The coronary sinus: an alternate channel for administration of arterial blood and pharmacologic agents for protection and treatment of acute cardiac ischemia

Signs and signals limiting myocardial damage using PICSO: a scoping review decoding paradigm shifts toward a new encounter

Background

Inducing recovery in myocardial ischemia is limited to a timely reopening of infarct vessels and clearing the cardiac microcirculation, but additional molecular factors may impact recovery.

Objective

In this scoping review, we identify the paradigm shifts decoding the branching points of experimental and clinical evidence of pressure-controlled intermittent coronary sinus occlusion (PICSO), focusing on myocardial salvage and molecular implications on infarct healing and repair.

Design

The reporting of evidence was structured chronologically, describing the evolution of the concept from mainstream research to core findings dictating a paradigm change. All data reported in this scoping review are based on published data, but new evaluations are also included.

Results

Previous findings relate hemodynamic PICSO effects clearing reperfused microcirculation to myocardial salvage. The activation of venous endothelium opened a new avenue for understanding PICSO. A flow-sensitive signaling molecule, miR-145-5p, showed a five-fold increase in porcine myocardium subjected to PICSO.Verifying our theory of "embryonic recall," an upregulation of miR-19b and miR-101 significantly correlates to the time of pressure increase in cardiac veins during PICSO (r2 = 0.90, p < 0.05; r2 = 0.98, p < 0.03), suggesting a flow- and pressure-dependent secretion of signaling molecules into the coronary circulation. Furthermore, cardiomyocyte proliferation by miR-19b and the protective role of miR-101 against remodeling show another potential interaction of PICSO in myocardial healing.

Conclusion

Molecular signaling during PICSO may contribute to retroperfusion toward deprived myocardium and clearing the reperfused cardiac microcirculation. A burst of specific miRNA reiterating embryonic molecular pathways may play a role in targeting myocardial jeopardy and will be an essential therapeutic contribution in limiting infarcts in recovering patients.

Non-surgical stem cell delivery strategies and in vivo cell tracking to injured myocardium

Heart failure is a major economic and public health problem. Despite the recent advances in drug therapy and coronary revascularization, the lost cardiomyocytes due to necrosis and apoptosis are not replaced by new myocardial tissue. Cell therapy is an interesting therapeutic option as it potentially improves contractility and restores regional ventricular function. Early clinical data demonstrated that cell transplantation, mainly delivered through non-surgical methods, is safe and feasible. However, several important issues need to be elucidated. This includes, next to determining the best cell type, the optimal delivery strategy, the biodistribution and the survival of implanted stem cells after transplantation. In this view, pre-clinical animal experiments are indispensable. Reporter genes, magnetic or radioactive labeling of stem cells have been developed to observe the fate and the distribution of transplanted cells using non-invasive imaging techniques. Several studies have demonstrated that these direct and non-direct labeling techniques may become an important tool in cell therapy. Integration of cell delivery and cell tracking will probably be a key for the success of cell therapy in patients. This review will provide a comprehensive overview on the various cell tracking and non-surgical cell delivery techniques, which are highly important in view of experimental and clinical studies.

Selective regional myocardial infiltration by the percutaneous coronary venous route: A novel technique for local drug delivery

Recent advances in the treatment of heart disease, in particular cardiovascular gene therapy and therapeutic angiogenesis, highlight the need for efficient and practical local delivery methods for the heart. We assessed the feasibility of percutaneous selective coronary venous cannulation and injection as a novel approach to local myocardial drug delivery. In anesthetized swine, the coronary sinus was cannulated percutaneously and a balloon-tipped catheter advanced to the anterior interventricular vein (AIV) or middle cardiac vein (MCV). During balloon occlusion, venous injection of radiographic contrast caused regional infiltration of targeted myocardial regions. Complete AIV occlusion had no impact on LAD flow parameters. Videodensitometric analysis following venous injection showed that radiographic contrast persisted for at least 30 min. Selective regional myocardial infiltration is feasible by this approach, targeting selected myocardial beds, including the apex, anterior wall, septum, and inferoposterior wall. This novel technique has potential application for local myocardial drug or growth factor delivery. Cathet. Cardiovasc. Intervent. 51:358-363, 2000.

