Background: Targeted therapies have expanded treatment options for mRCC. Using a joint community/academic center database, we determined outcomes for “real world” mRCC patients (pts) in order to understand current treatment patterns and their effectiveness. Methods: A retrospective cohort dataset included all identified adult mRCC pts receiving care in Duke University Health System and 11 community oncology practices, diagnosed and alive since Jan. 2007, and not enrolled in a trial. Standardized chart abstraction supplemented the electronic medical record. Demographics, comorbidities, tumor characteristics, treatment patterns and outcomes were recorded. Pts were grouped by exposure sequence (TKI [sorafenib, sunitinib or pazopanib], mTOR [temsirolimus or everolimus], TKI/mTOR, etc.) for 9 sequences. Outcomes analysis for overall survival (OS) included Cox regression. Results: The cohort included 385 pts, 64 yrs old (±10.1), 66% male, 72% Caucasian; 64% had clear cell histology (25% unknown). Median OS was: mTOR (N=33; 5.4 mo), mTOR/TKI (N=11; 9.3 mo), TKI/mTOR (N=33; 13.7 mo), TKI (N=108; 21.1 mo), TKI/mTOR/TKI (N=18; 29.8 mo), TKI/TKI/mTOR (N=11; 33.1 mo), Other (N=40; 38.2 mo), TKI/TKI (N=32; 42.9 mo), No Therapy (N=99; not reached). “Other” pts received a combination or a non-TKI, non-mTOR monotherapy in the 1st three exposures. “No Therapy” pts had received no systemic therapy as of data analysis; most were alive and censored. Pts in the mTOR only sequence had trends toward worse disease at baseline with more metastatic sites (median 2.0; p=.067) and impaired performance status (18%; p=NS). When significant covariates were controlled, OS (vs. No Therapy) was poorest for pts receiving mTOR only (Hazard Ratio[HR]=2.3) and best for pts receiving TKI/TKI (HR=0.50). Significant risks were evident with liver metastasis (HR=1.7) and impaired performance status (HR=3.1). Conclusions: In the real world setting pts receiving mTOR as the 1st treatment exposure had worse OS, likely because pts with more aggressive disease were preferentially treated with mTOR. Outcomes were consistent with contemporary trials; however, several sequences (e.g. TKI/TKI, No Therapy) had surprisingly long OS, which requires further evaluation.