So if 4.60% of active cases die, then as a first-order approximation would this be correct?:

418,000 (still-active cases worldwide)*0.04597=19,219 more deaths worldwide after 27 Mar?

94,091 (still-active US cases)*0.04597=4,326 more US deaths after 27 Mar?

Or will the number of deaths be even larger before the curves flatten out?

A more sophisticated model currently predicts 81,114 US deaths by 01 Aug (range 38,242–162,106).