Translational Pathway for Ventricular Assist Devices
Preclinical Evaluation of VADs
Changfu Wu, PhD
With some risk of oversimplification, the dawn of a ventricular assist device (VAD) product development effort can typically be attributed to either an unmet customer (patient or clinician) need, sometimes accompanied by a complementary technological advance, or a reaction to a newly appreciated shortcoming in an existing product. Accordingly, we can conceptually separate new product development (NPD) (next-generation, “game changing” field advances) from product-sustaining efforts (customer-responsive adaptations or improvements). A motivation for classifying efforts into these two groups is the difference in certain aspects of the development path. However, new product development projects often offer more definitive solutions to shortcomings in existing products than sustaining projects can provide. There is, therefore, a need for discipline in neither under-pursuing nor over-pursuing individual field observations associated with a VAD system in order to optimize patient benefit.
New Product Development
NPD in VADs is typically characterized by relatively long durations, driven by:
— the intrinsic challenges of designing a reliable system with implanted parts exposed to the hostile human homeostatic environment, for which uninvited titanium and polyester guests are not welcome in an evolutionary sense, and the frequently more hostile vagaries of human interaction with the nonimplanted parts among an increasingly active patient population with longer support durations;
— the relatively long duration of mission-critical verification and validation testing, which includes some tests that run for extended periods (e.g., cycle testing to cover the equivalent of years of use, sometimes involving tens of millions of cycles) plus the general need to acquire extensive product performance experience internally prior to product launch; and
— the most stringent regulatory requirements due to a VAD’s classification as an active, implantable medical device.
Thus, generational product life cycles are on the order of a decade, give or take a few years. This leads to pressure to include as many features and technological advances as possible within the scope of an NPD project. General technological advances that can be adapted to medical devices often enable reduced device size and enhanced capability. Collective clinical experience with existing products is a fertile source of unmet needs, especially as medical management improves concomitantly with technological advances and improved patient outcomes drive elevated demand for refined quality and new features. Organizations typically devote a portion of their engineering and science resources to research that is strategically targeted to address anticipated NPD needs. Such groups are charged with converting innovative concepts to a level of technical feasibility and practicality appropriate for transfer into an NPD program. Discoveries and breakthroughs so attained, for example, obviating mechanical bearings in left VADs by magnetically levitating the left VAD’s rotor, have led to some of the most meaningful advances in reducing mechanical circulatory support-related adverse events and enhancing patient quality of life.
Commercial LVADs are subject to change for various reasons. Component obsolescence is a constant source of supply chain risk, magnified by the much shorter life cycles of the consumer products, such as cellular phones and computers, that drive frequent revision of components that are also utilized in VAD systems. Anticipating looming unavailability of a component, identifying a substitute, adapting the device for the new component (e.g., modifying a circuit board layout for a new integrated circuit that does not fit into the space vacated by the obsolesced one), and verifying functionality is a continuous effort. Organizations that develop and manufacture VADs have structured systems for capturing and managing customer complaints, a term used without pejorative to include remarks, dissatisfactions, mishaps with the medical device, injuries to patients, adverse events, and so forth. Downstream systems to evaluate imminent or future hazards and potential corrective actions can lead to projects for design and development improvements.
Preclinical VAD Evaluation
New product development and sustaining projects proceed in accordance with internal procedures that simultaneously satisfy business and regulatory requirements that, because the long-term interests of a business so greatly depend upon the quality of its products, do not conflict in terms of satisfying the patient and clinician. Although the nomenclature, categorization, and format will differ among medical device companies, some fundamental aspects of the development process, described below, are typical.
- Formal design reviews by responsible staff at defined project phases, for instance, between a design phase and a verification phase, provide an opportunity to scrutinize work already accomplished (in terms of quality, rigor, etc.) and evaluate whether the existing plan should be adjusted (an option that could be as extreme as project cancellation). Such reviews are mandated by Design Control regulations, such as FDA 21 Code of Federal Regulations (CFR) Part 820 and European Union Medical Device Regulation.
