My Masters' Thesis, MSQA, 2009.

Hi Everybody!  I've decided to post my Masters' Thesis from 2009, when I received my Masters' Degree in Quality Assurance from Cal State Dominquez Hills.  No need to have this document just sitting in my folders; better to get it out there, and maybe it can be to some use.  I spent some 10-11 months writing this Thesis, and I'm quite proud of my accomplishment.  Enjoy!  

CHAPTER 1

INTRODUCTION

In July of 1979, a Bjork-Shiley Convexo-Concave prosthetic heart valve failed catastrophically when the outflow strut had fractured, which resulted in the escape of the tilting occluder disc (see Figures 1, 2, and 3). The incident, which took place in Oslo, Norway, resulted in the death of the 72-year old female patient (Brubakk, Simonsen, Kallman, & Fredriksen, 1981). She had received her new mitral valve in April of 1979, the same month that the United States Food and Drug Administration (USFDA) had approved the valve for sale in the United States (US; The Bjork-Shiley Heart Valve, 1990). This would be one of many reports of failures of the outflow strut; by December 2003, there had been 633 reports of strut fractures, wherein two-thirds of the fractures resulted in the death of the patient (Blot et al., 2005).

                                                                              

Figure 1. Outflow strut fracture, as seen from the outflow side. Source: Article by O’Neill et al., in The New England Journal of Medicine (1995).

Figure 2. Outflow strut fracture, as seen from the inflow side. Source: Article by O’Neill et al., in The New England Journal of Medicine (1995).

 

 

Figure 3. Outflow strut fracture (white arrow). Note the bottom of the Pyrolytic disc, wherein the outflow strut hook fits into the recess. Source: Article by van Gorp et al., in Radiology (2004).

Between May of 1979 and November of 1986, some 86,000 Bjork-Shiley Convexo-Concave heart valve prostheses were implanted in patients, approximately half in the US and the rest in Europe, Australia, Canada, and Japan. There were two distinct designs, depending on the opening angle of the occluder disc; the 60-degree and the 70-degree heart valve (Stevenson, Yoganathan, & Franch, 1982). The number of 60-degree models sold was about 82,000. In addition, another 4,000 of the 70-degree valve were sold outside of the US, since the 70-degree was never approved here. The 60-degree valve, especially in the 29-33 mm mitral size, would fracture about 1.0% of the time; however, the 70-degree would fracture in 5.0 to 10.0% of implanted valves (sales of the 70-degree valve were discontinued in 1983; The Bjork-Shiley Heart Valve, 1990). Due to the outflow strut fractures (see Figure 4), the Bjork-Shiley Convexo-Concave valve would become one of the most problematic medical devices ever introduced and the valve would change the regulatory landscape of the medical device industry.

 

 

 

Orifice ring and integral inflow strut   Completely fractured outflow strut

 

Figure 4. Valve with complete outflow strut fracture in patient. White arrow indicates position of the escaped disc. Note that the round ghost image to the left of the spine is not the disc. Source: Article by O’Neill et al., in The New England Journal of Medicine (1995).

 

Donald Shiley had started Shiley, Inc. in 1966, after having served as chief engineer at Edwards Laboratories. By 1968, Don Shiley had developed the Kay-Shiley floating disc heart valve in partnership with Dr. Jerome Kay, of Los Angeles. The first Bjork-Shiley tilting disc heart valve was developed with Dr. Viking Bjork, a Swedish thoracic surgeon, in the late 1960s. The tilting disc was followed by the development of the Bjork-Shiley Convexo-Concave around 1975 (Bjork, 1978; Derloshon, 1983; Westaby, 1997).

In 1979, Shiley Inc. was sold to HomeMedica, a subsidiary of the drug giant Pfizer. Due to hundreds of law suits, bad press and flat sales, Pfizer sold Shiley Inc. in 1992 to Sorin Biomedica, a subsidiary of the Italian conglomerate Fiat (“Mergers, Acquisitions,” 1992). However, Sorin wanted no part of the Bjork-Shiley Convexo-Concave and Pfizer was left with hundreds of millions in legal liabilities (Tidmarsh, 1998).

As of 2005, due to the high mortality rate among people who receive heart valve prostheses, the world-wide population of survivors has been reduced to some 22,000 people (Blot et al., 2005). This is still a large number of people, who in some cases are faced with weighing the risk of sudden valve failure versus the risk of re-operation. As early as 1986, Dr. Bjork and others had discussed the prophylactic removal and exchange of the Bjork-Shiley Convexo-Concave valve, but made no recommendation (Lindblom, Bjork, & Semb, 1986). Later, after 1990, many authors tried to put together decision models that could aid in the determination of who should receive a new replacement valve; however, none of the models were truly satisfactory (e.g., Ericsson et al., 1992; Kallewaard, Algra, Defauw, & van der Graaf, 1999; Omar et al, 2001; Omar, Morton, Murad, & Taylor, 2003; Steyerberg, van der Meulen, van Herwerden, & Habbema, 1996; van der Graaf, de Ward, van Herwerden, & Defauw, 1992; Walker, Funch, Bianchi, & Blot, 1997; Walker, Funch, Sulsky, & Dreyer, 1995).

To the casual reader, the numbers, dates, and names may not have any significance. Nevertheless, for professionals involved with medical devices, the parallel history of the Bjork-Shiley Convexo-Concave heart valve failures and the ever increasing Food and Drug Administration (FDA) involvement and regulations serve as a lesson in ethics, device design, and politics. Therefore, a study of Shiley Inc., the Bjork-Shiley Convexo-Concave heart valve prosthesis, and the FDA during the years from 1970 to 1996 can illustrate how medical devices should be introduced into the market place; and once on the market, how a manufacturer and a regulatory agency should react to catastrophic failures. Furthermore, this study will answer the central question of this thesis, as outlined below.

                                                                     

Statement of the Problem

Can the new Quality System Regulation, including design controls, as formulated in 21 Code of Federal Regulations (CFR) 820 (enacted in 1997), ensure that another poorly designed medical device such as the Bjork-Shiley Convexo-Concave heart valve will not enter the market?

In 1990, the FDA enacted the Safe Medical Devices Act. This act gave the FDA the authority to add pre-production design controls to the Current Good Manufacturing Practices regulation. One of the main reasons for adding design controls was the belief that a large proportion of medical device recalls were due to faulty design (U. S. Department of Health and Human Services, “Device Recalls,” 1990). The Current Good Manufacturing Practices regulation, which was enacted in 1978, had focused mainly on production and inspection, as a means for controlling the quality of medical devices. Although, as the Bjork-Shiley Convexo-Concave heart valve problem had shown, focusing on the control of manufacturing and inspection of a medical device was not enough (The Bjork-Shiley Heart Valve, 1990).

Consequently, by October, 1996, the revised Current Good Manufacturing Practices Regulation, 21 CFR 820, was published, which became known as the “Quality System Regulation.” The added design control elements were largely borrowed from the ISO 9001:1987 quality management system standard and contained such items as design planning, design input, design output, design verification, design review, and design validation (Trautman, 1997). The main purpose of design controls is to introduce a mechanism that will ensure that a medical device is designed in a methodical and controlled fashion, such that only a safe and effective device is introduced to the market. At the heart of design control is design planning, which should guide the whole design process. Given a robust design plan, a medical device design effort will go through all of the main phases, with numerous interspersed gates and reviews. After the design has been completed, the design effort is summarized in the design history file, a living document that should contain the full history of the device, from design inputs to design transfer.

                                                                                       

Purpose of the Study

The intent of this thesis is to investigate the above stated problem and to answer either in the affirmative that design controls are effective, or conclude that the regulation is ineffective. In this investigation, the failure of the Bjork-Shiley Convexo-Concave valve will be used as an example of a medical device that did not go through a proper design control effort, due to the age of the device. One must keep in mind that when the Bjork-Shiley Convexo-Concave heart valve was introduced to the market (1979), design controls as we know them today did not exist; consequently, the failure of the valve cannot be entirely blamed on Shiley. Also, by researching the Bjork-Shiley Convexo-Concave device, questions of business ethics and how personal choices can have a tremendous impact on others will be illustrated.

 

 

Theoretical Basis and Organization

            This study is an attempt to either assure the public that the current Quality System Regulation and especially the design control component, will ensure that only safe and effective medical devices are designed, manufactured, and marketed, or show that the design control regulation is ineffective. The study will focus on the history of the Bjork-Shiley Convexo-Concave valve, the various FDA guidance documents published between 1980 and 1997 and the 510(k) filing that Shiley submitted to the FDA in 1978 (FDA 510(k) file no. K780772). Also, the FDA-prepared “Statement of Safety and Efficacy” will serve as a proxy for the original Shiley Pre-Market Approval file for the Convexo-Concave heart valve (FDA PMA number P780008). The central hypothesis of this study, wherein it is assumed that design controls are effective and could have prevented the Bjork-Shiley Convexo-Concave heart valve from entering the market, will be tested by performing a gap analysis, consisting of the following elements:

1)      An outline of the design control requirements;

2)      the interpretation of the requirements;

3)      an evaluation of the fulfillment of these requirements using the information provided  by Shiley in the 510(k) filing and the “Statement of Safety and Efficacy” and

4)      evaluating a “what if” scenario, assuming that Shiley would have fulfilled the design control requirements and if so, would the additional requirements have prevented the Bjork-Shiley Convexo-Concave from entering the market place.    

 

Limitations of the Study

            This study consists of a thorough documentation review of the readily available literature, such as journal articles, material available from the FDA website, newsletters, and books. Furthermore, it will review Shiley’s 1978 510(k) filing (obtained from Freedom of Information Services, a private information provider located at 701 Quince Orchard Road, Gaithersburg, MD 20878) and the FDA-prepared “Statement of Safety and Efficacy” which summarized the testing done by Shiley in support of their original Pre-Market Approval Filing (also obtained from FOI Services). It will not involve interviews. It will only focus on Shiley, Inc. and the FDA. Due to the amount of time a request takes through the Freedom of Information Act (up to 18 months from request to fulfillment by the FDA) it will not review Shiley’s original Pre-Market Approval application that was filed in November of 1978 (only the “Statement of Safety and Efficacy,” as outlined above). Lastly, it will not involve any other entities or medical device manufacturers; however, it must be noted that from 1979, Shiley Inc. was a subsidiary of Pfizer (The Bjork-Shiley Heart Valve, 1990; Derloshon, 1983).

                                                                                     

Definition of Terms

Anticoagulant:            A pharmaceutical agent which impedes the clotting of the blood

Aortic valve:               The heart valve that connects the heart and the aorta; ensures that oxygen-rich blood is flowing through the circulatory system, without flowing back into the heart 

Calcification:               A process whereby calcium phosphate form within implanted or natural tissue

Class III device:          The highest risk medical device as defined by the FDA. A device which supports life, prevents health impairment, or presents an unreasonable risk of illness or injury to a patient

Convexo-Concave:     Pertaining to the shape of the disc (Occluder, see below) in the Bjork-Shiley Convexo-Concave heart valve – convex on the outflow side of the disc and concave toward the inflow side of the disc. Shiley claimed that the shape of the disc improved the hemodynamics of the heart valve

Delrin:                         DuPont’s trade name for a formaldehyde-based polymer (polyacetal) used to form the disc in the first Bjork-Shiley tilting disc heart valves

Dendritic structure:     A mineral crystallizing in a pine tree configuration as a result of cooling after heat application (welding)

Diastole:                      The normal rhythmically occurring relaxation and dilation of the heart cavities during which the cavities are filled with blood

Embolism:                   A detached blood clot that closes a blood vessel; a common complication of mechanical prosthetic heart valves

Endocardium:             The innermost layer of tissue that lines the chambers of the heart

Explantation:               The operation of removing a prosthetic heart valve from an implantee

FDA:                           The United States Food and Drug Administration, also known as “The Agency”

Flange:                        Metal ring used as a housing for the disc (occluder, see below); the outside of the flange is surrounded by a cloth-covered nylon ring (sewing ring, see below), which is used in anchoring the prosthetic heart valve to the heart 

GMP:                          Good Manufacturing Practices (1976). The framework for developing individualized quality assurance programs; includes controls over manufacturing specifications and processing procedures, device components, packaging, labeling, manufacturing equipment, and records

Hemodynamics:          The study of the circulation of blood. In reference to a heart valve, the flow pressure characteristic of the valve

Implantee:                   Recipient of a prosthetic heart valve

Inlet strut:                    An integral strut of the convexo-concave heart valve upon which the disc (occluder, see below) pivots. The inlet strut was machined as a part of the flange or valve ring in the Convexo-Concave valve

Laminar:                      A type of flow which is smooth, with parallel stream-lines

Mitral valve:                The heart valve that connects the left atrium and the left ventricle. The mitral valve prevents oxygen-rich blood from flowing back into the lungs

Monostrut valve:         Mechanical heart valve manufactured by Shiley, Inc., marketed outside of the US from 1982. The Monostrut did not contain any welded parts, instead the flange and the inlet and outlet struts were machined as one component

Occluder:                    The vane-like disc which tilts back and forth within the prosthetic heart valve. The main opening-closing component of the heart valve

Outlet strut:                 A small strut or hook which prevents the disc from escaping during opening and closing of the disc (occluder, see above). The outlet strut was welded to the flange of the Bjork-Shiley Convexo-Concave valve

PMA:                          Pre-Market Approval, a regulatory process by which a medical device is approved as safe and effective by the FDA prior to full-scale marketing. The PMA process is mandatory for Class III medical devices

Process validation:      Establishing documented evidence which provides a high degree of assurance that a specific process will consistently produce a product meeting its pre-determined specifications and quality characteristics; a requirement of the Good Manufacturing Practice (1978) regulation for medical devices (21 CFR Part 820)

Radiopaque-Spherical:  A Bjork-Shiley prosthetic heart valve. Predecessor to the Convexo-Concave valve; consists of a flange with two welded struts (inflow and outflow) and a flat occluder disc made from pyrolytic carbon with an embedded Tantalum ribbon, which gave the valve its radiopaque property

Stellite:                        A cobalt-based alloy used for the Bjork-Shiley flange

Suture ring:                 A fabric ring attached to a heart valve which is sutured (sewn) into the patient’s heart

Systole:                       The rhythmic contraction of the heart, especially of the ventricles, by which blood is driven through the aorta and pulmonary artery after each dilation or diastole

Thromboembolism:     General term used to refer to thrombus and embolism complications

Thrombosis:                 The formation of a blood clot

Thrombus:                   A blood clot formed on a mechanical heart valve which grows to such a size that it inhibits operation of the valve

CHAPTER 2

 

LITERATURE REVIEW

                                                                       

In order to gain an understanding of this subject, numerous sources were reviewed, such as journal and newspaper articles, books, newsletters, and government documents. However, the sources below have proven to be of special interest and value.      

In 2005, Dr. Eugene H. Blackstone, MD, wrote in the journal Circulation: “Could It Happen Again? The Bjork-Shiley Convexo-Concave Heart Valve Story.” In this article, Dr. Blackstone discusses his involvement with the Convexo-Concave heart valve. He gives a cursory background and description of the Convexo-Concave valve, the involvement of the FDA and the role of self-regulation. He also touches on the role of professional journals, how to make life and death decisions, and finally discusses the central question, “Could It Happen Again?” Dr. Blackstone’s assertion is that:

It is happening again! All of the elements of the Bjork-Shiley story remain in place, although the recent controversy over withdrawing drugs from the market presents fewer tough decisions than replacing a functioning prosthetic valve. (p. 2718)

However, he makes no reference as to what “elements” are in place, nor does he offer any examples of devices that have failed as catastrophically as the Bjork-Shiley Convexo-Concave valve. Therefore, it appears that his article only offers a seemingly vague, though expert opinion, rather than fact. However, this article has served as the catalyst for this thesis and ultimately the thesis will either prove Dr. Blackstone wrong, or concede that he is right.

In 1989, John Dingell, US Representative (D-Michigan), started a round of internal investigations of the FDA, in order to uncover what he believed were serious shortcomings at the FDA. In so doing, he used the Bjork-Shiley Convexo-Concave heart valve as his “stick,” with an embattled FDA receiving the punishment. The hearings culminated with a report titled “The Bjork-Shiley Heart Valve: ‘Earn as You Learn.’ Shiley Inc.’s Breach of the Honor System and FDA’s Failure in Medical Device Regulation.” The name refers to what was deemed an unethical practice of trying to correct the outflow strut fractures by changing manufacturing and inspection practices, while continuing to market the device to an unsuspecting and trusting public. Although the document is out of print at the U.S. Government Printing Office, copies are still available at some university libraries, such as University of Southern California, in Los Angeles. “The Bjork-Shiley Heart Valve: ‘Earn as You Learn’” appears to have been written in haste, mistakes are evident, and it looks as though it was typeset using an antiquated typewriter. The document oscillates from the Shiley story to the internal failures at the FDA and all of the FDA’s shortcomings are illuminated. Nevertheless, the story of Shiley from about 1978 to 1986 is clearly illustrated, accompanied with copies of pertinent Shiley internal memos and documents. The infamous telex from Shiley to Dr. Bjork is repeated in all of its lurid detail:

 

 

REF:  YOUR TELEX NO 954 DATED DEC 17, 1980

WE WOULD PREFER THAT YOU DID NOT PUBLISH THE DATA RELATIVE TO STRUT FRACTURES. WE EXPECT A FEW MORE AND UNTIL THE PROBLEM HAS BEEN CORRECTED, WE DO NOT FEEL COMFORTABLE. WE WOULD LIKE TO DISCUSS THE STRATEGY WITH YOU DURING YOUR JANUARY [19]81 VISIT.

I LOOK FORWARD TO OUR JANUARY MEETING.

MERRY CHRISTMAS

PAUL MORRIS

SHILEYINC. (p. 108)      

As such, the “The Bjork-Shiley Heart Valve: ‘Earn as You Learn’” report has proven to be an invaluable resource, especially in telling the Shiley story. Many references to “Earn as You Learn” can be found in Appendix D “Shiley Timeline.”

In order to investigate the history of the Food and Drug Administration, two books proved to be interesting resources. Philip J Hilts’ Protecting America’s Health: The FDA, Business and One Hundred Years of Regulation (2003) and Herbert Burkholz’s The FDA Follies: An Alarming Look at Our Food and Drugs in the 1980s (1994) both look at the FDA in depth, but with two different viewpoints. Hilts traces the history of the FDA from its very early days, back to 1885 and the days of Dr. Wiley, on through the present day. The author provides ample detail into the various crises and scandals that have rocked the FDA and how most often regulation only follows serious catastrophes, such as the Elixir of Sulfanilamide and Thalidomide. However, the book is singularly focused on drugs and the drug industry; medical devices are scarcely mentioned.

Burkholz’s book focuses on the tumultuous 1980s, when deregulation was purported as the answer to Carter’s economic stagnation and inflation. Burkholz makes no excuses for his dislike of the Reagan years and President Reagan’s policies. Nevertheless, in the chapter “Toys R Us,” Burkholz focuses on the history of medical devices, starting with the Dalcon Shield and moves on through the Bjork-Shiley Convexo-Concave heart valve and silicone breast implants. As such, the book gives a vivid backdrop to the Bjork-Shiley saga, which played out during this time. Both books also serve as references for each other; even though Burkholz is more of a populist and Hilts is more scholarly, both are able to hold the attention of the lay person.  

