Fax: (212) 717-3342
Intensive chemotherapy with cyclophosphamide, doxorubicin, high-dose methotrexate/ifosfamide, etoposide, and high-dose cytarabine (CODOX-M/IVAC) for human immunodeficiency virus–associated Burkitt lymphoma
Version of Record online: 1 AUG 2003
Copyright © 2003 American Cancer Society
Volume 98, Issue 6, pages 1196–1205, 15 September 2003
How to Cite
Wang, E. S., Straus, D. J., Teruya-Feldstein, J., Qin, J., Portlock, C., Moskowitz, C., Goy, A., Hedrick, E., Zelenetz, A. D. and Noy, A. (2003), Intensive chemotherapy with cyclophosphamide, doxorubicin, high-dose methotrexate/ifosfamide, etoposide, and high-dose cytarabine (CODOX-M/IVAC) for human immunodeficiency virus–associated Burkitt lymphoma. Cancer, 98: 1196–1205. doi: 10.1002/cncr.11628
- Issue online: 5 SEP 2003
- Version of Record online: 1 AUG 2003
- Manuscript Accepted: 16 JUN 2003
- Manuscript Revised: 10 JUN 2003
- Manuscript Received: 1 APR 2003
- National Institutes of Health. Grant Number: CA-09207-24
- human immunodeficiency virus;
- Burkitt lymphoma;
- intensive chemotherapy;
- non-Hodgkin lymphoma
In the era of highly active antiretroviral therapy (HAART), standard-dose chemotherapy for human immunodeficiency virus (HIV)-associated diffuse large B-cell lymphoma is becoming the standard of care. In contrast, the safety and efficacy of intensive regimens have not been established for Burkitt lymphoma (BL), a highly aggressive lymphoma for which moderate-dose chemotherapy is substandard in the HIV-negative population.
To evaluate the feasibility of intensive chemotherapy in HIV-associated BL, the authors retrospectively reviewed 14 HIV-positive adults with BL treated at their center between 1988 and 2000. Eight patients were treated between 1996 and 2000 with cyclophosphamide, doxorubicin, high-dose methotrexate/ifosfamide, etoposide, and high-dose cytarabine (CODOX-M/IVAC), one of the currently preferred intensive-dose chemotherapy regimens for BL. Six received other chemotherapy. Outcomes were compared with those of 24 HIV-negative adult patients with BL who had similar patient characteristics and were treated concomitantly (13 with CODOX-M/IVAC; 11 with other regimens).
Of the 14 HIV-positive patients, 63% had a complete response after CODOX-M/IVAC treatment, compared with 67% of patients receiving other chemotherapy. The 2-year event-free survival (EFS) rates were 60% and 60% after CODOX-M/IVAC or other regimens, respectively. Similar outcomes were seen despite the fact that 88% of CODOX-M/IVAC-treated HIV-positive patients had Stage IV disease, compared with one-third of HIV-positive patients treated with other chemotherapy. HIV status did not adversely affect long-term EFS independent of the treatment regimen (P = 0.88). When EFS was evaluated according to chemotherapy regimen independent of HIV status, CODOX-M/IVAC was found to be associated with improved EFS (P = 0.05) in all patients, and particularly those at high risk. HIV-positive patients treated with CODOX-M/IVAC tolerated chemotherapy well with similar rates of myelosuppression and infectious complications as HIV-negative patients.
The current nonrandomized retrospective study suggested that intensive chemotherapy with CODOX-M/IVAC is feasible and well tolerated in HIV-positive adults with BL. In the post-HAART era, intensive chemotherapy such as CODOX-M/IVAC may be appropriate in all adult patients with BL, and especially those with poor prognostic factors, regardless of HIV status. Cancer 2003;98:1196–205. © 2003 American Cancer Society.