Utility of standby cardiopulmonary support for elective coronary interventions

At our institution, elective coronary interventions are performed without formal surgical backup. Instead, a policy of "standby cardiopulmonary support" (CPS), and "next-available operating room" is used. Standby CPS requires a perfusionist dedicated to the catheterization laboratory with immediate access to CPS apparatus. Between January 1989 and June 1994 we performed 2,850 elective coronary interventions. Eleven patients (0.4%) required emergency CPS. None of these patients fell into a high-risk category for PTCA (i.e., sole circulation, ejection fraction <20%, unprotected left main). Eight of these (73%) had completion of their coronary intervention while on CPS in the catheterization laboratory. Three patients were sustained on CPS until an operating room became available. All patients required blood transfusions and sustained non-Q-wave myocardial infarctions. Two late in-hospital deaths occurred. Nine patients (82%) were successfully discharged. Standby CPS provides hemodynamic support for patients who sustain a potentially catastrophic event during coronary intervention. Our data suggest that this modality should not be limited to high-risk patients.

Infarct size reduction with coronary venous retroinfusion of diltiazem in the acute occlusion/reperfusion porcine heart model

Calcium channel blockers are commonly used for the treatment of ischemic heart disease, but their effects on myocardial infarct size (IS) after reperfusion are not well known. Enflurane-anesthetized open-chest pigs subjected to 60-min left anterior descending coronary artery (LAD) occlusion followed by 3-h reperfusion were referred to one of the four experimental groups. Beginning 10 min before the onset of reperfusion, pigs in group A received diltiazem (7.5 micrograms/kg/min) retrogradely infused into the coronary vein for 30 min. In group B, the same amount of diltiazem was infused into the right atrium. A corresponding volume of saline was infused into the coronary vein in group C or into the right atrium in group D. IS expressed as a percentage of the myocardium at risk was significantly smaller (p < 0.01) in group A (33 +/- 14%; mean +/- SD) than in groups B (58 +/- 11%), C (58 +/- 11%), and D(63 +/- 10%). After reperfusion, functional recovery of the ischemic myocardium, determined by ultrasound crystals, was significantly more improved (p < 0.05) in group A as compared with other groups. The ischemic and nonischemic regional myocardial blood flow (RMBF), determined by radioactive microspheres, did not differ between four groups. Coronary venous retroinfusion of diltiazem had a myocardial protective effect in the porcine experimental model of acute coronary occlusion/reperfusion, whereas intravenous drug administration was not effective. The protective effect could not be attributed to washout of toxic metabolites from the ischemic area or to improved microcirculation. It was probably related to a pronounced accumulation of the calcium-channel blocker diltiazem in the ischemic myocardium achieved by the coronary venous route of delivery.

Differential effects of synchronized coronary sinus retroperfusion on regional myocardial function during brief occlusion of the left anterior descending and circumflex coronary arteries

To test the hypothesis that coronary sinus retroperfusion would preserve regional myocardial function during either left anterior descending or circumflex occlusion, sonomicrometer crystals were implanted in the midmyocardium of five chronically instrumented dogs. Regional fractional shortening was measured during 5 min of coronary occlusion with and without retroperfusion. Percent fractional shortening in the left anterior descending region fell from 18% at baseline to -4%(dyskinesis) after 3 min of left anterior descending occlusion. With coronary sinus retroperfusion, the percent fractional shortening declined from 16% at baseline to 0 (akinesis) during occlusion. A modest but significant improvement in percent fractional shortening in the ischemic region during left anterior descending occlusion occurred with retroperfusion (p < .05). By contrast, no amelioration of ischemic dysfunction occurred with retroperfusion during circumflex occlusion. Coaxial flow into the great cardiac vein during retroperfusion may provide preferential protection to ischemic myocardium supplied by the left anterior descending coronary artery. However, it is unlikely that the modest degree of improvement in regional function observed during acute left anterior descending occlusion would be mechanically important in the presence of significant ischemic dysfunction.

