In past times, what surgeons did was basically removing infected body parts and things that didn’t belong in our bodies, like tumors and foreign bodies. For most of the history of the practice of medicine, surgeons were admired for their manual dexterity rather than for their skill and intelligence in diagnosing and treating disease. The surgeon who removed your appendix was supposed to be sure to hold on to the slippery bit of gut and not lose it in your innards, and then to stitch you up quickly and neatly. Until just about a hundred years ago, surgery was a procedure of last resort. But then, with great improvements in anaesthesia and sterilization of surgical equipment as well as the operative theater, elective surgery began to take off. Surgeons operated to address conditions that were not immediately life-threatening, but conditions that had negative effects on quality of life. And patients were willing to entrust themselves to surgeons to have those conditions remedied, even if it meant that they would go “under the knife.”

Trust in surgery has been strongly enhanced by a number of technical innovations that facilitated the work of surgeons. Imaging has been transformative. Ultrasound, magnetic resonance imaging (MRI), and computerized tomography (CT) scans let surgeons get a clear picture of the inside of the patient’s body before they pick up their scalpels. Laparoscopic surgery lets the surgeon insert the cutting instrument into the patient’s body by means of a thin, flexible tube that permits the surgeon to see and cut at the same time, without making more than a very small incision. Endoscopes, which have long been used in colonoscopies, now have laser cutting facilities, so that the removal of polyps from the colon is now almost a routine matter.

The surgical procedures we’ve been talking about up to this point all consist of taking something out. However, what about surgery in order to put something in? The earliest surgeries of this kind were performed in order to replace body parts that were subject to a great deal of physical stress, and under certain circumstances just wore out.

The earliest attempt to replace a hip joint took place in Germany in 1891, when Professor Themistocles Glück advocated the use of ivory to replace parts of the hip joints of patients whose joints had been destroyed by tuberculosis. Later, surgeons experimented by placing tissues such as the fascia that envelopes the whole calf muscle or parts of pig bladders between the surfaces of the arthritic hip joint.

The part of the hip joint that is most subject to frictional wear is the ball-and-socket joint. The ball is at the top of the femur, which is the upper leg bone, while the socket, called the acetabulum, is in the hip bone itself. An English surgeon, George McKee, began in the 1950s to use a one-piece cobalt-chrome socket as a new acetabulum. This prosthesis worked fairly well. A recent study reported that it has a 28-year survival rate of 74%. However, by the mid-1970s, this method fell out of favor. The friction, under considerable pressure, led to the formation of metal particles which eventually impeded the performance of the joint.

The types of hip prostheses used in the present day were first introduced in the 1960s. They consisted of three parts. The femoral stem, which includes the ball part of the joint, is metal. The acetabular part – the socket – is polyethylene. The ball and socket are smaller with less surface area, which reduces wear. And the sections are fastened together using acrylic bone cement, which is what dentists use.

The success of hip joint replacement has been echoed in many other parts of the body. In the year 2014 (the most recent year for which I have found reliable data), surgeons replaced 522,800 hips, 723,100 knees, 90,000 shoulders, 15,000 elbows, 16,000 finger joints, 12,000 toe joints, 2,000 ankles, and 2,000 wrists – a total of nearly 1.4 million joint replacement procedures. Estimates are that by the year 2030, there will be four million such procedures per year.

In addition to these implants which are essentially mechanical in nature, there are electronic implants and biologic implants. Cardiac pacemakers, which are clearly lifesaving, have been around for about 75 years. The earliest pacemakers were huge machines that were connected to the patient’s heart by means of a needle. These were clearly not implantable. The first implantable cardiac pacemaker was surgically implanted in Arne Larsson on 8 October 1958. It was about the size and shape of a can of Kiwi shoe-polish. That first pacemaker lasted 8 hours. But Larsson, who was having 20 to 30 Stokes-Adams heart attacks per day, kept getting successive pacemakers, eventually 11 different devices in total. (A sudden decrease in heart rate can cause a person to lose consciousness; this is called a Stokes-Adams attack). Larsson died, more than 40 years after receiving his first pacemaker, in 2001 at age 86.

