Endocardial excision versus encircling endocardial ventriculotomy. A comparison of effects on ventricular structure and function.

Although endocardial excision and encircling endocardial ventriculotomy are being performed in patients with extensive triple-vessel disease and compromised ventricular function, long-term effects of the operative intervention on structure and function of the left ventricle have not been determined. These procedures were performed in healthy dogs in three groups: control (ventriculotomy alone), endocardial excision, and encircling endocardial ventriculotomy (five dogs per group). Six weeks later, through a left thoracotomy, an arterial line, left atrial line, and Swan-Ganz catheter were inserted. Cardiac output measurements permitted calculation of left ventricular stroke work index, and gated nuclear ventriculograms permitted calculation of left ventricular volume indices. Myocardial performance (stroke work index/end-diastolic volume index relation), systolic elastance (systolic blood pressure/end-systolic volume index relation) and diastolic pressure-volume relationship (left atrial pressure/end-diastolic volume index relation) were determined from volume loading studies. In the endocardial excision group, the left atrial pressures were increased at similar end-diastolic volumes (p less than 0.05 by performance and systolic elastance were similar in the three groups. On completion of hemodynamic studies, the hearts were excised. Gross and light microscopic examination showed that the inner layer of myocardium was scarred in the area of intervention after both endocardial excision and encircling endocardial ventriculotomy. In neither group was there significant morphologic change elsewhere in the myocardium. Both endocardial excision and encircling endocardial ventriculotomy have little effect on long-term structure and function when performed in healthy canine hearts.

Myocardial metabolism and ventricular function following cold potassium cardioplegia.

Transient alterations in myocardial metabolism and ventricular function were observed after elective coronary bypass grafting despite apparently adequate intraoperative protection with cold potassium cardioplegia. Ninety patients had serial hemodynamic measurements and coronary sinus catheters inserted. Thirty-three patients had thermodilution coronary sinus flow catheters inserted to measure coronary sinus blood flow and to evaluate the myocardial utilization of oxygen and lactate. Nuclear ventriculograms were performed in 43 patients to assess ventricular function. Cardiac index fell after discontinuation of cardiopulmonary bypass and then rose between 2 and 24 hours postoperatively. Myocardial oxygen consumption steadily increased during this period. Myocardial lactate production reverted to lactate extraction 30 minutes after reperfusion. Reactive hyperemia was present during the first 10 minutes after cross-clamp release, and coronary sinus blood flow increased gradually during the first 24 hours postoperatively. The response to the stress of volume loading (the infusion of 250 to 500 ml of a colloid solution) and atrial pacing (at a rate of 110 beats/min) was evaluated 2 to 4 hours postoperatively (EARLY) and between 4 to 6 hours postoperatively (LATE). Volume loading resulted in a decrease in lactate extraction EARLY and an increase LATE (EARLY: -0.07 +/- 0.35 mmol/L; LATE: 0.08 +/- 0.32 mmol/L, mean +/- standard deviation not significant). Atrial pacing resulted in a decrease in lactate extraction EARLY and an increase LATE (EARLY: -0.11 +/- 0.34 mmol/L; LATE: 0.14 +/- 0.36 mmol/L, p less than 0.05). Diastolic compliance (the relation between the end-diastolic volume index) decreased between EARLY and LATE. Systolic function (the relation between the systolic blood pressure and the end-systolic volume index) and myocardial performance (the relation between the left ventricular stroke work index and the end-diastolic volume index) were unchanged. Ejection fraction correlated inversely with the end-diastolic volume index and did not represent an independent index of contractility. After elective coronary bypass grafting and cold crystalloid cardioplegia, myocardial metabolism recovered slowly. Hemodynamic stresses should be avoided in the early postoperative period to prevent progressive ischemic injury.

Surgical aspects of regional myocardial blood flow and myocardial pressure.

The surgical technique of cardiopulmonary bypass with either an empty beating or an empty fibrillating ventricle produces marked changes in the regional blood flow and oxygen demand of the left ventricle. This paper describes the changes which occurred in the regional perfusion of both the normal and the hypertrophied left ventricle during these conditions and relates them to the known changes in oxygen demand. It also correlates the changes in flow with the measurable changes in myocardial tissue pressure-systolic when the heart is beating and continuous when fibrillating. The various types of filbrillation had identical effects on both regional tissue pressure and regional flow. The subendocardial blood supply was adequate or more than adequate under each of these conditions so long as the coronary perfusion pressure was maintained at an adequate level. A low perfusion pressure during ventricular fibrillation of any type led to a marked reduction in flow to the subendocardial portion of the left ventricle: The presence of ventricular hypertrophy accentuated this danger.

