Background: Radium 223 (Ra-223) was approved for the treatment of metastatic castration resistant prostate cancer (mCRPC) patients with bone-only disease, following demonstration of significant improvement in overall survival (OS). To date, there are no validated prognostic factors useful in predicting outcome of mCRPC patients treated with Ra-223. Our retrospective study aims to evaluate the prognostic role of treatment discontinuation due to adverse events in mCRPC patients treated with Ra-223, and to identify which factors correlate with the toxicity onset. Methods: We performed a retrospective analysis of all consecutive mCRPC patients treated with Ra-223 from September 2013 to December 2019 at our institute. Patients were divided in 2 groups according to the reason of Ra-223 therapy discontinuation: toxicity versus other causes. Outcome measures were progression-free survival (PFS) and OS. Results: In the overall population (75 patients) median PFS and OS were 5.46 months and 11.15 months respectively. Patients who discontinued treatment due to toxicity had a lower median PFS (3.49 vs 5.89 months, HR: 1.88, 95% CI: 1.14-3.12, p = 0.014) and OS (8.59 vs 14.7 months HR: 3.33, 95% CI: 1.85-6.01, p < 0.001) than patients who discontinued therapy due to other causes. The risk of Ra-223 discontinuation due to toxicity correlates with the number of previous treatments ( p = 0.002), previous chemotherapy treatment ( p = 0.039), baseline LDH ( p = 0.012), Hb ( p = 0.021) and platelet-to-lymphocyte ratio ( p = 0.024). Conclusions: Discontinuation due to toxicity is associated with worse outcomes in mCRPC patients treated with Ra-223. To reduce the risk of developing toxicities that may compromise treatment efficacy, Ra-223 should be used early in mCRPC patients.