No, the 42% is correct. You have to compare the proportion who got longcovid taking the drug, vs those who got longcovid without taking it, i.e. (10.6-6)/10.6.

No, the 42% is correct. You have to compare the proportion who got longcovid taking the drug, vs those who got longcovid without taking it, i.e. (10.6-6)/10.6. Also, the study just looked at obese patients, who are at high risk of both diabetes and covid complications. So this likely just reduced those complications, and not "longcovid" as a whole (which can mean a number of different things). I would imagine it had no effect against the ME/CFS type of longcovid.