Comparison between antibiotic irrigation and mobilization of pectoral muscle flaps in treatment of deep sternal infections.

Between January of 1978 and December of 1983, 41 patients developed deep sternal infections with mediastinitis after cardiac operations. Between January of 1978 and December of 1981, 19 of these patients were treated with débridement, primary wound closure, and mediastinal antibiotic irrigation (Group I). Between January of 1982 and December of 1983, 22 patients were treated with débridement, open "clean" packing, and delayed wound closure by the technique of pectoral muscle flap mobilization, which preserves the thoracoacromial pedicles and the pectoral humeral attachments (Group II). The purpose of this study was to compare the results of the treatment of deep sternal infections after cardiac operations with these two techniques. The perioperative hemodynamic, operation, functional, and pathological profiles of both groups of patients were the same. The cosmetic and functional results were the same in both groups as were shoulder girdle and torso mobility. We conclude that either technique is equally effective in the management of patients in whom the serious complication of deep sternal infection with mediastinitis develops after cardiac operation, and we now recommend débridement and pectoral muscle flap closure in one stage.

Diltiazem cardioplegia. A balance of risk and benefit.

Calcium channel blockers may prevent myocardial injury during cardioplegia and reperfusion. A prospective, randomized trial was instituted to evaluate the hemodynamic and myocardial metabolic recovery in 40 patients undergoing elective aorta-coronary bypass with either diltiazem in crystalloid potassium cardioplegia (n = 20) or crystalloid potassium cardioplegia (n = 20). In a preliminary trial, doses between 150 and 250 micrograms/kg reduced the period of heart block after cross-clamp removal (90 +/- 110 minutes) from that found with higher doses and improved myocardial metabolism. In the randomized trial, diltiazem cardioplegia (150 micrograms/kg) produced coronary vasodilatation during cardioplegia and produced less reactive hyperemia during reperfusion. Myocardial oxygen extraction was lower and myocardial lactate production was less after diltiazem cardioplegia during reperfusion. Tissue adenosine triphosphate and creatine phosphate concentrations were preserved better after diltiazem cardioplegia. The postoperative creatine kinase MB levels were less (p less than 0.05) after diltiazem cardioplegia, which indicated less myocardial injury. Postoperative volume loading demonstrated that systolic function (the relation between systolic blood pressure and end-systolic volume index) was depressed after diltiazem cardioplegia compared to crystalloid cardioplegia, but cardiac index was higher because afterload (mean arterial pressure) was lower and preload (end-diastolic volume index) was higher. Diltiazem cardioplegia preserved high-energy phosphates, improved postoperative myocardial metabolism, and reduced ischemic injury after elective coronary bypass. However, diltiazem was a potent negative inotrope and produced prolonged periods of electromechanical arrest. Diltiazem cardioplegia may be of value in patients with severe ischemia but should be used with caution in patients with ventricular dysfunction, and a dose-response relation must be established at each institution before clinical use.

Comparison of alternative cardioplegic techniques.

Cold potassium cardioplegia provides adequate protection for coronary bypass operations, but severe coronary stenoses limit cardioplegic delivery to ischemic regions. The traditional technique delivers cardioplegic solution into the aortic root during the performance of distal anastomoses. The proposed alternative technique constructs proximal as well as distal anastomoses during a prolonged cross-clamp period, but permits more uniform cooling. The two techniques were compared in a prospective concurrent trial of 45 patients undergoing elective coronary bypass grafting. The traditional technique was employed in 26 patients (Group A) and the alternative technique in 19 patients (Group B). In both groups, 700 to 1,000 ml of a crystalloid cardioplegic solution was infused into the aortic root after application of the aortic cross-clamp. In Group A (traditional technique), 500 ml was infused into the aortic root after each distal anastomosis. In Group B (alternative technique), cardioplegic solution was administered through the vein graft after each distal anastomosis, and a proximal anastomosis was constructed after distal anastomoses to the most ischemic regions to permit continued cardioplegic delivery to these regions. The cross-clamp period was shorter in Group A than in Group B (44 +/- 15 versus 60 +/- 18 minutes, p less than 0.01), but the mean temperature in the most ischemic region was warmer (Group A, 19 degrees +/- 3 degrees C; Group B, 15 degrees +/- 3 degrees C, p less than 0.05). The postoperative CK-MB was higher in Group A (Group A, 47 +/- 36; Group B, 21 +/- 9 IU/L, p less than 0.01). Cardiac lactate production persisted longer in Group A (Group A, 4 +/- 1; Group B, 1 +/- 1 hours postoperatively, p less than 0.05). Volume loading 4 hours postoperatively produced a similar increase in left atrial pressure and cardiac index in both groups. In response to volume loading, Group A patients produced lactate, but Group B patients extracted lactate (change in cardiac lactate extraction: Group A, -1.7 +/- 2.3; Group B, +2.5 +/- 5.1 mg/dl, p less than 0.05). The construction of proximal as well as distal anastomoses during a prolonged cross-clamp period permits more uniform cooling and immediate reperfusion. This alternative technique resulted in less injury (CK-MB release) and more rapid recovery of myocardial metabolism.

