Gemcitabine added to doxorubicin, bleomycin, and vinblastine for the treatment of de novo Hodgkin disease

Unacceptable acute pulmonary toxicity




Gemcitabine is an effective treatment for recurrent Hodgkin disease (HD), with relatively minimal associated toxicity. The authors conducted a trial substituting this drug for dacarbazine in the standard regimen to form ABVG (doxorubicin, bleomycin, vinblastine, gemcitabine) for patients with newly diagnosed, high-risk HD.


Twelve patients (median age, 34 years) with advanced-stage de novo HD were enrolled. Standard doses of doxorubicin, bleomycin, and vinblastine were given for six cycles. Cohorts of three patients were enrolled and the dose of gemcitabine was escalated to identify the maximally tolerated dose in this combination.


The maximally tolerated dose of gemcitabine was determined to be 800 mg/m2 in this combination. Five patients developed clinically significant pulmonary toxicity. Three required hospitalization during the final two cycles of treatment. Pneumonitis could not be predicted with serial diffusion capacity for carbon monoxide (DECO) evaluations, and reversed after discontinuation of bleomycin in three patients and steroid therapy in two patients. All 12 patients are alive to date, and 4 patients have experienced disease progression.


The bleomycin/gemcitabine combination should not be pursued for de novo HD due to significant pulmonary toxicity. Cancer 2003;98:978–82. © 2003 American Cancer Society.

DOI 10.1002/cncr.11582

For the past decade, the combination of doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) has been the standard treatment regimen for patients with advanced-stage Hodgkin disease in the United States.1 Long-term follow-up of a Cancer and Leukemia Group B (CALGB) study randomizing patients among mechlorethamine, vincristine, procarbizin, and prednisone (MOPP), ABVD, and MOPP-ABVD indicates improved failure-free survival with ABVD-containing regimens compared with MOPP. However, only approximately 60% of patients are cured, and there was no unique benefit in overall survival associated with ABVD.2, 3 High-risk subgroups of patients with advanced-stage Hodgkin disease with cure rates less than 50% can now be identified using clinical criteria.4 Clearly, improvements in therapy are needed, particularly for this group of patients.

Dose-intensive regimens, including mechlorethamine, bleomycin, doxorubicin, vincristine, vinblastine, etoposide, and prednisone (Stanford V),5 and bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine and prednisone (BEACOPP),6 and consolidation with high-dose chemotherapy and autologous stem cell transplantation,7, 8 have high disease-free survival rates in Phase II studies. Long-term follow-up of ongoing randomized trials with these approaches will be necessary because second malignancies and other late toxicities have emerged as important competing causes of death in patients cured of Hodgkin disease, particularly in patients treated with alkylating agents or radiotherapy.9

Gemcitabine (Gemzar; Eli Lilly and Company, Indianapolis, IN) is a deoxycitidine analog, which requires intracellular phosphorylation to inhibit DNA synthesis.10 Several studies have suggested single-agent activity in patients with recurrent or refractory Hodgkin disease.11, 12 A multicenter trial of 23 patients with recurrent or refractory disease demonstrated an overall response rate of 39%, with minimal nonhematopoietic toxicity.13 Given the limited single-agent activity of dacarbazine,14 we performed a clinical trial substituting gemcitabine for dacarbazine in the standard ABVD regimen for patients with de novo Hodgkin disease (ABVG).


The protocol was approved by the Human Protection Committee of Dana-Farber/Partners Cancer Care, and written informed consent was obtained from all participants before enrollment.


Eligible patients had histologically confirmed Hodgkin disease, either bulky Stage II (as defined by a mediastinal mass larger than one-third the maximal diameter of the chest) or advanced-stage disease defined by the Cotswold revision of the Ann Arbor staging criteria.15 Patients were excluded if they received any previous treatment for Hodgkin disease, had evidence of left ventricular dysfunction, or severe pulmonary function abnormalities, including FEV1 less than 50% of predicted on mandatory baseline pulmonary function testing.