Burkitt lymphoma (BL) is an aggressive non-Hodgkin' lymphoma (NHL) characterized by rapid proliferation and a nearly 100% growth phase fraction. Originally identified as ‘undifferentiated lymphoma’, it has since been classified as diffuse small noncleaved cell lymphoma in the Working Formulation1 and as Burkitt lymphoma in the revised European-American Lymphoma2 and World Heath Organization (WHO) classification systems.3 In the general population, BL is more common in children and adolescents, but it accounts for 1–2% of NHL in adults. Most patients present with advanced disease involving multiple extra lymph node sites, most commonly in the abdomen. Although BL was one of the first cancers in which a response to chemotherapy was observed,4 early recurrences occurred, often in the central nervous system (CNS), with rapid disease progression.5 Cure rates of 50–60% were achieved in early-stage patients treated with chemotherapy regimens containing high-dose cyclophosphamide, antimetabolites, and prophylactic intrathecal chemotherapy. However, only 20% of patients with bone marrow (BM) or CNS involvement achieved durable responses.6–8 In 1996, Magrath et al.9 reported a 2-year event-free survival (EFS) rate of 92% in HIV-negative adult and pediatric patients after intensive chemotherapy with cyclophosphamide, doxorubicin, high-dose methotrexate/ifosfamide, etoposide, and high-dose cytarabine (CODOX-M/IVAC). Results were particularly impressive in patients with BM and/or CNS involvement, with a 2-year disease-free survival rate of 80%.9 A recent international multicenter Phase II study confirmed that CODOX-M/IVAC led to high cure rates, particularly in the international prognostic index stratified high-risk patients with a reported 2-year EFS rate of 65% and a 2-year overall cure rate of 72.8%, although results were inferior to the previous McGrath et al. report.9, 10
BL comprises 25–40% of NHL in human immunodeficiency virus (HIV)-infected individuals.11–13 Clinically, HIV-associated BL closely resembles BL in the general population. A recent review of 75 adults with BL (46 HIV-positive) found no difference between HIV-positive and HIV-negative patients in terms of disease stage, BM (33–35%) and CNS (17–19%) involvement,14 and histology.15
The optimal treatment for adult patients with HIV-associated BL is not clear. In the era before highly active antiretroviral therapy (HAART), combination chemotherapy was less successful for patients with HIV-associated NHL than for HIV-negative patients with similar NHL histopathology. Early deaths were often due to opportunistic infection.16 Improved immune function in the HAART era has led to a reevaluation of full-dose chemotherapy for HIV-associated lymphoma. It has been proposed that HIV-positive patients with BL and preserved immune function may represent a subset of patients with acquired immunodeficiency syndrome (AIDS)-related lymphoma who would benefit from more aggressive chemotherapeutic approaches with acceptable toxicities.6 However, because of the perceived risk of increased hematologic and infectious complications, most patients with HIV-associated BL continue to be treated with cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) and other moderate-dose chemotherapy regimens, despite the fact that such treatment is known to be inferior to intensive chemotherapy in HIV-negative patients with BL.
At the Memorial Sloan-Kettering Cancer Center (MSKCC), we routinely treat patients with BL with chemotherapy regimens irrespective of HIV status. To evaluate the feasibility of intensive chemotherapy on BL in the HAART era, we retrospectively reviewed 14 HIV-positive adult patients with BL treated at MSKCC center between 1988 and 2000. Eight patients received an intensive regimen (CODOX-M/IVAC) between 1996 and 2000, and six patients were treated with nonintensive regimens. Treatment outcomes and toxicities were compared with 24 concomitantly treated HIV-negative patients with BL.
MATERIALS AND METHODS
Between March 1988 and June 2000, 14 HIV-positive and 24 HIV-negative adults (older than 18 years old) with previously untreated, newly diagnosed Burkitt lymphoma were treated at MSKCC. For patients known or suspected to be HIV-positive, the presence of HIV infection was assayed by enzyme-linked immunosorbent assay and confirmed by Western blots. Viral loads (VL) and T-cell subset analysis were available for some patients. The study was reviewed by the institutional review board and determined to be exempt research.
All standard-dose treatment regimens were administered as previously described. The CODOX-M/IVAC regimen was administered according to the original report by Magrath et al.9 For example, patients with a single extra abdominal mass or completely resected abdominal disease and a normal serum level of lactate dehydrogenase (LDH) were considered to be at low risk and received three cycles of CODOX-M. All other patients were designated as being at high risk and received four alternating cycles of CODOX-M, IVAC, CODOX-M, and IVAC. CODOX-M consisted of cyclophosphamide (800 mg/m2 on Day 1, 200 mg/m2 on Days 2–5), doxorubicin (40 mg/m2 on Day 1), vincristine (1.5 mg/m2 on Days 1 and 8 during Cycle 1 and on Days 1, 8, and 15 during Cycle 2), and high-dose methotrexate (1200 mg/m2 over 1 hour followed by 240 mg/m2/hour over 23 hours). All patients received intrathecal cytarabine 70 mg on Days 1 and 3 and intrathecal methotrexate 12 mg on Day 15. IVAC consisted of ifosfamide (1500 mg/m2 on Days 1–5 with sodium mercaptoethanesulfonate), etoposide (60 mg/m2 on Days 1–5), high-dose cytarabine (2000 mg/m2 every 12 hours on Days 1 and 2 for 4 doses), and intrathecal methotrexate (12 mg on Day 5). Patients with CNS disease received additional intrathecal cytarabine and methotrexate.
