Chronic hypoxia induces adaptive metabolic changes in neonatal myocardium.

The effect of chronic hypoxia on neonatal myocardial metabolism remains undefined. With a new neonatal piglet model, we determined changes in myocardial metabolism during global ischemia after chronic hypoxia. Five-day-old piglets (N = 30) were randomly assigned to two groups and exposed to an atmosphere of 8% oxygen or to room air for 28 days before they were killed. Left ventricular myocardium was then analyzed at control and at 15-minute intervals during 60 minutes of global normothermic ischemia to determine high-energy phosphate levels, glycogen stores, and lactate accumulation. Time to peak ischemic myocardial contracture was measured with intramyocardial needle-tipped Millar catheters as a marker of the onset of irreversible ischemic injury. Results showed an initially greater level of myocardial adenosine triphosphate in the hypoxic group (27 +/- 1.2 vs 19 +/- 1.8 micromol/gm dry wt, p = 0.001) and a delay in adenosine triphosphate depletion during 60 minutes of global ischemia compared with the control group. Initial energy charge ratios (1/2 adenosine diphosphate + adenosine triphosphate/adenosine monophosphate + adenosine diphosphate + adenosine triphosphate) were also greater in the hypoxic group (0.96 +/- 0.01 vs 0.81 +/- 0.04, p = 0.01) and remained so throughout global ischemia. Initial glycogen stores were greater in the hypoxic group (273 +/- 13.3 vs 215 +/- 14.7 micromol/gm dry weight, p = 0.02) when compared with the control group. Lactate levels in the hypoxic group were initially higher (19.1 +/- 6.4 vs 8.9 +/- 3.1 micromol/gm dry weight, p = 0.001) compared with control levels and remained elevated throughout 60 minutes of ischemia. Time to peak ischemic contracture was prolonged in the hypoxic group (69.5 +/- 1.8 vs 48.9 +/- 1.4 minutes, p = 0.001) compared with the controls group. These data show that chronic hypoxia results in significant myocardial metabolic adaptive changes, which in turn result in an improved tolerance to severe normothermic ischemia. These beneficial effects are associated with elevated baseline glycogen storage levels and an accelerated rate of anaerobic glycolysis during ischemia.

Direct mechanical ventricular actuation for cardiac arrest in humans. A clinical feasibility trial.

Direct mechanical ventricular actuation (DMVA) is a non-blood-contacting method of biventricular cardiac massage which may be applied expediently for total circulatory support. The purpose of this study was to assess the feasibility of DMVA application for patients suffering refractory cardiac arrest. Following informed consent, DMVA was applied in 22 patients. Vascular access for hemodynamic monitoring was possible in only 12 patients, whose outcomes serve as the basis for this report. The mean age of the patients was 48.2 +/- 4.2 years (seven men; five women). The average time from witnessed cardiac arrest to DMVA application was 81 +/- 9 minutes. Application took less than two minutes from the time of skin incision and resulted in immediate hemodynamic improvement. Systolic and diastolic blood pressures averaged 78 +/- 4 and 41 +/- 4 mm Hg, respectively, with a mean cardiac output of 3.14 +/- 0.18 L/min during a mean of 228 +/- 84 minutes of circulatory support (range, 25 minutes to 18 hours). In selected cases the device was temporarily removed for 2 to 3 minutes and open-chest cardiac massage (OCCM) performed at similar compression rates. DMVA increased arterial pressures 65 percent and cardiac output 190 percent compared to OCCM. Initial arterial pH (7.12 +/- 0.04) improved by the time the device was removed (7.24 +/- 0.05). Serum lactate levels decreased from 18.0 +/- 2.3 mumol/L to 14.9 +/- 2.9 mumol/L. Four patients were successfully defibrillated: two had inadequate cardiac function and died within 1 h, and two were successfully resuscitated, but later died from cardiac failure and respiratory insufficiency. Another patient regained normal neurologic function during DMVA and was successfully bridged to cardiopulmonary bypass for emergent coronary artery bypass grafting, but died later from myocardial infarction. There were only two complications: (1) a cardiac laceration during pericardiotomy (1/22 patients); and (2) a ventricular rupture during OCCM (1/22). No complication resulted from the device. We found DMVA to be a feasible method for acute cardiovascular stabilization in victims suffering refractory cardiac arrest. Human clinical trials employing earlier DMVA application are required to determine its resuscitative potential.

