[Federal Register: July 27, 1998 (Volume 63, Number 143)]
[Rules and Regulations]
[Page 40025-40041]
From the Federal Register Online via GPO Access [wais.access.gpo.gov]
[DOCID:fr27jy98-7]

-----------------------------------------------------------------------

DEPARTMENT OF HEALTH AND HUMAN SERVICES

Food and Drug Administration

21 CFR Part 888

[Docket No. 95N-0176]
RIN 0910-ZA12


Orthopedic Devices: Classification and Reclassification of
Pedicle Screw Spinal Systems

Agency: Food and Drug Administration, HHS.

Action: Final rule.

-----------------------------------------------------------------------

SUMMARY: The Food and Drug Administration (FDA) is classifying certain
previously unclassified preamendments pedicle screw spinal systems into
class II (special controls) and reclassifying certain postamendments
pedicle screw spinal systems from class III (premarket approval) to
class II. FDA is taking this action because it believes that special
controls would provide reasonable assurance of safety and
effectiveness. This action is being taken under the Federal Food, Drug,
and Cosmetic Act (the act), as amended by the Medical Device Amendments
of 1976 (the 1976 amendments), the Safe Medical Devices Act of 1990
(the SMDA), and the Food and Drug Administration Modernization Act of
1997 (FDAMA).

EFFECTIVE DATE: August 26, 1998.

FOR FURTHER INFORMATION CONTACT: Aric D. Kaiser, Center for Devices and
Radiological Health (HFZ-410), Food and Drug Administration, 9200
Corporate Blvd., Rockville, MD 20850, 301-594-2036.

SUPPLEMENTARY INFORMATION:
Table of Contents
I. Background
II. Regulatory History of the Device
III. Summary of the Final Rule
IV. Proposed Rule Clarifications
V. Analysis of Comments and FDA's Response
    A. Issues Relating to the Recommendations of the Panel, FDA's
Tentative Findings, and Summary of the Data Upon Which FDA's
Findings Were Based
    B. Issues Relating to Information Published in the 1994
Supplementary Issue of the Journal Spine (vol. 20S, 1994)
    C. Issues Relating to the January 1995, 510(k) Substantial
Equivalence Determination for a Pedicle Screw Spinal System Intended
for Severe Spondylolisthesis
    D. Issues Relating to Misstatements or False Statements
Appearing in the Proposed Rule
    E. Issues Relating to the FDA's Issuance of Regulations
    F. Responses to Comments Which Contained Clinical Data
    G. Requests for Additional Pedicle Screw Clinical Trials and
Data Analyses
    H. Issues Relating to Indications for Use
    I. Issues Relating to Special Controls
    J. Other Issues
    K. Labeling of Bone Screws
    L. Review of New Pedicle Screw Spinal System 510(k)'s
    M. Review of New Information Published and Submitted After
Publication of the Proposed Rule: Pedicle Screw and Related
Literature and MedWatch and MDR System Reports
VI. References
VII. Environmental Impact
VIII. Analysis of Impacts

I. Background

    The act (21 U.S.C. 331 et seq.), as amended by the 1976 amendments
(Pub. L. 94-295), the SMDA (Pub. L. 101-629), and FDAMA (Pub. L. 105-
115), established a comprehensive system for the regulation of medical
devices intended for human use. Section 513 of the act (21 U.S.C. 360c)
established three categories (classes) of devices, depending on the
regulatory controls needed to provide reasonable assurance of their
safety and effectiveness. The three categories of devices are: Class I
(general controls), class II (special controls), and class III
(premarket approval).
    Under section 513 of the act, devices that were in commercial
distribution before May 28, 1976 (the date of enactment of the 1976
amendments), generally referred to as preamendments devices, are
classified after FDA has: (1) Received a recommendation from a device
classification panel (an FDA advisory committee); (2) published the
panel's recommendation for comment, along with a proposed regulation
classifying the device; and (3) published a final regulation
classifying the device. FDA has classified most preamendments devices
under these procedures.
    Devices that were not in commercial distribution prior to May 28,
1976, generally referred to as postamendments devices, are classified
automatically by statute (section 513(f) of the act) into class III
without any FDA rulemaking process. Those devices remain in class III
and require premarket approval, unless and until: (1) The device is
reclassified into class I or II; (2) FDA issues an order classifying
the device into class I or II in accordance with new section 513(f)(2)
of the act, as amended by FDAMA; or (3) FDA issues an order finding the
device to be substantially equivalent, in accordance with section
513(i) of the act, to a predicate device that does not require
premarket approval. The agency determines whether new devices are
substantially equivalent to previously offered devices by means of
premarket notification procedures in section 510(k) of the act (21
U.S.C. 360(k)) and part 807 of the regulations (21 CFR part 807).
    A preamendments device that has been classified into class III may
be marketed, by means of premarket notification procedures, without
submission of a premarket approval application (PMA) until FDA
promulgates a final regulation under section 515(b) of the act (21
U.S.C. 360e(b)) requiring premarket approval.
    Reclassification of classified preamendments devices is governed by
section 513(e) of the act. This section provides that FDA may, by
rulemaking, reclassify a device (in a proceeding that parallels the
initial classification proceeding) based upon ``new information.'' The
reclassification can be initiated by FDA or by the petition of an
interested person. The term ``new information,'' as used in section
513(e) of the act, includes information developed as a result of a
reevaluation of the data before the agency when the device was
originally classified, as well as information not presented, not
available, or not developed at that time. (See, e.g., Holland Rantos v.
United States Department of Health, Education, and Welfare, 587 F.2d
1173, 1174 n.1 (D.C. Cir. 1978); Upjohn v. Finch, 422 F.2d 944 (6th
Cir. 1970); Bell v. Goddard, 366 F.2d 177 (7th Cir. 1966).)
    Reevaluation of the data previously before the agency is an
appropriate basis for subsequent regulatory action where the
reevaluation is made in light of newly available regulatory authority
(see Bell v. Goddard, supra, 366 F.2d at 181; Ethicon, Inc. v. FDA, 762
F. Supp. 382, 389-91 (D.D.C. 1991)), in light of changes in ``medical
science.'' (See Upjohn v. Finch, supra, 422 F.2d at 951.) Regardless of
whether data before the agency are past or new data, the ``new
information'' on which any reclassification is based is required to
consist ``valid scientific evidence,'' as defined in section 513(a)(3)
of the act and Sec. 860.7(c)(2) (21 CFR 860.7(c)(2)). (See, e.g.,
General Medical Co. v. FDA, 770 F.2d 214 (D.C. Cir. 1985); Contact Lens
Assoc. v. FDA, 766 F.2d 592 (D.C. Cir.), cert. denied, 474 U.S. 1062
(1985).) FDA relies upon ``valid scientific evidence'' in the
classification process to determine the level of

[[Page 40026]]

regulation for devices. For the purpose of reclassification, the valid
scientific evidence upon which the agency relies must be publicly
available. Publicly available information excludes trade secret and/or
confidential commercial information, e.g., the contents of a pending
PMA. (See section 520(c) of the act (21 U.S.C. 360j(c).)

II. Regulatory History of the Device

    Consistent with the act and the regulation, FDA referred the
proposed classification and reclassification of pedicle screw spinal
systems to the Orthopedic and Rehabilitation Devices Panel (the Panel),
an FDA advisory committee, for its recommendation on the requested
classification and change in classification.
    The Panel reviewed complication type and rate data present in the
literature, a meta-analysis of the literature; a nationwide,
retrospective Cohort study of patients treated with the devices;\1\ and
a review of publicly released investigational device exemptions (IDE)
data from patients treated with pedicle screw spinal systems. The Panel
recommended that the postamendments pedicle screw spinal systems
intended to treat spinal fracture and degenerative spondylolisthesis of
the thoracic, lumbar, and sacral spine, be reclassified from class III
into class II.
---------------------------------------------------------------------------

    \1\ The Cohort study was an open, nonblinded, historical Cohort
study designed to recruit the maximum number of surgeons to provide
clinical data on patients who had undergone spinal fusion surgery.
Three hundred fourteen surgeons were recruited through announcements
at professional society meetings and direct mailings to professional
society memberships. Only clinical data from spinal fusion surgeries
intended to treat degenerative spondylolisthesis or spinal trauma
(fracture) that were performed between January 1, 1990, and December
31, 1991, were used in the analysis. This was done in an effort to
maximize the number of patients with a minimum of 24 months
followup. Data from 3,498 patients were collected.
---------------------------------------------------------------------------

    In January, 1995, a manufacturer was able to demonstrate
preamendments status for pedicle screw spinal systems intended to
provide immobilization and stabilization of spinal segments as an
adjunct to spinal fusion in the treatment of grades 3 or 4 severe
spondylolisthesis at the fifth lumbar-first sacral (L<INF>5</INF>-
S<INF>1</INF>) spinal level. In an April 1995, homework assignment, FDA
requested that the Panel recommend a classification for this
unclassified preamendments device. The Panel recommended that the
unclassified preamendments pedicle screw spinal systems intended to
provide immobilization and stabilization of spinal segments as an
adjunct to spinal fusion in the treatment of grades 3 or 4 severe
spondylolisthesis at the L<INF>5</INF>-S<INF>1</INF> spinal level be
classified into class II.
    In the Federal Register of October 4, 1995 (60 FR 51946), FDA
published a proposed rule to classify certain unclassified
preamendments pedicle screw spinal systems (for use in certain types of
severe spondylolisthesis ) into class II, to reclassify certain
postamendments pedicle screw spinal systems (for use in fracture and
other conditions) from class III to class II, and to retain in class
III other postamendments pedicle screw spinal systems. The proposed
rule reflected FDA's belief that the clinical outcomes and
complications described in the literature, clinical data, and MDR and
MedWatch surveillance data bases,\2\ described patient risks and
benefits of pedicle screw spinal systems comparable to other class II
spinal fixation devices and that special controls have been identified
which would provide a reasonable assurance of safety and effectiveness,
i.e., compliance with material standards, mechanical testing standards,
biocompatibility standards, and special labeling requirements.
Initially, FDA provided for interested persons to submit comments on
the proposal by January 2, 1996. Subsequently, in the Federal Register
of December 29, 1995 (60 FR 67345), FDA extended the comment period
until March 4, 1996, in response to several requests for extension of
the comment period.
---------------------------------------------------------------------------

    \2\ MDR and MedWatch data bases are two reporting systems that
FDA uses to track adverse events, e.g., injuries, deaths, and device
malfunctions, related to medical devices. The information consists
of a combination of mandatory and/or voluntary adverse event reports
from manufacturers, distributors, user facilities, healthcare
professionals, as well as consumers.
---------------------------------------------------------------------------

    FDA received 4,060 comments in response to the proposed rule. These
comments were submitted by physicians, patients, lawyers, device
manufacturers, trade associations, and other interested parties. The
overwhelming majority of these comments were in favor of the proposed
rule, although some comments were opposed to the proposed rule, and a
few were both in favor of some aspects of the proposed rule and opposed
to others.
    In response to comments received on the proposed rule, FDA
reanalyzed the meta-analysis of the literature, the Cohort study, and
the publicly released IDE data for the indications of spinal fractures
and degenerative spondylolisthesis. The reanalysis of the meta-analysis
of the literature consisted of a review of the summary data and
conclusions from the original, published analysis. The review of the
Cohort study consisted of an audit (Ref. 1) of a structured sample of
all 377 patients enrolled by 21 of the 314 participating surgeons, a
reanalysis (Ref. 2) of all of the data from the audit, and a comparison
to the data from unaudited surgeons. The Division of Bioresearch
Monitoring (BIMO) in the Office of Compliance performed the data audit,
while the Office of Device Evaluation and the Office of Surveillance
and Biometrics performed the reanalyses. This audit found records were
incomplete and investigators had not followed the protocol. In review
of the audit, the agency concluded that the disparities and
irregularities were consistent, with respect to both type and scope,
with other audits of similar studies. After careful reanalysis of the
potential impact of the ``problem'' records, the agency concluded that
they could not account for the favorable results reported in this
study.
    The review of the Cohort study in the context of the audit findings
yielded results that supported the safety and effectiveness of these
devices. For spinal fracture, pedicle screw spinal systems presented
risks and benefits that were comparable to those presented by
nonpedicle screw instrumented spinal fusion. The devices used in the
comparison group are class II medical devices. For spondylolisthesis,
the review in the context of the audit findings described an advantage
for pedicle screw spinal systems with regard to the clinical outcome
parameters of fusion and improvement in neurological status when
compared to noninstrumented spinal fusions. For the other parameters
that were analyzed, e.g., pain, function, and reoperation rate, pedicle
screw spinal systems did not always demonstrate an advantage compared
to noninstrumented spinal fusion. When compared to instrumented spinal
fusions, however, results among pedicle screw spinal system patients
for these parameters were not statistically equivalent and not worse.
Thus, FDA has concluded that the results from the review of the Cohort
study are consistent with those reported in the literature and the
publicly released IDE data.
    The reanalysis of the meta-analysis of the literature describing
experience with pedicle screw spinal systems in treating spinal
fracture and degenerative spondylolisthesis found that pedicle screw
spinal systems present risks and benefits that are comparable to those
presented by nonpedicle screw spinal systems and noninstrumented spinal
fusions. For degenerative spondylolisthesis, the reanalysis found