Complete response (CR) was defined as resolution of all measurable disease as determined by clinical and radiologic evaluation for at least 4 weeks. No response was defined as less than CR or overt disease progression during chemotherapy. EFS was defined as the date of lymphoma diagnosis to the date of the first observed progressive disease or death due to any cause. These response criteria are more stringent than those recently defined by the National Cancer Institute (NCI) Working Group.17
Survival curves were generated using the Kaplan–Meier method.18 When disease recurrences or progressive disease occurred in patients with BL, death usually followed within 3–12 months. Patients surviving more than 1 year after treatment were likely to be cured of disease. Consequently, EFS and overall survival were very similar. For simplification of presentation, we chose to present only EFS. The log-rank19 and Fisher exact tests20 were used to compare EFS and CR rates between patients treated with CODOX-M/IVAC and patients treated with other regimens and between HIV-positive and HIV-negative patients. The Cox proportional hazards ratio model was used to find the treatment effect after adjusting clinical variables.
The diagnosis of BL was confirmed at MSKCC (J.T.-F.). Thirty-seven patients had BL and one patient had Burkitt-like lymphoma, as defined by the WHO classification criteria.3, 21 Immunohistochemistry demonstrated a B-cell phenotype in all patients. Cytogenetic analysis was performed on soft tissue specimens from four patients, three of which revealed a t(8;14), and on two involved BM specimens, neither of which showed evidence of the translocation. Epstein–Barr virus status was not determined. Pretreatment evaluation included hematologic and blood chemistry tests, including LDH level, chest radiography, computed tomography (CT) scan of the abdomen and pelvis, BM biopsy and/or aspiration, and lumbar puncture. Elevated serum LDH levels were defined as greater than 200 U/L. Additional tests, including CT/magnetic resonance imaging (MRI) scans of the brain, spinal cord, or thorax, gallium scans, bone scans, and evaluation of pleural and ascites fluid, were performed as clinically indicated. Patients were staged according to the Ann Arbor classification system.22 CNS involvement was defined as malignant pleocytosis and/or evidence of leptomeningeal and/or epidural involvement on CT/MRI scans with focal neurologic signs.
Human Immunodeficiency Virus–Infected Patients with Burkitt Lymphoma
Fourteen HIV-positive patients with BL were treated between 1988 and 2000. With one exception, all patients presenting with BL after January 1, 1997 received CODOX-M/IVAC regardless of HIV status. Therefore, eight HIV-positive patients were treated with CODOX-M/IVAC9 from 1996 to 2000. Three HIV-positive patients were treated with CHOP-based chemotherapy.23 Of the remaining three HIV-positive patients with BL, one received the L-20 protocol for acute lymphoblastic leukemia (ALL),24 one received m-BACOD,25 and one received NHL-15.26
Table 1 summarizes the characteristics of the 14 HIV-positive adult patients with BL. Among the HIV-infected patients, the median age was 40 years (range, 19–61 years). The median overall follow-up period was 34.3 months (range, 4–115 months). Eight patients were known to be HIV-positive before the BL diagnosis. Of these, four patients were known to have been taking antiretroviral therapy before chemotherapy for BL. One patient discontinued zidovudine (AZT), lamuvidine, and protease inhibitors before chemotherapy due to gastrointestinal involvement by BL and did not restart HAART. Two patients continued antiretroviral medications (lamivudine/stavudine/nelfinavir and lamivudine/stavudine/nevirapine, respectively) during chemotherapy. One patient discontinued didanosine and AZT during chemotherapy for unknown reasons and died. Four patients known to be HIV-positive before BL were not receiving antiretroviral medications at BL diagnosis. Of these, two died of refractory lymphoma without starting HAART, one patient did well without HAART, and one began HAART (AZT/ritonavir/lamivudine) after CODOX-M/IVAC. Five patients were diagnosed with HIV infection in the setting of BL. Of these five patients, three began antiretroviral therapy after the first cycle of CODOX-M/IVAC and continued throughout intensive chemotherapy: two received HAART (stavudine/lamivudine/indinavir and stavudine/lamivudine/nelfinavir, respectively) and one received AZT/lamivudine/efavirenz but was switched to stavudine/lamivudine/efavirenz due to leukopenia. Two of the five patients diagnosed with HIV in the setting of BL did not receive HAART during chemotherapy. One of the 14 HIV-positive patients in the current study was diagnosed with HIV infection 4 years after chemotherapy for BL.