Dr. William S. Stoney’s Pioneers of Cardiac Surgery (2008) is a well written and informative introduction to the history of cardiac surgery in the 20^th century. The book is primarily a compilation of interviews with some of the most famous cardiac surgeons, conducted by Dr. Stoney, including Dr. Starr and Dr. Bjork, both who are featured in this thesis. In addition, the book contains two important sections, an introduction to “Significant Events in Cardiac Surgery” and chapter one “A Short History of Cardiac Surgery.” Furthermore, the chapter “Valvular Heart Surgery” contains interviews with valvular surgery pioneers such as Dr. Alain F. Carpentier, in addition to Dr. Bjork and Dr. Starr. Therefore, this book is an important contribution to the history and craft of cardiac surgery.

To compliment Pioneers of Cardiac Surgery, the Mayo Clinic Heart Book, the Ultimate Guide to Heart Health, edited by Bernard J. Gersh, M.D. (2000) provides an easy to read guide to the cardiovascular system. With its numerous color and black-and-white illustrations, coupled with in-depth but easy to grasp information, the Mayo Clinic Heart Book fills the gaps left by the more academic Pioneers of Cardiac Surgery. As such, these two books have served as very informative resources, especially concerning heart valve disease.

Lastly, Gerald Derloshon’s book One for the Heart. The Story of the Professor Viking O. Bjork, M.D (1983) has provided a reasonable but biased biography of Dr. Viking Bjork. This book is unique in that it was never intended for the public; instead, it was written and published by Shiley Inc. in order to commemorate Dr. Bjork’s retirement in 1983. Gerald Derloshon was a Shiley employee at the time, engaged in writing such things as How We Get There From Here, Reflections on Shiley’s Sixteenth Birthday, in 1982. One for the Heart is lightly written, with no conflicts or crises. It serves as a travel log for Dr. Bjork, discussing his extensive travel, especially during the 1970s. However, given the scant information that is otherwise available concerning Dr. Bjork, One for the Heart gives an insight into Dr. Bjork’s personal life, even though the writing is somewhat superficial. Also, by combining the interview with Dr. Bjork that Dr. Stoney conducted for Pioneers of Cardiac Surgery with the information given in One for the Heart gives the reader an introduction to Dr. Viking Bjork, one of the most influential cardiac surgeons of the 20^th century. Dr. Bjork, after all, was a brilliant surgeon whose actions helped save and prolong the lives thousands of patients. At the same time, his inaction in not alerting his fellow cardiac surgeons regarding the outflow strut fractures that plagued the Convexo-Concave valve may have cut short the lives of hundreds of other patients.           

CHAPTER 3

 

METHODOLOGY

 

This thesis will review the available design information regarding the Bjork-Shiley Convexo-Concave heart valve, such as journal articles, the original patent filing, the 510(k) filing from 1978, and the “Statement of Safety and Efficacy,” which was prepared by the FDA after the Pre-Market Approval application was submitted by Shiley in November of 1978. Also, it will review the history of design controls, starting with the 1976 Good Manufacturing Practices, the 1984 FDA guidance on process validation, the 1990 FDA report “Device Recalls: A Study of Quality Problems,” the 1996 Quality System Regulation, the FDA “Preamble to the Final Rule” (regarding the implementation of the 1996 Quality System Regulation), and the 1997 FDA “Design Control Guidance for Medical Device Manufacturers.” Furthermore, it will review the requirements for prosthetic heart valves outlined in the FDA document “Replacement Heart Valve Guidance” which was released as a draft on October 14, 1994. As has been explicitly promulgated by the FDA, device design encompasses both the medical device itself and also the manufacturing methods that must be utilized to manufacture the device.

This review will attempt to answer the question can design controls ensure that another Bjork-Shiley Convexo-Concave heart valve does not enter the market? The method for the review will be a gap analysis, wherein the requirements of 21 CFR 820.30 “Design Controls” will be outlined along with an interpretation of the requirements. Further, a review of Shiley’s 510(k) filing from 1978, and the Statement of Safety and Efficacy from 1978 will be conducted, in order to evaluate if the design control requirements were met. Lastly, a “what if” evaluation of the design control requirements will be performed, in order to answer the question if design controls had been followed by Shiley, would the Bjork-Shiley Convexo-Concave valve still have reached the market? As such, this thesis will evaluate the effectiveness of the design control requirement as stated in 21 CFR 820.30.         

 

 

 

 

 

 

 

 

 

 

CHAPTER 4

 

VALVULAR DISEASE

 

 

Figure 5. The human heart (n. d.). Source:  www.nhf.org.nz.  

 

 

The heart is without comparison when discussing the human organs. It relies on the heart muscle or myocardium, heart valves, a network of vasculature, and nerves to function as expected and to provide the body (and paradoxically itself) with a continuous and consistent blood flow. On the average, the heart beats 60-70 beats per minute, which would amount to some 2.5 billion-plus heart beats in the lifetime of an octogenarian. This durability is not often matched by machines or other human inventions, yet it forms naturally in almost all animals. Unfortunately, it is also among the organs that have no redundancy (as opposed to the eyes, ears and kidneys); therefore, if the heart does not function properly, no backup system exists. Consequently, if the heart fails, unless intervention is close at hand, death is most certain.

The heart is a powerful muscle called the myocardium, which is surrounded by the pericardial sac. The heart is composed of four chambers; the left and the right atrium and the left and the right ventricle. As such, the heart is separated into two pumps; the right heart which receives oxygen-poor blood from the extremities and the head through the superior vena cava and inferior vena cava; the right heart pumps this oxygen-poor blood into the pulmonary artery and into the lungs to be oxygenated. As the oxygen-rich blood leaves the lungs through the pulmonary veins, it enters into the left heart, which pumps this oxygenated blood back into the body through the aorta. Also, the internal pressure in the left heart is approximately six times as great compared to the right heart (Gersh, 2000; Guyton & Hall, 2005).

The cyclic function of the heart is called the cardiac cycle. It starts with a period of relaxation (diastole), when the left and right ventricles are being filled with blood. Subsequently, during the contraction period (systole), the ventricles contract and push blood out of the ventricles into the aorta and the pulmonary artery. As seen from Figure 5, oxygen-poor blood from the body enters the right atrium from the superior and inferior vena cava, passes through the tricuspid valve into the right atrium (diastole), gets squeezed by the right ventricle (systole) through the pulmonary valve, and into the lungs for oxygen-exchange. As the oxygen-rich blood leaves the lungs, it passes into the left atrium (diastole) through the mitral valve and into the left ventricle. As the left ventricle contracts (systole), it pushes the oxygen-rich blood through the aortic valve, through the aorta into the body. One must remember that the function is not like that of a four-stroke engine; instead, the left and right atria are filled at the same time and then pump simultaneously into the ventricles. Likewise the left and right ventricles contract in parallel (Gersh, 2000; Guyton & Hall, 2005).

The four valves (the tricuspid and the pulmonic in the right heart and the mitral and aortic in the left heart), ensure one-way blood flow through the heart, with no backward leakage. All four valves consist of fibrous tissue flaps called leaflets which are covered by endocardial cells. At the base of each leaflet, the fibrous tissue flap merges with the myocardium to form a flexible hinge, also known as the annulus. The leaflets are attached to each other at the commissures, which form vertical “posts,” to give shape and stability to the valve structure, see Figure 6 (Gersh, 2000; Guyton & Hall, 2005).  

 

 

 

 

 

Commissure posts

 

Figure 6. Superior view of the aortic valve showing the commissure posts. Note the significant calcification of the valve leaflets (n.d.). Source: www.clevelandclinicmeded.com

                                                                          

The Atrioventricular (A-V) Valves

The mitral and the tricuspid valves are also known as the atrioventricular valves, since these two valves separate the left and the right atriums from the left and the right ventricles. Furthermore, both the mitral and tricuspid valves are attached in the ventricles by strong fibrous strings called cordae tendenae, which are anchored to small muscles called the papillary muscles. The cordae tendenae and papillary muscles keep the leaflets stable against any backward blood flow. The A-V valves resemble sail-cloth, tethered to the heart by the cordae; their rhythmic movement gives the appearance of a sail being blown back and forth by the alternating pressure of the blood. The A-V valves are open during diastole when the ventricles are being filled with blood and closed during systole, when the heart muscle contracts. The mitral valve is the only valve in the heart with only two leaflets, while the tricuspid valve has three leaflets (Gersh, 2000; Guyton & Hall, 2005). 

The Semilunar Valves

The semilunar valves ensure that the blood being pumped out of the heart does not flow back into the ventricles. Both valves are closed during diastole and open during systole. As compared to the A-V valves, the semilunar valves are composed of “half-moon like” leaflets and each semilunar valve is composed of three leaflets; however, they have neither chordate tendinae nor papillary muscle attachments. Consequently, their fundamental structures are simpler than that of the A-V valves. Their movement during opening and closing appear more orderly as compared to the “flapping” of the A-V valves; the semilunar valves are either fully closed or fully open, in order to assure the one-way movement of blood out of the heart (Gersh, 2000; Guyton & Hall, 2005).

                                                                       

Diseases and Symptoms

There are primarily two dysfunctions that affect the valves; stenosis and regurgitation. When a heart valve cannot open fully, the condition is called stenosis. For example, a stenotic or calcified mitral valve does not pass blood into the left ventricle as expected; instead, the narrow valve causes blood to back up in the left atrium; consequently, pressure builds up in the lungs. This back up of blood can lead to a number of heart problems, such as congestive heart failure, heart chamber enlargement, overgrowth of heart muscle, and various heart rhythm disorders. Conversely, a valve where the leaflets do not meet correctly is called regurgitation, which also causes a leakage back into the heart. Again, if the mitral valve is regurgitant, the leaky valve causes a back flow upon the contraction of the left ventricle and again blood is forced back into the left atrium and back into the lungs. The most common early symptom of valvular disease is shortness of breath or fatigue, especially following exercise (Gersh, 2000; Guyton & Hall, 2005; Westaby, 1997).

The main pathologies that lead up to valvular disease are congenital defects (which are rare), degenerative disease, rheumatic fever (usually caused by a bacterial throat infection), and also bacterial endocarditis, which can be caused by a remote infection, such as a tooth abscess. Also, calcium will deposit on the valve leaflets due to aging. Furthermore, the normal wear and tear we experience with age plays a factor; in fact, most recipients of prosthetic heart valves are people over the age of 65 (Gersh, 2000).

                                                                       

Rheumatic Fever

Rheumatic fever appears to be a reaction against certain strains of the streptococcal bacteria and symptoms are generally evident two to four weeks after an un-treated strep throat infection. The damage to heart valves is believed to be related to antibodies which are produced to fight the throat infection, or possibly toxic substances produced by the bacteria. Most often, rheumatic fever will affect young people aged five to fifteen and unfortunately once stricken, one is more susceptible to a recurrence. During the last 50 years, due to the widespread use of penicillin, rheumatic fever is no longer common in the developed world; however, the disease continues to plague undeveloped third-world countries. The effects of rheumatic fever do not affect the valves equally; the mitral valve is most often affected (85% of cases), secondly the aortic valve (44% of cases), and thirdly the tricuspid valve (16% of cases). However, the pulmonary valve is rarely affected. As can be seen from above, more than one valve is usually affected. Also, the damage to the valves from rheumatic fever may take 10 to 30 years to become symptomatic, at which time surgery is most often the only remedy (Gersh, 2000; Guyton & Hall, 2005; Westaby, 1997).

                                                                           

Endocarditis

Endocarditis is an infection of the endocardium, the membrane that lines the inside of the heart. As bacteria and other microorganisms enter the blood stream, they tend to settle and multiply on abnormal or damaged heart valves (such as valves damaged by rheumatic fever), especially if the blood flow is turbulent. This will cause even further destruction of a weakened heart valve and if left un-treated the condition is most often fatal (Gersh, 2000; Guyton & Hall, 2005; Westaby, 1997).

                                                                         

Diagnosis

The various valvular diseases are most commonly confirmed using echocardiography (ECHO), trans-thoracic echocardiogram (TTE) or the more invasive but more accurate trans-esophageal echocardiogram (TEE). The TTE is an ultrasound apparatus used specifically for cardiology; using a probe placed on the chest, the cardiologist can “see” the overall function of the heart, the valve movements and can also scan for tumors and other growths. The output from the probe is input into the echocardiograph (the receiver and computer) and the image (echocardiogram) is displayed on a monitor and the output can be digitally recorded. Also, by using the TTE’s Doppler effect, the cardiologist can determine the actual cardiac output of the heart, which aids in the diagnosis of both stenosis and regurgitation. Furthermore, the echocardiogram can show the anatomy and the movement of the valves, if they are opening and closing properly, congenital defects, deposits of calcium, and wear and tear. However, since the ultrasonic waves must travel through the skin, fat, bone, and lungs before bouncing back on the heart structure, traditional TTE or ECHO is not the preferred method of diagnosis for the structures on the back of the heart, such as the left atrium (Gersh, 2000).

TEE is an ultrasonic diagnostic tool that relies on a probe that is inserted into the esophagus. Since the heart rests on the esophagus, the distance traveled by the ultrasonic waves is much reduced and the images from TEE are much superior as that of TTE or ECHO. Also, since the mitral valve is placed between the left atrium and left ventricle, TEE is most often used for assessing the function of the mitral valve and also commonly used to assess the function of a prosthetic valve (Gersh, 2000).

                                                                             

Early Valvular Disease Treatments

Early 20^th century attempts at providing surgical relief for valvular disease most often involved blind surgical experiments, since the operations were carried out in a beating heart. These early surgeries consisted of either cutting a leaflet of a stenotic valve, or breaking apart the commissures in order to provide some movement of the leaflets. These primitive and risky surgeries would induce regurgitation, however, at the time (circa 1920-1940), this was considered an improvement over an immovable valve. Also, most cardiologists believed that the damage from rheumatic fever affected the myocardium as opposed to the valves. Consequently, not much research was devoted to valvular disease (Stoney, 2008; Westaby, 1997).

During the period between 1945 and 1955 advancements were made, primarily in the technique of cutting stenotic valves. However, the patient survival rates were quite dismal, often less than 50.0%. Nevertheless, it must be pointed out that the poor survival rates were often attributed to the end-stage disease state of the patient, not necessarily due to the surgery itself. Not until the advent of the heart-lung machines in the mid-1950s and the general availability of prosthetic heart valves in the late-1950s and early 1960s did heart-valve surgery become an effective means of correcting heart valve disease. Dr. Charles Hufnagel’s ball-and-cage valve, which was designed to be implanted in the descending aorta, provided effective relief for aortic regurgitation. Hufnagel implanted his first series in humans starting in September of 1952 and many patients were greatly improved. By 1960, the heart-lung machine had been perfected and made more available, which converged with the advent of the Starr-Edwards ball-and-cage mitral prosthesis. These two complementary technologies ushered in the dramatic advancements in heart valve surgery which have saved and prolonged the lives of millions of patients during the last 50 years; advancements which continue into the 21^st century (Stoney, 2008; Westaby, 1997).             

 

 

CHAPTER 5

THE HISTORY OF SHILEY, INC. AND THE BJORK-SHILEY

CONVEXO-CONCAVE HEART VALVE

 

 

Background

The story of the modern prosthetic heart valve started in 1958 in Oregon, US. Lowell Edwards, a semi-retired engineer in his 60s with a number of patents to his name, contacted Dr. Albert Starr to inquire about the feasibility of designing an artificial heart. Dr. Starr, a young surgeon from the East Coast, had arrived in Portland in 1957, to start a cardiac surgery unit at the University of Oregon Medical School. Dr. Starr kindly informed Edwards that an artificial heart was not feasible; for one thing, there were no reliable artificial heart valves, let alone the technology to develop an artificial heart. However, if Edwards agreed to help Starr to develop an artificial mitral heart valve, they would have a deal (Matthews, 1998; Stoney, 2008).

Lowell Edwards had enjoyed a rewarding career as an engineer and an inventor, holding patents on a hydraulic lumber-debarking system which was widely used in the Pacific Northwest. Edwards also held a patent on a fuel injection system that was used on fast climbing propeller-powered WWII aircraft, which helped the US win the war. These and other patents were providing Edwards royalties and this income supported the Edwards Development Laboratories in Portland, Oregon. Also at this time, Don Shiley was Edward’s chief engineer and Don Shiley would later develop the Starr-Edwards aortic ball-and-cage prosthetic heart valve (Matthews, 1998; Stoney, 2008).

After two years of intense development work and animal studies, Edwards and Dr. Starr had what they believed to be a clinically viable artificial heart valve. This valve was a crude ball-and-cage valve, molded from acrylic and using an acrylic ball for the valve poppet. This design proved successful, even though the materials were subsequently changed. In 1960 Edwards moved south to Santa Ana, California, to start Edwards Laboratories. Here Edwards began commercial manufacturing of the Starr-Edwards ball-and-cage valve. The cage was a highly polished cage made from cast Stellite 21 (62% Cobalt, 28% Chromium, 6% Molybdenum, and 3% Nickel), a non-ferrous metal used in orthopedic implants. The sewing ring was made from a Teflon ring, covered with Teflon cloth. The poppet, or closure mechanism, was a ball made from medical grade silicone rubber, see Figure 7 (Matthews, 1998; Starr et al., 1966; Stoney, 2008).

                                                                    

Figure 7. Starr-Edwards ball-and-cage mitral heart valve (n.d.). Source:  http://img.medscape.com

 

One of Edwards’ employees that followed from Oregon to Santa Ana was Don Shiley. A year or so later, Don Shiley and Dr. Starr collaborated on the Starr-Edwards aortic ball-and-cage valve, with Edwards supervising the development and Don Shiley producing all of the drawings for the valve (Stoney, 2008). There were many early adopters of the Starr-Edwards prosthetic valve, including Dr. Viking Bjork of Sweden. Dr. Bjork had started to use the Starr-Edwards ball valve around 1961 and had achieved satisfactory results, both in the mitral and aortic position. However, for patients with narrow aortic roots, Dr. Bjork found that the tall ball-and-cage valve obstructed the aorta too much, which created a pressure gradient across the artificial valve. Consequently, Dr. Bjork continued to look for an alternative (Bjork & Cullhed, 1967).

                                                                   

Shiley Laboratories and Dr. Bjork

In 1964 Don Shiley left Edwards Laboratories, to set up his own medical device company, Shiley Laboratories. Don Shiley started working with Dr. Jerome Kay at St. Vincent’s Hospital in Los Angeles and by 1966 the Kay-Shiley floating disc prosthetic heart valve was commercially available. The Kay-Shiley valve consisted of a metal ring, a plastic disc held in place by two u-shaped metallic struts and a cloth-covered sewing ring (Derloshon, 1983; Westaby, 1997). In August of 1966, Dr. Bjork started to use the Kay-Shiley floating disc in the aortic position, even though the Kay-Shiley valve was developed to be used in the mitral position. By April of 1968, Dr. Bjork had implanted a total of 60 Kay-Shiley valves with reasonable results (Bjork, Olin, & Astrom, 1969). Don Shiley was astonished that Dr. Bjork had used the Kay-Shiley in the aortic position; however, after having been shown the x-rays and some of the actual patients, Shiley was convinced (Derloshon, 1983).

Nevertheless, even the Kay-Shiley floating disc valve introduced a high pressure gradient and Dr. Bjork informed Don Shiley that he would no longer use the Kay-Shiley valve in the aortic position (Derloshon, 1983). Dr. Bjork had visited Dr. Juro Wada in Japan, where he was introduced to the Wada-Cutter pivoting disc valve. The Wada-Cutter valve was manufactured by Cutter Laboratories in Berkeley, California. The Wada-Cutter valve had a titanium ring for the housing, a Teflon cloth sewing ring, and a hard Teflon disc. Dr. Bjork had started to use the Wada-Cutter valve and initially had very good results, especially regarding the low pressure gradients across the valve. Even so, the Wada-Cutter valve had problems; fibrin formation on the hinge would cause the disc to become immobilized and Dr. Bjork had experienced a hinge fracture, which made him leery of the Wada-Cutter valve (Derloshon, 1983; Westaby, 1997).

