All-trans retinoic acid (ATRA), alone or combined with chemotherapy (CHT) is widely used to induce complete remission (CR) in newly diagnosed acute promyelocytic leukemia (APL). If used alone, ATRA results in a substantial proportion of CRs. To maintain remission further, ATRA is commonly used with cycles of CHT, frequently followed by autologous (auto) or allogeneic (allo) stem cell transplantation (SCT), as early reports have shown that the continuous administration of ATRA as single therapy almost invariably leads to relapse in a short period of time (months). Pharmacokinetic studies have shown that induced resistance to ATRA is frequently suppressed by the intermittent use of the drug. In this study we applied an intermittent therapeutic protocol with ATRA in five APL patients who were either molecularly refractory after combined ATRA/CHT treatment, or relapsed, or at diagnosis, but not eligible for the combination treatment because of previous toxicity. They were treated with ATRA (45 mg/m2/day) for 21 days. The treatment was then prolonged continuously for 1 week every 2 weeks. Molecular analysis was performed by qualitative and quantitative reverse transcription-polymerase chain reaction (RT-PCR). All patients obtained molecular remission, as assessed by qualitative RT-PCR, in a median of 3 months (range 1-15). Quantitative RT-PCR confirmed these data, showing a progressive reduction (1 or 2 logs) to a 'negligible quantity' of PML-RARalpha fusion transcript (ratio PML-RARalpha/ABL x 10(4) ABL < 10(-1)) in all but one patient treated with pulsed ATRA therapy. These data were confirmed with qualitative and quantitative RT-PCR. After a median follow-up of 17 months from the start of ATRA therapy, 4/5 patients (80%) are in continuous complete molecular remission. To our knowledge, this is the first clinical observation that intermittent ATRA therapy (without chemotherapy) is effective not only in inducing but also in maintaining long-term molecular remission in APL patients. This approach could therefore be effective, if confirmed in larger series, in relapsed/refractory patients unsuitable for high-dose therapy and SCT; it may be proposed as induction therapy for selected older APL patients if considered not to be eligible for combined ATRA/CHT due to inadequate performance status or concurrent disease.