Prophylactic versus standby cardiopulmonary support for high risk percutaneous transluminal coronary angioplasty

OBJECTIVES

Data from a national registry of 23 centers using cardiopulmonary support (CPS) were analyzed to compare the risks and benefits of prophylactic CPS versus standby CPS for patients undergoing high risk coronary angioplasty.

BACKGROUND

Early data from the CPS registry documented a high angioplasty success rate as well as a high procedural morbidity rate. Because of this increased morbidity some high risk patients were placed on standby CPS instead of prophylactic CPS.

METHODS

Patients in the prophylactic CPS group had 18F or 20F venous and arterial cannulas inserted and cardiopulmonary bypass initiated. Patients in the standby CPS group were prepared for institution of cardiopulmonary bypass, but bypass was not actually initiated unless the patient sustained irreversible hemodynamic compromise.

RESULTS

There were 389 patients in the prophylactic CPS group and 180 in the standby CPS group. The groups were comparable with respect to most baseline characteristics, except that left ventricular ejection fraction was lower in the prophylactic CPS group. Thirteen of the 180 patients in the standby CPS group sustained irreversible hemodynamic compromise during the angioplasty procedure. Emergency institution of CPS was successfully initiated in 12 of these 13 patients in < 5 min. Procedural success was 88.7% for the prophylactic and 84.4% for the standby CPS group (p = NS). Major complications did not differ between groups. However, 42% of patients in the prophylactic CPS group sustained femoral access site complications or required blood transfusions, compared with only 11.7% of patients in the standby CPS group (p < 0.01). Among patients with an ejection fraction < or = 20%, procedural morbidity remained significantly higher in the prophylactic CPS group (41% vs. 9.4%, p < 0.01), but procedural mortality was higher in the standby group (4.8% vs. 18.8%, p < 0.05).

CONCLUSIONS

Patients in the standby and prophylactic CPS groups had comparable success and major complication rates, but procedural morbidity was higher in the prophylactic group. When required, standby CPS established immediate hemodynamic support during most angioplasty complications. For most patients, standby CPS was preferable to prophylactic CPS during high risk coronary angioplasty. However, patients with extremely depressed left ventricular function (ejection fraction < 20%) may benefit from institution of prophylactic CPS.

Supported angioplasty with synchronized retroperfusion in high-risk patients with left main trunk or near left main trunk obstruction

To test the feasibility of synchronized retroperfusion (SRP) as a support device of percutaneous transluminal coronary angioplasty (PTCA) for high-risk patients, 10 patients with left main trunk or near left main trunk obstruction underwent PTCA with SRP. An 8.5F retroperfusion catheter was inserted from the antecubital vein into the coronary sinus. Arterial blood was supplied through the catheter into the myocardium with a retroperfusion pump during the diastolic phase by means of ECG triggering. In all patients, the narrowings were successfully dilated and an improvement of more than 20% in the luminal diameter stenosis was achieved; however, narrowing of more than 50% (58%) remained in one patient. In all patients, systemic hemodynamics was maintained for more than 30 seconds during balloon inflation. In seven patients, a 60-second balloon inflation was possible without any collapse of systemic hemodynamics. To test the protective effect of SRP on myocardial ischemia and impairment of systemic hemodynamics, balloon inflation without SRP was performed in eight patients after successful dilatation. The duration for balloon inflation with SRP (71 +/- 30 seconds; n = 8) was significantly longer than that without SRP (56 +/- 30 seconds; n = 8). The decrease in systolic aortic pressure, the increase in pulmonary diastolic pressure, and ST-T segment elevation in the precordial lead of ECG during balloon inflation with SRP were less than those during balloon inflation without SRP. After PTCA, angina was not provoked by exercise stress testing in any of the 10 patients. We concluded that SRP is a beneficial support device of PTCA for high-risk patients.

