Reduction of infarct size--a rational aim?