapy approach in terms of administered dose intensity of chemotherapy and toxicity. The secondary end-point was the evaluation of the clinical presentation of the tumour according to the response to HAART.
An intention-to-treat approach was adopted. The analysis concerning the RDI was restricted to the patients who underwent regimens including cyclophosphamide and doxorubicin, whereas patients who did not receive HAART concomitantly with chemotherapy were considered as HAART failures.
At univariate analysis, the statistical association between discrete variables was assessed by means of the Fisher's exact test . A stepwise multiple logistic regression model, including variables significantly (P < 0.05) associated with the outcome at univariate analysis, was used to assess the independent role of each variable on the outcomes [24,25]. Odds ratios (OR) and their 95% confidence intervals (CI) were computed.
The mean values of HIV-RNA and CD4 cell count at follow-up times was statistically evaluated by means of the Student's t-test for paired cases; the independent Student's t-test was used to assess differences in the mean values of HAART responder and non-responder groups . Overall survival and disease-free survival rates were estimated by means of the Kaplan–Meier product-limit method . At univariate analysis, the effect of each variable was assessed using the log-rank test . The Cox proportional hazards regression model  was used to assess the independent effect of selected variables in determining the probability of death.
During the study period, 44 HIV-infected patients with NHL matching the inclusion criteria were observed and included in the study. Their general characteristics are listed in Table 1. Thirty-five out of these 44 patients (79.6%) had a diffuse large B cell lymphoma, 13.6% had a Burkitt's lymphoma, and 6.8% a Burkitt-like lymphoma.
Antiretroviral treatment: characteristics and response
At the time of NHL diagnosis, 19 out of 44 patients (43.2%) had already been on HAART for a median time of 385 days (range 120–1095). Eight of these 19 patients (42%) had a virological response to HAART at the time of lymphoma presentation. These eight patients presented an increased probability over non-responders of 1–2 ECOG performance status (100 versus 56%, respectively) (OR not evaluable;P = 0.036) and of 1–2 Ann Arbor stage (37 versus 8%, respectively) (OR = 6.6; 95% CI 1.0–42.2;P = 0.06).
In addition to the 19 patients continuing HAART during the study period after NHL diagnosis, 17 patients started HAART when NHL diagnosis was performed. Eight patients (18.2%) did not take HAART concomitantly with the chemotherapy phase because of the patient's refusal or for gastrointestinal intolerance at assumption of antiretroviral drugs. This was not related to the chemotherapy-induced adverse events.
Among the 36 patients who received HAART concurrently with chemotherapy, stavudine was included in the regimen in 26 patients (72%), lamivudine in 23 (64%), didanosine in 17 (47%), zidovudine in two (6%), abacavir in one (3%), efavirenz in four (11%), nevirapine in three (8%), indinavir in 17 (47%), ritonavir in one (3%), and nelfinavir in 12 (33%).
During the study period, 24 out of 36 patients who received HAART (67%) had a virological response: of these, 18 (75.0%) also had an immunological restoration with a CD4 cell count increase above 30% of the baseline value. No significant associations were observed between virological response to HAART and sex, age, HIV transmission category, previous AIDS diagnosis, baseline CD4 cell count and HIV-RNA level, previous HAART experience, Ann Arbor stage, the number of extranodal sites involved, haematological and systemic toxicities. A significant association was found between 1–2 ECOG performance status and the virological response to HAART after lymphoma diagnosis (OR 7.0; 95% CI 1.45–33.7;P = 0.02).
Effect of highly active antiretroviral therapy on non-Hodgkin's lymphoma response to chemotherapy
A CR was achieved in 23 patients (52.2%), whereas five patients (11.4%) had a PR and 16 patients (36.4%) showed PD. When NHL patients were stratified according to prognostic factors, a non-statistically significant higher percentage of CR (65.2%) was registered in the low–intermediate risk group, compared with the high-risk group (38.1%) (P = 0.07).