Pharmacokinetic analysis of coronary venous retroinfusion: a comparison with anterograde coronary artery drug administration using metoprolol as a tracer

Plasma and myocardial tissue concentrations of metoprolol were studied in ischemic and nonischemic areas of 22 pigs after 90 (n = 19) and 16 (n = 3) min of left anterior descending coronary artery occlusion. Group A (n = 6) received simultaneous intravenous metoprolol (0.2 mg/kg body weight) and tritium-labeled (3H)-metoprolol (0.2 mg/kg) retrogradely into the coronary vein. In group B (n = 5), metoprolol and 3H-metoprolol were administered in the same way, but at half the volume to study the influence of derived coronary venous pressure on the myocardial concentration of drug. In group C (n = 3), metoprolol was given retrogradely and saline solution was infused into the left anterior descending artery before induced death to wash out metoprolol from the coronary veins. To rule out a possible influence of the development of myocardial necrosis on drug distribution, metoprolol was retroinfused after 1 min of arterial occlusion in three pigs (group D). In group E (n = 5), metoprolol (0.2 mg/kg) was infused anterogradely into the left anterior descending artery. Peak plasma concentration was significantly higher after intravenous infusion of metoprolol (1,188 +/- 503 nmol/liter) than after coronary venous infusion (417 +/- 155 nmol/liter; p less than 0.001). In groups A and B, the nonischemic myocardial concentration of metoprolol was 250 to 300 pmol/g, whether the drug was infused intravenously or into the coronary vein. Coronary venous retroinfusion, however, resulted in a substantial accumulation of metoprolol in the ischemic myocardium. In group A pigs, subendocardial myocardial concentration was 16,800 +/- 7,774, mid-myocardial 39,590 +/- 18,043 and subepicardial 57,143 +/- 29,030 pmol/g (mean +/- SE). The ischemic myocardial concentration in pigs from group B was somewhat less pronounced, probably secondary to a lower coronary venous pressure (15 +/- 3 mm Hg) with the lower volume of infusion (6.1 +/- 0.3 ml) in group B compared with 32 +/- 5 mm Hg with a 14 +/- 1 ml infusion in group A. Coronary artery anterograde administration resulted in myocardial ischemic and nonischemic zone drug concentrations similar to those observed after retroinfusion into the coronary vein. With both modes of administration, there was a transmyocardial gradient from a somewhat lower drug concentration in the subendocardium, toward an increasing level in the mid-myocardium, to the highest concentration in the subepicardial zone of the ischemic myocardium. Coronary venous retroinfusion resulted in pronounced drug accumulation in the ischemic myocardium. The derived coronary venous pressure during infusion influenced the concentration of drug.(ABSTRACT TRUNCATED AT 400 WORDS)

Retrograde coronary venous administration of recombinant tissue-type plasminogen activator: a unique and effective approach to coronary artery thrombolysis

Recent studies of interventional therapy by way of the coronary venous system have demonstrated that it can protect acutely ischemic myocardium. To evaluate the efficacy of coronary venous retroinfusion compared with systemic intravenous administration of recombinant tissue-type plasminogen activator (rt-PA), 14 dogs were studied with a copper coil-induced thrombus in the left anterior descending coronary artery. The rt-PA (24,000 fluorescence units/kg) was administered continuously, either intravenously (n = 8) or retrogradely (n = 6), for 30 min beginning 60 min after coronary occlusion. Thrombolysis was determined by repetitive coronary angiography. All dogs were killed 3 h after termination of rt-PA infusion and infarct size was measured by the triphenyltetrazolium chloride staining technique. Complete thrombolysis occurred in five of the six dogs in the retroinfusion group and four of the eight dogs in the systemic intravenous infusion group. Partial lysis was achieved in two dogs treated by intravenous infusion. Lysis did not occur in one dog treated with retroinfusion and in two dogs treated with intravenous infusion. Time to thrombolysis was 13.4 +/- 2.3 min in the retroinfusion group versus 27.8 +/- 4.8 min in the intravenous group (p less than 0.001). Myocardial functional recovery in the ischemic zone measured by two-dimensional echocardiography 60 min after reperfusion was significant only in the retroinfusion group (p less than 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)

