Water (well earth's anyway and cometary ice, IIRC) actually is relatively rich in Deuterium - compared to GGs, like Jupiter, which aren't... so IMTU I fiddled with harnessing the neutron and Tritium from D+D fusion to create more D as well as harness the byproducts for D-T, D+He3, etc. (not that the ratios, energy yields would work out in RL) to rationalize using skimmed GGs. Smaller GGs (Ice giants) probably have a higher percentage of D, but worlds and asteroids would likely be a better source.
For a long time this is the model I used - with a rule mechanic that favored water/ice based refueling vs GGs - especially considering something had to be done with the He or whatever byproduct, so not using all the L-Hyd seemed reasonable.
