STEVENSON AND KORMOS, ET AL., MECHANICAL CARDIAC SUPPORT
2000
JACC Vol. 37, No. 1, January 2001:340-70
V.
Future Devices Entering Clinical Development
A.
Existing Minimum Standards for Pre-Clinical Device Evaluation
There
is presently no standard for the pre-clinical evaluation
of devices used in mechanical circulatory support systems.
The FDA Office of Device Evaluation still provides useful
information and interaction for blood pump developers,
but officially, there is no existing standard for the
pre-clinical evaluation of these devices. Consequently,
it is recommended that circulatory support system developers
schedule a pre-investigational device exemption (IDE)
submission meeting with the FDA to educate the reviewers
in advance on the specifics of their system and to receive
feedback from the FDA on the appropriate criteria for
the review of their system. Two guidelines for pre-clinical
device evaluation do exist. First, the Preliminary Draft
Guidance for Ventricular Assist Devices and Total Artificial
Hearts issued by the FDA in December 1987 is the original
document. Although it is useful in presenting criteria
for device evaluation, it is considered obsolete. It
also needs to be recognized that the document was issued
early in the clinical experience of using VADs and total
artificial hearts for bridging to transplantation. The
full extent of the circumstances in which these devices
would be used (i.e., in and out of the hospital and
for durations of months to over a year) could not be
fully anticipated by that document. Hence, the periodic
revision of the criteria for evaluation became both
necessary and appropriate for the evaluation process
and a source of frustration for device developers and
investigators.
The
second guideline comes from a joint paper developed
by an ASAIO and the STS interdisciplinary working group
(including participants from academia, industry, the
NIH and the FDA). This working group jointly published
a reliability recommendation for long-term blood pump
systems in 1998.100
This recommendation has been used to guide the reliability
evaluations for blood pump systems that are currently
under development or that have recently entered clinical
trials. It needs to be emphasized, however, that this
recommendation is limited to reliability concerns for
long-term devices, so there is still a need for a more
comprehensive standard with specific criteria for pre-clinical
in vitro and in vivo testing and evaluation of devices.
As
long-term clinical experience has been gained with circulatory
support systems in bridge-to-transplant, bridge-to-recovery
and alternative-to-transplant settings, it has become
clear that the performance goals for these systems needs
to be revised from values stated in or related to the
FDA Preliminary Draft Guidance. Controversy has existed
over the required duration of pre-clinical animal implantation
tests and reliability mission life duration. Concern
has been expressed over the recommended duration of
pre-clinical reliability mission life duration (some
consider the recommended minimum of one year to be too
short for a long-term system) and the duration of the
animal implantation trials (some consider the recommended
90 days to be too long), but there is insufficient evidence
to address these concerns at this time. It also needs
to be recognized that although the longer use of these
circulatory support systems is the primary motivation
for updating minimum criteria for pre-clinical device
evaluation, the pre-clinical criteria for devices intended
for short-term use (i.e., post-cardiotomy CS and transient
right heart failure after LV assist implantation) and
bridge-to-recovery also need to be examined and accommodated
in a new standard. The revision of these guidelines
becomes even more crucial as the definitions for short-
and long-term devices become less clear based on clinical
applicability. Previously, patients undergoing post-cardiotomy
support were felt to require periods of support not
extending beyond 10 days to 2 weeks. There are now anecdotal
reports showing that recovery has, in fact, been seen
with periods of support extending several weeks to several
months. In addition, there is the distinct possibility
that the patient may become device-dependent, changing
what was originally anticipated to be a short-term support
period to an extended period as either destination therapy
or a bridge to cardiac transplantation. Another perspective
to consider is that devices need to be specifically
designed to meet the needs of the identified patient
population.
The
FDA Preliminary Draft Guidance Document and the ASAIO/STS
Reliability Recommendation are still considered to be
useful documents by several blood pump development groups.
However, the need for a current and comprehensive standard
for pre-clinical evaluation of devices remains. To begin
to address this need, the Association for the Advancement
of Medical Instrumentation (AAMI)2 is presently
leading the interdisciplinary development (including
participation by the FDA3 of a Technical
Information Report (TIR). The AAMI TIR is in the final
development stages. It is expected to be available from
the AAMI by the end of the summer of 2000. It must be
recognized that due to the uniqueness of each blood
pump system, this document provides a comprehensive
review of blood pump system issues to be evaluated and
considered for inclusion in a FDA IDE submission, but
it does not provide a checklist of specific performance
requirements. However, the AAMI document does provide
several references to guidelines and standards on specific
topics related to blood pump systems. Ultimately, the
comprehensive design, implementation and documentation
of a blood pump system development program with validated
in vitro and in vivo testing using sound scientific
protocols for data collection and analysis will lead
to a successful FDA device review.