Percutaneous cardiopulmonary bypass support in the catheterization laboratory: technique and complications

A safe and easily applied technique of percutaneous cardiopulmonary bypass support has been developed for use in the cardiac catheterization laboratory. The importance of this technique lies in its ability to maintain hemodynamic stability during high risk interventional procedures regardless of intrinsic cardiac function. Venous and arterial cannulas (18F) are inserted percutaneously over a stiff guide wire after sequential dilatation with 12F and 14F dilators. Bypass flow rates of up to 5 L/min can be achieved. This technique can be applied to support patients with cardiac arrest, hemodynamic collapse after abrupt closure during coronary angioplasty, and cardiogenic shock, as well as those undergoing high-risk elective coronary angioplasty. This form of support also permits transport of the patient to the operating room in a stable condition after an unsuccessful angioplasty. The complications are mostly related to cannula removal and can be minimized by the use of a proper technique. Although the ultimate role of this new technique remains to be completely defined, it appears that it will expand the patient population for whom coronary interventions can be applied.

Initial report of the National Registry of Elective Cardiopulmonary Bypass Supported Coronary Angioplasty

Relative contraindications to coronary angioplasty have been large amounts of jeopardized myocardium and poor left ventricular function. To prevent possible hemodynamic collapse after balloon occlusion or acute vessel closure in such high risk patients, a cardiopulmonary bypass system capable of providing up to 6 liters/min output was employed prophylactically. This technique, termed supported angioplasty, results in reductions of preload and afterload and allows prolonged balloon inflations in critical coronary vessels. A National Registry of 14 centers performing elective supported angioplasty was formed to collate the initial experience with high risk patients. Suggested indications were ejection fraction less than 25% or a target vessel supplying more than half the myocardium, or both. During 1988, the data from 105 patients (mean age 62 years) undergoing supported angioplasty were entered into the Registry. This group included 20 patients whose disease was deemed too severe to permit bypass surgery and 30 patients who had dilation of their only patent coronary vessel. Seventeen patients had stenosis of the left main coronary artery and 15 underwent dilation of that vessel. Chest pain and electrocardiographic changes occurred uncommonly despite prolonged balloon inflations. During the trial, there was a progressive change from cutdown insertion to percutaneous insertion of the circulatory support cannulas. The angioplasty success rate was 95% for the 105 patients, who underwent an average of 1.7 dilations per patient. Morbidity was frequent (41 patients), in most cases due to arterial, venous or nerve injury associated with cannula insertion or removal, or both.(ABSTRACT TRUNCATED AT 250 WORDS)

Reversal of chronic regional myocardial dysfunction (hibernating myocardium) by synchronized diastolic coronary venous retroperfusion during coronary angioplasty