ABVG was administered as a combination chemotherapy regimen (like ABVD) on Days 1 and 15 for six full cycles. The standard doses of doxorubicin (25 mg/m2), bleomycin (10 units/m2), and vinblastine (6 mg/m2) were fixed. Cohorts of three patients were enrolled to define the dose-limiting toxicity (DLT) of gemcitabine. The first cohort received gemcitabine 800 mg/m2, the next cohort received 1000 mg/m2, and the final cohort received 1200 mg/m2. Gemcitabine was infused during a period of 30–60 minutes. Once the DLT was defined, an additional 10 patients were to be enrolled at the maximum tolerated dose. DLTs were defined within the first 2 cycles (4 courses) of therapy only, and included need to delay therapy for at least 1 week, due to absolute neutrophil count (ANC) less than 250 neutrophils per microliter or a platelet count less than 75,000 platelets on the planned day of treatment; Grade 4 elevation of bilirubin; Grade 4 elevation of creatinine (per the NCI Common Toxicity Criteria, version II); or death within the first 2 cycles of treatment. The use of hematopoietic growth factors, including granulocyte–colony-stimulating factor (G-CSF), was not permitted during the first two cycles of therapy.

Statistical Methods

A Phase I design that identified DLT only during the first two cycles of therapy was selected to allow timely progression to the next cohort. Patients then were observed carefully for any cumulative toxicity. Association between categoric variables was assessed by the Fisher exact test. If one variable was continuous, the Wilcoxon rank sum test was used. Exact binomial confidence intervals (CIs) are provided. Progression-free survival was estimated with the Kaplan–Meier method. CIs were calculated using the Greenwood formula.


Patients and Therapy

Twelve patients (including 7 females) with a median age of 34 years (range, 24–60 years) were enrolled. Four patients had bulky Stage II disease, and six patients had B symptoms. All patients had nodular sclerosis histology. Using the international scoring system for advanced Hodgkin disease,4 four patients had more than three risk factors. The first patient in the second dose cohort had the DLT of a low ANC on Day 15 of Cycle 1. Therefore, the maximum tolerated dose of gemcitabine was established as 800 mg/m2 on Days 1 and 15 of each cycle of ABVG.


Infusional toxicity was minimal, with only a single episode of severe (Grade III) nausea. The ABVG infusion time was shorter than the infusion time for standard ABVD. Four patients developed catheter-related thrombosis and were treated with anticoagulation. There was no evidence of embolism, and thromboses resolved in all cases. A single episode of transaminase elevation (Grade III) occurred in one patient that was self-limited. Three patients required G-CSF support during Cycles 3–6.

Five patients (only one of whom smoked tobacco and only one of whom required G-CSF) developed clinically significant pulmonary toxicity: one patient in Cycle 4, two patients in Cycle 5, one patient in Cycle 6, and one patient immediately after completion of Cycle 6 (Table 1). Two of these patients had diffuse radiographic pneumonitis that required hospitalization, steroid therapy, and oxygen. Two patients had cough and dyspnea on exertion with declining diffusion capacity for carbon monoxide (DLCO). One additional patient required hospitalization for a lobar infiltrate and associated hypoxia. No patient underwent lung biopsy, but sputum cultures for both routine and opportunistic infections were negative. Therapy included corticosteroids and oxygen in two patients and discontinuation of bleomycin in three patients. Radiographic studies included computed tomographic scans demonstrating characteristic ‘ground-glass’ opacities in two patients, a lobar infiltrate on chest X-ray in one patient, and subtle, diffuse plain radiographic changes in two patients. Symptoms resolved completely with conservative therapy and no patient has required long-term supplemental oxygen.

Table 1. Pneumonitis (n = 5) during Therapy with Doxorubicin, Bleomycin, Vinblastine, and Gemcitabineea
Age (yrs)GenderSmoker?DLCO baseline (%)bDLCO Midtherapy (%)bCycle no.Clinical courseOutcome
  • DLCO: declining diffusion capacity for carbon monoxide; PD: progressive disease; CR: complete response; ASCT: autologous stem cell transplant; XRT: external beam radiation therapy; CXR: chest x-ray; abx: antibiotics.

  • a

    All patients recovered pulmonary function within 2 months of pneumonitis.

  • b

    DLCO values were obtained at baseline and after Cycle 3.

  • c

    Patient subsequently received two cycles of chlorambucil, vinblastine, procarbazine, and predinsone, followed by ASCT and XRT. Patient currently is in CR.

  • d

    Completed therapy with two cycles of chlorambucil, vinblastine, procarbazine, and prednisone.