Improved donor myocardial recovery with a new lazaroid lipid antiperoxidant in the isolated canine heart.

U74006F is a new 21-amino steroid (lazaroid) that prevents lipid peroxidation without glucocorticoid or mineralocorticoid side effects. Reperfusion injury has been reduced by the addition of various free radical scavengers and antiperoxidants to the reperfusate. To assess the effect of U74006F on reperfusion of donor hearts subjected to prolonged hypothermic ischemia, 21 isolated canine hearts were divided into three groups: control (group 1), drug (2 mg/kg) injected into the oxygenated blood perfusate immediately before 4 hours of preservation (group 2), and drug (2 mg/kg) injected 1 hour before heart isolation and again 15 minutes before reperfusion (group 3). After control left ventricular function studies (with an intraventricular balloon) and biopsy for high-energy phosphates and dry/wet ratios, the hearts were arrested with cold cardioplegia and cooled for 4 hours then reperfused for 3 hours. Left ventricular work was calculated by systolic and diastolic pressure curves, which showed a better return of function in group 3 hearts (1625, 2150, and 3493 mm Hg/ml in groups 1, 2, and 3, respectively, at 180 minutes of reperfusion; p = 0.02). This was likely the result of improved diastolic compliance in group 3. Dry/wet ratios showed increased tissue edema in all hearts at the end of reperfusion. Although high-energy phosphate concentrations were not different between groups, adenosine was best preserved in group 3 (p = 0.03), suggesting reduced washout of this precursor. In conclusion, administration of U74006F before preservation and reperfusion may be useful for donor heart protection.

Mechanical cardiac actuation achieves hemodynamics similar to cardiopulmonary bypass.

The disadvantages of blood/interface interactions and difficult installation are common to current circulatory support devices. Direct mechanical ventricular actuation (DMVA) is a method of biventricular cardiac massage that avoids contact of blood with various surfaces. The purpose of this study was to compare hemodynamic responses and organ perfusion between DMVA and cardiopulmonary bypass (CPB). Twenty adult mixed-breed dogs randomized for DMVA or CPB were anesthetized with alpha-chloralose. During 4 hours of ventricular fibrillation, animals received either DMVA or CPB with aortic arch perfusion (90 to 120 ml/kg/min), bicaval venous return, and full left ventricle venting. Hemodynamics and organ perfusion were assessed by multivariant analysis of variance with repeated measures. Blood flow was similar to normal sinus rhythm (control) with either method; however, average CPB flows (control, 110%) were increased significantly over DMVA flows (control, 75%) (p = 0.016). The resulting mean arterial pressures were significantly greater during DMVA (control, 66%) compared to CPB (control, 49%) (p = 0.0011). Radiolabeled microspheres were the measure of organ perfusion during sinus rhythm and at 2 and 4 hours of circulatory support. Myocardial blood flow was equal to control in all regions during DMVA; CPB resulted in increased flows to the left ventricular epicardium, septum, and right ventricle. DMVA generated significantly greater flows to the renal cortex. All other organs demonstrated similar perfusion with either method. However, CPB displayed declining cerebral flows at 4 hours compared to DMVA (42% vs 55% control, respectively). Overall, DMVA provided hemodynamic stability equal to that of CPB. Rapid application and avoidance of blood/surface contact make DMVA a favorable method of temporary circulatory support.

Current morbidity, mortality, and survival after bronchoplastic procedures for malignancy.

The number of patients reported to have undergone bronchoplastic procedures has increased nearly fourfold in the past decade. These techniques represent excellent surgical therapy for patients with benign endobronchial lesions, traumatic airway disruptions, or tumors of low-grade malignant potential, and for select patients with surgically resectable lung cancer. Eighty-nine percent of bronchoplastic procedures are performed for malignancy. We reviewed 1,915 bronchoplastic procedures for carcinoma reported over the past 12 years to determine the incidence of complications and survival. Complications included local recurrence (10.3%), 30-day mortality (7.5%), pneumonia (6.7%), atelectasis (5.4%), benign stricture or stenosis (5.0%), bronchopleural fistulas (3.5%), empyema (2.8%), bronchovascular fistulas (2.6%), and pulmonary embolism (1.9%). Results were further stratified into sleeve lobectomy and sleeve pneumonectomy groups. Five-year survivals for stage I, II, and III carcinoma were 63%, 37%, and 21%, respectively. Sleeve lobectomy for carcinoma extends surgical therapy to select patients with complication rates comparable to pneumonectomy and long-term survival similar to that for conventional resections.