Finally,
some criteria need to be developed to clearly identify
system standards for devices that can be used in different
situations for variable clinical indications as the
definitions of bridge-to-transplantation, bridge-to-recovery
and destination therapy become less distinct. It is
not uncommon for example, for a device to be implanted
for a post-cardiotomy indication, and then removed much
later (three to six months) than intended, because the
recovery process may be longer than anticipated. In
addition, at some point if the patient cannot be weaned,
he or she can be converted to a transplant candidate.
On the other hand, if adverse events occur that preclude
transplantation, the device may have to perform in the
mode of destination therapy. Thus, reliability requirements,
which may have been sufficient for post-cardiotomy use,
are now ill-defined for permanent use.
The
development of a comprehensive standard for the pre-clinical
evaluation of blood pump systems, though needed, is
not presently being planned. The effort to create such
a standard would require a rigorous interdisciplinary
effort over a period of three to five years. Until such
a standard is developed, it is incumbent upon the members
of the blood pump development community and the FDA
Device Evaluation staff to share the lessons they have
learned to advance the understanding of the pre-clinical
blood pump evaluation process. It is also incumbent
upon the FDA Device Evaluation staff to continue their
difficult job of fairly and expeditiously submitting
reviews, while being cognizant of the need to revise
their criteria as the clinical experience with circulatory
support systems grows. Because of the uniqueness of
each blood pump system and its intended use, the development
of a fixed true standard may be an unachievable goal.
A more farsighted approach may be a continuing, interdisciplinary
revision of a guidance document for blood pumping systems.
B.
Devices Currently in Clinical Development
The
first section of this conference document reviewed the
devices currently available in the U.S. for intermediate
or long-term support. This section reviews the mechanical
circulatory support systems that are likely to enter
clinical trials as chronic support devices in the U.S.
within the next five years. Such devices fall into four
major categories: 1) continuous flow LVADs (including
axial flow and centrifugal flow pumps), 2) pulsatile
LVADs, 3) the total artificial heart and 4) devices
without blood contact.
In
general, these new devices first undergo extensive ex-vivo
reliability testing followed by chronic animal implantations.
The third phase is human trials, which generally begin
with a single site and then expand to five to twenty
centers, testing the device initially as either a bridge
to transplantation or as a chronic implant. Clinical
trials are then performed to obtain PMA.
1.
Continuous flow left ventricular assist devices.
Continuous
flow, or rotary devices, are currently of two basic types:
axial flow pumps and centrifugal flow pumps. They have
several potential advantages over current pulsatile pumps:
1) they are smaller devices and therefore can be used
in smaller patients (less than the 1.5 m2
body surface area (BSA) required for most pulsatile devices);
2) they are relatively simple, have fewer moving parts
than pulsatile pumps and thus may be less prone to mechanical
failures, although this is unproven; 3) because of the
continuous flow characteristics, they do not require a
compliance chamber in the system; 4) they have lower energy
requirements; and 5) the small size of the device and
the pocket may decrease the risk of infection, although
this is also unproven. These devices also have potential
disadvantages that remain to be quantified: 1) current
axial flow pumps use bearings lubricated by blood, and
this area of relative stasis is a potential source of
in-situ thrombus or thromboemboli; 2) chronic anti-coagulation
is necessary; 3) some degree of hemolysis is common, the
long-term effects of which are unknown; 4) the long-term
effects of non-pulsatile (or essentially non-pulsatile)
flow are unknown; and 5) feedback control mechanisms for
pump speed are complex and unproven.
a.
Axial flow pumps. Three axial flow pumps are likely
to undergo “first generation” chronic device
trials in the U.S., with several trials underway in
Europe. They include the Nimbus/TCI IVAS, the Jarvik
2000 IVAS and the DeBakey/MicroMed IVAS. The axial flow
motor is small and contains rotary blades that spin
at 10,000 to 20,000 rpm and can pump approximately five
to six l/min. Because of the continuous flow properties
of the axial flow pumps, there are no valves in the
system.
The
Nimbus IVAS (HeartMate II) is a small (7 cm length)
axial flow pump that connects to the LV apex for inflow
and the ascending aorta for outflow.101
Under normal operation, the inlet pressure to the axial
flow pump will be cyclical, varying with the systolic-diastolic
phases of the LV, creating some degree of pulsatility.
An electromagnetic motor (pump rotor) turns the turbine.
