Draft Guidance for Industry and Food and Drug Administration Staff

Comments and suggestions regarding this draft document should be
submitted within 90 days of publication in the Federal
Register of the notice announcing the availability of the
draft guidance. Submit written comments to the Division of Dockets
Management (HFA-305), Food and Drug Administration, 5630 Fishers
Lane, rm. 1061, Rockville, MD 20852. Alternatively, electronic
comments may be submitted to http://www.regulations.gov.
All comments should be identified with the docket number listed in
the notice of availability that publishes in the Federal
Register.

For questions regarding this document, contact Charles Zimliki,
Ph.D., 301-796-6297,
Charles.Zimliki@fda.hhs.gov.

U.S. Department of Health and Human
Services

Food and Drug Administration

Center for Devices and Radiological Health

Preface

Additional copies are available from the Internet. You may also
send an e-mail request to
dsmica@fda.hhs.gov to receive an electronic copy of the
guidance or send a fax request to 301-827-8149 to receive a hard
copy. Please use the document number (1786) to identify
the guidance you are requesting.

Introduction
Background
1. Overview
2. Currently Marketed Devices to Treat DM
3. Basic Design of an APDS
4. Different APDS Types
Scope
Device Description
1. APDS System Level Description
2. Glucose Monitoring Functional Components
  1. CGM Component
  2. BGD Component (if applicable)
3. Control Algorithm & Signal Processing Functional
  Component
4. Infusion Pump Functional Component
5. Communication Pathway Functional Component
Indications for Use
APDS Performance
1. Software
2. Report of Prior Investigations (If Applicable)
3. Biocompatibility
4. Sterility
5. Shelf Life
6. Electrical Safety
7. Magnetic Resonance (MR) Imaging Safety
8. Quality of Results from Diagnostic Devices Used During the
  Clinical Study
9. Human Factors
10. Glucose Monitoring Functional Component
  1. CGM Component
  2. BGD Component
11. Control Algorithm/Signal Processing Functional Component
12. Infusion Pump Functional Component
Clinical Study Progression
1. Early Feasibility Study
2. Transitional Clinical Study
3. Pivotal Clinical Study
  1. Patient Population
  2. Study Endpoints
  3. Statistical Analysis
Labeling
Post-Approval Study

Appendix A: IDE Content for APDS Studies

Background
Device Description
Letters of Authorization
Indication for Use
Nonclinical studies/Prior Investigations
1. Algorithm
  1. Definition of Algorithm(s)
  2. Definition of Algorithm Symbols/Parameters
  3. Identification of Fixed Parameters
  4. Identification of Adjustable Parameters that May be Modified
    During the Study
  5. Parameter Sensitivity Analysis
  6. Summary of the Verification Activities for the Control
    Algorithm
2. Software Documentation
3. Summary of System Communication
4. Safety Measures for Dosing
5. Biocompatibility Testing
6. Electrical Safety
7. Animal/In-Silico Testing
8. Human Studies
9. Human Factors/Usability Testing
10. Catheter Occlusion Bench Testing
11. Dose Accuracy Bench Testing
12. Diagnostic Devices To Be Used During the Clinical Study
13. Drugs Used During the Study
Bibliography
Clinical Study
1. Purpose/Objective(s)
2. Study Design
3. Sample Size and Investigational Sites
4. Study Duration
5. Inclusion Criteria
6. Exclusion Criteria
7. Study Timeline
8. Safety Monitoring/Risk Analysis
9. Stopping Rules
10. Endpoints
11. Success Criteria/Goal
12. Statistical Analysis Plan
Informed Consent
Patient Case Report Form(s)
Investigator Agreement Forms
Monitoring Information
Institutional Review Board (IRB) Information
Labeling
Anticipated Changes
Manufacturing

Appendix B: Glossary

Draft Guidance for Industry and Food and Drug
Administration Staff

The Content of Investigational Device Exemption
(IDE) and Premarket Applications for Artificial Pancreas Device
Systems

This draft guidance, when finalized, will represent
the Food and Drug Administration’s (FDA’s) current thinking on this
topic. It does not create or confer any rights for or on any person
and does not operate to bind FDA or the public. You can use an
alternative approach if the approach satisfies the requirements of
the applicable statutes and regulations. If you want to discuss an
alternative approach, contact the FDA staff responsible for
implementing this guidance. If you cannot identify the appropriate
FDA staff, call the appropriate number listed on the title page of
this guidance.