|Characteristic||HIV-positive patients (%)||HIV-negative patients (%)|
|No. of patients||14||24|
|Male||12 (86)||15 (63)|
|IV||9 (64)||13 (54)|
|Karnofsky PS > 70%||3||9|
|CNS disease||3 (21)||7 (29)|
|BM disease||4 (29)||10 (42)|
|CODOX-M/IVAC||8 (57)||13 (54)|
|Poor prognosisa||11 (79)||13 (54)|
|CD4 count (cells/μL)|
|Viral load (copies/mL)c|
|BL = AIDS-defining disease||—e||—|
Due to the timing of the current retrospective study (1988–2000), some patients were diagnosed and treated for BL before the routine use of CD4 counts and VL in HIV-infected individuals, i.e., before 1995. Therefore, total CD4-positive cells were available for 9 of the 14 HIV-positive patients at the time of BL diagnosis. CD4 counts ranged from 10 to 555 cells, with a mean of 185 ± 65 cells per microliter and a median of 149 cells per microliter. Three patients had CD4 counts less than 50 cells per microliter. The remaining patients had CD4 counts from 52 to 555 cells per microliter. Follow-up data for 5 HIV-positive patients (4 treated with CODOX-IVAC, one with other chemotherapy) showed an increase in CD4 counts from a mean of 105 ± 82 cells per microliter at the start of chemotherapy to 182 ± 38 cells per microliter at 8–12 months follow-up and to 386 ± 71 cells per microliter 2 years after therapy.
HIV VL data were available for 9 of the 14 HIV-positive patients with BL and for only 3 patients at the time of BL diagnosis. One patient had a VL of less than 500 copies per milliliter, and 2 patients had VLs of 438,646 and 6357 copies per milliliter, respectively. Following chemotherapy trends in HIV VL correlated with clinical outcome. In 1 patient, VL decreased from 438,646 copies per milliliter at diagnosis to 3576 copies per milliliter after 2 months and to an undetectable level 5 months after the start of concomitant CODOX-M/IVAC and HAART. Another patient who was diagnosed with HIV during chemotherapy received HAART and experienced a decrease in HIV VL from 6761 to 1433 copies per milliliter after 3 months and undetectable levels after 6 months. Both of these patients have remained in CR. In contrast, 2 patients with persistently high VL (219,895 and 92,897 copies per milliliter, respectively) after 3 months of chemotherapy did poorly and died of lymphoma and HIV-related thrombotic thrombocytopenic purpura (TTP). Six long-term HIV-positive survivors still were receiving HAART and had undetectable VLs at their last follow-up visit.
At the time of BL diagnosis, the mean Karnofsky performance status (KPS) of HIV- infected patients ranged from 50% to 90%, with a mean value of 70%. Of 14 HIV-positive patients with BL, 10 had a KPS less than or equal to 70%. Of the eight patients known to be HIV-positive before the BL diagnosis, four had existing AIDS-defining illnesses such as viral gastroenteritis (not otherwise specified), shingles, oral thrush, molluscum contagiosum, and immune thrombocytopenic purpura.
All patients had measurable lymphomatous disease at the time of initial presentation. Nine HIV-positive patients (64%) presented with Stage IV disease. Three patients had known CNS disease and four patients had BM involvement. Nine patients had evidence of extranodal involvement. Seven patients had elevated serum LDH levels (range, < 150–2220 U/L). Overall, 11 patients (78%) had high-risk disease.
Outcome by treatment group for HIV-positive patients with BL is outlined in Table 2. Eight HIV-positive patients (58%) received CODOX-M/IVAC chemotherapy for treatment of BL. Of these, 7 patients (88%) had Stage IV disease. Five patients (63%) achieved CR, and 3 patients died (1 due to progressive lymphoma [within 4 months]; 1 due to toxicity [respiratory failure due to Pneumocystis carinii pneumonia acquired during chemotherapy-induced Neutropenia, 4 months after diagnosis]; and 1 due to TTP unrelated to lymphoma, 12 months after BL diagnosis, during CR). Six HIV-positive patients received chemotherapy other than CODOX-M/IVAC. Of these patients, only 2 (30%) had Stage IV disease. Four patients (66%) achieved CR, and 2 died of progressive lymphoma (1 patient who never achieved CR, at 6 months, and 1 patient with recurrent disease, 15 months after BL diagnosis). Similar CR (63% vs. 67%) and 2-year EFS rates (60% vs. 60%) were noted after CODOX-M/IVAC and non-CODOX-M/IVAC regimens despite the disproportionate high-risk features in the former group.
|Characteristic||CODOX-M/IVAC (%)||Other (%)|
|No. of patients||8||6|
|Stage IV disease||7 (88)||2 (33)|
|No. of deaths||3||2|
|CR||5 (63)||6 (67)|
|2-yr EFS||4 (60)||3 (60)|
Outcomes Compared with Concurrently Treated Human Immunodeficiency Virus-Negative Patients with Burkitt Lymphoma
Outcomes of the HIV-infected patients were compared with the outcomes of 24 HIV-negative patients with BL treated over the same time interval. With one exception, all patients presenting with BL after January 1, 1997 received CODOX-M/IVAC. Of the 24 HIV-negative patients with BL, 13 received CODOX-M/IVAC.9 Ten HIV-negative patients received other first-line chemotherapy including CHOP23 (six patients), L-2024 (two patients), the Vanderbilt regimen27 (one patient), doxorubicin and cyclophosphamide (one patient), and a pediatric Burkitt regimen (one patient).