A 62 year old man with previous myocardial infarction, an occluded right coronary artery and a 90% stenosis of the left anterior descending coronary artery underwent angioplasty with the support of coronary venous retroperfusion of arterial blood during the procedure. In two of four angioplasty balloon dilations of the left anterior descending coronary artery, synchronized diastolic retroperfusion of the coronary veins with arterial blood was applied to protect the severely dysfunctioning myocardium from additional ischemia. Two-dimensional echocardiography was used to monitor and quantitate alterations in left ventricular function. Retroperfusion of arterial blood resulted in immediate improvement in ischemic zone wall motion despite the totally occluded artery during balloon dilation. Echocardiographic images recorded after angioplasty showed a marked improvement in contraction of the previously dyskinetic segments, with changes similar to those seen during balloon dilations with synchronized diastolic coronary venous retroperfusion. Thus, in this patient, viability of chronically dysfunctioning myocardium could be demonstrated by the improvement in regional wall motion during retroperfusion. This technique could eventually be of value to elucidate the anatomic location of viable myocardium while maintaining adequate left ventricular systolic function during coronary artery interventions in the catheterization laboratory.

Emergency cardiopulmonary bypass support in patients with cardiac arrest in the catheterization laboratory

Cardiac arrest in the catheterization laboratory is fatal if unresponsive to advanced cardiac life support (ACLS). Seven patients not responding to ACLS following cardiac arrest in the catheterization laboratory underwent percutaneously instituted cardiopulmonary bypass support. Cardiac arrest occurred following abrupt closure postcoronary angioplasty in three patients, during cardiogenic shock in three patients, and during diagnostic angiography in one patient. Cardiopulmonary bypass was instituted 10-45 min (mean, 21 min) following the onset of cardiac arrest. Flows on bypass ranged from 4.0 to 5.2 liter/min. Mean blood pressure ranged from 70 to 110 mm Hg on bypass. Six of the seven patients regained consciousness after the institution of bypass. Acid-base balance was normalized in all patients. Coronary bypass surgery was subsequently performed in three patients and coronary angioplasty in two. Four patients survived. One patient died following coronary bypass surgery. Two patients, who were not suitable candidates for revascularization, expired. Total bypass time was 1.5-8.5 hr (mean, 2.7 hr). At a mean follow-up of 6 months, all four survivors are alive and asymptomatic or NYHA class 1. We conclude that cardiopulmonary bypass support 1) can stabilize patients following cardiac arrest in the catheterization laboratory, 2) can facilitate emergency coronary angioplasty or transfer to the operating room for coronary bypass surgery, and (3) can improve survival in patients unresponsive to ACLS when instituted early following cardiac arrest in the catheterization laboratory.

Percutaneous cardiopulmonary bypass support in high-risk patients undergoing percutaneous transluminal coronary angioplasty

Fifty-one consecutive patients in whom percutaneous cardiopulmonary bypass support was instituted to enhance the safety of high-risk elective coronary angioplasty were studied. All patients had a low ejection fraction, a large amount of viable myocardium perfused by the target artery(s) or both. Thirty-five men and 16 women, mean age 63 years, with Canadian Cardiovascular Society class III angina (23 patients) or class IV (28 patients) were studied. There was a history of myocardial infarction in 45 (88%), bypass surgery in 14 (27%) and congestive heart failure in 17 (33%). Forty-six (90%) had impaired left ventricular function. Twenty (39%) had an ejection fraction of less than or equal to 25%. Left main stenosis was present in 9 (18%), 3-vessel disease in 48 (94%) and 2-vessel disease in 2 (4%). Twenty (39%) were considered at a prohibitive risk for bypass surgery (14 were turned down for surgery). Bypass was instituted percutaneously with flows ranging from 2 to 5 liters/min (mean 3.6). Angioplasty was successful in 115 of the 117 lesions attempted with the culprit vessel dilated in all. Dilatation of the only remaining vessel was performed in 14 (27%). Inflation times up to 10 minutes were well tolerated. Bypass was discontinued after a mean bypass time of 37 minutes. Hemostasis was achieved by external clamp compression in 50. There were 3 hospital deaths unrelated to bypass. Patient follow-up at 2 to 8 months (mean 4.9) disclosed 1 late death, 31 (66%) asymptomatic patients, 12 (26%) patients in class I and 4 patients (9%) in class II. Thus, this study demonstrates the safety and efficacy of percutaneous bypass support in selected patients undergoing high-risk coronary angioplasty.

