Hemodynamic predictors for weaning patients from ventricular assist devices (VADs)

In order to find hemodynamic parameters that can accurately predict whether patients can be successfully weaned from ventricular assist devices (VADs), we studied data from 17 patients supported with Pierce-Donachy VADs [11 left VAD (LVAD); 6 right VAD (RVAD)] following cardiogenic shock for periods from 1.3 to 22 days (mean 5.4). Myocardial recovery was determined by daily measurements of "pump on pump off" parameters, and the data from the 8 LVAD patients and 3 RVAD patients whose hearts recovered were compared to the data from those whose did not. In this study, daily pump on pump off hemodynamic measurements were found to be predictive of success for weaning patients from VADs. In particular, the most significant predictors were: increases in mixed venous oxygen saturation, cardiac index, mean arterial pressure and ventricular ejection fraction, as well as decreases in atrial pressures. An index for measuring hemodynamic function with the VAD off is proposed, as are models of recovery. Seventy percent of the patients weaned from VADs survived, indicating that patients appropriately weaned from VAD support have a reasonable chance for survival.     

Total heart replacement with dual centrifugal ventricular assist devices

In an ovine feasibility study, we implanted two HeartMate-III centrifugal ventricular assist devices (VADs) for total heart replacement. With cardiopulmonary bypass support, both ventricles were transected at the atrioventricular groove, preserving a rim of ventricular tissue. The atrioventricular valves were excised, and the aorta and pulmonary artery were transected above the ventriculoarterial valves. An interatrial septal window was created by excising the foramen ovale. The VADs' sewing rings were attached to the left and right ventricular remnants, respectively. Outflow grafts were anastomosed to the aorta and pulmonary artery. The left VAD operated continuously at 4,500 rpm. Right VAD speed increased from 2,000 to 4,500 rpm in 500 rpm increments. Outflow graft flow, pressure, oxygen saturation, and shunt direction were recorded. The pulmonary artery to aortic ratio of flow and pressure increased from 0.26 and 0.15 (at 2,000 rpm) to 1.21 and 0.53, respectively (at 4,500 rpm). The interatrial shunt, which was right to left at lower right VAD speeds, progressed to bidirectional, then to left dominant as right VAD speed increased. Outflow-graft oxygen saturation was reflective of the shunt direction. In this acute experiment, total heart replacement with continuous flow VADs satisfactorily balanced left and right ventricular flows and preserved the physiologic circulatory response.     

Pathological findings in explanted prosthetic heart valves from ventricular assist devices

AIMS: Ventricular assist devices (VADs) are now a mainstay in the management of patients with end-stage heart failure. An important consideration in the long-term durability of these devices is the structural integrity of the prosthetic valves. Herein, we report the morphological findings in inflow and outflow explanted bioprostheses from seven such devices. METHODS: The porcine bioprostheses (n = 7; HeartMate, Novacor) were examined from inflow and outflow valve conduits. Cusp tears were assessed on gross examination. Tissues were then processed for histology and graded for pannus, thrombus, and calcification. Immunohistochemistry was performed using anti-CD68 (macrophages), CD45 (leukocytes) and CD31 (endothelial cells) antibodies to assess inflammation. RESULTS: There was no evidence of infection, host tissue growth, or calcification in either the inflow or the outflow valves in any case. A mild-to-moderate mononuclear cell 'deposit' was present on all porcine bioprostheses, largely on the non-flow surface of the valve cusps. In the case of the longest implant (HeartMate, duration 567 days), a significant mononuclear cell infiltrate was seen on the flow surface, the non-flow surface, as well as the base of the cusp tissue. Variably sized cusp tears were found in all inflow porcine bioprostheses at and beyond 3 months post-implantation, with the longest duration implant showing multiple tears. No tears were identified in the outflow valves. Histology revealed thrombus deposition in all inflow and outflow porcine valves. In addition, inflow valve cusps were characterised by the presence of longitudinally running 'cystic' spaces, which seem to increase in size with increasing implant duration. CONCLUSION: Bioprosthetic heart valves in VADs show significant changes which appear to correlate with duration post-implantation. These changes suggest that haemodynamic forces and the inflammatory reaction may play a significant role in the long-term durability of the porcine bioprostheses in these devices.     

The isoprostane 8-epi-PGF(2alpha) is a valuable indicator of oxidative injury in human heart valves.

