The requirements of the first stage are that it deliver about 3340 m/s of delta v to a 690 ton payload (the upper stages and spacecraft), with an initial thrust-to-weight ratio of at least 1.16:1.
The best candidate for a first-stage hydrogen engine in the Saturn era would be the never-completed M-1. At sea level, it would be much less powerful than the F-1, but with better specific impulse - 310 seconds vs 263. Without going into too much detail, the result is that the first stage carries 1730 tons of LH2/LOX rather than 2160 tons of kerosene/LOX. However, depending on the exact mixture ratios in use, kerosene/LOX is about 3.5 times denser than LH2/LOX, so even with less fuel mass, the stage needs to get a lot bigger. If the 10 meter diameter is retained, the stage gets stretched from 42m to 92m -- a very long and skinny stage. More reasonable would be a 12-meter diameter, 63 meter long first stage. The overall rocket, all-up for an Apollo J mission, masses 2593 tons instead of 2970 tons. It might look something like this:
The drag penalty wouldn't be terribly severe. About 50 m/s of the Saturn V's total potential is lost to drag; the larger diameter stage would probably incur another ~25m/s hit, but I don't want to go back and recalculate.
To achieve the initial TWR requirement, 8 M-1s are needed, granting 1.21:1 TWR. As fuel mass was consumed and specific impulse (and thus thrust) increased, 2 or 4 of them would likely be shut off over the course of the burn to limit acceleration for crew comfort, with the remainder left running til fuel-out. Acceleration-by-time curves are shallower for higher Isp engines, so we may have a less efficient ascent than the Saturn V. Engine shutdown might therefore be delayed to make up some speed; this could mean a higher peak acceleration than Saturn V.
Total hydrogen carried by the rocket would be about 4x that of the Saturn V; presumably that would quadruple the ground storage requirements.
Bear in mind that even though this rocket masses less in total, it probably costs much more because of first stage construction and transport logistics, engine complexity, and hydrogen handling, and on that basis it's somewhat inferior to the Saturn V.
Another possibility would be using 16x SSME instead of 8x M-1. This probably couldn't have flown until the 1980s, but the sea level specific impulse is far better, 366s instead of 310s. This would further reduce the first-stage fuel mass but probably add a lot of cost -- the entire shuttle program only built 42 of those engines.