Despite widespread use of decitabine to treat acute myeloid leukaemia (AML), data on its effectiveness and safety in the real-world setting are scanty. Thus, to analyze the performance of decitabine in clinical practice, we pooled together patient-level data of three multicentric observational studies conducted since 2013 throughout Italy, including 306 elderly AML patients (median age 75 years), unfit for intensive chemotherapy, treated with first-line decitabine therapy at the registered schedule of 20 mg/m2/iv daily for 5 days every 4 weeks. Overall response rate (ORR), overall survival (OS) curves, and multivariate hazard ratios (HRs) of all-cause mortality were computed. Overall, 1940 cycles of therapy were administered (median, 5 cycles/patient). A total of 148 subjects were responders and, therefore, ORR was 48.4%. Seventy-one patients (23.2%) had complete remission, 32 (10.5%) had partial remission, and 45 (14.7%) had haematologic improvement. Median OS was 10 months for patients with favourable-intermediate cytogenetic risk and 6 months for those with adverse cytogenetic risk. Median relapse-free survival after CR was 10.9 months (95% confidence interval [CI]: 8.7-16.0). In multivariate analysis, mortality was higher in patients with adverse cytogenetic risk (HR=1.58; 95% CI: 1.13-2.21) and increased continuously with white blood cell (WBC) count (HR=1.12; 95% CI: 1.06-1.18). A total of 183 infectious adverse events occurred in 136 patients mainly (>90%) within the first five cycles of therapy. This pooled analysis of clinical care studies confirmed, outside of clinical trials, the effectiveness of decitabine as first-line therapy for AML in elderly patients unfit for intensive chemotherapy. An adverse cytogenetic profile and a higher WBC count at diagnosis were, in this real life setting, unfavourable predictors of survival. © 2019 John Wiley & Sons, Ltd.