After meeting in Stockholm in May of 1968, Dr. Bjork had convinced Don Shiley that a hinge-less tilting disc valve was needed. By December of 1968, Don Shiley had produced a workable prototype of a tilting disc valve (Derloshon, 1983). The initial Bjork-Shiley tilting disc valve had a Stellite 21 ring with a Teflon sewing ring and two opposing struts made from Haynes 25, a similar material to the cobalt-based Stellite. The struts were welded on to the Stellite ring; one on the inflow side of the valve and one on the outflow, to restrain the disc. The disc was originally made from Delrin, a hard plastic, with a circular recess in the outflow side. The outflow strut was shaped like a hook, which fit into the recess in the outflow side of the disc, effectively capturing the disc (see Figures 8 and 9). The hook would also limit the disc opening translation and rotation about the inlet strut (Bjork, 1970a; Bjork, 1977).  

Hook-shaped outflow strut

 

Figure 8. Drawing of the Bjork-Shiley tilting disc heart valve, from the outflow side (n.d). Note the outflow strut (hook) and recessed disc. Source:  www.wikimedia.org.

 

 

 

 

Bjork-Shiley tilting disc valve had a Delrin occluder disc   Bjork-Shiley tilting disc valve had welded inflow struts. Note the white arrows pointing to the outflow struts

           

Figure 9. Bjork-Shiley Convexo-Concave heart valve. Note that the Bjork-Shiley tilting disc valve is of a similar design, the major difference being welded inflow struts and a Delrin occluder disc. The Delrin disc was later replaced by a pyrolytic carbon disc. Source: Article by van Gorp et al. in Radiology (2004).

 

By January of 1969, Dr. Bjork started to implant the Bjork-Shiley tilting disc valve, which was later to be known as the Bjork-Shiley Radiopaque-Spherical prosthetic heart valve. By 1970, Dr. Bjork had implanted 99 Bjork-Shiley valves; 77 in the aortic position and 22 in the mitral position (Bjork, 1970a; Bjork, 1970b; Bjork, 1971). In 1971, the Delrin disc was replaced by a conical pyrolytic carbon disc (Bjork, 1972) and by 1975, the conical disc was replaced by a spherical pyrolytic disc with a radiopaque marker, hence the name Bjork-Shiley Radiopaque-Spherical (Bjork, Henze, & Hindmarsh, 1977). The Bjork-Shiley Radiopaque-Spherical prosthetic heart valve proved very successful, selling some 90,000 valves by 1975 and close to 255,000 by 1983 (Bjork, 1983).

            However, the Radiopaque-Spherical was not without problems. In 1975, Shiley recalled three lots of 29 and 31 mm mitral valves, after reports of inflow strut fractures (Elliott, 1976). Also, due to the design of the Radiopaque-Spherical, the occluder disc created a pseudo-hinge with the valve ring, which could lead to thrombus formation. Consequently, design work was started on an improved valve, which would ultimately become the Bjork-Shiley Convexo-Concave valve (Bjork, 1983; Bjork, 1984).

The Convexo-Concave valve was an improvement over the Radiopaque-Spherical valve in three aspects; the inflow strut was no longer welded to the valve ring; instead the inflow strut was machined as an integral part of the flange. Furthermore, the pyrolytic carbon occluder disc had been redesigned as a convex-concave disc, resembling that of a saucer. According to Shiley, the new occluder design would aid in the washing of the disc by the pumping blood, thereby reducing the risk of thrombus. Lastly, the outflow strut design had been changed, such that the occluder disc would now move down and away from the valve ring during opening and eliminating the pseudo-hinge, again in an effort to reduce thrombus formation (Bjork, 1983; Bjork, 1984).

As such, the patent application for the Convexo-Concave was filed in 1975 and clinical trials started in June 1976. Dr. Bjork implanted 297 valves at the Karolinska Institute in Stockholm. Also, Drs. Ehrenhaft, LeMole, and Fernandez implanted another 89 valves in the US (U.S. Department of Health and Human Services, Shiley Statement of Safety and Efficacy, FDA PMA780008, 1978).   

Originally, Shiley tried to file a 510(k) application for the 60-degree Convexo-Concave valve, claiming substantial equivalence to the Radiopaque-Spherical valve (the 510(k) approval process is less stringent than the Pre-Market Approval route). However, the FDA determined that the Convexo-Concave valve should go through the new PMA process (U.S. Department of Health and Human Services, FDA 510(k) file no. K780772, 1978). In November of 1978, Shiley filed a PMA for the 60-degree Convexo-Concave valve with the FDA. Incidentally, this was the first ever heart valve to go through the PMA process (U.S. Department of Health and Human Services, FDA PMA no. 780008, 1978).

                                                                                

Outflow Strut Fractures

            In September of 1978, there had been an outflow strut fracture during the Convexo-Concave clinical trials. However, Shiley claimed that the strut failure was an “anomaly” and that nothing in the welding had changed when comparing the Radiopaque-Spherical with the Convexo-Concave valve, which would have seemed plausible at the time, given the excellent durability of the Radiopaque-Spherical valve (The Bjork-Shiley Heart Valve, 1990). Also, following the argument above, it would appear that most of the force would rest on the inflow strut during opening, with the occluder acting as a lever on the inflow strut. Therefore, it would seem that the outflow strut failure was a chance event. Consequently, the FDA approved the new Convexo-Concave valve in April 1979, around the same time that Don Shiley sold Shiley, Inc. to the pharmaceutical giant Pfizer (The Bjork-Shiley Heart Valve, 1990).

            After the strut fractures started to manifest themselves during the 1980s, there were many attempts to explain the failures. (a) The welding was thought to be faulty (Bjork, 1985; Sacks, Harrison, Bischler, Martin, Watkins et al., 1986); however, this does not appear to be plausible, again given the excellent durability of the Radiopaque-Spherical valve, from where the welding technology was adapted. (b) The size and position of the implanted valve was implicated, since the failures occurred most often in the mitral position and in the large size valves (29 – 31 mm; Lindblom, Bjork, & Semb, 1986). (c) The size and age of the prosthetic valve recipient was suspected; it was found that most often the failures would occur in a younger patient, combined with a large body surface area (i.e. a large patient; Lindblom et al., 1986). (d) The opening angle of the valve was implicated; in late 1980 a 70-degree opening valve was introduced, at least in part to alleviate the strut failures with the 60-degree valves (Bjork, 1981). Unfortunately, the 70-degree valve proved to be even worse than the 60-degree valve; the 70-degree would have almost 10 times the failure rate as compared with the 60-degree valve (The Bjork-Shiley Heart Valve, 1990). (e) In 1985, Dr. Bjork even made a feeble attempt at blaming the surgeon, writing that:

A: Obstruction of the disc movement in the aortic area has been due to faulty implantation technique, i.e. a suture cut too long and caught between the disc and the valve ring. Disc obstruction in the mitral area has also been attributable to technique – leaving chordate tendineae too long. B: Disc escape by strut dislocation or strut fracture is a risk which can be reduced by care during valve implantation, always using the holder to orient the valve. Forceful use of forceps, or use of the disc itself as a handle, may produce a crack in the weld which can propagate a strut fracture in the course of a year. (p. 6)    

However, again pointing to the success of the Radiopaque-Spherical valve and surmising that the same thoracic surgeons were using the same implantation technique, this argument was hollow.

            As early as 1981, Shiley was made aware that a probable root cause for the Convexo-Concave outflow strut failures was bimodal closure, wherein the occluder would pivot on the inflow strut during closing and put extreme pressure on the outflow strut (The Bjork-Shiley Heart Valve, 1990). Nevertheless, it was not until 1999 that this failure mode would be rigorously tested and proved (Wieting et al., 1999). The testing was initiated by The Shiley Heart Valve Research Center, an organization that was set up by Pfizer in 1992 to investigate the failures. The bimodal closure failure would then finally point to a design problem, rather than a manufacturing problem (such as welding), or any of the other issues discussed above.

            The Convexo-Concave valve continued to be a problem for Shiley and as a counter measure to the outflow strut failures, the 70-degree Convexo-Concave valve was introduced in 1980 (Bjork, 1981). However, the 70-degree valve was only sold outside the US, since the FDA was reluctant to approve the valve for sale in the US without any clinical data. Nevertheless, the FDA did approve the exportation of the 70-degree valve. Consequently, some 4,000 of the 70-degree valves were implanted; primarily in Europe, Canada, and Australia. Unfortunately, the 70-degree valves proved to have a substantially higher fracture rate as compared to the 60-degree valve; the large size mitral valves would fracture 5.0 – 10.0% of the time (The Bjork-Shiley Heart Valve, 1990).

            Due to reports in 1982 from Australia and Sweden regarding strut fractures in the 70-degree valve, the FDA withdrew their permission for Shiley to export the 70-degree valve in early 1983. Nevertheless, instead of recalling the 70-degree valves from their European distributors, Shiley urged them to “contact the Brussels office in order to make use of the [70-degree] valves before ordering the 60-degree Convexo-Concave or [Radiopaque] Spherical design” (The Bjork-Shiley Heart Valve, 1990, p. 33).

                                                                            

Product Recalls

            Meanwhile, the 60-degree valve continued to suffer outflow strut failures. Consequently, the valve was recalled several times during the 1980s, as outlined in the government report “The Bjork-Shiley Heart Valve, ‘Earn as You Learn’”:

1.      On February 1, 1980, Shiley announced a product recall: All 60-degree valves that had been welded from June-1976 to August 1979 were recalled.

 

2.      On May 18, 1982, Shiley again announced a product recall: This time, selected 29, 31, and 33mm 60-degree valves that had been shipped between May 1980 and February, 1982 were recalled.

 

3.      On July 6, 1983, again a product recall: These were valves from additional selected lots of 29, 31, and 33mm 60-degree valves that had been welded between February 1, 1980 and March, 1982.

 

4.      On October 14, 1985, yet another recall and product withdrawal: Size 29, 31, and 33mm 60-degree valves that had been made between March 1, 1981 and June 30, 1982 were recalled. Those valves that were made after June 30, 1982 were withdrawn from the market and production of the 29, 31, and 33mm valve sizes ceased.

 

5.      On November 23, 1986, the final product withdrawal: Lastly, all 19, 21, 23, 25, and 27mm size valves were withdrawn from the market and production of these valves ceased.

 

            During this time, the relationship between Shiley and the FDA became increasingly strained. Consequently, the Los Angeles District Office of the FDA performed numerous inspections of Shiley’s Irvine and Puerto Rico facilities, as outlined in “The Bjork-Shiley Heart Valve: ‘Earn as You Learn’ (1990)”:

1.      March 1979: The FDA found “Discrepancies between the PMA submitted ‘Manufacturing Process’ and actual operations were noted.” (p. 25)

 

2.      February 1982: The FDA found that “The exact cause of the fracture is purportedly unknown but they seem to be occurring just above the weld that fuses the strut to the flange…Management claims the frequency of fractures is actually declining…A total of ten fractures have been reported…No corrective action was indicated.” (p. 25)

 

3.      March 1983: the FDA found that “GMP deficiencies include failure to investigate or record complaint for failure of a 70-degree valve, no failure evaluations of 60-degree and 70-degree convexo-concave valves, indications that two scrapped units were included with acceptable units…” (p. 25)

 

4.      February 1984: The FDA found that “some manufacturing changes made the valves less safe, subjecting the patient to increased risk… It is recommended that the approval of the firm’s premarket approval for the convexo-concave valves be withdrawn and that all un-implanted valves be recalled until such time as the firm demonstrated efficacy of the device and safety of the device.” (p. 26)

 

            Unfortunately, FDA headquarters did not follow the Los Angeles District Office’s suggestion. Instead, FDA’s Centers for Devices and Radiological Health (CDRH) was concerned about the possible negative patient impact if the Convexo-Concave was withdrawn from the market, since the valve had considerable market share (The Bjork-Shiley Heart Valve, 1990).

            In 1984, George Sherry, a former tool design engineer at Shiley, contacted the Public Citizen Health Center consumer advocate group in Washington. Dr. Sidney Wolfe of Public Citizen petitioned the FDA to have the Convexo-Concave valve withdrawn from the market; however, the FDA took no action. Finally, in November of 1986, after numerous law suits and bad publicity, Shiley withdrew the Convexo-Concave heart valve (The Bjork-Shiley Heart Valve, 1990).

 

Monostrut heart valve. Note the integrated inflow struts and integrated outflow strut

 

Figure 10. Bjork-Shiley Monostrut heart valve (n.d.). Note that figure has been enhanced to show the single-strut feature and the occluder. Source:  www.absoluteastronomy.com.              

            As a result of the outflow strut fractures in the Convexo-Concave valve, Shiley had designed a new heart valve around 1980, the Monostrut valve, see Figure 10. The Monostrut had no welded components, instead both the inflow and the outflow struts were machined into the valve ring as integral components (Bjork, 1983; Bjork, Lindblom, & Henze, 1985). The Monostrut was marketed outside of the US with excellent results (Lindblom D, Lindblom U, Henze, Bjork, & Semb, 1987; Maddern et al., 1990); by 1990 more than 100,000 Monostrut valves had been distributed, with no known valve failures. It is truly unfortunate that Monostrut never reached the US market, due to the fact that the FDA had become leery of Shiley for reasons outlined above. Consequently, a heart valve with an excellent safety record was kept off the US market, due to the shortsighted actions (or inactions) of Shiley.

                                                                                 

The Aftermath

            Between 1986 and 1989 things were relatively quiet for Shiley and during this time they were focusing on other products, such as oxygenators. However, during 1989, Representative John Dingell, who chaired the powerful Subcommittee on Oversight and Investigations of the Committee on Energy and Commerce, started a series of congressional hearings aimed at investigating the actions and inactions of the FDA during the 1980s. In order to drive home his point, Dingell used the Convexo-Concave debacle as one of the most serious examples of medical device failures and Shiley was now again in the public eye. The investigations culminated with the publication of the report “The Bjork-Shiley Heart Valve: ‘Earn As You Learn’ Shiley Inc.’s Breach of the Honor System and FDA’s Failure in Medical Device Regulation.” This report outlined the actions taken by Shiley and the inaction of the FDA.

            In 1990, Pfizer conceded that during the latter part of the 1980s, they had settled some 200 lawsuits related to the Convexo-Concave heart valve. Most of these suits had been settled quietly out of court (Pratt, 1990). Nevertheless, the public hearings and the Dingell report sparked a new interest in the Convexo-Concave; consequently, more bad press and more lawsuits followed. By now, Shiley had become a liability for Pfizer and by 1992 Shiley was sold to Sorin Biomedica, a subsidiary of Italian Fiat (“Mergers, Acquisitions,” 1992). Also, the lawsuits culminated with the Bowling-Pfizer settlement. As a consequence of this class-action suit, Pfizer was forced to put aside hundreds of millions of dollars for patients and research (Tidmarsh, 1998).    

                                                                               

 

 

Figure 11. Dr. Viking Bjork. Source: Article by Robicsek, in the Journal of Thoracic and Cardiovascular Surgery (2009).      

 

Dr. Bjork Retires

At the age of 65, Dr. Bjork (see Figure 11) had retired in 1983 from the Karolinska Institute in Stockholm. He settled in Rancho Mirage, in Southern California, to do research at the Eisenhower Institute (Meier, 1990). After having an extraordinary career as a world-renowned thoracic surgeon, lecturer, teacher, and administrator, Dr. Bjork was maligned by the Swedish press due to the publicized Convexo-Concave strut fractures. In 1985, he was also severely censured by the Swedish National Board of Health and Welfare, for having failed his duty to warn other Swedish thoracic clinics about the failures of the Convexo-Concave heart valve (The Bjork-Shiley Heart Valve, 1990). His conflict of interest was quite obvious; during the 1970s he received a 6.5% royalty on each valve sold, which would have earned him about $13 million by 1979. Also, after Pfizer took over Shiley in 1979, Dr. Bjork received over a million dollars in consulting fees (Zoroya & Hirsh, 1993). Furthermore, from 1967 to 1983, Dr. Bjork was chief editor of the Scandinavian Journal of Thoracic and Cardiovascular Surgery. Consequently, he published numerous articles regarding the Bjork-Shiley valves, exhorting the virtues of the heart valve, while never reporting any failures.

            However, Dr. Bjork’s quest for the perfect mechanical heart valve did not stop; during the latter part of the 1980s he experimented extensively on goats with a modified Monostrut valve, which, due to its special coating, did not require anticoagulants (Bjork, Wilson, Sternlieb, & Kaminsky, 1988). This study culminated with a small human clinical trial in Brazil, which was reported during the 1990s. In this trial, several young women had successful pregnancies after having received the modified Monostrut. Since they were not required to take anticoagulants, they could carry their pregnancies to term without the risk of birth defects or excessive bleeding during child birth (Bjork, Riberio, & Canetti, 1999; Bjork, Riberio, Canetti, & Bomfim, 1994). Dr. Bjork, 90, died in Danderyd, Sweden in February of 2009. Note:  For further information regarding Shiley, see appendix D, “Shiley Timeline.”

  CHAPTER 6

 

THE HISTORY OF THE FOOD AND DRUG

ADMINISTRATION, 1906-19974

 

 

Background

In 2006, the Food and Drug Administration (FDA) celebrated its Centennial. The agency had chosen the passage of the 1906 Federal Food and Drugs Act as its marker from when the agency itself believes it came into being. However, the beginnings of the FDA are usually traced back to 1883, when Dr. Harvey W. Wiley became chief chemist at the U.S. Department of Agriculture’s Division of Chemistry. Dr. Wiley had an interest in the adulteration of foods, which at the end of the 19^th century had reached fantastic proportions. As one of his first duties, Dr. Wiley had started to determine the sugar content in syrups and honey and found that most of them were cut with cheap, laboratory-produced sugar. He expanded the division’s research into this area and ran a study from 1887 to 1902, which was published in 10 parts, the “Food and Food Adulterants.” The compendium listed an array of food additives used at the time, such as copper sulfate (which would make wilted vegetables appear fresh and also added to flour to hold moisture), sodium benzoate (which stops the rotting of tomatoes), borax (used as a preservative), and also other fillers such as chalk, clay, and plaster of Paris. This compendium was presented to Congress and in 1902, with the help of Teddy Roosevelt, Dr. Wiley had appropriated $5,000 to study the health effects of preservatives. These studies were carried out on human volunteer subjects, who were given increasing doses of preservatives such as borax, salicylic acid, benzoic acid, sodium benzoate, and formaldehyde. The experiments, although flawed (there was no control group), changed Dr. Wiley’s earlier belief that any substance could be added to a food, as long as it was labeled accordingly (FDA Backgrounder, 1999; Hilts, 2003; Swann, 1998).

In late 1905 and early 1906, muckraker Samuel Adams wrote a series of articles in Collier’s, exposing patent medicines. The medications of the turn of the last century were for the most part concoctions of alcohol and opiates, if they contained any substance at all. However, the “patenting” of medications only involved the labeling and the advertising, which was misleading at best and outright fraudulent at worst. Concoctions such as Peruna (28 percent alcohol and 72 percent water) and Liquizone (99 percent water and 1 percent sulfuric acid) were sold to the masses, promising to cure colds, tuberculosis, inflamed appendix, dysentery, and dandruff. However, it was Upton Sinclair’s The Jungle, wherein the deplorable conditions of the meatpacking industry were exposed, that had the most profound effect. Shorty after the publication of the book, meat sales fell by half and the nation was ready for a federal law concerning food and drugs. From spring to early summer 1906, the various regulations were written and were formulated as the Food and Drug Act of 1906. Finally, the bill was signed into law by President Theodore Roosevelt on June 30^th. Unfortunately, the law was written in such a fashion that the onerous responsibility was on the Bureau of Chemistry to first identify problems, then trying to persuade the food and drug industries to change their ways. If the industries were recalcitrant, the Bureau’s inspectors were forced to take the culprits to court, which often resulted in protracted and unsuccessful court battles (FDA Backgrounder, 1999; Hilts, 2003; Swann, 1998).  

