The efficacy and tolerance of azacitidine in higher-risk myelodysplasia with hypocellular bone marrow (BM) are unknown. This post hoc AZA-001 trial analysis assessed whether baseline BM cellularity affected the overall survival (OS) advantage demonstrated with azacitidine versus conventional care regimens (CCR). Baseline BM biopsies of <30% cellularity were considered hypocellular with data evaluable from 299 patients (azacitidine n = 154, CCR n = 145); 13% (n = 39) hypocellular, 87% (n = 260) non-hypocellular. Patient characteristics were balanced between cellularity and treatment groups. Most patients (90-100%) had 2-3 cytopenias at baseline. Median (range) azacitidine treatment cycle lengths were 35·5 (28-54) and 33·0 (15-75) d in hypocellular and non-hypocellular groups, respectively. At 33 months, median OS was not reached (NR) [95% confidence interval (CI): 19·2, NR] in hypocellular patients receiving azacitidine versus 16·9 months (95% CI: 11·1, 19·3) with CCR (P = 0·001); and in non-hypocellular patients, it was 21·1 months (95% CI: 16·2, 34·7) versus 15·3 months (95% CI: 9·3, 17·6) (P = 0·012). Azacitidine tolerance was similar regardless of cellularity. Grade 3-4 thrombocytopenia and neutropenia occurred similarly in hypocellular patients treated with azacitidine versus CCR (80% vs. 92% and 88% vs. 75%). Azacitidine OS results are consistent with those from AZA-001, regardless of cellularity, and demonstrate its safety and efficacy in higher-risk myelodysplasia with hypocellular BM.