IIRC (and I may not be), they found something like a 50% increased risk of heart attacks in those that took Vioxx for more than 18 months straight.
Really though, the recall of Vioxx is no big deal to the overwhelming majority of arthritis sufferers. Vioxx, Celebrex and other members of that class of medicines are not any better at controlling arthritis pain than plain old Advil or Aleve. Their only real benefit is that they don't cause upset stomachs in patients prone to that sort of thing.
A better, more accurate, statement would be:
"They do not cause or aggravate bleeding ulcers, as do NSAIDS."
Hope that helps.
DG
This sort of thinking is the problem with the way the FDA approves pharmaceuticals and the way trials are done. Medications can behave very differently for different people. Aleve is wonderful for my migraines, but did little for bursitis. Celebrex was wonderful. However, for my father, Celebrex was worthless and caused problems. [He went onto Vioxx for a couple of years...and then had a heart attack... :-( ]
The statistics and theories don't account for the personal variation in efficacy. While the average advantage for most people is merely the lack of aggravating an ulcer, there is a great difference in individual response in terms of effectiveness, too--varying from patient to patient for a single medication, as well as from one medication to another in a single patient.