[[Page 40027]]

that patient results with pedicle screw spinal systems were comparable
to those with noninstrumented spinal fusions; it did not find a
clinically significant improvement in results at followup obtained with
instrumented spinal fusions over noninstrumented spinal fusions.
    The reanalysis of the publicly available IDE data supports the
Panel's recommendation for the classification and reclassification of
pedicle screw spinal systems intended to treat spinal fractures and
severe spondylolisthesis. It also supports the use of pedicle screw
spinal systems when intended to provide immobilization and
stabilization of spinal segments in skeletally mature patients as an
adjunct to fusion for the treatment of the following acute and chronic
instabilities or deformities of the thoracic, lumbar, and sacral spine:
degenerative spondylolisthesis with objective evidence of neurologic
impairment, fractures, dislocations, scoliosis, kyphosis, spinal
tumors, and failed previous fusion (pseudarthrosis).
    When all of these data are viewed in conjunction with the medical
literature and the MDR and MedWatch surveillance data, no new issues
relating to the safety or effectiveness of pedicle screw spinal systems
are raised. Therefore, the agency has concluded that these data provide
valid scientific evidence that certain special controls in conjunction
with the general controls applicable to all devices, will provide a
reasonable assurance of the safety and effectiveness of pedicle screw
spinal systems for L<INF>5</INF>-S<INF>1</INF> use and for use at other
levels for the treatment of degenerative spondylolisthesis with
objective evidence of neurologic impairment.
    The agency also reviewed whether the Panel was properly
constituted. Investigation of alleged undisclosed and unwaived
conflicts of interest held by Panel members found either no omissions
of current interests or omissions of minor interests for all but one of
the Panel members. The agency has concluded that the minor omissions
are insignificant and do not constitute a financial conflict of
interest that would credibly influence the members' actions in forming
the Panel's recommendations.
    The agency has found that one voting Panel member did have
significant undisclosed financial conflicts. However, because the
recommendation of the Panel, both in the July 23, 1994, meeting and on
the subsequent homework assignment, was unanimous and this individual
was not controlling, or unduly influential, of the votes of the other
Panel members and was not necessary to constitute a quorum, after
expunging the participation of this Panel member, FDA has concluded
that this Panel, both in the meeting and on the subsequent homework
assignment, was a valid scientific Panel to make recommendations to the
agency.
    The agency's reanalysis of these data has confirmed its original
conclusion, reflected in the proposed rule, that the risks and benefits
of pedicle screw spinal systems are comparable to those of other class
II spinal fixation devices. FDA's decision to classify and reclassify
these devices into class II is based upon valid scientific evidence
establishing that the special controls described above, along with the
general controls applicable to all devices under the act, provide a
reasonable assurance of the safety and effectiveness of pedicle screw
spinal systems.

III. Summary of the Final Rule

    In this final rule, FDA is classifying into class II the
unclassified preamendments pedicle screw spinal systems intended for
treatment of severe spondylolisthesis (grades 3 and 4) of the
L<INF>5</INF>-S<INF>1</INF> vertebra in skeletally mature patients
receiving fusion by autogenous bone graft having implants attached to
the lumbar and sacral spine with removal of the implants after the
attainment of a solid fusion. In addition, FDA is reclassifying into
class II the postamendments class III pedicle screw spinal systems
intended to provide immobilization and stabilization of spinal segments
in skeletally mature patients as an adjunct to fusion in the treatment
of the following acute and chronic instabilities or deformities of the
thoracic, lumbar, and sacral spine: Degenerative spondylolisthesis with
objective evidence of neurologic impairment, fracture, dislocation,
scoliosis, kyphosis, spinal tumor, and failed previous fusion
(pseudarthrosis). Pedicle screw spinal systems intended for any other
uses are considered postamendments class III devices for which
premarket approval is required. The following four special controls
apply to the devices being classified and reclassified into class II:
(1) Compliance with materials standards, (2) compliance with mechanical
testing standards of performance, (3) compliance with biocompatibility
standards, and (4) adherence to labeling requirements.

IV. Proposed Rule Clarifications

    FDA is taking this opportunity to clarify that neither well-
controlled investigations nor valid scientific evidence relating to
pedicle screw spinal systems intended for use in the cervical spine is
available and, therefore, the safety and effectiveness of these devices
for this intended use have not been demonstrated. As a result, pedicle
screw spinal systems intended for use in the cervical spine are
excluded from this classification and reclassification and are
considered postamendments class III devices for which premarket
approval is required.
    In addition, although not specifically stated in the preamble to
the proposed rule, all valid scientific evidence reviewed by the Panel
and FDA were obtained from skeletally mature populations. To date, the
safety and effectiveness of pedicle screw spinal systems in pediatric
populations have not been demonstrated. Consequently, pedicle screw
spinal systems intended for use in pediatric populations are
postamendments class III devices for which premarket approval is
required.

V. Analysis of Comments and FDA's Response

A. Issues Relating to the Recommendations of the Panel, FDA's Tentative
Findings, and Summary of the Data Upon Which FDA's Findings Were Based

    1. Several comments believed that valid scientific evidence was not
presented to the Panel or used in formulating the proposed rule. These
comments argued that only prospective, randomized, concurrently-
controlled clinical trials constitute valid scientific evidence and
that anything else is insufficient to support device reclassification.
    FDA disagrees that only data from prospective, randomized,
concurrently-controlled clinical trials can constitute valid scientific
evidence. Although prospective, randomized, concurrently controlled
clinical trials have the potential to produce the most convincing and
reliable data, e.g., all sources of bias have been reduced to a
minimum, such clinical trials are not the only type of study that can
produce data adequate to support a determination that there is
reasonable assurance that a device is safe and effective for its
conditions for use. In fact, Sec. 860.7(c)(2) defines valid scientific
evidence as
    * * * evidence from well-controlled investigations, partially
controlled studies, studies and objective trials without matched
controls, well-documented case histories conducted by qualified
experts, and reports of significant human experience with a marketed
device, from which it can fairly and responsibly be concluded by
qualified experts that there is reasonable assurance of the safety
and effectiveness of a device under

[[Page 40028]]

its conditions of use. The evidence may vary according to the
characteristics of the device, the existence and adequacy of
warnings and other restrictions, and the extent of experience with
its use * * *.
From this definition, it is clear that there is a hierarchy of data
that fits the definition of valid scientific and that, while well-
controlled, prospective clinical trials are at the top of the
hierarchy, they are not the only source of data that may support a
determination regarding reasonable assurance of safety and
effectiveness for purposes of classifying and/or reclassifying a
device.
    FDA also disagrees that valid scientific evidence was not presented
to the Panel or used in support of the proposed rule. The three sources
of data presented to the Panel and used in support of the proposed rule
were: (1) Reformatted IDE data, (2) a meta-analysis of the literature,
and (3) the Cohort study. The reformatted IDE data represent data from
well-controlled investigations, while the meta-analysis of the
literature and the Cohort study represent studies and objective trials
without matched controls or well-documented case histories or reports
of significant human experience. All three sources of data used in
support of the classification and reclassification of pedicle screw
spinal systems clearly fall within the definition of valid scientific
evidence in Sec. 860.7(c)(2).
    2. One comment objected that, in addition to not being valid
scientific evidence, the three sources of data, i.e., the meta-analysis
of the literature, the reformatted IDE data and the Cohort study, were
flawed.
    The comment noted the following deficiencies with the meta-
analysis:
    <bullet>  FDA previously determined that the available literature
on pedicle screw spinal systems could not be used to support device
reclassification.
    FDA disagrees. FDA made that statement prior to January, 1993, when
no adequate analysis of pedicle screw literature had been provided to
the agency. FDA believes that, while individual literature articles
describing the use of pedicle screw spinal systems would be
insufficient to support reclassification of a device, group analysis of
relevant articles may be adequate, especially where, as here, the group
analysis is considered in conjunction with other supporting data.
Furthermore, after noting the limitations of the individual studies
reported in the literature, FDA concluded that the literature, taken as
a whole and used in conjunction with the other sources of data,
provided adequate information to support the reclassification of
pedicle screw spinal systems intended to treat degenerative
spondylolisthesis with objective evidence of neurologic impairment or
spinal trauma.
    <bullet>  The meta-analysis is not an appropriate scientific
technique, as applied to retrospective studies, because different
studies have different parameters, biases, and strengths and
weaknesses, all of which invalidate the pooling of data.
    FDA disagrees. Although meta-analysis of literature may be less
rigorous than other forms of scientific research, it still provides
useful information. As discussed in section V.A.1 of this document,
Sec. 860.7(c)(2) defines ``valid scientific evidence'' to include many
types of evidence of varying degrees of scientific rigor, including
meta-analysis of literature. FDA participated in the development of the
meta-analysis because the agency believed that this analysis could
produce data meeting the definition of valid scientific evidence.
Finally, the inherent limitations of a literature meta-analysis were
discussed during the presentation of this analysis at the July 23,
1994, Panel meeting and in the preamble to the proposed rule (60 FR
51946).
    <bullet>  The meta-analysis actually lent support to the conclusion
that pedicle screw fixation is less effective than other methods of
treating degenerative spondylolisthesis and spinal fracture and that it
may present the patient with more risks.
    FDA disagrees. With respect to degenerative spondylolisthesis,
there was no statistically significant difference in fusion rates
between the control and the pedicle screw spinal system treatment
groups. This is supportive data that clarifies the relative safety and
effectiveness of pedicle screw spinal systems for this use. With
respect to spinal fracture, significantly higher fusion rates were
achieved in the pedicle screw spinal system treatment group than in the
nonpedicle screw treatment groups. Thus, the meta-analysis confirmed
the comparability of pedicle screw spinal systems to other class II
devices used to treat spinal fracture in terms of safety and
effectiveness.
    <bullet>  Fifty-five of the 58 studies in the meta-analysis were
nonexperimental case-series having no validity as scientific evidence.
    FDA disagrees. As discussed in section V.A.1 of this document,
Sec. 860.7(c)(2) states that valid scientific evidence may include ``*
* * well-documented case histories conducted by qualified experts * *
*''. Moreover, these well-documented case studies, which were conducted
by qualified experts, were not the sole basis for the proposed
classification/reclassification, but rather were considered in
conjunction with data from various other sources.
    The comment also noted the following deficiencies with the
reformatted IDE data:
    <bullet>  The reformatted IDE data are not appropriate for
classifying and reclassifying pedicle screw spinal systems because FDA
previously had determined that these data could not support PMA's for
these devices.
    FDA disagrees in part. Prior to the August 20, 1993, Panel meeting,
FDA had determined that data from individual IDE's were insufficient to
support PMA's for those devices. Nevertheless, FDA recognized that the
IDE data could still be valuable. In 1993, after receiving permission
from nine IDE sponsors to publicly release and use their combined data,
FDA determined that the data, reviewed as a whole, corroborated the
results of other available data sets demonstrating the safety and
effectiveness of pedicle screw spinal systems.
    <bullet>  The reformatted IDE data are inherently suspect because
they (1) were reformatted by the sponsors and not by FDA, (2) were not
provided for public scrutiny during the Panel meeting or at any other
time, and (3) may have omitted poor results.
    FDA disagrees that the reformatted data were suspect because they
were reformatted by the sponsors and not by FDA. If IDE data are not
properly formatted, FDA requests the sponsor to reformat its data for
proper presentation to the agency. Furthermore, data in all marketing
applications are formatted by the sponsor. Therefore, the simple fact
that the IDE data were reformatted by the sponsor, not by FDA, does not
make these data inherently suspect.
    FDA also disagrees that the data were suspect because they were not
presented for public scrutiny. For reclassification purposes, the valid
scientific evidence upon which the agency relies must be publicly
available Sec. 860.5(e) (21 CFR 860.5(e)). Publicly available
information excludes trade secret and/or confidential commercial
information (21 CFR 20.61). IDE data typically contain trade secret
and/or confidential commercial information and, consequently,
ordinarily may not be publicly disclosed by the agency to support
reclassification of a device (49 FR 17523 at 17531 and 17532, April 24,
1984). In fact, under Sec. 812.38(a) and (b)(3) (21 CFR 812.38 (a) and
(b)(3)), FDA generally does not acknowledge the existence of an IDE or
disclose any

