Abstract Background Although abiraterone acetate (abiraterone) has proven efficacy in two randomised phase 3 trials in metastatic castration-resistant prostate cancer (mCRPC), patients who had an Eastern Cooperative Oncology Group (ECOG) performance status (PS) ≥2 were either excluded or under-represented in these trials. Objective To compare outcomes in ECOG PS 0–1 and ≥2 in mCRPC patients treated with abiraterone. Design, setting, and participants Cancer registries from three Canadian centres were used to retrospectively identify mCRPC patients (postdocetaxel and docetaxel-naïve) treated with abiraterone. ECOG PS, clinicopathologic characteristics, prostate-specific antigen (PSA) response, and survival data were collected. Outcome measurements and statistical analysis Survival outcomes were estimated using the Kaplan-Meier method and compared using the log-rank test. Cox proportional hazards modelling was used to examine the effect of clinicopathologic characteristics on overall survival (OS) and time to PSA progression. Results and limitations A total of 519 patients were identified; 61% (n=318) and 39% (n=201) were ECOG PS 0–1 and ≥2, respectively. ECOG PS 0–1 patients were significantly more likely than PS ≥2 patients to achieve a PSA decline ≥50% from baseline (45% vs 32%; p=0.003, Fisher exact test) and had significantly longer median time to PSA progression (5.2 mo vs 4.1 mo; p=0.023), median treatment duration (7.4 mo vs 4.5 mo; p<0.001), and median OS (20.0 mo vs 9.1 mo; p<0.001). On multivariate analysis, ECOG PS was a significant factor for OS (p<0.001), time to PSA progression (p=0.043), and PSA decline (p=0.002). Potential limitations include the retrospective study design and subjective nature of ECOG PS classification. Conclusions ECOG PS ≥2 mCRPC patients treated with abiraterone have inferior outcomes compared with ECOG 0–1 patients, especially in regard to OS. These data indicate that early initiation of abiraterone prior to a decline in PS may be warranted. Patient summary We found that advanced prostate cancer patients who have worse performance status (PS) derive less benefit from abiraterone, indicating that earlier treatment before PS declines could improve outcomes.