+ + + + +






+ + + + +



JUNE 3, 2004


+ + + + +




            The above‑entitled meeting was conducted

at 8:00 a.m., at the Gaithersburg Marriott, Salons A,

B, C, D, 9751 Washingtonian Boulevard, Gaithersburg,

Maryland, Dr. Michael J. Yaszemski, Panel Chairperson,






MICHAEL J. YASZEMSKI, M.D., Ph.D., Chairperson, Mayo

   Clinic Graduate, School of Medicine

JANET L. SCUDIERO, M.S., Acting Executive Secretary

MAUREEN A FINNEGAN, M.D., Voting Member, University

   of Texas, Southwestern Medical Center

JOHN S. KIRKPATRICK, M.D., Voting Member, University

   of Alabama, School of Medicine

KINLEY LARNTZ, Ph.D., Voting Member, Private


SANJIV. S. NAIDU, M.D., Ph.D., Voting Member,

   Pennsylvania State College of Medicine

SALLY L. MAHER, Esq., Industry Representative,

   Smith & Nephew Endoscopy

LEELEE DOYLE, Ph.D., Consumer Representative,

   University of Arkansas for Medical Sciences




MARCUS P. BESSER, Ph.D., Deputized Voting Member,

   Thomas Jefferson University

CHOLL W. KIM, M.D., Ph.D., Deputized Voting Member,

   University of California, San Diego

JAY D. MABREY, M.D., Deputized Voting Member,

   University of Texas, Health Science Center

MICHAEL B. MAYOR, M.D., Deputized Voting Member,

   Dartmouth Hitchcock Medical Center

CELIA WITTEN, M.D., Ph.D., FDA Division Director,

   General Restorative and Neurological Devices







TONI R. KINGSLEY, Ph.D., Zimmer, Inc, Warsaw,


GREG MAISLIN, M.S., M.A., Biomedical Statistical

   Consulting, Wynnewood, Pennsylvania

JAMES B. STIEHL, M.D., Medical College of Wisconsin

   and Columbia and St. Mary's Hospital, Milwaukee,


PETER S. WALKER, Ph.D., New York Medical Center, New

   York, New York




JOEL BATTS, Corin, USA, Tampa, Florida

JOSHUA JACOBS, M.D., Rush Medical Center, Chicago,


BERNARD N. STULBERG, M.D., Cleveland Center for

   Joint Reconstruction, Cleveland, Ohio






PETER G. ALLEN, M.S., Orthopedics Devices Branch







BARBARA D. BUCH, M.D., Orthopedics Devices Branch

PHYLLIS M. SILVERMAN, M.S., Orthopedics Devices







I.    Call to Order                                5


II.   Open Public Hearing                         16


III.  Industry Presentation                       46


IV.   FDA Presentation                            84


V.    Panel Deliberation and Recommendation      114


VI.   Reclassification Questionnaire and         189

      Supplemental Data Sheet, and Vote




VII.  Open Public Hearing                        247


VIII. Industry Presentation                      248


IX.   FDA Presentation                           269


X.    Panel Deliberation                         300


XI.   Adjournment                                388



                                         (8:01 a.m.)

            DR. SCUDIERO:  Good morning, everyone.

            We're ready to begin this meeting of the

Orthopedic and Rehabilitation Devices Panel.  I'm Jan

Scudiero, the Acting Exec Sec of the panel, while the

Exec Sec is on detail.

            If you haven't already signed in, please

do so.  I'm sure most of you already have.

            I would like to announce that the

tentatively scheduled meetings for this panel are ‑‑

for the year 2004 remaining are August 12th and 13th,

and December 2nd and 3rd.  Please monitor the Center's

web ‑‑ panel website for updated information on this.

            Before I turn the meeting over to Dr.

Yaszemski I'm required to read two statements ‑‑ the

appointment to temporary voting status statement and

the conflict of interest statement.

            Pursuant to the authority granted under

the Medical Devices Advisory Committee charter dated

October 27, 1990, and amended April 20, 1995, I

appoint the following as voting members of the

Orthopedic and Rehabilitation Devices Panel for the

duration of this meeting on June 2nd and 3rd:

Marcus P. Besser, Ph.D., for June 2nd and 3rd.

            Brent A. Blumenstein was deputized for

yesterday as was Fernando G. Diaz and ‑‑ were

deputized for yesterday.  And for today Choll W. Kim,

M.D., Ph.D., Jay D. Mabrey, M.D., and Michael B.

Mayor, for the morning session.

            For the record, these people are special

government employees and are consultants to this panel

or another panel under the Medical Devices Advisory

Committee.  They have undergone the customary conflict

of interest review and have reviewed the material to

be considered at this meeting.

            Daniel G. Schultz, M.D., Acting Director,

Center for Devices and Radiological Health, on

May 28th.

            And now the conflict of interest

statement.  The following announcement addresses

conflict of interest issues associated with this

meeting and is made part of the record to preclude

even the appearance of an impropriety.  To determine

if any conflict existed, the agency reviewed the

submitted agenda for this meeting and for all

financial interests reported by the panel


            The conflict of interest statutes prohibit

special government employees from participating in

matters relating ‑‑ that could affect their or their

employer's financial interests.  However, the agency

has determined that the participation of certain

members and consultants, the need for whose services

outweigh the potential conflict of interest involved,

is in the best interest of the government.

            Therefore, waivers were granted for Drs.

Choll Kim and Jay Mabrey for their interest in firms

that could be affected by the panel's recommendations.

Dr. Kim's waiver entails consulting on a creditor's

unrelated product.  Dr. Mabrey's waiver involves

consulting with an unaffected division of the

sponsor's firm on matters unrelated to today's agenda.

            We would like to note for the record that

the agency took into consideration certain matters

regarding Drs. Maureen Finnegan, Choll Kim, John

Kirkpatrick, and Jay Mabrey. 

            Each of these panelists reported current

or past interest in firms at issue, but in matters not

related to today's agenda.  The agency has determined,

therefore, that they may participate fully in today's


            In the event that the discussions involve

any other products or firms not already on the agenda,

for which an FDA participant has a financial interest,

the participant should excuse himself or herself from

such involvement, and the exclusion will be noted for

the record.

            With respect to all other participants, we

ask in the interest of fairness that all persons

making statements or presentations disclose any

current or previous financial involvement with any

firm whose products they may wish to comment upon.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks, Ms.


            Good morning.  I'm Dr. Michael Yaszemski.

I'm the Chairperson of the Orthopedic and

Rehabilitation Devices Panel.  I'm an orthopedic

surgeon and a chemical engineer at Mayo Clinic in

Rochester, Minnesota.  My areas of interest are spinal

surgery and polymeric biomaterials.

            I'd like to note for the record that the

voting members present constitute a quorum as required

by 21 CFR Part 14. 

            At this meeting, the panel will be making

a recommendation to the Food and Drug Administration

on an OSMA‑initiated reclassification proposal for

mobile bearing knee prostheses.  We will also consider

a draft guidance on performance criteria for hip joint


            Before we begin the meeting, I'd like to

ask our distinguished panel members who are generously

giving their time to help the FDA in the matter being

discussed today, and other FDA staff seated at the

table to introduce themselves. 

            Please state your name, your area of

expertise, your position, your institution, and your

status on the panel, whether a voting member, a

deputized voting member, consumer rep, or industry

rep.  Let's start to my left with Dr. Mayor.

            DR. MAYOR:  Thank you, Mike.  Dr. Michael

Mayor from the Dartmouth Hitchcock Medical Center in

Hanover, New Hampshire.  I'm an orthopedic surgeon and

the co‑director of the Retrieval Laboratory at the

Dartmouth Thayer School of Engineering.  I'm a

consultant to the panel and a voting member.

            CHAIRPERSON YASZEMSKI:  Thank you.  May I

also mention, before we move on, that Dr. Mayor is a

former Chairperson of this panel.

            Dr. Larntz?

            DR. LARNTZ:  Kinley Larntz.  I'm a

statistician, professor emeritus, University of

Minnesota.  And I'm now doing independent consulting,

and I'm a voting member.

            DR. BESSER:  Marcus Besser, associate

professor in the Department of Physical Therapy at

Thomas Jefferson University, but my background and

training was in mechanical engineering and

biomechanics.  I am a deputized voting member.

            MS. MAHER:  Sally Maher, group director,

regulatory and clinical, for Smith & Nephew Endoscopy.

I'm the industry representative.

            DR. WITTEN:  Celia Witten.  I'm the

division director of the Division of FDA that reviews

orthopedic products.

            DR. KIRKPATRICK:  I'm John Kirkpatrick.

I'm an associate professor at the University of

Alabama‑Birmingham in orthopedic surgery, and I have

a special interest in spine surgery.  I am a panel


            DR. MABREY:  Jay Mabrey.  I'm at Baylor

University‑Dallas.  I'm a ‑‑ my area of specialty is

total joint replacement, wear debris, and particle

analysis, and I'm a deputized voting member.

            DR. FINNEGAN:  Maureen Finnegan.  I'm an

associate professor at UT‑Southwestern.  I'm an

orthopedic surgeon, and I'm director of the Orthopedic

Research Laboratory, and I am a voting member.

            DR. KIM:  I'm Choll Kim.  I'm an assistant

professor at the University of California‑San Diego.

My clinical interest is in spine surgery.  I'm the

director of the Spine Research Lab at UCSD.

            DR. NAIDU:  Sanjiv Naidu.  I'm an

associate professor of orthopedic surgery at Penn

State College of Medicine.  My interests are in

orthopedic surgery and material science, and I am a

voting panel member.

            CHAIRPERSON YASZEMSKI:  Thanks, everybody.

            Today the panel will deliberate on and

provide recommendations to FDA on a reclassification

petition for mobile bearing knee joint prostheses and

a draft guidance on performance criteria for hip joint

prostheses.  Both documents were submitted by members

of the Orthopedic Surgical Manufacturers Association.

            In the morning, after the open public

hearing, we'll first deliberate on the

reclassification petition.  Representatives of OSMA

will present, followed by the FDA, then we'll have the

panel deliberation portion of the meeting, beginning

with an introduction of today's topic led by Dr. Mayor

and by Dr. Larntz.

            After having a general discussion, the

panel will address the FDA questions.  Then the ODE

classification/reclassification coordinator ‑‑ Ms.

Shulman ‑‑ will guide the panel on completion of two

forms ‑‑ the reclassification questionnaire and

supplemental worksheet.  The panel's vote on these two

documents will constitute our recommendation to the


            In the afternoon, we'll follow a similar

agenda for the draft guidance document.  This is the

first industry group prepared draft guidance document.

After the open public hearing, representatives of OSMA

will again present, followed by FDA.  In the panel

deliberations, Dr. Mabrey and Dr. Larntz will provide

their perspectives to start the panel deliberations.

There will be no panel vote on this topic.

            Our response to the FDA questions will

constitute our consensus recommendations on the draft

guidance document. 

            We're now going to proceed to the open

public hearing.  We ask at this time that all persons

addressing the panel speak clearly into the microphone

as the transcriptionist is dependent on this means to

provide an accurate record of the meeting.

            I'll apologize ahead of time, if you

forget to this and identify yourself, then I'll ask

you to do so when you come up to speak.

            Ms. Scudiero will now read a statement

prepared for open public hearings.

            DR. SCUDIERO:  Both the FDA and the public

believe in a transparent process for information‑

gathering and decision‑making.  To ensure such

transparency at the open public hearing session of the

Advisory Committee meeting, FDA believes it is

important to understand the context of any

individual's presentation. 

            For this reason, FDA encourages the open

public hearing speaker, at the beginning of your

statement, to advise the panel of any financial

relationship you may have with the sponsor, its

products, and, if known, its direct competitors.  For

example, the financial information may include the

sponsor's payment of your travel, lodging, or other

expenses in connection with your attendance at this


            Likewise, the FDA encourages you at the

beginning of the statement to advise the committee if

you do not have any such financial relationship.  If

you choose not to address this issue of financial

relationships at the beginning of your statement, it

will not preclude you from speaking.

            I would like to note for the record that

the American Academy of Orthopedic Surgeons has sent

a statement to the agency for the record of this

meeting, and it's signed by its President, Roger W. ‑‑

or Dr. Roger W. Bocholz.  He stated that the

association is pleased to express support for the

reclassification petition for mobile bearing knees for

medical device Class III, intermedical device Class


            CHAIRPERSON YASZEMSKI:  Prior to the

meeting, FDA received four requests to speak in the

open public hearing.  We'll start now with these four

people, and I'll identify the amount of time allotted

for each of them.  Just before the meeting started, we

had two additional requests, and we'll add two minutes

for each of those two people to come up, two minutes


            The first speaker will be Dr. Steve

Peoples, scheduled for five minutes.  Dr. Peoples?

Just to help all of the speakers as you're timing your

speech, I'll have the light go from green to yellow

when there's two minutes left.

            DR. PEOPLES:  Good morning.  I'm Steve

Peoples, and I am an employee of DePuy.  Thank you for

the opportunity to provide comments regarding this

reclassification petition.

            You will hear this morning that the

sponsors of this petition believe that the information

supplied provides strong evidence of the safety and

effectiveness of mobile bearing knees and that the

risks associated with them are now adequately defined,

justifying reclassification.

            The petition further proposes that FDA can

regulate these devices adequately under Class II

controls.  You will also hear that the proposed

reclassification meets the least burdensome

requirement.  However, least burdensome does not

preempt the underlying principle that the level of

control must be appropriate to the level of risk posed

by the device.

            The petition offers clinical and

laboratory data as justification for general

reclassification of mobile bearing knees and

identifies almost 50 different mobile bearing designs

as representing the spectrum of mobile bearing knees.

            Basically, the proposed reclassification

would move mobile bearing knees from the current

requirement for valid scientific evidence of safety

and effectiveness provided via the pre‑market approval

process to the level of Class II controls and

substantial equivalency under the Section 510(k) pre‑

market notification process.

            We believe that the petition fails to

justify a general reclassification and that an

examination of the information upon which the proposal

was based reveals why.  And that reason is that the

vast majority of the mobile bearing clinical

literature and information presented as justification

for reclassification is in regard to a single total

knee system and a single unicompartmental device.

            For example, 86 percent of the total knee

survivorship literature cited in the petition is on

the LCS mobile bearing knee system.  Only two

survivorship articles on two other total knee designs

are included, one of which ‑‑ the Accord knee ‑‑

deserves special comment. 

            Although the Accord design underwent

extensive preclinical laboratory and finite element

analyses, that testing did not predict the almost

50 percent failure rate encountered in actual clinical

use.  Similar observations can be made in regard to

the clinical results reviewed. 

            The petition presents very limited data on

a limited number of total knee designs to substantiate

the safety and effectiveness of mobile bearing knees

as a generic type of device.  And even this limited

data indicates that there is a very large variation in

revision rate from design to design.

            The clinical outcomes information provided

is no different.  And although IDE data on six total

knee mobile bearing designs is included in the review,

the vast majority of that data is a very short

followup and very small populations.

            The petition clearly demonstrates what is

known today about mobile bearing knees, and in doing

so also demonstrates what is not known.  Absolutely no

clinical data is presented for over 60 percent of the

mobile bearing knee designs identified in the

petition, designs which presumably would be covered by

the proposed reclassification.

            Regulations require that a proposed

reclassification describe how the new classification

will provide reasonable assurance of safety and

effectiveness.  The petition proposes that the

controls already established for Class II fixed

bearing knees are sufficient.

            However, most of the recommended special

controls are only standard or unvalidated non‑standard

test methods.  The petition offers no performance

criteria for these tests and provides no guidance on

the predicate control to be used, other than that it

be clinically successful, which the petition also

leaves undefined.

            No recommendations are made to

specifically address significant polyethylene

performance issues, such as the effects of multi‑

directional movement and cross‑shear, or knee

stability, which was the unpredicted mode of failure

of the Accord knee.

            Both of these issues are unique to mobile

bearing design and cannot be evaluated or controlled

using methods employed for fixed bearing knees. 

            The significance of mobile bearing knee

kinematics and polyethylene wear, in relation to

mobile bearing design, was reported on at this year's

Orthopedic Research Society meeting.  The authors

concluded, and I quote, "This study shows that minute

differences in mobile bearing prostheses may have a

major affect on their wear behavior."

            The petition under consideration is

thorough, and it does employ sophisticated analytical

techniques.  It is deep but very narrow.  It is

essentially a review of the results of a single total

knee system and a single unicompartmental device that

the petitioners claim represents the safety and

effectiveness of mobile bearing knee designs in

general.  We do not believe that such a generalization

is valid.

            Your requirements for reclassification

that you must consider and answer today are:  does the

petition provide adequate and valid scientific

evidence that mobile bearing knees in general are safe

and effective and thus can be reclassified to

Class II?  And does the petition identify the special

controls for Class II necessary to assure the safety

and effectiveness of these devices?

            We do not believe that the petition meets

either of these requirements, and based on the limited

evidence provided in the petition and the significant

risk involved, general reclassification of mobile

bearing knees simply is just not justified.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr.


            The next speaker will be Dr. John Fisher,

also scheduled for five minutes.  Dr. Fisher?

            DR. FISHER:  Good morning.  My name is

John Fisher.  I'm director of the Institute of Medical

and Biological Engineering at the University of Leeds.

I have 15 years' experience in wear testing of

artificial joints and run an academic laboratory with

over 100 stations of wear simulation capacity for

joint replacement.

            Our work in the laboratory is supported by

government, by a range of different companies,

including DePuy, and DePuy is supporting my attendance

at this meeting.

            CHAIRPERSON YASZEMSKI:  Thank you.

            DR. FISHER:  The LCS rotating platform

mobile bearing knee is very special and unique.  The

bearing design decouples motions and allows rotation

at the tibial tray, which then predominantly allows

linear motion at the femoral interface.  Both these

motions are unidirectional, which has a substantial

reduction in polyethylene wear.

            However, not all mobile bearings are the

same.  Unconstrained bearings have multi‑directional

motion, and, therefore, higher wear.  That has been

shown by RADCA at two presentations this year and has

been confirmed in our own laboratory.

            So wear is dependent on interfaced

kinematics and is design‑specific.  So are kinematics

the same, in fact, to the mobile bearing knees?  Well,

the clinical studies show the overall kinematics of

the whole joint, but what we must be concerned about

is the kinematics at the individual interfaces.

            Mobile bearings are complex systems.

Motion ‑‑ individual wear interfaces ‑‑ is design‑

dependent, and, therefore, cannot be predicted from

whole joint kinematics.  Small changes in interface

kinematics can have a major effect on the wear in

mobile bearing knees.

            Do lower contact stresses in mobile

bearing knees reduce wear?  Lower contact stresses

certainly reduce the lamination fatigue failure, but

there is no increasing evidence that lower contact

stresses and larger wear areas actually increase

surface wear and micro and macro wear debris


            However, this is not the case in rotating

platform designs, which have unidirection of motion

and, therefore, much lower wear than fixed bearing

knees.  So, again, the effects of contact stress on

wear is design‑dependent in mobile bearing knees. 

