Background: Two PARP inhibitors (olaparib, rucaparib) are now FDA approved for mCRPC patients with mutations in BRCA1/2, but the relative efficacy of PARP inhibition in BRCA1- vs BRCA2-altered mCRPC is understudied. Methods: We conducted a multi-center retrospective analysis involving 11 academic sites. We collected genomic and clinical data from 123 BRCA1/2-altered mCRPC patients receiving PARP inhibitor treatment. The primary efficacy endpoint was PSA50 response rate (≥50% decline in PSA level). Secondary endpoints were PFS (clinical, radiographic or PSA progression, whichever occurred first) and overall survival (OS). We also captured information on other concurrent genomic alterations. We compared genomic characteristics and clinical outcomes among BRCA1- vs BRCA2-altered mCRPC. Results: A total of 123 patients (13 BRCA1, 110 BRCA2) were included. PARP inhibitors used were olaparib (n = 116), rucaparib (n = 3), talazoparib (n = 2) and veliparib (n = 2). All BRCA1 and 71% of BRCA2 patients had Gleason 8-10 disease. Compared to BRCA2 patients, BRCA1 patients were more likely to have received prior taxane chemotherapy (77% vs 62%). Age, baseline PSA, and prior enzalutamide/abiraterone treatment were similar between groups. PSA50 responses in BRCA1- vs BRCA2-altered patients were 38% vs 65% respectively (P= 0.06). Median PFS in BRCA1 vs BRCA2 patients was 3.0 mo (95%CI, 0.9–5.1) vs 8.0 mo (95%CI, 5.3–10.6) respectively (HR 0.42, P= 0.01). Similarly, median OS in BRCA1 vs BRCA2 patients was 11.0 mo (95%CI, 9.9–12.1) vs 24.0 mo (95%CI, 18.2–29.8) respectively (HR 0.33, P= 0.005). There were roughly equal proportions of germline mutations (50% vs 45%) and biallelic mutations (31% vs 25%) in the BRCA2 and BRCA1 groups, respectively. There were numerically more TP53 mutations in the BRCA1 vs BRCA2 group (40% vs 29%, P= 0.45), but an equal number of mutations in other key genes (PTEN, RB1 and AR). Conclusions: PARP inhibitor activity is diminished in BRCA1- vs BRCA2-altered mCRPC. In our cohort, this differential activity was not explained by mutation origin (germline vs somatic) or allelic status (mono- vs biallelic). These findings warrant validation.