ICDs: Dealing With Less Than Perfect
ICDs: Dealing With Less Than Perfect
Implantable defibrillator (ICD) therapy is high-profile life and death therapy, set up by a long series of impressive randomized clinical trials dramatically demonstrating improved survival. Such high stakes therapy and striking trial results bring visibility and scrutiny. From the beginning, ICD therapy has had its scientific and emotional aspects, clinical and ethical issues, technical and economic considerations and its supporters and detractors. Lost in all of this is the fact that the therapy is and will continue to be less than perfect. Sometimes, as demonstrated by the recently reported tragic loss of life with ICD device failure, the lack of perfection transcends an academic discussion. Unfortunately, this was not the first time nor will it be the last that we need to confront similar situations. Maisel and his colleagues reported on the increasing number of ICD device failures that produced safety alerts and recalls between 1990 and 2000. As we consider this case and these issues we need to step back much further and see what we think the correct actions ought to be and the consequences of these actions.
There have been scientific and regulatory systems set up to monitor the design, clinical evaluation, manufacturing, marketing and sales of medical devices. It is crucial that both the devices and the systems designed to monitor their use maintain the highest standards of excellence. It is for that purpose -- the highest standards of excellence -- that the US Food and Drug Administration (FDA) has been charged by Congress to regulate and monitor the safety and effectiveness of ICD therapy and all medical devices in the United States. However, at no step is perfection a possibility.
The manufacturing and sale of ICDs are under the scrutiny of the FDA, which is responsible to certify the facilities, general manufacturing processes and every device model. New products are tested and certified, on the basis of data collected during carefully designed and monitored clinical trials, to be safe and effective when used in accordance with their labeling. It is not possible for these systems or the preclinical studies to detect the infrequent deficiencies that are inherent to the therapy much less the defects that develop due to component failures, whether they occur during manufacture, during surgical implantation or are related to the patient's anatomy or disease. These failures are often extremely infrequent and difficult to analyze. Consequently, it is mandated that the manufacturer, the hospital and the physician report all situations where a device malfunction may have played a role. These reports are publically available, in the Manufacturer and Users Facility Device Experiences (MAUDE) database. The electrical overstress situation of the recently publicized death was reported, as were the prior similar events in regard to the same device model. Every day failures of pulse generators, lead insulation, conductor failures, device software and firmware failures and programmer software lockups are reported since the regulations state that every incident should be reported by the physician/hospital and the manufacturer. Unfortunately, the physicians are both the weakest and most important link in this system. On the other hand, the manufacturers take this responsibility very seriously and report every failure for which they are aware. There is no doubt that the current systems need dialogue, review and reform.
The physician community has a profound and important responsibility in this scenario. However, we have hardly been engaged in the process. Frequently we have not reported failures to the MAUDE database, not returned devices or leads for analysis, not understood that with the exponential increase in implants, the absolute failure rate is becoming relatively common, and we have failed to understand that we must constantly examine the performance of the devices in our patients. Just as important, we need to put into perspective these blemishes in device performance and produce a clinically appropriate response. Often our response has been worse than the problems that we have tried to ameliorate. Inherent in all device therapy is a failure rate, which is part of the initial risk undertaken when the device is implanted. Our society and physician community seem to be under the false impression that perfection is a possibility. A more accurate representation is that ICD therapy is a specific treatment with inherent flaws, but one that produces an effective strategy that reduces sudden cardiac death despite its limitations. There are multiple assumptions made when we implant a device. We assume that the device: 1) is implanted properly, 2) is given to an appropriate patient; 3) is properly programmed; 4) continues to function properly; and 5) that the patient's situation wasn't significantly changed. None of these assumptions are 100% correct.
A recent example of the problem is exemplified in the response to the recent Medtronic recall related to an increased risk of sudden battery failure. I am certain that many more people have been harmed than helped by our collective response and will even die instead of being alive due to the removal and replacement of these devices. Since over 11,000 devices have been exchanged, if we even apply a low estimate of surgical complications it should be expected that there will be over 200 infections alone associated with these procedures. In addition, it is likely that the number of complications of various levels will approximate 500, mostly minor. However, even after the ICD exchange the patients will have a new ICD device capable of failure at a rate not much smaller, and if unfortunate the same or greater than the ICD that was removed. The substitution of one very small risk (battery failure) for another larger risk (infection and extraction) so that the patient can have a device with a similar failure rate does not sound like a good idea to me. As a community we must learn from this experience, consider each situation separately, and craft appropriate responses to the recent Guidant, St. Jude and future recall situations.