Surgical treatment of automatic atrial tachycardias.

From 1979 to 1989, 18 patients were seen in the Electrophysiology Service, Duke University Medical Center, with automatic atrial tachycardia. There were 8 male and 10 female patients with a mean age of 28.1 +/- 2.9 years. Electrophysiological mapping localized automatic foci to right atrial sites (14 patients) and left atrial sites (4 patients). Depending on origin of the focus, patients were further diagnosed as having either chronic ectopic atrial tachycardia or inappropriate sinus tachycardia. Of the 15 patients with chronic ectopic atrial tachycardia, 6 responded to medical treatment; in 9, the tachycardia was not adequately controlled. Six of them were referred for surgical intervention. All 3 patients with inappropriate sinus tachycardia underwent operative therapy. In the surgical group of patients with chronic ectopic atrial tachycardia, all 6 had a tachycardia-induced cardiomyopathy with ejection fractions ranging from 14% to 27% (mean ejection fraction, 21% +/- 2.7%). Surgical techniques used (alone or in combination) included an isolation procedure in 1 patient, cryoablation in 4 patients, and excision of atrial appendages or portions of atrial free walls in 7. Normal sinus rhythm developed in all surgical patients except 1 patient who had intractable congestive heart failure preoperatively and died of this condition and stroke. The overall success rates for medical and surgical therapy were 33.3% and 88.9%, respectively (p less than 0.01). Long-term follow-up was possible for 7 (87.5%) of 8 patients 3 to 7 years after operation. All patients with chronic ectopic atrial tachycardia were cured, but only 1 of 3 patients with inappropriate sinus tachycardia was in sinus rhythm.(ABSTRACT TRUNCATED AT 250 WORDS)

Argon beam coagulation compared with cryoablation of ventricular subendocardium.

The Argon Beam Coagulator uses radiofrequency energy to excite argon gas that may be used for ventricular ablation. The effects of power level and number of applications of the Argon Beam Coagulator were compared wtih cryothermia. Ten mongrel dogs underwent cardiac extirpation. The endocardial surfaces of 5 hearts were used for the creation of lesions using the Argon Beam Coagulator at five power levels with either one or two applications. Five hearts were used for endocardial and epicardial lesions using cryothermia (15-mm-diameter probe at -70 degrees C) for 1, 2, 3, or 4 minutes. The Argon Beam Coagulator lesions showed an increase in depth with increasing power levels (2.25 +/- 1.05 mm at 50 W to 6.64 +/- 0.75 mm at 150 W) and number of applications (maximum depth of 6.64 +/- 0.75 mm with one application, 11.2 +/- 1.1 mm with two applications). Cryothermia lesions were similar in depth regardless of duration or site of application (range, 6.1 to 10.2 mm). Both techniques resulted in homogeneous and well-demarcated lesions. These data show that the Argon Beam Coagulator results in discrete endocardial lesions, which may be created quickly and reproducibly. This may be a useful alternative for the operative ablation of endocardial scar in the treatment of ventricular tachycardia.

First successful bridge to cardiac transplantation using direct mechanical ventricular actuation.

Currently available ventricular assist devices are technically difficult to implant, require continuous anticoagulation, and are associated with hemorrhagic and thromboembolic complications. Direct mechanical ventricular actuation is a biventricular assist device that can be applied in 3 to 5 minutes through a left anterior thoracotomy and has no direct blood contact or need for anticoagulation. The present study was designed to determine the effects of direct mechanical ventricular actuation in total biventricular circulatory support. Cardiogenic shock refractory to standard therapy developed in 2 patients awaiting cardiac transplantation. Direct mechanical ventricular actuation was applied and provided immediate hemodynamic stabilization in both. All inotropic agents and intraaortic balloon support were then discontinued. Fifty-six hours of circulatory support bridged the first patient to successful cardiac transplantation without complication. The patient is alive and well more than 1 year later without incident of infection or rejection. The second patient suffered cardiac arrest and required closed chest cardiopulmonary resuscitation before device application. After 45 hours of support, it was determined that irreversible neurologic injury had occurred and direct mechanical ventricular actuation was discontinued. Neither patient's native heart exhibited any histologic evidence of device-related trauma. Direct mechanical ventricular actuation has undergone limited clinical investigation since its original description 25 years ago, but in these initial trials, the device has proved effective. The concept of mechanically actuating the ventricles appears to be a valuable, yet under-utilized method of total circulatory support.