[[Page 40029]]

of the collected data. However, on August 13, 1993, after receiving
permission from nine IDE sponsors to publicly release and use their
combined data, the Commissioner of Food and Drugs (the Commissioner)
exercised his discretionary authority under Sec. 812.38(b)(2) and
publicly released the data from nine IDE's, redacted of the
identification of the IDE sponsors, institutional review boards,
investigators, and patients. Although FDA did not make publicly
available the unformatted data from the IDE studies or the
identification of the IDE sponsors, institutional review boards,
investigators or patients, the agency did provide the public with a
detailed report of the combined IDE data (60 FR 51946 at 51961, ref.
173). This information was publicly available for analysis for more
than 2 years before the publication of the proposed rule.
    Finally, FDA disagrees that the data were suspect because they may
have omitted poor results. Nine of fourteen sponsors provided their
reformatted IDE data for analysis. There is no evidence that the five
sponsors who did not offer their data did so because the data reflected
adversely on the performance of their products. They may not have
provided their data for any number of reasons. For instance, the
sponsors may have believed that they had an inadequate amount of data
to contribute to the effort or that the data may not have been in a
readily accessible format. Regardless of the reason, the publicly
available reformatted IDE data corroborate the results of other studies
that demonstrate the safety and effectiveness of pedicle screw spinal
systems. Specifically, the fusion rates, complication rates, and
reoperation, revision, and removal rates attained under publicly
available IDE studies were consistent with what was observed in the
literature for such devices.
    <bullet>  The 12-month followup time period was inadequate to
support any conclusions. Specifically, the comment stated that the
Panel was not supplied with any information on the safety and
effectiveness of these devices at more than 1 year following surgery.
The comment continued that, without a minimum followup period of 2
years, it is impossible to make appropriate conclusions with regard to
the longer-term safety and efficacy of these devices in accordance with
accepted scientific convention.
    FDA agrees that a 12-month followup time period would be inadequate
and, therefore, selected a 24-month followup period for analysis. The
24-month followup period was also supported by the Panel and the
literature. Contrary to the comment's statement, the Panel was supplied
with information on the safety and effectiveness of pedicle screw
spinal systems at more than one year following surgery. Spinal fusion
generally occurs within 6 to 18 months after surgery. The majority of
post-operative complications occur by the 18th month time point. For
these reasons, FDA concluded that a 24-month followup period was
adequate. FDA recognizes that not all of the reformatted IDE data were
from a 24-month followup examination. However, a sufficient amount of
data from a 24 month followup evaluation was examined for the Panel to
make a recommendation about the reasonable assurance of safety and
effectiveness of pedicle screw spinal systems for their class II
intended uses.
    <bullet>  The comment stated that the lost-to-followup rate was too
high.
    FDA agrees that the lost-to-followup rate was high. FDA believes
that patients with poor results tend to either return to their surgeons
more frequently or go to other caregivers, attempting to receive the
pain relief and return of function that they were originally seeking.
It cannot be determined whether the patients who were lost-to-followup
had acceptable results or went to other caregivers. However, FDA does
not believe that this theoretical weakness in the data is of such a
magnitude as to justify rejecting the studies. Thus, both the Panel and
FDA believe that the lost-to-followup rate was not unacceptably high.
    The comment noted the following deficiencies with the Cohort study:
    <bullet>  The Cohort study did not constitute valid scientific
evidence.
    FDA disagrees. As described above, valid scientific evidence
encompasses a wide variety of data. The Cohort study satisfies the
definition of valid scientific evidence because it consisted of data
from well-documented case histories conducted by qualified experts and
reports of significant human experience.
    <bullet>  The sample size and statistical power used in designing
the Cohort study were inadequate and, therefore, no reliable
conclusions can be drawn from the study. Another comment attempted to
rebut this allegation.
    FDA believes that the sample size and statistical power
calculations that were performed in the Cohort study were accurate and
appropriate and, consequently, that the conclusions drawn from the
study had a sound basis.
    <bullet>  The Cohort study was biased and the data were not
independently audited.
    FDA disagrees. While the potential for bias exists in any study, it
was of particular concern in the design of the Cohort study due to its
retrospective nature. As described at the July 22, 1994, Panel meeting
and in the preamble to the proposed rule (60 FR 51946 at 51954),
various steps were taken to minimize the potential effects of bias due
to the study design. In addition, contrary to the comment's assertion,
there was a review of the data by an independent auditor and a
subsequent FDA BIMO audit and review. The review by the independent
auditor was not extensive and no definitive conclusions can be drawn
from its analysis of the Cohort study data. Although both audits
uncovered instances of protocol departures, recordkeeping
inconsistencies, or a lack of clear understanding or unfamiliarity with
the protocol requirements on the part of a participating surgeon, these
inconsistencies and protocol departures did not affect the reliability
of the data. For example, one type of reported protocol recordkeeping
departure was that some data forms were incomplete. In some instances,
the data forms simply omitted the patient's weight, but not the
patient's fusion status. The absence of that piece of information,
while rendering the form incomplete, clearly did not affect the
clinical outcome analysis. A more significant protocol departure
related to the inclusion and analysis of data from patients whose
diagnosis did not meet patient eligibility criteria. However, no
obvious pattern that would improve overall patient outcomes was
identified because these departures included indications for surgery
both more and less severe than those targeted by the protocol.
    The data retrieved from the BIMO audit were analyzed to determine
if the major outcomes of the Cohort study were significantly different
(statistically or clinically) with or without the presence of protocol
departures, with or without the presence of recordkeeping
inconsistencies, or at sites where the participating investigator,
based on the audit, was or was not familiar with the protocol
requirements. While some differences were noted between sites with and
without inconsistencies, in most cases, these were not statistically
significant and no consistent or clinically relevant patterns were
noted. The analysis of the audited data did not find systemic bias in
either the conduct of the study or its reported results. None of the
analyzed audit data contradicted the published results of the Cohort
study. Finally, the data audit analysis concluded that the audited data
were consistent with other publicly available

[[Page 40030]]

data and that the Cohort study data could be used as part of a larger
body of data to support the classification and reclassification of
pedicle screw spinal systems.
    <bullet>  Documents relating to the Cohort study were destroyed.
    FDA disagrees. All Cohort study data were maintained in a master
file. Only extra copies of information were destroyed in an effort to
maintain the confidentiality of the identities of the participating
surgeons and their patients. In addition, as a matter of course, FDA
routinely assists Panel members in destroying copies of documents
containing trade secret and/or confidential commercial information that
they have received from FDA as preparatory material for a Panel
meeting.
    <bullet>  Certain FDA employees had inappropriate relationships
with pedicle screw manufacturers and others involved in the Cohort
study.
    This allegation, which has two parts, is unfounded. FDA performed
an internal affairs investigation of the employees about whom
allegations were made. This investigation showed that their attendance
at a health professional meeting was properly paid for by the agency,
not subsidized by the regulated industry. Also in the case of one
employee, FDA's investigation showed that negotiations regarding
outside employment with the regulated industry had been properly
reported to the employee's supervisors and immediate colleagues in all
instances.
    <bullet>  The Scientific Committee and the Spinal Implant
Manufacturers Group (SIMG) were not independent.
    FDA disagrees. The preamble to the proposed rule and the subsequent
correction (60 FR 51946 and 60 FR 66227, December 21, 1995) described
the makeup of the Scientific Committee and SIMG. SIMG consisted of
representatives of manufacturers who provided funding to support a
nationwide analysis of clinical data relating to pedicle screw spinal
systems. SIMG did not participate in the design of the study. The study
was designed and implemented by the Scientific Committee with input
from FDA as to the feasibility of various clinical study design
parameters. The Scientific Committee was formed by five professional
medical societies. Although two SIMG representatives were part of the
Scientific Committee, they were nonvoting members. Furthermore, even if
there were not independence between the Scientific Committee and SIMG,
there is no requirement that clinical studies be performed by parties
independent of device manufacturers. In fact, FDA routinely receives
and relies upon studies performed by manufacturers.
    3. Several comments contended that financial conflicts of interest
were present in the three sources of data relied on by FDA to support
the classification/reclassification of pedicle screw spinal systems.
The comments claimed that, in the meta-analysis of the literature, the
authors of the individual articles had financial conflicts of interest
due to their relationships with device manufacturers. With respect to
the analysis of the reformatted IDE data and the Cohort study, the
comment stated that the IDE investigators and Cohort study participants
had significant financial interests in the companies whose devices they
were using and, therefore, had a strong financial incentive to report
only successful results. Similar objections were raised about the
authors of the 206 articles cited as constituting the body of medical
literature bearing on pedicle screw fixation. The comments stated that
almost all of the surgeons who authored these articles failed to
disclose their financial connections to manufacturers. The comments
stated that such interests raise serious concerns about researchers'
motivation to perform the research, the propriety and importance of
research questions and research designs, the adequacy of protection of
human subjects, lack of bias, and veracity in collecting and analyzing
the data and reporting the results.
    FDA recognizes that some of the clinical investigators involved in
the three sources of data, as well as some of the authors of the 206
literature articles used to support classification and reclassification
of pedicle screw spinal systems, had financial interests in the devices
they were studying. FDA disagrees, however, that these financial
interests resulted in biased or unreliable data. Regardless of the
source of the data, the meta-analysis, the reformatted IDE data, the
Cohort study, or the collection of cited literature, the conclusions
were similar, i.e., that pedicle screw spinal systems are safe and
effective for the uses examined. Because of this, even if financial
conflicts of interest were present, they did not affect the resulting
data and the conclusions. Moreover, the agency has concluded that,
despite the failure to disclose the financial interests of clinical
investigators, the sponsors of these investigations and/or articles
took reasonable steps to minimize potential bias.
    Furthermore, the fact that some spine surgeons were compensated by
industry for research or consulting services, or were reimbursed for
expenses incurred in connection with continuing medical education
courses, did not affect the validity of any of the data. Moreover, many
of the grants to support research were made directly to university
accounts for general research and development, not directly to
individual investigators. Consequently, the existence of a financial
relationship between some surgeons and manufacturers did not
necessarily result in biased case selection or reporting. Finally, FDA
notes that research used to support a medical device marketing
application has always been supported by the sponsor of the device and
there is neither an expectation of nonsupport nor a requirement of
disclosure of such support.
    4. Several comments stated that pedicle screw spinal systems
present different safety and effectiveness issues than do either class
II spinal devices using hooks and/or wires or noninstrumented spinal
fusions. One comment identified the following areas of concern as
having the potential of presenting unreasonable danger for patients:
    (1) Difficulty in placing screws completely within the walls of the
pedicle;
    (2) Inability to determine screw placement postoperatively using
radiographic techniques;
    (3) Damage to nerve tissue as a result of transient contact with a
screw during screw placement;
    (4) Nerve root damage (irritation or compression) as a result of
screw malposition;
    (5) Device failure;
    (6) Loss of bone density as a result of stress shielding;
    (7) Foreign body tissue response;
    (8) Crevice corrosion;
    (9) Fretting corrosion;
    (10) Fibrosis;
    (11) Bone fracture, particularly that of the pedicles;
    (12) Nerve root or spinal cord compression as a result of fibrosis
or foreign body tissue response;
    (13) Chronic irritation;
    (14) Spine destabilization possibly leading to nonunion;
    (15) Increased venous pressure as a result of blocked venous
channels within the bone;
    (16) Increased risk of infection;
    (17) Loss or decrease of sensory and/or motor function;
    (18) Loss of bowel or bladder control; and
    (19) Loss of sexual function.
    FDA agrees that pedicle screw spinal systems have some potential
risks that