There is another truth that needs to be balanced in this polemic. What about the patient's and physician's "right to know". When should we know? What should we know? How would we respond? How many letters would be placed on our desk every day if we received one every time that there was a report to the MAUDE database, updated to let us know when there were duplicate situations, or every time even without a clinical event that a manufacturing change was made to improve a product due to a theoretical future risk from a device observation. Our society has strongly embraced the era of full disclosure. If every time there is a case report, particularly one put into the lay press, of a failure especially with a death, will there be any implantable devices produced, failures investigated or manufacturing processes improved?
It is clear that our current system is not complete and not yet adequate. However, the system is also not completely broken. There is a system in place and overall we have reliable devices, effective ICD therapy and mechanisms that have, until now allowed us to detect, quantify and fix the imperfections in our devices. Part of the problem is the system and part of the problem is our expectations.
Finally, significant contributors to this problem are the interrelated issues of trust and conflict of interest. Clearly, manufacturers have primary incentives to make business decisions that are in the best interests of the organization and their investors. A zero tolerance for defects is impossible, but if we escalate the consequence of a defect then the cost of responding even to small problems will become prohibitive. Secondly, physicians who are consultants to device companies may either have or be perceived to have a higher tolerance for device imperfections, because of their company relationships, even though their patient-centered interest and interpretations are genuine. Practicing physicians also have financial incentives, and although I am reticent to believe that this is a significant factor, there is additional personal income associated with the increased procedural volume during a device recall. Finally, government is extremely risk adverse and will respond to the loudest outrage of its constituents whether it is the right or wrong action. Not to sound too cynical, I truly believe that all players truly want excellent health care at reasonable cost. For this very reason alone we must bring this discussion out in the open and come to reasonable standards that balance the right to know with an understanding that there is no such thing as a perfect device.
Sudden cardiac death is too frequent, too severe of an outcome and too effectively treated with ICDs to let these issues stand unresolved. We must finish the job of proving the role of ICD therapy. The medical community has overwhelmingly demonstrated that ICDs, even with their imperfections, meaningfully reduce total mortality. There is still a long way to go. We must also minimize the morbidity of ICD therapy. Often we think of shocks, infection and inappropriate therapy as the morbidities of device therapy. However, the perception that ICD therapy is either unsafe or unreliable would be the worst morbidity. Safety and reliability are foundational and are the substance that enables us to trust in any therapy, and trust is firmly established only when the entire community develops common expectations. When you think about it, less than perfect is OK, because there is no other choice.
CLICK HERE for subscription information about this journal.
Implantable defibrillator (ICD) therapy is high-profile life and death therapy, set up by a long series of impressive randomized clinical trials dramatically demonstrating improved survival. Such high stakes therapy and striking trial results bring visibility and scrutiny. From the beginning, ICD therapy has had its scientific and emotional aspects, clinical and ethical issues, technical and economic considerations and its supporters and detractors. Lost in all of this is the fact that the therapy is and will continue to be less than perfect. Sometimes, as demonstrated by the recently reported tragic loss of life with ICD device failure, the lack of perfection transcends an academic discussion. Unfortunately, this was not the first time nor will it be the last that we need to confront similar situations. Maisel and his colleagues reported on the increasing number of ICD device failures that produced safety alerts and recalls between 1990 and 2000. As we consider this case and these issues we need to step back much further and see what we think the correct actions ought to be and the consequences of these actions.
There have been scientific and regulatory systems set up to monitor the design, clinical evaluation, manufacturing, marketing and sales of medical devices. It is crucial that both the devices and the systems designed to monitor their use maintain the highest standards of excellence. It is for that purpose -- the highest standards of excellence -- that the US Food and Drug Administration (FDA) has been charged by Congress to regulate and monitor the safety and effectiveness of ICD therapy and all medical devices in the United States. However, at no step is perfection a possibility.
The manufacturing and sale of ICDs are under the scrutiny of the FDA, which is responsible to certify the facilities, general manufacturing processes and every device model. New products are tested and certified, on the basis of data collected during carefully designed and monitored clinical trials, to be safe and effective when used in accordance with their labeling. It is not possible for these systems or the preclinical studies to detect the infrequent deficiencies that are inherent to the therapy much less the defects that develop due to component failures, whether they occur during manufacture, during surgical implantation or are related to the patient's anatomy or disease. These failures are often extremely infrequent and difficult to analyze. Consequently, it is mandated that the manufacturer, the hospital and the physician report all situations where a device malfunction may have played a role. These reports are publically available, in the Manufacturer and Users Facility Device Experiences (MAUDE) database. The electrical overstress situation of the recently publicized death was reported, as were the prior similar events in regard to the same device model. Every day failures of pulse generators, lead insulation, conductor failures, device software and firmware failures and programmer software lockups are reported since the regulations state that every incident should be reported by the physician/hospital and the manufacturer. Unfortunately, the physicians are both the weakest and most important link in this system. On the other hand, the manufacturers take this responsibility very seriously and report every failure for which they are aware. There is no doubt that the current systems need dialogue, review and reform.