A low-pulse mode produced by variable motor speed will
also be available. Two cup-socket ruby bearings support
the pump rotor. The outer boundary of the bearing’s
adjacent static and moving surfaces is washed directly
by blood flow. The pump’s speed can be controlled
manually and by a proposed auto-mode that relies on
an algorithm based on pump speed, inherent native cardiac
pulsatility and current. A first version of this device
is powered through a percutaneous small-diameter electrical
cable connected to the system’s external electrical
controller. A fully implantable system is under development.
The
Jarvik 2000 Heart is a similar, compact (5.5 cm length,
85 gm weight) axial flow pump that receives inflow from
the LV apex and outflow through a Dacron graft anastomosed
to the descending thoracic aorta.102
The rotor constitutes the only moving part of the device
and is supported at each end by tiny blood-immersed
ceramic bearings.103
The currently existing device is tethered to an external
electrical power source through a percutaneous wire,
but a subsequent totally implantable version will contain
a microprocessor-based controller that can sense and
change pump speed according to different phases of the
cardiac cycle and receive power via a transcutaneous
energy transfer system coil.
The
MicroMed DeBakey Axial Flow Pump is an electromagnetically
actuated, implantable titanium axial flow pump that
connects to the LV apex and ascending aorta. The pump
is designed to produce flows of 5 l/min against 100
mm Hg pressure with a rotor speed of 10,000 rpm.104
The currently existing design of this pump includes
a fixed rpm rate that can be adjusted through an external
device. During periods of patient mobilizations, power
can be supplied by two 12-volt DC batteries for several
hours.
b.
Centrifugal flow pumps. Centrifugal flow devices
are somewhat larger than axial flow pumps and provide
nonpulsatile flow, but the rotational speeds are much
slower (about 2,000–4,000 vs. 10,000–20,000
rpm). The same general advantages and disadvantages
apply to centrifugal flow pumps as to axial flow pumps.
The
AB-180 Circulatory Support System is a small, durable
implantable centrifugal pump that receives inflow from
the left atrium and empties into the ascending aorta.105,106
The rotor is powered by electromagnetic coupling. A
solution of distilled water and heparin provides a high
local concentration of anticoagulant within the pump.
An occluder device prevents retrograde flow from the
aorta to the left atrium in the event of pump failure.
Although it is potentially useful for long-term support,
the AB-180 CSS will first be tested as a support device
for post-cardiotomy shock.
The
HeartMate III LVAD is a centrifugal pump powered by
magnetic levitation, a process that combines the functions
of levitation and rotation in a single magnetic structure.
The small pump rotor does not contain bearings and is
completely encased in titanium.
The
CorAide™ centrifugal blood pump is an implantable
LVAD with a suspended rotor that is noncontacting. The
pump produces 8 liters/min flow at 6.5 W.
2.
Pulsatile flow devices. Excluding
the Novocor and TCI HeartMate (discussed under “Current
State of Devices”), pulsatile LVADs likely to
enter long-term clinical trials within the next five
years are the Thoratec Intracorporeal Ventricular Assist
Device (IVAD), the Novacor II, the Worldheart HeartSaver
VAD and the Arrow Lionheart VAD. Each of these chronic
LVADs requires chronic anti-coagulation with coumadin.
The
Thoratec IVAD is designed as a small lightweight device
for left or biventricular support.107,108
This IVAD maintains the same blood flow path, valves
and polyurethane blood pump sac as the paracorporeal
Thoratec device. The major advantage of this IVAD is
its relatively small size (339 gm) and simplicity in
a pulsatile system that can be implanted in patients
ranging in weight from 40 to ≥100 kg. Only the small
blood pump is implanted in a pre-peritoneal position
with a small (9 mm) percutaneous pneumatic drive line
for each VAD connected to a more complex control unit
externally, where it can be serviced and replaced. The
pump is controlled with a small briefcase-sized, battery
powered pneumatic control unit.
The
Novacor II miniaturized pulsatile pump is an extension
of the current Novacor technology that substantially
reduces pump size. The single pump is replaced by two
small sac-type pumps, each driven by a central pusher
plate mechanism, supporting the LV output through multiple
pump cycles. The pusher plate is driven by direct electromagnetic
actuation, resulting in a simple bearingless system.
The
Worldheart HeartSaver VAD was designed as a totally
implantable chronic VAD and has several major attributes:
1) the device is totally implantable and requires no
percutaneous connections; 2) it is designed for implantation
in the left hemothorax adjacent to the natural heart
and can be anchored to the rib cage; 3) the device is
remotely monitored and controlled; 4) an internally
implanted and rechargeable battery allows the patient
to partake in a variety of activities, unencumbered
by any external components; and 5) the device can be
implanted without cardiopulmonary bypass. The blood
contact surface of the sac is fabricated from polyurethane
and the valves are porcine tissue valves. An electromagnetic
coupling device transfers power across the intact skin
and tissue. Wireless monitoring and control of the device
is provided by a transcutaneous infrared biotelemetry
system.
