Aspirin for people with diabetes: a misleading inference in a recent meta-analysis?