Implanted insulin pumps are not lifesaving, since persons with insulin-dependent diabetes are able to give themselves insulin injections without the need for any kind of automatic device, and external insulin pumps have been available. But implanted pumps are convenient and probably also salutary, since individuals with diabetes who self-inject can miss doses, give themselves too little, or – more dangerous – too much insulin. The first implantation of insulin pumps, made by an outfit called MiniMed, took place in 1980.

The picture changed considerably when, in 2001, MiniMed was bought by Medtronic, which discontinued the implanted insulin pump in 2007. This left those persons who had implanted insulin pumps in a precarious situation. Physicians were able to refill the pump, which had to be done every three months. But care and maintenance of the device was up to the persons with the implants. A number of possible alternatives emerged, but the overall picture has been confusing. Medtronic has gone on to introduce its own implanted insulin pumps, but these, in turn, were recalled in October 2021.

The predicament of a patient who has come to rely on a medical device – especially an implanted medical device – which then ceases to be supported by the manufacturer, was described in a Doc Gumshoe piece about epilepsy, which posted on 4 November 2021. The piece described a woman with epilepsy who had an implanted medical device which could accurately predict an epileptic seizure sufficiently in advance of its onset so that this woman could take necessary precautions – i.e., not be driving a car or going down the stairs when the seizure struck.

Here’s what I said about the device:

“A small US company called NeuroVista developed a device specifically to detect such signs and communicate them to the patient. It consisted of 16 electrodes implanted in specific areas of the brain, which communicated with an implanted device which had the capacity of translating the impulses into the brain wave patterns as recorded on an EEG, and assessing whether these brain wave patterns were signs of an impending seizure. In order to accomplish this, a data set of brain waves patterns from 200 patients were recorded and analyzed, such that the patterns that predicted a seizure could be recognized in real time. The implanted device had the capacity to communicate with an external device which could then transmit information to the NeuroVista computer system, which would in turn send a warning to the patient if a seizure was impending. The external device was about double the size of a cell phone, which the patient could wear in something like a holster. The patient was supposed to keep this device on his or her person at all times. The seizure warning consisted of a flashing red light and a fairly loud beep.”

The seizure advisory system (SAS) designed by NeuroVista worked reasonably well in 10 of the 15 patients enrolled in a study. The patient discussed in the Doc Gumshoe piece received timely seizure warnings on several occasions and came to rely on the SAS. She lived with this implanted device for three years, but then NeuroVista ran out of funding for the continued research on the device and simply stopped supporting the SAS. Having had the SAS surgically implanted, the patients now had to experience a second brain surgery to have the device removed.

As it happened, the particular patient in the Doc Gumshoe piece was able to avoid the consequences of seizures even after the device had been removed. Over time, she had learned to associate her own perceptions of what was going on in her brain with the warnings, such that she became able to anticipate impending seizures on her own and avoid the worst consequences.

However, the discontinuation of the NeuroVista SAS system serves to illustrate the tricky area that patients enter when they agree to have devices implanted into their bodies. Most of the time, they are taking the prospective benefit of the implant on faith. It may be the case that they have little choice – the alternative may be to do without the treatment course that could deliver significant benefit. In my case – I have had two total knee replacements – the likelihood of success is supported by the statistics, specifically that hundreds of thousands of total knee replacements are performed every year in the US, and failures due to deficiencies in the implants are very rare.

But those mechanical implants are simple affairs, compared with the implants that are meant to detect and interact with complex dynamic body functions, which, like the brain activity that predicts epileptic seizures, can be detected electronically.

Medtronic is by far the leading global manufacturer of electronically assisted medical devices. The company produces a multitude of devices in 16 categories covering virtually every medical specialty, including 18 ear, nose and throat devices, 19 general surgery devices, 22 neurological devices, and 20 spinal and orthopedic devices.

The Medtronic roster of devices has been affected by safety problems. In 2021 alone, the FDA posted 10 Class I recall notices for Medtronic products, including the above mentioned insulin pumps, but also several recalls for their HeartWare Ventricular Assist Device, which was pulled from the market in June of that year.