            Now let me turn to third body damage and

wear debris.  Mobile bearing knees are more prone to

damage and destruction by third body damage,

particularly on the tibial counterface.  Particles get

trapped in that counterface and remain there for a

substantial period of time. 

            However, the wear that is produced by

third body damage is, again, design‑dependent, as

linear scratches in rotating platform bearing knees do

not accelerate wear, whereas they would do in multi‑

directional designs.

            And what about wear debris?  It is really

a very important issue.  Is wear debris from mobile

bearing knees more reactive?  There are significant

studies now in the laboratories and clinically that

shows that wear debris from fixed bearing knees is

larger and less reactive than debris from hips.

            It has been speculated that mobile bearing

knee debris is more like hip debris.  This may well be

the case in multi‑directional designs due to cross‑

shear as found in the head will produce fragmentation

of fibrils and smaller particles. 

            However, we've shown that with

unidirectional motion the debris that is produced in

the polymer bearing is actually larger and less

reactive.  So, once again, the reactivity of the

debris produced in mobile bearing knees will be


            Can we effectively determine wear?  We all

know there are two separate standards for knee joint

simulators at the moment ‑‑ force control and

displacement control machines.  And these have

produced different results in comparison to mobile

bearing and fixed bearing knees.

            In our own laboratories, we developed a

special methodology for rotating platform design,

which allowed a combination of both force and

displacement control testing for the rotating platform

mobile bearing knee.  This is not an ideal test

methodology, and it is not currently available in

other joint simulation systems.

            Secondly, can wear be determined from

clinical measurements?  It is not easy.  Penetration

can be measured as reported in the knee study, but, of

course, volume depends on penetration and wear areas,

and these are difficult to determine and compare.

            The inherent nature of mobile bearing

knees introduces greater variability.  Any bearing has

six degrees of freedom.  Mobile bearing knees have 12

degrees of freedom.  Meniscal bearings have 24 degrees

of freedom at four different interfaces.  Meniscal

bearings also typically have smaller polyethylene

inserts, which are more prone to edge loading and,

therefore, cracking and fragmentation.

            For the rotating platform LCS mobile

bearing knee, there is over 20 years' experience.

There are some serendipitous design features that have

resulted in low contact stress, low fatigue,

unidirectional interface motion, and a stable low‑

wearing bearing. 

            Many of these design features have not

been replicated in other mobile bearing knee designs.

Not all mobile bearing designs are the same, and we do

not understand the result and impact of the numerous

design variables encountered in different mobile

bearing knees.

            To conclude, the reclassification petition

does not consider or address the effect of the

aforementioned design variables on the performance of

mobile bearing knees, and the special controls

proposed by the petition cannot assure us that the

various designs of mobile bearing knees are both safe

and effective.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thank you,

Professor Fisher.

            Our next speaker will be Dr. Doug Dennis.

Dr. Dennis?  Dr. Dennis is scheduled also for five

minutes speaking time.

            DR. DENNIS:  I'm Dr. Douglas Dennis from

Denver.  I serve as an adjunct professor in the

Department of Biomedical Engineering at the University

of Tennessee, medically direct the Rocky Mountain

Musculoskeletal Research Laboratory.  I do serve as a

consultant for DePuy.  They have provided my travel

expenses here. 

            Over the last 10 years, my laboratory has

received orthopedic industry support from many

different companies.  It's a privilege to present.

            As I have reviewed the reclassification

petition, it is to reclassify all mobile bearing total

knees from Class III to Class II devices.  Therefore,

I think this assumes that all available mobile bearing

designs will demonstrate similar efficacy and safety

as those designs that have been evaluated by pre‑

market IDE studies.

            There are many fears associated with

mobile bearing knees.  These include the potential for

increased polyethylene wear, as you now have two

articulating surfaces on both the top and bottom side

of the bearing.  It requires a more demanding surgical

technique.  It is less tolerable of instability, as

demonstrated by Dr. Jack Burt and a 9.3 percent

bearing subluxation or spinout rate.

            There are also fears about increased wear

that are created from tibial tray post bumpers,

etcetera, which try to control the boundaries of

bearing mobility.  Therefore, all mobile bearings are

not the same.

            There are numerous differences.  Knee

kinematic patterns have shown various kinematic

differences on both the top and bottom side of the

bearing.  There are geometry differences of both

femoral and tibial components, and the bearing

stabilizing mechanisms of these various designs are


            My concerns about the petition primarily

are centered on the bottom side of the bearing as

underside motion pattern differences among differing

designs can be quite substantial, and, therefore,

create a potential for premature polyethylene wear and

periprosthetic osteolysis from increased undersurface

wear versus currently approved designs.

            Another concern has already been mentioned

about the size of the microparticulate polyethylene

debris in mobile bearings.  It is more similar to hip

replacement in that the particles are smaller, more

reactive.  Will this result in more osculysis,

particularly in designs which permit multi‑directional

motion on the underside of the bearing?

            If we look at some of these design issues,

as has been stated, in a rotating platform design you

have unidirectional motion patterns on the

undersurface of the bearing.  While there are many

designs that allow both rotation as well as antero‑

posterior translation, these will, therefore, have

multi‑directional motion patterns on the bottom side

of the bearing.

            In the 1970s, Pooley & Tabor showed us,

when dealing with polyethylene, if this material is

exposed to unidirectional motion, the molecules align,

decreasing the coefficient of friction and the wear

rates, whereas with multi‑directional motion patterns

you increase the shear forces and wear. 

            And this has been shown in multiple

studies since then, both by Wang and Marrs, which have

shown increased wear with multi‑directional motion and

reduced wear when you have mono‑directional motion


            Our previous speaker, in an elaborate

laboratory analysis studying undersurface wear, has

shown that unidirectional type of patterns

demonstrated .23 millimeters cubed per million cycles.

When exposed to multi‑directional wear patterns,

nearly a 10‑fold increase in wear.

            The last 10 years of my laboratory life

has been studying knee kinematics.  We have done over

75 different knee kinematic studies, over 40 fixed and

mobile bearing designs.  And to summate this 10 years

of work, we have found that knee kinematics vary

widely among differing fixed and mobile bearing

designs, and numerous adverse kinematic patterns have

been identified, which can adversely affect

polyethylene wear.

            This has been shown in the study

previously mentioned this year at the ORS where minor

differences in kinematics have resulted in major

differences in wear, particularly when dealing with

unconstrained tibial bearings that have multi‑

directional wear patterns.  They had double the wear

rates versus those with unidirectional wear.

            So, in summary, all mobile bearings are

not the same.  I do think the petition clearly

demonstrates the safety and efficacy of a single

rotating platform design that demonstrates

unidirectional underside motion.  It has been shown

that mobile bearing kinematics vary widely, which can

affect polyethylene wear rates, inherent implant

stability such that precise surgical technique is


            There are numerous laboratory studies that

suggest the potential of substantial increases in

polyethylene on the bottom side of the bearing if

multi‑directional motion patterns are present, yet no

good long‑term clinical data is available to document

the safety of this concept.

            I think that it is wise for the committee

to proceed with caution in grouping all mobile bearing

total knees as Class II devices as only the rotating

platform concept has proven clinically successful.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks very much,

Dr. Dennis.

            Our next speaker is going to be Dr.

Ruddlesdin.  As Dr. Ruddlesdin is coming up, I'd like

to note that we've been joined by another

distinguished panel member. 

            Dr. Doyle has kindly agreed on short

notice to serve as our consumer representative today,

and I'd like to take the opportunity and ask her to

introduce herself.  Ma'am, give your institution, your

area of expertise, please.  And welcome.

            DR. DOYLE:  I'm LeeLee Doyle.  I am the

assistant dean for faculty development and a professor

emeritus of OB/GYN and currently a professor of

maternal and child health at the University of

Arkansas for Medical Sciences College of Medicine.

            I have a very personal interest in this

particular area, because I already have two hips that

have been replaced, and I'm looking forward to two

possible knee replacements.  So I have spent a great

deal of time considering these things.

            CHAIRPERSON YASZEMSKI:  Thank you, and


            Dr. Ruddlesdin, you're scheduled for 10

minutes, and please begin.

            DR. RUDDLESDIN:  Thank you, Chairman.

It's a privilege to address this panel this morning.

            My name is Cris Ruddlesdin.  I'm a full‑

time orthopedic surgeon working with the National

Health Service in the United Kingdom.  I receive no

remuneration from Corin, apart from my travel expenses

to come here today.  My NHS organization has been

using Corin products from 1989, and we purchase these

products on a competitive tendering basis.

            I think it is important to stress that

knee replacement is trying to replace a normal knee,

and for many, many years fixed bearing knees have

relied on either a hinge or a rotating device.  But

they're trying to reproduce multi‑directional knee

movement at one bearing surface.

            The mechanics have been covered in some

detail by my colleagues, and I will not go into any

further detail beyond that.

            The rotor‑glide knee, as marketed by

Corin, first started life in 1977 as an experimental

design.  It was first implanted in 1988 in clinical

patients.  Over 20,000 of these joints have been

implanted worldwide to date, and the joint relies on

a polyethylene bearing, which both rotates and glides

anteriorly and posteriorly on the tibial tray.

            I have a knee implant here, if the panel

wishes to see it.  And if I can pass it around ‑‑

            CHAIRPERSON YASZEMSKI:  Thank you, sir.

We'd like to do that.

            DR. RUDDLESDIN:  ‑‑ as I talk, you can see

how the implant works.

            The tibula tray is a metal base plate with

two metal bullards, and this is a cross‑section

through the knee.  And the white is the polyethylene.

The polyethylene can rotate 12.5 degrees either side

of the midline and can glide backwards and forwards by

five millimeters.  And, therefore, we are reproducing

the rotation of the knee as it achieves full


            You will notice as the implant comes

around that it is fully congruous through 90 degrees

of flexion from full extension to 90 degrees of

flexion at the bearing surface between the femur and

the polyethylene insert.  And this means that we have

very low point loading in that situation.

            My personal experience over the last four

years ‑‑ I have implanted 119 joints in four years.

Prior to 2000, I had used the fixed bearing version of

the same knee marketed in the United Kingdom as the

Nufield knee, and subsequently we changed to the

mobile bearing knee, mainly because one of my newly‑

appointed colleagues got considerable experience in

this implant in Scotland.  The only difference is that

the insert locks onto the pole shroud and being

constrained by them.

            The average followup in my series is two

years.  I have one gross dislocation, which I do not

blame on the implant.  It was an overenthusiastic

surgical soft tissue release on my part, with a very

deformed knee, and this allowed the femoral component

to dislocate from the top of the polyethylene.

            It is interesting that despite the fact

that his knee was grossly unstable that the

polyethylene bearing remained contained on his tibial


            In a report in the Journal of Arthroplasty

in 1996, from the Solihol Group, they had 161 patients

with 171 knees, and this shows that the knee has been

in clinical use for the last 17 years.  They had an

average followup of 3.1 years, which is relatively

short in orthopedic terms, because we don't expect to

see problems, even in the worst design of joint, until

at least five years.

            They had a series of complications, as you

would expect of any major joint replacement, but they

had no dislocations or spinouts of the inserts.

            Further work from Stuart Brooks at Solihol

reviewed 136 patients between 1989 and 1991, and this

had a much longer followup ‑‑ minimum followup of

seven years and a maximum followup of nine years.  Two

revisions for one loose femoral component and one

fractured insert in a knee which had been implanted in

extreme varus ‑‑ the other complications, again, as

expected in any major joint replacement.

            Clinical data from Hayward's Heath

Hospital in United Kingdom shows a very large series

‑‑ nearly 900 patients, who have been operated on in

the last 10 years with followup for five years or more

‑‑ 13 deep infections, eight aseptic loosenings, and

two gross dislocations, which yields a very low

percentage of 0.2 percent.

            Since 1988, this knee has been implanted

using exactly the same instrumentation used in fixed

bearing designs.  So the discussion that these are

different types of knees is not valid.

            Ligament balancing is equally important in

any knee, be it a fixed bearing or a mobile bearing

knee.  A badly performed operation, whether it be

fixed or mobile, will fail, and I would put it to the

panel that a well prepared knee will survive whatever

the bearing type.

            I have converted from fixed bearing to

mobile bearing, and I have found no difference in the

level of difficulty to implant either a fixed or a

mobile bearing knee.  The Corin instrumentation is

extremely simple and straightforward.  I think levels

of difficulty often relate to the instrumentation as

designed by the manufacturer.

            A lot of discussion about the bone clips

and ligament balancing.  These obviously play a role.

And as I've said, any knee which is ‑‑ any knee

surgery which is well performed will do well.  Any

badly performed knee surgery is likely to fail.  And

this is equally important, be they fixed or mobile


            And at the end of the day, it's the

surgeon's technical skill and expertise, and

particularly his feel for the operation and his

experience, that will be the indicator of success.  No

instrumentation can compensate for poor surgical


            The earlier instrumentation involved

tensing devices to try and balance the ligament on the

line of the Freeman knee.  But subsequently

instrumentation has been simplified to what I can only

describe as a lollipop.  This is a spacer that is

inserted into the knee in both flexion and extension

to ensure that the ligaments are balanced in both

flexion and extension, as this is vitally important.

            If the cuts are not correct ‑‑ and, in my

experience, they almost always are correct ‑‑ then, if

they are not correct, then further adjustment to the

bone cuts at this stage is essential.  Unless people

are overly enthusiastic at resecting one or other of

the bone surfaces, either the proximal tibia or the

distal femur, then the bone cuts usually come out


            But it is vitally important to check the

gaps after the cuts have been made.  And as I've said,

if the cuts are not correct, then it is important to

correct them at this stage.  And this really does not

need to be different between the fixed and mobile

bearing knees, as I keep emphasizing.

            So the take‑home points basically are

insert dislocations are not a clinical issue, and

gross dislocations are not happening at a significant

rate ‑‑ less than one percent.  And this is contrary

to some of the evidence that has been produced.

            There is no difference in the operative

technique unless it is imposed upon or desired by the

surgeon.  In other words, a fixed bearing and a mobile

bearing knee require the same level of surgical skill,

and, in the case of the Corin, exactly the same

instrumentation are used.

            And whether fixed or mobile bearings, the

final arbiter to a good fit is the surgeon's tactile

field during the trial reduction.

            Thank you very much.

            CHAIRPERSON YASZEMSKI:  Thank you very

much, Dr. Ruddlesdin.

            Is there anyone else who would like to

address the panel at this time?  If so, please

identify yourself, be recognized, and come forward.


            DR. FERRING:  My name is Tom Ferring.  I'm

an orthopedic surgeon, Charlotte, North Carolina.  I'm

adjunct professor of mechanical engineering at

University of North Carolina‑Charlotte.  And I'm a

consultant for DePuy, who paid my way here.

            I have a special interest in osteolysis.

I've been published on this subject.  And I'm

concerned about the potential for a significant

difference in wear debris with the new designs of

mobile bearings.  I am also concerned about the long‑

term effects that small changes in knee design will


            One only has to look at the effects of

moving a posterior stabilized post a few millimeters

in a fixed bearing knee, which led to instability of

certain designs that have been marketed or the

addition of a posterior stabilized post in the LCSPS,

which is a mobile bearing knee that led to early


            Therefore, it is counterintuitive to me to

conclude that a variety of new mobile bearings with

back side stops and different pivot points are

equivalent to proven designs.  As one of the speakers

earlier said, in orthopedics we rarely see failures

before five years.  I would urge you to recommend that

new mobile bearing implants be placed through standard

scientific methods to prove their efficacy and safety

for our patients.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks very much,

Dr. Ferring.

            Would other people like to speak?  Sir?

            DR. SORRELS:  Good morning.

            CHAIRPERSON YASZEMSKI:  Good morning.

            DR. SORRELS:  I am Barry Sorrels from

Little Rock, Arkansas.  I thank DePuy for inviting me

to this meeting.

            I had the pleasure of serving as one of

the original clinical investigators with the LCS in

the beginning of 1980.  In 1980, there were over 300

knee designs on the market, and this was the first and

only knee to undergo an FDA evaluation.

            The reason for that is it's more

complicated and it's fraught with potential

complications for the patient.  My interest in this ‑‑

I have never been funded by the company nor paid for

the investigation.  But my interest is on the part of

the patient.

            Our patients were served very well with

the IDE.  There were 15 of us.  We collaborated

frequently, we met, we discussed complications,

discussed instrumentation, and ultimately I think the

patient was much better served as a result of going

through this process.

            When I look back on my clinical results of

over 3,000 cases, I realize that 50 percent of my

complications were technical error committed by me,

and half of those were in the first three years of

this procedure. 

            There's a steep learning curve with these

mobile bearing knees, and I think if we can

collaborate, if we can meet together, if we can

exchange our problems and our ideas, I think

ultimately the patient is much better served.  And I

think that there's a real purpose served for the

patient by this process.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks very much,

Dr. Sorrels.

            Would someone else like to speak?  Sir?

            DR. FITZPATRICK:  Good morning.  My name

is David Fitzpatrick.  I'm a bioengineer.  I'm a

faculty member at the Department of Mechanical

Engineering in University College‑Dublin.  I have

research interests in knee joint kinematics modeling

and in vitro analysis of implant design. 

            DePuy has financially supported my costs

associated with attending this meeting, and I would

just like to concentrate on the issue of special

controls relating to these mobile bearing knee joint


            It is clear that total knee operations are

very dependent on the management of the soft tissues,

and the level of success is directly related to the

operative technique and surgeon experience.  And when

compared with fixed bearing knee designs, the reduced

level of constraint within mobile bearing knees places

a higher level of demand on the surgeon and the

operative technique.

            Post‑operative joint stability is a

critical factor in clinical success, and the

sensitivity of various mobile knee designs to

operative technique is highly variable.  The clinical

history of mobile bearing knees has shown that device

redesign or revised clinical indications is a common

outcome following initial clinical experience.

            I would propose that the existing

preclinical tools, such as those proposed within the

special controls in the petition, do not have

sufficient capabilities to assess the kinematic

performance or the ability to predict the clinical

performance of mobile bearing knees in use.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks very much,

Dr. Fitzpatrick.

            Would someone else like to make a


            Seeing no other speakers, I'd like to

suggest that we proceed now to the OSMA presentation.

The OSMA representatives will now present on their

proposed reclassification position for the total and

the unicompartmental mobile bearing knee joint


            I'd like to remind public attendees now

that the meeting is still open for public observation,

but public attendees may not participate except at the

specific request of the panel.

            The first OSMA presenter is Dr. Toni

Kingsley.  Dr. Kingsley?

            DR. KINGSLEY:  Good morning.  I am Toni

Kingsley, and I am representing OSMA, the Orthopedic

Surgical Manufacturers Association.  OSMA is an

organization of manufacturers of both medical devices

and biological products used in the treatment of

orthopedic pathologies.

            OSMA has sponsored a number of

reclassification petitions in recent years for the

purpose of bringing safe and effective medical devices

to the public through the least burdensome regulatory

path possible.  We are here this morning as the

sponsor of the petition to request the

reclassification of mobile bearing knees ‑‑ MBKs ‑‑

from Class III to Class II.