To date, little information is available concerning oxidative injury in human cardiac valves. Therefore, we sought to investigate whether the isoprostane, 8-epi-PGF(2alpha), a novel oxidative stress marker, is localized in aortic and pulmonary valves derived from explanted hearts of patients suffering from idiopathic dilative cardiomyopathy (IDC). By using semiquantitative immunohistochemistry, we demonstrated that 8-epi-PGF(2alpha) is localized in both valves with pulmonary valves accumulating more of this isoprostane compared to aortic valves (36.69+/-12.04% vs. 31.54+/-11.49%, P<.05). These results were confirmed by a radioimmunoassay (RIA) analysis showing a similar, but not significant, difference between the two valves (288.50+/-72.18 pg/mg protein in the pulmonary valves and 267.30+/-58.77 pg/mg protein in aortic valves, P=.09). Considering the data presented in this study, we suggest that 8-epi-PGF(2alpha) is a valuable indicator of oxidative injury in human semilunar valves.     

Human factors issues in ventricular assist device recipients and their family caregivers

Ultimately, for ventricular assist devices (VADs) to be acceptable as permanent alternatives to heart transplantation, patients' and their families' satisfaction with specific features and risks of VADs must be addressed. Of 42 eligible patients who received VADs between February of 1996 and December of 1998, we interviewed 37 patients (17 Novacor, 18 Thoratec, 2 with both devices) and 20 of their primary family caregivers about device related concerns and reactions. Demographic and health related correlates of respondents' concerns were examined. Eleven patients discharged from the hospital with the VAD in place were then reinterviewed 1 month after discharge. At baseline, patients' general perceptions of the VAD were positive, although 22-52% reported specific concerns, including most often worry about infection (52%), difficulty sleeping due to the position of the driveline (52%), pain at the driveline exit site (46%), worry about device malfunction (40%), and being bothered during the day by device noise (32%). The prevalence of most concerns rose with duration of VAD support. Caregivers' perceptions did not differ significantly from patients' perceptions. Outpatients were somewhat more concerned than inpatients about device noise and risk of stroke, but were markedly less concerned about infection. Across all patients, higher levels of device related concerns were correlated with more physical functional limitations and more psychological distress, and reduced quality of life. Demographic characteristics and device type were not uniformly related to device concerns.     

In vivo performance evaluation of the innovamedica pneumatic ventricular assist device

We evaluated the short- and mid-term in vivo performance of the Innovamedica ventricular assist device (VAD), a new, low-cost, paracorporeal, pneumatically actuated, pulsatile blood pump. We implanted the VAD in six healthy sheep by inserting the stainless-steel inflow cannula into the left ventricular apex and suturing the outflow graft to the descending thoracic aorta. The anesthetized animals were supported for 6 hours, and pump performance, hemodynamic parameters, and hemolysis were monitored. The pump maintained a blood flow of 4.4 ± 0.8 L/min and an arterial blood pressure of 76 ± 15 mm Hg. At 6 hours, the plasma free hemoglobin concentration was 5.11 ± 0.6 mg/dl (baseline value, 4.52 ± 0.7 mg/dl). The VAD was easy to implant and deair and performed well during the 6 hour period. After successful short-term results, we similarly implanted the VAD in two healthy sheep for 30 days. The animals reached the scheduled end point without device-related problems. Postmortem examination of the explanted organs revealed small infarcted areas in the kidneys of one animal, but renal function was unaffected; the animal also had two thrombi (3 and 7 mm) on the outlet valve. This device may offer a simple, economical alternative to currently available VADs.     

Mechanical circulatory support devices for bridge to heart transplantation, bridge to recovery, or destination therapy

Both the ventricular assist device (VAD) and the total artificial heart (TAH) have been effective in supporting circulation of end-stage cardiac patients and in bridging to heart transplantation. However, because of a shortage of donor hearts and age limitations, destination therapy with the completely implantable VAD has also been started. The totally implantable TAH in the United States is in the final stage of development and will go into preclinical trials in 2004. In Japan, heart transplantation has been re-instituted since last year, but because of shortages of donor hearts the waiting time prior to transplantation is fairly long. To date, six heart transplantations have been carried out, of which four have been bridged transplantations, using extracorporeal or implantable VADs. With the extracorporeal VADs, patients cannot be discharge home, which increases the hospital expenses. With the implantable VADs such as Novacor and HeartMate imported from the USA, patients can be discharged home, but major threats with these devices are thromboembolic, complications and infection. These devices are also fairly large, being designed for 80-kg patients, and are thus difficult to implant in patients of 50 to 60 kg, including women. Because of these limitations, there is a strong clinical demand for a compact, high-performance, implantable, permanent-use VAD. This paper addresses the current status of the artificial heart research and development program at the Tokyo Medical and Dental University, which was started in May 1999.     

