Arsenic trioxide (ATO) is currently regarded as the best treatment option in relapsed acute promyelocytic leukemia (APL). The long-term outcome of salvage therapy using an ATO-based approach compared with chemotherapy based regimens is not well established. We analyze the clinical outcome of 151 APL patients relapsing after front-line therapy with ATRA and anthracycline, who received second-line therapy with chemotherapy or ATO-based regimens.


From June 1997 to May 2013, 151 patients (94 M/57 F; median age: 42 years, 2-81) relapsed after front-line therapy with PETHEMA trials. Patients presented with either molecular relapse (n=47) or hematological relapse (n=104). Sixty-seven patients (44%) followed salvage therapy with chemotherapy-based regimens (chemotherapy group) consisting of induction with mitoxantrone plus cytarabine plus ATRA (n=45), EMA (n=7), or other regimens (n=15). Patients not eligible for stem-cell transplantation (SCT) received consolidation with or without maintenance therapy. From October 2003, 84 patients (56%) received salvage therapy with ATO-based regimens (ATO group), comprising induction with ATO (0.15 mg/kg intravenously until CR, with ATRA [n=19] or chemotherapy [n=6]) followed by one consolidation cycle with ATO plus ATRA. If the post-consolidation bone marrow PCR was negative an autologous (auto)-SCT was recommended, when the PCR was still positive an allogeneic (allo)-SCT was planned. Patients not eligible for SCT received maintenance therapy with ATO plus ATRA with or without low-dose chemotherapy.


Baseline characteristics, including sex, relapse-risk at primary diagnosis, morphologic and BCR subtype, as well as age at relapse and type of relapse, were similar in both cohorts of patients. Although not significant, patients rescued in the chemotherapy group presented earlier relapses (<18 months after initial APL diagnosis) (55% vs. 43%, P=0.13). CR rates were 85% in the chemotherapy group (8 deaths and 2 resistances) and 92% in the ATO group (4 deaths and 3 resistances) (P=0.11). In patients achieving second CR, a molecular CR was achieved after consolidation in 78% and 84%, respectively (P=0.38). Twenty-two patients in the chemotherapy group (39%) and 16 in the ATO group (21%) were not transplanted in second CR (P=0.04). The reasons for not SCT in the chemotherapy and in the ATO group were ineligibility for SCT (6 vs. 8 patients), early relapse before planned SCT (10 vs. 7 patients), and mobilization failure (6 vs. 1 patients). Overall, 34 patients underwent SCT in the chemotherapy group (auto-SCT, 20; allo-SCT, 14), and 57 underwent SCT in the ATO group (auto-SCT, 47; allo-SCT, 10). The median follow-up in the chemotherapy group was 95 months (range, 24-167), and 33 months (range, 3-100) in the ATO group. The 5-year overall survival (OS), disease-free (DFS), and relapse-free survival (RFS) in the chemotherapy group and in the ATO group were 40% vs. 56% (P=0.01), 31% vs. 39% (P=0.07), and 34% vs. 48% (P=0.09), respectively. For patients not receiving SCT because of mobilization failure or ineligibility, the 5-year OS, DFS, and RFS in the chemotherapy group and in the ATO group were 56% vs. 19% (P=0.11), 42% vs. 19% (P=0.49), and 42% vs. 29% (P=0.63), respectively. For patients receiving auto-SCT, the 5-year OS, DFS, and RFS in the chemotherapy group and in the ATO group were 55% vs. 80% (P=0.04), 40% vs. 54% (P=0.06), and 44% vs. 57% (P=0.13), respectively. For patients receiving allo-SCT, the 5-year OS, DFS, and RFS in the chemotherapy group (P=0.48), and 31% vs. 57% (P=0.34), respectively. All but one patient underwent auto-SCT with negative PCR (this patient relapsed rapidly after auto-SCT). Regarding allo-SCT, 7 patients were PCR+ and 17 were PCR negative before SCT (5-year DFS 0% vs. 41%, respectively, P=0.01).


This study performed in a large series with prolonged follow-up of APL patients relapsing after upfront therapy with ATRA and anthracycline shows high rates of CR either with ATO (92%) or chemotherapy regimens (85%). Salvage therapy with ATO-based regimen allowed performing more frequently auto-SCT with negative PCR, and this strategy resulted in an overall improvement of the 5-year OS, DFS, and RFS.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.