            The petition requests reclassification of

mobile bearing total knees and unicompartmental knees,

both cemented and uncemented.  Included within these

general categories are a number of subcategories

listed on this slide.  Any given mobile bearing knee

on the market today will exemplify several of these


            At FDA's request, OSMA considered each of

these subcategories in evaluating the risks and

special controls specific to mobile bearing knees.

            MBKs have been on the market for nearly 25

years.  The first mobile bearing knee approved in the

United States was J&J DePuy's rotating platform LCS

knee approved in 1985.  Since that time, there has

been considerable technical progress leading to

development of second and third generation devices.

            A review of the devices on the market

today reveals that approximately 46 mobile bearing

knee designs are available around the world.  In the

United States, there are six designs currently

approved by the FDA, including five marketed by J&J

DePuy and one recently approved meniscal bearing

unicompartmental device marketed by Biomed.

            It should be noted that FDA previously, in

January 1998, considered the reclassification of

mobile bearing knees as part of the petition to

reclassify uncemented porous knees.  At that time, the

panel recommended reclassification only of those MBKs

that were tricompartmental, cemented, and had a

rotating or translating base.

            However, FDA subsequently chose to

recommend submission of a new reclassification

petition for the entire class of mobile bearing knees

rather than reclassify specific subcategories. 

            In the six years since these decisions,

considerable additional information on mobile bearing

knees has become available.  OSMA will present today

summaries of clinical data, including seven ongoing or

completed IDE studies, two international clinical

outcome studies, and published data on approximately

14 knee designs, most with two‑year or longer


            OSMA will also present information on

risks specific to mobile bearing knees, with a focus

on the issues of wear, including back side wear, and

the potential for bearing dislocation.  We will

discuss special controls, including existing FDA

guidance and standard test protocols.  Where no

standard test protocol exists, we will present

suggestions to be developed further for inclusion in

an FDA guidance document.

            We believe these special controls will be

adequate to enable the FDA to regulate mobile bearing

knees as Class II devices.  OSMA will also present a

summary of two meta‑analyses on published data that

support the claim that the clinical performance of

mobile bearing knees is not different from fixed

bearing knees, and that survivorship of the various

subcategories of mobile bearing knees are similar to

one another.

            OSMA believes that the information

presented will establish that mobile bearing knees of

all the subcategories included in the petition ‑‑ for

all of these there is well documented, successful

clinical history, design requirements are well

understood, associated risks are well defined, and

special controls either already exist or can be

developed for inclusion in FDA guidelines.

            Because of these factors, OSMA believes

that mobile bearing knees should be reclassified to

Class II, since the special controls either currently

available or to be developed are sufficient to provide

the required reasonable assurance of safety and


            Presenting for OSMA today will be Dr.

James B. Stiehl, clinical associate professor of

orthopedic surgery, Medical College of Wisconsin, who

will discuss the clinical data.  Dr. Peter S. Walker,

director of minimally‑invasive surgery, laboratory,

and professor of orthopedic surgery, at the New York

University Medical Center, will discuss risks and

special controls.  And the meta‑analyses will be

discussed by Greg Maislin, principal biostatistician

of Biomedical Statistical Consulting.

            In addition, there are a number of

representatives of contributing OSMA companies present

together with expert surgeons who are expert in the

specific knee designs manufactured by those companies.

            Dr. Stiehl?

            CHAIRPERSON YASZEMSKI:  Tanks very much,

Dr. Kingsley.

            Dr. Stiehl, welcome.

            DR. STIEHL:  Thanks.

            Dr. Jim Stiehl.  I'm from Milwaukee,

Wisconsin.  I am a private surgeon in orthopedic

practice.  I have worked for a number of years in the

area of orthopedic biomechanics.  I have over 10

years' experience with mobile bearing total knee

arthroplasty, specifically the LCS.  And I continue to

have a very active practice with that particular


            My financial interest ‑‑ I am a consultant

for Zimmer, not specifically with mobile bearing total

knee arthroplasty.  I currently do not have any vested

interest in any mobile bearing knee implant.

            The information that will be presented in

this petition is extensive.  There is a comprehensive

literature review and summary, both of unpublished and

published clinical studies, and I would point out that

the unpublished studies currently come from the OSMA

companies that are supporting this petition. 

            They include IDE clinical trials that are

under current process of investigation and two

important international clinical outcome studies that

have been compiled and are currently ongoing.  We do

a comprehensive literature review from 1977 until

2002, which includes clinical outcome scores and


            This slide summarizes the current IDE

studies that are ongoing.  There are seven different

series of studies.  The Oxford Meniscal Bearing Uni is

currently approved through the PMA process, and that

IDE study is completed.  The remaining devices, as you

can see ‑‑ the Genesis, Profix, and Scorpio MBK and

NexGen ‑‑ are mobile bearing devices.

            And I would state that the ‑‑ there are a

couple of devices in this group ‑‑ the Genesis,

Profix, and Scorpio ‑‑ that are fixed bearing devices

that have been adapted for mobile bearing application.

            The revision rate is being looked very

carefully at for these devices.  As you can see, the

Oxford, which is a completed study, had a 6.8 percent

revision rate.  After approximately four years'

followup, the numbers towards the bottom of the other

devices have relatively short followup to date.  But

as you can see, the clinical outcome has been very

successful to date with nil or very few revisions.

            There are two international outcome

studies that are currently being collected by Zimmer.

These specific studies are with mobile bearing

devices.  The European and Asian and other groups

around the world do not necessarily require the

control that the U.S. FDA requires, so it's able ‑‑

we're able to get these studies done.  These are open

enrollment studies, open surgeon participation.  They

literally come from around the world.

            We believe these studies to be very

important, because their results are generalizable,

and they really measure the performance in the hands

of general orthopedic surgeons who have new experience

with a new device.

            The MBK is the first study that we would

cite.  There are 1,254 cases collected at the point of

data collection from 22 surgeons from seven countries.

To date there have been eight revisions.  The overall

revision rate with this particular device has been

.6 percent through an average of two‑year followup.

            The NexGen LPS flex mobile knee, again,

has a similar study that's ongoing, and at the point

of data collection there were 390 cases from 19

international surgeons from 17 centers around the

world.  In this particular study, there were two

revisions, one for infection and one for instability.

The overall revision rate, again, was .5 percent

through two years of followup.

            We have also collected data of the

outcomes, and I think this is an important issue

because most clinical series have global knee rating

scores that are evaluated, and they will be discussed

in the statistics section where we look at the good to

excellent results.  And as you can see for those of

these, the devices at two years ‑‑ that ranges from 85

to 100 percent.  There have been no red flags with

either of these devices.

            We then look at the literature that is

collected from 1977 until 2002.  There are 274

articles available that talk about mobile bearing

devices in some form or another.  We specifically look

at the papers that have clinical results.  There were

57 such articles; 48 summarized clinical outcome with

data.  And this included the broad spectrum of devices

available ‑‑ multi‑directional rotating platform,

meniscal bearing, and unicondylar.  And there are nine

review articles.

            Summarizing this data, the multi‑

directional platform devices ‑‑ and, basically, this

is a polyethylene bearing that allows both rotation

and anterior motion in the transverse plane.  The

number that I highlight is 91.7 percent with 5.6‑year

followup, and that excludes the results of the Johnson


            The Johnson Accord was an early device.

It comes from the '70s.  It was abandoned, really,

quite a number of years ago because of significant

design problems, and really doesn't meet current

criteria and from the knowledge that we have.  So I

think it was appropriate to remove that device from

this conclusion.

            The rotating platform, which really is the

hallmark of certainly the LCS experience, has been a

very, very durable implant.  And as you can see, of

the four clinical trials that we cite, 96.5 percent

survivorship at 9.3 years.  The meniscal bearing is a

rotating bearing that has two bearings that slide in

tracks.  This implant has a long experience from the

LCS, really nearly 25 plus years of experience.  In

the eight clinical trials that we cite, 97.4 percent

survivorship at 8.2‑year followup. 

            And then we have a large group of

unicondylar knees that are reported from 21 series

that offer 92.1 percent survivorship at 8.8 years.

            One of the articles in our review is from

Dr. Callahan, and in that article they cite a number

of series.  The one series from Murray quotes 144

Oxford medial unis with 98 percent survival at 10

years.  He was a designer involved with that series.

            But we also have the experience of Price,

et al.  This study was done in Sweden by private

practitioners ‑‑ 378 Oxford medial unis with 95

percent survivorship at 10 years.  We have the

experience of Dr. Jordan, et all, with 473 LCS

meniscal bearing ‑‑ again, a private surgeon, non‑

designer, 94.6 percent survival at eight years. 

            And we have the extensive experience of

Dr. Sorrels, who you just heard from a moment ago.  He

reported 665 LCS rotating platform, cemented version,

with 95 percent survivorship at 11 years; and 119 LCS

rotating platform, uncemented, with 100 percent

survival at 12 years.

            There is a very large study that comes

from the Norwegian Arthroplasty Registry.  This had

7,174 total knees, of which 982 were the LCS with a

small addition of interacts 23 cases.  This group

showed slightly better results with the mobile bearing

as opposed to the general fixed bearing group ‑‑ 97.2

percent survivorship after five years. 

            However, they concluded that there was no

statistical differences between mobile and fixed

bearing designs pertaining to either survivorship or


            I have published a number of studies on

mobile bearings, and I have one review article where

I review the literature on this subject.  From all of

the studies that I have looked at, the mobile bearing

revision rate approximates one percent per year.  This

is analogous to the fixed bearing rates of failure

that are described virtually across the literature.

            The surgical technique is similar to fixed

bearing knees.  And despite the unique elements of the

LCS technique that we all learned and developed years

ago, most current mobile bearing knees are falling

along surgical technique which is similar to other

fixed bearing knees.  Bearing‑related complications

with the LCS are .5 percent overall for rotating

platforms, and 2.5 percent for meniscal bearings.

            Most of us believe that the LCS rotating

platform is a very safe design.  And our technique, as

Dr. Sorrels mentioned, relates specifically to this


            I think the most important issue of mobile

bearings ‑‑ and this is the reason why we are here

today, this is the reason why I believe these devices

should be available in general for our patients ‑‑ is

that the current experience looking at osteolysis and

the complications with these devices are minimal.  It

really is an advance in orthopedic surgery to remove

osteolysis, because this is the primary cause, in my

view, of late failure of total knees.

            Vertullo has outlined the potential

advantage of mobile bearings, and I think this really

carefully gives us some idea of what this is all

about.  Axial rotation decreases loosening due to

axial torque.  This was certainly one of the early

design parameters that we felt important.  It has been

improved over the years. 

            Actual rotation may account for self‑

correction of some of the tibial component

malrotation.  Tibial component malrotation is a

significant issue in current technique of any total

knee.  And if you've got a device that gives you some

freedom with that particular issue, it is a help ‑‑

and we believe that these devices reduce contact

stresses, which is important to wear.

            I don't think you have to have all three

of these design advantages in a specific mobile

bearing knee to make it optimal.  I think you can have

a mobile bearing that gives one or two of these and

certainly makes it worth the trouble.  But these are

the advantage of a mobile bearing design, and I

believe that's why these devices are so important.

            This author concludes by stating

hypothetically longer term followup of mobile bearing

knee arthroplasty results may reveal a significant

difference from fixed bearing total knee as a fatigue

failure threshold of incongruent polyethylene is

exceeded.  Really, this is the subject.  This is why

we believe these devices are so good.

            In summary, the literature suggests that

mobile bearing knee devices performed similarly to

well designed fixed bearing knees in terms of

survivorship and clinical function.  Current IDE and

international outcome studies suggest that other

mobile bearing knee designs are clinically successful

and comparable to fixed bearing designs.

            The potential benefit of this technology

is improved long‑term clinical performance and

longevity for our patients.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks very much,

Dr. Stiehl.

            The next presenter will be Dr. Walker.

            DR. WALKER:  Good morning, panel members.

I am pleased to address this issue on behalf of OSMA,

and they are paying my expenses.  And I ought to

declare that as a knee designer for many years I do

receive financial benefit from two companies for knee

designs, notably Stryker and Zimmer.

            I'm a biomechanical engineer.  I've worked

on the biomechanics of the knee and knee design and

knee testing methods since 1966.  I've worked at

notable institutions in the UK and in the USA.

            I'm addressing risks and special controls.

In other words, the purpose of tests is to ensure a

device is safe and effective, so I will talk about

what kinds of tests are appropriate for mobile bearing

knees.  And I will also reference fixed bearing knees,

which are already approved devices.

            The risks are divided into known and

potential.  Now, known mobile bearing design risks

have been elucidated in the previous talks, but also

in the medical device registry, which reports

significant problems, and there have been reports.

But of tens of thousands of knees, they have received

385 reports, of which a small number are to do with

the mobile bearing metal‑poly separation and


            Now, further known risks with mobile

bearing knees include dislocation and subluxation

because the bearing is mobile and not fixed.  And

also, another known potential risk is, of course, wear

because of the extra bearing surface in the design.

            Other potential mobile bearing risks with

the different kinds of designs have been studied by

OSMA by looking at all of the available 46 designs,

looking at them and grouping them, as was shown in the

very first talk, and then identifying potential risks

and mechanical problems with these different designs.

            So let me just summarize these different

risks.  The first group of risks has to do with fixed

and mobile bearing knees in common.  In other words,

these are risks which are well known and fixed

designs, as well as mobile bearing designs.  The

second group includes risks and special controls that

have specific considerations when applied to mobile

bearing knees.  And I will focus my talk on those.

            But, however, I will just mention the

first group.  These are risks and special controls

common to fixed and mobile bearing knees.  So any new

fixed bearing design has to satisfy the FDA on these

different aspects.

            Just looking down the left‑hand column if

you will, the risk, and the special control means the

accepted tests which have to be provided to the FDA

for approval.  So we have sterility, biocompatibility,

metal compatibility, metal corrosion, delamination of

porous coating.

            Now let's look at the second group.  These

are risks and special controls that have specific

considerations when applied to mobile bearing knees.

We are not saying that all mobile bearing knees are

alike.  Clearly, they are not alike. 

            What we are saying, however, is that the

following risks have been evaluated and identified,

and tests have also been identified which will

evaluate these particular risks for any new design

that is submitted to the FDA.  These tests are not

intended to be rubber stamps.  They're intended to be

tough standards, tough tests, which any new design has

to pass before being accepted.

            So, first of all, tibial tray fracture ‑‑

this, of course, applies to fixed bearing designs.

But in mobile bearing designs we could, for example,

have features such as tracks, stops, and other things

that could increase the stress on the tibial tray.

There's a very well accepted standard ‑‑ ASTM and ISO

standard, as well as articles in the literature, which

can reliably test this aspect, and also stress

analysis using computer technology.

            The other very important issue is wear of

the articulating surfaces.  Of course, it is probably

the case that a multi‑directional mobile bearing knee

may wear more than a unidirectional mobile bearing

knee.  It is quite possible.  However, the bottom

line, in fact, is wear testing.  That's what the tests

are supposed to do.  They find out whether a

particular design has more wear than is acceptable.

            Now, there are two tests ‑‑ the ASTM and

ISO tests ‑‑ and there are well developed simulating

machines which have been developed here and in Europe

for testing the wear of any kind of knee.  The total

wear is measured, and the extra test, if you like,

that can be applied to mobile bearing knees is to make

sure any mobile bearing knee is compared to a well‑

known fixed bearing knee design with a long clinical

history.  All wear testing should have a standard like

that to make it credible.

            Now, back side wear ‑‑ if one wanted to

look at the back side wear as well, because this may

be a concern ‑‑ certainly in testing little marks or

engravings can be put on the back side of the plastic.

And at the end of the wear test one can look at the

wear associated with just the back side as well as the

top side.  Obviously, the simulator measures the total

wear, and we can now measure the wear on both sides of

the bearing.

            The wear particles is a concern.  We've

heard earlier that the particles may resemble hip

particles.  The point of this is that hip particles

tend to be submicron and are more reactive to tissues.

Knee replacements generally have larger particles, and

they are probably less reactive to tissue.  And there

is also less wear associated with knees anyway.

            However, again, that's what the tests are

for.  A particular mobile bearing knee design would be

tested.  The particle distribution, the particle size,

would be measured by these well accepted test methods

looking at fibrils, flakes, and granules, and looking

at the size distribution of these different particles.

            Now, spinout of the insert ‑‑ you've heard

about this, and I think the panel members have seen

this implant being passed about ‑‑ it is possible to

lift the insert right off the tray, but also it's

possible that the femoral component could roll off the

plastic.  This has been recorded, particularly with

the Oxford meniscal knee where the meniscus is a

freely‑floating anterior/posterior piece of plastic.

            This is definitely an issue with mobile

bearing knees.  However, the incidence, as you have

heard, is very, very small, for example, with the LCS

knee.  It's a very, very small incidence indeed.  In

order to test for this, the constraint testing, as you

see here ‑‑ spinout of the insert ‑‑ constraint

testing, where you put the thing ‑‑ the device in a

machine, apply A/P, internal/external, medial/lateral,

and various analogous forces, and look at the

stability of the femoral component on the plastic

insert ‑‑ now, again, the additional tests that should

be done with mobile bearing knees is not only to do

that but to measure the actual jump height of the

femoral on the plastic.  This can be added by the FDA

in their new guidance document.

            So, in other words, the lift ‑‑ the amount

of lift that the femoral component has to go to

dislocate from the plastic can be measured at all

different angles of flexion ‑‑ extension and high


            Insert/tray disassociation ‑‑ again, this

device that was passed around ‑‑ you can actually lift

the device off.  This is an issue with fixed bearing

designs as well.  The plastic is snapped into a metal

tray.  Occasionally, a plastic insert, if it's thick,

for example, can flip out of the metal tray.  It's a

very rare, but occasional, risk.

            This risk also exists with mobile bearing

knees.  Now, there are tests ‑‑ the FDA guidance

document component interlock strength section ‑‑ if

you look at the underlined piece, component interlock

strength section, encourages tests where we load the

front and the back of the insert at the extremes of

loading as if a patient was getting up from a very low

position or extending very violently.

            So these instances have been identified,

and tests have also been identified by the industry ‑‑

and are well‑known tests ‑‑ to test for plastic coming

loose from the metal tray.

            Insert defamation or fracture ‑‑ nobody

mentioned so far the rotating platform, of course,

because it's rotating, the plastic can overhang from

the edge of the metal.  You could say, "Well, this

might increase the stresses on the plastic bearing and

be dangerous."

            Now, again, in modern designs, this is not

an issue.  This is no failure that I'm aware of,

certainly, that the literature has identified, since

the very early meniscal bearing knees, fracture of

inserts.  It's a very rare occurrence.  However, it

needs to be tested.

            And, again, the FDA guidance document on

contact area and stresses for mobile bearing knees,

the new guidance document for mobile bearing knees,

should include a section on measuring the contact

areas and stresses in a mobile bearing knee at all the

degrees of rotation of the insert on the metal tray to

ensure safety and avoidance of excessive stresses.

            Soft issue impingement or joint not

balance ‑‑ you've heard from several surgeons that in

their opinion the surgical technique is very similar.