Blood biocompatibility analysis in the setting of ventricular assist devices

Ventricular assist devices (VADs) are increasingly applied to support patients with advanced cardiac failure. While the benefit of VADs in supporting this patient group is clear, substantial morbidity and mortality occur during the VAD implant period due to thromboembolic and infective complications. Efforts at the University of Pittsburgh aimed at evaluating the blood biocompatibility of VADs in the clinical, animal, and in vitro setting over the past decade are summarized. Emphasis is placed on understanding the mechanisms of thrombosis and thromboembolism associated with these devices.     

Neuropathology associated with ventricular assist devices: an autopsy series of 33 patients

The ventricular assist device (VAD) is a mechanical pump that has been shown to be an effective modality of cardiac support in patients with heart failure refractory to pharmacologic intervention and who are awaiting cardiac allograft transplantation. Neuropathologic findings in these patients have not been well described. We retrospectively reviewed 2,632 autopsy reports (between 1990 and 2000) and found 64 patients who received VADs. Of these 64 patients, brain and spinal cord tissue was available for review in 33 patients (25 males and eight females; age range, 4 to 69 years; mean age, 52 years). The study group was composed of these 33 patients. Ventricular assist devices were in place from one to 603 days (mean 49 days). Twenty-five patients had left VAD, three had right VAD, and five had biventricular VADs. Brain weights ranged from 928 g to 1,740 g (mean 1,325 g). The most common central nervous system pathologic findings included infarct (N = 23; 70%), acute neuronal necrosis (N = 22; 67% focal and N = 1; diffuse anoxic encephalopathy), hemorrhage (N = 14; 42%), and herniations (N = 7; 21%). Two patients had no neuropathologic findings at autopsy. Cause of death was central nervous system-related in eight patients (24%) including six with massive parenchymal hemorrhage and herniations, one with brainstem infarction, and one with air embolism (radiographically diagnosed). The most common causes of death in the remaining 25 patients included sepsis (n = 10; 30%), pneumonia (n = 4; 12%), and embolic events with widespread infarcts (n = 4; 12%). The most common neuropathologic findings in patients with VAD were related to ischemia and infarction. In a significant subset of patients, central nervous system pathology, particularly hemorrhage with herniation, was the primary cause of death. Ann Diagn Pathol 5:67-73, 2001.     

The effect of aortic valve incompetence on the hemodynamics of a continuous flow ventricular assist device in a mock circulation

There is evidence that the incidence of aortic valve incompetence (AI) and other valvular pathologies may increase as more patients are submitted to longer periods of ventricular assist device (VAD) support. There is a need to better understand the mechanisms associated with the onset of these conditions and other possible complications related to the altered hemodynamics of VAD patients. In this study, the effect of AI on the hemodynamic response of continuous flow VAD (C-VAD) patients was measured in a mock loop over a range of pump speeds and level of native cardiac function. Our results showed that, in the presence of sufficient ventricular function, decreasing the C-VAD speed can allow a transition from series to parallel flow. Our study demonstrated that AI reduces the aortic pressure and flow when system impedance is unchanged. AI produces wasteful recirculation that substantially increases the pump work and decreases systemic perfusion, in particular during series flow conditions coupled with higher C-VAD speeds. The hematologic consequence of this regurgitant flow is a much higher exposure to shear for the blood, increasing the likelihood of hemolysis and thrombosis. While a certain degree of AI can be tolerated by a heart with good cardiac function, the consequences of AI for patients with VADs and poor cardiac function are much greater. Valve dysfunction in VAD patients may be related to structural changes in the tissue induced by altered biomechanics and excessive stress.     