- Formal inputs (requirements) and outputs must be established (21 CFR 820.30), with clear traceability between input(s) and output(s). Inputs may directly address specific user needs or be derived from system architecture, international standards, or risk management activities. It is critical that requirements be adequately defined, which requires engineering judgment. Requirements are typically quantitative, often with tolerable deviations, ranges, and pre-existing conditions specified. For example, the required runtime of a battery may be at least a certain period of time after the battery has been aged a certain number of cycles. Each input requirement must be verified (21 CFR 820.30(f)), confirming that the product has been made properly to meet requirements. Formal user needs must be established and validated (21 CFR 820.30(g)), confirming that the right product has been produced to meet needs. Validation testing is done in actual or simulated use conditions with test articles that represent actual production processes or equivalents. The design verification and validation studies of VADs typically include, but are not limited to, the following:
- Performance characterization
- Mechanical safety
- Mechanical strength and integrity
- Leakage and pressurization
- Fluid dynamics and flow visualization
- Packaging and shipping
- Environmental use and exposure
- Heat and particulate generation
- Hemocompatibility (e.g., mechanical hemolysis)
- Electromagnetic compatibility (EMC)
- Wireless compatibility
- Electrical safety (including battery safety and function)
- Software (including cybersecurity)
- Human factors
- Shelf life
- Animal study
While the specifics of the individual tests are particular to the individual VAD design and are beyond the scope of this chapter, the device manufacturers might find the following overview helpful:
- Some general guiding principles are as follows:
- Test the final, finished device (unless differences between the finished and tested device are irrelevant to the test attribute);
- Include germane environmental conditions, with documented rationale, and test details. For example, in testing the hydraulic characteristics of a VAD, it is likely important to specify the elemental constituency, temperature, viscosity, and density of the fluid required to meet a certain pressure and flow, while disregarding the environmental humidity and temperature (because it is a closed system). Conversely, in testing the seal integrity of the VAD’s sterile package, environmental temperature and humidity may be critical factors.
- Precondition the test articles as expected clinically (e.g., sterilization, environmental conditioning, shipping/handling);
- Determine “worst-case” simulated use conditions expected for each particular test;
- Adopt adequate, and where applicable, statistically-derived sample sizes; and
- Prespecify acceptance criteria, if applicable, with appropriate clinical justifications based upon the intended use of the device. For example, if the axial strength requirement of a cable is greater than or equal to a certain force, the rationale should explain why that force value was specified, which may involve reasoning about surrounding objects, patient activities, factors of safety, and so forth.
- Performance characterization testing will demonstrate that the complete system meets its design specifications over its full range of operating conditions for all possible configurations. It is critical that the test parameters not be limited to the nominal operating conditions. Any time-dependent performance characteristics of the system will need to be reported in the data. In addition, performance characterization testing can be used to determine the appropriate “worst-case” test conditions and parameters for the in vitro reliability testing.
- Mechanical safety testing is conducted to show that all device components perform to design specifications and meet applicable acceptance criteria. The testing of the pump, conduits, cannulae, cables, leads, bend reliefs, and connectors is usually conducted separately on each component or assembly under either static or cyclic loading. Testing includes but is not limited to leakage, pressurization, mechanical strength, and mechanical integrity. The mechanical safety testing may seem to be simple and straightforward, but in reality, many adverse events related to these components have been observed in investigational device exemption (IDE) trials and in commercial use, which often prompts necessary design and manufacturing modifications. Accurate characterization of the loading is the key. It is worth emphasizing that testing performed under single-mode motion may not reflect the real-life scenario and thus multiple-mode motion may be more appropriate. Additional mechanical safety testing includes but is not limited to fluid dynamics and flow visualization of the entire blood pathway, environmental and expected patient transport testing (e.g. ambulance, fixed-wing aircraft, and helicopter), packaging and shipping testing, heat generation, and particulate generation in pumps that experience wear of any moving components over time.
- In vitro reliability testing is conducted to demonstrate that a VAD can last the duration of intended use. As VAD therapy is increasingly an option for patients ineligible for cardiac transplant, support durations are increasing, and product reliability and component durability are increasingly important characteristics. With VAD systems, a multifaceted approach is routine. Testing typically includes accelerated testing in which loads or conditions are exerted upon components or subassemblies at a rate often far exceeding real-world conditions to relatively quickly acquire results. The extent to which the acceleration unrealistically affects results, the exaggeration of loads imposed to establish a factor of safety where appropriate, and the calculation of suitable life requirements (i.e., number of cycles) and sample sizes are the responsibility of experienced engineers using solid judgment. A historical example illustrates this point: in evaluating the wear-out of the contact elements of a ball bearing, an apparatus may be devised to impose much faster than normal cycles on the bearing in order to reach, say, a million cycles in a matter of days instead of months. However, if the dynamics of what is occurring during the stop and start at each cycle and the speed of the rotating elements are supra-normal, the accelerated results may not reflect what would have occurred at real-world speeds; indeed, the failure mechanism may be entirely different. In this example, it may be appreciated that a faster relative speed between lubricated rolling elements may be protective against wear, and that accelerated test results could undesirably overestimate expected life.