Balancing the knee, making the ligaments just the

right tightness in extension and in flexion, indeed at

all angles of flexion, is a prerequisite for any

design.  And that applies equally to mobile bearing

knees as well as to fixed bearing designs.

            This is primarily the job of the surgeon

community itself ‑‑ in training, in lecturing.  And,

of course, surgical technique by the experts who

develop these devices, and also training courses which

are appropriately put on by different companies.

            One more thing ‑‑ damage to the insert of

the rotational stop.  We've heard about this.  Some

devices, including the one that was passed around, has

stops to stop it rotating.  Again, the standard wear

testing can be used to evaluate this.  This will stop

the form of the plastic and present the bearing from

functioning correctly.

            What I'm trying to get over in these

discussions is the tests that already exist ‑‑ well

established or else they're in the literature and can

relatively easily be devised by the FDA where

necessary to augment the tests for mobile bearing


            Patella wear ‑‑ I won't dwell on that.  If

a mobile bearing patella was submitted to the FDA ‑‑

and to my knowledge there is only one ‑‑ similar kinds

of tests can be applied to patellas. 

            So, in summary, I believe the mobile

bearing knee risks are well understood and are similar

for the most part to fixed bearing designs.  However,

special controls for the risks associated with mobile

bearing knees are either commonly used in industry,

exist as ASTM and ISO standards, or can be adapted for

any unique characteristic of a specific mobile bearing

knee design.

            Every mobile bearing knee design must pass

a range of tests appropriate for that design.  So I

believe that a new special controls ‑‑ if you'll look

at this here.  A new special controls guidance

document we believe is needed from the FDA to

recognize these extra risks, or extra nuances if you'd

like, of mobile bearing knees, that just describe very

clearly each test and each test parameter that is

required for mobile bearing knees.

            And I believe that the literature, current

practice, experts abound in order to come up with

suitable tests for mobile bearing knees.  And there is

enough knowledge, I believe, today to be assured that

the tests are reasonable.

            OSMA, which I represent, believe that

special controls, when combined with the general

controls, will be sufficient to provide reasonable

assurance of the safety and effectiveness of mobile

bearing knees.

            Thank you very much for your attention.

Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks very much,

Dr. Walker.

            Mr. Maislin, you're going to speak next.

            MR. MAISLIN:  Good morning to the panel.

My name is Greg Maislin, and I am the principal

biostatistician of Biomedical Statistical Consulting.

We're a contract research organization that

specializes in randomized clinical trials for

regulatory support.  I'm also an adjunct faculty

member at the University of Pennsylvania School of

Medicine, where I serve as the director of the

Biostatistics and Patient Recruiting Corps at the

Center for Sleep and Respiratory Neurobiology.

            I have the privilege today of having OSMA

support my appearance here to summarize their work in

summarizing the current literature.  And I also should

say that Biomedical Statistical Consulting has or has

had several of the petitioning sponsors as clients.

            It's important to have a feel for what the

current literature looks like when we try to extract

literature to compare the mobile bearing and fixed

bearing clinical outcomes.  Randomized clinical trials

comparing mobile bearing implants to fixed bearing

implants are largely unavailable in the literature.

            Therefore, methods of meta‑analysis that

are appropriate for observational studies were

utilized as opposed to those that might be appropriate

for randomized clinical trials where two treatments

are compared head to head. 

            The authors of the study utilized methods

that were in the literature to produce their

literature summary meta‑analysis, particularly a

summary from Callahan that was published in JAMA in

1994, which included a meta‑analysis of fixed bearing


            There were two separate meta‑analyses that

were performed.  One of them will give rise to an

estimate of what the clinical outcomes look like,

comparing fixed bearing to mobile bearing.  And the

other specifically addressed what does the profile of

implant survival look like. 

            And I'm going to jump in a sense to a last

slide.  I'll come back during the next few moments and

try to justify this.  But essentially, at the end of

the day, what the outcome of the first meta‑analysis

said was that the proportion of knees that are

expected to have good to excellent clinical results

are very similar in this particular set of studies

that were summarized.

            Moreover, that the cumulative survival was

also very similar.  There will be claim that the

survival is better for mobile bearing compared to

fixed in this set of studies, even though it was 93

percent survival versus 91 percent survival.  But what

we will claim is that there is good evidence that the

survival is similar between mobile bearing and fixed

bearing, at least in the current set of studies that

were revised.

            An important additional aspect that I'll

bring out in implant survival is that the variability

in implant survival in the set of studies that were

summarized was very similar to the variability in

revision rates in the set of fixed studies that were


            So it didn't appear that the several

different designs that were summarized for mobile

bearing had any impact on variability in revision

rates.  Moreover, an earlier speaker indicated that

the wear characteristics of mobile bearing devices

cannot be predicted from fixed bearing devices.

            But to the extent that the most important

consequence of wear is increased revision rates, the

literature doesn't support that contention.  It seems

to support the contention that in terms of the wear

effect on revision, the revision rates of mobile

bearing knees can be predicted from the fixed bearing


            I'm going to briefly just summarize how

those few numbers came to be.  It was an English

language articles summary.  There was an

identification of 22 cohorts from 21 studies that ‑‑

that met the criteria and contained the mobile bearing

devices.  The reference group for the first comparison

of clinical outcomes was extracted from the Callahan

meta‑analysis of fixed bearing. 

            A separate set of criteria, which I'll

detail in just a second, summarize the experience from

mobile bearing devices, and compared it to another set

of studies that produce 30 fixed bearing survival

estimates.  These were obtained in 16 articles.

            In general, for both studies, the mobile

bearing knee designs that were contained in this

literature summary meta‑analysis included both

cemented and uncemented designs, unicompartmental,

bicompartmental, tricompartmental replacements, multi‑

directional platforms, rotating platforms, meniscal

bearing articulations, both PCL sacrificing and PCL

sparing.  All mobile bearing articles were included

without regard to the cement technique, number of

components replaced, or mobile bearing type and PCL


            This is the same approach that Callahan

used to construct their fixed bearing meta‑analysis,

and that was used as the control group for the first


            So, in summary, in order to include a

sufficient number of mobile bearing knee articles for

the meta‑analysis, the bearing type and number of

compartments replaced was not used to exclude studies.

All were included.

            Specifically, for the first meta‑analysis,

the only criteria was that every study had to have at

least 10 or more patients.  They had to report post‑

operative clinical results, and specifically they had

to report post‑operative clinical results in terms of

a 100‑point rating score, so that those could be


            In total, the population had about 3,000

knees, about 2,500 patients.  It was a typical

population.  There were almost two‑thirds female, 82

percent osteoarthritis, and 13 percent of the knees

were in bilateral patients.  So this is the mobile

bearing cohort that was constructed on the basis of

those criteria. 

            Two outcomes were compared from that

cohort to the Callahan cohort.  There was a mean

percent improvement in the ‑‑ in a post‑operative

global rating, and, in particular, I typically used

percent of cases with good to excellent post‑operative


            Also, following the Callahan technique,

the studies were weighted in terms of the contribution

to the overall success by the number of knees, by the

size of the study. 

            This is a summary of the primary results,

and I'll call your attention to this row here.  This

is the numbers that I quoted at the beginning on the

first ‑‑ on average among those studies weighted by

the number of knees per study.  The success rate was

90.3 percent, comparable to 89.3 percent.

            In several places I will call attention to

two devices which are obsolete and no longer used in

current implantation.  It's the Oxford Phase I and the

Accord.  If you were to remove these outliers ‑‑ and

they look like outliers along a number of dimensions

‑‑ the success rate goes up.  But to be conservative

we could just note that these success rates between

the mobile bearing and fixed bearing are virtually


            I'll note that the ‑‑ that in this first

meta‑analysis the weighted mean followup was six

years.  It was four years in the Callahan.  The

revisions are going to be more formally analyzed in a

second, but I'll just note that if the revision rate

is six percent in the mobile and four percent in the

fixed, and if the revision rate is about one percent

per year, which Dr. Stiehl noted in his review it is

for mobile, and also it has been ‑‑ that approximates

what we know about fixed, then this two percent

increase in the revision rate would be explained by

the two‑year increase in mean followup.  And we'll see

a reverse relationship shortly.

            The other meta‑analysis focused on

survival.  For the mobile bearing, 10 patients were

required.  They had to report an estimate of implant

survival.  For the fixed bearing ‑‑ this is important

to note ‑‑ that a criteria that was used was that

there had to be at least one cohort in the study that

had at least 10 years of followup, and that the study

had to be cited at least twice in peer reviewed

journals as having high durability and clinical

success.  In other words, the attempt was to create a

fixed bearing control group of high quality.

            The details were that there were 37

articles that met these criteria ‑‑ 16 and 21.  There

were 111 survival estimates.  There were multiple

estimates in each of the studies, because of various

definitions.  For studies that reported multiple

estimates, the estimate that was most consistent ‑‑

that had a most consistent definition of revision ‑‑

for example, revision for any reason ‑‑ and the

longest followup was retained.  This was a consistent

rule used, and it was a priori specified.

            What it culminated in was 30 estimates of

fixed bearing implant survival and 26 estimates of

mobile bearing survival.  And these were the studies

that were compared.

            Among those studies, I'll note that the

mean followup was about six in the mobile bearing.  It

was about eight in the fixed bearing as a consequence

of requiring at least 10 years in any one cohort in

the studies.  And what we end up seeing is this

picture ‑‑ I don't want to dwell on this picture,

except to note the following characteristic.

            First of all, these confidence intervals

here are only reported in the mobile for primarily

compared with fixed, because these studies were later,

and statisticians finally had their way and induced

authors to report confidence intervals. 

            These are the confidence intervals that

are reported in the articles themselves.  The earlier

studies didn't report confidence internals. 

            Besides that editorial, I just want to

comment on these blue dots.  These are the revision

estimates.  The main finding is that if you look at

the weighted average, the centerline here, these are

about the same in the fixed cohort studies that met

the inclusion criteria compared to the mobile studies

that met the inclusion criteria. 

            And, moreover, if you look at the

variability around the centerline, and you look at the

variability around the centerline, generally speaking

they look very, very close.  Even though there was a

number of different design features, those design

features didn't translate into increased variability

in revision rates. 

            The variability in revision rates seems

comparable to fixed.  So both the average is the same

and the variability around the average is the same.

Those two red dots are highlighted, because those are

the two outliers that I mentioned earlier as being

obsolete and no longer used designed.

            So the bottom line is that the ‑‑ that in

this cohort of studies that met the criteria that was

put out by the authors of the study, there was 93

percent survival in the mobile bearing and 91 percent

survival in the fixed.  But note again that there's

about a two‑year difference in average followup, but

this time the mobile bearing had less, and that

translated into about a two percent increase in


            In conclusion, that the ‑‑ I'll point out

very quickly that this is the statistical techniques

that were used to take all those numbers and reduce

them down to those two numbers ‑‑ weighted least

squares, weighting by nears, resampling for robust

confidence interval estimation. 

            The model assumptions were tested and

verified from the analysis of variance that was used,

and heterogeneity and survival was estimated with

formal testing with chi square statistics, and there

was no heterogeneity in either group.

            I'll summarize and say that the meta‑

analysis results found that the mobile and fixed

bearing implants are similar in both effectiveness and

survival.  The clinical outcome was about 90 percent

success in both mobile and fixed.  The implant

survival was about the same ‑‑ about 90, 91 to 93

percent cumulative survival in both the mobile and the


            The mobile characteristics ‑‑ cemented

versus uncemented, rotating platform versus meniscal

bearing, etcetera ‑‑ did not demonstrate significant

differences in clinical outcomes, and the results

appeared to favor the downclassification of mobile

bearing from Class III to Class II, given the

similarity in clinical outcomes and homogeneity among

device survivals.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks very much,

Mr. Maislin.

            What I'd like to do now is ask Mr. Allen

to come up and give the FDA presentation.  We'll have

an opportunity for panel questions of both OSMA and


            MR. ALLEN:  Good morning.  My name is

Peter Allen.  I'm a biomedical engineer in the

Orthopedics Branch in the Office of Device Evaluation

at FDA.  I would like to thank OSMA for their

presentation and for their efforts in preparing this

reclassification petition.

            Today I will present a summary of FDA's

review of this petition, along with some questions we

would like our panel members to discuss and answer.

            Before we get to that, I'd like to provide

a bit of an overview on the current regulatory status

and marketing history of mobile bearing knees.  First,

I'll cover some general background in the medical

device classification process.  Then I'll briefly

discuss the mobile bearing knees that have been

approved for use and are currently on the market in

the United States.

            After that I'll discuss the information

provided in the petition and how it fits in with the

requirements for reclassification.  And then I'll

identify the questions we'd like our panel to discuss.

            The 1976 amendments to the Food, Drug, and

Cosmetic Act provided regulations for the

classification and regulation of medical devices.  The

Act established three classes of medical devices

dependent on the regulatory controls needed to provide

a reasonable assurance of safety and effectiveness.

            To provide a reasonable assurance of

safety and effectiveness for Class I devices, general

controls are considered adequate.  This slide

identifies some examples of general controls.

            When general controls alone are

insufficient, a device may be classified into

Class II.  Class II devices require additional special

controls.  Special controls include guidance

documents, performance or consensus standards,

labeling, and possibly even clinical data.  The

special controls issue will be the focus of much of

our later discussion, and we will have a few of our

panel questions directed to this issue.

            Devices classified into Class III ‑‑ pre‑

market approval ‑‑ when it cannot be classified into

Class I or Class II, because general and special

controls are insufficient to provide reasonable

assurance of safety and effectiveness.  In addition,

any new device type first introduced into commerce

after the 1976 amendments to the Act, commonly

referred to as post‑amendments devices, are by statute

automatically classified into Class III.

            Class III devices are regulated using the

validated scientific evidence that is presented to FDA

in a pre‑market approval application, or PMA, to

establish the safety and effectiveness of the device.

Typically, this requires the submission of clinical


            Just like classification, FDA will

reclassify a device into Class I, II, or III,

depending on the level of regulatory control needed to

provide reasonable assurance of safety and

effectiveness.  FDA may initiate the reclassification,

or any person, manufacturer, or importer may submit a

petition for reclassification.

            A Class III device may be reclassified in

Class II when FDA can identify the risks associated

with the device and the manner in which these risks

can be controlled by general and special controls.

            A special controls guidance document is

one way in which risks can be controlled.  This is a

document created by FDA that provides acceptable

methods for controlling the risks identified for a

given device type.  It is intended to provide guidance

by conveying FDA's current thinking about a specific

device type.  And it provides recommendations on how

to address the issues presented in the guidance, such

as the use of performance and consensus standards, the

use of labeling for those instances when clinical data

may be deemed necessary.

            A company need only demonstrate that their

Class II device meets the recommendations of the

special controls guidance document to receive FDA

clearance for marketing. 

            So there's a quick overview of the

reclassification process.  Now let's turn our

attention to the devices being considered in this

reclassification petition ‑‑ the mobile bearing knees.

            I'll discuss the current classification

status of these devices and look at their marketing

history here in the U.S., as well as the indications

for use for which they were originally approved.

            Mobile bearing knees were first introduced

for commercial distribution in the U.S. after the 1976

amendments to the Act.  Therefore, by regulation, they

are Class III post‑amendments devices and require an

approved PMA prior to marketing.

            To date, FDA has approved three mobile

bearing knee PMAs.  The first PMA for the DePuy LCS

total knee system was approved in 1985.  The original

approval included two design versions ‑‑ a rotating

platform design, which was crucial at sacrificing, and

a meniscal bearing design was crucial at retaining.

            In addition to the knees approved in the

original PMA, over the years numerous design

modifications and variations have been approved

through supplements to that PMA.  So there are

actually multiple mobile bearing knee designs on the

market resulting from this one PMA, but they are all

offshoots from the same device system.

            The second PMA was for the LCS

unicompartmental knee, which was approved in 1992.

And the third PMA was for the Oxford unicompartmental

knee from Biomed, which was just approved this past

April, although older versions of the device have been

on the market in Europe for over 25 years. 

            All three of these systems are well

described in the published literature.  Each of these

three devices have their own approved indications for

us.  Indications approved for the original rotating

platform and meniscal bearing versions of the LCS are

listed here ‑‑ the only difference between the two

being that the rotating platform version was approved

for use in revision procedures, while the meniscal

bearing was not.

            The additional design variations, since

added to the system over the years, may have slightly

different implications from those approved with these

original designs. 

            Here are the indications for use approved

for the LCS and the Oxford unicompartmental knee

systems ‑‑ the main differences between the two

systems being that the LCS was indicated for

uncemented use on either condyle in older patients,

and the Oxford was indicated for cemented use only on

the medial condyle with no limitations in patient age.

            Now I'll move on to the current petition

in which OSMA has provided a proposal to reclassify

the mobile bearing knees from Class III to Class II.

I have already reviewed for you what's required to

reclassify a device from Class III to Class II.  That

is, you must identify the risks and the manner in

which these risks can be controlled by general and

special controls. 

            So I will now review the following items

included in the petition ‑‑ most important being the

risks identified for mobile bearing knees and the

special controls proposed to address these risks.

            The petition is split into two groups of

mobile bearing knee designs.  The first consists of a

total knee design, which contains patella, femoral,

and tibial components, and is intended to replace the

entire knee joint.  The second consists of a

unicompartmental design and contains only femoral and

tibial components intended to replace either the

medial or lateral compartment of the knee.

            The proposed classification descriptions

for both designs are listed here.  Further description

of the classification definitions are included with

the panel questions in the presentation packet you

received this morning.  I note this simply because the

adequacy of these classification definitions is

included as one of the panel questions.

            Both device type are available in many

design variations, depending, for example, on the

directional mobility of the bearings, the type of

constraint, levels of congruence, management of the

patella, and management of the posterior crucial


            Each group of colors here represents a

different design variable, such as the bearing type,

method of constraint, amount of constraint, etcetera.

As you can see, there are a large combination of

variables that can affect the design of a mobile

bearing total knee.  Reclassification of the currently

approved devices, which by regulation is what we are

actually doing, would potentially provide for the

reclassification of these design variables, many of

which are incorporated in the approved devices.

            Again, please note that questions

regarding the adequacy of the data in the petition

supports these multiple design types.  That is, the

identification of the risks and the appropriate

special controls are included in the list of panel


            Although much fewer in number, there are

also various combinations of design variables that go

into the development of a unicompartmental knee, as

you can see here. 

            Turning now to the indications for use ‑‑

the proposed indications for the mobile bearing total

knee are listed here.  These devices are also

indicated for use with or without bone cement.  The

proposed indications for use for the mobile bearing

unicompartmental knees are listed here as well.

Again, these devices are indicated for use with or

without the use of bone cement.

            The sponsor has provided over 230

published references in support of the preclinical and

clinical issues in this petition.  Some of the

preclinical issues addressed include evaluation of

device kinematics, wear of the mobile bearings, and

device biomechanics.  With regards to the clinical

data, the sponsor summarized a series of 48 studies

which evaluated the various types of bearings listed


            Data presented for each study included

study design, demographics, safety, effectiveness, and

survivorship.  The majority of these studies focused

on those devices already approved for use in the

United States ‑‑ that is, the LCS and the Oxford

mobile bearing devices. 