By 1927, the Bureau of Chemistry was reorganized into two separate entities; the regulatory functions were now located in the Food, Drug, and Insecticide Administration and non-regulatory research located in the Bureau of Chemistry and Soils. By 1930, the name of the Food, Drug and Insecticide Administration was shortened to the Food and Drug Administration (FDA). By 1933, the FDA’s Walter G. Campbell (named head of the Bureau of Chemistry in 1921), had realized that the 1906 law had become ineffective. Consumer items such as cosmetics, which had a very limited market by the turn of the century, were by the 1930s big business; some, such as Lash Lure, were so toxic they caused blindness in hundreds of women. Pesticides were another product that the 1906 law did not cover; consequently, there was no oversight regarding pesticide residuals. Even more alarming, the 1906 Food and Drug Act had failed to curtail the usage of toxic substances in drugs, such as narcotics and even a few poisons. Therefore, anybody could brew a “cure,” with no testing required and put the “medication” on the market. As long as the maker made a claim that his intention was to cure a disease, the concoction could be sold as a medicine, regardless of deadly side-effects. At most, the FDA could begin a court action, which, if successful, would result in nothing but a small fine for the drug maker (FDA Backgrounder, 1999; Hilts, 2003; Swann, 1998).

                                                                           

Elixir of Sulfanilamide

The new proposed law was hotly contested by the patent medicine industry, chiefly by the Proprietary Association and the Institute for Medicine Manufacturers. The FDA fought back, using the “Chamber of Horrors” (a graphic exhibit of various harmful drugs and concoctions used by Campbell to emphasize his message). However, at the end of summer, 1937, the Massengill Company of Bristol, Tennessee, was marketing Sulfanilamide, a highly successful and potent antibiotic. The drug was originally sold in a tablet form; however, field salesmen reported back to Massengill that patients would be happier if the drug would be available in a liquid, more palatable form. Harold C. Watkins, chief chemist at Massengill, set out to put sulfanilamide into solution, trying a variety of solvents. He settled on diethylene glycol, a mildly sweet fluid. Initially, the drug was made in a few batches, totaling 240 gallons. The liquid antibiotic was called Elixir of Sulfanilamide and shipped to druggist in the fall of 1937. Almost immediately, reports started coming in to the FDA about deaths, most of them children. No testing of the Elixir had been done and none was in fact required. The FDA field agents acted swiftly and were able to track down most of the Elixir. However, by November of 1937, over 100 people had died from being poisoned by the Elixir. It can also be argued that the Elixir killed Harold Watkins, who committed suicide (FDA Backgrounder, 1999; Hilts, 2003; Swann, 1998).     

The Elixir of Sulfanilamide gave the impetus to pass the new law, which was propelled through Congress and signed into law by President Roosevelt on the 25th of June, 1938. The law became known as the Food, Drug, and Cosmetic Act. The new law required that drugs be labeled with adequate directions for use and that no false therapeutic claims could be made. More importantly, the law required that new drugs be approved for sale by the FDA, before being put on the market. The framework for this approval was the Pre-Market Approval mechanism, which would include safety testing. This testing would be conducted using the scientific approach, not anecdotal or even authoritative opinion. The law also touched on foods, such as packaging and quality of foodstuffs. In addition, the law formally authorized factory inspections and added injunctions as an enforcement tool. However, one weakness of the bill was lax clinical trials, since the law gave a company the ability to give drugs to patients for experimental purposes, without the patients even knowing what was given. Also, the drug companies and the doctors were not required to keep records of these trials (FDA Backgrounder, 1999; Hilts, 2003; Swann, 1998).

During the 1950s, following hearings under Representative James Delaney, a series of laws were added; pesticide residues (1954), food additives (1958), and color additives (1960). These new laws gave the FDA much tighter control over chemicals that were entering the food supply; now the manufacturer had to prove the safety of additives before they were marketed.

In late 1959, Estes Kefauver, Senator from Tennessee, started hearings on the excessive profits enjoyed by drug companies. The Kefauver hearings continued through 1960 and uncovered price fixing, substantially higher prices for drugs in the US and an industry where massive sales promotions and huge markups were the norm. Senator Kefauver wanted legislation to deal with the drug profits; however, he found no allies at the FDA. George P. Larrick, who had started with the FDA in 1923 as a food and drug inspector, had been commissioner since August of 1954; he did not believe that the FDA should be involved with the pricing of drugs, only chemical purity and truthful labeling. Also, the Kennedy White House was not supportive of a bill to change the drug industry. By spring 1962, Kefauver had almost lost the battle; however, the thalidomide drug disaster would change how the US and the world looked at drug safety (FDA Backgrounder, 1999; Hilts, 2003; Swann, 1998).

                                                                          

Thalidomide

Thalidomide was first sold in Germany in 1957, marketed by Chemie Grunenthal. The drug was sold as an over-the-counter product, as a strong but safe sedative. Thalidomide was purported to alleviate anxiety, sleeplessness, and morning sickness associated with pregnancy. It was soon for sale in most of Europe; however, throughout 1958 and 1959, reports of serious side effects from Thalidomide started to reach Grunenthal, which did nothing. Complaints such as dizziness, nervousness, wakefulness, and giddiness in patients were followed by more serious complaints, such as peripheral neuropathy, a nerve affliction that leaves a patient with a debilitating loss of strength and numb hands and feet.

Thalidomide had received the interest from Richardson-Merrell, Inc, a subsidiary of the Vick Chemical Company in the US. Richardson-Merrell wanted to put Thalidomide on the US market and started a vigorous campaign to have the drug cleared by the FDA. By early 1959, Richardson-Merrell had signed the licensing agreement with Grunenthal and shortly thereafter started giving patients in the US the drug as part of a “clinical study.”  The application for clearing the drug in the US was given to Dr. Frances Kelsey, M.D., Ph.D. for review. Dr. Kelsey asked Richardson-Merrell for proof that the drug was safe for pregnancy; however, the company had no such information, since it had not conducted any studies. By fall of 1961, Grunenthal had started settling lawsuits brought by European patients, especially suits involving birth defects. It now became clear that Thalidomide causes phocomelia (seal-limbs), wherein the baby is born without the long bones in the extremities. Also, oftentimes these children would be born with hands and feet attached right at the trunk, or sometimes just fingers and toes attached at the shoulders and torso. By November of 1961, Grunenthal had withdrawn the drug from the German market and the news soon reached Dr. Kelsey at the FDA. Unfortunately, Commissioner Larrick’s actions were slow; not until August 1962 did he order FDA field investigators to remove Thalidomide from the US. In the end, some 8,000 grossly deformed babies were born in Germany and other European countries. It is also estimated that some 5,000 to 7,000 babies were still-born due to their deformities. In the US only some 40 cases may have occurred (FDA Backgrounder, 1999; Hilts, 2003; Swann, 1998).     

However, the Thalidomide catastrophe resurrected the Kefauver bill. In October of 1962, Congress passed the Kefauver-Harris amendment to the US food and drug laws. President Kennedy signed the bill on October the 10th. The amendment closed the un-scientific “clinical trial” provision in the 1938 act; instead the amendment made it mandatory for drug companies to conduct proper clinical trials. Now, the burden of proof lay with the drug companies and the FDA was put in charge to evaluate the safety and effectiveness of the new drugs (FDA Backgrounder, 1999; Hilts, 2003; Swann, 1998).     

                                                                   

Medical Device Law

Even though the 1938 Federal Food, Drug, and Cosmetic Act ostensibly treated a medical device as a drug, the FDA provided very little oversight for devices. There was also confusion as to what government agency was responsible for what. In 1948, as a result of the bombings of Hiroshima and Nagasaki and its aftermath, the Radiological Health Unit was established by the Bureau of State Services, U.S. Public Health Service. By 1966, the Division of Radiological Health was renamed the National Center for Radiological Health and again renamed in 1968, at this time the Bureau of Radiological Health. By 1984, another name change took place, which is also the center’s current name; the Center for Devices and Radiological Health (CDRH; Center for Devices and Radiological Health [CDRH] Milestones, 2006; FDA Backgrounder, 1999; Swann, 1998).

In the fall of 1969, President Nixon issued a message to Congress calling for certain minimum standards and for pre-market clearance for some medical devices. Shortly thereafter, Theodore Cooper, then assistant secretary of the Department of Health, Education, and Welfare, was assigned to investigate whether or not a federal program to require pre-market clearance of medical devices was needed. The committee, known as the Cooper Committee, completed an extensive literature search regarding faulty or dangerous medical devices. The Cooper Committee reported that over a 10-year period, some 10,000 device-related injuries had taken place, of which some 700 had resulted in death. The study further classified that some 500 deaths were attributable to prosthetic heart valves, some 190 injuries and 90 deaths from the usage of pacemakers and some 8,000 injuries and 10 deaths resulting from intra-uterine devices (IUDs). The most egregious of the IUDs was the Dalcon Shield, manufactured by the A. H. Robins Company (Burkholz, 1994; CDRH Milestones, 2006; Swann, 1998).

 

 

The Dalcon Shield

The Dalcon Shield had been developed in 1968 by Hugh Davis, a gynecologist and Irwin Lerner, an electrical engineer. In order to overcome the uterus’ natural tendency to expel any foreign object, the Dalcon Shield was designed with prongs to hold it in place; however, the prongs often got embedded in the uterine wall, resulting in problems as outlined below. One of these unwanted results occurred because the tail string of the device was made from a multi-filament string, as opposed to the mono-filament that was used in most other IUDs at the time. The multi-filament would cause bacteria to travel up into the uterus, causing infections. In early 1970, Davis and Lerner sold the device to A. H. Robins and the IUD was put on the market without any pre-market safety testing (Burkholz, 1994; CDRH Milestones, 2006; Swann, 1998).

During 1972 and 1973, A. H. Robins received numerous reports of uterine infections and spontaneous septic abortions, after women had become pregnant in spite of using the device. As Burkholz points out (1994), in 1973, A.H. Robins added a warning label, instructing women of the possibility of “severe sepsis with fatal outcome, most often associated with spontaneous abortion following pregnancy with the Dalcon Shield in situ” (p. 66). Due to negative publicity, A. H. Robins finally withdrew the Dalcon Shield from the market in April of 1975, after some 15 deaths and 245 non-fatal septic abortions had been reported. In the end, A. H. Robins filed for bankruptcy and was left with an estimated liability of $2.4 billion (Burkholz, 1994; CDRH Milestones, 2006; Swann, 1998).

Under the law at the time, the FDA had only limited authority to regulate the Dalcon Shield. However, by 1976, the Medical Devices Amendment was enacted, as a supplement to the 1938 Food, Drug and Cosmetic Act. The Medical Devices Amendment gave the FDA responsibility to classify the thousands of medical devices on the market as either Class I (low risk), Class II (medium risk), and Class III (high risk, usually involving implantable and or lifesaving devices). As a consequence of the 1976 amendments, in 1977, the Bureau of Medical Devices and Diagnostic Products was re-named the Bureau of Medical Devices, in order to implement the amendments. The 1976 amendment also required a medical device manufacturer to establish a quality system, at the time known as Good Manufacturing Practices, or GMPs. The GMP regulation became effective in late 1978. The GMPs were also specific regarding the manufacturing controls needed for the manufacture of Class III (implantable and lifesaving) devices. As a consequence of these changes, the Bjork-Shiley Convexo-Concave became the first prosthetic heart valve to undergo the new Pre-Market Approval (PMA) process and Shiley Inc. submitted their PMA in November of 1978 and got approval by April of 1979 (The Bjork-Shiley Heart Valve, 1990; Burkholz, 1994; CDRH Milestones, 2006; Swann, 1998).

 

 

 

 

 

 

 

 

 

 

The 1980s and the FDA

 

 

Table 1

FDA Commissioners, 1954-1997

Commissioner Name (President served):	Date in office, beginning:	Date in office, end:	Tenure, months:
George P. Larrick (Eisenhower, Kennedy, Johnson)	8/12/1954	12/27/1965	136
James L. Goddard, MD (Johnson)	1/17/1966	7/1/1968	31
Herbert Ley Jr. MD (Johnson)	7/1/1968	12/12/1969	17
Charles Edwards, MD (Nixon)	12/13/1969	3/15/1973	39
Shervin Gardner, Deputy (Nixon)	3/16/1973	7/19/1973	4
Alexander M. Schmidt, MD (, Nixon, Ford)	7/20/1973	11/30/1976	40
Shervin Gardner, Deputy (Ford)	12/1/1976	4/3/1977	4
Donald Kennedy, Ph.D. (Carter)	4/4/1977	6/30/1979	26
Shervin Gardner, Deputy (Carter)	7/1/1979	10/20/1979	4
Jere E. Goyan, Ph.D. (Carter)	10/21/1979	1/20/1981	15
Mark Novitch, Deputy (Reagan)	1/21/1981	4/12/1981	3
Arthur Hull Hayes Jr. MD (Reagan)	4/13/1981	9/11/1983	29
Mark Novitch, Deputy (Reagan)	9/12/1983	7/14/1984	10
Frank E. Young, MD (Reagan, Bush)	7/15/1984	12/17/1989	65
James Benson, Deputy (Bush)	12/18/1989	11/7/1989	11
David A. Kessler, MD (Bush, Clinton)	11/8/1990	2/28/1997	75

 

Source: www.fda.gov/oc/commssioners/

 

 

As can be seen from Table 1, George P. Larrick ended an era at the FDA of the tenured Commissioner. After Larrick, the office of the Commissioner became a political appointment and only a handful of commissioners would cross two presidential administrations. In between Larrick and Kessler, the average tenure for the Commissioner (including the Deputies) would be less than 22 months, which in of itself would cause the agency to become demoralized. However, even more important, during the Carter years and continuing through the Reagan administration, the FDA received less and less funding and resources. In 1978 the FDA had about 7,850 employees; then President Carter cut the number to 7,500. In his first budget, President Reagan cut the number to 6,800. Given this backdrop, there were three major challenges the FDA faced during the 1980s; the AIDS crisis, the generic drug scandal and the Bjork-Shiley Convexo-Concave heart valve debacle (Burkholz, 1994; Hilts, 2003; Swann, 1998).

                                                                  

The AIDS Crisis

The first signs of the new epidemic started to surface in the summer of 1981, when the Center for Disease Control (CDC) in Atlanta saw an increase in pheumocystis carinii, a fatal form of pneumonia. Soon after came reports of Kaposi’s sarcoma, a rare cancer that usually only affected people with seriously compromised immune systems. By the middle of 1982, the cases of this new and deadly epidemic had spread from the homosexual communities in Los Angeles, San Francisco, and New York to Haitian immigrants and hemophiliacs who had received tainted blood transfusions. By May of 1983, some 1,450 cases of the new disease had been reported, with approximately 560 dead. As the new disease was named Acquired Immuno-Deficiency Syndrome (AIDS), there were again changes at the FDA. Commissioner Art Hayes had left in the fall of 1983, his successor Mark Novitch (registered Democrat), lasted 10 months. As Hilts (2003) pointed out:

The new commissioner would face slashed budgets, an epidemic in the making, an overburdened workforce, and a whole lot of partisan politics. And for a public health official, needless to say, the pay is poor, and the person who holds down the job can assume that there will be regular attacks on his or her reputation, intelligence, and integrity. At that moment [summer 1984], the agency was unable to fulfill the missions given to it by Congress. At least six people turned down the job. (p. 238) 

Finally, Frank Young accepted the position and was sworn into office on July 15, 1984. Dr. Young was the former dean of the medical school at University of Rochester. He had been recommended to the Reagan administration by New York conservatives and after having been asked three times accepted the position. Dr. Young would stay with the FDA through the end of 1989, which were some of the most difficult times that the FDA would experience. The questions of blood banks and possibly tainted blood, the lack of any viable anti-AIDS drugs, and the more and more vocal AIDS activists, would soon consume the FDA (Burkholz, 1994; Hilts, 2003; Swann, 1998).

ATZ became available for human use by July-1985. Azidothymidine had been developed by Dr. Jerome Horwitz by the Michigan Cancer Institute in 1964, operating under a grant from the National Cancer Institute (NCI). However, since the compound was not effective against cancer, the drug was shelved and since Dr. Horwitz never applied for a patent, the compound would enter into the public domain. Some 20 years later, Dr. Samuel Broder of NCI persuaded Burroughs Wellcome, an American subsidiary of Wellcome PLC of Great Britain, to start manufacture of ATZ. The clinical trial started in February of 1986 and after three months, the initial results were clearly promising. Out of the 145 patients who took ATZ, only one had died of the disease. However, out of the 137-patient control group that had taken a placebo, 19 had died. Due to these results, the phase III trials (which usually take years and involve some 1,000 patients) were waived by the FDA, since the patient group was in desperate need for any drug that would show promise. Burroughs Wellcome submitted their New Drug Approval (NDA) in December of 1986 and ATZ was cleared for sale by the FDA by March of 1987 (Burkholz, 1994; Hilts, 2003; Swann, 1998).

Equally distressing, the question of the blood banks and tainted blood became another serious challenge for the FDA. In December 1982, the CDC had been alerted to cases of the new immunological disorder affecting people that had received blood transfusions, such as hemophiliacs. At the time, the organizations that collected blood fell into two groups; the non-profits Red Cross, American Association of Blood Banks (AABB), and the Council of Community Blood Centers (CCBC). The other group was made up of commercial companies that focused on collecting blood plasma from paid donors. Unfortunately, it was primarily the non-profit groups that evaded their responsibilities during the early years of the AIDS crisis. In March of 1983, the FDA had issued screening recommendations to the non-profit blood banks; however, these recommendations were largely ignored. Not until March of 1985 did the FDA take any substantial action; at this time the FDA approved and enforced the first AIDS antibody screen test. Unfortunately, this action would later pit the FDA against the Red Cross, when it became clear that the Red Cross had serious problems with its procedures for testing and keeping track of donated blood. Not until 1991, when the Red Cross announced sweeping changes in its handling of blood products, did some of the problems at the Red Cross get under control (Burkholz, 1994; Hilts, 2003; Swann, 1998).

 

The Generic Drug Scandal

In 1984 Congress passed the Drug Price Competition and Patent Extension Act (Hatch-Waxman). The law was passed in order to expedite the approval process for generic drugs through the FDA and to make these generics available to the public at substantially lower costs. Instead of lengthy clinical trials and New Drug Applications, the makers of generic drugs only had to show that their drugs were the chemical equivalent to the brand name drugs. The FDA would review the generic drug manufacturer’s Abbreviated New Drug Application (ANDA) along with samples of the generic drug and the FDA would either approve or deny the application (Burkholz, 1994; FDA Backgrounder, 1999; Hilts, 2003; Swann, 1998).

