According to the information on the referencerecordings' site the so called HRx files are in the 24-bit/176.4 kHz wave format.
I just tested a couple of 24-bit/176.4 kHz/2-channel wave files. MC didn't have any problems with them. My sound card can handle up to 96 kHz so the Windows kernel mixer "helpfully" resampled the output to 96 kHz. Naturally I couldn't use the ASIO output mode, which bypasses the kernel mixer, because the sample rate is not supported by my HW.
In any case, I think using such a sample rate is waste of storage space. It can only provide a wider frequency range that can extend up to 88.2 kHz. It doesn't make the audible part of the frequency range technically better. If the audio tracks are mastered better the user may get an impression of better audio quality, but that impression would not be any different if the sample rate were 44.1 kHz or 48 kHz.
The same is true with the bit depth -- the 24-bit mode in an end user product is mostly waste of space. A 16-bit audio signal can provide a dynamic range of 96 dB. A very silent room may have a 30dB SPL background noise level. 30 dB SPL + 96 dB = 126 dB SPL. That should be enough for everyone. Once again, if the signals are produced correctly the only difference between 16-bit and 24-bit is in the maximum available dynamic range. 16-bit is measurably noisier, but that noise is normally inaudible.
The 24-bit mode (or even more preferably the 32-bit floating point mode) is useful in a production environment because the additional headroom is needed for multiple file operations.