Increased risk of urgent revascularization.

To determine the risk factors for operative mortality and morbidity, we performed a prospective analysis of 1,980 patients undergoing isolated coronary artery bypass operations between 1982 and 1984. The operative mortality was 3.5%, and the incidence of perioperative myocardial infarction was 8.6% and low output syndrome, 12.0%. Stepwise logistic regression identified sex, preoperative left ventricular ejection fraction, and the urgency of operation as independent risk factors for postoperative mortality. Urgent revascularization was performed in patients with unstable angina refractory to maximal medical therapy. In these patients the operative mortality was 8.5%. Independent risk factors of postoperative morbidity, in addition to sex, ejection fraction, and urgent revascularization, included a previous bypass procedure, age, and New York Heart Association functional class. Unstable angina unresponsive to medical therapy contributed significantly to the operative risk. Interventions to reduce perioperative ischemic injury, such as improved methods of myocardial protection, may improve the results in high-risk patients.

Optimal postoperative volume loading.

Intravenous infusions are required to maintain ventricular preload after uneventful coronary bypass operation. During the early postoperative period, when myocardial metabolic recovery is incomplete, volume loading is intended to stabilize ventricular function and metabolism and to prevent progressive ischemic injury. This study attempts to define the optimal preload for both metabolism and performance. Thirty-seven patients recovering from elective coronary bypass operations and cold potassium cardioplegia underwent volume loading with whole plasma. The initial response (VLA) from a low left atrial pressure (LAP = 7.3 +/- 3.3 mm Hg) was compared with the subsequent response (VLB) from a higher filling pressure (LAP = 10.9 +/- 2.7 mm Hg). Both VLA and VLB produced a similar increase in cardiac index, stroke work index, and end-diastolic volume index (EDVI), and a decrease in ejection fraction (measured by nuclear angiography). Myocardial lactate extraction increased with VLA, but myocardial lactate production resulted with VLB. A careful analysis of these volume loading studies suggested that myocardial performance and compliance were not altered in the early postoperative period. The decrease in ejection fraction with volume loading may have resulted from a combination of increased wall tension and decreased inotropic stimulation. After uneventful coronary bypass surgery, an LAP between 5 and 12 mm Hg corresponded to an EDVI between 30 and 80 ml/m2 and produced adequate cardiac index, stroke work index, and lactate extraction. A lower or higher preload did not improve function and resulted in abnormal metabolism.

Effects of postoperative hypertension and its treatment.

Hypertension following aorta-coronary bypass operations can contribute to myocardial ischemia. Nitroprusside therapy will reduce afterload, preload, and coronary perfusion pressure. Since both hypertension and its treatment can result in ischemic injury, nitroprusside must be carefully titrated to optimize cardiac function and metabolism. Thirty-one patients undergoing elective coronary bypass grafting were studied during a hypertensive episode (mean arterial pressure [MAP] = 119 +/- 18 mm Hg) and during nitroprusside therapy at an MAP of 97 +/- 11 mm Hg and at an MAP of 80 +/- 11 mm Hg (normotension). Nitroprusside also produced a significant (p less than 0.05) decrease in left atrial pressure (LAP), left ventricular end-diastolic volume index (EDVI) (stroke index divided by ejection fraction by nuclear angiography), stroke index, and stroke work index (SWI). Cardiac lactate extraction (LEx) and the ratio LEx/SWI increased (p less than 0.05) with the initial nitroprusside therapy, but lactate production resulted when the MAP was lowered to 80 mm Hg. Volume loading studies were performed during hypertension in four patients and during nitroprusside therapy in 15 patients. Neither performance nor compliance was significantly altered at an MAP of 97 mm Hg, but compliance decreased at normotension. Both hypertension and its treatment can result in inadequate myocardial metabolism. Nitroprusside should be titrated to maintain MAP between 90 and 100 mm Hg.