Patient characteristics and outcomes of these HIV-negative patients are compared with those of HIV-infected patients in Table 1. HIV-positive patients had a median age of 41 years (range, 19–61 years), which is comparable to the HIV-negative patients (median, 37 years; range, 19–66 years). Sixty-four percent of the HIV-positive patients had Ann Arbor Stage IV disease compared with 54% of HIV-negative patients. Although 10 HIV-negative patients had BM involvement, high-risk disease was more prevalent in the HIV-positive patients (79% vs. 54%) at presentation.
When outcomes of the HIV-positive and HIV-negative patients were grouped and analyzed without regard to chemotherapy, HIV status did not have a significant impact on long-term EFS (Fig. 1). However, when the EFS of all 38 (14 HIV-positive, 24 HIV-negative) patients with BL was evaluated by chemotherapy regimen (CODOX-M/IVAC vs. non-CODOX-M/IVAC), the CODOX-M/IVAC regimen was associated with an improved EFS (P = 0.05; Fig. 2). The CODOX-M/IVAC regimen resulted in a CR rate of 81% versus 41% after other regimens (Fisher' exact test, P = 0.02). The 2-year EFS rate after CODOX-M/IVAC was 80% compared with 38% after other regimens. The 2-year EFS rates were 59% and 92% for HIV-positive and HIV-negative patients, respectively. However, 7 of 8 HIV-positive patients had Stage IV disease, whereas only 6 of 13 HIV-negative patients had Stage IV disease. HIV-positive and HIV-negative patients with BL achieved similar CR rates (64% vs. 63%) after all chemotherapy programs with about one-third of all patients failing primary treatment (Table 3).
|Characteristic||All patients (%)||HIV status||Chemotherapy|
|Stage IV (%)||22||9 (64)||13 (54)||13 (62)||9 (53)|
|CNS or BM||18||6||12||11||7|
|CR (%)||24 (63)||9/14 (64)||15/24 (63)||17/21 (81)||7/17 (41)|
|2EFS (%)||14 (62)||5 (59)||9 (62)||4 (80)||10 (38)|
Sixty-nine percent (n = 9) of patients with Stage IV disease and 64% (n = 7) of patients with CNS and/or BM disease achieved a CR after receiving CODOX-M/IVAC chemotherapy, compared with 22% and 0%, respectively, for patients receiving other regimens. The median survival of all patients with Stage IV BL who were treated with CODOX-M/IVAC regardless of HIV status was 31 months, compared with a median survival of 7 months for all patients with Stage IV disease who were treated with non-CODOX regimens. Two-thirds of all patients with high-risk disease achieved CR and were alive 2 years after completion of CODOX-M/IVAC, compared with less than 15% of patients treated with other regimens (Table 3).
Because of the aggressive nature of this high-grade lymphoma, the majority of patients with BL were either cured of their malignancy or died of rapidly progressive refractory lymphoma within 6–12 months of diagnosis. Of the 3 HIV-positive patients with BL who died due to lymphoma, only 1 patient died of recurrent disease refractory to second-line therapy, 15 months after diagnosis. Of the 10 HIV-negative patients who died, 9 died in less than 11 months due to progressive lymphoma and 1 died of unknown causes 32 months after the BL diagnosis.
Toxicities Associated with CODOX-M/IVAC
NCI Grade 3/4 toxicities after CODOX-M/IVAC therapy for HIV-positive (8) and HIV-negative (14) patients with BL are summarized in Table 4. Toxicities after CODOX-M/IVAC were primarily hematologic, with almost universal pancytopenia and high incidences of neutropenic fever, documented infection, and mucositis. All HIV-positive patients treated with CODOX-M/IVAC tolerated intensive chemotherapy with incidences of myelosuppression and mucositis similar to those of HIV-negative patients. The time to neutrophil recovery was affected by the use of granulocyte–colony-stimulating factor after chemotherapy in many patients. Measurement of the time to platelet recovery was logistically difficult to perform given the multiple drug-dosing regimen inherent to CODOX-M/IVAC. However, among patients treated with CODOX-M/IVAC, HIV-positive patients experienced significantly longer thrombocytopenic periods (mean, 117 ± 29 days [n = 7]) than did HIV-negative patients (mean, 42 ± 7.6 days, [n = 11]). These calculations exclude 1 HIV-positive patient who developed TTP after chemotherapy and 3 HIV-negative patients who developed chronic mild thrombocytopenia lasting up to 3 years after CODOX-M/IVAC. Due to confounding factors, as well as the small number of patients, the effects of HIV status and HAART on BM recovery could not be meaningfully determined.
