United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT
Argued September 21, 2012 Decided March 22, 2013
No. 11-1268
CYTORI THERAPEUTICS, INC.,
PETITIONER
v.
FOOD & DRUG ADMINISTRATION,
RESPONDENT
Consolidated with 11-1279
On Petitions for Review of Orders
of the Food & Drug Administration
Andrew S. Ittleman argued the cause for petitioner. With
him on the briefs was Mitchell Fuerst.
Adam C. Jed, Attorney, U.S. Department of Justice,
argued the cause for respondent. With him on the brief were
Stuart F. Delery, Acting Assistant Attorney General, and
Scott R. McIntosh, Attorney. Douglas N. Letter, Attorney,
entered an appearance.
Before: BROWN and KAVANAUGH, Circuit Judges, and
SENTELLE, Senior Circuit Judge.
2
Opinion for the Court filed by Circuit Judge KAVANAUGH.
KAVANAUGH, Circuit Judge: In Administrative Procedure
Act cases alleging arbitrary and capricious agency action,
courts must be careful not to unduly second-guess an agency’s
scientific judgments. That basic principle of administrative
law controls this case.
The Food and Drug Administration must approve certain
medical devices before they are marketed. Here, Cytori
Therapeutics applied to FDA to market two new medical
devices, the Celution 700 and the StemSource 900. Those
two devices use adipose tissue – that is, fat – as a source of
stem cells that could later be used in lab analysis or,
potentially, in regenerative medicine. The most similar
devices on the market extract stem cells from blood or bone
marrow.
Federal law establishes two basic paths for FDA approval
of new medical devices. One is the “premarket approval”
process. See 21 U.S.C. § 360e. That process generally
requires extensive clinical research on a new device to ensure
the device’s safety, and it often takes significant time. The
other is the streamlined “premarket notification” process,
which simply requires that the new device be “substantially
equivalent” to another device already on the market. See 21
U.S.C. §§ 360(k), 360c(i).
Here, FDA concluded that the Celution and the
StemSource were not substantially equivalent to any device
already on the market. The FDA reasoned, in essence, that
using fat rather than blood as a source of cells made those new
devices different from existing devices. Therefore, FDA
ruled that Cytori must go through the extensive premarket
approval process.
3
Cytori argues that FDA’s decision was arbitrary and
capricious under the Administrative Procedure Act. Cytori
contends in particular that FDA acted unreasonably in
rejecting Cytori’s substantial equivalence application and that,
in any event, FDA did not reasonably explain its decision. In
response, FDA first raises a jurisdictional argument: that
Cytori must file its petition in the district court rather than in
this Court. On the merits, FDA argues that it reasonably
determined and explained that the Celution and the
StemSource were not substantially equivalent to any device
already on the market, meaning that Cytori must go through
the more extensive premarket approval process.
On the threshold jurisdictional issue, we conclude that
this Court is the proper forum for direct review of FDA’s
substantial equivalence determination. On the merits, we
conclude that FDA’s determination was reasonable and
reasonably explained for purposes of the Administrative
Procedure Act. We therefore deny the petitions for review.
I
Cytori Therapeutics manufactures medical devices,
including devices for use in cell therapy and other forms of
regenerative medicine. In a typical cell therapy treatment,
doctors introduce stem cells 1 or other regenerative cells into
the patient’s body to treat a disease. The cells may come
from the patient or from a donor. For example, there are
several devices on the market that draw and concentrate blood
or bone marrow in order to treat leukemia and blood-borne
diseases, among other things. The healthy donor cells are
1
Stem cells are a cellular “blank slate” that can change into a
variety of other kinds of cells and generate additional stem cells.
Those cells can be used to regenerate and repair damaged tissue.
4
used to replace the diseased or damaged cells and regenerate
new tissue.
Cytori is anticipating a major breakthrough in
regenerative medicine: the expanded use of adipose tissue –
that is, fat – as a source of stem cells for therapy and other
medical uses. Cytori recently developed technology to
harvest and concentrate stem and regenerative cells from fat
via its Celution system. The Celution and the StemSource
are two versions of this broader Celution system.
But before a new medical device such as the Celution or
StemSource may be marketed in the United States, the
manufacturer must obtain approval from FDA. In many
cases, premarket clearance is obtained by submitting a
“premarket notification.” The premarket notification process
requires that FDA find the new device “substantially
equivalent” to a device that is currently on the market. See
21 U.S.C. § 360c(i). Once FDA makes that finding, the
device may be marketed.
Some devices – in recent years, a low percentage of all
devices marketed in the United States – are not “substantially
equivalent” to existing devices and must go through FDA’s
more extensive “premarket approval” process. Premarket
approval entails scientific review of a device and often
requires clinical studies.
Cytori recently submitted premarket notifications for two
of its cell-harvesting devices. Although the notifications
both referred to virtually identical physical devices, each
notification corresponded to a different marketing version of
the device that would be sold for different medical purposes.