            In addition, there are also nine published

review articles on mobile bearing knees, information

from seven ongoing FDA approved clinical trials, two

international clinical outcome studies, two meta‑

analyses, one comparing clinical outcomes of mobile

bearing knees of different types, and the second

comparing survivorship of mobile bearing knees versus

fixed bearing knees.

            These clinical experiences underscore the

strong influence of the technical performance of the

operation on the long‑term success of a new device.

Properly aligned knee replacements that have restored

ligament balance appear to have survival rates of 10

years or greater, irrespective of bearing mobility.

These data indicate that when provided with

medial/lateral stabilization, mobile bearing knees

provide equivalent results to fixed bearing knees.

            The sponsor has also provided information

on adverse events.  This includes data gathered from

searches of FDA's medical device reporting program, or

MDRs, reports from the published literature, and data

from manufacturers on their FDA approved clinical


            Again, the vast majority of this

information relates to the DePuy LCS devices and the

Oxford unicompartmental device from Biomed.  The MDRs,

in particular, relate specifically to the LCS devices.

            Here I have listed the most commonly cited

adverse events that were associated with revision,

although they're not listed in any particular order.

It is noted that the three most common adverse events

cited in the MDR database for the LCS knees were pain

accompanied with swelling, fractured bearings, and

loosening, respectively.

            As you can see, the patient‑related

adverse events are fairly typical of the type of

events you might see with any total joint replacement

procedure.  And the device‑related adverse events are

consistent with those types of complications often

seen with fixed bearing knees, although there does

appear to be a tendency to see a greater number of

bearing dislocations, subluxations, and impingement

with the mobile bearing knee designs.  And the way in

which these events occur may be somewhat different

from the mobile bearing knee versus the fixed bearing


            The sponsors proposed methods to control

for these potential device‑related risks.  Again, I

alert the panel that there will be some questions

forthcoming regarding these risks and the adequacy of

proposed special controls.  High wear rates of

polyethylene bearings can lead to particle‑induced

osteolysis, which can in turn lead to loosening of

device components.

            To address the risk of wear, which

includes all of the different modes of wear envisioned

with the mobile wearing knees ‑‑ that is, designs with

tibial posts, rotational stops, grooved tracks, multi‑

directional platforms, patella bearings, etcetera ‑‑

the petitions have suggested the following two

standards:  ASTM F1715 and/or ISO 14243‑1.

            The ASTM standard is a general guideline

for establishing test conditions and obtaining wear

measurements for wear simulation of the femoral/tibial

components of knee joint prostheses.  As such, it does

not provide any specifics on wear testing of the

patellar/femoral compartment of the knee joint.

            The ISO standard is an international

standard for wear testing of total knee joint

prostheses, which includes loading, displacement, and

environmental testing parameters.  Of concern to FDA

is whether such tests can provide results that

reproduce clinical wear behavior of the many mobile

bearing knee designs.

            One way to determine this is through

analysis of the wear surfaces and wear particles.  To

evaluate the wear of the bearing surfaces, the sponsor

has suggested two additional methods of analysis,

including the use of coordinate measuring machines to

quantify the volume of wear on the articulating

surface, and the measurement of changes and engraved

markings on the back side of the bearing.

            Back side wear measured by this method on

bearings tested in knee joint simulators has been

shown to correlate well with wear measured and

clinically retrieved specimens according to the


            The sponsor has also proposed a means to

evaluate wear particles.  For analysis of the wear

particles, they recommend ASTM F2025 and/or ISO

14243‑2.  The ASTM standard uses a weight loss method

of wear determination for polymeric components used in

human joint prostheses.

            The ISO standard also employs gravimetric

methods for measurement of wear in total knee joint

prostheses, as well as dimensional methods using a

coordinate measuring machine to determine biometric

wear rate.

            To evaluate or mitigate the risk of tibial

insert or patellar bearing deformation or fracture,

which may result from overhang with respect to the

metal tibial base plate, the sponsor has proposed

utilizing the wear test just previously mentioned.

            However, the wide variety of mobile

bearing knee designs proposed ‑‑ however, with the

wide variety of mobile bearing knees proposed, it is

not clear if such wear tests can provide results that

reproduce the bearing deformation and/or fracture that

is seen clinically.

            Contact area changes and stress changes on

the insert can change dramatically as the insert moves

throughout its available range of motion.  To evaluate

the effect of this, the sponsor again proposes using

wear simulator testing standards to address the

potential risk that damage due to the changing load

profile might impart on the components.

            In addition, they also propose using

existing recommendations for contact area stress

evaluation and the current FDA guidance documents for

fixed bearing knees.  These include evaluations of the

tibial/femoral and patella/femoral interfaces, several

different angles of flexion, using relevant

physiologic loads.

            A copy of the fixed bearing knee guidance

document was provided in the panel presentation

package you received this morning. 

            Separation of the tibial bearing from the

metal base plate is an inherent risk associated with

the design of most mobile bearing knees.  This may

lead to dislocation, subluxation, instability, or

impingement.  Due to their requirement for mobility,

tibial bearings cannot be held rigidly in place, and,

therefore, are susceptible to separation as a result

of the shear and torque forces experienced in the


            To address this risk, the sponsor has

proposed a characterization of the component interlock

strengths as recommended in the current fixed bearing

knee guidance.  This includes anterior/posterior and

medial/lateral shear testing and/or static tensile

pulloff testing of the tibial bearing.

            But as a number of mobile bearing designs

have no means or limited means of fixation between

components, the value of such testing appears of

questionable use for some of the designs.

            Tipping of the tibial insert is a risk

that has been identified for the rotating platform

design with the Cohen and Cohen configuration.  Edge

loading of the insert can lead to the tipping of the

insert and partial dislocation or subluxation of the

component.  Repeated tipping of the insert may lead to

deformation, wear, or fracture of the component.

            The sponsor recommends characterization of

the component interlock strengths as recommended in

the current fixed bearing knee guidance.  However,

this design type has no permanent connection between

the components and relies on the soft tissue

structures to hold the components in place.

            In addition, it's not clear if the

suggested anterior/posterior, medial/lateral, shear,

and/or static tensile pulloff testing will

characterize the clinical propensity for tipping.  So,

once again, the value of such testing appears to be

questionable for some designs.

            Spinout of a tibial insert or patella

bearing has been identified as another risk that is

unique to the mobile bearing knees.  Within the

petition, spinout is defined as the excessive rotation

of the polyethylene insert resulting from at least one

femoral condyle riding up and over the lip of the

insert, such that the femoral condyle is no longer in

contact with the insert's articular surface.  This can

lead to dislocation, subluxation, wear, impingement,

and instability.

            To mitigate the risk for bearing spinout,

mobile bearing devices should be evaluated to limit or

eliminate the potential for spinout.  Currently, there

is no standard for spinout testing.  However, the

sponsor has suggested that the potential for tibula

insert spinout may be assessed using a modified knee

constraint testing standard after adapting for

physiologic compressive loads, rotary torques, and

various moments that are deemed to be causative of

insert spinout.

            This ASTM standard, which is a standard

test method for the determination of total knee

replacement constraint, is one of the same special

controls used for the evaluation of fixed bearing

knees.  To evaluate the risk of the patella bearing,

they suggest a patella/femoral lateral stability test,

as recommended in the current fixed bearing knee


            The sponsor believes this testing should

provide reasonable assurance that the insert bearing

does not have an increased spinout risk as long as it

does not spin out under normal physiological loads.

However, it is not clear to FDA how this modified

physiological testing would correlate with the

clinical mechanisms of this type of failure mode.

            Separation of the patellar bearing from

the metal base plate is an inherent risk associated

with the design of the mobile bearing patellas.  It is

also a risk that is associated with fixed bearing

patellas.  As a result of the shear and torque forces

experienced in the patella/femoral joint of the knee,

there is a risk of patella bearing separation. 

            To address this risk, the sponsor has

proposed characterization of the component interlock

strengths as recommended in the current fixed bearing

knee guidance.  These include static tensile pulloff

testing, shear fatigue testing, and evaluation

according to ASTM Standard F1672, which is the

standard specification for resurfacing of patellar

prostheses, although it is noted that this standard

has no device‑specific test methods to evaluate the

performance of patellar prostheses.

            It is well known that successful

implantation of mobile bearing knees is highly

technique‑sensitive.  Without proper attention given

to soft tissue balancing, instability of the implanted

joint is a very real risk.  To minimize this risk, the

sponsor suggests that special attention should be

given to providing appropriate instructions for use of

the device in the product labeling.

            The sponsor believes surgeon training and

detailed surgical techniques that include instructions

for proper soft tissue balancing will provide

reasonable assurance of safety and effectiveness.

            Although no specifics were given for these

recommended controls, such as the requirements for

training, it appears that these would include the same

type of information currently provided for the fixed

bearing knees.

            The only risk identified as specifically

unique to unicompartmental knees is that

unicompartmental devices require an intact anterior

crucial and posterior crucial ligament.  To mitigate

the risk of these devices being implanted in patients

without functional crucial ligaments, the sponsor has

recommended appropriate instructions for use in the

product labeling and surgeon training in the proper

surgical technique as ways to control this risk.

            And, lastly, the risk of prosthesis or

soft tissue impingement has been identified for mobile

bearing knees.  Impingement of the soft tissues can

lead to soft tissue irritation, swelling, bleeding,

and pain, and, again, this has to do with the mobile

bearing nature of the tibula insert and patella


            To control for this risk, the sponsor

believes wear testing in a knee simulator and

appropriate surgeon training with detailed surgical

techniques that include proper instructions for use in

the product labeling will provide a reasonable

assurance of safety and effectiveness.

            It is the potential occurrence of these

adverse events or risks that we at FDA are responsible

for evaluating.  As such, we must have methods at hand

to evaluate or mitigate these risks for all device

types that are to be reclassified.  Remember, to

reclassify into Class II, we must have reasonable

assurance of the safety and effectiveness of these


            And this is where we come back to the

issue of special controls.  The challenge for FDA will

be to develop a special controls guidance document

that can adequately address the risks identified for

all mobile bearing knee types being reclassified.  If

special controls are not available or are insufficient

to control these risks, then reclassification may not

be an option.

            FDA's primary concern is whether the

proposed special controls are adequate to properly

evaluate all of the mobile bearing knee designs

covered by this reclassification petition, and whether

they provide that reasonable assurance of safety and


            Based on the information provided in the

petition, FDA has some questions we'd like the panel

to address in order to help us reach a decision on

this reclassification petition.  We ask for your

recommendations on these questions after you have

completed your discussion.

            The questions focus on four general areas.

I'll present the specific questions in just a moment,

but the four areas of focus are the proposed

classification definitions, the risks to health

presented by mobile bearing knees, the ability of

special controls to adequately control the risks

associated with these devices, and the data presented

and whether it supports the reclassification of the

mobile bearing knees.

            I will run through all of these questions

for you now, and then come back to them one by one if

needed later on as you address them following your


            I will refer the panel members to their

copy of the panel questions in the presentation packet

you received this morning for the complete text of the

questions.  A few of these are somewhat lengthy and

are paraphrased on some of these slides.

            Question number 1:  do you believe the

proposed classification definitions for the following

device configurations recommended for reclassification

adequately describe the devices?  If not, what changes

in the definitions do you recommend?  And, again,

we're specifically talking about total mobile bearing

designs and the unicompartmental designs.

            A copy of the proposed classification

definitions are included with a copy of the panel

questions, and I have also put them on the following

two slides here.  I'll read them to you now, and we

can come back to them after the discussion, if we need

to look at them.

            The sponsor has proposed the following

classification description for a total mobile bearing

knee.  A knee joint patellar/femoral, tibula, metal

polymer, mobile bearing, cemented, or porous‑coated

uncemented prosthesis is a device intended to be

implanted to replace the knee joint.  The device

permits either unconstrained or constrained rotation

of the articular surface in the transverse plane, and

may or may not permit limited anterior/posterior

and/or medial/lateral movement of the articular

surface upon the tibular component.

            It has no linkage across the joint.  The

device may use a fixed structural porous metal in

place of a porous coating.  This generic type of

device is designed for use with bone cement and/or to

achieve biological fixation to bone without the use of

bone cement.

            And the following classification

description has been proposed for the unicompartmental

mobile bearing knee.  It's a knee joint

patellar/femoral, tibula, metal polymer, mobile

bearing cemented or porous‑coated uncemented

prosthesis.  It's a device intended to be implanted to

replace part of a knee joint.

            The device permits either unconstrained or

constrained rotation of the articular surface in the

transverse plane, and may or may not permit limited

anterior/posterior and/or medial/lateral movement of

the articular surface upon the tibular component.

            It has no linkage across the joint, and

the device may use a fixed structural porous metal in

place of the porous coating.  This generic type of

device is designed for use with bone cement and/or to

achieve biological fixation to bone without the use of

bone cement.

            And on to question 2:  do you believe the

risks to health of the following device configurations

proposed for reclassification are adequately

described?  If not, what additional risks do you

believe should be included?

            Question 3:  special controls have been

proposed to address the risks to health identified for

both of the above‑referenced device configurations and

all related subconfigurations.  Please respond to the

following questions regarding specific risks and/or

special controls.

            3A.  Dislocation and subluxation of mobile

bearing knee components have been cited as common

complications in the literature.  Do you believe

appropriate special controls have been identified to

adequately address these risks?  If not, what

additional controls, if any, do you recommend to

address these risks?

            3B.  A reduction in wear is often cited as

a theoretical advantage of mobile bearing knees over

fixed bearing knees.  However, this has not been

consistently demonstrated clinically, and it is not

clear how well pre‑clinical wear testing of mobile

bearing knees correlates to the clinical situation.

In fact, the potential for third body wear appears

greater due to the fact that you have the two moving

interfaces instead of one.

            Currently, the state of development of

knee simulator wear testing has not yet been

standardized or clinically validated for all design

types of mobile bearing knees, and, therefore, may not

be applicable for all of the various mobile bearing

knee types identified in this petition.

            In light of the fact that wear appears to

be in part design‑dependent, do you believe

appropriate controls have been identified to

adequately address the risk of wear ‑‑ that is,

osteolysis and loosening ‑‑ of the various mobile

bearing knee designs under consideration in this

petition?  If not, what additional controls, if any,

do you recommend to address this risk?

            3C.  Labeling has been cited as a method

with which to control some of the identified risks to

health.  Proposed labeling requirements are consistent

with those generally found in current fixed bearing

knee package labeling. 

            Such labeling typically includes adequate

instructions for use, device description, indications

for use, contraindications, adverse events,

precautions, warnings, listing of compatible

components, and sterility information.  What

additional labeling, if any, do you recommend for

these mobile bearing knee devices?

            3D.  Do you believe appropriate special

controls have been identified to adequately address

the risks to health for each of the above device

configurations and the various subconfigurations?  If

not, what other special controls do you recommend to

address the risks presented by these devices?

            Question 4:  do you believe the data

presented in this petition supports the

reclassification of all total mobile bearing knee

prostheses identified in the petition?  If not, which

types of total mobile bearing knees do you believe are

inappropriate for reclassification?  And why?

            And do you believe the data in the

petition supports the reclassification of all

unicompartmental mobile bearing knee prostheses

identified in the petition?  If not, which types of

unicompartmental knees do you believe are

inappropriate for reclassification?  And why?

            At this time, I would like to thank the

panel for your time and attention.  This concludes

FDA's presentation.  I will now turn the floor over to

the Chair for discussion. 

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks very much,

Mr. Allen.

            We're going to take a break in a minute,

but I'd like to say before we break that when we

return we'll start with the panel's review of this

reclassification position led by Dr. Mayor and by Dr.

Larntz.  And we'll have the panel discussion and go to

the questions.

            It's now almost 10:05.  Let's start again

at 10:15, a 10‑minute break.

            (Whereupon, the proceedings in the

            foregoing matter went off the record at

            10:05 a.m. and went back on the record at

            10:18 a.m.)

            CHAIRPERSON YASZEMSKI:  Let's get back to

the meeting. Dr. Michael Mayor is the former

chairperson of this panel.  He is going to open this

part of the meeting with his remarks.

            Then Dr. Kinley Larntz, our statistician,

will give his remarks on the proposed reclassification

from the statistician's perspective.  The panel will

then have a general discussion, after which they will

focus their deliberations on the four FDA questions

that you saw a minute ago.  Then Ms. Shulman will

guide us in the completion of two documents:  the

reclassification questionnaire and the supplemental

data sheet forms.  We will then conclude our


            I will also mention that after Dr. Mayor

and Dr. Larntz speck, I am going to ask Dr. Witten to

make some general comments about reclassification so

that the panel can consider them as we go forth with

our discussions.

            I am going to ask Dr. Mayor to begin and

give us his presentation.  Dr. Mayor?

            MEMBER MAYOR:  Thank you, Dr. Yaszemski.


            MEMBER MAYOR:  These observations that I

will make are my own but are offered for consideration

by the panel and FDA.

            My sense is that the state of the art is

comparable for both "fixed" ‑‑ and I have used "fixed"

in quotes ‑‑ and mobile bearing total knees.  Fixed

and mobile bearing total knees as devices had with

some exceptions proven to provide some of the most

predictable and cost‑effective interventions available

to medical practice.  Consensus standards have evolved

to enhance the likelihood of evaluation of devices

submitted for determination can be assessed


            Many of the considerations regarding where

are common to both fixed and mobile bearings since

unintended motion and wear have emerged as significant

factors for the "fixed" bearing knees as well.  It is

not clear that mobile bearing designs emerged as a

source of excessive risk regarding wear.

            Stability represents another source of

concern.  In an effort to gain the advantages offered

by mobile bearings, do we expose the general public to

unnecessary, unacceptable risk?  Past experience

suggests that these risks are being addressed.

            We, the panel, are being asked to

determine if safety and effectiveness of mobile

bearing knee devices can be adequately assured were

they to be reclassified into the class II category.

            With the means available to FDA including

and not limited to available performance standards and

test methods, extensive clinical experience documented

in the literature, and established procedures for

evaluation of devices brought to FDA for approval, it

seems prudent and in adequate defense of the

well‑being of the general public to recommend to FDA

the reclassification of mobile bearing knees that is

the subject of this petition.

            Additionally, the development of special

controls and an appropriate guidelines document would

be strongly supported.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks very much,

Dr. Mayor.

            Dr. Larntz?

            MEMBER LARNTZ:  I'm going to make a few

scattered comments as I might.  Let me say, first off,

I think statistical analysis is a wonderful subject.

I think it is useful.  And no matter what you do, you

can be criticized, guaranteed.

            Some of what I am going to say is in the

form of some criticism, but if the positions were

reversed, ‑‑ and sometimes they are ‑‑ the same kinds

of criticisms might be leveled at work I would do.  So

I don't want to take this as too negative.  Okay?