At the moment, a congressional oversight committee is investigating the FDA’s oversight of that HeartWare device, specifically citing safety issues. The device, designed to treat patients with severe heart failure, stopped meeting key federal standards in 2014, but the FDA took no decisive action. Thousands of persons in the US continued to be implanted with the pump. By the end of 2020, the FDA had received reports of more than 3,000 deaths in individuals implanted with the HeartWare Ventricular Assist Device. It should not be assumed that all of those deaths were due to pump malfunctions, although some specific cases were cited when a patient died during – and likely as a result of – pump failure. The FDA’s oversight – or lack of oversight – of medical devices is being increasingly questioned.

In spite of those incidents, there is very little doubt that the development of these devices has been transformative for medical practice. However, determining the effectiveness and safety of these implanted devices is a difficult process, and the results are far from certain, in comparison with the verification process used in bringing drugs to market.

As you know, getting drugs approved by regulatory bodies such as the Food and Drug Administration (FDA) in the US requires going through a process of clinical trials, in which the efficacy and safety of the candidate drug is compared with that of both one or more alternative drugs and with a placebo. The trial needs to be conducted in a sufficient number of subjects such that the results cannot be attributed to chance. If the probability is that the results are just due to chance, the trial is said to lack statistical significance. The minimum threshold for statistical significance is that the probability that the results are due to chance is less than 1 out of 20, which is expressed as P < 0.05. If the probability that the results were due to chance were 1 out of 100, that would be expressed as P = 0.01. In some clinical trials, the probability of chance is even lower. P values as low as 0.001 are not unknown.

This cannot happen in studies of an implanted medical device. It would be out of the question to compare a cardiac pacemaker with a non-working “placebo” cardiac pacemaker. Studies of medical devices are of necessity observational. The performance of the device in a patient is compared with the general population. In some cases, a new candidate device may be compared with the performance of an already existing device, but the comparison is not likely to be carried out in a typical double-blind clinical trial fashion, in which neither the treating physician nor the patient knows which device is being tested.

The FDA as we know it came into existence in 1938 with the passage of the Federal Food, Drug, and Cosmetic Act, which authorized product oversight by the FDA. However, the FDA did not regulate medical devices of any kind until 1976, after a great many complications resulting from a contraceptive device known as the Dalkon Shield came to light. The Dalkon Shield was supposedly a huge improvement over birth control pills. It was inserted into a woman’s uterus for pregnancy prevention. As many as 200,000 American women testified that they were injured by the device and have filed claims against its maker, the A.H. Robins Company, which sold it to an estimated 2.5 million women during a four-year period in the early 1970’s. Lawyers representing the women say the device was dreadfully dangerous, causing serious pelvic infections leading to infertility or even death.

The need for some kind of regulation of medical devices was emphasized by the Dalkon Shield story, which attracted a great deal of attention. In response, the Medical Device Amendments were enacted in 1976, which assigned ultimate regulatory authority for devices to the FDA. A three-category risk-based classification system was established for devices, and two regulatory pathways were created.

The classification system, which today remains the principal means of regulating implants, divided medical devices into three classes. Class I consists of “low-risk items, such as bandages and material for surgical stitches. Class II includes “medium-risk” items, such as most artificial knees and hips. Class III, the “high-risk” items, are pacemakers, ventricular assist devices, implanted insulin pumps, and such devices.

The FDA created two regulatory pathways for devices. For Class I and II devices, pre-market notification, known as the 501(k) pathway, is sufficient, meaning that the manufacturer needs to do no more than notify the FDA that their intended device is substantially equivalent to a predicate device. According to the FDA, a predicate device is a previously legally marketed device cleared through a regular path, which is used to determine substantial equivalence. The FDA has an 11 step process for the determination of substantial equivalence, but it does not involve anything like a clinical trial. In other words, when a manufacturer comes up with something like an artificial hip joint, perhaps made of some material that promises less frictional wear, all that the manufacturer has to do is demonstrate that this new candidate artificial hip is about the same (“substantially equivalent”) as artificial hip joints already on the market. (Most of the information in the next few paragraphs comes from a paper published in the American Medical Association’s Journal of Ethics (Pisac A, “FDA Device Oversight from 1906 to the Present,” AMA J Ethics 2021;23.9:E712-720)

In 1990 the Safe Medical Devices Act (SMDA) stated that as a standard for clearing devices using the 501(k) pathway, the manufacturer must provide evidence that the device has the same intended use as a predicate device and also have the same technical characteristics. If these technical characteristics are different, the manufacturer must provide performance data to the FDA demonstrating comparable safety and effectiveness.

