Background: Venetoclax combined with hypomethylating agents is a new standard of care for newly diagnosed patients with acute myeloid leukaemia (AML) who are 75 years or older, or unfit for intensive chemotherapy. Pharmacodynamic studies have suggested superiority of the longer 10-day regimen of decitabine that has shown promising results in patients with high-risk AML in phase 2 trials. We hypothesised that venetoclax with 10-day decitabine could have improved activity in patients with newly diagnosed AML and those with relapsed or refractory AML, particularly in high-risk subgroups. Methods: This single centre, phase 2 trial was done at the University of Texas MD Anderson Cancer Center (Houston, TX, USA). The study enrolled older patients (aged >60 years) with newly diagnosed AML, not eligible for intensive chemotherapy; secondary AML (progressed after myelodysplastic syndrome or chronic myelomonocytic leukaemia); and relapsed or refractory AML. Patients were required to have an Eastern Cooperative Oncology Group (ECOG) performance status of 3 or less, white blood cell count less than 10 × 109 per L, and adequate end-organ function. Patients with favourable-risk cytogenetics (eg, t[15;17] or core-binding factor AML) or who had received previous BCL2-inhibitor therapy were excluded. Patients received decitabine 20 mg/m2 intravenously for 10 days with oral venetoclax 400 mg daily for induction, followed by decitabine for 5 days with daily venetoclax for consolidation. The primary endpoint was overall response rate. The secondary endpoints analysed within this report include safety, overall survival, and duration of response, in keeping with recommendations of European LeukemiaNet 2017 guidelines. All patients who received at least one dose of treatment were eligible for safety and response assessments. The trial was registered on ClinicalTrials.gov (NCT03404193) and continues to accrue patients. Findings: Between Jan 19, 2018, and Dec 16, 2019, we enrolled 168 patients; 70 (42%) had newly diagnosed AML, 15 (9%) had untreated secondary AML, 28 (17%) had treated secondary AML, and 55 (33%) had relapsed or refractory AML. The median age was 71 years (IQR 65–76) and 30% of patients had ECOG performance status of 2 or higher. The median follow-up for all patients was 16 months (95% CI 12–18; actual follow-up 6·5 months; IQR 3·4–12·4). The overall response rate was 74% (125 of 168 patients; 95% CI 67–80) and in disease subgroups were: 89% in newly diagnosed AML (62 of 70 patients; 79–94), 80% in untreated secondary AML (12 of 15 patients; 55–93), 61% in treated secondary AML (17 of 28 patients; 42–76), and 62% in relapsed or refractory AML (34 of 55 patients; 49–74). The most common treatment-emergent adverse events included infections with grades 3 or 4 neutropenia (n=79, 47%) and febrile neutropenia (n=49, 29%). 139 (83%) of 168 patients had serious adverse events, most frequently neutropenic fever (n=63, 38%), followed by pneumonia (n=17, 10%) and sepsis (n=16, 10%). The 30-day mortality for all patients was 3·6% (n=6, 95% CI 1·7–7·8). The median overall survival was 18·1 months (95% CI 10·0–not reached) in newly diagnosed AML, 7·8 months (2·9–10·7) in untreated secondary AML, 6·0 months (3·4–13·7) in treated secondary AML, and 7·8 months (5·4–13·3) relapsed or refractory AML. The median duration of response was not reached (95% CI 9·0–not reached) in newly diagnosed AML, 5·1 months (95% CI 0·9–not reached) in untreated secondary AML, not reached (95% CI 2·5–not reached) in previously treated secondary AML, and 16·8 months (95% CI 6·6–not reached) in relapsed or refractory AML. Interpretation: Venetoclax with 10-day decitabine has a manageable safety profile and showed high activity in newly diagnosed AML and molecularly defined subsets of relapsed or refractory AML. Future larger and randomised studies are needed to clarify activity in high-risk subsets. Funding: US National Institutes of Health and National Cancer Institute.