|Total (%)||HIV-positive (%)||HIV-negative (%)|
|No. of patients||21||8||14|
|Neutropenia||21 (100)||7 (88)||14 (100)|
|Thrombocytopenia||15 (71)||6 (75)||9 (64)|
|Anemia||21 (100)||8 (100)||13 (93)|
|Nadir fever||20 (95)||7 (88)||13 (93)|
|Nonneutropenic infection||7 (33)||5 (63)||2 (14)|
|Culture-positive infection||14 (67)||3 (38)||11 (79)|
|Sepsis||5 (24)||3 (38)||2 (14)|
|Ileus||4 (19)||0||4 (29)|
|Neuropathy||5 (24)||3 (38)||2 (14)|
|Mucositis||14 (67)||6 (75)||8 (57)|
|Tumor lysis||3 (14)||3 (38)||2 (14)|
|Renal toxicity||5 (24)||1 (13)||2 (14)|
|Toxicity-related deaths||2 (9)||1||1|
Although nonneutropenic fever was more common among HIV-infected patients compared with HIV-negative patients (63% vs. 14%), HIV-negative patients had more culture-positive infections (79% vs. 38%). Among this small cohort, more HIV-infected individuals experienced peripheral neuropathy, mucositis, sepsis, tumor lysis syndrome, and renal and gastrointestinal side effects than HIV-negative individuals. Two patients (one HIV-positive and one HIV-negative) died of CODOX-M/IVAC-related toxicity.
Similar rates of opportunistic infections were observed in HIV-positive patients treated with either CODOX-M/IVAC or non-CODOX regimens. Two episodes of P. carinii pneumonia, one cytomegalovirus (CMV) pneumonia, and one herpes simplex virus (HSV) infection occurred in the CODOX-M/IVAC-treated HIV-positive patients compared with one P. carinii pneumonia, one CMV pneumonia, and one HSV infection among the non-CODOX–treated HIV-positive patients.
We reviewed 14 HIV-positive adult patients with BL treated at MSKCC between 1988 and 2000. Despite reluctance in the oncologic community to apply intensive regimens in patients with HIV infection, it has been our policy to administer chemotherapy for BL irrespective of HIV status. Accordingly, eight patients received an intensive regimen (CODOX-M/IVAC) between 1996 and 2000, and six patients were treated with nonintensive regimens, primarily before 1996. Although the overall CR (63% vs. 67%) and 2-year EFS rates (57% vs. 60%) were similar for the CODOX-M/IVAC versus other chemotherapy groups, the CODOX-M/IVAC group was composed disproportionately of high-risk patients (88% vs. 33% in the CODOX-M/IVAC vs. other treatment group). Consequently, CODOX-M/IVAC overcame features associated with high risk, leading to remarkably high 2-year EFS rates compared with other regimens: Stage IV disease (66% vs. 13%), BM involvement (67% vs. 0%), CNS disease (67% vs. 0%), and elevated LDH levels (68% vs. 14%).
The clinical outcome of CODOX-M/IVAC treatment in HIV-positive patients with BL was compared with the clinical outcome of 24 concomitantly treated HIV-negative patients with BL. Among all 38 patients independent of treatment regimen (14 HIV-positive, 24 HIV-negative patients), HIV status did not adversely affect the long-term EFS (P = 0.88). Overall CR (HIV-positive patients, 64%; HIV-negative patients, 63%) and 2-year EFS rates (HIV-positive patients, 59%; HIV-negative patients, 62%) were comparable after all chemotherapy. When EFS was evaluated by chemotherapy independent of HIV status, CODOX-M/IVAC was associated with improved EFS (P = 0.05) in all patients, particularly those with high-risk factors. In addition, the HIV-positive patients in the current study had similar outcomes to those of a recent multicenter Phase II trial, United Kingdom Lymphoma Group LY06.10 That multicenter trial evaluated CODOX-M and CODOX-M/IVAC in 52 adult patients with BL (33%; age range, 16–21 years) and reported a 2-year EFS of 64.6% with an overall survival rate of 72.8%. For high-risk patients, the 2-year EFS rate was 59.5% and the overall survival rate was 69.9%.10
Despite the increased toxicity associated with CODOX-M/IVAC, HIV-positive patients treated with this regimen tolerated chemotherapy almost as well as HIV-negative patients. The high rates of therapy-related myelosuppression and infectious complications after CODOX-M/IVAC in HIV-positive patients were similar to the toxicities seen in HIV-negative patients treated with CODOX-M/IVAC, both in the current study and in the larger Phase II study of 52 adult HIV-negative patients.10 Among eight HIV-positive patients treated with CODOX-M/IVAC, one death was attributable to toxicity. However, one death due to toxicity also occurred among the 14 HIV-negative patients treated with CODOX-M/IVAC. In contrast to previous studies, outcomes for patients with HIV infection were not affected adversely by increased infections during intensive (CODOX-M/IVAC) chemotherapy.
