Summary of MedSun Reports Describing Adverse Events With Ventricular Assist Devices
MedSun: Newsletter #35, April 2009

A ventricular assist device (VAD) is a mechanical pump that helps a heart that is too weak to pump blood through the body. It is sometimes referred to as "a bridge to transplant" since it can help a patient survive until a heart transplant can be performed. A VAD is used to aid the pumping action of a weakened heart ventricle (a major pumping chamber of the heart). VADs were originally intended for short-term use to support failing hearts until donor hearts became available. Some VADs are now used for long-term (destination) therapy in severe heart failure patients who are not candidates for heart transplants [1].

Over the past 2 years, MedSun has received 19 adverse event reports associated with the Ventricular Assist Device. The reports represent 4 manufacturers: 15 reports for Thoratec Corp, 2 reports for Cardiac Assist Inc., 1 for Berlin Heart Inc., and 1 for ABIOMED Inc. The reports were submitted by 10 hospitals between March 6, 2007 and March 6, 2009. The most frequently reported device problems were:

• Battery problems (4)
• Pump failure (4)
• Electrical issue (3)
• Tubing Crack/Hole (2)
• Percutaneous lead fracture (1)
• Human factors (1)
• Device interaction with an Implantable Cardiac Defibrillator (1)
• Filling malfunction in the stroke volume limiter (1)
• Cannula disconnection (1)
• Console leak (1)

A total of 4 reports involved a patient death. Nine patients had no injuries or the injuries were unknown. The patient injuries listed below were reported in 6 of these 15 reports.

• 1 patient had delay of treatment
• 5 patients required additional surgery/removal of device

Of the reports that listed patient age, 1 had patient age listed as less than 21 years and 18 had patient age listed as greater than 21 years. Of the reports that listed patient gender, a total of 6 reports involved female patients, and a total of 12 reports involved male patients.

These MedSun reports contributed to FDA awareness of VAD device problems. The following recall, describing problems with the percutaneous lead, was for the Thoratec HeartMate II Left Ventricular Assist Device.

Recall number: Z-0496-2009
A recall was initiated on October 24, 2008 by Thoratec Corp for the HeartMate II Left Ventricular Assist System (LVAS), Model/Catalog Number 1355 and Model/Catalog Number 102139, all serial numbers.

The reason for the recall was that wear and fatigue of the percutaneous lead connecting the HeartMate II Left Ventricular Assist System (HM II LVAS) blood pump with the external system controller could result in damage that had the potential to interrupt pump function and require a reoperation to replace the pump.

FDA is currently looking into the other listed device problems described above.


[Note: The reports have been edited for clarity]
_________________







Ventricular Assist Devices
Device Device Identifiers Event Description
Thoratec/ HeartMate Catalog #: 1270/ Lot # 42526
The patient had the device implanted one year ago. Recently, the patient had low flows on device, was clammy and pale with Low BP, and had an acute drop in hemoglobin/hematocrit. The patient was admitted for concerns that the LVAD was failing.
Thoratec/No brand No Device Identifiers Blood pump of ventricular assist device failed.
Thoratec/Circulatory Assist Dual Drive Unit Model 2600 & Model 20002-0000-010 Surgeons were placing the BiVAD devices on the patient. The right VAD wiring and pneumatic line on the first module were placed and, when connected to the first Thoratec dual drive console, functioned as expected. However, when staff went to place the left VAD, it failed to function properly. A second Thoratec dual drive console was brought in with both top and bottom modules and was used to attempt to get the LVAD to work, while the RVAD remained on the first console. However, when both of the LVAD modules in the second console were connected to the patient, the device failed to function properly. After three hours of attempting to troubleshoot the problem, a portable console/driver device was tried, which appropriately operated both the RVAD and LVAD.

