L-arginine, a substrate in the production of endotheliumderived nitric oxide (NO), may stimulate the release of NO. L-arginine improves the coronary blood flow response to acetylcholine in patients with normal coronary arteries and hypercholesterolaemia. However, it is unknown whether local administration of L-arginine dilates coronary stenoses in patients with stable angina. This study assessed the effects of L-arginine and D-arginine at the stenotic site in patients with coronary artery disease and stable angina. 15 patients with chronic stable angina, coronary artery disease, and a positive treadmill exercise test ( 0·1 mV ST segment depression) at between 5 and 7 METS with the modified Bruce protocol, were studied in the cardiac catheterisation laboratory of the Hippokration Hospital, Athens University. The protocol was approved by the research ethics committee of the Hippokration Hospital and each patient gave written informed consent. Antianginal medication was stopped 24 h before the study. The patients were allowed to use sublingual glyceryl trinitrate as necessary, but no study was done within 3 h of its administration. None of the study patients had taken glyceryl trinitrate on the day of the study. Following the diagnostic coronary angiogram, an optimum radiographic projection was selected and kept constant for subsequent angiograms. In ten patients (seven male, three female, mean age 58 [SD 8] years) intracoronary infusion of 0·9% saline (2 mL/min) for 2 min was followed by intracoronary infusion of 150 μmol/min of L-arginine for 8 min. In five patients (four male, one female, mean age 61 [10] years) the same protocol was used, substituting 150 μmol/min of D-arginine for L-arginine. The chosen dose and the infusion time of L-arginine administration was consistent with a study showing a response after L-arginine administration in the brachial artery. The arterial segments in each frame were analysed at the Hammersmith Hospital in random order with quantitative computerised analysis, with an automated edge contour detection analysis system. The percentage change in luminal diameter from baseline was calculated at the end of the 8 min infusions and the values for L-arginine and D-arginine and normal saline were compared with Student’s t-test. The infusions of L-arginine and D-arginine did not affect systolic blood pressure (mean [SE]) (L-arginine: mean 152 [7·1] vs 150 [7·2] mm Hg, p=0·82 D-arginine: 154 [11·2] vs 157 [11·6] mm Hg, before and after infusions, respectively, p=0·27). L-arginine administration was associated with significant dilatation of stenoses (11·7 [2·3]%: p=0·004 vs D-arginine, p=0·001 vs normal saline) (figure), but there was no significant change with D-arginine (2·0 [0·04]%) (figure). No patients were receiving longacting nitrates, and the previous antianginal treatment did not affect the response to L-arginine ( -blockers: 11·1 [6·0]%; calcium antagonists 10·5 [4·7]%; -blockers and calcium antagonists 12·5 [3·4]%). Arginine, a semiessential aminoacid serves as the substrate for the enzyme NO synthase, which converts arginine to citrulline and NO, reducing vascular tone. A recent study in experimental models suggested possible beneficial effects of L-arginine administration. Long-term oral therapy with L-arginine reduced endothelial dysfunction and inhibited the development of atherosclerosis, but the mechanisms by which it exerts its vasodilator effect are not completely understood. Possible explanations include an effect on endothelial synthesis and release of NO, and direct or indirect physiological actions on the vascular smooth muscle to augment its responsiveness to NO. In our study a significant dilation was found in coronary stenotic segments after L-arginine administration in patients with coronary artery disease and stable angina. These data suggest that the 25