CR was more frequently achieved in patients who had a virological response to HAART (71%) than in HAART-failed/naive patients (30%) (OR 5.67; 95% CI 1.54–20.78). An overall response was ultimately achieved in 22 out of 24 (92%) and six out of 20 (30%) of the HAART-responder and HAART-failed/naive patients, respectively (OR 25.67; 95% CI 4.53–145.49). Furthermore, among patients with an intermediate–high-risk NHL, a CR was achieved in 71% of patients responding to HAART and in 26% of patients who failed HAART or were treatment naive (OR 7.00; 95% CI 1.74–28.17). The overall response rates among these intermediate–high-risk patients were 90 and 26% in HAART-responder and non-responder patients, respectively (OR 26.60; 95% CI 4.49–157.61).
At univariate analysis, the presence of CR was associated with having a good ECOG performance status, a virological response to HAART and with an RDI higher than 95% (Table 2). At logistic regression analysis, the virological response to HAART was the only variable independently associated with CR (OR 6.36; 95% CI 1.56–25.81). For each log10 of HIV-1-RNA decrease during HAART, HAART responders showed a 1.7 times higher probability of achieving CR (95% CI 1.14–2.64).
HIV-RNA and CD4 cell count changes during chemotherapy
Overall, in the 36 patients who received concomitantly HAART and chemotherapy, there was a mean change in plasma HIV-RNA between lymphoma diagnosis and the end of chemotherapy of −1.61 log10 copies/ml (P = 0.0001); at 6 months after NHL diagnosis the mean change was −2.82 log10 copies/ml (P = 0.0001), whereas at 9 months it was −2.91 log10 copies/ml (P = 0.0001) (Fig. 1a). Despite comparable mean baseline HIV-RNA levels between patients achieving CR, PR and PD at chemotherapy (4.84, 4.54 and 5.26 log10 copies/ml, respectively), significant differences were found between patients with CR or PD (mean difference −1.48 log10 copies/ml;P = 0.04 at Bonferroni test) and between PR or PD (mean difference −2.21;P = 0.03) at the end of chemotherapy. Moreover, after 9 months of follow-up, statistically significant differences emerged between patients with PD or CR (mean difference +2.46 log10 copies/ml;P = 0.0001) and between those with PD or PR (mean difference +2.64 log10 copies/ml;P = 0.002).
In the 36 HAART-treated patients, a slight reduction of the mean CD4 cell count was registered after the completion of chemotherapy (−47 cells/mm3;P at paired t-test 0.016). CD4 cell counts returned to values higher than baseline at the 6 month assessment (+25 cells/mm3;P = 0.19), and increased significantly at the 9 month follow-up (+131 qcells/mm3;P = 0.0001). Moreover, after 9 months of follow-up a mean increase in the CD4 cell count of 178 cells/mm3 with respect to the time of completion of chemotherapy (P = 0.0001), and of 105 cells compared with the 6 month assessment (P = 0.001) was observed (Fig. 1b). Despite comparable baseline CD4 cell counts between patients achieving CR, PR and PD after chemotherapy, a significant difference was detected at the 9 month follow-up between patients who had PD compared with those with CR (mean difference 173 cells/mm3;P = 0.04 at Bonferroni test).
Dose intensity and toxicity
The mean received dose intensity resulted in 85% of the projected dose (25th–75th percentiles: 69–100): 24 patients (57%) received more than 95% of the projected dose and 30 patients (71%) more than 50%. A significant difference in terms of mean RDI and of the number of patients receiving more than 95% of the projected dose were found comparing patients with virological response to HAART to those without (Table 3). The mean of WHO haematological toxic effects was 1.00 events (25th–75th percentiles: 0.66–1.33) for each cycle of chemotherapy, and the mean of any toxic effect was 1.08 (25th–75th percentiles: 0.70–1.46). In this regard, no differences were found comparing patients with virological response to HAART to those who failed or were HAART naive (Table 3).
The estimated overall median survival of these 44 patients was 360 days (95% CI 81–639), with a 1 year cumulative survival probability of 0.49. In HAART-treated patients, the estimated overall median survival was 425 days (95% CI 241–609), with a 1 year cumulative probability of survival of 0.51. Table 4 reports the stratified results of survival analysis. The 1 year estimated probability of survival was 0.78 in patients with virological response to HAART and 0.84 in those who presented immunological restoration (Fig. 2). At univariate analysis, factors significantly associated with a higher 1 year probability of survival were: RDI > 95%, good performance status at the time of lymphoma diagnosis, virological and immunological response to HAART, CR to chemotherapy. RDI, immunological response to HAART and CR to chemotherapy were independently predictive of a reduced risk of death, as indicated by the results obtained at Cox regression analysis (Table 4). The 1 year cumulative probability of disease-free survival in the 23 HAART-treated patients with CR to chemotherapy was 0.88.