The
Arrow LionHeart VAD is another totally implantable LVAD
system with tilting disc valves in which transcutaneous
energy is transferred to implanted batteries.109
The energy converter is based on a roller screw mechanism,
which in turn causes linear motion at a circular pusher
plate that compresses the polyurethane blood sac during
systole. In diastole the motor reverses to withdraw
the pusher plate. An intrathoracic compliance chamber
maintains near-thoracic pressures in the energy converter
airspace. External electronics consist of the energy
transmission source, a power pack, a battery charger
and portable power supplies.
3.
Total artificial hearts. Two total artificial heart
systems are expected to enter clinical trials in the
U.S. within the next five years. They include the Abiomed
Total Artificial Heart and the Penn State Total Artificial
Heart. Both pumps require chronic anticoagulation with
warfarin ± anti-platelet agents.
The
Abiomed Total Artificial Heart (AbioCor) is a completely
implantable system that can generate cardiac output
in excess of 10 liters/m. Powered by transcutaneous
energy via coils, an internal battery is included for
20 to 40 min of tether-free time. All blood-contacting
surfaces, including the two blood pumps and four tri-leaflet
valves, are fabricated from seamless polyurethane (angioflex).
Blood flow is maintained by a high-efficiency miniature
centrifugal pump, which operates unidirectionally, while
a cylindrical rotary valve alternates the direction
of the hydraulic fluid flow between the left and right
pumping chambers. Left/right balance is achieved by
adjusting the right prosthetic ventricle stroke volume
via a hydraulic shunt mechanism that incorporates a
balancing chamber attached to the left prosthetic ventricle
inflow port.110
The
Penn State/3M Total Artificial Heart is a totally implantable
device based on a rotor screw mechanism that produces
8 liters/min with a stroke of 64 ml.111
Circular pusher plates are attached to the two ends
of the rotor screw shaft, and a brushless DC electric
motor rotates the screw 6.3 revolutions to provide a
full pusher plate stroke with 1.9 cm linear motion.
One pump empties while the other fills, and the motor
then reverses to eject the opposite pump. A seamless
polyurethane blood sac fits within each titanium pump
case, and Bjork-Shiley convexo-concave or Delrin monostrut
valves (2.5 mm inlet, 27 outlet) provide unidirectional
flow. Left/right balance is achieved by the use of estimated
end-diastolic volume from motor speed and voltage. A
compliance chamber is coupled to the housing to accommodate
volume changes caused by gas diffusion from the blood
and changes in atmospheric pressure. Energy is passed
through a transcutaneous system to an implanted controller
box and Nilco rechargeable battery (45 min tether-free).
There is a subcutaneous port for access to the compliance
chamber.
4.
Devices without blood contact. Currently existing
devices without blood contact are designed for short-term
support. However, the development of similar devices
for chronic therapy appears likely. The Abiomed Heart
Booster combines an LV volume constraining device with
a contractile component. Control of LV dilatation is
effected by a conical “jacket” that fits
over the apex of the heart. The contractile component
is based on a change in the shape of multiple thin-walled
tubes from a circular cross-section to a highly elliptical
or flat cross-section, and vice versa. Rapid hydraulic
inflation of the tubes (toward a circular shape) results
in a smaller enclosed volume, and rapid deflation of
the tubes (toward highly elliptical shape) results in
a larger enclosed volume. When negative pressure is
applied to the tubes during diastole, the tubes collapse
completely in such a way that the pericardial wrap becomes
a thin structure that is relatively pliable and does
not impede diastolic filling. The device wraps around
the apex of the heart and, like other volume constraining
devices, does not require cardiopulmonary bypass for
implantation. A smooth outer surface is used to prevent
tissue ingrowth around the outer surfaces of the device
and reduce diastolic dysfunction.
C.
Conclusions
Results
and lessons learned from trials such as the REMATCH
trial will inevitably influence future trial design
in the field of mechanical circulatory support. As the
field moves ahead, it has become clear that no one trial
design will be ideal or appropriate for all devices,
populations and stages of development. A variety of
research designs will be necessary. Creation of a national
outcomes database for advanced HF will facilitate effective
trial design and identify populations that may potentially
benefit.
Responsible
progress in this field requires the establishment and
maintenance of a mandatory registry that includes all
implantable devices, both before and after approval.
The combined effort of the various stakeholders is required
to address issues of funding, data format and management,
compliance and access, while balancing proprietary concerns.
A major achievement of this conference is the recognition
that the field will advance further and more rapidly
if the various groups involved in developing and testing
new devices can collaborate effectively in the future.
|