The response of diastolic myocardial tissue pressure and regional coronary blood flow to increased preload from blood, colloid, crystalloid.

An increase in the left ventricular end-diastolic pressure (LVEDP), resulting from the intravenous administration of blood, a colloid solution, or a crystalloid solution, was accompanied by an increase in both the inner:outer left ventricular wall flow ratio and the diastolic myocardial tissue pressure gradient. With a normal left ventricular end-diastolic pressure, the normal inner:outer flow ratio was just above unity and the minimum diastolic myocardial tissue pressure in the subepicardium was twice as high as in the subendocardium. When the left ventricular end-diastolic pressure was raised, the subepicardial tissue pressure rose earlier and to a higher degree than the subendocardial; thus a gradient which encourages subendocardial flow was increased. This paper correlates the changes in diastolic myocardial tissue pressure and in inner:outer flow ratios which occurred when eight dogs were transfused with blood, seven with colloid, and seven with lactated Ringer's solution. Regional diastolic myocardial tissue pressure was measured by the flow-cessation technique and regional flow by radioactive microspheres. The gradient in diastolic myocardial tissue pressure, which is a major contributory factor to adequate subendocardial perfusion under normal conditions, is of even greater importance when LVEDP is raised.

Optimal myocardial protection.

The low mortality and perioperative infarction rates for aortocoronary bypass (ACB) make them unsuitable for evaluating the adequacy of myocardial protection. Enzymatic and functional measurements were found to be sensitive and specific indicators of myocardial injury. A prospective concurrent study of 78 patients undergoing triple ACB was conducted to evaluate the effectiveness of three popular methods of myocardial protection. Group I (32 patients) had a single dose of cold (4 degrees C) potassium cardioplegic (CPC) solution infused inducing a mean myocardial temperature (MMT) of 31 +/- 4 degrees C/min. Group II (23 patients) had multiple doses of CPC solution 8nducing a MMT of 22 +/- 2 degrees C/min. Group III (23 patients) had intermittent anoxic arrest at a MMT of 28 +/- 1 degrees C. The groups were not randomized but had comparable clinical symptoms and catheterization findings. Serial measurements of cardiac specific creatine kinase (CK-MB) revealed a peak in enzymatic activity occurring 60 minutes following ACB. The highest CK-MB was significantly (P less than 0.01) lower in group II (25 +/- 8 IU/liter) than group I (50 +/- 8 IU/liter), or group III (68 +/- 14 IU/liter). Myocardial performance was evaluated after ACB by serially measuring left ventricular stroke work index (SW) and left atrial pressure (LAP) in response to volume loading. The rise in SW was significantly (P less than 0.01) greater in group II (3.0 +/- 0.7 gm.m/sq m/mm Hg) than in group I (1.4 +/- 0.7) or group III (1.8 +/- 0.9). The highest SW attained was higher (P less than .01) in group II (43 +/- 7 gm.m/sq m) than group I (19 +/- 6) or group III (34 +/- 8) at comparable LAP values (group I: 20 +/- 5 mm Hg; group II: 18 +/- 3; group III: 18 +/- 4). Post-operative clinical evaluation failed to differentiate among the three groups. The more sensitive indices, however, demonstrated the superiority of cold, multidose cardioplegia in providing optimal myocardial protection.

Identification and successful treatment of congenital microfistulas with the aid of directional Doppler.

Hitherto unidentifiable and therefore untreatable congenital microfistulas were detected with the aid of continuous-wave directional Doppler and a fine-beam pencil probe. The microfistulas formed part of the Klippel-Trenaunay (K-T) syndrome. The management of one case is reported in detail and in three others salient features are touched upon. Careful clinical and radiological examination failed to demonstrate any arteriovenous microfistulas. When Doppler ultrasound scanning was carried out, two discrete fistulas were discovered. Their extent and direction were mapped out accurately. Incisions were made directly over the markings displaying a pulsating capillary tuft of vessels. Further dissection exposed a feeding arteriole which was less than 1 mm. in diameter. Excision of the vascular malformations resulted in the cure of the patient. In another patient with the K-T syndrome in whom a cutaneous hemangioma involved the whole lower limb, in spite of a thorough and systematic search with a Doppler, no microfistulas could be demonstrated. It is suggested that all patients suffering from the K-T syndrome should be examined by Doppler ultrasound in the hope that microfistulas which elude radiodiagnostic techniques might be detected and treated surgically.