The results of this study indicate that patients with AIDS NHL who respond to HAART show a better response to chemotherapy and have a longer survival than patients who fail or do not receive potent antiretroviral therapy. On the basis of these results, to treat lymphoma patients with HAART starting from the diagnosis of NHL and concomitantly with chemotherapy could be a feasible and tolerable approach, also because it does not seem to be associated with increased toxicity.
With the partial exception of performance status and B symptoms, the prognosis of AIDS lymphoma in our HAART-treated patients was not related to any previously reported prognostic factor, such as older age, CD4 cell count less than 100 cells/mm3, previous AIDS diagnosis, bone marrow involvement or extranodal disease [12,13,27,28]. However, we demonstrated that virological response to HAART is the only factor independently associated with a higher response rate to chemotherapy, by determining a more than sixfold increased probability of CR. Moreover, the likelihood of tumour regression increased 1.7 times for each log10 decrease of plasma viraemia during therapy.
This is the first report on the changes in the natural history of AIDS NHL not simply derived from data regarding exposure to antiretroviral agents, but obtained in a prospective manner by assessing the effect of the response to HAART. The improvement of survival and of the remission rate in AIDS lymphomas developing in the HAART era has already recently been documented by two studies [9,10], but were not confirmed in another cohort study . The lack of a control group represented by HAART non-responder patients in these studies makes it difficult to draw definitive conclusions regarding the real impact of virological control and immune restoration on the clinical course of AIDS lymphoma. In a recent study , the use of HAART had a protective effect on the development of NHL and the nadir CD4 cell count was an independent predictor of tumour occurrence. Nevertheless, no survival differences were detected according to time periods or HAART exposure, even though the same authors suggested that this may simply reflect insufficient sample size and follow-up. In the same study, no data according to virological or immunological response after HAART are reported.
In the pre-HAART era, several studies [13,15,27–30] reported different rates of response to chemotherapy for AIDS-related NHL, ranging from 14 to 70%, mainly depending on the presence of risk factors given by general health status, immunological parameters, pre-existing AIDS, tumour stage and some histological properties [15,27,28]. It has been observed that high-risk patients defined by a CD4 cell count of less than 100 cells/mm3, ECOG performance status greater than 1, and previous AIDS-defining events, have a very low CR rate (14%) . Although our study population was largely represented by high-risk patients (48%), an overall response of 92% to chemotherapy was obtained in HAART-responder patients, with a 1 year survival probability of 78%, which is clearly higher than the rates reported in all previous major studies [13,17,18] in which the concomitant use of potent combination antiretroviral treatment was not planned. Furthermore, the CR rate of 71% found in our HAART-responder patients was one of the best results ever reported. Also, considering only high-risk patients with a response to HAART, CR to chemotherapy was 57% and overall response was 84%. Because the distribution of the type of chemotherapeutic regimen (CHOP versus non-CHOP) was similar among the different NHL prognostic groups, our findings could not be explained by the administration of a more aggressive chemotherapy in low-risk patients. Moreover, virological response to HAART did not differ among low- versus intermediate–high-risk NHL groups (data not shown). These results strongly suggest that response to HAART is essential to treat with a curative intent patients who in the past would have been selected for palliative or less aggressive chemotherapy. To date, the concomitant administration of HAART and chemotherapy represents a question of importance , and a higher risk of toxicity, mainly haematological or neurological, has previously been observed . In our study, combining HAART with chemotherapy was not associated with a higher risk of toxicity. Furthermore, it might be hypothesized that the response to HAART ameliorates the tolerability of chemotherapy, as suggested by the finding that patients with virological response to HAART were able to receive a significantly higher dose intensity of chemotherapy. As an explanation, haematopoiesis could be improved by means of the inhibition of HIV-1 replication itself, but also by the inhibition of indirect myelosuppressive mechanisms, and by a direct or indirect stimulation of the bone marrow . The higher chemotherapy dose intensity has been correlated with a good prognosis in immunocompetent patients with lymphomas  and might, at least partly, explain the favourable effect of HAART on response rate and survival.