[[Page 40031]]

are different from those of other class II spinal devices. However, the
majority of the potential risks presented by these devices, e.g., bone
fracture, foreign body tissue response, loss or decrease in sensory
and/or motor function, and device failure or corrosion, are also
associated with class II spinal devices which use hooks and/or wires
for the same intended uses. Similarly, potential risks such as nonunion
and instability are also associated with noninstrumented spinal
fusions. Moreover, as described in the proposed rule, the incidence of
these adverse outcomes is no greater when a pedicle screw spinal system
is used than when other types of spinal fusions, instrumented and
noninstrumented, are performed in appropriately selected patients (60
FR 51946 at 51957). Finally, FDA believes that the potential risks that
are unique to pedicle screw spinal systems, e.g., difficulty in placing
screws completely within the walls of the pedicle, inability to
determine screw placement postoperatively using radiographic
techniques, damage to nerve tissue as a result of transient contact
with a screw during screw placement, and nerve root damage (irritation
or compression) as a result of screw malposition, can be adequately
addressed by the identified special controls and proper surgeon
training and surgical technique.
    5. One comment asserted that the supposed advantages of pedicle
screw spinal systems are largely theoretical. The comment stated that,
while some investigators have shown that instrumented fusions increase
the likelihood of obtaining a solid fusion, others have demonstrated
that there is no significant increase in fusion rates performed with
instrumentation as compared with noninstrumented fusions performed with
bone graft alone.
    FDA agrees that the data do not always support the theoretical
advantages of using pedicle screw spinal systems compared to alternate
methods of achieving spinal fusion. However, in forming its
recommendations, neither FDA nor the Panel is required to analyze the
theoretical behavior of a given device. It is only required to
determine whether the data demonstrate that there is a reasonable
assurance of safety and effectiveness for its intended uses.
    6. The same comment stated that spinal fusion surgery is usually
performed because of the belief that spinal instability results in
pain. The clinical indicators used to determine which patients have
spinal instability and, therefore, are candidates for spinal fusion
surgery, are not clearly defined and are often not measurable. Because
the results of spinal fusion surgery are also dependent on measurements
of instability, a determination of success is difficult, if not
impossible.
    FDA agrees that the methods used to measure instability are not
perfect and that several working definitions of instability exist.
Nevertheless, instability is measurable. In addition, the same
instability definitions and measurement techniques are used in
determining how a patient is to be treated, i.e., with pedicle screw
spinal systems, class II spinal devices using hooks and/or wires, or
noninstrumented fusions. FDA agrees that the determination of success
of spinal fusion surgery is often difficult, but disagrees that it is
impossible to determine. In fact, the success rates from using the
three treatment methods described above have been determined and found
to be reasonably equivalent (60 FR 51946 at 51954).
    7. Three comments questioned the most appropriate classification
for pedicle screw spinal systems. One comment proposed that pedicle
screw spinal systems be classified into class I and two comments
suggested placing them in class III.
    FDA disagrees. Based on the available information, both the Panel
and FDA concluded that general controls alone are not sufficient to
provide a reasonable assurance of the safety and effectiveness of
pedicle screw spinal systems intended to provide immobilization and
stabilization of spinal segments as an adjunct to fusion in the
treatment of the following acute and chronic instabilities or
deformities of the thoracic, lumbar, and sacral spine: Degenerative
spondylolisthesis with objective evidence of neurologic impairment,
fractures, dislocations, scoliosis, kyphosis, spinal tumors, and failed
previous fusion attempts (pseudarthrosis). The Panel and FDA also
concluded that premarket approval was not necessary to provide such
assurance. Both the Panel and FDA identified the potential risks
associated with the use of these devices and concluded that sufficient
information exists to establish special controls to provide reasonable
assurance of their safety and effectiveness. As a result, FDA is
classifying and reclassifying these devices in class II.
    8. A comment believed that classification and reclassification of
pedicle screw spinal systems into class II is inappropriate because FDA
was correct in its prior determination that basic principles of
physiology, anatomy, biology, and biomechanical engineering demonstrate
that pedicle screw spinal systems present a serious risk of injury to
the spinal nerves, nerve roots, and surrounding vascular structures,
and increase the risk of pseudarthrosis. According to this comment,
these risks are not posed by existing spine fusion technology and
pedicle screw spinal systems are of questionable efficacy in comparison
to existing methodologies of treatment.
    FDA disagrees. FDA did not determine that basic principles of
physiology, anatomy, biology, and biomechanical engineering demonstrate
that pedicle screw spinal systems present a serious risk of injury.
Rather, in 1984, FDA determined that a multiple component device system
intended for attachment to the spine via the pedicles was not
substantially equivalent to any legally marketed predicate device, in
accordance with section 513(i)(1) of the act. FDA's decision was based
on the fact that: (1) The sponsor did not identify a legally marketed
preamendments device incorporating pedicle screw components and (2) the
device posed potential risks not exhibited by other legally marketed
predicate spinal fixation systems, such as a greater chance of
neurological deficit due to imprecise screw placement or the event of a
screw failure; pedicle fracture during placement of screws; soft tissue
damage or inadequate fusion due to bending or fracture of device
components; and greater risk of pseudarthrosis due to instability of
the device design (60 FR 51946 at 51947). As stated previously, FDA
believes that the risks to health presented by pedicle screw spinal
systems intended to provide immobilization and stabilization of spinal
segments as an adjunct to fusion in the treatment of the following
acute and chronic instabilities or deformities of the thoracic, lumbar,
and sacral spine: Degenerative spondylolisthesis with objective
evidence of neurologic impairment, fractures, dislocations, scoliosis,
kyphosis, spinal tumors, and failed previous fusion (pseudarthrosis)
can be adequately addressed by special controls. Consequently, FDA is
classifying and reclassifying these devices into class II.
    9. One comment argued that manufacturers of pedicle screw spinal
systems are seeking to have FDA down classify the device into class II
because the manufacturers are unable to prove that pedicle screws are
safe and effective for posterior implantation into the spine.
    FDA disagrees. First, contrary to the comment's statement, this
classification and reclassification proceeding was

[[Page 40032]]

initiated by FDA; it is not in response to a petition for
reclassification. Second, under section 513 of the act, devices are
classified/reclassified into one of three classes based on reasonable
assurance, not ``absolute proof,'' of their safety and effectiveness.
Contrary to the comment's statement, it was not pedicle screw spinal
system manufacturers, but rather the Panel and FDA, that concluded that
pedicle screw spinal systems should be classified and reclassified into
class II because they determined that premarket approval is not
necessary to provide reasonable assurance of the device's safety and
effectiveness, general controls alone are insufficient to provide such
assurance, and there is sufficient information to establish special
controls to provide such assurance.
    10. According to another comment, by classifying and reclassifying
pedicle screw spinal systems into class II, FDA is acknowledging that
there is no need for the manufacturers of pedicle screw spinal systems
to prove that the devices are safe and effective.
    FDA agrees. The agency has determined that sufficient information
exists to establish special controls to provide reasonable assurance of
the safety and effectiveness of pedicle screw spinal systems intended
to provide immobilization and stabilization of spinal segments in
skeletally mature patients as an adjunct to fusion in the treatment of
the following acute and chronic instabilities or deformities of the
thoracic, lumbar, and sacral spine: Degenerative spondylolisthesis with
objective evidence of neurologic impairment, fracture, dislocation,
scoliosis, kyphosis, spinal tumor, and failed previous fusion
(pseudarthrosis). FDA has determined that premarket approval is not
necessary to provide such assurance.

B. Issues Relating to Information Published in the 1994 Supplementary
Issue of the Journal Spine (vol. 20S, 1994)

    11. One comment objected that the manner in which the Scientific
Committee communicated to the public the results of the Cohort study
and related meta-analyses of the literature lacked scientific
integrity. According to the comment, the articles were not peer
reviewed, but rather they were accepted for publication solely by the
Editor-in-Chief of the peer-reviewed journal Spine. The comment
contended that publication of the articles without peer review
prevented the studies from being submitted to the usual critical
scrutiny of any peer review in the future.
    While the articles describing the Cohort study and related meta-
analysis were not peer-reviewed in the usual manner, they were
subjected to a review process and published in an October 19, 1994,
Special Supplement of Spine. The editorial at the beginning of the
supplement states that,
    The members of the Scientific Committee and editors of Spine
felt it important that presentations from the (July 1994, Panel)
meeting be available to the readers of Spine in an expedited manner.
The articles have been reviewed by the Scientific Committee, but
have not gone through the normal review process of the Spine
Editorial Board. However, it has been prepared, written, re-written,
and critiqued by all members of the Scientific Committee and member
of the Spine Editorial Board, as well as presented in an open public
forum to the scientists who comprised the Orthopedic and
Rehabilitation Devices Advisory Panel to the FDA. Weaknesses and
strengths of the studies are readily apparent and have been
addressed by each author, as well as in my summation.
    12. The comment also stated that the articles should not have been
accepted for publication because the editorial policy of the journal
requires that the recommended minimum followup period for studies
should be 24 months.
    FDA disagrees. Under Spine policy, a sufficient length of time for
followup of articles is necessary for publication. While the
recommended time period for surgical procedures is 2 years, the policy
does not state that studies with less than 2-year followup will not be
published.

C. Issues Relating to the January 1995, 510(k) Substantial Equivalence
Determination for a Pedicle Screw Spinal System Intended for Severe
Spondylolisthesis

    13. A comment stated that the circumstances surrounding the first
510(k) clearance of a pedicle screw spinal system in January 1995, were
highly suspect because, until that time, FDA consistently had found
bone screws for use in the pedicles to be not substantially equivalent
to the identified predicate device, the lag screws used by Dr.
Harrington. The comment also stated that the lag screws were
manufactured as a custom device and used under a funded research grant
and, therefore, were not in commercial distribution prior to 1976.
    FDA disagrees. The 510(k) applicant provided new evidence
documenting, for the first time, that: (1) A medical device company had
manufactured and shipped in interstate commerce bone (lag) screws
intended for use in the pedicles of the spine prior to May 28, 1976;
(2) the devices were marketed to physicians, including, but not limited
to, Dr. Harrington; and (3) the devices were not used solely for
research purposes.
    14. The same comment also argued that the two devices had different
technological characteristics because the lag screws attach to fixation
constructs by wires whereas the pedicle screws attach directly to
plates or rods. The comment concluded that the applicant could not
demonstrate that its device did not raise different questions of safety
and effectiveness compared to the predicate device because the lag
screws were used on an extremely limited basis and were abandoned
because of a lack of effectiveness.
    FDA disagrees. The presence of technological differences does not
preclude a finding of substantial equivalence under section 513(i) of
the act. In accordance with section 513(i)(1)(A)) of the act and
Sec. 807.100(b)(2)(ii)(B), for purposes of determining substantial
equivalence, manufacturers have to demonstrate that their device (1)
Has the same intended use as a predicate device and (2) if it has
different technological characteristics than the predicate device, that
the device is as safe and as effective as a legally marketed device,
and it does not raise different questions of safety and effectiveness.
The relative extent of use of one device compared to another is not
relevant.
    In making its decision, FDA analyzed all of the data provided by
the sponsor. This included reports describing the clinical and
mechanical behavior of the device, in addition to affidavits. From
these data, the Panel and FDA determined that the complications were
similar to those of a predicate device and that the technological
differences raised no new questions relating to safety or
effectiveness.
    15. The comment also stated that FDA's reversal of its position
with regard to the preamendments status of pedicle fixation devices was
insupportable and a clear violation of its own regulations.
Specifically, the comment stated that the agency took the unprecedented
step of determining the existence of commercial distribution based
solely on the affidavit of a former employee of a pedicle screw
manufacturer. According to the comment, this was not sufficient
evidence to demonstrate that the device was in commercial distribution
prior to 1976.
    The use of affidavits to document the preamendments status of a
predicate device is not unprecedented. In fact, FDA routinely allows
affidavits to be used to document the preamendments status of a device.
FDA recognizes that obtaining labeling, advertising, and

[[Page 40033]]

other records concerning the marketing status of a device dating back
more than 20 years is often difficult, if not impossible. Therefore,
FDA allows sponsors to rely on alternative methods to demonstrate
interstate commerce. Moreover, contrary to the comment's statement, the
preamendments status of the device was established by much more than a
single affidavit. In fact, the 510(k) submission contained several
affidavits from individuals other than the sponsor, correspondence, and
other documents, e.g., shipping documentation, that demonstrated the
preamendments status of the Harrington lag screws for use in a limited
area of the spine, i.e., L<INF>5</INF>-S<INF>1</INF>, and for a
particular indication, i.e., severe spondylolisthesis.
    16. Finally, the comment alleged that FDA changed its regulatory
position regarding pedicle screw spinal systems after it made a
``deal'' with the affected industry on or about June 15, 1993. The
comment stated that, if manufacturers funded a retrospective study, FDA
provided assurances that it would (1) Refrain from taking criminal,
regulatory, or other legal actions against them; and (2) reclassify
pedicle screw spinal systems without requiring prospective studies and
without regard to the quality of any of the retrospective data.
    FDA disagrees. Prior to its January 1995, 510(k) decision and the
publication of this classification and reclassification regulation, FDA
consistently maintained that pedicle screw spinal systems, except when
intended for a very limited use, were class III devices requiring
premarket approval. The purpose of FDA's meeting with the affected
industry and the orthopaedic professional societies was to request that
these groups submit to the agency all available clinical data on the
performance of pedicle screw spinal systems. FDA, at no time, agreed to
change the regulatory status of these devices without regard to the
quality of the data or to refrain from taking regulatory action if a
retrospective study were funded.