Upon inspection after the event, it was noted that there were incorrectly installed components on all three of the different Thoratec modules that were tried and failed to work. The first module was found to have an incorrectly wired solenoid valve which reversed the pressure and vacuum features of the module. Both the second and third modules were found to have incorrectly installed Humphrey valves, which also resulted in a reversal of the pressure and vacuum cycles in the LVAD. This produced insufficient filling and emptying intra-operatively. Though the perfusionists ran a Thoratec designed self pre-use check, the problem was not identified. Another contributing factor to this event was that during the preventative maintenance process in clinical engineering, staff members failed to notice (during the tubing verification check) that the incorrect lamp lit during that test. Careful and thorough analysis of the root cause of this incident revealed poor human factors design of the valve mechanisms. The solenoid valve installation problem on the first module was most likely caused by two of wires feeding into the module accidentally getting crossed. Internal examination of the component revealed a plethora of unlabeled wires used to connect the electric valve and control the Humphrey valve, which are all white and could easily be confused and erroneously crossed. The second and third modules were both found to have the Humphrey valves installed upside down. These valves are identical both right side up and upside down. They have no indication or directions as to the proper installation technique, and could easily be placed incorrectly within the device during regular PM routines. The manufacturer sent a representative to look at the device. There has been no response to date. However, after the root cause analysis, one action item to plan on working with the manufacturer on identifying human factors related issues in equipment design.
Thoratec/HeartMate Battery Lot # MG087009 Patient was going from batteries to Power Base Unit (PBU) in the clinic. When the patient was removed from the first battery, the patient got a red battery alarm, which then went to a red heart alarm in approximately two seconds. The pump was off for three seconds until the power was reconnected.
Thoratec/HeartMate II Lot # MG 264 The patient, who was at home, called the Ventricular Assist Device (VAD) Coordinator, and stated that there were three (3) light bars of power on the controller. Additionally, the patient noted that a yellow battery alarm occurred which was immediately followed by a red battery alarm. The patient said that alarms had been ringing for 5 minutes. The patient was instructed, per the VAD Coordinator, to change her power source. The patient changed one battery and alarms then stopped. The coordinator told the patient to change the second battery, no alarms occurred again. The patient was encouraged to change the batteries to the power base unit (PBU). Again, there were no alarms and the self test was ok. The patient came to the clinic four days later to pick up new batteries for her left ventricular assist device (LVAD); the old batteries where given to Bio-Med. The manufacturer Thoratec told us, "As part of our efforts to continuously improve the quality and reliability of HeartMate batteries we have implemented an improved screening process at our battery supplier, which has three essential features: A simple cycle process that the supplier has run all along, which involves running each battery twice through a charge/discharge cycle at loads representative of end-use; a rapid, partial cycle course at relatively high current specifically designed for high sensitivity to 'bad' cells and/or poor workmanship; improved metrics and acceptance limits based on our experience and understanding (e.g. 'negative voltage change (dV)', end-of-charge current, etc.), some of which are similar in some ways to the way in which [you] characterize batteries at your institution. Since we introduced HeartMate batteries from the improved screening process late last year we have not confirmed any reduced support time issues with these batteries to date".
Thoratec/HeartMate II Lot # MG264 Patient brought 2 batteries into clinic; stated that the batteries would not hold a charge. The 2 batteries involved in this incident were exchanged with the new HeartMate batteries from the company. The manufacturer Thoratec told us, "As part of our efforts to continuously improve the quality and reliability of HeartMate batteries we have implemented an improved screening process at our battery supplier, which has three essential features: A simple cycle process that the supplier has run all along, which involves running each battery twice through a charge/discharge cycle at loads representative of end-use; a rapid, partial cycle course at relatively high current specifically designed for high sensitivity to 'bad' cells and/or poor workmanship; improved metrics and acceptance limits based on our experience and understanding (e.g. 'negative voltage change (dV)', end-of-charge current, etc.), some of which are similar in some ways to the way in which [you] characterize batteries at your institution. Since we introduced HeartMate batteries from the improved screening process late last year we have not confirmed any reduced support time issues with these batteries to date".