Notwithstanding the importance of these numerous accelerated tests, a “real-time life test,” in which an implanted VAD and its controller are run in real-world modes, speeds, and conditions, has been historically important in identifying real-world failure modes, and continues to be recommended in this era in which VADs have fewer moving parts and less wear. It is typically run indefinitely until failure even as its predictive power of reliability becomes eclipsed by that of actual clinical experience. The interim result of the “real-time life test” at a prespecified interval is often used to support regulatory submissions for clinical trials and product approvals.
Highly accelerated life tests or HALTs, in which unnaturally hostile loads and conditions are imposed, are often utilized in the early stages of VAD and controller design to elicit weaknesses and drive design improvements for robustness. As such, they contribute to product reliability and durability, but because they are not used to verify requirements, they are conceptually distinct from the accelerated and real-time tests mentioned above.
The duration for in vitro reliability testing is commensurate with the longest duration of intended use with an adequate safety margin. For temporary support VADs, the test duration is expected to be two times the intended use period. For durable VADs, the test duration for the short-term (e.g., bridge-to-transplant) indication is 6 months for the IDE study and 1 year for the premarket approval (PMA) application; and the test duration for the long-term (e.g., destination therapy) indication is 1 year for the IDE and 2 years for the PMA. The in vitro reliability testing is often performed under “worst-case” simulated use conditions, not nominal conditions. It is also critical to prespecify what constitutes a failure in order to avoid ambiguity in interpreting the test results. Another important aspect of in vitro reliability testing is the sample size, which is typically calculated based on a prespecified reliability goal and associated confidence level. For VADs, it has been recommended that a minimum of 80% reliability be demonstrated at an 80% confidence level (1).
- Biocompatibility testing is typically conducted on the final, finished device. Details of the test recommendations can be found in the FDA guidance document on the use of international standard ISO 10993-1 (2).
- In vitro short-term blood damage testing (up to 6 h) may complement the hemocompatibility testing in the animal study and clinical study for long-term VADs. It is typically more informative to perform parallel comparative testing against an appropriate predicate or control device, including comparing the new design and the old design. During the tests, it is imperative to use appropriate flow rates, pressures, and temperature(s); to optimize test sensitivity (e.g., blood volume, test duration, and hematocrit); to use blood from multiple subjects; to justify an adequate sample size to demonstrate reproducibility of results; and to demonstrate device safety over the entire operating range, not just the nominal range.
- Electrical safety and EMC testing is conducted in accordance with IEC 60601-1, IEC 60601-1-2, and other applicable standards on the whole system to prevent, for example, electrical shock, explosion, fire, and interference with other electronic equipment and devices. All power sources and device configurations need to be evaluated. Battery depletion time, recharge time, and cycle life are measured under “worst case” expected loads. It is important to pay special attention to the batteries, as many adverse events associated with batteries have been reported. If the controller has wireless capability, it needs to be tested for wireless compatibility in order to prevent any interference and ensure that it maintains a robust and safe wireless connection (3).
- Software documentation is commensurate with the level of concern of the software (4). All alarm functions need to be properly simulated and tested because nuisance and false alarms are common in clinical use. Additional recommendations can be found in FDA’s guidance document on general principles of software validation (5).
- Human factors or usability study is an important aspect of consideration for home use devices (6). Although the emergence of VADs as a viable therapy for end-stage heart failure was precipitated by breakthroughs in pump development, most notably the advent of compact, reliable, continuous flow pumps, it is not the implanted pump that garners the attention of the patient daily; rather, it is the non-implanted equipment, such as controllers and batteries, that require patient comprehension and interaction. Human factors and usability engineering techniques have emerged as principally important in averting real-world usage problems (e.g., connector pin misalignment during controller exchange) that can sometime lead to nonobvious hazards. All aspects of VAD therapy are important, of course, and human factors and usability engineering is applied across the board in optimizing surgical steps, patient management by health care workers, and product labeling. For VADs, AAMI/ANSI HE75:2009 and ISO/IEC 62366:2007 are the applicable standards.
