The clinical association of diabetes mellitus and coronary disease has been an accepted fact for many years. Levine,1 in a study of 103 patients with angina pectoris, found that 6.8 per cent showed well defined diabetes. (In a second series of 145 similar cases he2 found 23.7 per cent to have either glycosuria or known diabetes.) A study by Kahn3 showed that 12 per cent of eighty-two patients complaining of angina pectoris had true diabetes, and a similar investigation by Conner4 revealed that 10 per cent of his patients with coronary thrombosis had diabetes.
The tendency of the arteries of diabetic patients to undergo marked sclerosis has long been known, attention having been called to the frequency in diabetes of peripheral gangrene and the presence of sclerotic vessels in limbs amputated for this reason. Warren5 found thrombosis of the coronary arteries with recent or healed