This draft guidance is intended to provide recommendations to
Sponsors or Applicants
¹ planning to develop and submit an Investigational
Device Exemption (IDE) or premarket approval (PMA) application for
an Artificial Pancreas Device System (APDS) for single patient use
in the home environment. FDA recognizes the need for guidance on
the least burdensome means of development for these innovative
device systems. Due to the evolving nature of these device systems,
it is expected that they will develop incrementally. The
recommendations contained in this guidance are intended to provide
adequate guidance and instruction to facilitate the development and
marketing of the APDS while, at the same time, adopting a flexible
approach.

This guidance discusses the development and evaluation of APDS.
We describe both nonclinical and clinical approaches to
establishing the safety and effectiveness of an APDS, and suggest
areas where there is flexibility in the pathway to market for these
devices. Specifically,
Section VII and
Appendix A of this guidance provide detailed information to
assist Sponsors in assembling information to support an IDE
submission, while Sections IV – VI set out the criteria FDA will
use in evaluating an APDS for premarket approval (PMA).

This document does not provide guidance on the evaluation of low
glucose suspend (LGS) systems. We issued
draft guidance in June 2011 that provided recommendations for
Sponsors or Applicants planning to develop and submit an IDE or PMA
for an LGS system for use in the home environment .

FDA’s guidance documents, including this guidance, do not
establish legally enforceable responsibilities. Instead, guidances
describe the Agency’s current thinking on a topic and should be
viewed only as recommendations, unless specific regulatory or
statutory requirements are cited. The use of the word
should in Agency guidances means that something is
suggested or recommended, but not required.

A. Overview

This guidance lays out many of the possible options for the
design, testing and marketing of an APDS, but it is by no means
all-inclusive. Our intent is to provide the most flexible
recommendations to guide Sponsors in designing and testing these
devices that are consistent with the least burdensome principle to
provide options while assuring that testing is adequate to support
marketing approval. In particular, the guidance:

Provides an approach to allow Sponsors to proceed to outpatient
studies as quickly as possible;
Provides maximum flexibility in determining appropriate size
and duration of clinical studies;
Gives Sponsors the option to prove non-inferiority to standard
therapy, but also describes study criteria to support superiority
claims if the Sponsor prefers to make a superiority claim; and
Describes approaches to leveraging existing data about the
safety and effectiveness of the devices already on the market
collected from studies done within and outside of the US, which
minimizes the need for preclinical data.

Use of Continuous Glucose Monitor Data in the Evaluation
of APDS. We have placed the primary focus for glucose
measurement on a Continuous Glucose Monitor (CGM). Because patients
currently need to periodically calibrate their CGM using a blood
glucose measurement from a blood glucose device (BGD), we have kept
the BGD as part of the APDS, but not the primary focus. Recognizing
that, over time, improved CGM performance may obviate the need for
periodic blood glucose checks with a BGD, we have built in the
flexibility to eventually allow for the approval of APDS that do
not use a BGD.

Flexibility with respect to Endpoints. We have
also introduced a number of suggested examples of primary endpoints
that can be used to measure the safety and effectiveness of an APDS
and support a successful PMA. In addition to these endpoints, we
believe that there may be alternative acceptable primary endpoints.
We encourage Sponsors to discuss their choice of primary endpoint
and study design with us.

Flexibility with respect to Indication. We give
examples of indication statements that we believe will be the most
likely indications for early APDS, and those that would be
supported by the endpoints we suggest, but do not foreclose the
possibility of other indication statements that can be
appropriately supported. We have also set out criteria that would
support superiority claims for Sponsors who wish to be able to
claim that their device improves outcomes compared to other
therapies.

Reasonable Study Progression allowing for Quick
Outpatient Use. We have set out a clinical study
progression that will move the APDS to outpatient use as quickly as
possible . Each step is designed to test specific aspects of the
APDS functionality and performance. The guidance describes how
sponsors who believe they already have sufficient, valid scientific
evidence that fulfills the purpose of a particular study phase and
that justifies moving to the next study phase may do so, using
clinical and non-clinical evidence, and evidence that was obtained
from studies performed outside the US.

Flexibility with respect to Study Size and
Duration. We have built in maximum flexibility regarding
the size and duration of each study phase, while also aiming to
take the least burdensome approach. We recognize that study size
and duration is entirely dependent on the design and features of
the APDS and its proposed indication. We recognize that each APDS
will likely have unique features that affect study design. In
addition, because some APDS may be composed of parts that have
already been approved or cleared by FDA, we encourage Sponsors to
leverage what we already know about the safety and effectiveness of
the individual components to streamline the clinical testing of
such a system.