Another debated argument is the impact of the combined approach chemotherapy-HAART on the virological and immunological control of HIV infection. It has previously been reported that the temporary discontinuation of antiretroviral therapy during chemotherapy does not prevent the control of HIV replication and the subsequent immune recovery and allows the administration of a more adequate dose intensity . The results of our study clearly demonstrate that the administration of HAART during chemotherapy does not negatively affect the virological response to antiretroviral agents, and that even though it causes a decline of CD4 cells strictly limited to the first months after the diagnosis of lymphoma, it is followed by a significant and sustained improvement of the immune parameter. Moreover, the control of underlying HIV-1 replication and CD4 cell restoration were more prominent in patients who had a response to chemotherapy.
The results of our study suggest that the concurrent administration of HAART and chemotherapy could deeply modify the natural history of NHL in HIV-infected patients, increasing the response rate to chemotherapy and prolonging survival. Moreover, HAART might increase the tolerability of a higher chemotherapy dose intensity. On the basis of these results, the combination of HAART and chemotherapy may be an effective and feasible approach.
Preliminary results of this work were presented at the 7th Conference on Retroviruses and Opportunistic Infections, San Francisco, CA, USA, 30 January–2 February 2000.
1. Palella FJ, Delaney KM, Moorman AC. et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. N Engl J Med 1998, 338: 853–860.
2. Jacobson MA, French M. Altered natural history of AIDS-related opportunistic infections in the era of potent combination antiretroviral therapy. AIDS 1998, 12: S157–S163.
3. Jacobson LP, Yamashita TE, Detels R. et al. Impact of potent antiretroviral therapy on the incidence of Kaposi's sarcoma and non-Hodgkin's lymphomas among HIV-1-infected individuals. J Acquir Immune Defic Syndrome 1999, 21: S34–S41.
4. Rabkin CS, Testa MA, Huang J. et al. Kaposi's sarcoma and non-Hodgkin's lymphoma incidence trends in AIDS Clinical Trial Group Study participants. J Acquir Immune Defic Syndrome 1999, 21: S31–S33.
5. Jones JL, Hanson DL, Dworkin MS. et al. Effect of antiretroviral therapy on recent trends in selected cancers among HIV-infected persons. J Acquir Immune Defic Syndrome 1999, 21: S11–S17.
6. Buchbinder SP, Holmberg SD, Scheer S. et al. Combination antiretroviral therapy and incidence of AIDS-related malignancies. J Acquir Immune Defic Syndrome 1999, 21: S23–S26.
7. Ledergerber B, Telenti A, Egger M. Risk of HIV related Kaposi's sarcoma and non-Hodgkin's lymphoma with potent antiretroviral therapy: prospective cohort study. BMJ 1999, 319: 23–24.
8. Matthews GV, Bower M, Mandalia S, Powles T, Nelson MR, Gazzard BG. Changes in acquired immunodeficiency syndrome related lymphoma since the introduction of highly active antiretroviral therapy. Blood 2000, 96: 2730–2734.
9. Evison J, Jost J, Ledergerber B. et al. HIV-associated non-Hodgkin's lymphoma: highly active antiretroviral therapy improves remission rate of chemotherapy. AIDS 1999, 13: 732–734.
10. Thiessard F, Morlat P, Marimoutou C. et al. Prognostic factors after non-Hodgkin's lymphoma in patients infected with the human immunodeficiency virus: Aquitaine cohort, France, 1986–1997. Cancer 2000, 88: 1696–1702.
11. Elbeik T, Charlebois E, Nassos P. et al. Quantitative and cost comparison of ultrasensitive human immunodeficiency virus type 1 RNA viral load assays: Bayer bDNA Quantiplex versions 3.0 and 2.0 and Roche PCR Amplicor Monitor Version 1.5. J Clin Microbiol 2000, 38: 1113–1120.
12. Straus DJ, Huang J, Testa MA. et al. Prognostic factors in the treatment of human immunodeficiency virus-associated non-Hodgkin's lymphoma: analysis of AIDS Clinical Trials Group protocol 142 – low dose versus standard m-BACOD plus granulocyte–macrophage colony-stimulating factor. National Institute of Allergy and Infectious Diseases. J Clin Oncol 1998, 16: 3601–3606.