Numerical comparison of hemodynamics with atrium to aorta and ventricular apex to aorta VAD support

We report the first attempt to study with numerical methods ventricular assist device (VAD) models and the effects of various inlet VAD cannulations, coupling physical explanations and numerical investigation conclusions with clinical research results. We compared the hemodynamic response with VAD support by using two distinct VAD-inlet cannulation configurations: left atrium to aorta and left ventricular apex to aorta. Impeller pump and displacement pump VADs are considered. Constant VAD flow rate and counterpulsation motion models are simulated. The native cardiovascular system is modeled using the concentrated-parameter method by considering the flow resistance, vessel elasticity, and inertial effect of blood flow in cardiovascular system individual segments. Impeller and displacement pump dynamic models are represented by corresponding inlet and outlet flow rate changes in the VADs. Results show that the two VAD inlet cannulation configurations produce similar cardiac response (flows, pressures, volumes), except that when the VAD flow approaches the 100% assisting condition, the peak left ventricular systolic pressure and diastolic volume increase slightly in the left atrial cannulation, whereas they drop markedly in the left ventricular apex cannulation, suggesting increased ventricular wall tension and ventricular dilatation in the left atrial cannulation and that hemodynamically the left ventricular apex cannulation is more advantageous.     

Paracorporeal ventricular assist device as a bridge to transplant candidacy in the era of implantable continuous-flow ventricular assist device

Ventricular assist devices (VADs) have long been used as bridge to transplant therapy (BTT). Nipro-Toyobo paracorporeal pulsatile-flow VAD (nt-VAD) was the only device available until April 2011, when implantable continuous-flow VADs (cf-VADs) became available. Although cf-VADs are central to BTT, nt-VAD remains a necessary option. We aimed to clarify the role of nt-VAD in an era of increasing cf-VAD use. We retrospectively reviewed patients who underwent VAD implantation at the National Cerebral and Cardiovascular Center from May 2011 to March 2013. Characteristics were compared between the nt-VAD and cf-VAD groups. Twenty-nine patients (mean age 37.7 ± 11.1 years, 23 males) underwent VAD implantation. Fifteen patients initially received nt-VADs, although 4 were converted to cf-VADs. Of these 15 patients, 3 were too small for cf-VADs and 2 needed bilateral ventricular support. The remaining 10 patients received nt-VADs (7 patients at INTERMACS level 1 and 3 at level 2). The nt-VAD group patients had significantly more preoperative mechanical circulatory support and were in a more critical condition before VAD implantation than the cf-VAD group. The 2-year survival rate was not significantly different. Despite the critical conditions of nt-VAD patients, their overall survival is not statistically inferior to that of cf-VAD patients. nt-VAD is a good option as a BTC for the patient with urgent and critical condition.     

Blood biocompatibility analysis in the setting of ventricular assist devices

Ventricular assist devices (VADs) are increasingly applied to support patients with advanced cardiac failure. While the benefit of VADs in supporting this patient group is clear, substantial morbidity and mortality occur during the VAD implant period due to thromboembolic and infective complications. Efforts at the University of Pittsburgh aimed at evaluating the blood biocompatibility of VADs in the clinical, animal, and in vitro setting over the past decade are summarized. Emphasis is placed on understanding the mechanisms of thrombosis and thromboembolism associated with these devices.     

Improved biocompatibility of heparin surface-coated ventricular assist devices.

Heparin surface coated ventricular assist devices (VADs) and cannulas were evaluated in comparison to uncoated VADs in 10 bovine experiments (body weight 77 +/- 6 kg). All systems were primed with cristalloid solution. No systemic heparin was given. Left ventricular assist was started with a blood flow of 4.2 +/- 0.4 l/min and maintained over 6 hours. Besides hemodynamic monitoring, blood samples were taken at regular intervals for blood gas, hematological, biochemical and coagulation studies. All animals in the study group (coated) were assisted for the scheduled 6 hours without device failure. In the control group, however, total occlusion occurred in 1 VAD after 1 hour of left ventricular assist whereas the other 4 VADs remained functional throughout the protocol. Mixed venous oxygens saturation was preassist 56 +/- 12% for coated versus 63 +/- 11% for uncoated and the final value at 60 minutes after weaning was 58 +/- 16% versus 59 +/- 5% (NS). Mean hematocrit dropped from a baseline value of 33 +/- 4% for coated versus 29 +/- 8% for uncoated to 29 +/- 7% versus 30 +/- 5% (NS) after 6 hours of assist. There was no significant difference between the baseline values (5.7 +/- 3.0 mumol/l for coated versus 4.6 +/- 3.1 mumol/l for uncoated) and the 6-hour values (3.8 +/- 3.7 mumol/l versus 7.6 +/- 6.4 mumol/l) for mean plasma hemoglobine. The normalized platelet levels dropped after 10 minutes of assist to 91 +/- 21% for coated versus 94 +/- 49% for uncoated (NS) and 89 +/- 29% versus 65 +/- 44 at 6 hours (NS).(ABSTRACT TRUNCATED AT 250 WORDS)     