Pulmonary mucormycosis: results of medical and surgical therapy.

Mucormycosis is an opportunistic fungal infection that commonly begins by invading the respiratory tract. The purpose of the present study was to define the clinical presentation of pulmonary mucormycosis and to evaluate current treatment regimens. Thirty patients treated at our institution and 225 cases reported in the literature were reviewed. For the combined groups, the mean age at presentation was 41 +/- 21 years and associated medical conditions included leukemia or lymphoma (37%), diabetes mellitus (32%), chronic renal failure (18%), history of organ transplantation (7.6%), or a known solid tumor (5.6%). The in-hospital mortality was 65% for patients with isolated pulmonary mucormycosis, 96% for those with disseminated disease, and 80% overall. The mortality in patients treated surgically was 11%, significantly lower than the 68% mortality in those treated medically (p = 0.0004). The most common causes of death were fungal sepsis (42%), respiratory insufficiency (27%), and hemoptysis (13%). Pulmonary mucormycosis has a high mortality; however, antifungal agents appear to improve survival. In addition, surgical resection may provide additional benefit to patients with pulmonary mucormycosis confined to one lung.

Pulsatile versus nonpulsatile reperfusion improves cerebral blood flow after cardiac arrest.

Cardiopulmonary bypass using nonpulsatile flow (NF) is currently advocated for treating refractory cardiac arrest. Although the heart can be revived using cardiopulmonary bypass support, the brain must recover if such therapy is to be considered successful. Previous studies have demonstrated that pulsatile flow (PF) reperfusion can improve neurologic outcome compared with NF reperfusion after cardiac arrest. The purpose of this study was to assess cerebral perfusion and oxygen consumption during either PF or NF reperfusion after cardiac arrest. Dogs (n = 22) underwent a 15-minute cardiac arrest followed by 1 hour of either PF or NF reperfusion. Microsphere techniques were used to assess cerebral perfusion and oxygen consumption at 3, 15, and 60 minutes of reperfusion. Mean arteriovenous gradients and total brain flows were similar in both groups. However, cerebral oxygen consumption was significantly improved at 3 minutes of reperfusion with PF versus NF (1.8 +/- 0.3 versus 0.9 +/- 0.3 mL O2.dL-1.min-1, respectively; p < 0.05). These results were coincident with improved gray-to-white flow ratios at 3 minutes of PF versus NF reperfusion (5.2 +/- 1.0 versus 2.0 +/- 0.3, respectively; p < 0.05). There were no statistically significant differences in brain perfusion variables by 15 minutes of reperfusion. However, a relative hyperemia was exhibited at 15 minutes of NF versus PF reperfusion, which suggests nutrient flow was insufficient during early NF versus PF reperfusion. In conclusion, PF reperfusion can better restore cerebral blood flow and oxygen consumption than can NF reperfusion.(ABSTRACT TRUNCATED AT 250 WORDS)

Direct mechanical ventricular actuation: a review.

Direct mechanical ventricular actuation (DMVA) is a non-blood contacting method of circulatory support that can be rapidly instituted for resuscitation. DMVA is superior to conventional methods (open and closed-chest cardiac massage) in providing reliable cardiovascular stabilization for resuscitation following cardiac arrest. Furthermore, DMVA has important advantages including rapid application, technical simplicity, and avoidance of blood contact compared to other resuscitation devices (cardiopulmonary bypass and blood pumps). This review summarizes laboratory and clinical applications of DMVA.

Polytetrafluoroethylene hemoaccess site infections.

The advent of permanent hemodialysis access has made possible the use of chronic hemodialysis in patients with end-stage renal disease. Although autogenous arteriovenous fistulae remain the conduit of choice, their construction is not always feasible. Prosthetic grafts made of polytetrafluoroethylene (PTFE) are typically the second-line choice for hemoaccess. However, these grafts suffer from decreased rates of patency and an increased number of complications. Although thrombosis is the most common complication, infection of PTFE grafts accounts for a significant number of hospitalizations and uses a large amount of healthcare resources. In this monograph, we address infectious complications of PTFE hemoaccess grafts and present a review of the recent literature.

