In free space, these mics are effectively fully omnidirectional up to a very high frequency. 12 years ago I convinced myself I could hear a slight top end increase directly on-axis, using a test source with lots of HF and ultrasonic content a few feet away from the microphone. In actual use, I've never been able to hear a difference and no longer try to point them directly at the source. Could I still hear a difference today with 10 year older ears? Was my 10 year old test really sound, or might I have been fooling myself? [shrug]
However, once you mount them to something, that changes things significantly. Not in terms of which way the mic itself is oriented on the surface to which it is mounted, but in terms of which direction the surface itself is facing, the size and geometry the surface, the acoustic properties of the surface material, and in some cases how close/tight-coupled the mic is to that surface. Spacing between the two microphones is always a variable of key importance, and angle (of the surface at each microphone, not of the mic-grid on that surface in this case) also becomes equally important once they are mounted to something. My advice is to focus on getting the spacing and angle you want using whichever mic-orientation works most practically to achieve that. And when fine-tuning your setup, adjust those variables without worrying overly much about mic-grid orientation.
That said, in the hierarchy of what matters, the direction the mic grid points probably has more influence than choice of recording sample rates above 48kHz, IMHO.