One version of the device, labeled as the Celution 700, is
intended to harvest and prepare stem cells for clinical
5
laboratory analysis. The other, labeled as the StemSource
900, is also intended to harvest stem cells but for storage, so
they can be used or analyzed in the future, potentially for
therapeutic purposes. Neither version of the device is
expressly intended for a specific cell therapy treatment, at
least not yet.
In general, federal law requires a new device to meet two
criteria to be considered “substantially equivalent” to a
previously marketed device. First, the new device must have
the same intended use as the predicate device. Second, the
new device must use the same basic technology as the
predicate device – or, if not, the materials submitted must
establish that the devices are equally safe and effective, and
the technological differences must not raise any different
questions of safety and effectiveness.
In its premarket notifications, Cytori claimed that the
Celution and StemSource were substantially equivalent to
currently marketed devices, including devices that harvest
cells from blood and bone marrow. FDA’s basic response
was simple: Fat is not blood. And a device meant to derive
cells from fat does not have the same intended use as a device
meant to derive cells from blood. FDA also determined (as
an alternative basis for the “not substantially equivalent”
finding) that the devices had different technological
characteristics or posed different safety concerns. In
particular, FDA highlighted risks posed by an enzyme that the
Celution and the StemSource use to separate the useful cells
from other tissue. FDA stated that this enzyme posed new
safety questions based on its effect on the harvested cells. 2
2
Cytori correctly notes that its devices are not yet labeled for
use in cell therapy. However, because the StemSource 900 is
designed for cell banking and cryopreservation – that is, for storing
6
Moreover, FDA said that the testing data for the Celution
were based on a study of only 12 donors and thus not
sufficient to demonstrate substantial equivalence. Therefore,
FDA concluded that Cytori’s devices would need to complete
the more extensive premarket approval process.
Cytori contests FDA’s “not substantially equivalent”
determination. Cytori claims that the Celution and the
StemSource share an intended use with other predicate
devices already on the market: They all process tissue
samples and isolate cells. In addition, according to Cytori,
the devices share basic technological characteristics. Cytori
therefore claims that FDA acted unreasonably in rejecting its
premarket notification.
FDA contends that this Court does not have jurisdiction
to hear Cytori’s petitions and, alternatively, defends its
determination on the merits.
II
As a preliminary matter, FDA asserts that this Court lacks
jurisdiction to hear Cytori’s petitions. In particular, FDA
argues that the relevant statutes establish the district court as
the proper forum for initial review of Cytori’s petitions.
In general, initial review “occurs at the appellate level
only when a direct-review statute specifically gives the court
of appeals subject-matter jurisdiction to directly review
agency action.” Watts v. SEC, 482 F.3d 501, 505 (D.C. Cir.
2007). The medical device section of the Food, Drug, and
Cosmetic Act contains such a direct-review provision. See
cells – those cells could foreseeably be used for treatment at a later
date.
7
21 U.S.C. § 360g. That provision allows “any person
adversely affected by” a specified regulation or order to file a
petition in the U.S. Court of Appeals for the D.C. Circuit. 21
U.S.C. § 360g(a). The specified orders in Section 360g
include “an order pursuant to section 360c(i) of this title” –
that is, an order determining whether a new device is
substantially equivalent to an existing device. 21 U.S.C.
§ 360g(a)(8).
This Court may therefore review an “order” that is made
“pursuant to section 360c(i)” of Title 21. We thus must
decide (i) whether a “not substantially equivalent”
determination is an “order” for purposes of the Act’s
direct-review provision; and, if so, (ii) whether such an order
is “pursuant to” Section 360c(i).
First, a “not substantially equivalent” determination is
plainly an “order” for purposes of the direct-review provision.
Because the Food, Drug, and Cosmetic Act does not define an
“order,” we look to the Administrative Procedure Act’s
definition. See Watts, 482 F.3d at 505. The APA provides
that an ‘‘order’’ is “the whole or a part of a final disposition,
whether affirmative, negative, injunctive, or declaratory in
form.” 5 U.S.C. § 551(6) (emphasis added).
In this case, FDA’s decision was the “final disposition”
of the issue. FDA’s letter to Cytori stated that, after
consideration, FDA had determined that the devices were not
substantially equivalent to any device currently on the market
and would need to go through the premarket approval process.
The letter did not state that FDA was still considering the
applications or that the decision was preliminary. FDA said
that the devices were not substantially equivalent. End of
story. And as the APA’s definition makes clear, a final
8
disposition may be either affirmative or negative. The
disposition here was negative.
Second, a “not substantially equivalent” order is made
pursuant to Section 360c(i). Section 360c(i) sets forth the
criteria for determining whether a new medical device is
substantially equivalent to a predicate device already on the
market.
FDA contends that findings of equivalence – but not
findings of non-equivalence – are made “pursuant to” Section
360c(i). According to FDA, Section 360c(i) explicitly refers
to only one kind of order, in which “the Secretary by order has
found that the device” is substantially equivalent. 21 U.S.C.