Cardiac pathology following resuscitative circulatory support. Direct mechanical ventricular actuation versus cardiopulmonary bypass.

Cardiopulmonary bypass (CPB) is currently advocated for treating refractory cardiac arrest. Direct Mechanical Ventricular Actuation (DMVA) is an alternative method that does not contact the blood and has other unique advantages for providing resuscitative circulatory support, including rapid application and relative technical simplicity. The purpose of this study was to assess pathologic changes in the heart following resuscitation with either CPB or DMVA. Dogs (n = 22) received 1 hr of CPB (n = 11) or DMVA (n = 11) following a 12.5 min cardiac arrest. All deaths [4/11 (CPB) vs. 2/11 (DMVA), p = 0.31] occurred during the initial 24 postoperative hours. At 7 days, survivors had magnetic resonance imaging to determine cardiac ejection fraction [46% (CPB) vs. 51% (DMVA), p = 0.39], as well as the presence of cardiac wall motion abnormalities [50% (CPB) vs. 33% (DMVA), p = 0.57] and gross cardiac lesions [17% (CPB) vs. 17% (DMVA)]. The survivor's hearts were then extirpated, fixed, and examined for gross lesions [2/7 (CPB) vs. 0/9 (DMVA), p = 0.17]. Transmural sections of the anterior and posterior papillary muscles were histologically evaluated. The severity and extent of epicardial fibrosis and focal myocyte necrosis did not differ between groups. These data demonstrate that DMVA does not cause more myocardial trauma than CPB when used to provide resuscitative circulatory support. Therefore, the unique attributes of DMVA may improve resuscitation outcome in patients who suffer refractory cardiac arrest, without additional risk of cardiac injury.

Intraoperative timing may provide criteria for use of post-cardiotomy ventricular assist devices.

When surgeons consider ventricular assist devices (VADs) for post-cardiotomy support, unnecessary delays and early use can have detrimental effects on patient outcome. The authors analyzed the timing of intraoperative events in all patients receiving post-cardiotomy VAD support at their institution during a 2-1/2 year period (N = 17). They used ability to wean patients from VAD support as a measure of outcome. Neither preoperative risk factors nor the timing of distinct intraoperative events (e.g., cross-clamp time, total bypass time, delay to VAD) significantly differed between those patients able (Group I, n = 9) and those unable (Group II, n = 8) to be weaned from VAD support. The authors did find, however, that the time intervals from completion of the cardiac procedure to insertion of either an intra-aortic balloon pump (time to IABP) or VAD (time to VAD) were predictive of outcome when normalized to the duration of the cardiac procedure (DCP). [Time to IABP]/DCP ratios of < 1.0 versus > 1.0 (p = 0.02) and [time to VAD]/DCP ratios of < 2.5 versus > 2.5 (p = 0.10) each segregated Group I and II patients, respectively. Appropriate timing criteria for VAD insertion may be predicted during surgery by consideration of the duration of the cardiac procedure. This approach may attenuate tendencies to delay VAD use without leading to premature VAD insertion in the post-cardiotomy setting.

Prophylactic implantable defibrillator patches in patients at high risk for malignant ventricular dysrhythmias.

The indications for prophylactically placing implantable cardioverter defibrillator (ICD) patches at cardiac surgery are unclear. Over the past 4 years, 25 patients have undergone placement of prophylactic ICD patches at the authors' institution. Indications were severe coronary artery disease with ventricular tachycardia (VT) (n = 14) or ventricular fibrillation (VF) (n = 2), and left ventricular aneurysmectomy and/or subendocardial resection (SER) (n = 9). Four patients died in-hospital, three from refractory VT (which could be transthoracically cardioverted until recurrence was unremitting), and one from congestive heart failure (CHF). All of the remaining 21 patients underwent post-operative electrophysiologic studies (EPS), and 12 had inducible VT (8/14 CABG, 4/7 SER). Eight of the 12 inducible patients had generators implanted, whereas 3 patients were controlled medically. One patient refused generator implantation and died at home from sudden cardiac death, and one initially non-inducible patient required late ICD generator placement, yielding a total of nine patients who received generators. All nine of these patients are currently alive at 14 +/- 7 months follow-up, and five have subsequently received appropriate ICD discharges. In patients undergoing cardiac surgery considered preoperatively or intraoperatively to be at increased risk for VT/VF and too unstable for preoperative EPS, prophylactic ICD patches should be considered.