We believe the recommendations contained in this guidance will,
when finalized afford Sponsors the flexibility they need to design
innovative, safe and effective systems to treat diabetes mellitus
(DM). If you believe an alternative, less burdensome approach to
investigating and developing premarket applications for these
devices can satisfy regulatory requirements for investigation and
approval of APDS, we encourage you to discuss that approach with
the FDA.

B. Currently Marketed Devices to Treat DM

Today, patients with DM utilize a variety of devices to monitor
and manage their blood glucose levels:

Hand-held portable BGDs which have been
cleared by FDA for home-use, allow patients to determine their
blood glucose levels using blood from a finger stick. Patients use
BGDs multiple times a day to help make decisions regarding insulin
administration and diabetes management around meals, exercise, and
other activities of daily living.
Some patients also use continuous subcutaneous insulin infusion
(CSII) via an insulin pump to manage their
disease.
Some patients use a CGM system, which uses a
sensor inserted into the subcutaneous tissue and continuously
(meaning, at a consistent interval) measures the concentration of
glucose in the interstitial fluid. While CGM devices have not yet
reached a performance level that would make them an adequate
substitute for BGDs, they do allow patients to monitor trends and
patterns of glucose levels in their bodies.

Even with the aid of these devices, maintaining blood glucose
concentrations within a suggested optimal range is a daily struggle
for people living with DM, and the risk of hypoglycemia associated
with attempts at improved glycemic control remains an ever-present
danger.

C. Basic Design of an APDS

APDS link a CGM to an insulin pump and automatically reduce or
increase insulin infusion based upon specified thresholds of
measured interstitial glucose. The APDS parts are designed to
communicate with each other to automate the process of maintaining
blood glucose concentrations at or near a specified target or range
and minimize the incidence and severity of hypoglycemic
(dangerously low blood sugar) and hyperglycemic (dangerously high
blood sugar) events.

The illustration below describes the parts of an APDS and
depicts how they work together.

* Blood Glucose Device: BGDs are currently used to calibrate the
CGM, but we anticipate that future improved CGM performance may
obviate the need for a BGD in APDS.

(1) Continuous Glucose Monitor (CGM). A CGM
provides a steady stream of information that reflects the
patient’s blood glucose levels. A sensor placed under the
patient’s skin (subcutaneously) measures the glucose in the fluid
around the cells (interstitial fluid) which has been found to
correlate with blood glucose levels. A small transmitter sends
information to a receiver. A CGM continuously displays both an
estimate of blood glucose levels and their direction and rate of
change of these estimates.

Blood Glucose Device (BGD). Currently, to get
the most accurate estimates of blood glucose possible from a CGM,
the patient needs to periodically calibrate the CGM using a blood
glucose measurement from a BGD; therefore, the BGD still plays a
critical role in the proper management of patients with an APDS.
However, over time, we anticipate that improved CGM performance may
obviate the need for periodic blood glucose checks with a BGD.

(2) Control algorithm . A control algorithm is
software embedded in an external processor (controller) that
receives information from the CGM and performs a series of
mathematical calculations. Based on these calculations, the
controller sends dosing instructions to the infusion pump.

(3) Infusion pump. Based on the instructions
sent by the controller, an infusion pump adjusts the insulin
delivery to the subcutaneous tissue.

(4) The Patient. The patient is an important
part of the APDS. The concentration of glucose circulating in the
patient’s blood is constantly changing. It is affected by the
patient’s diet, activity level, and how his or her body
metabolizes insulin and other substances.

D. Different APDS Types

Although the fundamental parts described above are common to all
APDS, different device designs, algorithms, and patient management
strategies create the potential for different APDS types
including:

A Control-to-Range (CTR) system that reduces
the likelihood or severity of a hypoglycemic or hyperglycemic event
by adjusting insulin dosing only if a person’s glucose levels
approaches predetermined thresholds. When a patient’s blood glucose
concentration is within the specified range, the infusion pump will
not take any action based upon CGM readings. Patients using this
system still need to monitor their blood glucose concentration, set
appropriate basal rates for their insulin pump and give pre-meal
bolus insulin to maintain control of their glucose levels.
A Control-to-Target (CTT) system that sets
target glucose levels and tries to maintain these levels at all
times. This system is fully automated and requires no interaction
from the user (except for calibration of the continuous glucose
monitoring system). There are two subtypes of CTT systems currently
being investigated (i.e., insulin-only and bi-hormonal) and a
hybrid system option (patient administration of a pre-meal or
partial pre-meal insulin bolus) that can be used in either of the
system types.
CTR and CTT System Subtypes are
dependent upon the drug or drugs being delivered and how each drug
affects blood glucose concentrations. Subtypes may include:
An insulin-only system that achieves a target
glucose level by increasing or decreasing the amount of insulin
infused.
A bi-hormonal control system that achieves a
target glucose level by using two algorithms to instruct an
infusion pump to deliver two different hormones — one hormone
(insulin) to lower glucose levels and another (glucagon) to
increase blood glucose levels. The bi-hormonal system mimics the
glucose-regulating function of a healthy pancreas more closely than
an insulin-only system.