Coronary sinus interventions: from concept to clinics

For almost 100 years the coronary venous system has attracted researchers as an access route to deprived myocardium. Different concepts have been tried and numerous experimental studies have been performed to evaluate whether coronary sinus occlusion, retroinfusion, and retroperfusion of arterial blood via the coronary sinus represent an effective treatment of myocardial ischemia. The early successful studies of Drs. Gregg, Eckstein, Beck and others led to the application of the Beck II procedure, a permanent retroperfusion technique used in the early 1950s to revascularize patients with diffuse atherosclerosis. The coronary sinus route was also used to retroperfuse blood intraoperatively during opening of the heart to keep the heart beating. Excessive mortality and severe side effects, such as myocardial edema and hemorrhage, and an insufficiently advanced technology resulted in the temporary demise of the coronary sinus approaches. Recently, however, the need to improve myocardial protection, despite enormous advances in coronary bypass surgery and interventional cardiology, has resulted in renewed interest in the coronary sinus as an access route to deprived myocardium. Advances in technology such as percutaneous catheter techniques have improved access to the coronary venous system and allowed for a physiological adaptation of coronary sinus retroperfusion techniques. Today the three major coronary sinus techniques, i.e., sychronized retroperfusion (SRP), retroinfusion of cardioplegia during cardiac arrest and retroinfusion of pharmaceutical agents in the normal working heart, and pressure-controlled intermittent coronary sinus occlusion (PICSO) have been documented as providing superior protection of jeopardized myocardium in selected subsets of patients. All of these techniques currently are under clinical consideration; the retroinfusion of cardioplegia has already found wide clinical acceptance, SRP and PICSO have only recently been tested in first clinical trials. Due to a vast resurgence in interest, it is desirable to survey the results obtained with each of these techniques, to discuss the pathophysiology and mode of action of coronary sinus interventions, and to assign them a place in the perspective of conventional therapies. Furthermore, an attempt will be made to weigh individual coronary sinus techniques against each other, discuss in which clinical settings each of them may be most effective, and define issues facing current development.

Selective perfusion of ischemic myocardium during coronary venous retroinjection: a study of the causative role of venoarterial and venoventricular pressure gradients

Coronary venous retroinjection is often associated with preferential distribution of flow to ischemic myocardium. The purpose of this study was to define the mechanism of such retrodistribution of flow. In 24 anesthetized open chest dogs, Monastral blue dye (10 ml) was injected by way of a balloon catheter in the distal great cardiac vein as a marker for retrograde flow distribution. The injection rate (0.6 to 2.4 ml/s) was adjusted such that systolic pressure in the anterior interventricular vein ranged between 60 and 85 mm Hg. In 11 dogs with no ischemia and normal myocardial perfusion pressure (96 +/- 8 mm Hg), no myocardial staining occurred despite retrograde filling of epicardial veins. One minute after occlusion of the left anterior descending coronary artery, dye injections caused selective staining of the cyanotic area in 15 of 18 episodes, sparing the normal myocardium within the zone of retroperfused veins. In five dogs, with the arterial pressure less than 55 mm Hg, retroinjection resulted in homogeneous staining of all the myocardium drained by the retroperfused veins. Selective staining of the ischemic myocardium caused by retroinjection was associated with the following pressure gradients: during systole from the anterior interventricular vein to the occluded coronary artery, 31 to 58 mm Hg, and during diastole from the retroperfused veins to the left ventricular chamber, 9 to 28 mm Hg. There was no diastolic venoarterial gradient in the ischemic myocardium. In normal myocardium, retroinjection did not reverse the arteriovenous pressure gradient. In conclusion, retrograde flow is primarily directed to myocardium with low anterograde perfusion pressure. Selective retrograde penetration of acutely ischemic myocardium can thus be achieved by a mechanism consistent with the development of venoarterial and venoventricular pressure gradients.