The observation of the high-temperature (≳10 MK) Fe XXI 1354.1 Å line with the Interface Region Imaging Spectrograph has provided significant insights into the chromospheric evaporation process in flares. In particular, the line is often observed to be completely blueshifted, in contrast to previous observations at lower spatial and spectral resolution, and in agreement with predictions from theoretical models. Interestingly, the line is also observed to be mostly symmetric and significantly broader than expected from thermal motions (assuming the peak formation temperature of the ion is in equilibrium). One popular interpretation for the nonthermal broadening is the superposition of flows from different loop strands. In this work, we test this scenario by forward-modeling the Fe XXI line profile assuming different possible observational scenarios using hydrodynamic simulations of multi-thread flare loops with the 1D RADYN code. Our results indicate that the superposition of flows alone cannot easily reproduce both the symmetry and the significant broadening of the line and that some other physical process, such as turbulence, or a much larger ion temperature than previously expected, likely needs to be invoked in order to explain the observed profiles.