For Class III devices – the highest risk category – premarket approval (PMA) is required. The manufacturer needs to provide some safety and effectiveness data. In addition, the SMDA introduced postmarket surveillance requirements for the manufacturers, mandated adverse event reporting, established penalties for violations, and granted the FDA authority to recall devices.

Those requirements strike me as completely reasonable. Of course adverse events should be reported, and if a device was to be associated with a large number of adverse events, of course the FDA should have the authority to get those devices off the market. But to the medical device industry, and to many in the healthcare field that made use of those devices in patient care, the requirements were concerning. Stringent regulation could work against innovation and also deprive patients of access to devices that in some cases were lifesaving and in many cases made life much less burdensome.

As a result, subsequent legislation introduced an approach to premarket review that was meant to be “least burdensome.” A program, entitled “De Novo,” was introduced for moderate risk devices, and manufacturers were relieved of some of the cost burden of the PMA process. In 2016, more efficient and flexible approaches were outlined in the 21st Century Cures Act, which provided support for breakthrough devices and expanded criteria for humanitarian-use devices.

One of the factors that affect medical devices in ways that do not affect drugs is that a device can change over its lifetime, due to wear and tear in normal use. These changes can have an effect not only on the effectiveness of the device, but also on its safety. Metal-on-metal hip implants can lead to the accumulation of toxic metal particles in the joint. Cardiac defibrillator leads can fail, breast implants can cause anaplastic large-cell lymphoma, and gynecological mesh implants can cause complications. In patients with peripheral arterial disease, stents coated with paclitaxel have been shown in some studies to increase mortality by as much as 38%. (Note, other studies have reported no increased risk of mortality, but an increased risk of limb-amputation.)

In spite of these limitations, the PMA process has led to devices of much greater reliability than the 510(k) pathway. Devices approved via the 510(k) pathway between 1992 and 2012 were 11.5 times more likely to face recalls than devices approved through the PMA process. Only 16% of 510(k) devices were supported by publicly available documentation of scientific evidence used to establish substantial equivalence, in spite of the FDA requirements to provide such evidence. The 501(k) pathway is how most medical devices enter the market; for example, in 2017, 3,173 devices used that pathway. This was 82% of the total number of devices entering the market that year.

The safety of medical devices was the subject of a book, The Danger Within Us, by Jeanne Lenzer, published in 2017. Ms Lenzer observed that in 2015 the FDA received about 16,000 reports of deaths associated with medical devices. A study by the Government Accountability Office estimated that 99% of such adverse event go unreported, and stated that “the more serious the event, the less likely it was to be reported.” This means that the total number of deaths associated with medical devices could be much, much higher.

Between 2003 and 2010, more than 90,000 patients worldwide, about one-third of them American, were given the DePuy Articular Surface Replacement hip implant. In the US, this device was recalled in 2010 because metal particles created by friction in the joint could lead to intense inflammation, sometimes resulting in the destruction of tendons, ligaments, muscles, and bone. And, of course, the failure of the hip implant required further surgery, including the insertion of a new artificial hip, which the patient and surgeon alike hoped would cause no problems. But the hip bone itself was no longer intact. Replacing the hip implant was essentially the repair of a repair.

There are many, many more instances of medical device failures. For example, in 2007 the FDA recalled the Sprint Fidelis defibrillator, which had been implanted in more than a quarter million patients, two-thirds of them in the US. In 2013, 33,000 filters which had been implanted in the inferior vena cava were recalled. These filters were supposed to prevent blood clots from reaching the heart. Instead, it turned out, the filters caused blood clots. In 2015 Medtronic paid a $2.8 million fine to the Justice Department for selling a spinal-cord stimulator without going through the FDA’s premarket approval (PMA) process. Medtronic admitted that during a five-year period they had failed to report more than one thousand adverse events related to the spinal-cord implant.