13. Tirelli U, Spina M, Gabarre J. et al. Treatment of HIV-related non-Hodgkin's lymphoma adapted to prognostic factors [Abstract]. J Acquir Immune Defic Syndrome 1999, 21: A32.A32.
14. Harris NL, Jaffe ES, Stein H. et al. A revised European-American classification of lymphoid neoplasms: a proposal from the International Lymphoma Study Group. Blood 1994, 84: 1361–1392.
15. Gisselbrecht C, Oksenhendler E, Tirelli U. et al. Human immunodeficiency virus-related lymphoma treatment with intensive combination chemotherapy. Am J Med 1993, 95: 188–196.
16. Kaplan LD, Kahn J, Crowe S. et al. Clinical and virologic effects of recombinant human granulocyte macrophage colony stimulating factor in patients receiving chemotherapy for human immunodeficiency virus-associated non-Hodgkin's lymphoma: results of a randomized trial. J Clin Oncol 1991, 9: 929–940.
17. Sparano JA, Wiernik PH, Hu X. et al. Pilot trial of infusional cyclophosphamide, doxorubicin, and etoposide plus didanosine and filgrastim in patients with human immunodeficiency virus-associated non-Hodgkin's lymphoma. J Clin Oncol 1996, 14: 3026–3035.
18. Kaplan LD, Straus DJ, Testa. et al. Low-dose compared with standard-dose m-BACOD chemotherapy for non-Hodgkin's lymphoma associated with human immunodeficiency virus infection. N Engl J Med 1997, 336: 1641–1648.
19. Tirelli U, Errante D, Spina M. et al. Second-line chemotherapy in human immunodeficiency virus-related non-Hodgkin's lymphoma: evidence of activity of a combination of etoposide, mitoxantrone, and prednimustine in relapsed patients. Cancer 1996, 77: 2127–2131.
20. Antiretroviral therapy in adults. Updated recommendations of the International AIDS Society – USA Panel. JAMA 2000, 283:381–390.
21. Hryniuk WM, Bush H. The importance of dose intensity chemotherapy of metastatic breast cancer. J Clin Oncol 1984, 2: 1281–1288.
22. Meyer RM, Hryniuk WM, Goodyear MDE. The role of dose intensity in determining outcome in intermediate-grade non-Hodgkin's lymphoma. J Clin Oncol 1991, 9: 339–347.
23. Armitage P, Berry G. Statistical methods in medical research. Oxford: Blackwell Scientific Publications; 1987.
24. Breslow NE, Day NE. The analysis of case–control studies. In:Statistical methods in cancer research, Vol. 1. IARC Scientific Publication No. 32, Lyon; 1980.
25. Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 1958, 53: 457–481.
26. Cox DR, Oakes D. Analysis of survival data. London, United Kingdom: Methuen; 1984.
27. Kaplan LD, Abrams DI, Feigel E. et al. AIDS-associated non-Hodgkin's lymphomas in San Francisco. JAMA 1989, 261: 719–724.
28. Levine AM, Sullivan-Halley J, Pike MC. et al. Human immunodeficiency virus-related lymphoma. Cancer 1991, 68: 2466–2472.
29. Little RF, Pearson D, Franchini G. et al. Dose-adjusted EPOCH chemotherapy in previously untreated HIV-associated non-Hodgkin's lymphoma: preliminary report of efficacy, immune reconstitution, and HIV control following therapy [Abstract]. J Acquir Immune Defic Syndrome 1999, 21: A33.A33.
30. Tirelli U, Errante D, Oksenhendler E. et al. Prospective study with combined low-dose chemotherapy and zidovudine in 37 patients with poor prognosis AIDS-related non-Hodgkin's lymphoma. Ann Oncol 1992, 3: 843–847.
31. Vaccher E, Spina M, Talamini R. et al. Concomitant CHOP chemotherapy and highly active antiretroviral therapy in patients with HIV-related non-Hodgkin's lymphoma [Abstract]. J Acquir Immune Defic Syndrome 1999, 17: A30.A30.
32. Sparano JA, Hu X, Wiernik PH. et al. Opportunistic infection and immunologic function in patients with human immunodeficiency virus-associated non-Hodgkin's lymphoma treated with chemotherapy. J Natl Cancer Inst 1997, 89: 301–307.
© 2001 Lippincott Williams & Wilkins, Inc.