We have retrospectively evaluated a cohort of 144 patients (including 17 pediatric ones) with de novo acute promyelocytic leukemia registered in databases of institutions cooperating within the CELL group (The Czech Leukemia Study Group for Life). The patients were diagnosed according to WHO criteria from 1989 until 2006. The aim was to check how well fared the patients, the majority of whom was not included into clinical trials, in real life. Of 140 evaluable patients, 97 (69.3%) attained complete remission (CR). The projected overall survival (OS) 4 years after diagnosis was 58.9%, and 55.3% at 6 years. In 8 patients (6.0%), no antileukemic therapy at all was given (either they died shortly after admission to the ward or therapy was not feasible due to their clinical status). Of 125 patients with documented commencement of some kind of therapy, 96 (76.8%) achieved CR. Of 102 patients with induction treatment with a combination of anthracycline and tretinoin (ATRA), 84 individuals (82.4%) attained CR (typically, this cohort might have been subjected to clinical trials). This result was better than that of patients treated by chemotherapy only (n = 15; CR 46.7%; P = 0.003) or by ATRA monotherapy (n = 13; CR 62.5%; P = 0.17). Another parameter with a significant impact on attaining CR was the leukocyte (WBC) count at diagnosis: its median values in patients achieving and not achieving CR were 2.1 and 24.0 x 10(9)/l, respectively (P < 0.0001). The WBC counts affected OS as well (P = 0.0001). However, when only patients after attaining CR were evaluated, the initial WBC counts no longer affected OS (P = 0.18). Achieving CR was also influenced by the performance status (PS) 0-1 (P = 0.005), which was in turn closely correlated to WBC counts (P = 0.0006). Additional factors (most likely connected with leukocytosis) influenced attaining CR with borderline statistical significance: e.g. FAB M3v morphology, LDH serum level, fibrinogen level, presence of internal tandem duplication (ITD) of the FLT3 gene (which was strongly associated with leukocytosis and also with the short PML/RARalpha transcript resulting from the bcr3 break in the PML gene). It may be speculated that FLT3-ITD is just one of the possible factors that lead to leukocytosis. The platelet counts at diagnosis had no impact on entering CR. Thus, we have not validated the current PETHEMA risk stratification in distinguishing intertermediate and low risk patients. Our study points to a significant difference of the results obtained in real life and of the results that could be achieved in patients who were fit to enter clinical trials. Among the prognostic factors, the most important one was the WBC count, the PS (which is highly affected by the WBC count), and feasibility of administration of the most potent induction therapy with anthracyclines and ATRA.