However, during the later part of the decade, it became apparent to some of the generic drug manufacturers that bribery and favoritism would influence the ANDA approvals. Mylan Laboratories of Pittsburg and its chairman Roy McKnight had heard rumors of payoffs and other favors that were expected by some of the ANDA reviewers at the FDA. Charles Y. Chang, a group leader within the generic drug division at the FDA, was one of the key suspects. In May of 1989, hearings held by the Subcommittee on Oversight and Investigations of the House Energy and Commerce Committee, lead by Representative John Dingell, revealed that Chang and others had taken bribes in the form of money, free weekend trips, and a fur coat. The hearings also revealed that many of the generic drug manufacturers had falsified their chemical equivalency tests, in order to expedite the approval process, only to market less-potent or dangerous drugs after they had gained approval (Burkholz, 1994; Hilts, 2003; Swann, 1998).

In the end, some fifty-five employees of fifteen generic drug manufacturers would be convicted of felonies, including five FDA officials. Most notably, Dr. Marvin Seife, then director of the Division of Generic Drugs at the FDA, would inadvertently receive the harshest sentence. After having been convicted of perjury for lying about some free lunches received from his industry contacts, he was sentenced to five months in a work-release facility. Due to federal sentencing guidelines, Dr. Seife turned himself in at the federal prison in Three Rivers, Texas, in January of 1992. Unfortunately, Seife’s paperwork was missing at the prison and he was put into solitary confinement for twelve days. During this time, having had his shoes confiscated and exchanged for boots that were too small, his blistered left foot got infected, which ultimately lead to gangrene. In the end, in order to save his life, his left leg had to be amputated below the knee (Burkholz, 1994).

The generic drug scandal lead to the enactment of the Generic Drug Enforcement Act of 1992, which is often called the “debarment act.” This act gave the FDA the ability to debar (exclude) individuals from working in the drug industry by publishing the individuals name in the debarment list (which can be found at the FDA’s webpage). On top of the current list is the infamous Charles Y. Chang, followed by a host of individuals, many who were debarred right after the passage of the act, presumably all involved in the generic drug scandal. The act is enforced by having the drug companies file a statement along with a new drug application which affirms that no debarred individuals worked on the application. Fines are steep; if a drug company is found to associate with a debarred individual, it can be fined up to $1 million; furthermore a debarred individual can be fined up to $250,000 if found working in the drug industry, in any capacity between janitor and CEO (FDA Backgrounder, 1999; Hilts, 2003).  

                                                                 

The Bjork-Shiley Convexo-Concave Heart Valve

In 1984, Dr. Sidney Wolfe of the Public Citizen Health Research consumer advocate group in Washington petitioned the FDA to have the Bjork-Shiley Convexo-Concave valve removed from the market. In spite of Wolfe’s urging, the FDA took no action, save for pressuring Shiley to write “Dear Doctor” letters, in order to inform the medical community of the fractures. The local Los Angeles FDA office also inspected Shiley numerous times and suggested that Rockville (FDA headquarters) take action; however, none were taken specifically against Shiley (The Bjork-Shiley Heart Valve, 1990). However, it can be argued that the Convexo-Concave valve problems were at least in part responsible for the 1984 Medical Device Reporting (MDR) regulation, which was published in September of 1984. The MDR regulation requires that a manufacturer or importer of medical devices maintain files when one of their devices may have caused or contributed to a patient death or serious injury and report these incidents to the FDA in a timely manner (CDRH milestones, 2006).  

Due to mounting law suits and negative publicity, by November of 1986 Shiley voluntarily withdrew the Bjork-Shiley Convexo-Concave valve from the market. In 1989, the Subcommittee on Oversight and Investigations, of the Committee on Energy and Commerce under Representative John Dingell started an investigation into the Convexo-Concave valve and the FDA’s apparent lack of action. This investigation culminated in the report “The Bjork-Shiley Heart Valve: ‘Earn as You Learn.’ Shiley Inc.’s Breach of the Honor System and FDA’s Failure in Medical Device Regulation.” This report re-awakened the interest in the heart valve failure and a series of highly-publicized lawsuits followed. By 1992, Pfizer had enough and sold Shiley to Sorin Biomedica, a subsidiary of the Italian industrial giant Fiat (“Mergers, Acquisitions,” 1992). The Bjork-Shiley Convexo-Concave heart valve story culminated in 1992 with the Bowling-Pfizer settlement, wherein Pfizer set aside hundreds of millions of dollars for patients and research (Tidmarsh, 1998).

During the 1980s, the FDA had suffered numerous congressional investigations, setbacks, and crises, such as AIDS, Bjork-Shiley, and the generic drug scandal, while working with a decimated staff. Nevertheless, during this time the FDA’s workload had increased dramatically, with applications for drug and medical device approvals nearly tripling between 1970 and 1989 (from 4,200 to 12,800 in the time period). Furthermore, between 1980 and 1990, Congress had passed twenty-four new laws giving the FDA new responsibilities. Also, the Freedom of Information Act (FOIA) had added another burden; as a consequence the FDA was responding to some 40,000 FOIA requests in 1989. Due to the generic drug scandal, commissioner Young had left in late 1989, after almost five years in office (Burkholz, 1994; Hilts, 2003). In the CDRH, Dr. John Villforth had also left, after having served as Center Director since 1982. Villforth’s style of focusing on engineering solutions to medical device problems as opposed to legal actions, primarily due to a chronic shortage of staff, was not popular after the 1989 Bjork-Shiley hearings (Perrone, 2006). Villforth was replaced with Walter Gundaker (1990-1991), James Benson (1991-1992), and not until Bruce Burlington (1993-1999) took over in 1993 did the Center receive some long-term stability (CDRH Milestones, 2006).  

However, there had been one positive outcome of the scandals in the 1980s; the passage of the Safe Medical Devices Act (SMDA) in 1990, which was in no small part a result of the Bjork-Shiley failures. The SMDA requires facilities that use medical devices (such as hospitals) to report to the FDA any incident where a medical device may have contributed to a serious injury or death of a patient. The SMDA also requires manufacturers of critical implantable medical devices to track them, in order to be able to notify patients. Lastly, the act also gave the FDA the power to order medical device recalls (CDRH Milestones, 2006).

                                                                               

1990 Through 1997

Dr. David Kessler, MD, JD, took over as commissioner of the FDA in November of 1990. He had received his medical degree from Harvard Medical School and his law degree from the University of Chicago. He was only 39 when he took over and from the start of his tenure he signaled a new course for the FDA. In short, deregulation would be no more; instead, enforcement of the law would be Dr. Kessler’s message to the food, drug, and medical device industries. In April of 1992, the FDA took action against Procter and Gamble, over P&G’s claims that its Citrus Hill Fresh Choice was indeed “fresh” (the drink was largely made from orange pulp, water, oils, and flavorings, sometimes months after harvest). The FDA seized some 24,000 cartons of the drink in Minneapolis, charging that the product was misbranded. This started Dr. Kessler’s quest for truthful product labeling that could be useful for consumers, which would result in the nutrition facts labeling in 1992 (FDA Commissioners, 2009; Hilts, 2003).

Medical device manufacturers were also being pursued for wrong-doing. In 1993, C.R. Bard Inc., pleaded guilty to fraud after a 1990 FDA investigation had found evidence that the company had marketed unapproved angioplasty catheters. In this case, Bard paid $61 million in fines; at the time this was the highest penalty ever imposed for health-care fraud. The illegal activities had started around 1987, when Bard redesigned a balloon catheter due to failures in the field. The redesigned catheter had been used in illegal clinical trials, during which it became evident that the redesigned catheter tip would break off in two out of every 100 patients. This information was kept secret from the FDA, who approved the redesigned catheter in 1989. After a series of unethical acts by Bard, the FDA seized some 1,800 catheters in February of 1990; by November of the same year the FDA ordered the destruction of these catheters. In addition to the parent company paying a very substantial penalty, in 1995 the FDA went after some of the responsible top executives at Bard, who were put on trial and found guilty of conspiring to defraud the FDA. Consequently, in 1996 three high-ranking Bard managers were sentenced to 18 months in prison each; David Prigmore, corporate executive vice president; John Cvinar, president, and Lee Leichter, director of regulatory affairs and quality assurance (Kurtzweil, 1996).              

The FDA also became alarmed when reports of leaking silicone breast implants started to become public. Silicone breast augmentation can be traced back to post-world war II, when Japanese prostitutes, along with back-alley doctors, started to use crude silicone implants in order to attract business from the US occupying force. The modern form of the silicone breast implant with its gel center and elastomeric envelope was developed in the early 1960s and subsequently manufactured by Dow Corning. As early as the 1970s, studies had been conducted that showed a relationship between breast implants and an increase in antibodies, which, in some cases, can lead to relatively benign problems such as joint pain, rashes, and flu symptoms. Furthermore, in some cases more serious problems would surface, such as arthritis and lupus. The issue had started to surface as early as 1982, when the FDA had suggested that implant manufacturers start collecting safety data, a request that was largely ignored (Burkholz, 1994). By 1988, after a series of highly publicized law suits, the FDA recognized that breast implants were devices that needed safety studies. However, by 1991, the data coming from implant manufacturers pointed mostly to problems, which had been known to the manufacturers for decades. This led to a heated battle between plastic surgeons and the breast implant manufactures on one side and the FDA on the other. Dr. Kessler was publicly maligned and after much publicity, the FDA choose a middle ground, wherein the silicone implants could still be used as long as women and their surgeons would agree to be part of a large safety study (Hilts, 2003).

The Safe Medical Devices Act of 1990 also gave the FDA the authority to add pre-production design controls to the Good Manufacturing Practices regulation. In January 1990, the CDRH had published “Device Recalls: A Study of Quality Problems” (document FDA 90-4235). This report looked at the various reasons for recalls of medical devices, during the period from October 1983 through September 1988. The report found that 44% of all recalls were attributed to “Preproduction” problems, which included device design, component design, and process design. Another 47% of recalls were due to problems with Good Manufacturing Practices, while the remaining recalls were due to other causes, such as failure to follow standards promulgated under the Radiation Control for Health and Safety Act of 1968 (the most problematic device during this time was suntan beds, followed by respirators). Implantable pacemakers were also on the list, but prosthetic heart valves were not, possibly due to the fact that most of Shiley’s Convexo-Concave recalls had taken place between February 1981 and July 1983 (U.S. Department of Health and Human Services, Device Recalls, 1990).   

Largely based on the ISO 9001:1987 quality management system standard, the new Current Good Manufacturing Practices (CGMP) including design controls were published as the Quality System Regulation (QSR) in the Federal Register in October of 1996 and became effective on June 1, 1997 (CDRH Milestones, 2006). By this time, Dr. Kessler had left the FDA, after more than six years as commissioner. The 1996 Quality System Regulation is still in effect un-changed as of 2009 and serves as the back-bone to the medical device industry. Through its structure, the QSR regulation in addition to the FDA’s right to inspections will, more often than not, ensure that only safe and effective medical devices are released to the market. However, as has been shown throughout the FDA’s 100-year history, many companies do not behave ethically and the need for a strong FDA has not diminished.      

 

 

CHAPTER 7

 

DESIGN, MANUFACTURING, AND FAILURE MODE OF THE

BJORK-SHILEY CONVEXO-CONCAVE HEART VALVE

 

 

Teflon fabric- coated suture ring   Pyrolytic carbon Convexo-Concave disc   Integral inflow struts   Haynes 25 orifice ring

 

 

 

 

 

 

 

Figure 12. Bjork-Shiley Convexo-Concave heart valve prosthesis, showing the orifice ring, welded outflow struts (white arrows), the integral inflow struts and the suture ring. The prosthetic valve is shown from the inflow side. Source: Article by van Gorp et al. in Radiology (2004). 

 

The Bjork-Shiley Convexo-Concave Heart Valve

The Shiley Convexo-Concave Prosthetic Heart valve was developed around 1975 and first implanted clinically in 1976. The Convexo-Concave valve was largely designed on the platform of the Bjork-Shiley Radiopaque-Spherical prosthetic heart valve; however, there were some significant design differences. Even though the Radiopaque-Spherical valve had shown very little thrombus formation (the bane of all mechanical heart valves), the Convexo-Concave valve was designed to minimize thrombus as much as possible (Bjork, 1978). Consequently, the disc shape was changed from a spherical shape to a saucer-like convex-concave shape, where the concave shape was oriented toward the inflow of the valve and the convex shape oriented toward the outflow side (see Figure 12). Furthermore, the strut design was changed, so that during opening the disc would not only open, but move slightly down in order to allow as much blood flow as possible. The valve ring was now made from Haynes 25 (51% Cobalt, 10% Nickel, 20% Chromium, and 15% Tungsten) and the outflow strut was still welded to the valve ring. However, the inflow strut was now machined as an integral part of the Haynes 25 ring (Bjork, 1978, United States Patent 4057857, 1975). The inflow strut change was instituted as an improvement, since the Radiopaque-Spherical prosthesis had experienced a very limited number of inflow strut fractures (between 1969 and 1975, around 10 fractures in an estimated 90,000 valves implanted, or 0.011%; Elliot, 1976). The Bjork-Shiley Convexo-Concave prosthesis would eventually be made available in two models, the 60-degree Convexo-Concave and the 70-degree Convexo-Concave, depending on the opening angle of the tilting disc (Bjork, 1981; The Bjork-Shiley Heart Valve, 1990). It was made for both mitral and aortic positions, ranging in odd millimeter sizes from 17 mm to 33 mm (measured across the suture ring, see Figure 12). The 29 – 33 mm sizes had the same size orifice ring and disc, only the size of the suture ring differed (The Bjork-Shiley Heart Valve, 1990).   

The Haynes 25 orifice ring for the Convexo-Concave valve was machined from bar stock, which was purchased and stored by Shiley. The bar stock would be released and subsequently machined into orifice rings by Grindley Manufacturing Inc., located in Los Angeles, CA. Grindley had been making the rings for the Bjork-Shiley Radiopaque-Spherical valve, which required much ingenuity on the part of Grindley, since there were no commercial cutting tools available to cut the very hard and abrasive Haynes 25 material. Having designed and manufactured their own cutting tools, they were able to provide Shiley with the required ring design for the Radiopaque-Spherical valve. The Convexo-Concave valve introduced a new challenge for Grindley, since the new design required the inflow strut to be an integral part of the ring. Grindley turned to Electro Chemical Machining (ECM) in order to machine the part and had to acquire a new machine and new tooling in order to supply Shiley with the new Convexo-Concave ring (Grindley Manufacturing, Inc., [n.d.]).

The disc was made from pyrolytic carbon, made by CarboMedics Inc. in Austin, Texas. The disc started as a graphite core, with an imbedded, radio-opaque tantalum ring, which would enable the disc to be visualized radio-graphically. The tantalum marker ring had a notch, so that the metal could expand during the coating process (Wieting et al., 1999). The graphite core substrate was subsequently placed in a fluidized bed at about 1,500º C. The graphite core was then exposed to Low-Temperature-Isotropic (LTI) carbon, which cracked at the high temperature, depositing carbon on the disc in the bed. The carbon would be deposited as large planar molecules, which would condense on the surface of the graphite disc. The porous coating would then be polished, which would render the pyrolytic-coated disc impermeable (Bjork, 1972).

At Shiley, the ring would be polished to a high gloss finish, in order to remove any scratches or tool marks. After ring polishing, a Haynes 25 strut made from wire purchased from Cabot Corporation, would be welded onto the ring, using Tungsten Inert Gas (TIG) welding (Wieting et al, 1999). The welding would melt the two ends of the outlet struts legs and the surrounding orifice ring to form integral welds, without filler material. The TIG welding was performed in a rigid fixture bathed in argon gas, at proscribed electrical current and gas flow. Due to the rigid fixture, the welder’s control was limited to duration and times of current application, slight variations in the angle of electrode application, and minor movements of the strut within the fixture (Wentzel, Manning, Chandler, & Williams, 1999). After welding, the orifice ring and the welded struts would then be hand-polished to smooth out the weld areas and give the blood-flow interfaces a mirror finish (Omar et al., 2001). These activities were considered a batch process and were documented on a fabrication order as a lot.

After welding and polishing, the completed and polished orifice ring (with the integral inflow strut and welded outflow strut), would be moved to the disc fit department. In disc fit, the completed orifice ring would be matched up with a disc, in order to meet specified tolerance specifications (The Bjork-Shiley Heart Valve Story, 1981). After the disc was selected, the welds would be inspected by quality control for cracks and voids in the weld areas. If the welds were considered acceptable, the welded outflow strut would be subjected to flexibility testing (Omar et al., 2001).

                                                                          

Flexibility testing

As reports of outflow strut fractures started to reach Shiley in 1979-1980, the company turned to inspection and testing in an attempt to identify weak outflow struts (Omar et al., 2003). Some of these tests had been described by Shiley in their 1978 PMA filing, such as stress tests on valve struts. These tests had been developed in an attempt to perform non-destructive testing in order to determine weak welds; however, at that time it was determined that “Load versus deflection tests are incapable of distinguishing between valves with normal weld joints and those with abnormal weld joints” (U.S. Department of Health and Human Services, Shiley Statement of Safety and Efficacy, 1978, p. 9). Even so, between 1981 and 1984, Shiley performed the load deflection tests or flexibility tests as described below, at times repeating testing on lots that had been recalled twice by the company (Omar et al., 2001, 2003).       

There were two types of flexibility tests; the Hook Deflection Test (HDT) and the Load Deflection test (LDT). As reported by Omar et al (2003):

The hook deflection test was introduced to test for deficient welds. In this test, the valves were placed in a testing apparatus, a 5-kg weight was suspended from the outlet strut for less than 21-mm valves and a 7-kg weight for more than 21-mm valves and the deflection of the outlet strut measured. Subsequently, the test was allowed to be performed more than once after rework (usually re-weld) done on the valve. HDT was later replaced with the load deflection test (LDT), in which valves were placed in a specialized testing apparatus and a 5-kg load gradually was placed on the outlet strut, the weight unloaded and then it was re-applied. Recalled valves and valves manufactured during the period of changeover from HDT and LDT may have undergone both tests. The time periods for flexibility tests in the manufacturing process are summarized below:

1.       Before May 1980:  no flexibility test was performed.

2.      May 1980 to August 1981: HDT was performed only once.

3.      August 1981 to February 1982: one or more HDT was performed.

4.      February 1982 to April 1984: one or more LDT was performed, HDT and LDT were performed if the valve was recalled or manufactured during the period of changeover.

5.      April 1984: LDT performance continued as well as new checking procedures for disc-to-strut gap (DSG). (p. 833)

Nevertheless, the flexibility testing proved to have no effect in reducing the number of fractured outflow struts, since the testing itself never corrected the root cause of the failures. As has been described above, this root cause of the outflow strut fractures continues to be debated; however, it is the premise of this thesis that the root cause was faulty design, something that no testing can compensate for.

