Matt Walker makes this point in his book (at the risk of sounding like a shill). One group of people who sleep abnormally long hours are those suffering from various illnesses - they sleep more either as a result of the illness or while in recovery. Some of them may die. This group potentially skews results and increases the mortality rate of long sleepers. There is no evidence I'm aware of (nor does the book give any) that oversleeping is harmful in itself.

That book is so good that I feel like a shill when I recommend because I recommend it so enthusiastically.

Every single causal link made in epidemiologic studies will have potential confounders that need to be adjusted for. I'd be shocked if anyone could think of a causal link without a potential confounder, never mind that we may have unknown/unmeasured confounders that aren't even top of mind.