            A basis for the reclassification we are

going to do or some of the statistical evidence for

that has to do with some meta analyses.  Meta analysis

is hard work.  I have done enough of it to know that

your work is never done.  And wherever you stop, you

could always do more.  And you could always be

criticized.  So that is where we are with the meta

analysis we have here, particularly meta analyses that

are from literature that has got to be selective.

            My gosh.  How do these papers get

published anyway?  I was a university faculty member

for 27 years.  Boy, there was pressure on me to

publish those papers, no matter what, no matter where.

But it's selective.  What gets published is selective.

There's no question about that.

            The literature, as was indicated in some

of the presentations, is incredibly variable,

incredibly variable.  I would say, however, that that

is the way life is.  That is why statistics are so

wonderful.  Everything is incredibly variable.

            By that token, we are looking at a

comparison to say that mobile bearing and fixed

bearing give similar results.  Similar within the

context of lots of variation is actually very easy to

achieve.  Do you hear what I am saying?

            Similar in the context of lots of

variation is very easy to achieve.  So saying that

they are similar doesn't mean that there is no

difference between them.  It means that we don't have

evidence, enough evidence, of difference if there are


            There clearly is not enough or the

literature is such that, as literature often is

composed of many different studies, often quite

scattered and published for different reasons.  And,

thus, it is very difficult to do a meta analysis that

will give you any firm, firm results.  So that, by the

way, was preamble in some sense but conclusion also,

by the way.  So maybe I should stop.  But I won't.

            So, on the whole, I believe that the

differences we see, the amount of difference we see,

actually is an understatement of the variation that

actually exists.

            There are a couple of technical reasons.

Those of you who aren't statisticians can turn off now

for a second.  I will tell you when to tune back in.

            A couple of technical reasons.  One is the

analyses, the meta analyses, were done in what is

called a fixed effects context.  Each study has its

own component.

            If you look at tables ‑‑ and I'm not going

to ask you to turn to page 277, but if you look at

tables there are clearly variations.  For instance,

the percentage good, excellent varies from in the 30s

to 100.  Well, that's different.  Thirty percent to

100 percent.  Do you believe that's different?

            That's the kind of variation we have in

these studies.  So these studies are clearly

statistically different.  That is, they have random

components that have to do with the studies


            And they are probably inexplicable but

need to be accounted for in any measure, any measure,

of variation that is given.  And in some sense, some

of the statistical analyses do that, but doing a true

random effects I think would answer that better,

rather than a fixed effects meta analysis.  That being

said, I don't think that's too terribly important, but

that is a technical comment.

            There is also some amputation of data.

Non‑statisticians don't have to tune in yet.  In

addition, there is some amputation of data based on

eight data points using regressions with five

variables.  Now, actually, non‑statisticians tune back

in because this might be something you might try to


            If you have eight data points and you use

five variables to predict them, guess what.  You can

do a very good job of prediction.  And so when they

report r2 values, which you have all heard r2, that is

a big number, right, like .8?  With that kind of

prediction, to me, that is the same as nothing.  Okay?

It's not statistically appropriate to make a big deal

of that, and I will just leave that alone.  Okay.  So

those are some small points.  So understate the

difference in variation.

            I don't believe we have evidence that

there are differences in anything with respect to

these subjective scales of goodness.  I worry a little

bit about the survival meta analysis.  Let me tell you

a couple of things why.

            In fact, we heard some interesting points

about the survival.  We said it is one percent per

year.  We heard that.  In fact, we saw some very

interesting documentation that that is about what it

is.  And, yet, the survival analysis itself, the

actual meta analysis, as near as I can tell, takes no

account of the individual follow‑up time in the

studies, as far as I can tell.

            I looked and tried to figure that out.  In

fact, there are some answers to some questions the FDA

raised about their analysis, and it looked like the

actual study length time is not accounted for.  And

survival should be related to time, yes?  But why

wasn't it related to time in this?  There's so much

variation.  You could find it's not statistically

related to time because of the amount of variation in

these studies.

            Did you see the side by side comparison

that was up there?  And you saw a whole bunch of

things.  Maybe you said, "I will tune out."  It was

kind of hard to see, but if you looked, a lot of the

data points were to the right, toward 100, right,

toward 100 percent survival.  But there were a few.

In fact, a couple were highlighted in red, right?  Do

you remember that?  There were a few that were on the

other side.

            These are very different values, and they

are all being combined.  They are all being combined.

This is the kind of variation that is being combined.

            So that is just a couple of minor points.

I don't know that there is no difference in survival.

Look, I am here criticizing.  My gut feeling is they

are probably pretty much the same.  Okay?  But that is

my gut feeling.  I am a statistician.  I am not

allowed to have gut feelings.  So ignore that comment.

            Let me make a couple of other comments.

And then I will be quiet.  There were some bootstrap

technical comments.  If bootstrapping is done, I would

rather do a randomization test.  That is a minor

detail.  Okay.  Leave that alone.

            One meta analysis that was not done that

I worried about is a meta analysis on adverse event

rates.  It looked like to me ‑‑ now, maybe you have

done something, but I didn't see a meta analysis of

adverse event rates.

            I think adverse event rates would be very

important, but there was also some information in my

adverse events with respect to some of the studies, at

least that I could tell.  And there was one group of

studies that had much higher adverse event rates, if

the medical people would help me understand why that

is true.  There were studies called PCL sacrificing;

that is, devices that do whatever that is.  Okay?

            They look to be much worse.  That's at

least from the data that was provided, much worse.

And, yet, when the meta analyses were put together, as

near as I can tell, they were lumped in with

everything else, as near as I could tell.  And, yet,

there is a variable or a kind of device or a kind of

study or whatever it is.  And I am just a

statistician.  So I don't know what.  They called it

PCL sacrificing.  Okay?  That, in fact, looked to be

a variable that was a very big effect on adverse event

rates, on survival, and so on.  But as far as I could

tell, that wasn't taken account of in the meta


            Now, all of that said, I think it is also

interesting that we actually only have three approved

PMAs for this device already.  That is not a very big

experience set.  I will just say that.  What I am

going to do is ‑‑ oh, you are all waiting for me to


            There is one last group of reporting that

actually is very minor, but when I do meta analysis,

I try to identify the specific studies that are in the

meta analyses, right, specific articles.  I will give

a good name, like it's done by Larntz, et al.  That

would be a good study, I am sure, a perfectly good

study.  Okay?

            But in the reporting, the number of the

reports were study 1, study 2, study 3, study 22.  For

instance, without great effort, ‑‑ and I didn't make

that great effort ‑‑ I can't tell which of those 22

studies were PCL sacrificing.  For instance, in that

reporting, I think when we do a meta analysis, it

behooves you to make sure you keep emphasizing the

source of your information.

            And the way you do that is by keeping

reminding people which of these studies are all the

labels.  When you do the 37 in the survivorship, 22 in

the first one, all of these should be labeled.  Minor

point in the end.

            I am not sure I said anything except there

is a lot of variation.  I think it is very difficult

to make any conclusions from the meta analysis.

Saying that everything is the same or statistically

saying that there is no difference is not the same as

saying there is no difference.

            I will be quiet.  Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr.

Larntz.  I will thank both Dr. Mayor and Dr. Larntz

for giving their reviews.

            What we are going to do now is I am going

to ask Dr. Witten to give some general comments from

the FDA's perspective on classification so that we all

understand what we are being asked to do.

            Then we are going to have a session ‑‑ and

I will suggest it be a short session ‑‑ on any general

comments before going specifically to the questions.

During either the general comment session or the

specific question session, any panel member may

request that FDA representatives or sponsor

representatives give an answer to specific questions

that they may have.

            Let's start now.  Dr. Witten, could I ask

that you make some comments about reclassification?

            DR. WITTEN:  Yes.  I just want to clarify.

Thank you.

            I just want to clarify what our goal is

here today.  Hopefully these were clarify it more and

not confuse it more.  Let me just say that when you

consider reclassifying a device type, you reclassify

based on the devices that FDA has approved.

            So the ones we would be reclassifying are

based on what we have approved.  But all of the other

designs that have been presented would be devices that

then potentially would be eligible for coming in under

this regulatory route of class II devices.

            So our discussion or what we are hoping

for from you is not really so much a discussion about

whether or not these devices are safe and effective

but whether or not we understand enough about the test

methodology and its ability to predict device

performance to be able to adequately control the risks

and regulate these devices.  So it's more the risks

and the test methods, rather than, are the devices

safe and effective.

            Maybe that was already clear.

            CHAIRPERSON YASZEMSKI:  Thank you.  Thank

you for the clarification.

            Let's go around the table one time with a

general discussion.  In a moment, I am going to ask,

Dr. Kirkpatrick, that we start with you.  I am going

to pose a topic for the discussion.  It seems that

through these presentations by the FDA and by the

sponsor, there does exist a question as to whether all

mobile bearing knee devices can be considered alike;

that is, those that have unidirectional motion versus

those that have multidirectional motion.

            I want to go around the table and ask

everybody what they think about that.  Should we

address all of these as a single class or might they

be separated?  Let's get that out before we start

answering the questions.

            Dr. Kirkpatrick?

            MEMBER KIRKPATRICK:  So you want to answer

that question before I ask a question of the sponsor

that may be related?

            CHAIRPERSON YASZEMSKI:  You decide how you

want to go, but I would like to hear your comments on

that sometime before we go to Dr. Mabrey.

            MEMBER KIRKPATRICK:  So your question

specifically was, are mobile bearing unicondylars the

same as mobile bearing total ‑‑

            CHAIRPERSON YASZEMSKI:  No.  Are any

mobile bearing devices in which there is only motion

type; i.e., rotation only, appropriate to be

considered together with all other types that may

rotate and translate and have multidirectional motion,

as we have heard from many of the presenters this

morning that sometimes the multidirectional mobile

bearing devices more are like a hip device in their

wear characteristics than a single bearing type mobile

bearing device?

            I would just like to have everybody

comment on that, in addition to their questions.  So

go ahead with your questions, but do comment on that.

            MEMBER KIRKPATRICK:  My comment on that

would be I think you have got to test in more than one

direction if the device allows more than one

direction.  I think that a device that allows more

than one direction is somewhat different than one that

just allows one direction.

            CHAIRPERSON YASZEMSKI:  Thank you.  That

is what I wanted.

            MEMBER KIRKPATRICK:  Now, I do have a

question for the sponsor if it's okay.  And I think it

relates somewhat to this.  I have to apologize.  I am

not quite as smart as the sponsor's people or perhaps

even the panel, but I am having a difficulty

understanding one concept with what you presented.  To

put it into some understanding, I have to relate to

something I can understand.

            I cover a football team in the fall.  We

were the state champions this past year, but we lost

our opening game against a cross‑town rival because on

two defensive plays, a single missed assignment by one

player on one play and another player on another play

resulted in two scores and we lost by those two


            Now, you have presented an extensive

amount of data.  However, you have eliminated two

devices that are both mobile bearing knees.  As I

understand, the Oxford one and the Accord were

excluded from your device analysis.  And then you take

after the elimination of that all of the remaining

data and say everything else is similar.

            My concern is that if we go with this

reclassification, then we won't be including some of

the devices that should be included.  Can you help me

understand the rationale for excluding those two

devices so that I can better put this into context?

            DR. STIEHL:  I think that is a very, very


            CHAIRPERSON YASZEMSKI:  Dr. Stiehl, excuse

me.  For the transcription, would you just state your


            DR. STIEHL:  Sure.  Dr. Jim Stiehl.

            As I understand the issue here, we are

trying to figure out, can we pick those two problems

out of this group and say, "All right.  These are

going to be problem implants.  You are going to have

to look at these implants very carefully"?

            There were issues with the Accord device

a number of years ago that made it unsuitable as a

mobile bearing device.  I mean, it is old.  I don't

even know what it looks like.  I have been told it

relates to the stability device and that sort of


            I think the challenge here for us is to

decide if we can pick out that particular device as a

problem device.  The Oxford I, for example, is a

unicondylar device.  They kind of sprayed it at first.

But then they figured out the device worked very, very

well if they had fairly stringent criteria of a

retained ACL, certain amounts of deformity not used on

the lateral side and that sort of thing.  So they

refined their experience with it so then they could

pick up what that problem was.  Then it is okay.  Now

you see a free market approval of that particular


            I think the challenge here is, can we look

at devices and know they are going to be okay or not?

That is really what this special controls issue is.

            We have an extraordinary amount of

knowledge about mobile bearing knees because they have

been out there for 30 years.  In the last five to ten

years, my colleagues have shown an extraordinary

amount of information about how total knees in general

work.  I mean, my area and Doug Dennis' has been in

the area of kinematic research.  We have learned

things looking at these mobile bearings that we didn't

know before, but they certainly will guide us in what

we will be thinking about in the future.

            For example, all mobile bearings aren't

alike.  I mean, there are certain features that cause

some concern, like this multidirectional issue.  Now,

it may be fine with the Corin knee, the Rotaglide.

The Rotaglide has an extraordinary and a long‑term

experience.  I mean, it is a good device.  The AP

Glide, on the other hand, where it can just slide

forward as far as it wants to go without any checks,

may be a problem.

            Soft tissue impingement really wasn't

recognized with any knee device of considerable

concern.  Now with mobile bearings, I think soft

tissue impingement is a concern.  So that special

control is a very important issue.

            And when we discussed our presentation

here before this panel, I was very adamant that that

particular concern has to be addressed as a special

control.  So we have to know that that is okay.

            I believe that we are close to knowing

pretty much how these knees work well.  We are

refining how we get surgeons to use these knees.  But

there is some ease with which we do this.

            I think in your experience, it would be

much easier to take a fixed bearing knee that you

understand very, very well and the technique and all

of that and then say, "Well, I am going to add a

mobile bearing to this device."  I mean, that is a

small increment of a step.  And you can understand how

that might work.

            That, in effect, is what has happened with

the PFC signal, which has a supplemental pre‑market

approval.  The surgeons that use the PFC signal can

essentially do that.  They can know their fixed

bearing technique very well.  Then they can add this

mobile bearing to that construct and make it work with

some ease.

            Other devices, it is not going to be as

easy to do, particularly the LCFs because they had a

unique technique.  It was a unique design.  And you

had no fallback if the mobile bearing didn't work.

            So these are issues that we understand.

And I honestly believe that I as a surgeon can pick up

a good mobile bearing when I see it and a bad mobile

bearing that I wouldn't like if I saw it.  And I think

the engineers have a lot of experience with testing to

show you some of these efficacy issues.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr.


            MEMBER KIRKPATRICK:  May I just follow up

with a quick question ‑‑


            MEMBER KIRKPATRICK:  ‑‑ of both the

sponsor and the FDA?  Can you provide me the specific

current standards that would have kicked out both the

Oxford one and the Accord?

            CHAIRPERSON YASZEMSKI:  Mr. Allen, do you

want to lead that off?  And I will ask somebody from

OSMA to be prepared.

            MR. ALLEN:  Pete Allen, FDA.

            As far as the Accord, as Dr. Stiehl said,

I am not real familiar with the design.  And I don't

know specifically what the issues were with that one.

As far as the early Oxford devices, as Dr. Stiehl

said, I think it was a learning curve of a design that

they discovered what some of the problems were as far

as the way they were indicating it for use.

            I can't think of any specific preclinical

tests that could have caught that ahead of time.

            MEMBER KIRKPATRICK:  So if I could

summarize my understanding, we do not have current

specific standards which would have identified the two

devices that OSMA has eliminated from their


            MR. ALLEN:  Not to my knowledge.

            MEMBER KIRKPATRICK:  Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks.

            Dr. Walker, would you like to comment on

that?  Thanks, Mr. Allen.

            DR. WALKER:  That is an excellent

question, which gets to the heart of testing methods.

In other words, the test method must eliminate the

potentially defective devices.  In fact, a history of

defective devices is a very good way to validate a


            In other words, a test that we would come

up with to look at bearing dislocation, for example,

or instability would have to fail the Accord and

appropriately fail the early version of the Oxford.

            For example, without writing the test or

without writing the standard, if you like, or writing

the guideline at this time, one could say just quickly

that any device which is unconstrained in all

directions, for example, using the existing constraint

test, any device which is unconstrained in all

directions should simply not be allowed.  I mean, that

would be a no.  I mean, it's that simple because we

have evidence of that, a device which is unconstrained

and so on.

            So I think that the test has to be

devised, as I said, using the body of knowledge we

already have.  And I think that the Oxford is a

special case.  It has been approved already, the PMA.

            The Oxford is a special case in the

following way, that, really, if it does not have any

other stability associated with it, it would

automatically just dislocate front to back.  But the

surgeons who have used it and, in fact, the

specifications have been very, very clear on

specifying the surgical technique, that it has to have

ACL/PCL and that the surgical technique has to be very

rigorous, it has to be restricted to the medial side,

which is tighter.  The lateral side is not allowed if

you notice.

            So, again, I think using the experience

like that, I think one can, one could quite easily

eliminate designs that are simply too unstable.

            MEMBER KIRKPATRICK:  Your first comment I

can't agree with because we just voted approvable a

device that has no constraint in any direction


            DR. WALKER:  Well, I can take your point.

I did make a caveat of saying, though, that, you see,

the Oxford in a way is unique in the sense that it is

not just a device.  It is combined with the ACL and

the PCL and the medial side.  I think in a way, that

is a special case that would have to be made


            So, you know, I don't think any test can

cover every conceivable variation.  I mean, I think we

are kidding ourselves if we say that.

            MEMBER KIRKPATRICK:  My concern as a panel

member is, can we establish and define a special

control today that would cover those two devices?

            DR. WALKER:  I would say yes.  If we sat

down and we carefully documented it, I believe we


            MEMBER KIRKPATRICK:  But it does not exist

at this point?

            DR. WALKER:  It exists in the minds of

individuals who would have put it together.

            MEMBER KIRKPATRICK:  I think we will just

go back and forth if I continue.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr.


            Dr. Mabrey?

            MR. MAISLIN:  I just wanted a


            CHAIRPERSON YASZEMSKI:  Mr. Maislin, come

up.  Yes?

            MR. MAISLIN:  I just wanted to clarify

that.  And it doesn't go to your main point, which is

very well‑taken, that, in fact, the summaries that

were presented did not exclude those two.

            The analysis, the summary analysis, that

was on the first slide and the last slide included all

of the mobile bearings.  Analysis was repeated

excluding those, but those weren't highlighted.  The

summary statistics that were presented did include the

two devices that are now obsolete.

            CHAIRPERSON YASZEMSKI:  Thanks, Mr.


            MEMBER KIRKPATRICK:  That's my mistake if

I ‑‑

            MR. MAISLIN:  Right.

            MEMBER KIRKPATRICK:  If it was only for

certain parts of your analysis since you excluded

those, I apologize.

            MR. MAISLIN:  Yes, yes.  The 90‑some

percent success rate included them.  There was one

slide that excluded them as a secondary analysis.  And

the success rate increased to about 93 percent.  But

to be conservative, I included in my main presentation

all of the data that was available.

            CHAIRPERSON YASZEMSKI:  Thanks, Mr.

Maislin.  Sir?

