This one cracked me up. Okay, so I have a warped sense of humour.
The crooked origin of the lipid hypothesis generates an almost infinite number of paradoxes, here's a nice abstract about (yet) another paradox, this time in survival after ACS.
"The association of hypercholesterolemia with better outcomes highlights a major challenge in observational analyses"
Which prompted this anecdote from an eminent THINCS member:
A schizophrenic patient believes he is dead. The patient's psychiatrist, trying to cure the him of his delusion says, "Do dead people bleed?" The patient says, "No, they don't bleed." The Psychiatrist pricks the patient's finger and blood flows out. The patient says, "Well, this shows that dead people DO bleed".