Thoratec/HeartMate II No Device Identifiers In the postoperative period, the patient was noted to have intermittent pump stoppage consistent with a cracked percutaneous lead. The percutaneous lead underwent radiography which demonstrated a potential crack along its distal portion. The patient was taken emergently to the operating room for replacement of the HeartMate II left ventricular assist device.
Thoratec/HeartMate II No Device IdentifiersThe patient was implanted with a HeartMate II LVAD (HMII). Approximately 3 months later at approx. 3 am, the patient was lying on right side and had brushed sand or crumbs off the bed with the eft arm. The patient "felt like something dropped or pulled" on her side. The pain radiated to arm and chest. The patient took an ultram and norco and later took two nitroglycerin at 5 minute intervals. The pain was "like soreness in the chest and stomach". The patient became dizzy when moving from the bed to retrieve the HMII pump parameters manual. The patient also experienced nausea with dry heaving and a headache. A family member called EMS. The patient was transported to the ER. In the ER, the patient was evaluated with CT head, CXR, TTE, labs and was brought to CVICU for further management for possible acute LVAD hematoma. A CT scan demonstrated a large anterior mediastinal hematoma with compression of the right ventricle. Contrast enhancement demonstrated extravasation of blood into the hematoma along the outflow graft. The patient was taken to the operating room the day after the event. It was evident that the outflow graft connector had become disengaged from the outflow end of the HeartMate II pump. Outflow cannula was reconnected to HeartMate II pump. The patient is currently recovering in ICU.
Thoratec/HeartMate II Model 103593 The patient called with complaints of not getting enough time from the batteries for the LVAD. The patient was only getting approx. 45-60 min from a set of batteries. The patient was told to come to the clinic in CVC for loaner batteries and to have Biomedical look at the current batteries. The patient came in and was given loaner batteries. The patient was still having the same problem with the loaner batteries. Both set of batteries were tested and found to be in good condition. The decision was made to also change out the patient's primary controller to the back up and to give the patient a new back up controller. The procedure was done and the patient tolerated the procedure well. The new controller is working properly. Of note, we have had two other patients who have experienced similar problems. WE contacted the service rep and the rep acknowledged that they have been having problems like this with this controller.
Thoratec/HeartMate II Model 103593The patient paged the VAD Coordinator to report that the controller had been changed out at home at 11 pm. The patient reports changing over to PBU from batteries and then hearing a steady solid tone. All the lights on the controller were going off and the black power cable on the controller "lit up and burned a hole through the gray cable". The patient was on batteries when the coordinator returned the call. The coordinator walked the patient through transferring over to the PBU without problem. The patient did a self test on the controller, and had four green power lights on the controller. The patient was instructed to come in first thing in morning to get a new back up controller and return the malfunctioning one to us. The patient was also instructed to page us if he had any further problems or concerns.
Thoratec/Stroke Volume Limiter No Device IdentifiersThe Stroke Volume Limiter (SVL) on the patient was not filling completely. A vent of the SVL occurred and the patient “felt funny but was still awake.” Five minutes later, a check of the stroke volume limiter was done and it was not completely filling. A second venting of the SVL occurred and the patient then passed out. The SVL and console were changed out. The new SVL and console were vented and were completely filling and emptying. The patient sustained cerebral hypoxia and never fully recovered purposeful movement of all extremities. The patient remained on respiratory ventilatory support.
Thoratec/HeartMate II Model 103593The patient paged the VAD Coordinator and reported that when changing over to PBU from batteries they heard a steady solid tone and all the lights on the controller went off. Aditionally, the black power cable on the controller got hot. Just before calling the coordinator the patient changed back to batteries. With the VAD Coordinator on the phone, the patient transferred again to the PBU. This time the controller gave the backup controller alarm and the display read "change out controller.” Thoratec support was called by the coordinator out of concern for the percutaneous line (driveline). The patient was told to come in for controller change out and to bring in all the equipment. The waveform and data logger were sent to Thoratec for evaluation on both controllers. The patient was monitored in the clinic for two hours before being allowed to go home. From the waveform analysis, Thoratec felt that the controller change out and replacing the patient power line had solved the problem.
