Effect of African-American race on cancer specific mortality differs according to clear cell vs. non-clear cell histologic subtype in metastatic renal cell carcinoma

Aim: To test the effect of African-American race on cancer specific mortality (CSM) in clear cell metastatic renal cell carcinoma (ccmRCC) and non-ccmRCC. Patients and methods: Within Surveillance, Epidemiology and End Results registry (2001–2014), we identified patients with ccmRCC and non-ccmRCC. We relied on propensity score (PS) matching to reduce the effect of inherent differences between African-American vs. Caucasian patients. After PS matching that included access to cytoreductive nephrectomy (CNT), cumulative incidence, competing-risks regression (CRR) models and landmark analyses tested the effect of race on CSM. Results: Before PS matching, African-American patients accounted for 7.0 and 24.5% of respectively ccmRCC (N = 6742) and non-ccmRCC patients (N = 766). After PS matching, African-American patients accounted for 22.3 and 33.5% of respectively ccmRCC (N = 2050) and non-ccmRCC (N = 391) matched cohorts. In multivariable CRR models focusing on ccmRCC, higher CSM was recorded in African-Americans (HR:1.27, p < 0.001). Conversely, in non-ccmRCC, lower CSM was recorded in African-Americans (HR:0.54, p < 0.001). Landmark analyses rejected the hypothesis of immortal time bias. Conclusion: African-Americans experienced higher CSM in ccmRCC. Conversely, African-Americans experienced lower CSM, when diagnosed with non-ccmRCC. These differences are independent of access to CNT and warrant further study since they may have an impact on efficacy or access to systemic therapies. © 2018 Elsevier Ltd