This guidance is focused on the development, investigation and
approval requirements for APDS – autonomous systems that administer
insulin to maintain blood glucose concentrations within a
prespecified range to maintain glycemic control and minimize the
risk of hypoglycemia or hyperglycemia – for use outside a health
care facility. The guidance was written to both accommodate APDS
utilizing current technologies, and permit application of the
principles described in this guidance to newer technologies.

As described in Section II above, the wide variety of CGMs,
BGDs, insulin pumps, and control algorithms available allows for a
number of different types and designs of APDS. We anticipate that
some APDS will utilize already cleared or approved components.
Others may utilize components that have been modified in some way.
The information needed and the studies required when changes are
made to the different components, or when components are
substituted into an already approved APDS, will depend on the
effect the change is anticipated to have on system performance. For
example, additional clinical studies may not be needed if a Sponsor
is able to demonstrate that a newly introduced component is similar
to the previously approved version, e.g., its accuracy,
susceptibility to interferences, human factors, etc.

For purposes of this document, FDA defines an APDS as including
the following components:

Glucose Monitoring Devices – a CGM and
BGD used for calibrating the CGM (where
applicable) plus associated reagents/test strips;
Control algorithm ;
Infusion pump – Fluid infusion set for the
complete fluid pathway from, and including, the drug reservoir or
fluid source container (e.g., bag, cassette, vial, syringe),
infusion set, extension sets, filters and valves, clamps, up to and
including the patient connection;
Components and accessories (e.g., power cord,
wireless controller); and
Network (i.e., any device or system physically
or wirelessly connected to the APDS)

The primary product code for an APDS is ‘LHE’
(controller closed-loop blood glucose), which is regulated as a
class III device system.

This guidance applies only to APDS that use insulin products
that have been approved by the FDA for delivery via an infusion
pump, and that are used in accordance with their FDA-approved
labeling. This guidance does not address data requirements for a
drug labeling modification such as approval of a new drug
formulation or drug delivery method. Also, this guidance may not
apply to APDS that utilize synthetic or artificial cells, tissues
or organs nor does it address issues that are unique to combination
products.² Although elements of this guidance may be applicable to
these circumstances, additional considerations outside the scope of
this guidance may also need to be addressed.

IV. Device
Description

APDS currently consist of a number of device components that
communicate to form a complete system. The unique qualities of
these systems stem from the interaction of the various device
components to achieve the system’s intended use. To unify the
device description for all types of APDS, Applicants should
describe: (i) the device system as a whole, and (ii) each of the
functional components within the device system. FDA recommends
Applicants provide the following information as part of the APDS
device description:

A. APDS System Level Description

The Applicant should provide the following descriptive
information regarding the device system.

A clear statement of the intended use and indications for use
(see
Section V).
A picture or schematic of the entire system and how the
components interface. A listing of all the device functional
components and accessories that are part of the system (including
model numbers).
Because the system is intended for ambulatory use, a
description should be provided on features of the system designed
to address issues such as mobility, various environmental
conditions (e.g., water exposure, altitude, electromagnetic
interference), and ruggedness.
Because the system is intended for lay use, a description
should be provided of features of the system designed to address
how the device has been designed to be safely and effectively used
by the lay population, which often have limited or no clinical
background and may have functional limitations.
Detailed description of the technological features of the
system (e.g., alarms, etc.).
Detailed description of the training for all persons involved
with an APDS. See Part 8 of Section VI below for a full discussion
of Human Factors and other training considerations.

For each of the device functional components, the descriptive
information identified in the following sections should be
provided.

B. Glucose Monitoring Functional Components

Applicants should provide the following information for the
functional components of the APDS that serve to monitor glucose
levels in the patient.