D. Issues Relating to Misstatements or False Statements Appearing in
the Proposed Rule

    17. One comment alleged that the statement in the preamble to the
proposed rule regarding the conclusion of the August 20, 1993, Panel
meeting, i.e., that pedicle screw spinal systems appear to be safe and
effective when used as adjuncts to spinal fusion procedures, was
inaccurate.
    FDA disagrees. The description of the August 20, 1993, Panel
meeting contained in the preamble to the proposed rule states that the
Panel concluded that mechanical testing data demonstrated that pedicle
screw spinal systems exhibit adequate mechanical strength, rigidity,
and fatigue resistance (60 FR 51946 at 51948).
    18. The same comment alleged that neither the transcripts from the
two Panel meetings, nor the summary in the preamble to the proposed
rule accurately reflected the Panel's conclusions regarding potential
risks to health associated with the use of the pedicle screw spinal
system, special controls, development of performance standards,
mechanical performance of the device, and the Panel members' own
personal knowledge of, and clinical experience with, the device.
    FDA disagrees that the transcripts of the two Panel meetings did
not accurately reflect the Panel's conclusions. The proceedings from
the two meetings were verbatim stenographic transcripts of oral
testimony prepared by an independent transcriptionist. FDA also
disagrees that the preamble to the proposed rule did not accurately
reflect the Panel's conclusions. The preamble to the proposed rule
mirrors the transcripts of the meetings.
    19. The same comment alleged that the Panel members (voting members
and voting/nonvoting consultants), who met July 23, 1994, had
inappropriate relationships, e.g., financial arrangements and ex parte
communications, with pedicle screw spinal system manufacturers and had
participated substantially in the design of the Cohort study, thereby
compromising their impartiality.
    FDA disagrees in part. While it is expected that Panel members, who
are experts in a given field, will often have some financial interests
related to that field (e.g., certain arrangements with a manufacturer
(designing a device sold by a particular manufacturer; serving as a
consultant to a manufacturer; or receiving funding, directly or
indirectly, for research), the required FDA conflict-of-interest
questionnaire (FDA From 2725a) enables FDA to identify conflicts-of-
interest with a device or manufacturer that all substantial and/or
material to the subject of a particular Panel meeting, and thereby
facilitates the disclosure and possible waived for the Panel member(s)
in order to permit their participation in Panel deliberations.
    FDA performed an internal affairs investigation of the Panel
members regarding conflicts and ex parte communications . The agency
reviewed whether the Panel was properly constituted. Investigation of
alleged undisclosed and unwaived conflicts of interest held by Panel
members found minor disparities and reporting omissions for two voting
Panel members and one nonvoting consultant. The agency has concluded
these disparities and omissions were insignificant and did not
constitute financial conflicts of interest that would credibly
influence their recommendations.
    The agency has found that one other voting Panel member had
significant undisclosed financial conflicts. However, because the
recommendation of the Panel, both in the July 23 meeting and on the
subsequent homework assignment, was unanimous and this individual was
not controlling, of or unduly influential of, the votes of the other
Panel members and was not necessary to constitute a quorum, after
expunging the participation of this Panel member, FDA has concluded
that this Panel, both in the meeting and on the subsequent homework
assignment, was a valid scientific Panel for purposes of making
recommendations regarding classification and reclassification.

E. Issues Relating to FDA's Issuance of Regulations

    20. One comment argued that, in issuing a classification
regulation, FDA may not rely on a scientific study unless it makes
publicly available all study data, as well as the identities of the
persons who furnished the data. The comment cited 21 CFR 10.20(j),
20.63, and 860.5 as authority. In addition, the comment objected that
FDA refused to disclose the identities of the physician-investigators
who contributed data to the Cohort study, did not disclose the
reformatted IDE analysis, the IDE data, or internal information bearing
on the reliability of such data.
    FDA disagrees. Although the agency did not disclose the raw IDE or
the Cohort study data, or the identities of the clinical investigators
who furnished such data to the agency, FDA did provide a detailed
analysis of the Cohort Study, the clinical data released by the IDE
sponsors, and the meta-analysis (60 FR 51946 at 51960-51962; refs. 51,
65, 66, 119, and 201). FDA believes these publicly available data not
only satisfy the requirements under the statute, but provide the public
with at least the level of detailed information as that usually
available from published reports regularly relied upon to support
classification and reclassification.

F. Response to Comments Which Contained Clinical Data

    21. Several comments provided clinical information to support the
comment's position on the proposed

[[Page 40034]]

rule. The submitted clinical information consisted of literature
articles describing clinical trials and two questionnaires, a surgeon/
patient questionnaire and a lawyer/client questionnaire. The surgeon/
patient questionnaire provided mixed results, i.e., some patients were
satisfied with their clinical results and others were not satisfied,
whereas the lawyer/client questionnaire provided only negative results,
i.e., all clients were dissatisfied with their results.
    The majority of the articles submitted or referenced in these
comments were already reviewed by the Panel and used as part of the
basis for their recommendation to classify and reclassify pedicle screw
spinal systems into class II. The remainder of these articles were not
reviewed by the Panel because they were published after the July 1994,
Panel meeting. As described in section V.M of this document, these
articles did not raise new issues or concerns relating to the safety or
effectiveness of pedicle screw spinal systems. Because of the inherent
bias present in the questionnaires, e.g., the total number of
questionnaires sent to patients/clients in relation to the number
returned and the number included as part of the comment are unknown,
the data cannot be used in analyzing the success rate of pedicle screw
spinal systems. These data can be used, however, as part of an analysis
of the complications. As such, the questionnaires did not describe any
complications or raise any issues that had not already been reviewed by
the Panel and FDA in making their determinations with respect to the
classification and reclassification of pedicle screw spinal systems.

G. Requests for Additional Pedicle Screw Clinical Trials and Data
Analyses

    22. Ten comments requested that FDA require submission of
additional data before finalizing the classification and
reclassification of pedicle screw spinal systems. The comments
recommended that the following types of data be required: Studies to
analyze the long-term effects of the device, continuing evaluations,
collections of data using a recommended data report form for obtaining
data directly from patients rather than from their surgeons, studies
similar to the Cohort study but with larger sample sizes, comprehensive
reviews of the literature, and comprehensive reviews of all data. In
addition, one comment suggested that FDA was reclassifying these
devices without reviewing clinical trial data documenting their safety
and effectiveness.
    FDA disagrees. As previously explained, under section 513 of the
act, devices are classified and reclassified into one of three classes
based on reasonable assurance, not absolute proof, of their safety and
effectiveness. The Panel recommended, and FDA concurred, that pedicle
screw spinal systems intended to provide immobilization and
stabilization of spinal segments in skeletally mature patients as an
adjunct to fusion in the treatment of degenerative spondylolisthesis
and spinal fractures be classified and reclassified into class II
because they determined that premarket approval is not necessary to
provide reasonable assurance of safety and effectiveness; general
controls alone are insufficient to provide reasonable assurance of the
device's safety and effectiveness; and there is sufficient information
to establish special controls to provide such assurance. FDA also
determined that, when intended to provide immobilization and
stabilization of spinal segments in skeletally mature patients as an
adjunct to fusion in the treatment of the following acute and chronic
instabilities or deformities of the thoracic, lumbar, and sacral spine:
Degenerative spondylolisthesis with objective evidence of neurologic
impairment, fractures, dislocations, scoliosis, kyphosis, spinal
tumors, and failed previous fusion attempts (pseudarthrosis), special
controls would provide a reasonable assurance of safety and
effectiveness. The Panel and FDA reached these conclusions only after
considering a substantial amount of valid scientific evidence. As
described previously, this valid scientific evidence consisted of
clinical data collected from three sources--data from IDE's (the
reformatted IDE data), data from the literature (the meta-analysis),
and data collected directly from surgeons (the Cohort study). The IDE
data was prospective clinical data collected under the protocols of
FDA-approved clinical trials. The meta-analysis was retrospective
clinical data published in peer-reviewed literature. The Cohort study
consisted of retrospective nationwide clinical data collected from
surgeons of various experience levels from a patient population that
was homogeneous in terms of diagnosis, but mixed in terms of severity
of disease. In addition to these sources of clinical data, MDR and
MedWatch reports were analyzed for device problems. FDA does not
believe that it is necessary to require submission of additional data,
to conduct additional studies, or to re-review the literature before
classifying and reclassifying these devices. FDA does agree, however,
that the longer-term performance of these devices is not fully
characterized. For this reason, postmarket surveillance (PMS) studies
will be required.

H. Issues Relating to Indications for Use

    Over 200 comments addressed the various intended uses of pedicle
screw spinal systems.
    23. Twenty-three comments questioned FDA's authority to regulate
the indications for use of medical devices. They believed that,
although restrictions on the use of pedicle screw spinal systems may be
appropriate, this aspect of medical device regulation is outside the
scope of FDA's authority and should be decided by professional
societies, peer review groups, credentialing organizations, and
hospitals. One comment stated that FDA should regulate the safety of
medical devices only for certain indications. Several other comments
stated that there should be no restrictions on the use of pedicle screw
spinal systems. All of these comments argued that FDA's actions
interfered with the practice of medicine.
    FDA disagrees. In determining whether or not a device is safe and
effective, FDA first considers the intended uses for the device. Spinal
fusion is not a medical indication but a treatment option which can be
approached in a variety of ways. It is one of the desired outcomes from
using pedicle screw spinal systems. FDA recognizes, however, that
fusion in and of itself is not what patients with spinal disease are
seeking. They wish to be relieved of their symptoms, have their
objective impairment alleviated, and avoid more symptomatic or
functional impairment. Devices that share the same outcome for a given
condition do not necessarily share the same benefits and risks. One of
the aspects in determining if a device may be legally marketed is
deciding, based on the available data, what the appropriate indications
are. A device may be an appropriate treatment for one indication, but
not for another. In addition, to understand the evidence supporting a
device's safety and effectiveness, a distinct medical condition
requiring treatment must be identified. In reviewing the valid
scientific evidence, the Panel recommended and FDA found that the use
of pedicle screw spinal systems were safe and effective only for
certain indications. The valid scientific evidence did not support
unrestricted use of the device.
    In determining the safety and effectiveness of a device for the
purpose