Left ventricular outflow tract obstruction associated with chronic ventricular assist device support

Favorable long-term patient outcome after insertion of a left ventricular assist device (LVAD) as a bridge to recovery or destination therapy for the treatment of end-stage cardiomyopathy is adversely affected by pathophysiologic changes affecting the heart. Alterations in the native aortic valve apparatus, specifically aortic valve cusp fusion, is an example of such a phenomenon and may especially affect patients in cases of bridge to recovery, a rare but reported event. A retrospective review of the last 33 LVAD placements at our institution was conducted, including reviews of operative reports and pathologic examinations of the native hearts. Seven hearts were found to have varying degrees of aortic valve cusp fusion after chronic LVAD support (63-1, 339 days). Five of these patients had native aortic valves, and two had bioprosthetic valves. The left ventricular outflow tracts in two patients were surgically occluded at the time of LVAD insertion. Aortic valve cusp fusion occurs in roughly 25% of patients on chronic LVAD support. This phenomenon may prove to be clinically significant by creating a potential source of emboli and infection. In addition, in the case of myocardial recovery, left ventricular outflow tract obstruction could limit parallel flow and produce suprasystemic ventricular pressures that in turn would elevate left ventricular end diastolic pressures. The latter may contribute to further myocardial injury, ultimately limiting the ability of an otherwise recovered heart to be weaned from LVAD support.     

Retrospective analysis of adverse events in preclinical ventricular assist device experiments

Ventricular assist devices (VADs) are a widely acceptable therapeutic option for patients with end-stage heart failure. Data from preclinical animal trials provide important information about the efficacy, biocompatibility, and functioning of these devices. However, animal experiments have several limitations that may significantly affect outcomes even after implanting a properly functioning device. This retrospective analysis was performed to analyze, at explant, adverse events associated with VADs during the animal phase of testing. Data were collected from 141 VAD experiments performed in cattle that had received 11 different types of devices. A total of 294 adverse events were documented. Inflow cannula problems were cited in 79 implants (56%) and pump-related issues in 55 cases (39%). Adverse events associated with connectors were found in 53 implants (38%). Adverse events directly having to do with the animals represented only 17 implants (12%) in this series. These findings suggest that preclinical testing needs to place greater emphasis on the entire VAD system. In support of these results, human clinical data seem to indicate that actual VAD failure is rarely a cause for premature device removal.     

Inflammation and infection in nine surgically explanted Medtronic Freestyle stentless aortic valves.

BACKGROUND: The Medtronic Freestyle valve is fixed in glutaraldehyde at zero pressure on the cusps and treated with alpha-amino oleic acid. This valve reportedly has excellent clinical and hemodynamic results, but little has been reported about its long-term pathology. METHODS AND RESULTS: Nine Freestyle valves explanted between 2003 and 2005 were reviewed to assess the reasons for bioprosthesis failure (six implanted at our institution). All valves were examined in detail, using histochemistry and immunohistochemistry to identify the cellular response. One Freestyle valve, explanted for mitral valve endocarditis on the fifth postoperative day, was excluded from analysis. Average implant duration was 52.8+/-35.5 months. Four valves were explanted for infective endocarditis, three for aortic insufficiency, two for aortic stenosis with cusp calcification seen in five valves, pannus and thrombus in all valves and a chronic inflammatory reaction involving the xenograft arterial wall seen in eight of nine valves. This was associated with significant damage to the porcine aortic wall in seven cases, and cusp myocardial shelf damage in six cases. CONCLUSIONS: In this series of valves, we found (1) infective endocarditis; (2) pannus, thrombus, and calcification; and (3) unusual and significant inflammatory reaction and aortic tissue damage, which could by itself lead to aortic incompetence.     

