The purpose of this study was the evaluation of the impact of a variable relative biological effectiveness (RBE) compared to a constant RBE value of 1.1 in proton therapy prostate trials due to uncertainties in α/β ratio. Twenty patients receiving passive scattered proton therapy (PSPT) and fifteen patients receiving intensity modulated proton therapy (IMPT) were compared to twenty patients treated with 7-field intensity modulated photon therapy (IMRT). For proton beam therapy (PBT), the RBE was estimated using two different RBE models. Tumor control probabilities (TCP) and normal tissue complication probabilities (NTCP) were assessed. For one of the RBE models, dosimetric indices assuming a low α/β were ~10%-11% larger compared to using a fixed RBE. A different model resulted in 1%-3% lower values independent of α/β. Comparing PBT with IMRT revealed a negligible difference in TCP for a fixed RBE. Applying a variable RBE revealed an increase in TCP by 6% for PBT compared to IMRT for one model but a decrease of 2% for the other. Variable RBE values in PSPT resulted in an increase in NTCP for rectum from 7% to 11% for a fixed RBE with one model but a decrease to 6% for another. For IMPT, NTCP increased from 5% to 9% for a fixed RBE for one model but decreased to 3% using the other. The NTCP for bladder increased for PSPT for both models, from 11% to 19% and 14%, respectively. For IMPT, the NTCP increased from 12% to 17% using one model but decreased to 11% with the other. In radiation therapy for prostate cancer, disregarding variable RBE may lead to either underestimation or overestimation of the expected TCP and NTCP, depending on the RBE model and α/β. This should be considered when estimating uncertainties when comparing PBT and IMRT outcomes in clinical trials.