The physician community has a profound and important responsibility in this scenario. However, we have hardly been engaged in the process. Frequently we have not reported failures to the MAUDE database, not returned devices or leads for analysis, not understood that with the exponential increase in implants, the absolute failure rate is becoming relatively common, and we have failed to understand that we must constantly examine the performance of the devices in our patients. Just as important, we need to put into perspective these blemishes in device performance and produce a clinically appropriate response. Often our response has been worse than the problems that we have tried to ameliorate. Inherent in all device therapy is a failure rate, which is part of the initial risk undertaken when the device is implanted. Our society and physician community seem to be under the false impression that perfection is a possibility. A more accurate representation is that ICD therapy is a specific treatment with inherent flaws, but one that produces an effective strategy that reduces sudden cardiac death despite its limitations. There are multiple assumptions made when we implant a device. We assume that the device: 1) is implanted properly, 2) is given to an appropriate patient; 3) is properly programmed; 4) continues to function properly; and 5) that the patient's situation wasn't significantly changed. None of these assumptions are 100% correct.
A recent example of the problem is exemplified in the response to the recent Medtronic recall related to an increased risk of sudden battery failure. I am certain that many more people have been harmed than helped by our collective response and will even die instead of being alive due to the removal and replacement of these devices. Since over 11,000 devices have been exchanged, if we even apply a low estimate of surgical complications it should be expected that there will be over 200 infections alone associated with these procedures. In addition, it is likely that the number of complications of various levels will approximate 500, mostly minor. However, even after the ICD exchange the patients will have a new ICD device capable of failure at a rate not much smaller, and if unfortunate the same or greater than the ICD that was removed. The substitution of one very small risk (battery failure) for another larger risk (infection and extraction) so that the patient can have a device with a similar failure rate does not sound like a good idea to me. As a community we must learn from this experience, consider each situation separately, and craft appropriate responses to the recent Guidant, St. Jude and future recall situations.
There is another truth that needs to be balanced in this polemic. What about the patient's and physician's "right to know". When should we know? What should we know? How would we respond? How many letters would be placed on our desk every day if we received one every time that there was a report to the MAUDE database, updated to let us know when there were duplicate situations, or every time even without a clinical event that a manufacturing change was made to improve a product due to a theoretical future risk from a device observation. Our society has strongly embraced the era of full disclosure. If every time there is a case report, particularly one put into the lay press, of a failure especially with a death, will there be any implantable devices produced, failures investigated or manufacturing processes improved?
It is clear that our current system is not complete and not yet adequate. However, the system is also not completely broken. There is a system in place and overall we have reliable devices, effective ICD therapy and mechanisms that have, until now allowed us to detect, quantify and fix the imperfections in our devices. Part of the problem is the system and part of the problem is our expectations.
Finally, significant contributors to this problem are the interrelated issues of trust and conflict of interest. Clearly, manufacturers have primary incentives to make business decisions that are in the best interests of the organization and their investors. A zero tolerance for defects is impossible, but if we escalate the consequence of a defect then the cost of responding even to small problems will become prohibitive. Secondly, physicians who are consultants to device companies may either have or be perceived to have a higher tolerance for device imperfections, because of their company relationships, even though their patient-centered interest and interpretations are genuine. Practicing physicians also have financial incentives, and although I am reticent to believe that this is a significant factor, there is additional personal income associated with the increased procedural volume during a device recall. Finally, government is extremely risk adverse and will respond to the loudest outrage of its constituents whether it is the right or wrong action. Not to sound too cynical, I truly believe that all players truly want excellent health care at reasonable cost. For this very reason alone we must bring this discussion out in the open and come to reasonable standards that balance the right to know with an understanding that there is no such thing as a perfect device.
Sudden cardiac death is too frequent, too severe of an outcome and too effectively treated with ICDs to let these issues stand unresolved. We must finish the job of proving the role of ICD therapy. The medical community has overwhelmingly demonstrated that ICDs, even with their imperfections, meaningfully reduce total mortality. There is still a long way to go. We must also minimize the morbidity of ICD therapy. Often we think of shocks, infection and inappropriate therapy as the morbidities of device therapy. However, the perception that ICD therapy is either unsafe or unreliable would be the worst morbidity. Safety and reliability are foundational and are the substance that enables us to trust in any therapy, and trust is firmly established only when the entire community develops common expectations. When you think about it, less than perfect is OK, because there is no other choice.
CLICK HERE for subscription information about this journal.
Source...