[[Page 40035]]

of classification or reclassification, both the Panel and the agency
are to consider the persons for whose use the device is represented or
intended, the conditions of use for the device, and the probable
benefit to health from the use of the device weighed against any
probable injury or illness from such use (Sec. 860.7(b)). The device is
to be considered, not in a vacuum, but rather in the context of the
patient population for whose use it is intended. Accordingly, there is
reasonable assurance that a device is safe when it can be determined
that the probable benefits to health from use of the device for its
intended uses and conditions of use outweigh any probable risks
(Sec. 860.7(d)(1)). The benefits and risks to health presented by a
device depend, in large part, on the specific use for which the device
is intended. There may be reasonable assurance that a device is safe
for some, but not other, uses. Similarly, there is reasonable assurance
that a device is effective when it can be determined that, ``in a
significant portion of the target population,'' the use of the device
for its intended uses and conditions of use will provide clinically
significant results (Sec. 860.7(e)(1) (emphasis added)). It is clear,
then, that when making determinations regarding the classification or
reclassification of a device, it is appropriate for the agency to
consider the specific intended uses of a device, including the specific
patient populations for which it is intended. Consequently, the agency
disagrees that it does not have authority to regulate the indications
for use for pedicle screws and that it is interfering with the practice
of medicine.
    24. One comment objected that FDA's proposed reclassification
improperly exceeded the recommendations of the Panel.
    The Panel determined that the evidence demonstrated a reasonable
assurance of safety and effectiveness of pedicle screw spinal systems
intended for two severe and diagnostically distinct indications--
fracture and degenerative spondylolisthesis. Accordingly, the Panel
recommended that the device be classified and reclassified into class
II only when intended for these uses. FDA proposed that the device also
be classified and reclassified into class II when intended for the
following acute and chronic mechanical instabilities or deformities of
the thoracic, lumbar, and sacral spine: degenerative spondylolisthesis
with objective evidence of neurologic impairment, fractures,
dislocations, scoliosis, kyphosis, spinal tumors, and failed previous
fusion (pseudarthrosis). FDA disagrees that it exceeded its authority.
21 CFR 860.3(h) defines a classification panel as an advisory committee
established by the Commissioner for the purpose of making
``recommendations'' (emphasis added) to the Commissioner on the
classification/reclassification of devices. These recommendations are
designed to assist the Commissioner in the proper classification and/or
reclassification of a device. While FDA usually follows a Panel's
recommendations, it is not required to do so.
    As stated in the preamble to the proposed rule, FDA believes that
sufficient clinical data exist to classify and reclassify into class II
pedicle screw spinal systems intended for degenerative
spondylolisthesis with objective evidence of neurologic impairment,
fractures, dislocations, scoliosis, kyphosis, spinal tumors, and failed
previous fusion (pseudarthrosis). The medical literature and data from
IDE clinical investigations provide adequate evidence that the device
can safely and effectively stabilize the spine and maintain spinal
alignment while fusion takes place. The risks associated with the use
of pedicle screw spinal systems intended to provide immobilization and
stabilization of spinal segments as an adjunct to fusion in the
treatment of these acute and chronic instabilities or deformities of
the thoracic, lumbar, and sacral spine are similar to those associated
with other class II spinal implant devices, such as those classified in
Sec. 888.3050 (21 CFR 888.3050) (60 FR 51946 at 51956).
    25. Several comments advocated classifying and reclassifying into
class II pedicle screw spinal systems intended for additional uses,
including degenerative disc disease, degenerative deformities,
stenosis, iatrogenic instability and previous multiple laminectomies,
facet joint disease, pseudospondylolisthesis, low back pain, disc
herniation, arthritis, and osteomyelitis.
    FDA believes that valid scientific evidence does not currently
exist to support classifying and reclassifying into class II pedicle
screw spinal systems when intended for the indications listed above.
Neither the literature nor the clinical data establish the safe and
effective use of pedicle screw spinal systems for degenerative disc
disease, degenerative deformities, stenosis, iatrogenic instability and
previous multiple laminectomies, facet joint disease,
pseudospondylolisthesis, low back pain, disc herniation, arthritis, or
osteomyelitis. As stated in the preamble to the proposed rule, FDA has
determined that, when intended for use in conditions not categorized as
acute or chronic instabilities or deformities of the thoracic, lumbar,
and sacral spine, premarket approval is necessary to ensure the safety
and effectiveness of the device (60 FR 51946 at 51957). FDA-approved
clinical trials for some of these indications are ongoing. When data
from these or other studies become available for any of the indications
described above, they may be submitted in either an application for
premarket approval or reclassification petition.
    26. Eight comments advocated adding specific pediatric indications
and one comment advocated adding general pediatric use to the list of
indications. The specific indications included myelodysplasia, spina
bifida, cerebral palsy, muscular dystrophy, myelomeningocele, and
congenital subluxation.
    FDA disagrees. As stated previously, all valid scientific evidence
reviewed by the Panel and FDA were obtained from skeletally mature
populations. To date, the safety and effectiveness of pedicle screw
spinal systems in pediatric populations have not been demonstrated.
Therefore, this patient population is excluded from this classification
and reclassification. When intended for use in pediatric populations,
pedicle screw spinal systems are considered postamendments class III
devices for which premarket approval is required.
    27. Several comments addressed ways in which FDA should further
limit the indications for use of pedicle screw spinal systems, such as
by including specific patient evaluation criteria or by specifying the
severity of the condition.
    FDA disagrees that these actions are necessary. FDA classifies
devices based upon, among other things, patient selection, not
individual patient management. FDA notes that it is the responsibility
of individual surgeons to determine the appropriateness of using a
specific medical device for a given patient.
    28. Four comments stated that pedicle screw spinal systems should
not be allowed on the market for any use. Another comment requested
that an additional Panel meeting be convened to discuss further
restricting the intended uses of pedicle screw spinal systems.
    FDA disagrees. After reviewing all available data and information,
FDA believes that there is reasonable assurance that pedicle screw
spinal systems are safe and effective for certain intended uses. FDA
does not believe that pedicle screw spinal systems present a
substantial deception or an

[[Page 40036]]

unreasonable and substantial risk of illness or injury. Consequently,
FDA does not believe it would be appropriate to ban them under section
516 of the act (21 U.S.C. 360f).
    FDA also disagrees that an additional Panel meeting is necessary
because the relevant available data have been reviewed.

I. Issues Relating to Special Controls

    29. One comment asserted that PMS studies cannot legally be
required for pedicle screw spinal systems because the devices are not
intended for use in supporting or sustaining life and pose risks no
different from those associated with the use of other preamendments
class II spinal fixation devices.
    FDA disagrees. Under section 522 of the act (21 U.S.C. 360l),
postmarket surveillance is required for certain devices and may be
required for any device for which FDA determines that it is necessary
to protect the public health or to provide safety or effectiveness data
for the device. FDA has determined that PMS studies are necessary to
provide longer-term data on the safety and effectiveness of pedicle
screw spinal systems.
    Although originally proposed as a special control, FDA has
determined that PMS studies are best imposed by order in the
substantial equivalence determination letter for each device. This will
preserve the discretionary nature of the PMS studies and will allow the
agency to more easily remove the requirement once it determines that
these studies are no longer necessary to assure the safety and
effectiveness of pedicle screw spinal systems. The final regulation has
been modified to reflect that PMS studies are no longer one of the
special controls for these devices.
    30. One comment stated that PMS studies are appropriate only for
devices cleared for marketing with limited clinical performance data.
The comment noted that there now exists a vast amount of clinical
information gained from use of pedicle screw spinal systems in several
thousand patients. The comment also noted that, based on these data,
the Panel concluded that, with respect to safety and effectiveness,
these devices are comparable to, or better than, currently available
spinal systems. The comment concluded that this clinical information
and the conclusions drawn from this information provide sufficient
clinical data to adequately identify and characterize the performance
of pedicle screw spinal systems and the issues pertinent to safety and
effectiveness, thereby obviating the need to conduct PMS studies.
    FDA disagrees that PMS studies are appropriate only for devices
cleared for marketing with limited clinical performance data. Section
522 of the act allows FDA to require PMS studies for any device for
which it determines such studies would protect the public health or
provide safety or effectiveness data for the device. As stated in the
preamble to the proposed rule, FDA will require PMS studies in order to
address issues related to device specific design differences, surgical
techniques, and device usage (60 FR 51946 at 51955). Although there is
ample short-term clinical performance data for these devices, there
does not now exist sufficient longer-term, i.e., more than 24-month
followup, safety and/or effectiveness data regarding device specific
design differences, surgical techniques, and device usage.
    31. A second comment noted that components used to construct
pedicle screw spinal systems could be identical to those used to
construct either spinal interlaminal fixation orthoses (Sec. 888.3050)
or spinal intervertebral body fixation orthoses (21 CFR 888.3060).
Because PMS studies are not required for these devices, they should not
be required for pedicle screw spinal systems. A third comment believed
that PMS studies are inappropriate for well-established, standard of
care treatments involving medical devices that were in existence prior
to the 1976 amendments, including pedicle screw spinal systems.
    FDA disagrees that PMS studies are inappropriate for devices that
were in existence prior to the 1976 amendments. Section 522(a)(2) of
the act specifically authorizes FDA to require a manufacturer to
conduct PMS studies for any device, regardless of when it was first
introduced or delivered for introduction into interstate commerce, for
which FDA determines that PMS studies are necessary to protect the
public health or to provide safety or effectiveness data for the
device. Although, as the comment states, certain devices have been used
as pedicle screw spinal systems for some time, except for the limited
severe spondylolisthesis intended use available since January 1995,
pedicle screw spinal systems have not been legally marketed. Collection
of the PMS study data will allow FDA to analyze information on the use
of devices specifically intended, and legally marketed, for use as
pedicle screw spinal systems.
    32. Five comments believed that PMS studies are unnecessary and
will not further protect the public health because one or more of the
following current reporting systems already provides adequate
information on the performance of pedicle screw spinal systems: (1) The
MDR System, (2) Voluntary Reporting under MedWatch, (3) User Reporting,
and (4) Complaint Handling under the current good manufacturing
practices. One comment supported a requirement that labeling remind
surgeons they are required to report certain events under MDR. Two
comments suggested that a statement which encourages health care
professionals to submit MDR's under the Voluntary MedWatch System be
placed in the required package insert of the device. Two other comments
noted that no other class II spinal implant device is subject to PMS
studies. Three comments also stated that collecting additional
information will increase health care costs.
    FDA disagrees in part. The purposes of PMS studies and current
reporting systems are different. PMS studies are active investigations
of device performance during actual use, whereas other reporting
systems, i.e., MedWatch, MDR, User Reporting, and Complaint Reporting,
are passive reporting mechanisms. As such, these current reporting
systems would not provide the agency with clinical monitoring
information on pedicle screw spinal systems other than unexpected
problems in the marketplace. The PMS studies, in contrast, will provide
longer-term safety and effectiveness data for pedicle screw spinal
systems once the devices are distributed in the general population
under actual conditions of use. Finally, FDA is aware that PMS studies
might have an impact on health care costs. Although this is
unfortunate, the agency believes that it is necessary to impose this
requirement and collect this information in order to assure the safety
and effectiveness of pedicle screw spinal systems.
    33. A comment suggested that, due to the litigious climate
surrounding these devices, it may be very difficult for manufacturers
to recruit surgeons to participate in PMS studies.
    FDA recognizes the concern that there may be conditions which would
make the collection of the data somewhat difficult. However, FDA
believes that it is important that the data be obtained and that it is
possible to recruit a sufficient number of surgeons to participate in
PMS studies.
    34. One comment stated that the proposed identification for pedicle
screw spinal systems was inaccurate, or at least misleading. The
comment noted that, as proposed, a pedicle screw spinal system assembly
must contain all of the components listed as part of the pedicle screw
spinal system. The comment stated that, for any given assembly,