50MaleNo99885Hospitalized for hypoxia; received steroids and oxygen; bleomycin not given in Cycles 5 and 6PDc
38FemaleNo96946Cough and subjective dyspnea; bleomycin held in Cycle 6.CR
28FemaleNo104986Hospitalized for hypoxia; bleomycin held in Cycle 6B; received steroids and oxygenCR
58MaleYes53495Cough and dyspnea; bleomycin held in Cycles 5 and 6.CR
26FemaleNo84774Hospitalized for hypoxia; Infiltrate on CXR; received abx; taken off studydCR

DLCO was performed in all patients at baseline, after Cycle 3, and at the end of therapy. At least 1 value decreased by 15% or more in 7 patients. There was no significant association between the rate of decrease of DLCO and the subsequent development of clinically significant pulmonary complications (P = 0.93). There also was no association between a history of tobacco use and development of pneumonitis (P = 0.58). Because of this unexpected high rate of pulmonary toxicity, the trial was closed early. The final two patients received dacarbazine instead of gemcitabine during the last cycle of therapy.

Response and Outcome

All patients responded to therapy. Four patients, including two of the patients who experienced pneumonitis, received planned mantle consolidation radiotherapy to bulky mediastinal masses. One patient (without earlier pneumonitis) experienced a brief dyspnea reaction during radiation attributed to infection, which resolved with antibiotic therapy.

With a median follow-up of approximately 1 year, 5 patients experienced disease progression. All are alive and have received high-dose chemotherapy and autologous stem cell transplantation, including one patient who experienced pneumonitis during ABVG therapy. Progression-free survival at 1 year was 64% (90% CI = 40–87%).


The 1-year progression-free survival in the current study was 64%, somewhat lower than might be expected with ABVD combination therapy. However, the low number of patients in the current series precludes a definitive efficacy evaluation of ABVG. Additional study is required to determine the importance of dacarbazine in the ABVD combination. Both Stanford V5 and BEACOPP6 are active regimens for Hodgkin disease that do not include dacarbazine.

The trial was stopped prematurely due to an unexpectedly high incidence and severity of pneumonitis. Pulmonary toxicity associated with single-agent gemcitabine usually is related to the infusion and is self-limited. Severe pulmonary toxicity is rare and may be the result of a capillary leak phenomenon as observed with cytarabine,16 particularly in the setting of repeated administrations of gemcitabine.17 In one series of five patients, prompt initiation of corticosteroid therapy and discontinuation of the drug reversed symptomatic dyspnea and infiltrates.18

Previously identified risk factors for gemcitabine-induced pulmonary toxicity include previous chest irradiation and primary lung carcinoma.19–21 A recent series also has demonstrated the potential for radiation-recall reactions from gemcitabine administered to patients with solid tumors.22 This could have important implications in the future use of gemcitabine for Hodgkin disease as part of a combined modality approach.

The German Hodgkin study group recently conducted a trial substituting gemcitabine for etoposide in BEACOPP (to form BAGCOPP) for patients with de novo Hodgkin disease. In preliminary analysis, of 27 patients enrolled, 8 developed acute pulmonary toxicity, including 6 cases of pneumonitis not predicted by serial measurements of DLCO. The study was closed prematurely for safety concerns.23

Bleomycin has a known incidence of both severe acute24 and chronic25 pulmonary toxicity.1 In a CALGB trial comparing ABVD with MOPP/ABV, there was a significantly increased rate of pulmonary toxicity in patients older than age 40 years and in those who received previous localized radiation. The overall rate of acute pulmonary toxicity with ABVD was 25%, despite interval monitoring of pulmonary function.26 Similarly, in a randomized trial at Memorial Sloan-Kettering Cancer Center that compared 6 cycles of ABVD with 4 cycles of ABVD and radiation for patients with early-stage Hodgkin disease, bleomycin was discontinued in 22% of patients for decreasing DLCO, which was usually asymptomatic.27 Both in the BAGCOPP trial and in the current trial, there was an increased incidence and severity of pulmonary toxicity than associated with nongemcitabine-containing regimens, suggesting a possible interaction between gemcitabine and bleomycin. In addition, the significant pneumonitis could not be attributed to infection or predicted with serial monitoring of physical examination or DLCO. Future studies of chemotherapy treatment strategies for Hodgkin disease should avoid repeated administrations of the gemcitabine/bleomycin combination.