Thoratec/HeartMate No Device IdentifiersA patient with an ICD had a VAD implanted as bridge to transplant. After surgery, the MD was unable to interrogate the ICD. The next day, the tech was also unable to interrogate the ICD. One week later, the company rep attempted to interrogate and was also unsuccessful. This was followed by the VAD coordinator, company rep and techs repeated attempts to interrogate. The problem was determined to be that interaction between the VAD and ICD prevented the ICD from being turned on. This is a known interaction between the programmer and St. Jude ICD in conjunction with HeartMate II due to use of the same frequency. Since there was no work-around for the telemetry problem, the ICD remains in the patient, off, and unable to be interrogated. The patient is stable at this time.
Thoratec/LVAD Console No Device IdentifiersDuring implantation, there was a pop and then an electrical smell from the Thoratec DDC/VAD. The VAD alarmed "battery.” The patient flow was never disrupted. The backup VAD was readily available, was set up, and the devices were switched.
Thoratec/HeartMate II Lot # 71217The pump failed, but it was not able to be evaluated while in the patient. All cables were checked, with no evidence of failure. The patient was unable to withstand surgery to remove the device. We suspect either cable failure or that the patient developed a clot which caused the device to fail. The manufacturer would like the device so they can evaluate it. We have the device sequestered at this time.
Cardiac Assist, Inc/TandemHeart No Device IdentifiersThe patient was admitted from an outside hospital in cardiogenic shock, status post (s/p) MI to the catheterization laboratory. Stents were placed, and an IABP and Tandem ventricular assist device (VAD) were placed. About 18 hrs after VAD was placed, the IABP abruptly alarmed and stopped. The patient was already on pressors. The RN observed that the console wouldn’t run. The console was switched to back-up, and multiple attempts to restart were unsuccessful. There were expert clinicians at the bedside. The company was contacted for assistance to troubleshoot. The patient became less stable, requiring increased vasopressors. When the VAD could not be restarted within 15 minutes by troubleshooting at the bedside, the patient was rebolused with heparin. The VAD was clamped and the device was decannulated and removed. Subsequently the patient remained on a ventilator, IABP, and pressor support. The patient was stabilized over several days and was eventually weaned from support. The patient eventually was discharged 14 days later to rehabilitation with no neurovascular sequelae. After analysis of failed device by manufacturer (Cardiac Assist) we were instructed that if any high flow readings are seen in the Cath Lab on insertion, this should prompt immediate swap out of the VAD device.
ABIOMED, Inc./ BVS 5000 Bi-ventricular Support System Catalog # 0505-0032-MHThe patient was on an LVAD (left ventricular assist device) when blood started spurting out of a hole in the tubing.
Berlin Heart, Inc./ EXCOR Pediatric Model P25P-001A pediatric patient was admitted last year with cardiogenic shock secondary to an idiopathic dilated cardiomyopathy, likely due to viral etiology. The child was immediately cannulated for extracorporeal membrane oxygenation (ECMO) support. The child was eventually weaned off ECMO, but was then required to be placed back due to poor cardiac output. The child underwent implantation of a Berlin Heart left ventricular assist device approximately one month after admission. Despite optimal anticoagulation, the child has developed enlarging thrombus primarily on the outflow valve of the pump. A decision was made to proceed with a pump exchange three months after implantation. The Berlin Heart left ventricular assist device pump was changed two months later, after the pump developed increasing thrombus formation despite optimal anticoagulation. The pump had accumulating thrombus primarily in the outflow valve. The pump was exchanged again one year post-implantation. Two months later, the pump developed a crack in the extracorporeal portion of the apical cannula. This was controlled with digital pressure while the child was emergently resuscitated. The child was then taken emergently to the operating room for a pump change. A crack developed in the assist device tubing, which led to bleeding and probable CNS (air) insult leading to collapse. The child also had down time of about five minutes in the OR during pump change. The child has been hemodynamically stabilized on the ventilator but has been cooled and CNS status is unknown. Initial EEG tracing diffuse showing no seizure, but CT suggests right middle and posterior ischemic event with loss of white/gray differentiation. The child has been waiting for a transplant ever since. After a long discussion with the family, their wish was to withdraw further support, and to make the child’s organs available for donation. Hospital has the device sequestered. Manufacturer is awaiting return of the device for investigation.
Cardiac Assist, Inc./TandemHeart Model EC-1000Fluid was leaking from the housing of the Tandem Heart.



MedSun Newsletters are available at www.fda.gov/cdrh/medsun