§ 360c(i). Based on that passing reference to an affirmative
order, FDA concludes that negative, non-equivalence orders
are not made pursuant to Section 360c(i) because negative,
non-equivalence orders are not specifically mentioned.
We do not read the statute that way. Section 360c(i) sets
criteria for determining whether a device is substantially
equivalent to another device already on the market. As a
natural consequence of Section 360c(i)’s criteria, some orders
will confirm substantial equivalence, and some will not. As
the APA puts it, some orders may be “affirmative,” and others
may be “negative.” Either way, the order is “pursuant to”
Section 360c(i) because the criteria of that section – and that
section alone – guide the determination.
A non-equivalence determination, then, is an “order”
made “pursuant to” Section 360c(i). Under the text of the
Food, Drug, and Cosmetic Act, this Court therefore has
jurisdiction to directly review Cytori’s petitions.
9
III
Applying the arbitrary and capricious standard of the
Administrative Procedure Act, we next must determine
whether FDA’s non-equivalence decision was reasonable and
reasonably explained. See Motor Vehicle Manufacturers
Assn. v. State Farm Mutual Auto. Insurance Co., 463 U.S. 29
(1983).
Under Section 360c(i), a device must meet two core
criteria to be substantially equivalent to a currently marketed
device. First, the device must have “the same intended use as
the predicate device.” 21 U.S.C. § 360c(i)(1)(A). Second,
the new device must also have “the same technological
characteristics as the predicate device” or, if not, the
submitted data must establish that the new device is both
equally “safe and effective as a legally marketed device” and
“does not raise different questions of safety and effectiveness
than the predicate device.” Id.
Here, FDA reasonably determined – and reasonably
explained its determination – that the Celution and the
StemSource met neither the “intended use” criterion nor the
“technological characteristics” criterion.
First, FDA concluded that the intended uses of the
Celution and the StemSource are not the same as the intended
uses of the most similar predicate devices. The Celution and
the StemSource are designed to derive stem cells from fat
tissue. But the most similar devices currently on the market
are designed to derive cells from blood and bone marrow.
Extracting cells from fat, FDA reasoned, is different from
extracting cells from blood. Cytori, however, argues that
deriving cells and preparing cell concentrate – whether from
fat or blood – is the same intended use.
10
One of the factors FDA routinely considers regarding
intended use concerns the “types of tissue involved.” FDA,
Guidance on the Center for Devices and Radiological
Health’s Premarket Notification Review Program (1986). To
illustrate how different kinds of tissue can lead to different
intended uses, FDA’s guidance document offered the example
of (i) a device meant to process fat and (ii) a device meant to
process blood and other tissue. Id. A device designed
specifically to process fat, FDA explained, is not intended for
the same use as a device designed to process some other form
of tissue.
Here, using that same logic, FDA concluded and
explained that fat is not blood and that the difference matters.
A court is ill-equipped to second-guess that kind of agency
scientific judgment under the guise of the APA’s arbitrary and
capricious standard. After careful review, we find FDA’s
assessment both reasonable and reasonably explained.
Second, FDA concluded that, in any event, the Celution
and the StemSource did not meet the substantial equivalence
criteria for another, independent reason: They did not pass
the separate “technological characteristics” test for a
substantial equivalence determination.
To pass this prong of the substantial equivalence test, a
device ordinarily must have the same technological
characteristics as a predicate device. But as FDA explained,
the Celution and the StemSource use different technology
than blood processing devices use. The Celution and the
StemSource required new technology both to break down the
fat tissue and to harvest the useful cells. For example, the
Celution and the StemSource take advantage of the particular
buoyancy of fat cells to separate heavier stem cells from fat
tissue. In this way, the technology of the Celution and
11
StemSource differs from the technology of blood processing
devices.
Alternatively, even if it does not have the same
technological characteristics, a device may still satisfy the
technological characteristics component of the substantial
equivalence test if it is equally “safe and effective as a legally
marketed device” and does not raise different “questions of
safety.” However, FDA concluded that the Celution and
StemSource did not meet this prong of the test. As to the
StemSource, the FDA focused on one component, the enzyme
used to aid the separation of stem cells from fat tissue. The
enzyme, which is called Celase, was previously approved by
FDA for one particular use: After liposuction, it is used to
liquefy fat waste to simplify disposal. Because the enzyme
has been approved only for that use, the scientists at FDA
identified “different questions of safety” – and reasonably
raised concerns about the impact the Celase enzyme might
have on cells that may be reintroduced into the human body.
Regarding the Celution, FDA also reasonably determined
that the only safety study Cytori submitted – which had
merely a dozen participants – was insufficient to show that the
device was equally “safe and effective” as a “legally marketed
device.”
In short, FDA reasonably concluded and reasonably
explained that the Celution and StemSource did not meet
either the “intended use” requirement or the “technological
characteristics” requirement for a substantial equivalence
determination.
12
***
We have considered all of Cytori’s arguments. We deny
the petitions for review.
So ordered.