Hyper-CVAD therapy is superior to our previous regimens and should be compared with established regimens in adult ALL.

Hyper-CVAD is effective in adult B-ALL. Identification of patients with high risk for relapse and improved methods to detect residual disease may result in risk-oriented approaches.

Sixty-six adults with refractory acute lymphocytic leukemia received salvage therapy with the 'hyper-CVAD' regimen, consisting of eight courses of alternating intensive chemotherapy with growth factor support, followed by oral maintenance chemotherapy. Their outcome was compared with 63 prognostically similar historical control patients treated with high-dose Ara-C plus mitoxantrone with or without GM-CSF. Overall, the complete response rates were similar in the treatment and control groups (29 of 66 (44%) vs 24 of 63 (38%)). There were more patients in the current study with primary resistant disease (10 of 66 (15%) vs one of 63 (2%), P = 0.006), and conversely fewer patients with secondary resistance (19 of 66 (29%) vs 28 of 63 (44%), P = 0.06). Recovery of granulocyte counts above 500/l was significantly faster in the current study when compared to high-dose Ara-C-treated patients who were given GM-CSF (20 vs 25 days, P = 0.04). Survival was prolonged in the hyper-CVAD-treated patients, with most of the benefit seen in first salvage patients (42 vs 20 weeks, P = 0.016). When only first salvage patients were considered, there was a significant difference in disease-free survival in favor of hyper-CVAD (52 vs 20 weeks, P = 0.008). The hyper-CVAD regimen is a more effective and less toxic salvage regimen for relapsed acute lymphocytic leukemia than high-dose Ara-C-based regimens.

Central nervous system (CNS) involvement occurred in 45 of 222 (20.3%) leukemic adults achieving bone marrow (BM) complete remission (CR), including 12 of 23 (52%) acute undifferentiated leukemia (AUL), 12 of 32 (39%) lymphoma leukemia, 5 of 26 (19%) acute lymphoblastic leukemia, and 16 of 142 (11%) acute myelogenous leukemia. Risk factors for CNS disease were lactic dehydrogenase (LDH) greater than or equal to 25,000/mm3. AUL morphology, age less than 20 years, and extramedullary involvement were most significant. Pattern of CNS involvement varied with morphology. Survival after CNS relapse depended most on BM status and symptoms. Duration of CNS CR was longest for asymptomatic patients with low CSF cell counts. Also important were duration of first BM CR, ease of achievement of initial BM CR, and leukocyte count (original and at most closely antecedent BM involvement), reflecting the common origin of BM and CNS leukemic cells. Central nervous system relapse generally did not shorten BM CR or survival, although early primary CNS relapse was associated with early BM relapse.