Background: Loss of heterozygosity (LOH) of CEBPA has been sporadically reported in acute myeloid leukemia (AML), resulting in a complete or nearly complete homozygosity of CEBPA mutations ( CEBPA homo). However, its clinical significance has been underappreciated. Methods: We retrospectively analyzed the prognostic impact of CEBPA homo in 1223 adult AML patients, where CEBPA mutations were detected using whole exome sequencing (n=296) or targeted sequencing (n=927) technology. CEBPA homo is defined as the variant allele frequency (VAF) of a CEBPA mutation over 50%, and CEBPA heter is the opposite. Propensity score matching was used to reduce selection bias when comparing the survival difference between two groups. Results: CEBPA mutations were identified in 106 patients (8.7%), 47(44.3%) of which carried the C-terminal DNA-binding or basic leucine zipper region ( CEBPA bZIP). Compared to patients with other CEBPA mutations ( CEBPA others), patients carrying CEBPA bZIP were significantly younger (40.38±14.80 vs. 52.05±17.44 years, P<0.001) and had higher hemoglobin [89(65.5, 98) vs 70(58, 81) g/dL, P=0.001], lower platelet counts [35(17.5, 43) vs. 44(24, 78.5) 10 9/L, P=0.019) and higher proportion of peripheral blast cells [80(52.5, 90.5) vs. 51(14, 86.5) %, P=0.001] at diagnosis. CEBPA bZIP was associated with a higher complete remission rate (CR; 85% vs. 61%, P=0.012) in AML patients but did not reduce the relapse rate (32% vs. 39%, P=0.582). CEBPA bZIP significantly improved 4-year overall survival (OS; 0.70±0.09 vs. 0.30±0.13, P=0.013), 2-year event-free survival (EFS; 0.47±0.09 vs. 0.2±0.09, P=0.01), and 3-year pre-transplantation survival (PTS; 0.54±0.09 vs. 0.38±0.09, P=0.003) in AML patients compared to those with CEBPA others. These data are consistent with previous reports, suggesting the representative of this cohort. We further evaluated the clinical significance of CEBPA homo in this cohort and found that patients with CEBPA homo had higher white blood cell counts [29.37 (9.16, 76.11) vs. 10.37 (3.38, 43.43) 10 9/L, P=0.048] than those with CEBPA heter. Although CEBPA homo was not associated with CR rate (65% vs. 80%, P=0.145), it did reduce relapse rate (26% vs. 47%, P=0.045) in AML patients compared to those with CEBPA heter. CEBPA homo was significantly associated with improved 5-year OS (0.68±0.10 vs. 0.29±0.10, P=0.006), 5-year EFS (0.30±0.09 vs. 0.10±0.08, P=0.018), and 5-year PTS (0.40±0.10 vs. 0.14±0.07, P=0.046) compared to CEBPA heter. There is no difference in CEBPA VAF between patients carrying CEBPA bZIP and CEBPA other, suggesting similar distribution of CEBPA homo between these two groups. Conclusions: The clinical significance of CEBPA homo is different from CEBPA bZIP, which suggests a distinct biological meaning. Our study would add one new layer to the current stratification of CEBPA-mutated AML based on the prognosis. More cases are needed to further examine the clinical significance of the concurrency of CEBPA homo and CEBPA bZIP.