            DR. EMERSON:  Yes.  My name is Roger

Emerson.  I am an orthopedic surgeon.  I have come

here with Biomed.  I am familiar with the Oxford knee

and participated in the original IDE from this


            Just a point of clarification.  The design

of the Oxford has not changed that much over the three

phases, but the instrumentation, the understanding of

how to implement mobile bearing has become more

sophisticated.  So it was appreciated very quickly

that the phase I while the design could work was

unpredictable.  And precision had to be added.

            Basically, the phase II involved a change

in instrumentation that allowed incremental balancing

of the soft tissues and then an appreciation of the

role of the ACL in the stability of the

unicompartmental knee.

            CHAIRPERSON YASZEMSKI:  Thank you.

            MEMBER LARNTZ:  May I follow up?

            CHAIRPERSON YASZEMSKI:  Sure.  Go ahead,

Dr. Larntz.

            MEMBER LARNTZ:  I was just going to follow

up on your question.  At what point did you realize it

was unstable?  I am just asking the last, following up

on the last, comment.  You said it was clear it was

unstable.  At what point did you realize that?  When

it was in the lab?  When it was in the knee?  After


            DR. EMERSON:  Well, this was in the 1970s.

            MEMBER LARNTZ:  Right, sure.

            DR. EMERSON:  So it was a different era.

And there were being implanted both as a

bicompartmental and as a unicompartmental implant.  In

the bicompartmental situation, you are juggling both

sides of the knee.

            Basically, the survivorship of the

bicompartmental was in the range of 80 percent.

Survivorship of the unicompartmental back then was in

the range of the high 80s and 90 percent.

            MEMBER LARNTZ:  Excuse me.  At what time

period ‑‑

            DR. EMERSON:  This was at a time when we

didn't understand soft tissue balancing.

            MEMBER LARNTZ:  You said 80 percent.  Is

that 80 percent after a month, 80 percent after 5


            DR. EMERSON:  The implant came out in

1976.  And the phase II came out in the mid '80s.  So

it was over a ten‑year period.

            MEMBER LARNTZ:  So over a ten‑year period,

you realized it was unstable?

            DR. EMERSON:  Yes.  The survivorship ‑‑

            MEMBER LARNTZ:  That's what I need to


            DR. EMERSON:  Yes.

            CHAIRPERSON YASZEMSKI:  Okay.  Thanks very


            Let's move on, Dr. Mabrey.

            MEMBER MABREY:  Just for clarification,

you are asking me, do I think there is a difference

between mobile bearing knees and fixed?

            CHAIRPERSON YASZEMSKI:  No, that is not

what I am asking.  When we get to ‑‑

            MEMBER MABREY:  Okay.  The two ‑‑

            CHAIRPERSON YASZEMSKI:  When we get to

voting, is it appropriate in your opinion to vote for

all mobile bearing designs as a general design that

the FDA would then consider whether to allow as class

II's or might there be distinctions between mobile

bearing designs that we might recommend to the FDA

that we separate?  That's what I would like to hear

from you.

            MEMBER MABREY:  Okay.  I think in answer

to that, you have to take into account that these are

not just devices.  These are systems.  And, as some of

the previous speakers have pointed out, the actual

design of the implant may be ideal.  And it could have

performed quite well in the lab.  Yet, once it was

implemented with the early instrumentation, it proved

not to be so feasible.

            So to say that one design is equal to

another or essentially equivalent, you have to take

into account the history of those devices, the history

of the designers, and I would say that there are

subtle differences as far as the types of wear that

might be generated.  But I have to say that from my

personal experience, I haven't seen a clinical

difference between them.

            CHAIRPERSON YASZEMSKI:  Specifically, do

you think that we could recommend to FDA special

controls all of the possible types of mobile bearing


            MEMBER MABREY:  I think that might be

difficult based upon what we know today because we

need to anticipate all possible future



            MEMBER MABREY:  I am not supposed to say

that, am I?  Okay.  It cut me off.  I think that

getting back to Earth here, to anticipate current

designs, problems with current designs, I believe that

there are adequate tests available now that could pick

up gross problems with that design.  I don't think

that we can ever predict how it will perform

clinically until it's actually been in use for some


            CHAIRPERSON YASZEMSKI:  Thank you.  Other

comments, Dr. Mabrey?

            MEMBER MABREY:  No.

            CHAIRPERSON YASZEMSKI:  Dr. Finnegan?

            MEMBER FINNEGAN:  In answer to your

question, yes, I do think we probably do need to split

some things out.

            I have three questions.  The first

question is for Mr. Maislin.  On your scales, you

spoke about how you picked the global rating scale

that you used.  You sort of nicely skipped over the

fact that there is a statistically significant

difference between the two groups that you compared in

the percent global rating scale improvement.  It was

100 percent in the fixed.  It was, I believe, 91.4

percent in the mobile.  Can you explain that?

            MR. MAISLIN:  Yes.  That difference is

actually quite small.  It is not 100 percent of

patients and 91 percent of patients.  It's the percent

improvement.  It's as if there is a score that goes

from 45 to 90.  And that would be a 100 percent

increase or 45 to 87.  And that would be a 91 percent

increase.  And in those two cases, the clinical

significance of an increase from 45 to 87 is probably

trivial compared to one that went from 45 to 90.

            MEMBER FINNEGAN:  Did you have range or

median?  Because I noticed you start it.  So I'm

assuming that meant there was some statistical


            MR. MAISLIN:  No, no.  The table was

repeated from the reclassification petition as is.

The star indicated that that was raw data, that there

wasn't ‑‑ I remember the subject test indicated that

it was raw data.  It wasn't an indication of any

statistical significance.

            MEMBER FINNEGAN:  Did you have a range or

a median for either group?

            MR. MAISLIN:  I don't know what those

values are offhand.  The actual data is in the

reclassification petition, but I don't know.  I assume

that it's as variable as Dr. Larntz might suggest, and

it was probably comparable in variability to the

percent improvements.

            MEMBER FINNEGAN:  Okay.  Thank you.

            The next question is for Dr. Stiehl.  You

talked about in one of your slides the unis had a 6.8

percent revision rate at four years.  Can you ‑‑

            DR. STIEHL:  That was the phase II uni

experience that I guess has already had a PMA

approval.  That's the device that we're talking about.

            MEMBER FINNEGAN:  Right.  And was that

related to age?  Was it related to mechanical axis?

Do you have any idea what the cause of that was?

            DR. STIEHL:  I can only refer to Roger

behind me because he probably knows, but I think the

failure rate with that particular series relates to

its unis.  And they failed earlier because they get

lateral compartment disease and that sort of thing.

I honestly don't know what their results were related


            MEMBER FINNEGAN:  So, Dr. Emerson, was

this a pre‑PMA IDE or was this part of the PMA IDE?

            DR. EMERSON:  That was all in the PMA for

the original, for the phase II Oxford that has been

approved.  But the two figures were given.  One was

the figure for all revisions for any reason.  That

included lateral compartment disease, patella/femoral

problems, inflammatory issues.  And then there was a

statement about the percentage failure for a revision

rate simply for bearing‑related issues, which was the

smaller number.

            MEMBER FINNEGAN:  Correct.  But for all of

your knees, it was 6.8.

            DR. EMERSON:  That 6.8 was actually

reported at 2 years.  There were 8 out of 117.  It was

two‑year data.  Also in the petition, there is data

from two to eight years.  And the figure there is a

4.8 percent revision rate for bearing‑related problems

or implant‑related problems of any kind, excluding

lateral compartment, patella/femoral, inflammatory


            MEMBER FINNEGAN:  What is the whole for


            DR. EMERSON:  4.8 percent for the IDE out

to 8 years for bearing‑related.

            MEMBER FINNEGAN:  Right.

            DR. EMERSON:  15.7 percent.

            MEMBER FINNEGAN:  For total?

            DR. EMERSON:  For a total.

            MEMBER FINNEGAN:  Thank you.

            And then my general question to the

sponsor is Dr. Larntz brings up a wonderful point.

Can you pull out the PCL sacrificing problems?  And

are you aware of them?  And do you have any comments

on that?

            CHAIRPERSON YASZEMSKI:  Dr. Stiehl?

            DR. STIEHL:  Yes, Jim Stiehl back.

            Dr. Larntz, I think you are referring to

the Callaghan meta analysis that looked at PCL

sacrificing as being a problem area, as I recall.

            MEMBER LARNTZ:  I thought it was among

your 22 studies.  There were five of them that were

PCL sacrificing, if I remember right.  I mean, I was

looking at your table.  I think of the 22 studies, I

think there are mobile bearing studies.  There are 22

studies.  I will find the page in just a second.  I

will tell you.

            DR. STIEHL:  Well, the issue really is the

rotating platform, which from my knowledge, at least

of the LCS has been the standard of this standard

concept, is a PCL sacrificing system.

            Now, as I say, I'm aware in the Callaghan

study, they definitely flagged PCL sacrifice as a

lower number.

            MEMBER LARNTZ:  This is in your meta

analysis of your 22 studies.  It says 14 are PCL

sparing, 5 are PCL sacrificing.  I am on page 282 in

the document I have anyway if people want to find it,

282 in volume I.  And three are mixed prostheses.

            For instance, numbers that are there,

revision rates are 5.6 for the sparing and 9.2 for the

sacrificing.  Mechanical failure is 1.5 for the

sparing and 5.2 for the sacrificing.  There are

revisions for bearing location break and subtraction,

zero for the sparing, 2.8 for the sacrificing.

            This is your table in your analyses.  And,

as far as I know, nothing was sorted out with respect

to that covariate in the meta analyses except your

report.  Like I say, your report gave me information.

I couldn't have found this myself because I couldn't

have delved through all of that.  So it seems to me

there is something going on, at least with respect to

those studies.

            DR. STIEHL:  I am going to have to ‑‑

            CHAIRPERSON YASZEMSKI:  Does that satisfy

your question, Dr. Finnegan?

            MEMBER FINNEGAN:  Yes.

            CHAIRPERSON YASZEMSKI:  All right.

Thanks, Dr. Stiehl.

            Let's come back.  Let's wait until we come

to Dr. Larntz.  Then we'll have that discussion again.

Let's go to Dr. Kim.

            MEMBER KIM:  I'm going to try to address

the question that you asked, which is so that I

understand it.  These devices, can they be all grouped

together into a single ‑‑

            CHAIRPERSON YASZEMSKI:  Right.  Can we

adequately suggest special controls that will cover

all mobile bearing devices?

            MEMBER KIM:  That depends on the answer to

whether or not we can get special controls because in

my mind, I don't understand enough about the various

intricacies of the mobile systems to be able to

separate out one mobile system to another.

            So my feeling is that I think it is okay

to jump them all together.  What we need to work on is

whether or not we can devise special controls to look

at all of the special design, all of the different


            CHAIRPERSON YASZEMSKI:  Okay.  Thank you.

            Dr. Naidu?

            MEMBER NAIDU:  I'm not sure we can lump

them all together.  I think inherently these devices

are different systems, just like Dr. Mabrey has

referred to.  I think these devices have to be

addressed as a system.  I think they are all

inherently different.

            And from the presentations of the sponsor,

some of the multidirectional platforms' survival rates

are as low as 75 percent at 5 years.  And all of the

literature that Dr. Stiehl refers to, it seems like it

points to the LCS.

            So I think there is an inherent difference

between the two.  I am not sure that we could come up

with the special controls document.

            CHAIRPERSON YASZEMSKI:  Okay.  Thanks, Dr.


            Dr. Mayor?

            MEMBER MAYOR:  I'm going to suggest

something that might sound a little semantically

picky, but it comes from my experience with the ASTM

documents with relate to constraint.

            The documents that we have been reviewing

identify a class of devices which have no linkage

across the joint.  I would suggest we either change

that to these devices have no across the joint


            I recognize that some of you may think

that there is no difference in those two wordings, but

my feeling is that, first of all, it would be more

compatible with the ASTM documents as I remember them

contained.  And the term across the joint linkage is

more specific to a mechanical design which binds the

two parts together; whereas, linkage across the joint

could be interpreted more loosely and could be taken

to imply other kinds of mechanical interactions

between the two parts.  It's a small but I think

useful point.

            The other thing about the definitions is

that in relationship to the tricompartmentals, there

isn't a specific mention of whether or not the

patellar interface is going to be mobile bearing or


            I think the option needs to be included in

terms of the device design that the patella could

include a mobile bearing or might not.  So the

patellar devices that the designer might elect are

identified as applicable in the reclassification if,

in fact, that comes to be true.

            My sense is that issues of dislocation and

subluxation can probably be adequately addressed with

careful attention to the patient in certain variables,

the soft tissue issues that have been cited and

discussed fairly extensively in the presentations from

both sides of the issue.

            I think in terms of wear phenomena, the

thing I see as serious and missing in the discourse is

the variations in polyethylene, that we know from our

experience with all of the implants that have been

studied, that polyethylene behaves differently

depending on the resin from which it is derived, the

processing used to bring it to its final form, and the

steps that are taken subsequent to that to sterilize

it and store it.

            Without some acknowledgement of the impact

on survivability of these implant components

addressing the issue of polyethylene variability, I

think we have missed an opportunity to protect the

public from the kinds of things that we now begin to

understand better and better as they relate to

polyethylene and its variability.

            The labeling I think needs to be looked at

carefully because of the issues we have already

discussed in terms of predictability of outcome.  And

that includes issues related to both the patient and

the surgeon involved in the interaction.

            As regards configuration and the

subconfigurations, can they be adequately controlled

by special controls and the guidance documents?  I

think they can.  I think we need to discuss in some

detail how that might be done and how to address that

extra step that needs to be taken to make sure that we

are responsible in our deliberations.

            In response to Dr. Yaszemski's specific

question, do I think that wear issues need to be

addressed differently for "unidirectional" and

multidirectional mobility in bearing design, no.  I

think we can address with a fairly monolithic document

the issues that relate to both of those.

            CHAIRPERSON YASZEMSKI:  Thanks very much,

Dr. Mayor.

            Dr. Larntz?

            MEMBER LARNTZ:  On your primary question,

I don't have any evidence to give an opinion.  Do you

want me to follow up with these people and give them


            CHAIRPERSON YASZEMSKI:  At your


            MEMBER LARNTZ:  To talk to me about page

282?  We can do that for ‑‑ do you have a timer on? ‑‑

no more than three minutes.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr.


            MEMBER LARNTZ:  I will let you respond to

what is on page 282.

            CHAIRPERSON YASZEMSKI:  Mr. Maislin?

            MR. MAISLIN:  Thank you.

            I want to point out one item that the

stratification by PCL status was not something that

was uncovered.  It was an attempt by the investigator

who implemented this meta analysis to follow


            And since Callaghan did it, they did it,

just to let you know why it was stratified.  In fact,

the later analysis didn't emphasize it because it

wasn't something that was interesting.  It appears

primarily because they were following a recipe of an

analysis that was published.

            MEMBER LARNTZ:  And I think I understood

that.  I'm not sure that is always the best thing to


            MR. MAISLIN:  I agree.

            MEMBER LARNTZ:  But let me just point out

‑‑ and I think this is what Dr. Finnegan might want to

look at on page 282.  I mean, complication rate in

knees, any complication rate, 5 percent with sparing

and 14.1 percent, now, I don't know if those are

significant because there is tons of variation here.

There is tons of variation.

            But I am saying it seems to me ‑‑ and I

didn't have the comparable numbers from the Callaghan

fixed to compare that to, I don't think.  I don't

think I have that here.

            MR. MAISLIN:  I don't know.

            MEMBER LARNTZ:  I think this is your

study.  These are your man‑hours.  So there are not

comparable numbers.  I know if Callaghan found that in

the fix, then I think it would be important to think

about that kind of stratification here and kind of

comparison.  I didn't see that kind of comparison.

            So, anyway, I understand where it came

from.  I understand following the recipe.  I

understand all of that.  But in that sense, there is

no adjustment for that covariate except for these


            MR. MAISLIN:  Right.  The one mitigating

factor, just if I can respond, ‑‑ and I have the

tables from a preliminary manuscript.  So it doesn't

line up.  But it's the same tables.

            The proportion with good or excellent in

those 14 PCL sparing and 5 was 92.1 percent and 85.5.

There was a little bit less.  And the revision rate

was 5.6 vs. 9.2, 14 studies vs. 5 studies.

            Let me just point out that those two

studies that were identified as obsolete, one was in

the 5 and one was in the 14.  So that there is more

weight to the bad studies in the PCL sacrificing.

            MEMBER LARNTZ:  Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks, Mr.


            Dr. Walker, please?  Briefly, please.

            DR. WALKER:  I'll be very brief.  Just a

brief comment, particularly to Dr. Mayor's comment

about tests being monolithic and the words apply to

all devices.

            Yes.  I think we should not lose sight of

the fact that tests if they are any good should be

able to distinguish between different kinds of

devices.  I mean, if we just restrict ourselves to one

that we know they work, then the testing won't tell us

much extra.

            The purpose of the test, in fact, is to

separate the sheep from the goats, if you like.  It is

to identify designs which do not perform


            So in that sense, I am somewhat

comfortable about testing different kinds of designs

because I believe their deficiencies will be revealed

in the tests as in the special control guidance


            CHAIRPERSON YASZEMSKI:  Right.  Thank you,


            Dr. Besser?

            MEMBER BESSER:  A certain part of the

discussion so far seems to be concerned with whether

the success rate of the mobile bearing knees is the

same or different than the success rate of the fixed

bearing knees.  That is really I don't think the

question that is in front of us.

            I think, even if they are equal, the

question is whether we can decide whether we can

control manufacture of or designs of mobile bearing

knees using special controls, as opposed to going

through the PMA process.

            A lot of the data that has been presented

seems to be on, one, the LCS knee.  I am reminded of

the old Audi commercial that was very proud that 8 of

10 Audis that were on the road 15 years ago were still

on the road; whereas, 9 of those 10 Audis had been

sold in the last 5 years because Audis had just been

introduced in the U.S. market at the time.

            I think that we can come up with special

controls that would address wear aspects for different

configurations of a mobile bearing knee.  As long as

special controls can be written, if you'll allow me,

vaguely enough that we would like to address motion,

multi‑access motions, in all directions possible given

a specific design, I think that can address the wear

issues for a rotating vs. rotating and translating

design.  I don't think that's a problem.

            And I imagine that for all of these from

the standpoint of a mechanical engineer, we can test

the heck out of them and definitely eliminate some

that are destined to fail.  However, there will be new

and interesting manners of failure for any innovative

design that won't be discovered until after they are

implanted, unfortunately.

            CHAIRPERSON YASZEMSKI:  Thank you, Dr.


            Ms. Maher?

            MEMBER MAHER:  I actually don't have much

to say on your first question.  However, I just want

to remind the panel that we are looking at the devices

that are currently on the market.  And the FDA has a

strong skill set at looking at a 510K and determining

whether something is substantially equivalent or

whether the new changes and nuances of it fall and

kick it over into the class III product.  And that is

something that Celia and her group, Dr. Witten and her

group, have a very strong skill set at doing.  So we

need to look at the data we have now and make


            CHAIRPERSON YASZEMSKI:  Thanks, Ms. Maher.