Although the small number of patients and retrospective nature of the current study are undeniable limitations, the major conclusion is the similarity in outcomes and toxicities between HIV-positive and HIV-negative patients with BL treated with intensive chemotherapy. In the absence of a larger trial, the current data also suggest a possible benefit of CODOX-M/IVAC in all patients with BL. Multivariate analysis of the results in all patients with BL adjusted for HIV and stage effects favor CODOX-M/IVAC (P = 0.012). The hazard ratio (CODOX-M/IVAC vs. other chemotherapy) is 0.237, with a 95% confidence interval of 0.077–0.731.
The treatment of HIV-negative patients with BL recently has evolved. In the past, because of the highly aggressive biology of this tumor, CHOP or equivalent regimens were associated with rapidly fatal disease progression. In the early 1990s, distinct chemotherapy protocols for BL, including high-dose cyclophosphamide, antimetabolites, and prophylactic intrathecal chemotherapy, cured 50–60% of Stage I–III patients. However, patients with more-advanced-stage disease continued to fare poorly, with a 20% durable response rate overall.6, 28–31 In 1996, Magrath et al.9 reported the results of an intensive chemotherapy regimen, CODOX-M/IVAC, in 41 HIV-negative patients (20 adults and 21 children) with small noncleaved lymphoma. Ninety-five percent (n = 39) of patients achieved CR, with a 2-year EFS rate of 92%. Outcomes in Stage IV patients (7 children and 3 adults) with either BM and/or CNS involvement were particularly impressive, with a 2-year EFS rate of 80%. However, the inclusion of pediatric oncology patients and the exclusion of HIV-positive patients may have influenced the overall clinical outcomes in that study.9 As noted above, the United Kingdom Lymphoma Group LY06 study had inferior results,10 although 33% of the patients were 16–21 years of age, but those results were more consistent with the results reported in the current study.
BL' accounts for 35–40% of NHL cases in HIV-infected individuals.11–13 Despite the highly aggressive nature of this disease, HIV-positive patients with BL have traditionally been treated with reduced-dose chemotherapy regimens with resultant poor clinical outcomes. Much of the reluctance to pursue intensive regimens stems from early studies of standard therapy applied to AIDS-related NHL. These trials were associated with unfavorable outcomes due to increased toxicities and infections.16 Before HAART, the long-term disease-free survival rates for patients with AIDS-related lymphoma ranged from 30% to 50%.
