Purpose: We report the results of a prospective trial in which patients with relapsing non-Hodgkin's lymphomas were sequentially treated with two regimens (mesna, ifosfamide, mitoxantrone, and etoposide [MINE], and etoposide, methylprednisolone, cytarabine, and cisplatin [ESHAP]) if they had no history of disease resistance to these drugs. Patients and Methods: Ninety-two patients received MINE (mesna 4 g/m(2), ifosfamide 4 g/m(2), mitoxantrone 8 mg/m(2), and etoposide 195 mg/m(2)) for a maximum of six courses followed by ESHAP (etoposide 240 mg/m(2), methylprednisone 500 mg/d, high-dose cytarabine 2 g/m(2), and cisplatin 100 mg/m(2)) for three courses to consolidate complete response (CR) or for a maximum of six cycles after a partial response (PR) or no response to MINE. pretreatment serum levels of lactate dehydrogenase (LDH) and beta(2)-microglobulin (beta(2)M) were documented in 80 of 92 patients. Results: the response rate to MINE-ESHAP was 69% (48% CRs and 21% PRs), with a median survival time of 24 months and median time to treatment failure of 12 months. The median time to treatment failure according to histology was as follows: low-grade histologies, 16 months; low-grade transformed to intermediate-grade, 8 months; and intermediate-grade, 5 months, The most serious complication was myelosuppression, which resulted in two deaths due to neutropenic sepsis. A risk factor model based on beta(2)M and LDH levels before salvage treatment showed three categories of risk, with 36-month survival rates os follows: low (beta(2)M < 3 mg/dL and LDH normal), 61%; intermediate (beta(2)M greater than or equal to 3 mg/dL or LDH above normal), 23%; and high (beta(2)M greater than or equal to 3 mg/dL and LDH above normal), 0%. Conclusion: MINE-ESHAP is an effective salvage strategy for patients with recurrent lymphoma. Toxicity was acceptable. Factors that determine prognostic categories at relapse merit further study.