Comparison of conduits for leg revascularization.

The saphenous vein (SV) remains the conduit of choice for lower limb revascularization. When SV is unavailable, or unsuitable, two alternative conduits have been employed: gluteraldehydestablized human umbilical vein (HUV) and polytetrafluoroethylene (PTFE). In this study of the 218 patients who underwent lower limb revascularization, 3-year patency of 85 SV graft was 75% compared to 34% for the 66 HUV grafts and 33% for the 67 PFTE grafts. Three factors were found to independently influence patency: the indication for surgery, the site of the distal anastomosis, and the angiographic runoff. The SV group had significantly better patency than either HUV or PFTE in each of these subgroups. No consistent difference between HUV and PTFE was found. A risk score was obtained by assigning a value of 1 to 3 for each of the factors influencing patency--indication: 1 = claudication, 2 = rest pain, 3 = ischemic lesions; site: 1 = above knee (AK), 2 = below knee (BK), 3 = tibial; runoff 1 = good (two or three vessels), 2 = fair (one vessel), 3 = poor (no vessel). Patients with the lowest risk scores (3 to 4) had the best 3-year patency: SV, 78%; HUV, 44%; and PTFE, 48%. Patients with the highest risk scores (7 to 9) had the worst 3-year patency: SV, 68%; HUV, 32%; and PTFE, 28%. SVs had better patency under high- and low-risk conditions and remain the conduit of choice for lower limb revascularization. Both HUV and PTFE have equivalent and acceptable patency when SV is unavailable or unstable.

Transluminal dilation. An alternative?

In a prospective study, 384 peripheral arterial transluminal dilations were evaluated, using clinical and vascular laboratory criteria, and analyzed by the life-table method. The overall cumulative success rate was 58.9 +/- 3.3% (mean +/- SEM) after two years but was higher if the iliac segment was dilated, only one site was dilated, the clinical indication was claudication, the distal vessels were normal, the ankle-brachial BP ratio was more than 0.35, or if the patient was younger than 55 years. The complication rate was 3.9%. If the dilation failed, the symptoms were worse in 8% and the ankle-brachial BP ratio fell in 23%. After successful dilation in patients with normal distal vessels, residual claudication persisted in 34%, and the ankle-brachial BP ratio remained abnormal in 40%. Although the overall success rate of transluminal dilation was lower than for a comparable surgical procedure, dilation does have a role in the management of localized peripheral vascular disease.

Pericardial graft for intraoperative balloon pump insertion.

A technique that permits rapid insertion of an intraaortic balloon pump to support patients who cannot be successfully weaned from cardiopulmonary bypass is described. A pericardial patch is obtained while the sternum is still open, and is sutured to a common femoral arteriotomy. A 6-0 Prolene mattress stitch is inserted at the heel and continued along the sides. The ends are tied at the toe, and the same suture is used to construct a tube of pericardium over the balloon catheter. A single heavy silk suture is placed around the pericardial graft to prevent bleeding. The technique helps prevent thrombus and avoids infectious complications. It reduces the incidence of vascular complications and makes use of the Fogarty catheter after balloon removal unnecessary. We have used the method in 9 patients since November, 1979, without problems.

Ancrod anticoagulation for cardiopulmonary bypass in heparin-induced thrombocytopenia and thrombosis.

Heparin-induced thrombocytopenia and thrombosis was diagnosed in a 50-year-old man undergoing a repeat heart operation after heparinization led to microemboli and an eventual left transmetatarsal amputation. A third heart operation was aborted when anticoagulation with low molecular weight heparin produced intraoperative thrombi. The patient was referred to Toronto where ancrod (Arvin) was used to lower plasma fibrinogen level, allowing successful repair of a ventricular septal defect using cardiopulmonary bypass support. The patient made an uneventful recovery.

Cardioplegia for severe coronary artery disease: an improved technique using direct coronary artery infusion.

Conventional techniques of cardioplegic solution administration result in regional disparities in the level of myocardial protection in patients with severe coronary artery disease. This report describes a simple adjunct to conventional transaortic administration of cardioplegic solution, in which additional solution is introduced directly through the coronary arteriotomy used for the vein graft anastomosis. The supplemental infusate is delivered through s small-caliber flexible catheter. This technique permits effective perfusion of the coronary vascular bed distal to severely stenotic and occlusive lesions. Using an experimental model of physiologically significant coronary arterial stenosis, we compared the effectiveness of this technique with that of conventional techniques of cardioplegia.