Over the last several years, with the advent of HAART, standard-dose chemotherapy for HIV-associated diffuse large B-cell lymphoma is rapidly becoming the standard of care. A recent study by the AIDS Malignancy Consortium demonstrated that patients with AIDS-related lymphomas can be treated safely with modified or full-dose standard CHOP chemotherapy while receiving HAART.32 Another study has suggested that concomitant HAART and standard-dose CHOP chemotherapy may actually improve CR and overall survival rates in patients with AIDS-related lymphomas by reducing the incidence and morbidity of opportunistic infections.33
In the pre-HAART era, a few Phase II studies of aggressive chemotherapy regimens included HIV-positive patients with BL. At MSKCC, 3 of 9 HIV-positive patients (30%) with BL were cured with an aggressive ALL regimen.24 At the University of Texas M. D. Anderson Cancer Center (Houston, TX), 44 adults with BL (12 HIV-positive patients) were treated on protocol with 3 sequential chemotherapy combinations and intrathecal prophylaxis. Although 83% of the HIV-positive patients achieved disease-free status (identical to HIV-negative patients with BL), the overall 5-year survival for HIV-positive patients with BL was only 36%.29 In 1993, the French–Italian Cooperative Group treated 141 patients with HIV-associated NHL with LNH-84, an intensive regimen of doxorubicin, cyclophosphamide, vindesine, bleomycin, and prednisolone followed by high-dose methotrexate, ifosfamide, etoposide, L-asparaginase, and cytarabine. Among the 59 patients with BL, the CR rate was 67%, with a 2-year survival rate of 40%. Analyzed separately, HIV-positive patients with BL and with good prognostic factors (defined as CD4 counts > 100 cells per microliter, good performance status, and no previous AIDS-defining illnesses) achieved a 2-year overall survival rate of 50%. Most reported deaths resulted from recurrent/refractory disease and not from treatment-related toxicity.34, 35
Unlike other AIDS-related malignancies, BL usually occurs early in the course of HIV disease and often presents as the initial AIDS-defining illness in the setting of higher CD4 counts. In the curent study, BL was the AIDS-defining disease in one-half of the study population. Total CD4-positive cell counts, which were available for 9 of the 14 HIV-positive patients at the time of BL diagnosis, ranged from 10 to 555 cells per microliter, with a mean of 185 ± 65 cells per microliter and a median of 149 cells per microliter. A previous study reported the median CD4 counts of HIV-positive patients at cancer diagnosis to be 12 cells per microliter for cerebral lymphoma, 41 cells per microliter for Kaposi' sarcoma, and 120 cells per microliter for BL.36 In another series, the median CD4 count of HIV-positive patients with BL was 227.9 cells per microliter, compared with 95.3 cells per microliter for HIV-positive patients with immunoblastic lymphoma.37 The relatively good immune function possessed by patients with HIV-associated BL (independent of HAART) further suggests that this subset of patients may benefit from aggressive chemotherapy regimens that have acceptable toxicities.6, 27, 28, 31
Cortes et al.38 recently reported a retrospective study of 13 HIV-positive adult patients with BL who received intensive chemotherapy with hyperfractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone (hyper CVAD) alternating with high dose methatrexate and Ara-C from 1995 to 2000. Although they demonstrated an overall response rate of 92%, the median survival period was only 12 months. In comparison to the patients in the current study, only one-third had Stage IV disease, but the initial average CD4 count (77 cells per microliter) was lower than historically reported and lower than in the current cohort (210 cells per microliter). Six of the 7 HIV-positive patients with BL who received HAART during chemotherapy were alive after a median period of 29 months, whereas 4 patients who did not receive HAART during the same chemotherapy died.38
In summary, we have demonstrated that intensive chemotherapy with CODOX-M/IVAC can be administered safely to patients with HIV infection, that this regimen may abrogate poor prognostic features, and that the results are comparable with those observed in HIV-negative populations treated concomitantly. It should be emphasized that the current analysis was retrospective, nonrandomized, and included a relatively small number of patients. Therefore, it was not possible to determine conclusively whether one regimen (i.e., CODOX-M/IVAC) was superior to other regimens for HIV-positive patients with BL. Furthermore, these patients were treated over a long period of time, during which methods of supportive care (e.g., growth factor support, prophylaxis for opportunistic infections, HAART) changed significantly. Given the small number of patients analyzed, we were unable to assess the impact of these recent interventions separately. Although it is possible that the simple addition of HAART for HIV-positive patients accounts for the improved survival, HIV-positive patients with BL often have relatively preserved immune status with higher CD4 counts at diagnosis than do other HIV-positive patients with NHL,36 making it unlikely that HAART accounts for the entire difference. Larger prospective studies, particularly in HIV-positive patients with BL who have poor prognostic features, are warranted.
- 7Low-dose compared with standard-dose m-BACOD chemotherapy for non-Hodgkin's lymphoma associated with human immunodeficiency virus infection. National Institute of Allergy and Infectious Diseases AIDS Clinical Trials Group. N Engl J Med. 1997; 336: 1641–1648., , , et al.
- 20Hypothesis testing: categorical data. In: RosnerB. Fundamentals of biostatistics. Belmont, CA: Wadsworth Publishing Co., 1994: 345–442..
- 26An economic analysis of the NHL-15 protocol for diffuse aggressive lymphomas: the potential for substantially reducing cost by improving outcomes [abstract]. Am Soc Hematol. 1997; 90: 249a..
- 33Influence of highly active anti-retroviral therapy on response to treatment and survival in patients with acquired immunodeficiency syndrome-related non-Hodgkin's lymphoma treated with cyclophosphamide, hydroxydoxorubicin, vincristine and prednisone. Br J Haematol. 2001; 112: 909–915., , , et al.
- 38Hyperfractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone and highly active antiretroviral therapy for patients with acquired immunodeficiency syndrome-related Burkitt lymphoma/leukemia. Cancer. 2002; 94: 1492–1499., , , et al.