Effects of left ventricular assist device support and outflow graft location upon aortic blood flow

Although continuous flow (CFVAD) and pulsatile (PVAD) ventricular assist devices (VADs) are being clinically used, their effects upon aortic blood flow as a measure of overall blood distribution remain unclear. The objective of this study was to compare the effects of CFVAD and PVAD support for ascending (AscA) and descending (DA) aorta outflow cannulation upon mean aortic blood flow and waveform morphology. Six experiments were conducted in a normal, acute calf model, in which an inflow cannula was implanted in the left ventricle apex and outflow cannulae were anastomosed to both the AscA and DA. Flow probes were placed around the pulmonary artery, pump outflow, brachiocephalic trunk, and aorta proximal and distal to the DA outflow. For each acute experiment, calves received randomly selected levels of VAD support (0-100% of cardiac output) and pump failure (VAD off and outflow cannula unclamped) for each of four randomly selected test conditions: (1) PVAD and AscA, (2) PVAD and DA, (3) CFVAD and AscA, and (4) CFVAD and DA. Regardless of pump type or support level, proximal and distal aorta mean flows were lower (p < 0.05) for DA compared with the AscA. No differences in mean aortic flows between pump types at either outflow graft location were discerned. Differences in morphologic features of blood flow waveforms between PVAD and CFVAD were observed. During simulated pump failure, retrograde aortic blood flow in both the aortic arch and DA was observed. Partial ventricular suction was also observed during the greatest levels of CFVAD support and suggested pronounced effects upon both the right and left ventricle. Collectively, these findings imply that VAD outflow location may have an important role in patient response and recovery. Investigation of the long-term pathophysiologic responses to pump type and outflow location is ongoing.     

Hodgkin disease in adult and juvenile groups from two different geographic regions in Brazil: characterization of clinicopathologic aspects and relationship with Epstein-Barr virus infection

We analyzed clinicopathologic data, immunophenotype, and Epstein-Barr virus (EBV) status in 96 cases of Hodgkin disease (HD) in juveniles (younger than 20 years) and adults (20 years or older) from 2 distinctive states in Brazil. We studied 34 juvenile (group 1) and 16 adult (group 2) cases from Ceara and 31 juvenile (group 3) and 15 adult (group 4) cases from S?o Paulo. Ceara has a socioeconomic profile similar to a developing country; S?o Paulo is in better economic condition. Mixed cellularity (MC) was the major histologic subtype among groups 1 (22 [65%]), 3 (21 [68%]), and 4 (7 [47%]); nodular sclerosis (NS) was more frequent in group 2 (8 [50%]). EBV infection was observed in 61 cases (64%), including the following (among others): group 1, MC, 22 (65%) and NS, 4 (12%); group 2, NS, 3 (19%) and MC, 2 (12%); group 3, MC, 16 (52%) and NS, 1 (3%); and group 4, MC, 7 (47%). There was predominance of EBV+ HD cases in group 1 compared with group 3. HD in Brazilian patients is highly associated with EBV infection, but geographic differences reflect histologic subtypes and age distribution.     

Cost-benefit analysis of extended antifungal prophylaxis in ventricular assist devices

This report defines the cost and benefit of extended antifungal prophylaxis in ventricular assist device (VAD) patients (pts). Extended antifungal prophylaxis is defined as prophylaxis with fluconazole or nystatin that is given until pts are extubated and off antibiotics. These data are compared with that obtained from earlier VAD patients who only received anti-fungal drugs for documented fungal colonization or infection. Thirty-six patients had HeartMate (n = 15) or Thoratec (n = 21) VADs between 1989 and 1997. Cultures positive for fungus (n = 52 cultures) were obtained from 16 of 36 patients (44% of patients). Forty-three fungal cultures were in the preprophylaxis and nine in the postprophylaxis era. There was one death attributable to fungal sepsis in the preprophylaxis era and none in the postprophylaxis era. The total cost of antifungal drugs in the preprophylaxis era was $3,840 over 1,498 patient days (PD) (mean $2.56 per PD), versus $70,670 over 1,525 PD in the postprophylaxis era (mean $46.34 per PD). Extended antifungal prophylaxis was not cost effective in VAD patients at this institution. However, short-term perioperative antifungal prophylaxis was not addressed by this study. We are now using short-term antifungal prophylaxis with fluconazole and nystatin in VAD patients because of the potential for serious morbidity and mortality that is associated with fungal device infections. A future analysis will determine the usefulness of this change in strategy.