Well caps are not as ideal as one would think. There are a few different non-ideal characteristics that arise depending upon the type of capacitor. So if you are trying to maintain the ideal output impedance of the cathode follower you have to take those into account. For example if the cap was a large value say 100uF cap (sufficiently rated for the DC voltage on the cathode, 100-300V) then it would likely be an electrolytic capacitor. Electrolytics suffer from something known as ESR (equivalent series resistance). Usually at low frequencies for an electrolytic the ESR is like a few ohms, so in the case of AC coupling from circuit to circuit that means nothing. But at higher frequencies say 1kHz and up for example, the ESR increases. A cap with ESR is modeled as a capacitor with a series resistor so as you can see if you are adding series resistance off the cathode follower circuit, that will increase the output impedance as frequency increases!
Now lets assume instead of a electrolytic they used a 100nF film cap. Film caps suffer very little ESR, they are more ideal in that regards. However if you are trying to drive a load of around 400-600R as calculated by the followers intrinsic impedance, then the 100nF -400-600R RC filter results in lots of filtering of the low end. So effectively at low frequencies the output impedance is very very large in this case.
So one solution people often use for these things is to put a 100nF cap and a large electrolytic in parallel. In this case you can overcome the ESR at larger frequencies using the 100nF cap to shunt across the 100uF cap.
I cant say for sure if this is a problem with your preamp or not until I know the output cap used for the DI and the load that it is being placed under.