Improved myocardial protection during a prolonged cross-clamp period.

Severe coronary stenoses limit delivery of cardioplegic solution to ischemic regions in patients undergoing bypass operations. A prospective randomized trial was undertaken to determine whether the construction of proximal as well as distal anastomoses during a prolonged cross-clamp period would provide more uniform cardiac cooling and better myocardial protection. Ninety-one consecutive patients undergoing elective coronary bypass operations were randomized into two groups. The long cross-clamp technique was used in 46 patients (Group 1), and a proximal anastomosis was constructed after each distal anastomosis. The short cross-clamp technique was employed in 45 patients (Group 2), and distal anastomoses were constructed during aortic occlusion. Cardiopulmonary bypass time was identical, but the cross-clamp period was longer in Group 1 (59 +/- 15 minutes versus 46 +/- 17 minutes in Group 2; p less than 0.001). The mean temperature in the most ischemic region was colder with the long cross-clamp technique (12.5 +/- 3.1 degrees C in Group 1 versus 14.8 +/- 3.2 degrees C in Group 2; p less than 0.01). The total amount of the myocardial isoenzyme of serum creatine kinase released was greater in Group 2 than in Group 1 (332 +/- 34 IU/L per hour in Group 1 versus 469 +/- 45 IU/L per hour in Group 2). Thirty-six patients had coronary sinus catheters inserted (18 patients in each group). Myocardial lactate extraction returned to normal sooner in the patients who had a long cross-clamp period; time to a normal lactate extraction was 0.8 +/- 0.8 hours in Group 1 versus 2.2 +/- 2.1 hours in Group 2 (p less than 0.001). Volume loading and atrial pacing 2 to 4 hours postoperatively produced a similar hemodynamic response in the two groups, but myocardial lactate extraction increased in Group 1 and decreased in Group 2 (p less than 0.05). The construction of proximal as well as distal anastomoses during a prolonged cross-clamp period produced more uniform cooling and improved myocardial protection.

The determinants of mortality and morbidity after multiple-valve operations.

The factors predictive of hospital mortality and morbidity after contemporary multiple-valve surgical procedures were identified to develop strategies to improve the results of such procedures. Preoperative, intraoperative, and postoperative information was collected prospectively on 90 consecutive patients undergoing surgical procedures between 1982 and 1984. The operative mortality was 5.6%, and the incidence of postoperative low-output syndrome was 16.7%. Multivariate logistic regression analysis identified tricuspid regurgitation (p less than .03, improvement-of-fit chi square) and the aortic valve lesion (p less than .03) as the independent predictors of postoperative complications (mortality or low-output syndrome). Patients with tricuspid regurgitation and right ventricular decompensation and those with aortic stenosis and left ventricular hypertrophy had limited ventricular functional reserve and faced an increased risk. Improved methods of myocardial protection may reduce the risk in these patients.

A comparison of volume loading and atrial pacing following aortocoronary bypass.

Although cold potassium cardioplegia provides adequate myocardial protection, transient hemodynamic and metabolic instability occasionally occurs after uncomplicated coronary bypass surgery. Two methods to increase cardiac output were compared 2 to 6 hours postoperatively in 24 patients recovering from elective coronary bypass operation. Volume loading increased cardiac index (CI) from 2.1 +/- 0.5 to 2.7 +/- 0.6 L/min/m2 by increasing left atrial pressure (LAP) from 8.6 +/- 3.6 to 13.0 +/- 4.1 mm Hg. Atrial pacing at a rate of 112 +/- 8 beats per minute increased CI from 2.4 +/- 0.5 to 2.7 +/- 0.8 L/min/m2 without a change in LAP. Ejection fraction by nuclear angiography did not change, but the calculated left ventricular end-diastolic volume index (stroke index/ejection fraction) increased with volume loading and decreased with atrial pacing--a decrease in diastolic compliance. Myocardial oxygen extraction did not change, but myocardial lactate extraction increased with volume loading and decreased with atrial pacing. Coronary sinus blood flow was measured in 5 patients and increased with both methods studied. Volume loading demonstrated that myocardial performance was normal and myocardial metabolism increased commensurate with the increase in work. Atrial pacing increased CI but resulted in anaerobic metabolism and a decrease in diastolic compliance. Volume loading rather than atrial pacing will improve CI without producing ischemia in the early postoperative period.