The extensive timespan of modern radial velocity surveys have made the discovery of long-period substellar companions more common in recent years, however measuring the true masses of these objects remains challenging. Astrometry from the Gaia mission is expected to provide mass measurements for many of these long-period companions, but this data is not yet available. However, combining proper motion data from Gaia DR2 and the earlier Hipparcos mission makes it possible to measure true masses of substellar companions in favourable cases. In this work, we combine radial velocities with Hipparcos-Gaia astrometry to measure the true masses of two recently discovered long-period substellar companion candidates, HD 92987 B and HD 221420 b. In both cases, we find that the true masses are significantly higher than implied by radial velocities alone. A $2087 pm 19$ m s$^{-1}$ astrometric signal reveals that HD 92987 B is not close to its $17$ $M_J$ minimum mass but is instead a $0.2562 pm 0.0045$ $M_odot$ star viewed at a near-polar orbital inclination, whereas the $22.9 pm 2.2$ $M_J$ HD 221420 b can be plausibly interpreted as a high-mass super-planet or a low-mass brown dwarf. With semi-major axes of $sim$10 AU both companions are interesting targets for direct imaging, and HD 221420 b in particular would be a benchmark metal-rich substellar object if it proves possible to directly detect. Our results demonstrate the power of Hipparcos-Gaia astrometry for studying long-period planet and brown dwarf candidates discovered from radial velocity surveys.