After the flexibility testing, the selected disc would be assembled into the orifice ring. After the disc was fitted, the outlet strut would be cold-bent in order to ensure that the disc was properly captured between the inlet and outlet struts and that the correct opening angle would be achieved (Omar et al., 2001). The disc fit operator would make specific tolerance checks to ensure the proper clearance between the disc and the orifice ring (Wieting et al., 1999). Lastly, the disc fit operator would set the hook-to-well gap (the distance between the outflow strut’s hook or knuckle and the bottom of the disc well). The completed sub-assembly consisting of the orifice ring, the inlet strut, the outlet strut and the disc, would then be passed on to inspection (Omar et al., 2001). After the inspection of the sub-assembly, the orifice ring would be etched with a control number for traceability purposes. From outlet strut assembly the valve document consisted of a “baggy” or baggy card, including a welding chart, which would be unique to each valve. However, after the disc assembly and touch up, the valve would be given an individual data card or Device History Data Card (DHDC) and be uniquely identified. After the etching, the finished sub-assembly would be washed, pre-inspected and if found acceptable, move on to clean room assembly (The Bjork-Shiley Heart Valve Story, 1981).                                                                        

Final Assembly

In the clean room assembly, sewing operators would make suture rings and assemble them onto the valve sub-assembly, to complete the prosthesis. The suture ring was made from sterile Teflon, which was cut into the necessary sizes. The Teflon core would be covered with Teflon fabric, which was sewn onto the core. After the suture ring was assembled, it would be inspected and re-worked if necessary and subsequently matched up with a valve sub-assembly. After the matching between the sub-assembly valve and the suture ring, the suture ring would be attached to the sub-assembly using a machine called an assembler. The assembler would stretch the completed suture ring over the valve sub-assembly, fitting the suture ring into the outer flange of the orifice ring, much like that of stretching a tire over a rim. After final inspection, the valve would be sterilized by autoclaving. Lastly, the sterilized valves would receive their final outer packaging. In parallel, the DHDC would be sent to the device history records department, where a cross-check of the accuracy of the data cards would be performed. If no documentation discrepancies were found, the device history records department would authorize the release of the valve for shipment (The Bjork-Shiley Heart Valve Story, 1981).                                                                                                                                            

Testing and Measurement for a Failure Mode

Much has been written about the Bjork-Shiley Convexo-Concave outflow strut fractures. In fact, it can be argued that the Convexo-Concave is one of the most infamous of all medical devices, even worse than the disastrous Dalcon Shield, as described in chapter 6. Many authors have attempted to determine the actual failure mode of the Convexo-Concave and most investigators have been inclined to focus on the weld, the welder’s identities, patient factors (such as age, body size, and gender), and size of prosthesis (as indicated above, most of the strut fractures occurred in the 29-33 mm valves). Most of these investigations centered on statistical analysis of patient data and manufacturing records, in order to ascertain which valves were the most likely to fracture (e.g. Ericsson et al., 1992; Kallewaard et al., 1999; Omar et al., 2001; Omar et al., 2003; Steyerberg et al., 1996; van der Graaf et al., 1992; Walker et al., 1997; Walker et al., 1995). Nevertheless, as early as 1981, Dr. Bjork had indicated that over-rotation of the disc may have an influence on the strut fractures, but not until 1999 was the disc over-rotation hypothesis shown to have substantial merit (Wieting et al., 1999).   

After 1992, when Pfizer had sold Shiley Inc. to Sorin, the Shiley Heart Valve Research Center was opened in order to continue research of the valve failures and also to serve as a clearing house for the legal interactions. In 1999, the Shiley Research Center published two important articles in the Journal of Heart Valve Disease, volume no. 8: “Strut Fracture Mechanisms of the Bjork-Shiley Convexo-Concave Heart Valve” and “Welding Metallurgy’s Putative Influence on Bjork-Shiley Convexo-Concave Valve Outlet Strut Failures” were two highly sophisticated and well written articles; however, one must not forget that they were written by Shiley and are not completely un-biased. Nevertheless, the measurement methods utilized by the Shiley Research Center were a combination of the tried and true, such as micro-hardness testing (using a Buehler Micromet II micro-hardness tester and a Wickers indentor), to computerized pulse-duplication testing and strain-stress measurements, using miniature strain gages attached to the strut legs (Wentzel et al., 1999; Wieting et al., 1999).

The two articles attempted to prove one hypothesis; that disc over-rotation was the most probable cause of the strut failures, not the welding or the welder. In disc over-rotation, it is believed that the disc would pivot on the inflow strut during disc closure, past the normal closing position, and exert extraordinary pressure on the outlet strut hook. This over-rotation would be due to an abnormal closing event, most likely due to increased pressure in the left ventricle. This is plausible, since most of the strut fractures occurred in men under the age of 50, who are likely to have stronger heart functions as compared with most other heart valve recipients (Wentzel et al., 1999; Wieting et al., 1999).

 In order to prove this failure mode, valves were tested both in vitro and in vivo. The in vitro testing was performed using a pulse-duplicator test system developed by Helmholtz Institute for Bioengineering in Aachen (HIA), Germany. According to Wieting et al (1999), the HIA pulse-duplicator was a “computer-controlled, servo-hydraulic pulse duplication system that had been used to study other prosthetic valves” (p. 209). A pulse duplicator is a heart valve testing device, wherein an alternating pulse of fluid (most often saline), is pumped through the prosthetic heart valve, alternating between forward and back pressure. The HIA pulse-duplicator allowed the Bjork-Shiley Convexo-Concave valve to be tested over a range of simulated left ventricle pressures or pulses, ranging from a low of 500mmHg/second to a high of 4,000 mmHg/second. Also, the HIA pulse-duplicator was equipped with a light gate, which allowed the disc closing velocity to be measured as well. Furthermore, miniature strain gages were attached to the outlet struts, in order to measure the stress-strain during disc closure (Wentzel et al., 1999; Wieting et al., 1999).

The in vivo testing was performed using young adult sheep, which is the preferred animal model for cardiac research. The prosthetic heart valves, which were also equipped with miniature strain gages, were implanted in the mitral position. In addition, a left ventricle tip-manometer catheter was implanted and left in place. Then, during periods when the animals were exercised on a treadmill, intra-cardiac pressure readings were taken in conjunction with strut strain load information (Wentzel et al., 1999; Wieting et al., 1999).

Lastly, valves were sectioned at the orifice ring-strut interface, fixed in clear acrylic and subsequently lapped (polished), in order to provide a surface that could be investigated by microscopy and also by micro-hardness testing (see above).

In order for the testing and investigations to be performed as outlined above, many advances in technology had been employed that were not available in the early 1980s when the first strut failures occurred. As Wieting et al (1999) point out:

Understanding the mechanisms underlying OSF [Outlet Strut Fracture] had to await the development of computer-controlled pulse duplicators that could accurately reproduce animal and human cardiac dynamics, miniature strain gages and a means of applying them to outlet strut legs and reliable optic and electronic monitoring for detecting transient forces measured in fractions of milliseconds. (p. 206) 

The findings, as outlined in the two articles, give credence to the hypothesis that during certain cardiac pressures, for an extended period of time, the disc would over-rotate. This over-rotation would put pressure on the tip of the strut hook which exceeded the strength of the Haynes 25 wire itself, as opposed to the strength of the weldment, which was found to be adequate (Wentzel et al., 1999; Wieting et al., 1999). Also, it must be noted that when Shiley changed the design from a welded inflow strut (Radiopaque-Spherical) to an integral inflow strut (Convexo-Concave), the inflow strut was now about 3 times stiffer; therefore, it would not flex as much as that compared with the Radiopaque-Spherical valve during closing. Consequently, since new design involved more disc travel, the closing speed had been increased by about 15% (Bjork, 1976). Ultimately, this stress would cause a fatigue fracture in the outflow strut. One of the legs would usually fracture before the other and after a period of months, the other leg would fracture, with subsequent disc escape (de Mol et al., 1994, Sacks et al., 1986). As such, the failure mode was one of design and not necessarily a manufacturing problem.   

                                                                               

Final Thoughts on Design and Manufacturing

One can argue that knowing the “true” failure mode of the Bjork-Shiley Convexo-Concave valve will do little to nothing for the surviving patients who still have the valve implanted. However, it may lie to rest some of the sensationalist speculations that were put forth in the 1990s, such as the “Phantom Welder” (Carley, 1991; Omar et al., 2001) and that drug and alcohol use was rampant at Shiley, Inc. (Zoroya, 1993). Furthermore, it may offer some solace to the thousands of Shiley employees that worked day after day believing that they were truly helping patients, by providing the best possible products. Also, knowing the truth shows the medical device industry how truly important device design is and that nothing can substitute for the Hippocratic Oath: “Above Else, Do No Harm.”     

CHAPTER 8

 

DESIGN CONTROLS EXPLAINED

 

Background

Similar to the work undertaken by the Cooper Committee in 1969, wherein the dangers of faulty medical devices were catalogued and reported (Burkholz, 1994; CDRH Milestones, 2006; Swann, 1998), around 1989 the FDA undertook an effort to compile and evaluate the reasons behind medical device recalls. The work was summarized in the report “Device Recalls: A Study of Quality Problems,” published in 1990 (FDA document no. 90-4235). The study looked at device recalls from October of 1983 through September of 1988. The study found that “the data reveals that the recalls were caused predominantly by errors or deficiencies in design and by inadequate good manufacturing practice (GMP) controls” (p. 1). As tabulated in the report, some 47% of the device recalls were attributable to GMP problems and another 44% were due to “Pre-Production” problems or design problems. As outlined in the report:

Many of these [recalls] could have been prevented if adequate care had been exercised in establishing physical and operational requirements for the device, including consideration of the user and user environment. This should have been done before initiating the design effort. Many of the defects could have been detected if simulated use testing had been conducted before releasing the design to production. (p. 5)

The report further stratifies the identified design problems as inadequate component design, inadequate (manufacturing) process design, defective software design, poor package design, and non-compliant label design.

Consequently, the November 28, 1990 Safe Medical Devices Act, gave the FDA a mandate to establish design controls, as part of the updated 21 CFR 820 law. The update to the 1976 FDA Current Good Manufacturing Practices regulation was primarily based on the 1987 ISO 9001 quality management system standard (in an effort to harmonize world-wide medical device quality systems), and eventually became the Quality System Regulation (QSR). In November of 1993, the FDA published the proposed changes and asked for input from interest groups, such as the medical device industry. The comments from industry and the responses were published by the FDA in October 1996; this document has become known as the “The Preamble to the Final Rule.” or “Preamble.” As such, the “Preamble” serves medical device manufacturers with an in-depth interpretation of the Quality System Regulation, wherein the FDA essentially explains its thinking behind the new regulation. Regarding the new design controls, as outlined in the “Preamble”:

The intrinsic quality of devices, including their safety and effectiveness, is established during the design phase. Thus, FDA believes that unless appropriate design controls are observed during preproduction stages of development, a finished device may be neither safe nor effective for its intended use. (p. 52616)

Regarding design controls, the preamble discusses the various reservations from the medical device industry, which at the time seemed to focus on a notion that design is an art form. According to industry, due to the creative component, design cannot be “controlled.” Instead, manufacturers and designers need latitude in order to ensure that creativity is not hindered. The FDA did not agree; consequently, the new and improved Quality System Regulation including design controls was published in 1997.

To further aid the medical device industry, in March of 1997, the FDA also published “Design Control Guidance for Medical Device Manufacturers” a guidance document which supplemented the information given in the preamble. As explained the by the “Design Control Guidance”:

Design controls are an interrelated set of practices and procedures that are incorporated into the design and development process, i.e. a system of checks and balances. Design controls make systematic assessment of the design an integral part of development. As a result, deficiencies in design input requirements and discrepancies between the proposed designs and requirements, are made evident and corrected earlier in the development process. Design controls increase the likelihood that the design transferred to production will translate into a device that is appropriate for its intended use. (p. 1)

The emphasis is put on “checks and balances” in order to ensure that a multi-disciplinary team are involved in the design process. Therefore, in most medical device companies, design documents are usually signed by representatives from research & development, quality assurance, regulatory affairs, manufacturing, operations, and clinical affairs. It is an attempt to guarantee that a medical device is not developed in isolation, instead, the checks and balances fosters teamwork among the various disciplines. Also, an integral part of design controls is the proscribed design reviews, which should ensure that any design problems are discovered early in the design. A multi-disciplinary design review will take a critical look at the device design at predetermined stages or phases of the design, in order to uncover ambiguous design requirements and also to agree and sign off on the design effort and move the design to the next phase.   

A comparison of the 1978 Good Manufacturing Practices with the 1996 Quality System Regulation reveal the following (Trautman, 1997):

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 2

      Design Control. Elements, 1978 vs. 1996

Good Manufacturing Practices (GMP) regulation, 1978:  21 CFR 820:

Quality System Regulation (QSR) 1996:  21 CFR 820:

 

        None                                                                      820.30 Design controls

                None                                                                      820.30(a) General

                        None                                                                      820.30(b) Design and development planning

        None                                                                      820.30(c) Design input

        None                                                                      820.30(d) Design output

        None                                                                      820.30(e) Design review

        None                                                                      820.30(f) Design verification

        820.160 Finished device inspection                 820.30(g) Design validation

        (simulated use testing)

        820.100 Manufacturing specifications            820.30(h) Design transfer

        and processes

        820.100(a)(2) Specification controls               820.30(i) Design changes

        None                                                                      820.30(j) Design history file

      ___________________________________________________________________

 

As can be seen from the comparison in Table 2, the 1996 Quality System Regulation added such significant activities such as design planning, design input, design output, design review, and design validation (were design validation has come to encompass much more than simply “simulated use testing”). Today, in the medical device industry, design controls are the central part of medical the device design process and it is also a significant area regarding ISO audits and FDA inspections.

                                                                                 

Design Control Elements

FDA mandated design controls (in effect by 1997) added the following to the Good Manufacturing Practices: 

a) General information.

 

b) Design and development planning, wherein it is expected that a medical device developer will plan the design upfront, as much as possible, to avoid the often haphazard design attempts that had plagued the industry.

 

c) Design input, wherein it is expected that marketing will solicit input from the market, in order to understand the needs of both patients and practitioners.

 

d) Design output, where it is expected that the design inputs, which are often nebulous, will be translated into tangibles such as design drawings, design specifications, manufacturing specifications, and methods.

 

e) Design review, where it is expected that a cross-functional team comprised most often by members from marketing, clinical affairs, regulatory affairs, research and development, quality, and manufacturing along with one or more independent reviewers, will meet at pre-defined times during the design of the device, to discuss the design and add checks and balances.

 

f) Design verification, wherein it is expected that the device manufacturer will compare items such as prototypes, reports from animal trials, literature review and other sources, to ensure that the design output meets the design inputs, at least on paper. Design verification has come to mean, did we do it right and did we do what we said we would do?” (Meaning did we meet the specifications?)

 

g) Design validation, which has come to mean clinical trials. This is one of the most important aspects of design control, where a device is tested in the field. Design Validation answers the question did we make the right device? (Meaning, did we make a device that really met the user needs and intended uses?)   

 

h) Design transfer, wherein the design is transferred from the R&D lab into manufacturing. This usually involves a series of process validation activities and includes a completed design history file and device master record, the recipe for making the device.

 

i) Design changes, wherein it is expected that any changes to the design before transfer are approved by a multi-function review board, which could be as few as two during this development phase, depending on the component risk, complexity, etc.

 

j) Design history file, wherein the history of the design is kept in an organized and controlled fashion. The design history file is usually a compilation of all the work that took place during the design effort.

 

            Even though many comments on design controls from the medical device industry indicated that this would mean the end to creativity and ingenuity, design controls have served the medical device industry well for the last 11 years since its enactment. Design controls form an integral part of the development of medical devices. Through design controls, the safety and efficacy of the device is established and proven. Furthermore, the design history file serves as a map for future design efforts, while providing a central store of documentation relating to the device. Given the great diversity in medical devices that have been brought to market during the last 10 years, such as Magnetic Resonance Imaging (MRI) machines, Phaco Emulsifiers (used in ophthalmology to remove the natural lens of the eye due to cataracts), a host of orthopedic implants, and new and improved prosthetic heart valves, the concerns from industry regarding their expected loss of creativity and innovation seem unfounded.

                                                                                        

Central Question

            The central question of this thesis asks are design controls effective in preventing poorly designed medical devices from entering the market place? However, one must understand that no quality system is perfect; after all, the system is administered by people, individuals that have varying levels of experience and education, and varying levels of interest in complying with the regulations encompassed in the quality system. Also, regarding Shiley, Inc., the Radiopaque-Spherical, and the Convexo-Concave, it is also possible that no quality system, no matter how well intended, could have prevented the Convexo-Concave valve from being released to the market. First, the Convexo-Concave had been redesigned in order to correct the few inflow strut fractures that had plagued the Radiopaque-Spherical; therefore, it is possible that when the first few outflow strut fractures occurred in the Convexo-Concave, the engineers at Shiley were just dumfounded. It is also possible that greed and hubris blinded Shiley to the problems, or that being part of a large multi-national company such as Pfizer changed the culture at Shiley from one of patient-centric to money-centric. Nevertheless, this thesis will prove whether or not design controls are effective, given a company that is truly interested in regulatory compliance.  

 

 

CHAPTER 9

 

RESULTS AND DISCUSSION

 

As outlined in the Method section, the central work of this thesis is a gap analysis wherein the requirements of 21 CFR 820.30 design controls, the “Design Control Guidance,” the “Preamble,” and the “Replacement Heart Valve Guidance” (U.S. Department of Health and Human Services, Replacement Heart Valve Guidance, 1994) are compared with the information provided by Shiley to the FDA in 1978. Shiley’s documentation consists of the original 510(k) filing and the information prepared by the FDA as the “Statement of Safety and Efficacy,” a summary of the testing and information provided in the original PMA filing. This information was organized into the appropriate sections and evaluated for content (see attachment E). Further, a hypothetical “what if” evaluation has been performed, in which the various design control elements have been extended into virtual but plausible outcomes, if design controls would have been in place by the time that the Convexo-Concave was brought to market. As such, the hypothetical scenario has answered the central question as outlined in this thesis; can design controls, as formulated in 21 CFR part 820.30 ensure that another poorly designed medical device, like the Bjork-Shiley Convexo-Concave heart valve, does not enter the market place?

It must be noted that this analysis assumes that design controls do not exist in isolation; instead, they are part of comprehensive requirements for medical devices, which encompass the Quality System Regulation (21 CFR 820), the “Design Control Guidance,” the guidance in the “Preamble,” and requirements from the “Replacement Heart Valve Guidance.”                                    

                                                                

Results

As can be from attachment E, the following sections have been identified as elements wherein the design control requirements, had they been followed by Shiley, could have prevented the Convexo-Concave from entering the market:

                                                                        

Design Output

Design output is the element of design controls where design inputs are transferred into physical outputs such as product specifications, component, sub-assembly and final assembly drawings, manufacturing specifications, labeling, and Directions/Instructions for Use (DFU/IFU). It oftentimes also consists of sterilization methods, transport and storage requirements, and special handling specifications. Also at this stage, the various Failure Modes and Effects Analysis (FMEA) are at least in a draft form. At the present minimum, the FMEAs consist of the design FMEA, process FMEA, and system FMEA, wherein all of the accessories are evaluated along with the main device for its suitability to function as a system.

Attachment E Line 9 General Design Output

There is no evidence in the “Statement of Safety and Efficacy” that Shiley had provided the FDA with any drawings or other device specifications. Had detailed drawings been provided to the FDA, the agency may have questioned the new angle that the outflow strut was being welded at in the Convexo-Concave valve (different than the Radiopaque-Spherical). Also, if Shiley had conducted a thorough risk analysis, such as an FMEA, they may have at least considered if the new disc (Convexo-Concave) and the fact that the disc now moved down-stream during opening may have introduced new risks. In fact, it can be argued that this new feature, coupled with an integral inflow strut that was considerably stiffer, was the cause of the outflow strut failures. It had been identified that the closing speed of the convexo-concave disc was now 15% faster, which may have lead to the bi-modal closure phenomenon (Bjork, 1978).

Attachment E Line 13 Design Output for Manufacturing

In section IV, E, 1 of the “Replacement Heart Valve Guidance,” a heart valve manufacturer is required to perform manufacturing process validations and also provide information regarding the process of inserting the occluder disc. It is clearly evident that this section of the “Replacement Heart Valve Guidance” was written as a direct outcome of the Convexo-Concave problems, since it was believed that the manipulation of the outflow strut during disc insertion may have exacerbated the outflow strut fractures. However, it must be recognized that the predecessor valve (the Radiopaque-Spherical valve) also had its outflow strut manipulated during disc insertion, with no known outflow strut fractures. Nevertheless, it is at least possible that a thorough manufacturing process validation may have uncovered the vulnerability of the outflow strut in the Convexo-Concave valve, given a large enough sample size.

