In my experience, IN-18 do suffer cathode poisoning more than any other tube. But is is also hit and miss. Some batches seem better than others, but I have not been able to correlate this is terms of a specific set of dates or ages.
Its possible that the Soviets meddled with the gas / mercury fill over the production years, or even that they did not have close control over it.
Direct drive of these large tubes is important to give a nice bright display, and no noise. (large metal digits DO resonate and cause noise much more than in smaller tubes. ). However a direct drive methodology will not in itself stop cathode poisoning. A good firmware that regularly cycles all digits will undoubtedly help, although I have experienced that even this cannot stop a tube that is doggedly determined to get cathode poisoning. Note that the comparably sized European Z5660M tube has never presented me with a single case of cathode poisoning, even when driven by the identical circuit.
You asked for a circuit... .have a look at the back pages of this manual, and you will see the basic drive circuit and use of HV5622 ICs for the direct drive. Each has 32 outputs, so two ICs give the exact number of outputs for six tubes and two double-dot neon separators. How convenient!