A problem specific to implanted medical devices is that removing them requires a complex surgical procedure, and sometimes the complete removal is not possible. The wire leads that connect the devices to the patients may lodge in the heart, the brain, any of several nerves. They often become embedded in scar tissue and cannot be surgically removed. They interfere with CAT scans, and they may continue to be active electrically even when the device itself has been removed. Patients can be affected by these electric impulses, which can continue to cause pain.

One would think that multimillion dollar fines, such as the fine imposed on Medtronic, would serve as a warning to the device manufacturers that they needed to toe the line. This does not appear to be the case. Manufacturers probably view the fines as a cost of doing business. And the business they are doing is quite profitable. The profit margins of medical device manufacturers are generally higher than those of the major pharmaceutical companies – about 25% of their revenues are profit. In 2014, the overall revenues of the medical device manufacturers totaled over $136 billion.

Given the huge financial incentives, it is very hard to come up with an effective way to regulate the medical device area. Some commentators have recommended randomized clinical trials with appropriate control groups, with one group using sham devices whenever possible, to mimic the placebo cohort in a drug trial. “Whenever possible” here is a gigantic disclaimer. Are we to install sham artificial hips, or artificial pacemakers, or artificial insulin pumps? I would argue that the best the manufacturers could do is use the devices in volunteers and carefully report the results.

Commentators have also pressed to give patients the right to sue if they are harmed by a medical device. I am not a cynic, but it seems to me that such lawsuits would only be successful if the plaintiff could demonstrate that the manufacturer of the device was clearly aware of the danger posed by the device but concealed it from the patient, as in Medtronic’s failure to report the adverse events related to its spinal cord stimulator. Manufacturers could easily protect themselves by using the kind of language common in the Prescribing Information material given to patients with prescription drugs. Patients are notified that the drug in question has been associated with a long list of adverse effects, most of which occur rarely. But the patient has been warned.

Congressman Jerrold Nadler has attempted to pass a bill which would insure that information uncovered during product-liability suits cannot be kept secret under nondisclosure agreements. These non-disclosure agreements prevent attorneys from notifying the FDA of specific harms that their patients have suffered in connection with a specific medical device. Device manufacturers are, understandably, reluctant to reveal such information, because it is not always evident that the device caused the adverse event. However, it is legitimate information which should be available to persons who are considering whether to agree to go ahead with having the device implanted.

To reduce this matter to the simplest level: it is certain that no legislation can effectively prevent patients from being harmed by a medical device; however, an achievable goal of such legislation would be to prevent patients from being deceived.

The above-mentioned writer Jeanne Lenzer points out that independent studies of medical devices are much more likely to note potential harms from their use than studies conducted by the manufacturers. In an interview on National Public Radio’s “Fresh Air” she said:

“There are multiple studies that look at behavior and outcomes and interpretations of medical studies when they’re conducted by industry versus by truly independent sources. And invariably, what we find is that industry accentuates the positive and eliminates the negative. I mean, they have repeatedly been found guilty of suppressing bad outcomes, and they get fines for this regularly. In my book, I list numerous fines that companies have paid over and over again for concealing deaths and bad outcomes. And they just consider that part of the cost of doing business.”

Notwithstanding, the benefits brought to immense numbers of patients by medical devices are enormous. They save lives and they greatly ease the burdens of day-to-day living. They benefit the young and the old alike. My primary care physician years ago told me about a patient of his for whom he had implanted a replacement pacemaker at the age of 95. And teenagers with type 1 diabetes are frequently fitted with implanted insulin pumps.

Implantable artificial hearts?

I cannot imagine anyone willing to accept a procedure that removes his/her heart and replaces it with an artificial heart, unless this procedure was the only alternative to a disagreeable death. Most candidates for an artificial heart are on the waiting list for a heart transplant, but the number of persons in need of heart transplants greatly exceeds the number hearts available for transplantation. Therefore, there is a definite need for artificial hearts.