Attachment E Line 17 Design Output for Labeling

Section IV, E, 7 in the “Replacement Heart Valve Guidance” speaks of the requirements for the DFU/IFU. There is no indication that the outflow strut fractures in the Convexo-Concave valve were due to inadequate directions or instructions for use. However, if Shiley had included information regarding the outflow strut fracture that had occurred during the clinical trials, it is at least possible that some surgeons may have elected not to use the Convexo-Concave valve.      

                                                                      

Design Review

Design review is an integral component of design controls. This requirement is aimed at ensuring that the design team conducts timely reviews of the various design features of the device. As Trautman (1997) points out, the “design reviews should be conducted at major decision points throughout the design phase” (p. 62). As such, the design review should serve as an early warning system, wherein new or previously undetected design problems will be brought to light. Furthermore, with independent reviewers as currently required, some problems may have been uncovered.

Attachment E Line 19 General Design Review

The objective of a comprehensive design review is to discuss and review the design, such that any issues can be brought to the design team for resolution. It is at least possible that if a design review had taken place after the first Convexo-Concave outflow strut fracture during the clinical trials that this problem may have been discussed and addressed.

                                                                     

Design Verification

As outlined in the 1997“Design Control Guidance,” the “basis of verification is a three-pronged approach involving tests, inspection, and analyses” (p. 30). Design verification often consists of process validations, sterility validations, animal studies, first article inspections, bio-compatibility studies, shelf-life studies, and packaging validations.

Attachment E, Line 21 General Design Verification

The design verification activities carried out by Shiley appears to have been haphazard and not carried out in a planned fashion. In the “Replacement Heart Valve Guidance,” design verification activities play a central role and the guidance instructs a manufacturer to carry out a series of planned tests, using a pre-determined number of samples. Consequently, it is at least possible that if Shiley had complied with all of the prescribed design verification testing in the “Replacement Heart Valve Guidance,” they may have uncovered more outflow strut fractures. This, in turn, may have prevented Shiley from bringing the Convexo-Concave valve to the market.

Attachment E Line 25 Design Verification In-Vitro Studies

The testing proscribed in section VI, A, 2 of the “Replacement Heart Valve Guidance” is comprehensive, including both mitral and aortic valves, in sizes ranging from the smallest to the largest. Therefore, it is at least possible that outflow strut fractures may have been encountered during in-vitro testing, which may have precluded the Convexo-Concave valve from entering the market. In fact, as was outlined in the “Statement of Safety and Efficacy” (1978) during elevated pressure testing, three out of five valves tested experienced “lower leg failure” (Table 2 in the “Statement”) which is believed to be synonymous with outflow strut failure.

                                                                       

 

Design Validation

In the broadest of terms, design validation has become synonymous with clinical trials. Since design validation “shall ensure that devices conform to defined user needs and intended uses” (Trautman, 1997, p. 66), a clinical trial is oftentimes the only way to truly confirm that device functions as intended. Also, a clinical trial will test the device not only in the patient, but it will also show how the device functions in the hands of the physician. In the case of the “Replacement Heart Valve Guidance,” a clinical trial is now mandatory.   

Attachment E, Line 34 Design Validation Regulatory Issues

Having a clinical protocol that complies with the requirements of the “Replacement Heart Valve Guidance” could have ensured that a sufficient number of patients would have been enrolled in the clinical trial. Therefore, having a clinical protocol (assuming that the number of patients was insufficient, as was the case in the actual Convexo-Concave clinical trials), could have prevented the Convexo-Concave valve from entering the market.

Attachment E Line 36 Design Validation Study Design

Section VI, B, 3 of the “Replacement Heart Valve Guidance” speaks to conducting a statistical hypothesis testing. Consequently, it is at least plausible that if Shiley had conducted a statistical study, given a large enough sample, that they may have concluded that given the outflow strut fracture, the Convexo-Concave was no more safe and effective than the Radiopaque-Spherical valve. Instead, they may have concluded that the Convexo-Concave valve was more risky. Therefore, complying with this requirement may have prevented the Convexo-Concave valve from entering the market.

Attachment E Line 37 Design Validation Sample Size

Section VI, B, 3, a in the “Replacement Heart Valve Guidance” speaks to the sample size for clinical trials. The clinical trial for the Convexo-Concave valve was started in June of 1976 and reported in November of 1978, “the inducted in one year, then followed for an additional year” (Replacement Heart Valve Guidance, 1994, p. 35) model will be assumed. As such, instead of having 378 patients in their clinical trial, Shiley would have needed 556 patients (a 47% increase). Furthermore, instead of having an approximate 144 patients receiving a mitral valve, Shiley would have needed 278 mitral patients (a 90% increase). Therefore, it is at least plausible that having an increased number of patients and mitral implants would have resulted in additional outflow strut fractures. This fact may have led both Shiley and the FDA to conclude that the Convexo-Concave valve had a design problem and prevented the Convexo-Concave from entering the market.

Attachment E Line 40 Design Validation Outcome Variables

Section VI, B, 3, d in the “Replacement Heart Valve Guidance” speaks to the evaluation of clinical data. This section of the “Replacement Heart Valve Guidance” requires a heart valve manufacturer to carefully and thoroughly analyze a pre-determined set of clinical data on all patients. Since very little data was provided to the FDA regarding the Ehrenhaft and the LeMole/Fernandez implant studies, it is at least possible that more strut fractures may have occurred during the clinical trial. Consequently, complying with this requirement may have uncovered additional strut fractures, which may have prevented the Convexo-Concave valve from entering the market.

Attachment E Line 41 Design Validation Follow Up

Section VI, B, 3, e of the “Replacement Heart Valve Guidance” explains the requirements for patient follow up. Given the very lax follow-up conducted by the investigators, it is at least possible that more valve-related deaths may have occurred, unbeknown to the investigator(s). Therefore, it is at least plausible that if Shiley had complied with this requirement, they may have become aware of more valve-related deaths, which may have lead Shiley and the FDA to prevent the Convexo-Concave valve from entering the market.

                                                                          

Design Transfer

In the broadest terms, the design transfer element in design controls should ensure that all of the design requirements can be fulfilled by the manufacturing process. Even though design transfer is listed after design validation, many of the design transfer activities are often conducted earlier and the various process validations become part of the design verification process. As proof of a successful design transfer, many medical device manufacturers will conduct yield studies and manufacturing readiness studies as part of design transfer activities. Also, training of manufacturing personnel and the accompanying training records will become part of the design transfer documentation.

Attachment E Line 51 Design Transfer

The design transfer section of the “Design Control Guidance for Medical Device Manufacturers” speaks to the important requirement of ensuring that design specifications are accurately translated into production specifications. There was evidence collected by the FDA, by Pfizer, and by the Dingell subcommittee that manufacturing methods were not well established at Shiley (The Bjork-Shiley Heart Valve, 1990). Therefore, if one asserts that the welding of the outflow strut to the valve ring was a contributing factor to the outflow strut fractures (which has been asserted by numerous academic investigators such as Kallewaard et al., 1990; Omar et al., 2001; Sacks et al., 1986; Walker et al., 1995), it is at least possible that having fully defined and implemented manufacturing process could have prevented some defective Convexo-Concave valves from entering the market.                                                           

                                                                               

Design History File

The design history file requirement should ensure that all of the documentation that is generated during the design activities are collected, organized, catalogued, and stored in such a manner that the history of the design is well documented. As outlined in the FDA “Design Control Guidance for Medical Device Manufacturers,” “the primary beneficiary of the design history file is the device manufacturer” (p. 43).

Attachment E Line 54 Design History File General

Having a well developed Design History File, where all of the critical inputs and outputs, all verifications and where all validations are described, can be invaluable for troubleshooting a bad or troubled design. As design teams are disbanded and staff leaves for other employers, much design history can be lost unless it is well documented. Therefore, it is at least possible that having a complete design history file may have aided Shiley management in their attempts to correct the outflow strut fractures.

Discussion

This thesis has been written from the viewpoint that the failure of the Convexo-Concave was due to poor design. Even though many authors focus on the welding and the manufacturing, the Radiopaque-Spherical did not suffer from an equal rate of failure. Therefore, the discussions regarding the various manufacturing processes miss the point. Instead, the change in design from the Radiopaque-Spherical valve to the Convexo-Concave valve involved two crucial changes; the inflow strut was made an integral part of the valve ring, thereby increasing the stiffness of the inflow strut, on which the occluder disc would pivot during closing. Also, by changing the movement of the occluder disc during opening, wherein the disc would move down and tilt, increased the closing speed of the disc (Bjork, 1978). Consequently, it is the opinion of this thesis that these two design changes caused the outflow strut failures, along with little or no in-vitro testing of the Convexo-Concave valve before release to the market.

At this point, the central question of this thesis must be answered, could design controls have prevented a poorly designed medical device such as the Convexo-Concave from entering the market place? As can be seen from the identified sections in the gap analysis, it is possible that design controls, including the adherence to the “Replacement Heart Valve Guidance,” could have prevented the Convexo-Concave from being marketed as designed. Primarily, the “Replacement Heart Valve Guidance,” with its focus on design outputs, design verification, and design validation offers safeguards through its insistence on extensive in-vitro and in-vivo testing.

However, the intent of this thesis was to answer the larger question regarding the design of all medical devices and if design controls would be generally effective in ensuring that only safe and effective devices enter the market place. Therefore, one must remember that not all medical devices have accompanying guidance documents; however, most of the implantable and class III devices do, either through the FDA or through ISO. Furthermore, it must be pointed out that before a new class III device is approved, the FDA will conduct a Pre-Market Approval inspection of the medical device manufacturer, in order to ensure that all of the elements of the Quality System Regulation (QSR) are adhered to. Therefore, this thesis answers in the affirmative that adherence to design controls, coupled with the appropriate guidance documents and FDA inspections, can indeed prevent unsafe and ineffective medical devices from entering the market place.

Nevertheless, it must be remembered that the QSR and design controls must be instilled at a device manufacturer from top management down to the operators and assemblers on the manufacturing floor. If top management fails in its commitment to the QSR, it is quite likely that functions such as design, engineering, manufacturing, and quality assurance will suffer, especially under the pressures in today’s competitive environment. As has been uncovered in the Shiley Convexo-Concave case, if top management fails in ensuring the safety and efficacy of the medical devices it sells, the rank and file are quite powerless in affecting change. After all, quality is not a grass-roots activity and a device recall must be ordered by top management, not by the lowly tool engineer working on the manufacturing floor.

CHAPTER 10

 

SUMMARY

 

Faulty Design

As has been pointed out earlier in this thesis, the outflow strut fractures that occurred in the Convexo-Concave heart valve were due to faulty design, stemming from inadequate testing and, at the time, the lack of formal design controls. Even though many authors have attempted to blame the failures on faulty manufacturing processes and other factors such as patient age, gender, and size (e.g. Ericsson et al., 1992; Kallewaard et al., 1999; Omar et al., 2001; Omar et al., 2003; Sacks et al., 1986; Steyerberg et al., 1996; van der Graaf et al., 1992; Walker et al., 1995; Walker et al., 1997), these arguments are flawed, since the Radiopaque-Spherical valve did not suffer from any known outflow strut fractures. The Convexo-Concave was essentially manufactured in the same manner as the Radiopaque-Spherical, quite likely by the same manufacturing teams. However, the change from a welded inflow strut to a substantially stiffer integral inflow strut, coupled with a new pivot point for the occluder disc, changed the closing characteristics of the Convexo-Concave valve as compared with the Radiopaque-Spherical valve (Bjork, 1978).                                                                 

                                                                             

Design Controls are Effective

As has been shown by the central gap analysis, design controls, as required by 21 CFR 820.30 may well have been effective in preventing the initially designed Convexo-Concave valve from entering the market place. Nevertheless, it must be remembered that this thesis has encompassed design controls in the broadest of terms, including a reliance on the elements in the 1994 FDA “Replacement Heart Valve Guidance.” Furthermore, it is the opinion of this author that design controls are universally effective, coupled with the appropriate device specific guidance and standard(s).      

                                                                    

A Lapse of Ethics

In the end, the case of the Bjork-Shiley Convexo-Concave heart valve is not necessarily about design controls, regulations, and device design, nor is about the Food and Drug Administration. Rather, it is about personal choices and decisions, business ethics, and money. The central question is not about failure modes, welder identities, and whether or not one particular shop order is more prone to failure. Instead, the most fundamental question is why did Shiley keep the device on the market after they knew about the failures? After all, the Convexo-Concave was designed to correct some of the design flaws of the Radiopaque-Spherical valve, such as a very low inlet strut fracture rate and some thrombus formation. Why did Shiley redesign the Radiopaque-Spherical after 10 inlet strut fractures, but keep the Convexo-Concave on the market after reports of 100 or more outflow strut fractures? Unfortunately, this literature study points to two individuals, namely Bruce Fettel, President of Shiley at the time of the outflow strut fractures and Dr. Viking Bjork, chief medical spokesman for the Bjork-Shiley heart valves.

Bruce Fettel came from North American Rockwell Corporation and joined Shiley in 1968 as Chief Engineer (Derloshon, 1983). Fettel had been intimately involved in the design of the Bjork-Shiley tilting disc valve, which later became the Bjork-Shiley Radiopaque-Spherical heart valve. In 1975, Fettel filed the patent for the Convexo-Concave heart valve (United States patent no. 4057857, 1975) and in 1978 filed the 510(k) for the same valve. By 1978, Fettel was Executive Vice President and by 1980 he was named President of Shiley. Consequently, Fettel had the most intimate knowledge about both the design of the Convexo-Concave and also the outflow strut fractures. Due to his position at Shiley, he could have withdrawn the Convexo-Concave from the market and replaced it with the Monostrut, which had been marketed outside of the US since the early 1980s. Yet, he elected to keep the Convexo-Concave on the market and in so doing alienated the FDA to the point that they would never approve the Monostrut, even though this valve had an excellent safety record. Consequently, due to his decision not to withdraw the Convexo-Concave, hundreds of lives were cut short due to a poorly designed and poorly tested prosthetic heart valve. By March of 1985, Bruce Fettel was no longer President of Shiley; whether or not he resigned voluntarily is not known.

Dr. Viking Bjork had been affiliated with Don Shiley and Shiley, Inc. since the mid 1960s (Derloshon, 1983; Westaby, 1997). He had made millions of dollars on the Bjork-Shiley family of heart valves, due to his royalty payments and had become very wealthy, especially by Swedish standards. (Dr. Bjork received a 6.5% royalty on each valve sold and it is reported that the cost for each valve was about $2,000. By 1979, the Radiopaque-Spherical valve had sold some 100,000 units, which would have earned Dr. Bjork an estimated $13 million from 1969 through 1979.)  He had traveled the world as the ambassador for the Bjork-Shiley valves, presenting data at every major thoracic surgery meeting during the 1970s and into the early 1980s (Derloshon, 1983). Therefore, it may not be surprising that he elected to keep quiet, even though he himself had re-operated on a patient with a fractured 70-degree Convexo-Concave valve in March of 1982 (The Bjork-Shiley Heart Valve, 1990). In a 1990 interview with the New York Times, Dr. Bjork asserted “Should I have gone to all the hospitals in Sweden [to speak out about the outflow strut fractures] and then after that should I have gone to Norway?  What about Holland – I demonstrated valves there, too – or Belgium or Zurich?” (Barry, 1990). This makes Dr. Bjork appear like a victim; however, in reality he had numerous opportunities to speak out. For example, in 1976 Dr. Bjork had joined Pete’s Club, an association made up of European thoracic surgeons. During the Pete’s Club meetings, only mistakes and failures can be discussed (Derloshon, 1983); consequently, this could have been one venue where the flawed Convexo-Concave could have been debated. Further, in 1978, Dr. Bjork had been chosen as President Elect of the European Cardiovascular Surgical Society and later became President of the Society; as such he presided over the first joint European and Scandinavian Cardiovascular Congress in the fall of 1982 (Derloshon, 1983). Consequently, he did not need to travel to “all the hospitals in Sweden,” instead he could have spoken out at any of the meetings he attended. Lastly, Dr. Bjork had been Chief Editor of the Scandinavian Journal of Thoracic and Cardiovascular Surgery since 1967 through 1983. He had known about outflow strut fractures in the 60-degree Convexo-Concave since 1979. He had written and published numerous articles in the journal, most of them discussing the various virtues of the Bjork-Shiley heart valves. However, he failed to use the journal as a means to inform his fellow thoracic and cardiovascular surgeons; not until 1984 did he discuss the outflow strut failures, choosing the obscure Indian Journal of Thoracic and Cardiovascular Surgery periodical as his vehicle. Consequently, in 1985 Dr. Bjork was severely censured by the Swedish Board of Health for his failure to report the 70-degree strut fractures in a timely manner (The Bjork-Shiley Heart Valve, 1990). As the report points out, “It is the opinion of the Board that the lack of judgment in Professor Bjork’s conduct is especially serious, considering his position as one of the leading experts in Sweden in this field ([since he is] also a former member of the Board’s Scientific Council on the subject of thoracic surgery; p. 122).” Consequently, as was the case with Bruce Fettel, due to Dr. Bjork’s actions and inactions, lives were cut short due to a poorly designed prosthetic heart valve.                            

 

 

CHAPTER 11

 

RECOMMENDATIONS FOR FURTHER RESEARCH

 

Due to the lengthy and costly process of acquiring the original Bjork-Shiley Convexo-Concave Pre-Market Approval (P780008) though the Freedom of Information Act, this document was not reviewed for this thesis. A request for this document was made in March of 2009; as of May 2009, this author was informed that this request is number 104 in a queue of 114 requests with a particular FDA clerk. It is estimated that the request can take 12-18 months and even so the information will be scrutinized by the FDA before being made available, in order to obliterate any proprietary information. However, it is recognized that this document could have aided in this thesis; therefore, the request will stay open with the FDA and it is recommended that another study be instigated at the time the Pre-Market Approval file becomes available.  

Also, after reviewing the Bjork-Shiley history, it appears that Shiley had acted in an ethical manner at least before the sale of Shiley, Inc. to Pfizer in 1979. In response to early reports that the Delrin disc in the Bjork-Shiley tilting disc valve could swell during autoclaving, Shiley had replaced the Delrin disc with one made from pyrolytic carbon, which is impervious to sterilization (Bjork, 1972). Also, it can be argued that the Convexo-Concave valve, with its integral inflow strut, was a clear improvement over the welded inflow strut in the Radiopaque-Spherical valve, which had experienced a limited number of inflow strut fractures. Furthermore, the changed design wherein the Convexo-Concave disc moved away from the valve ring during opening, was a design change that was implemented in order to reduce the risk of thrombus (Bjork, 1978). Therefore, it can be concluded that the introduction of the 60-degree Bjork-Shiley Convexo-Concave prosthesis was thought of as an improvement over the Radiopaque-Spherical valve regarding patient safety and efficacy. Lastly, it can also be argued that Shiley again tried to improve on the Convexo-Concave valve with the introduction of the Monostrut valve, which had integrated inflow and outflow struts (Bjork, 1983; Bjork et al., 1985).

However, at the time of the introduction of the Monostrut valve in the early to mid 1980s, the relationship between the FDA and Shiley had deteriorated such that it was no longer possible for Shiley to introduce another Bjork-Shiley valve in the US market. Part of this was due to the fact that Shiley had misled the FDA, by not reporting outflow strut failures to the FDA and also the problem with the 70 degree Bjork-Shiley Convexo-Concave valve (The Bjork-Shiley Heart Valve, 1990).