Human factors or usability studies are conducted to demonstrate that the device has met the needs of the intended patient population and can be used by representative users without patterns of preventable failures or difficulties that could result in adverse clinical consequences including death, delayed therapy, or inadvertent injury (7). Such studies focus on the observed pattern of use errors while performing critical and essential tasks.
- Animal study is a key component of preclinical testing of VADs. The choice of animal model and the study design need to be justified. It is important to report the data from all animals tested, not just from the surviving animals. More detailed discussions on animal studies can be found in the FDA guidance document for animal studies involving cardiovascular devices (8). In addition, reliability data obtained from the animal study can be reported separately from the in vitro reliability test data.
- Computational modeling is an indispensable tool in the design of VADs. It can also play a role in device evaluation (e.g., pump hemocompatibility) when properly validated. Details on what to include in a report can be found in the FDA guidance document on reporting of computational modeling studies in medical device submissions (9).
- International standards provide a testing benchmark by which widely varying devices may be compared, to some degree. ISO 14708-5 is the applicable standard for VADs. In addition to compliance with some high-level standards, the appropriateness of applying particular standards of narrower scope is determined by engineering judgment in view of the clinical indication, habitus, environment, etc. Many aspects of VAD systems are not adequately covered by existing standards, so custom test protocols and apparatuses are often designed for verification purposes. It is the responsibility of the engineering team to ensure the adequacy of a particular test approach.
- Risk management is at once a process, a tool, and a mindset for VAD development engineers. It is ubiquitous in every phase of development. High-level product risk assessment, accompanied by low-level risk assessments such as failure mode and effect analysis, drive comprehensiveness in architecture and design considerations. Identified risk mitigations can lead to new requirements. Determined risk levels influence test sample sizes and confidence limits. Risk-to-benefit analyses are used to arbitrate tradeoffs. ANSI/AAMI/ISO 14971:2007 is the applicable standard for VADs.
Preclinical evaluation of VADs is based on the risk analysis of the whole system. It is conducted to properly mitigate all known potential risks associated with the device system and ensure that the residual risks post mitigations are as low as reasonably practicable. The risk analysis is an iterative process throughout the total product life cycle. As such, preclinical evaluation of VADs is an iterative process as well. Verification, validation, and qualification activities are carried out throughout the total product life cycle in response to design and manufacturing modifications to the device and additional information that becomes available (e.g., field events or complaints).
- Lee J. Long-term Mechanical Circulatory Support System reliability recommendation by the National Clinical Trial Initiative subcommittee. ASAIO J. 2009;55:534-42.
- U.S. Food and Drug Administration. Use of International Standard ISO 10993-1, “Biological evaluation of medical devices —Part 1: Evaluation and testing within a risk management process.” Guidance for industry and Food and Drug Administration staff.2016. Available at: https://www.fda.gov/media/85865/download. Accessed May 4, 2019.
- U.S. Food and Drug Administration. Radio Frequency Wireless Technology in Medical Devices. Guidance for Industry and Food and Drug Administration Staff. 2013. Available at: https://www.fda.gov/media/71975/download. Accessed May 4, 2019.
- U.S. Food and Drug Administration. Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices. Guidance for Industry and Food and Drug Administration Staff. 2005. Available at: https://www.fda.gov/media/73065/download. Accessed May 4, 2019.
- U.S. Food and Drug Administration. General Principles of Software Validation; Final Guidance for Industry and FDA Staff. 2002. Available at: https://www.fda.gov/media/73141/download. Accessed May 4, 2019.
- U.S. Food and Drug Administration. Design Considerations for Devices Intended for Home Use. Guidance for Industry and Food and Drug Administration Staff. 2014. Available at: https://www.fda.gov/media/84830/download. Accessed May 4, 2019.
- U.S. Food and Drug Administration. Applying Human Factors and Usability Engineering to Medical Devices. 2016. Available at: https://www.fda.gov/media/80481/download. Accessed May 4, 2019.
- U.S. Food and Drug Administration. Guidance for Industry and FDA Staff – General Considerations for Animal Studies for Cardiovascular Devices. 2010. Available at: https://www.fda.gov/media/79366/download. Accessed May 4, 2019.
- U.S. Food and Drug Administration. Reporting of Computational Modeling Studies in Medical Device Submissions. 2016. Available at: https://www.fda.gov/media/87586/download. Accessed May 4, 2019.