As others said, simply separating the CO2 away would mostly get rid of the problem - all non-stable isotopes of Oxygen have a max half life of 122 seconds, but, for the record, isotopic separation is easy when there is a very large difference in atomic weight.
In Uranium's case it's difficult because it's 235/238 = 1.28% (actually way worse - Uranium hexafluoride is used, which adds 114 units, bringing the ratio down to 0.86%)
In Oxygen's case the ratio would be at least 6.6% (15 vs 16).
Most importantly in Uranium you are interested in the tiny amount of U235, while in Oxygen you'd be interested in the huge bulk of O16, O17, O18, which are the stable isotopes. O16 alone is 99.762% of all Oxygen, and you can afford to lose half in your centrifuge if it spares you a few cycles, it's not exactly hard to come by.
