One honest way to deal with it is to write something like that it is "attributed by X to Y", where X is the author who made the unclear citation and Y is the cited paper that does not seem to you to contain this information. That casts some doubt on the attribution, since you are not endorsing it yourself, while giving the reader all the information you have.
One possibility is that the information really is there, but it takes some digging to extract it. In that case, flatly saying that it's not there would be awkward. On the other hand, you don't want to perpetuate a mistaken or dishonest citation without any warning for the reader. (Of course, you could simply replace it with a correct citation if you have the relevant expertise to do that, but then you wouldn't be asking this question.)
If X is still active, you could write and ask whether he/she accidentally cited the wrong paper, or whether you are missing something. That's the best way to find out what's behind this citation, since nobody else can really say for sure. If you can't contact the author, you can try asking other researchers you know.
Sometimes it's a random error, and sometimes it comes from copying incorrect citations from elsewhere. Occasionally there's a defensible historical reason for a confusing citation. For example, maybe the original author had an idea and but didn't explain it clearly at all, and later authors expressed it more clearly but cited the original paper (since it contained the key idea).