Sensing lead insulation fractures following implantable cardioverter-defibrillator placement.

The insulation of implantable cardioverter-defibrillator (ICD) sensing leads can fracture, causing device malfunction. Before March 1989, 32 patients received silicon CPI K54 sensing leads. Of these, four (12.5%) experienced discharges while in sinus rhythm. Subsequent ICD interrogation revealed that abdominal muscular contraction triggered discharges in two patients, but no abnormal sensing could be elicited in the other two patients. These patients subsequently were found to have sensing lead fractures. All sensing leads appeared intact radiographically but at reoperation were found to have insulation fractures 2-4 cm from the pin. The mechanism of fracture appears to be related to chronic intermittent lead compression between the anterior abdominal wall musculature and the ICD generator. Since March 1989, 134 patients have undergone ICD implantation with improved CPI 4312 sensing leads; none of these patients has experienced sensing lead fractures. The insulation of these leads is 2.5 times thicker than the K54 (1.0 mm vs. 0.4 mm). Patients with K54 sensing leads are at an increased risk for sensing lead insulation fracture and merit close surveillance. More importantly, sensing lead fractures must be considered in patients who present with inappropriate ICD discharges.

Myocardial tolerance to mechanical actuation is affected by biomaterial characteristics.

Direct mechanical ventricular actuation (DMVA) uses a pressure regulated heart cup, fabricated from silicone rubber (SR) for mechanical massage of the heart. Because DMVA has demonstrated potential for long-term circulatory support, investigations are currently exploring the use of more durable materials for fabricating DMVA heart cups. This study assessed the acute effects of heart cups fabricated from SR versus polyurethane (PU) on the myocardium. Dogs (n - 18) received DMVA for 4 hr of ventricular fibrillation (VF) using either SR (n = 10) or PU (n = 8) cups. Microspheres were used to determine perfusion during sinus rhythm (control) and at 2 and 4 hr of support. After support, myocardial biopsies were assayed for high energy phosphate content. Results demonstrated that PU cups required relatively frequent adjustments in drive line parameters that were likely due to material softening during PU cup support. Both PU and SR cups achieved similar hemodynamics during 4 hr of support. Myocardial perfusion, however, demonstrated a marked hyperemia at 4 hr of PU versus SR cup support. Regional high energy phosphate content was significantly decreased in hearts supported by PU versus SR cups. These results suggest that the relatively compliant characteristics of SR materials are important for achieving effective DMVA support without injuring the myocardium.

Prolonged total circulatory support using direct mechanical ventricular actuation.

Direct mechanical ventricular actuation (DMVA) is a unique, non blood contacting method for biventricular cardiac assist. Although DMVA has successfully provided cardiac assist for more than 7 days in humans, with long-term survival, its potential for long-term circulatory support has not been adequately investigated. DMVA has not been studied in the large ruminants commonly used to evaluate support devices. To develop a large animal experimental model of prolonged total circulatory support using DMVA, Suffolk sheep (n = 10) underwent sterile instrumentation for hemodynamic and chemistry monitoring. After baseline values were obtained, a left lateral thoracotomy and pericardotomy were performed. Upon electrical ventricular fibrillation (VF), DMVA was begun and the thoracotomy closed. Total circulatory support was continued until mean arterial pressure (MAP) persisted below 50% of the baseline value for more than 1 hr, with a goal of 7 days' support. Mean duration (plus or minus the standard deviation [SD]) of circulatory support was 65.9 +/- 56.8 hr (range, 10-168 hr). Pressors were not used during DMVA support. The subject supported for the maximal time (7 days) was defibrillated into sinus rhythm. No CK-MB fraction was greater than 1%, suggesting that DMVA, even with prolonged application during VF, does not result in myocardial injury. Blood urea nitrogen and creatinine levels indicate renal function was preserved. The model described represents the longest period any animal has been supported in VF using DMVA. This new model will be useful in determining what limitations, if any, exist to the prolonged use of DMVA for circulatory support.