Consequently, it is at least possible that the purchase of Shiley Inc. by Pfizer had influenced the corporate culture at Shiley such that a shift from a focus on patient safety to growth and market share had taken place. Nevertheless, given the sources that have been reviewed for this thesis, there is scant or no mention of this possible corporate culture shift within Shiley. Therefore, as a follow-up study to this thesis, a business case study could be carried out, in order to evaluate corporate mergers with the Shiley-Pfizer merger as a backdrop. Such a study could further illuminate the Bjork-Shiley Convexo-Concave story and offer valuable insights into all the causes for the catastrophic failure of the Bjork-Shiley Convexo-Concave heart valve.  

Lastly, there is surprisingly little written about Dr. Viking Bjork, even though Dr. Bjork himself wrote copiously, especially in the Scandinavian Journal of Thoracic and Cardiovascular Surgery. Save for the Shiley-produced biography One For The Heart; The Story of the Professor Viking O. Bjork, M.D., no other substantial biographical work regarding Dr. Bjork has been found. The Shiley-produced biography was written by Gerald Derloshon, a Shiley employee and presented to Dr. Bjork on the occasion of his retirement in 1983. As such, One For The Heart is highly biased and makes no mention of the Convexo-Concave valve. Nevertheless, the book does serve as a historical account of Dr. Bjork’s life up until 1983.

Consequently, there appears to be a biographical void regarding Dr. Bjork; therefore, a thorough study of his life and times could be of historical interest and serve to elucidate the Bjork-Shiley story even further.              

REFERENCES

 

Bjork, V. O. (1970a). Central flow tilting disc valve for aortic valve replacement. Thorax,

25,439-444.

Bjork, V. O. (1970b). The central flow tilting disc valve prosthesis (Bjork-Shiley) for

mitral valve replacement. Scandinavian Journal of Thoracic and Cardiovascular Surgery, 4(1), 15-23.

Bjork, V. O. (1971). Aortic valve replacement with the Bjork-Shiley tilting disc valve

prosthesis. British Heart Journal, 33( Suppl.), 42-46.

Bjork, V. O. (1972). The pyrolytic carbon occluder for the Bjork-Shiley tilting disc valve

prosthesis. Scandinavian Journal of Thoracic and Cardiovascular Surgery, 6(2), 109-113.

Bjork, V. O. (1977). The history of the Bjork-Shiley tilting disc valve. Medical

Instrumentation, 11(2), 80-81.

Bjork, V. O. (1978). The improved Bjork-Shiley tilting disc valve prosthesis.

Scandinavian Journal of Thoracic and Cardiovascular Surgery, 12(2), 81-84.

Bjork, V. O. (1981). The optimal opening angle of the Bjork-Shiley tilting disc valve

prosthesis. Scandinavian Journal of Thoracic and Cardiovascular Surgery, 15(3), 223-227.

Bjork, V. O. (1983). The development of the Bjork-Shiley disc valve. Indian Journal of

Thoracic and Cardiovascular Surgery, 2, 52-54.

Bjork, V. O. (1984). The development of the Bjork-Shiley disc valve. Clinical

Physiology, 4,1-4.

Bjork, V. O. (1985). Metallurgic and design development in response to mechanical

dysfunction of Bjork-Shiley heart valves. Scandinavian Journal of Thoracic and Cardiovascular Surgery, 19(1), 1-12.

The Bjork-Shiley heart valve: “Earn as you learn.” Shiley Inc.’s breach in the honor system and FDA’s failure in medical device regulation. A Staff Report prepared for the use of the Subcommittee on Oversight and Investigations of the Committee on Energy and Commerce, U.S. House of Representatives. 1 (February 1990). Washington, DC: U.S. Government Printing Office.

The Bjork-Shiley heart valve story. (1981, February/March). Shiley News & Views, 6(2), 2-3.

Bjork-Shiley Monostrut heart valve [Online image]. (n.d.). Figure 10, retrieved August 3,

2009,from www.absoluteastronomy.com. http://images.absoluteastronomy.com/images/topicimages/a/ar/artificial_heart_valve.gif

Bjork, V. O., & Cullhed, I. (1967). Functional results with aortic ball valve prostheses

(Starr-  Edwards) followed for two to three years. Thorax, 22, 21-24.

Bjork, V. O., Henze, A., & Hindmarsh, T. (1977). Radiopaque marker in the tilting disc

of the Bjork-Shiley heart valve. Evaluation of in vivo prosthetic valve function by cineradiography. Journal of Thoracic and Cardiovascular Surgery, 73(4), 563-569.

Bjork, V. O., Lindblom, D., & Henze, A. (1985). The Monostrut strength. Scandinavian

Journal of Thoracic and Cardiovascular Surgery, 19(1), 13-19.

Bjork, V. O., Olin, C., & Astrom, H. (1969). Results of aortic valve replacement with the

Kay-Shiley disc valve. Scandinavian Journal of Thoracic and Cardiovascular Surgery, 3(2-3), 93-106.

Bjork, V. O., Riberio, A., & Canetti, M. (1999). Mechanical mitral valves without long-

term anticoagulation up to thirteen years. Journal of Thoracic and Cardiovascular Surgery, 118, 767-769.

Bjork, V. O., Riberio, A., Canetti, M., & Bomfirm, V. (1994). Mitral mechanical valve

without long-term coagulation. Eight-year follow-up. Scandinavian Journal of Thoracic and Cardiovascular Surgery, 28(1), 19-24.

Bjork, V. O., Wilson, G. J., Sternlieb, J. J., & Kaminsky, D. B. (1988). The porous metal-

surfaced heart valve. Long-term study without long-term anticoagulation in mitral position in goats. Journal of Thoracic and Cardiovascular Surgery, 95(6), 1067-1082.

Blackstone, E. H. (2005). Could it happen again [Electronic version]?  The Bjork-Shiley

Convexo-Concave heart valve story. Circulation, 111, 2717-2719.

Blot, W. J., Ibrahim, M. A., Ivey, T. D., Acheson, D. E., Brookmeyer, R., Weyman, A., et

al. (2005). Twenty-five-year experience with the Bjork-Shiley Convexo-concave heart valve [Electronic version]. Circulation, 111, 2850-2857.

Brubakk, S., Simonsen, S., Kallman, L., & Fredriksen, A. (1981). Strut Fracture in the

new Bjork-Shiley mitral valve prosthesis. Thoracic and Cardiovascular Surgeon, 29, 108-109.

 

Burkholz, H. (1994). The FDA Follies: An alarming look at our food and drugs in the

1980s. New York, NY:  BasicBooks.

Carley, W. M. (1991, November 7). Fatal flaws. Artificial heart valves that fail are linked

            to falsified records. The Wall Street Journal, pp. A1, A6.

Center for Devices and Radiological Health (CDRH) milestones. (2006). Retrieved

            February 24,2009, from: www.fda.gov/cdrh/centennial/milestones.html

de Mol, B. A., Kallewaard, M., McLellan, R. B., van Herwerden, L. A., Defauw, J. J., &

van der Graaf, Y. (1994, January 1). Single-leg strut fractures in explanted Bjork-Shiley valves. The Lancet, 343, 9-12.

Derloshon, G. (1983). One for the Heart. The Story of the Professor Viking O. Bjork,

            M.D. Irvine, CA. Shiley Incorporated.

Drawing of the Bjork-Shiley tilting disc heart valve, from the outflow side [Online

image]. (n.d.). Figure 8, retrieved August 3, 2009, from www.wikimedia.org. http://upload.wikimedia.org/wikipedia/commons/thumb/8/85/Shiley-valve.png/250px-Shiley-valve.png

Elliot, R. A. (1976). Letter:  Bjork-Shiley prosthesis recall. Circulation, 53, 206.

Ericsson, A., Lindblom, D., Semb, G., Huysmans, H. A., Thulin, L. I., Scully, H. E., et al.

(1992). Strut fracture with Bjork-Shiley 70 degrees convexo-concave valve. An international multi-institutional follow-up study. European Journal of Cardio-Thoracic Surgery, 6, 339-346.

 

 

FDA Backgrounder:  Milestones in U.S. food and drug law history. 1999. Retrieved

February 24, 2009, from:  www.fda.gov/opacom/backgrounders/miles.html.

FDA Commissioners. 2009. Retrieved February 24, 2009, from:

www.fda.gov/oc/commissioners/

Gersh, B. J. (Ed.). (2000). Mayo Clinic Heart Book. New York, NY:  William Morrow

            and Company.

Grindley Manufacturing, Inc. (n.d.). Artificial heart rings – artificial heart valves.

            Retrieved February 17, 2009, from: www.grindley-mfg.com/HeartRingText.htm

Guyton, A. C., & Hall, J. E. (2005). Textbook of medical physiology (11th ed.). Maryland

Heights, MO: Saunders/Elsevier.

Hilts, P. J. (2003). Protecting America’s health:  the FDA, business and one hundred

            years of  regulation. New York, NY: Alfred A. Knoph/Random House.

The human heart [Online image]. (n.d). Figure 5, retrieved April 10, 2009, from

www.nhf.org.nz. http://www.nhf.org.nz/images/Heart%20diagram.jpg

Kallewaard, M., Algra, A., Defauw, J., & van der Graaf, Y. (1999). Which manufacturing

characteristics are predictors of outlet strut fracture in large sixty-degree Bjork-Shiley Convexo-Concave mitral valves? The Journal of Thoracic and Cardiovascular Surgery, 117, 766-665.

Kurtzweil, P. (1996). Ex-Bard executives sentenced to prison. FDA consumer –

investigator’s reports. December 1996. Retrieved May 15, 2009, from: ww.fda.gov/fdac/departs/096-irs.html

 

Lindblom, D., Bjork, V. O., & Semb, B. K. (1986). Mechanical failure of the Bjork-

Shiley valve. Incidence, clinical presentation and management. Scandinavian Journal of Thoracic and Cardiovascular Surgery, 92(5), 894-907.

Lindblom, D., Lindblom, U., Henze, A., Bjork, V.O., & Semb, B. K. (1987). Three-year

results with the Monostrut Bjork-Shiley prosthesis. Journal of Thoracic and Cardiovascular Surgery, 94(1), 34-43.

Maddern, G., Paterson, H. S., Craddock, D. R., Ross, I. K., Stubberfield, J., Picozzi, S., et

al. (1990). A 4-year experience with the Bjork-Shiley Monostrut valve. Texas Heart Institute Journal, 17, 216-218.

Matthews, A. M. (1998). The development of the Starr-Edwards heart valve. Texas Heart

Institute Journal, 25(4), 282-293.

Mergers, acquisitions and divestitures. (1992, February 15). The New York Times, p. 1.39.

Meier, B. (1990, April 17). Designer of faulty heart valve seeks redemption in new

device [Electronic version]. The New York Times. Retrieved January 27, 2009, from www.nytimes.com. http://www.nytimes.com/1990/04/17/science/designer-of-faulty-heart-valve-seeks-redemption-in-new-device.html?scp=1&sq=Designer%20of%20Faulty%20Heart%20Valve%20Seeks%20Redemption&st=cse

Omar, Z. R., Morton, L. S., Beirne, M., Blot, W. J., Lawford, P. V., Hose, R., et al.

(2001). Outlet strut fracture of Bjork-Shiley Convexo-Concave valves:  Can valve-manufacturing characteristics explain the risk [Electronic version]? The Journal of Thoracic and Cardiovascular Surgery, 121, 1143-1149.

Omar, Z. R., Morton, L. S., Murad, S., & Taylor, K. M. (2003). Use of flexibility tests in

the manufacturing process of 60º Bjork-Shiley Convexo-Concave valves and the risk of outlet strut fracture [Electronic version]. The Journal of Thoracic and Cardiovascular Surgery, 126, 832-836.

O’Neill, W. W., Chandler, J. G., Gordon, R. E., Bakalyar, D. M., Abolfathi, A. H.,

Castellani, M. D., et al. (1995). Radiographic detection of strut separations in Bjork-Shiley Convexo-Concave mitral valves. New England Journal of Medicine, 1995 (333), 414-420. Figure 1, 2 and 4, retrieved January 26, 2009, from http://content.nejm.org/content/vol333/issue7/images/large/03f2a.jpeg,  http://content.nejm.org/content/vol333/issue7/images/large/03f2c.jpeg and http://content.nejm.org/content/vol333/issue7/images/large/03f4.jpeg

Perrone, M. (2006, August 3). Former CDRH director put problem solving before data

monitoring. The Gray Sheet. Retrieved January 30, 2009, from:  www.fda.gov/cdrh/centennial/article-villforth.html

Pratt, E. T. Jr. (1990, February 22). Dear fellow employees. Pfizer Bulletin, 40(6), 1-2.

Robicsek, F. (2009). Tribute to a pioneer: Viking Olov Bjork, 1918-2009 [Electronic

version]. The Journal of Thoracic and Cardiovascular Surgery, 137(6), 1309-1310. Figure 11, retrieved October 30, 2009, from http://www.jtcvsonline.org/article/PIIS0022522309004814/fulltext

Sacks, S. H., Harrison, M., Bischler, P. J. E., Martin, J. W., Watkins, J., & Gunning, A.

(1986). Metallurgical analysis of failed Bjork-Shiley cardiac valve prostheses. Thorax, 1986 (41), 142-147.

Starr-Edwards ball-and-cage mitral heart valve [Online image]. (n.d.). Figure 7, retrieved

August 1, 2009, from http://img.medscape.com. http://img.medscape.com/pi/emed/ckb/emergency_medicine/756148-780702-359.jpg

Starr, A., Pierie, W. R., Raible, D. A., Edwards, M. L., Siposs, G. G., & Hancock, W. D.

(1966). Cardiac valve replacement: Experience with the durability of silicone rubber. Circulation, Vols. XXXIII and XXXIV (Suppl. 1, April 1966), 115-123.

Stevenson, D. M., Yoganathan, A. P., & Franch, R. H. (1982). The Bjork-Shiley heart

valve prosthesis. Flow characteristics of the new 70º model. Scandinavian Journal of Thoracic and Cardiovascular Surgery, 16(1), 1-7.

Steyerberg, E., van der Meulen, J. H. P., van Herwerden, L. A., & Habbema, J. D. F.

(1996). Prophylactic replacement of Bjork-Shiley Convexo-Concave heart valves: An easy-to-use tool to aid decision-making in individual patients. Heart, 76, 264-268.

Stoney, W. S. (2008). Pioneers of Cardiac Surgery. Nashville, TN: Vanderbilt University

Press.

Superior view of the aortic valve showing the commissure posts [Online image]. (n.d.).

Figure 6, retrieved August 11, 2009, from www.clevelandclinicmeded.com. http://www.clevelandclinicmeded.com/medicalpubs/diseasemanagement/cardiology/aortic-valve-disease/images/aorticvalvefig1_large.jpg

Swann, J. P. (1998). History of the FDA. Retrieved February 24, 2009, from:

www.gov/oc/history/historyoffda/fulltext.html;

Tidmarsh, J. (1998). Mass tort settlement class actions: Five case studies. Federal Judicial

Center. Retrieved March 15, 2009, from:  http://www.fjc.gov/public/pdf.nsf/lookup/Tidmarsh.pdf/$file/Tidmarsh.pdf

Trautman, K. A. (1997). The FDA and worldwide quality system requirements guidebook

 For medical devices. Milwaukee, WI: ASQ Quality Press.

United States Patent no. 4057857:  Heart valve with arcuate occluder. (1975). Retrieved

January 27, 2009, from: www.freepatentsonline.com/405787.html

U.S. Department of Health and Human Services, Food and Drug Administration. (1978).

Shiley Statement of Safety and Efficacy: Bjork-Shiley Prosthetic Heart Valve. Pre-Market Approval File no. P780008. Gaithersburg, MD. FOI Services.

U.S. Department of Health and Human Services, Food and Drug Administration. (1978).

Shiley 510(k): Bjork-Shiley Cardiac Heart Valve. 510(k) file no. K780772. Gaithersburg, MD. FOI Services.

U.S. Department of Health and Human Services, Food and Drug Administration. (1996).

The Preamble. Medical Devices: Current Good Manufacturing Practice CGMP) Final Rule; Quality System Regulation. Washington, D.C: U.S. Government Printing Office.

U.S. Department of Health and Human Services, Food and Drug Administration, Center

For Devices and Radiological Health. (1997). Design Control Guidance for Medical Device Manufacturers. Washington, D.C: U.S. Government Printing Office.

 

U.S. Department of Health and Human Services, Food and Drug Administration, Center

for Devices and Radiological Health, Office of Compliance and Surveillance. (1990). Device Recalls: A Study of Quality Problems. (FDA Document no. 90-4235). Washington, D.C: U.S. Government Printing Office.

U.S. Department of Health and Human Services, Food and Drug Administration, Center

for Devices and Radiological Health, Office of Device Evaluation, Division of Cardiovascular, Respiratory and Neurological Devices, Circulatory Support and Prosthetic Devices Group. (1994). Replacement Heart Valve Guidance, Draft Document. Washington, D.C: U.S. Government Printing Office.

van der Graaf, Y., Waard, F., van Herwerden, L. A., & Defauw, J. (1992, February 1).

            Risk of strut fracture of Bjork-Shiley valves. The Lancet, 339, 257-261.

van Gorp, M. J., van der Graaf, Y., de Mol, B. A. J. M., Bakker, J. G. C., Witkamp, T. D.,

Ramos, L. M. P., et al. (2004). Bjork-Shiley Convexo-concave Valves: Susceptibility Artifacts at Brain MR Imaging and Mechanical Valve Fractures. Radiology, 230(709). Figure 3, 9 and 12, retrieved March 26, 2009, from: http://radiology.rsna.org/content/230/3/709/F2.large.jpg and http://radiology.rsna.org/content/230/3/709/F1.large.jpg 

Walker, A. M., Funch, D. P., Bianchi, L., & Blot, W. J. (1997). Shop order fracture rate

as risk factor for strut fracture in Bjork-Shiley cc60º heart valves. The Journal of Heart Valve Disease, 6, 264-267.

 

 

Walker, A. M., Funch, D. P., Sulsky, S. I., & Dreyer, N. A. (1995). Manufacturing

characteristics associated with strut fracture in Bjork-Shiley cc60º heart valves. The Journal of Heart Valve Disease, 4, 640-648.

Wentzel, T. C., Manning, C. R., Chandler, J. G., & Williams, D. F. (1999). Welding

metallurgy’s putative influence on Bjork-Shiley Convexo-Concave valve outlet strut failures. Journal of Heart Valve Diseases, 8(2), 218-231.

Westaby, S. (1997). Landmarks in cardiac surgery. St. Louis, MO: Mosby-Year Book,

            Inc.

Wieting, D. W., Eberhart, A. C., Reul, H., Breznock, E. M., Schreck, S. G., & Chandler,

J. G. (1999). Strut fracture mechanisms of the Bjork-Shiley Convexo-Concave heart valve. Journal of Heart Valve Diseases, 8(2), 206-217.

Zoroya, G. (1993, August 20). Ex-worker says marijuana, alcohol were used at Shiley 

            Inc. plant. The Orange County Register, retrieved January 27, 2009, from:

http://findarticles.com/p/articles/mi_hb5553/is_199308/ai_n22292104?tag=rel.res4

Zoroya, G., & Hirsch, J. (1993, August 3). One of the two principal developers of a

Flawed heart valve that has killed 394 patients conceded in testimony that he abandoned using the faulty valve nearly six years before the Irvine-based manufacturer pulled it from the market. The Orange County Register, p. a.01.