[[Page 40037]]

some or all of the system components could be used.
    FDA agrees in part. As proposed, the identification could be
interpreted to require that all of the described components were
necessary to construct a pedicle screw spinal system assembly. FDA has
amended the identification of the device to clarify that not all of the
described components are required to be used in a pedicle screw spinal
system assembly.
    35. In the preamble to the proposed rule, FDA proposed two labeling
special controls. These controls described the intended uses and
indications for pedicle screw spinal systems and cautioned the user
about potential risks to health if the devices were used under certain
conditions. Three comments stated that the two labeling special
controls were incorrectly categorized as ``warnings'' according to
FDA's General Program Memorandum No. G91-1, ``Device Labeling
Guidance.'' They believed that these labeling requirements are more
appropriately described as ``precautions'' or ``important notes''
because they describe a particular patient population and not specific
risks or hazards associated with the use of a device. Four comments
objected that: (1) Use of the phrase ``* * * with significant potential
risk for serious injury to patients * * *'' in the second labeling
statement did not accurately reflect the data reviewed by the Panel to
make its recommendation,(2) references to training and experience
should not be part of the second labeling special control, and (3) the
controls containing the language referred to in (1) and (2) should be
removed or modified.
    FDA agrees with the comments that the two labeling special controls
should be rewritten, but disagrees with the specific reasons. General
Program Memorandum No. G91-1 states that ``A warning is appropriate
when the device is commonly used for a disease or condition for which
there is a lack of valid scientific evidence of effectiveness for that
disease or condition and such usage is associated with a serious risk
to health * * * Include an appropriate warning if there is reasonable
evidence of an association of a serious hazard with the use of the
device. A causal relationship need not have been proved.'' This is the
case when pedicle screw spinal systems are used for indications other
than significant mechanical instabilities or deformities of the
thoracic, lumbar, and sacral spine. Because valid scientific evidence
is not available to support a determination that a reasonable assurance
exists that pedicle screw spinal systems are safe and effective for
other indications, categorizing the first labeling special control as a
``warning'' is the appropriate mechanism to alert users to the
potential for injury to a patient.
    The second labeling special control does not warrant being
described as a ``warning'' because it does not meet the definition of
this term. It does not describe known serious adverse reactions or
known potential safety hazards; it does not provide specific steps to
be taken; it does not concern a use for which there is reasonable
evidence of association with a serious hazard. It does, however,
provide information on special care to be exercised by a practitioner,
although the need for special care is implied, not explicitly stated.
Accordingly, FDA concludes that it is more appropriately categorized as
a ``precaution''.
    After reviewing the proposed special controls regarding labeling,
FDA has concluded that the information should be stated more clearly.
FDA believes that the labeling special controls reflect the data
reviewed by the Panel. FDA also believes that the labeling special
controls are necessary to provide reasonable assurance of the safety
and effectiveness of the devices. Finally, as described in the next
section, the intent of the second control was not to specify the type
of training that should be available or to suggest that FDA would
provide or approve any training. Rather, it was intended to alert
surgeons to the necessity of receiving appropriate training in the use
of specific pedicle screw spinal systems. Because of concerns with the
proposed wording, the labeling special controls have been modified to
read as follows:
    ``Warning: The safety and effectiveness of pedicle screw spinal
systems have been established only for spinal conditions with
significant mechanical instability or deformity requiring fusion
with instrumentation. These conditions are significant mechanical
instability secondary to degenerative spondylolisthesis with
objective evidence of neurologic impairment, fracture, dislocation,
scoliosis, kyphosis, spinal tumor, and failed previous fusion
(pseudarthrosis). The safety and effectiveness of these devices for
any other conditions are unknown.''
    ``Precaution: The implantation of pedicle screw spinal systems
should be performed only by experienced surgeons with specific
training in the use of this pedicle screw spinal system because this
is a technically demanding procedure presenting a risk of serious
injury to the patient.''
    36. A number of comments stated that appropriate surgeon training
should be required prior to use of pedicle screw spinal systems and
that classification/reclassification into class II would make access to
training and device information easier. In addition, several comments
believed that professional societies and hospitals, not FDA or the
manufacturers, should determine what constitutes adequate training for
surgeons implanting pedicle screw spinal systems.
    FDA agrees that it is important that surgeons who use pedicle screw
spinal systems have proper training prior to using the device. FDA does
not believe, however, that it should identify who is most qualified to
provide such training or determine what constitutes adequate training.
The precaution statement is intended to inform surgeons (and patients)
of the possible effect the device could have on the patient if the
surgeon implanting the device is not trained or experienced in the
proper use of pedicle screw spinal systems. This includes knowledge of
the indications, patient selection criteria, and appropriate surgical
techniques.
    37. A comment questioned the proposed warning label because, in the
past, FDA has prohibited pedicle screw spinal system manufacturers from
supporting courses that described surgical techniques of ``off label''
uses demonstrating such uses or providing hands-on workshops to learn
such uses.
    FDA disagrees. Previously, the agency issued several warning
letters to pedicle screw spinal system manufacturers for participating
in or supporting the training of practitioners in the use of long bone
screw, pedicle fixation because, at that time, no long bone screw
devices had received FDA clearance for use in the pedicles of the
spine. As a result, FDA considered such use ``off label.'' Because the
association with these training programs was considered the promotion
of an ``off-label'' use, the agency stated that the manufacturers had
misbranded and adulterated the long bone screws in accordance with
sections 501(f)(1)(B) and 502(o) of the act (21 U.S.C. 351(f)(1)(B) and
352(o)) and promotion of this use was considered a major modification
of the intended use, requiring a new premarket notification (510(k))
submission under Sec. 807.81(a)(3)(ii). The regulations and the act are
clear that manufacturers must have clearance for the intended use for
which their device(s) are promoted, advertised, or held for sale.
    With the issuance of this final regulation, the agency now
encourages pedicle screw spinal system manufacturers to support
training for the class II intended uses. Such training, however, should
not be provided before FDA clearance is received. The above referenced
warning label will appear

[[Page 40038]]

only on devices that have been cleared for pedicle screw spinal
fixation.
    38. The comment also claimed that the right to free speech
guaranteed by the First Amendment to the U.S. Constitution should not
be restricted by FDA's suppression of training for ``off label'' use.
    FDA disagrees that its limitations on promotional training
conducted or sponsored by manufacturers for ``off label'' uses for
pedicle screw spinal systems violate the First Amendment. As described
above, the act requires that FDA regulate devices based on their
intended use. The term ``intended use'' is broadly defined and
encompasses the manner in which a company characterizes its product in
the marketplace. The intended use of a device refers to the objective
intent of the persons legally responsible for its labeling (Sec. 801.4
(21 CFR 801.4)). ``The intent is determined by such persons'
expressions or may be shown by the circumstances surrounding the
distribution of the article. This objective intent may, for example, be
shown by labeling claims, advertising matter, or oral or written
statements by such persons or their representatives.'' ( Sec. 801.4
(emphasis added)); (see e.g., Coyne Beahm, Inc. et al. v. United States
Food and Drug Administration, et al., 958 F. Supp. 1060 (M.D.N.C.
1997).) Consequently, oral statements and materials presented at
industry-supported training programs may provide evidence of a device's
intended use. If these statements or materials promote a use that has
not been approved by the agency, the device is misbranded under section
502(f)(1) of the act for failure to bear labeling with adequate
directions for all intended uses, and under section 502(o) of the act
because premarket notification was not provided as required under
section 510(k) of the act. The device is also adulterated under section
501(f) of the act for failure to have FDA approval. Thus, the various
means by which manufacturers and their representatives provide
information about their products to healthcare professionals and
consumers, including statements and materials presented at industry-
supported scientific and educational activities, directly bear on
whether a device is improperly promoted and, therefore, adulterated or
misbranded.
    Because the regulation of devices is an area of extensive Federal
regulation, the agency may regulate the communications at industry-
supported scientific and educational activities without violating the
First Amendment. (Cf. SEC v. Wall Street Publishing Institute, Inc.,
851 F.2d 365 (D.C.Cir. 1988), cert. denied, 109 S.Ct. 1342 (1989).)
Moreover, to the extent that such communications constitute protected
speech, they are commercial speech and FDA's regulation of such
activities does not violate the First Amendment. (See Bolger v. Youngs
Drug Products, 103 S.Ct. 2875 (1983); S.U.N.Y. v. Fox, 109 S.Ct. 3028
(1989); Cincinnati v. Discovery Network, 113 S.Ct. 1505 (1993).)
Industry-supported scientific educational activities refer to a
specific product, are economically motivated, and propose a commercial
transaction. These programs are intended to convince the audience to
prescribe, purchase, or otherwise use the particular product.
    The Supreme Court has afforded commercial speech limited
constitutional protection. (See, e.g., Virginia State Board of Pharmacy
v. Virginia Citizens Consumer Council, Inc., 96 S.Ct. 1817 (1976);
Central Hudson Gas & Electric Corp. v. Public Service Commission, 100
S. Ct. 2343 (1990).) In Central Hudson, the Supreme Court established a
four-prong test to determine whether limitations on commercial speech
are constitutional. The four prongs are: (1) Whether the speech
concerns lawful activity and is not misleading, (2) whether the
asserted government interest is substantial, (3) whether the limitation
directly advances the governmental interest asserted, and (4) whether
the limitation is not more extensive than is necessary to serve that
interest. The Court has clarified that the fourth prong requires that
the restriction be ``narrowly tailored'' to serve the asserted
government interest. Narrow tailoring means a fit that is reasonable.
(See S.U.N.Y. v. Fox, 109 S.Ct. 3028. 3035 (1989).)
    FDA's regulation of industry-supported scientific and educational
activities satisfies all four prongs. First, as previously discussed,
industry-supported scientific and educational activities that promote
an unapproved device, or promote an approved device for an unapproved
use, create an unlawful product--a misbranded or adulterated device.
Accordingly, industry-supported activities that promote unlawful
products concern illegal activity and may be prohibited. Second, FDA's
limitations on promotional activities with respect to off label uses
serve the substantial government interest of protecting the public
health and safety by helping to ensure the dissemination of truthful
and nonmisleading information about devices. The Supreme Court has
repeatedly held that the government's ``interest in the health, safety,
and welfare of its citizens constitutes a substantial interest.''
(Posadas de Puerto Rico Associates v. Tourism Co., 106 S.Ct. 2968, 2977
(1986); Rubin v. Coors, 115 S.Ct. 1585, 1591 (1995).) The limitations
also serve the second substantial government interest of protecting the
public health by preserving the integrity of the premarket approval
process under which manufacturers are required to establish that their
devices are safe and effective for each of their intended uses before
they may be marketed and promoted for those uses. Third, FDA's
limitations on promotional activities with respect to off-label uses
directly advance the government's substantial interests in protecting
the public health and safety by helping to ensure the dissemination of
truthful and nonmisleading information about devices and by preserving
the integrity of the premarket approval process by dissuading
manufacturers from using such activities as a means to promote
unapproved products and unapproved uses, thereby encouraging scientific
research and avoiding unnecessary harm to patients. Finally, FDA's
limitations on industry-sponsored training sessions are narrowly
tailored and are a reasonable approach to protect the public health and
safety by discouraging the dissemination of misleading or biased
information, and by maintaining the integrity of the premarket approval
process. FDA's limitations apply only to industry-supported activities
that relate to the supporting company's device or to competing devices.
They are directed to the regulated sponsors of such activities, and do
not apply to participating professionals or independent scientists and
organizations.
    39. Several comments believed that the device should be available
for use only by neurosurgeons or orthopaedic surgeons supervised by
neurosurgeons.
    FDA disagrees. According to section 520(e)(1)(B), FDA may not
restrict access to medical devices based on specialty or board
certification.

J. Other Issues

    40. Several comments objected that publication of the proposed rule
in the Federal Register was not appropriate because the general public
is not aware of the Federal Register. The comments noted that another
vehicle for disseminating the information would have been more
appropriate.
    FDA disagrees. The act (sec. 513(d)(1) and 513(e)(1)) requires that
a proposed rule be published in the Federal Register as the formal
mechanism to provide all interested parties an opportunity to submit
comments when

[[Page 40039]]

an advisory panel recommends an initial classification or change in
classification for a medical device. Comments are invited from anyone.
FDA recognizes that other mechanisms for distribution of this type of
information is also appropriate. One of the alternate mechanisms
currently being tested is electronic publication on the World Wide Web.
    41. Several comments objected to FDA's consideration of public
comments, which may contain only anecdotal information, in determining
the appropriate class for these devices.
    FDA agrees that comments provided by the public may contain
anecdotal information that does not meet the definition of valid
scientific evidence. However, FDA considers this information along with
the information provided in other comments. These anecdotal comments
did not raise any issues or comments that were not already addressed by
the information that the Panel reviewed in making its determination
that safety and effectiveness of pedicle screw spinal systems could be
assured by special controls.
    42. Six comments disapproved of the release of the PIN's which
identified the surgeons participating in the Cohort study.
    FDA regrets any problems that may have been caused by this
inadvertent release of information. However, release of this
information did not affect the quality, integrity, or value of the data
upon which the Panel's recommendation was based.
    43. A comment noted that there is no consensus among spine surgeons
that pedicle screw fixation has become the standard of care or the gold
standard for treatment of spinal instability so as to justify the
conclusion that the devices are safe and effective and to justify
abandonment of the randomized control trial in making such an
assessment.
    FDA agrees that there is no consensus among spine surgeons
regarding pedicle screw spinal systems. However, a medical device does
not need to be viewed as the ``gold standard'' in order for the agency
to determine that there is reasonable assurance of its safety and
effectiveness. Nor is it a requirement for the classification and
reclassification process that all members of a medical specialty agree
that a particular device should be used under all conditions. It is
recognized that certain devices provide their best outcome when used
for specific indications. This is one of the reasons why degenerative
disc disease is not included as one of the intended uses in the
classification and reclassification of pedicle screw spinal systems.
Finally, as described above, randomized clinical trials are only one of
the types of valid scientific evidence upon which FDA may rely in
support of a classification/reclassification determination. Many IDE
studies from which the reformatted IDE data came are still being
actively pursued by their sponsors and the patients are being actively
followed.