I will emphasize that point that if we recommend a

general down classification of mobile bearing knees,

the decision still has to be made in each individual

application by the FDA as to whether the product that

is being considered is substantially equivalent and

falls into that general class II classification.

Thank you for that clarification.

            Dr. Doyle?

            MEMBER DOYLE:  I guess I am filtering all

of the scientific evidence that has been offered here

through my personal consumer looking forward to

possibly having knee operations.

            I think the thing that concerns me the

most is basically no, I don't think all of the devices

are alike because otherwise they wouldn't have

designed them.  I think each design is different to

hopefully have some sort of improvements.  The

question I think is whether they are similar enough.

            I think what bothered me was talking about

the Oxford device that we don't have any special

controls that would have picked up a device, if I

understood correctly, that was not, that basically did

have a rather large fault.

            So looking down the road, I am a little

concerned.  And I do think that they are different

enough that they may have to be considered separately,

even those that have been looked at and even with the

constraints that are available.

            CHAIRPERSON YASZEMSKI:  Right.  Thank you,

Dr. Doyle.

            We have had a rather thorough discussion

in this preliminary phase prior to looking at the

questions.  So we are going to go to the questions

now.  I will ask that we consider each question

independently and try to not redo all of the things we

have done but to add any new information.

            Having said that, I want everybody to have

an opportunity to say what they think about the

questions.  We will start with question number one.

Mr. Melkerson, can you put question number one up?  We

will read question number one.  I will start with Dr.

Kim on question number one, and we will come clockwise

to Dr. Naidu next.

            Question number one is, do you believe the

proposed classification definitions for the following

device configurations recommended for reclassification

adequately describe the devices?  If not, what changes

in the definitions do you recommend for both total

mobile bearing knee prostheses and unicompartmental

mobile bearing knee prostheses?  Dr. Kim?

            Does anyone want the description?

Remember, Mr. Allen put those descriptions up.  If

anybody would like them, Dr. Witten will pass them

around.  Identify yourself and ask for them.

            MEMBER KIM:  The definition that was

proposed is broad for both the total and the

unicompartmental knees.  Thus, it allows for a wide

range of design features.  This is obviously desirable

if one wants to foster innovation.  However, it makes

the testing and implementation of special controls

much more difficult.

            So I don't have an exact answer for this

question.  I have more of a question for the question,

which is that an acceptable definition will depend on

our ability to have adequate special controls to test

the wide variety of design features that would be

allowed under this current definition.

            My gut feeling is that with some effort,

we could adopt those special controls.  And I don't

see any advantage in subcategorizing the various

design features.

            One change that I would recommend,

according to what was stated previously by one of the

panel members, that the patellar device was not

adequately described.  We need to address that issue

in the definition.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr. Kim.

            Dr. Naidu?

            MEMBER NAIDU:  Yes.  I think the

definitions are broad enough and all‑inclusive, and I

think they do adequately describe for both the total

and the uni knees.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr. Naidu.

            Dr. Mayor, your thoughts on question

number one?

            MEMBER MAYOR:  Basically consistent with

what I mentioned earlier in my earlier remarks, that

the issue of wording regarding joint linkage and the

issue of patellar design, being with or without mobile


            CHAIRPERSON YASZEMSKI:  Thanks, Dr. Mayor.

            Dr. Larntz?

            MEMBER LARNTZ:  I think definitions are

fine with the change Dr. Mayor said.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr.


            Dr. Besser?

            MEMBER BESSER:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you, Dr.


            Ms. Maher?

            MEMBER MAHER:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Doyle?

            MEMBER DOYLE:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Kirkpatrick?

            MEMBER KIRKPATRICK:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Mabrey?

            MEMBER MABREY:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Finnegan?

            MEMBER FINNEGAN:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Witten, the panel has considered

question number one.  In general, they feel that the

definitions are broad and cover the devices that may

come under this classification.

            There were a few concerns.  They included

the fact that the patellar device needs more

clarification with the wording referring to whether it

is mobile or not and additional wording regarding

joint loading.

            Have we adequately discussed this to FDA's


            DR. WITTEN:  Yes.  Thank you.

            CHAIRPERSON YASZEMSKI:  You're very


            Question two, Mr. Melkerson?  Question

two, do you believe the risks to health of the

following device configurations proposed for

reclassification are adequately described?  If not,

what additional risks do you believe should be

included for both the total mobile bearing and

unicompartmental?  Let's start with Dr. Finnegan this


            MEMBER FINNEGAN:  I'm going to actually

break this down into three parts.  I believe that the

total bearing knee prosthesis, which is

unidirectional, in fact, has been adequately described

and the risks to health have been adequately


            The second part I would like to say is

that the multidirectional total bearing knee

prosthesis I believe has been relatively

well‑described.  I think that there are, as has been

voiced, some risks to health that probably need to be

looked at in some extra special controls.

            My caveat for both of those is that I have

a feeling that the patellar sacrificing may have some

problems that have not been addressed by anyone.  And

I do think that probably needs to be either looked at

through the literature cases and/or looked at in the


            And the third component is

unicompartmental mobile bearing knee prosthesis.  I do

not think that the risks to health have been

adequately described.  The PMA is very young for the

most recent ones.  The indications for that one are

significantly different from the one that has been on

the market for a long time.

            The revision rates are high, as one would

expect.  And I think there is not enough literature.

There is not enough long‑term follow‑up.  And I

believe that that should be looked at separately.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr.

Finnegan.  May I ask for one point of clarification?

You said the "patellar sacrificing" were a special

case.  Did you mean patellar sacrificing?

            MEMBER FINNEGAN:  I'm sorry.  PCL


            CHAIRPERSON YASZEMSKI:  PCL sacrificing?

            MEMBER FINNEGAN:  Yes.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Kim?

            MEMBER KIM:  I have nothing to add to Dr.

Finnegan's comments.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr. Kim.

            Dr. Naidu?

            MEMBER NAIDU:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Mayor?

            MEMBER MAYOR:  The only other issue that

I think needs to be identified is that there are

actually three classes of knee replacement that have

been spoken of in the literature and in common

practice:  PCL retaining, PCL sacrificing, and PCL


            While I can't cite the statistical basis

on which the impression exists, in my own mind I think

there are some differences between those devices which

are truly PCL sacrificing, which the design may not

address the question of what the PCL function is

supposed to be, and designs which actually provide PCL


            Otherwise no additional items to add.

            CHAIRPERSON YASZEMSKI:  Thank you, Dr.


            Dr. Larntz?

            MEMBER LARNTZ:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Besser?

            MEMBER BESSER:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Ms. Maher?

            MEMBER MAHER:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Doyle?

            MEMBER DOYLE:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Kirkpatrick?

            MEMBER KIRKPATRICK:  I agree with what has

been said so far.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Mabrey?

            MEMBER MABREY:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Witten, the panel has considered

question number two.  They feel that in general, the

completeness of the risks to health with respect to

unidirectional mobile total bearing knees are okay;

that the multidirectional case perhaps needs

additional special controls, which has been a subject

of the discussion up to this point; that perhaps

unicompartmentals may need to be separated out and

needs further consideration before being included with

all mobile bearing knees.

            Have we adequately discussed this?

            DR. WITTEN:  Yes.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr.


            Let's move on to question number three.

Question number three, special controls have been

proposed to address the risks to health identified in

each of the above device configurations and all

related sub‑configurations.  Please respond to the

following questions regarding specific risks and/or

specific controls.  And there are several subparts.

            A)  Dislocation and subluxation of mobile

bearing components have been cited as common

complications in the literature.  Do you believe

appropriate special controls have been identified to

adequately address these risks?  And if not, what

additional controls would you recommend?

            Subpart B)  A reduction in wear is often

cited as a theoretical advantage of mobile bearing

knees over fixed bearing devices.  However, this has

not been consistently demonstrated clinically, and it

is not clear how well preclinical wear testing of

mobile bearing knees correlates to the clinical


            In addition, the potential for third body

wear appears greater and the potential for the amount

of third body wear also appears to be greater.

Currently, the state of development of knee simulator

wear testing has not yet been standardized or

clinically validated across all device types and,

therefore, may not be applicable for all of the

various mobile bearing knee types identified in the


            In light of the fact that wear appears to

be, in part, design‑dependent, do you believe

appropriate controls have been identified to

adequately address the risk of wear for the various

mobile bearing knee designs under consideration?  If

not, what additional controls do you recommend?

            Subpart C)  Although labeling has been

cited as a control with which to address risks to

health, the proposed labeling requirements are

consistent with those generally found in current fixed

bearing total and unicompartmental knee package

labeling.  Labeling typically includes device

description, type of material, indications for use,

contraindications, adverse events, precautions,

warnings, a listing of compatible components, and

sterility information.  What additional testing, if

any, do you recommend for these mobile bearing knee


            And part D)  Do you believe appropriate

special controls have been identified to adequately

address the risks to health for each of the above

device configurations and all sub‑configurations?  If

not, what additional special controls do you


            The summary of this is for each of parts

A through D, do special controls exist?  And if not,

which ones need to be specified?  Let's start with Dr.


            MEMBER MABREY:  Yes.  Thank you.

            I'll take those one at a time beginning

with A, dislocation and subluxation of mobile bearing

components have been cited as common complications in

the literature.

            I would contest that while these are the

most common complications, they are not common

complications.  They occur infrequently.  And those

have been identified primarily as a result of errors

in technique in many cases.

            At this point, I believe that there are

adequate controls to identify the inherent mechanical

problems within the device itself to address the risk

of dislocation of the mobile bearing.

            With respect to B, a reduction in wear is

often cited as a theoretical advantage, I agree at

this point it is theoretical, although our ability to

isolate wear debris and characterize it wear debris

has improved significantly over the last few years.

And I believe that those techniques are readily

available to the sponsors and should be employed in

the characterization of debris from their devices.

            With regards to the different types of

devices, unidirectional versus multidirectional, I

think that should be an important component of the

testing procedure looking at both unidirectional and

multidirectional wear patterns within the knee.

            Regarding labeling cited as a control

addressing the risk to health and the proposed

labeling requirements, I believe it's very difficult

to legislate or regulate against incompetence.

            We all know that these devices are

technique‑dependent.  I think it is imperative that

the implanting surgeon be familiar with the technique

and familiar with total knee replacement before he or

she even attempts that.  I think the labeling and

recommendations within the packaging are appropriate.

            Finally, D) Do you believe appropriate

special controls have been identified to adequately

address the risks to health for each of the above

device configurations and sub‑configurations?  And my

answer to that would be yes based upon the special

controls guidance document that we have been presented

with and also based upon data presented here by the

sponsors that those controls are available and


            CHAIRPERSON YASZEMSKI:  Thanks, Dr.


            Dr. Finnegan?

            MEMBER FINNEGAN:  I'm not going to add a

whole lot more.  I think for A, training probably

recommended at least is a necessity.  However, I do

think in B ‑‑ and we had a wonderful presentation by

Dr. Walker, but he kept sort of suggesting that maybe

there are some new tests that need to be developed.

And I think he is right for multidirectional wear,

that probably there needs to be some work done and

some new testing materials.

            I don't have any comments on the labeling.

And I think the risks to health we addressed in the

previous question.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr.


            Dr. Kim?

            MEMBER KIM:  I don't have any comments on

A or C, but I want to echo the comments of Dr.

Finnegan on B because it sounds like there are not

adequate special control systems to test all of the

different design configurations.  Dr. Walker actually

admitted that these exist in their minds but it's not

on paper yet.

            I would want some proof that a real

adequate special control design can be formulated in

this regard.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr. Kim?

            Dr. Naidu?

            MEMBER NAIDU:  Nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Mayor?

            MEMBER MAYOR:  I would simply cite the

earlier observations I made in my previous comments

and also add that while the goal of absence of risk is

unattainable, I think we can achieve a desirable

reduction of risk.

            The other observation is to identify the

experience that we have had with standards applicable

to materials that both the ISO and the ASTM standards,

where materials are standardized, set a floor below

which these materials should not fall without

achieving what may be an even more desirable goal of

identifying materials' qualities that may result in

optimal performance.  I think that is an important

issue to bear in mind.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr. Mayor.

            Dr. Larntz?

            MEMBER LARNTZ:  Well, I'm not sure how to

address the individual questions.  I will just say

that I think that from what I can see and the

variation in devices, the variation results, it is

going to be very difficult for me to think that we

could adequately address issues for these devices

without implantation.

            I think that we are going to have to have

a clinical study.  I realize clinical studies can be

part of special controls, but I don't think they were

mentioned in the special controls proposed.

            I do believe that a clinical study is

going to be necessary for new designs because I think

it is just going to be impossible given the way we

have seen.  I mean, the history is such that I just

can't imagine without a clinical study with some

reasonable follow‑up time.

            Typically I think that would stay as a

class III and go into it under an IDE.  But I realize

you can do clinical studies and require them as part

of this process.  So if a special control could be put

together that included a clinical study, then I would

guess that is what I would do.

            I think the mechanical, preclinical, all

of that testing seems to be amazingly good but unable

to identify what happens until in final implantation.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr.


            Dr. Besser?

            MEMBER BESSER:  I would echo Dr. Larntz's

comments.  I had been working with the assumption that

there would be clinical trials for these.  However,

that is not part of the class II requirements.  So

yes, I do believe that, in fact, the

dislocation/subluxation where can be handled by

special controls, the special controls that are

currently in place.  Plus, I would add some language

for the wear that would require multiple modes of

motion to be tested for wear at the same time.

            Don't do three separate tests, one for

translation, one for rotation, one for translation,

and another access, but to combine those during wear

testing and definitely would require clinical data

before approving device for market.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr.


            Ms. Maher?

            MEMBER MAHER:  I would follow up with what

Dr. Besser said.  I think the correct term would be

510k's may require clinical data, as opposed to a

clinical study, to allow the FDA and the sponsor of

the 510k's to determine what would be adequate if that

is what we are doing.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Doyle?

            MEMBER DOYLE:  I agree with everything

that has been said, particularly the emphasis on

clinical data, because I think studying something

under laboratory conditions that are ideal is very

different from seeing how it works in a 250‑pound man

who is clumsy.

            CHAIRPERSON YASZEMSKI:  Thank you, Dr.


            Dr. Kirkpatrick?

            MEMBER KIRKPATRICK:  I agree with Dr.

Finnegan on A that training is essential.  I think

special controls could be involved by adding it to C,

which is labeling; in other words, restricting it to

people who have been adequately trained if it is a

device that is that specific, as we have heard that

several are.

            I would also suggest that another control

on the training and insurance of adequate technique

can be restriction of the device to people that have

been trained, as we heard yesterday.  I think that

would be another option for the FDA to negotiate with

the companies on, a special control for that.

            As far as reduction in wear, I think that

we can establish special controls that can apply.  As

we heard, there may be or there is a concept of a

joint simulator test, which might pick up some of

these things.  If that is developed, obviously I think

FDA would automatically include that in their


            In the absence of that, I think a

post‑market analysis of dislocations of wear and of

any retrievals possible, although we can't mandate the

companies to get all of those retrievals, we can

certainly ask them to keep a very close eye on what is

published and any concerns that come into them so that

they can give that feedback as well.

            I also think on D that I agree with Dr.

Finnegan that it sounds like we need to separate out

the unicondylars in the totals as different devices.

            Thank you.

            CHAIRPERSON YASZEMSKI:  Thank you, Dr.


            Dr. Witten, may I ask you for

clarification from the FDA regarding clinical data in

class II devices because several members commented on

that and seemed to express some uncertainty about the


            DR. WITTEN:  Yes.  Well, as was mentioned,

it's not something that we would automatically request

or expect to see in a 510k.  But in evaluating the

device and the testing and the comparison to the

predicate device, which the sponsor would need to

provide, it could be that there would be a question

raised that we would suggest that clinical data be the

mechanism to address that particular difference or


            CHAIRPERSON YASZEMSKI:  Right.  Thank you,

Dr. Witten.

            Have we adequately discussed question


            DR. WITTEN:  Yes, you have.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Question four, do you believe the data

presented in this petition supports the

reclassification of:  A)  all total mobile bearing

knee prostheses identified in the petition?  And if

not, which types of total knees do you believe are

inappropriate for reclassification and why?  B)  All

unicompartmental mobile bearing knee prostheses

identified in the petition?  And if not, which types

do you believe are inappropriate for reclassification

and why?  I would like to start with Dr. Mayor this


            MEMBER MAYOR:  Thank you.

            Yes.  I think, as I have implied in my

previous remarks, I think the data has been presented

to support the reclassification of both the

bicompartmental, tricompartmental mobile bearing knees

and the unicompartmentals in the same reclassification


            CHAIRPERSON YASZEMSKI:  Thanks, Dr. Mayor.

            Dr. Larntz?

            MEMBER LARNTZ:  While I agree clinical

data can go in 510k's, it's not the usual thing.  I

think we do need clinical data.  And I think that has

to be the standard until we get more experience with

these devices.  And so I would be opposed to


            CHAIRPERSON YASZEMSKI:  Thank you, Dr.


            Dr. Besser?

            MEMBER BESSER:  I believe that both the

total and the unicompartmental devices, mobile bearing

devices, could be reclassified as class II.  However,

I would agree with Dr. Larntz that clinical data is

required.  And I would list that as a special control.

            CHAIRPERSON YASZEMSKI:  Thank you, Dr.


            Ms. Maher?

            MEMBER MAHER:  Nothing further to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Doyle?

            MEMBER DOYLE:  I concur with what has been


            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Kirkpatrick?

            MEMBER KIRKPATRICK:  I am most comfortable

with the tricompartmental devices.  I am a little

concerned about unicompartmental devices.  So I would

say a yes on the first and a no on the second.

            CHAIRPERSON YASZEMSKI:  Thanks, Dr.


            Dr. Mabrey?

            MEMBER MABREY:  I have nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Finnegan?

            MEMBER FINNEGAN:  Yes on A and no on B.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Kim?

            MEMBER KIM:  I have nothing to add.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Naidu?

            MEMBER NAIDU:  No on both.

            CHAIRPERSON YASZEMSKI:  Thank you.

            Dr. Witten, have we adequately discussed

question four?

            DR. WITTEN:  Yes.

            CHAIRPERSON YASZEMSKI:  Thank you.



            CHAIRPERSON YASZEMSKI:  Now that we have

addressed the FDA questions, we will complete the

classification questionnaire and supplemental data

sheet.  So our task now is to fill out two sheets.

            Ms. Shulman of the Office of Device

Evaluation will assist us.  After panel discussion of

each question, I will note our answer for each blank

on the data sheet.  And Ms. Shulman will record it on

the overhead for us.  What we will vote on is the

completed questionnaire and data sheet.  That vote

will become the panel's recommendation to the FDA.

            Does anyone have questions on how we are

about to proceed?  Can we get sheets for everybody,

please?  Once we distribute the sheets, I will note

that Dr. Mayor was the lead reviewer.  I am going to

ask his guidance on how to proceed with the answers.

            We will have discussion on each of the

answers and recognize that if there is disagreement on

the answers, we will fill the sheet out based upon our

impression of what the majority consensus opinion is.

And you can address disagreement, should you have it,