The evolution of artificial hearts has been in progress for about half a century. So far, the implanted artificial heart must be connected to an external pumping mechanism. One of the early, marginally successful procedures took place in 1982, when the Jarvik-7 heart was implanted into Barney Clark, a 61-year-old dentist whose heart had been operating at about one-sixth normal capacity. The implanted heart was connected to a 400-pound console and pumping mechanism. Clark lived for 112 days on his implanted heart.

A couple of years later, an improved version of the Jarvik-7 heart was implanted into 52-year-old William Schroeder. He was able to move out of the hospital and occasionally use a portable pumping unit with three hours of battery life, which permitted him to go on drives. Schroeder lived 620 days with his implanted heart.
The Jarvik hearts were replaced by implantable hearts made by HeartMate, approved by the FDA in 1994. These still needed to be powered by external pumping units, and they had a life of about a year and a half, after which they would need to be replaced, but the pumping units are much smaller and transportable. For most of the recipients of HeartMate devices, the artificial hearts are interim measures until the recipient gets to the top of the waiting list for a heart transplant.

There have been other attempts to develop and market artificial hearts. Symbion, the company that Jarvik helped to found, lost its FDA approval for the Jarvik-7 heart in 1990, but the technology was acquired by another firm, which, however, exhausted its funding in 2001. But two heart surgeons and a biomedical engineer raised enough venture capital to buy rights to the system, which they rebranded as the SynCardia Total Artificial Hearts (TAH).

SynCardia Systems (Tucson, AZ) is the only company that today actually makes artificial hearts. The SynCardia TAH is driven by external devices which provide the pumping energy to drive the blood through the circulatory system. Since the amount of blood that is pushed out of the heart and into the arteries is about five or six liters every minute, a considerable amount of pumping force is required. The TAH pumping units come in two sizes – one, the size of a small refrigerator, and the other the size of a toaster. These drivers need to be serviced every few months, so while one driver is being brought back to full operation, a caregiver unplugs the line that connects the driver to the implanted TAH and switches it to the other driver. The switch needs to take place very quickly, so that the TAH does not skip too many beats.

SynCardia is working on a next-generation heart that would use a new, battery-driven pump engine which would be implanted in the patient and connected to the artificial heart. In the meantime, they continue to manufacture and market the current model, which saves many lives.

Another outfit, Bivacor (Houston, TX) is developing an implantable heart in which the blood flow is controlled by a rotor system which at the same time pumps venous blood to the lungs to be recharged with oxygen and also pumps the blood returning from the lungs into the arterial system. At this point, the Bivacor heart is powered by an external battery, but the design team is cautiously working on a wireless, rechargeable heart.

In comparison with the extreme complexity of implanted artificial hearts and the high risk of the procedure, my two knee replacements were almost trivial. A couple of days in the hospital, a few weeks of physical therapy, and I was just about as good – mobile and pain-free – as I had been several years before the procedures. The great majority of medical devices should be like my two TKRs. By and large, medical devices contribute enormously to the lives and health of the public at large. But safety issues continue to emerge, and caution continues to be the watchword. All parties concerned – the manufacturers, the healthcare workers that deliver the goods, whether the surgeons that implant the devices or the physicians that recommend them, and the patients on the receiving end – need to keep a sharp eye on safety. Medical devices can certainly be extremely beneficial and even life-saving. But they are not exempt from the risk of causing harm.

*****

Not one single word about COVID-19 in this piece. You probably know these factoids, but the BA.2 strain is now the principal infective strain in the US and many other countries. However, the risk of serious illness and death appears to be way down, especially among the vaccinated and boosted. And in the US, booster shots for the boosted are now recommended and available for people aged 50 and up. Since Doc Gumshoe is in that category, he will certainly get a booster. There’s more COVID news, for instance that Biden is pushing for more investigation of “long COVID.” I’m not sure COVID is worth another whole piece, but I’ll try to keep you posted.

Stay well! Do please let me know what you want me to dig into. Best, Michael Jorrin (aka Doc Gumshoe)

[ed note: Michael Jorrin, who I dubbed “Doc Gumshoe” many years ago, is a longtime medical writer (not a doctor) and shares his commentary with Gumshoe readers a couple times a month. He does not generally write about the investment prospects of topics he covers, but has agreed to our trading restrictions.  Past Doc Gumshoe columns are available here.]