K. Labeling of Bone Screws

    44. A comment requested FDA to formally rescind its April 8, 1994,
and June 15, 1994, letters to manufacturers of bone screws and devices
classified under Secs. 888.3030 and 888.3040 (21 CFR 888.3030 and
888.3040), directing them to amend their labeling by including the
following: ``Warning: This device is not approved for screw attachment
or fixation to the posterior elements (pedicles) of the cervical,
thoracic, or lumbar spine.'' According to the comment, this labeling
requirement will become unnecessary when pedicle screw spinal systems
are classified into class II.
    FDA disagrees. In this final rule, FDA is classifying and
reclassifying only pedicle screw spinal systems intended for screw
attachment or fixation to the pedicles of the thoracic, lumbar, or
sacral spine for immobilization and stabilization of spinal segments
for the treatment of significant medical instability or deformity
requiring fusion with instrumentation. This classification and
reclassification in no way affects devices classified as single/
multiple component metallic bone fixation appliances and accessories
(Sec. 888.3030) or smooth or threaded metallic bone fixation fasteners
(Sec. 888.3040). Those devices are still not approved for screw
attachment or fixation to the posterior elements (pedicles) of the
cervical, thoracic, or lumbar spine. Hence, the 1994 amended labeling
remains appropriate for these devices.
    45. One comment noted that in January 1995, FDA began clearing
510(k)'s intended to treat grades 3 and 4 spondylolisthesis at the
L<INF>5</INF>-S<INF>1</INF> junction. The comment concluded that, by
default, grades 1 and 2 spondylolisthesis, less severe conditions, are
considered to be postamendments intended uses resulting in the device
being automatically classified into class III. According to this
comment, this means that FDA, through required manufacturer labeling,
is instructing physicians to wait until grades 1 and 2
spondylolisthesis develop into grades 3 and 4 spondylolisthesis before
employing treatments utilizing pedicle screw spinal systems, which is
not in the patient's best interests.
    FDA disagrees. FDA is not limiting physicians, through required
manufacturer labeling, to wait until grades 1 or 2 spondylolisthesis
develop into grades 3 or 4 spondylolisthesis. FDA is stating that the
preamendments documentation in the 510(k) described marketing of the
device only for the treatment of grades 3 and 4 spondylolisthesis at
L<INF>5</INF>-S<INF>1</INF>. Treatment of grades 1 or 2
spondylolisthesis does not have to wait until it progresses to grades 3
or 4. Legally marketed devices which do not utilize pedicle screws are
available for this purpose.

L. Review of New Pedicle Screw Spinal System 510(k)'s

    46. A comment pointed out that since FDA's January 1995,
determination regarding the preamendments status of pedicle screw
spinal systems in the treatment of severe spondylolisthesis, many
510(k) submissions have been cleared for this use. FDA's proposed rule
for pedicle screw spinal systems, once final, will essentially
represent a labeling change for these devices, requiring new 501(k)
submissions. The comment suggested that the new 510(k)'s should provide
a draft copy of the revised labeling and a statement that the
previously-cleared device has not been modified in any way that may
affect its safety or effectiveness. According to the comment, this
limited type of review would facilitate and expedite the review process
and would not unnecessarily burden FDA's device evaluation staff.
    FDA agrees with this approach and intends to apply it in its review
of 510(k)'s for pedicle screw spinal systems that were cleared
previously for use in severe spondylolisthesis. Pedicle screw spinal
systems which have not been previously reviewed, or that represent
significant modifications compared to the previously cleared device(s),
will require a complete 510(k) submission, including the device
labeling.

M. Review of New Information Published and Submitted After Publication
of the Proposed Rule: Pedicle Screw and Related Literature and MedWatch
and MDR System Reports

    FDA performed a comprehensive search of the English-language
medical literature published between 1994 and the present. Thirty-five
articles pertained to the clinical performance of pedicle screw spinal
systems. The clinical performance results, e.g., fusion rate and
complication types and rates,

[[Page 40040]]

from these peer-reviewed articles did not differ from those previously
reported in the preamble to the proposed rule for either pedicle screw
spinal systems or the group of class II spinal devices using hooks and/
or wires or noninstrumented fusions.
    FDA also performed a review of the MedWatch and MDR databases from
1994 to the present. The complications associated with pedicle screw
spinal systems during this period were comparable to those reported in
the preamble to the proposed rule for pedicle screw spinal systems and
the group of class II spinal devices using hooks and/or wires and
noninstrumented fusions.

VI. References

     The following references have been placed on display in the
Dockets Management Branch (HFA-305), Food and Drug Administration, 5630
Fishers Lane, rm. 1061, Rockville, MD 20852. These references may be
seen by interested persons between 9 a.m. and 4 p.m., Monday through
Friday.
     1. Rodgers, A. E., ``FDA Pedicle Screw Cohort Study: Audit
Findings,'' July 30, 1996.
     2. Richter, K. C., ``Assessment of the Impact of BIMO Audit
Findings for the Pedicle Screw Cohort Study on Study Results,'' August
29, 1997.

VII. Environmental Impact

    The agency has determined under 21 CFR 25.24(e)(2) that this action
is of a type that does not individually or cumulatively have a
significant effect on the human environment. Therefore, neither an
environmental assessment nor an environmental impact statement is
required.

VIII. Analysis of Impacts

    FDA has examined the impact of the final rule under Executive Order
12866 and the Regulatory Flexibility Act (5 U.S.C. 601-612) (as amended
by Subtitle D of the Small Business Regulatory Fairness Enforcement Act
of 1996 (Pub. L. 104-121), and the Unfunded Mandates Reform Act of 1995
(Pub. L. 104-4)). Executive Order 12866 directs agencies to assess all
costs and benefits of available regulatory alternatives and, when
regulation is necessary, to select regulatory approaches that maximize
net benefits (including potential economic, environmental, public
health and safety, and other advantages; distributive impacts; and
equity). The agency believes that this final rule is consistent with
the regulatory philosophy and principles identified in the Executive
Order. In addition, the final rule has been determined to be a
significant regulatory action as defined by the Executive Order and so
is not subject to review under the Executive Order.
    The Regulatory Flexibility Act requires agencies to analyze
regulatory options that would minimize any significant impact of a rule
on small entities. Classification and reclassification of the device
from class III to class II when the device is intended to provide
immobilization and stabilization of spinal segments as an adjunct to
fusion in the treatment of the following acute and chronic
instabilities or deformities of the thoracic, lumbar, and sacral spine:
Degenerative spondylolisthesis with objective evidence of neurologic
impairment, fractures, dislocations, scoliosis, kyphosis, spinal
tumors, and failed previous fusion attempts (pseudarthrosis) will
relieve all manufacturers of the device of the cost of complying with
the premarket approval requirements in section 515(b) of the act.
    Because classification and reclassification will reduce regulatory
costs with respect to this device, it will not impose significant
economic impact on any small entities, and it may permit small
potential competitors to enter the marketplace by lowering their costs.
The Commissioner of Food and Drugs, therefore, certifies that this
final rule will not have a significant economic impact on a substantial
number of small entities. In addition, this final rule will not impose
costs of $100 million or more on either the private sector or State,
local, and tribal governments in any one year, therefore, a summary
statement of analysis under section 202(a) of the Unfunded Mandates
Reform Act of 1995 is not required.

List of Subjects in 21 CFR Part 888

    Medical devices.
    Therefore, under the Federal Food, Drug, and Cosmetic Act and under
authority delegated to the Commissioner of Food and Drugs, 21 CFR part
888 is amended as follows:

PART 888--ORTHOPEDIC DEVICES

    1. The authority citation for 21 CFR part 888 continues to read as
follows:

    Authority: 21 U.S.C. 351, 360, 360c, 360e, 360j, 371.

    2. Section 888.3070 is added to subpart D to read as follows:

Sec. 888.3070  Pedicle screw spinal system.

    (a) Pedicle screw spinal systems--(1) Identification. Pedicle screw
spinal systems are multiple component devices, made from a variety of
materials, including alloys such as 316L stainless steel, 316LVM
stainless steel, 22Cr-13Ni-5Mn stainless steel, Ti-6Al-4V, and
unalloyed titanium, that allows the surgeon to build an implant system
to fit the patient's anatomical and physiological requirements. Such a
spinal implant assembly consists of a combination of anchors (e.g.,
bolts, hooks, and/or screws); interconnection mechanisms incorporating
nuts, screws, sleeves, or bolts; longitudinal members (e.g., plates,
rods, and/or plate/rod combinations); and/or transverse connectors. The
devices are intended to provide immobilization and stabilization of
spinal segments in skeletally mature patients as an adjunct to fusion
in the treatment of the following acute and chronic instabilities or
deformities of the thoracic, lumbar, and sacral spine: degenerative
spondylolisthesis with objective evidence of neurologic impairment,
fracture, dislocation, scoliosis, kyphosis, spinal tumor, and failed
previous fusion (pseudarthrosis).
    (2) Classification. Class II (special controls). Pedicle screw
spinal systems must comply with the following special controls:
    (i) Compliance with material standards,
    (ii) Compliance with mechanical testing standards,
    (iii) Compliance with biocompatibility standards, and
    (iv) Labeling which contains these two statements in addition to
other appropriate labeling information:
    ``Warning: The safety and effectiveness of pedicle screw spinal
systems have been established only for spinal conditions with
significant mechanical instability or deformity requiring fusion
with instrumentation. These conditions are significant mechanical
instability or deformity of the thoracic, lumbar, and sacral spine
secondary to degenerative spondylolisthesis with objective evidence
of neurologic impairment, fracture, dislocation, scoliosis,
kyphosis, spinal tumor, and failed previous fusion (pseudarthrosis).
The safety and effectiveness of these devices for any other
conditions are unknown.''
    ``Precaution: The implantation of pedicle screw spinal systems
should be performed only by experienced spinal surgeons with
specific training in the use of this pedicle screw spinal system
because this is a technically demanding procedure presenting a risk
of serious injury to the patient.''
    (b) Pedicle screw spinal systems for all other uses--(1)
Identification. Pedicle screw spinal systems for all other uses are
multiple component devices, made from a variety of materials, including
alloys such as 316L stainless steel, 316LVM stainless steel, 22Cr-13Ni-
5Mn stainless steel, Ti-6Al-4V, and unalloyed titanium, that allow

[[Page 40041]]

the surgeon to build an implant system to fit the patient's anatomical
and physiological requirements. Such an spinal implant assembly
consists of a combination of anchors (e.g., bolts, hooks, and/or
screws); interconnection mechanisms incorporating nuts, screws,
sleeves, or bolts; longitudinal members (e.g., plates, rods, and/or
plate/rod combinations); and/or transverse connectors.
    (2) Classification. Class III (premarket approval).
    (c) Date PMA or notice of completion of a PDP is required. An
approved PMA or a declared completed PDP must be in effect before
placing the device in commercial distribution. See Sec. 888.3.

    Dated: April 22, 1998,
Michael A. Friedman,
Lead Deputy Commissioner for the Food and Drug Administration.
Donna E. Shalala,
Secretary of Health and Human Services.
[FR Doc. 98-19944 Filed 7-23-98; 8:45 am]
BILLING CODE 4160-01-F