Fallacy of thoracic side-branch steal from the internal mammary artery: analysis of left internal mammary artery coronary flow during thoracic side-branch occlusion with pharmacologic and exercise-induced hyperemia.

In some patients, myocardial ischemia after coronary artery bypass graft surgery has been attributed to a coronary steal phenomenon through a thoracic side branch originating from the left internal mammary artery (LIMA), even in the absence of subclavian or LIMA stenosis. To demonstrate that coronary flow through the LIMA is unchanged by occlusion of a LIMA side branch, we examined LIMA coronary flow velocity measurements (0.014" Doppler flow wire) in three patients at rest, during adenosine hyperemia, and again during hyperemia induced by left arm exercise before and again after the balloon occlusion of the thoracic side branch. For the three patients, no significant changes in resting or hyperemic flow were noted due to side-branch occlusion. Before side-branch occlusion, pharmacologic intra-arterial (adenosine) coronary flow reserve (hyperemic-to-basal flow velocity ratio) was 2.6, 1.5, and 3.2 and exercise flow reserve was 2.1, 1.3, and 1.2, respectively. After side-branch occlusion, pharmacologic coronary flow reserve was 2.5, 1.8, and 2.7 with exercise flow reserve of 1.8, 1.1, and 1.3, respectively. Under most ordinary circumstances, thoracic side-branch steal does not exist and that side-branch occlusion does not alter LIMA flow at rest or during pharmacologic or exercise-induced hyperemia. These data further suggest that a demonstration of the physiologic value of side-branch occlusion should precede surgical or percutaneous interruption of the thoracic artery in such patients.