1. CGM Component

Applicants should indicate the regulatory status of the CGM
component.
- If a modified version of an approved CGM is used, the A
  pplicant should provide a comprehensive list and description of the
  modifications to the CGM and provide the rationale for the
  change(s). This might include instructions for use, such as the
  required run in period or calibration, or changes to the algorithm
  or physical structure.
- If it is an already approved CGM, applicants should provide:
  - The name of the CGM and the FDA document number under which it
    was approved (noting the appropriate Supplement number and the date
    of the Supplement).
  - If the Applicant wishes to rely on information previously
    submitted by a different Applicant, a letter of authorization
    granting access to the information .
Description of the function(s) the CGM performs in the
APDS.
Description of the methodology employed for glucose measurement
(e.g., electrochemical measurement).
Description of the sample matrix analyzed (e.g., interstitial
fluid)
Description of the anatomical site(s) into which the sensor is
inserted.
Description of the information provided by the CGM, such as the
frequency of reported glucose values, trending information and
alarms.
Device description, including a list of all device components
and accessories. As appropriate, this would include sensors,
display monitors, devices to aid in the insertion of the sensor,
quality control materials, standards (calibrators), and
software.

2. BGD Component (if applicable³)

The regulatory status of the BGD.
If modified, a comprehensive list and description of all
modifications and a rationale for the change(s).
Description of the function(s) performed by the BGD.
A list of all BGD functional components and accessories, as
appropriate. In addition to the instrument, reagents and quality
control materials, accessories might include standards
(calibrators), data transmitting equipment or software that
processes or stores data or quality control results.
A summary description of the measurement method utilized by the
BGD (e.g., electrochemical, spectrophotometric measurement).
A description of the test principle, i.e., all chemical
reactions and concentration of all reagent components.
Matrix of blood sample to be analyzed (e.g., fingerstick
capillary blood).

C. Control Algorithm & Signal Processing Functional
Component

The description of the control algorithm
functional component should include all computational steps,
including CGM signal and changes in the command for insulin
delivery. The control algorithm should include the following
information.

Description of how the algorithm addresses signal dropout and,
if applicable, a description of any analyses that occur to
determine if the CGM value is artifact or real (in addition to
processing performed as part of the CGM algorithm).
Description of the control algorithm that adjusts insulin
dosing. This description should be detailed sufficiently to allow
the recreation of the control algorithm. This should include:
- Defining the control algorithm equation(s) in symbolic
  form.
- Defining each symbol with a parameter name for the control
  algorithm equation(s).
- Defining all parameters that cannot be modified (fixed
  parameters) by the user and/or healthcare provider and their
  corresponding parameter value.
- Defining all parameters that are adjustable by the user and/or
  healthcare provider.
  - Define the parameter value range and the smallest increment
    that the parameter can be adjusted.
  - Identify the user that can adjust the parameter (i.e., patient
    or healthcare provider) and describe how the device secures these
    parameter values that are adjustable only by the specified
    user.
Summary of the minimum and maximum insulin delivery dose
recommendation of the control algorithm.
If applicable, a description of any safety check that the
system performs to ensure insulin infusion has delivered the
appropriate dose.
Description of the signal processing from the dosing
recommendation to the raw signal pump current command.
Summary of the verification activities to show the control
algorithm has been properly coded into the software.

If the CGM is not already approved or if the algorithm has been
modified and not previously reviewed by FDA, the Applicant should
also provide a description of the signal processing starting from
the raw CGM signal to the reported CGM value. This should describe
the method (e.g., signal averaging) of calculating the reportable
CGM value, the frequency of reporting the CGM value, and the signal
processing that is performed for the calibration method.

D. Infusion Pump Functional Component

The description of the infusion pump functional component should
include the following information: If the infusion pump is
labeled for use with a specific drug, the labeling should be
consistent with the approved indications and route of
administration. To facilitate FDA’s review, the FDA approved
labeling for that device or drug should be provided.

A detailed description should be provided (including, where
appropriate, assembly drawings, schematics, and/or specification
control documents) of the pump and its functional components, and
accessories including:
- The infusion delivery mechanism
- The bolus mechanism
- The drug reservoir
- Pump tubing and connectors (built-in or external to the
  pump)
- A user-interface, consisting of the programming unit, display
  unit, audio and tactile notification units
- Power supply or pump battery and circuitry to charge the
  battery
- A communication interface, including network components and
  interfaces to other devices and systems
- Refill frequency
The principle of operation of the infusion pump (i.e., the
scientific principles behind how the device achieves its intended
use).
Identification and description of particular infusion sets or
cassettes that will be provided or recommended for use with the
APDS, if any.
The user interface components of the pump, including keypads,
control menus, data entry screens, displays, indicator lights,
alarms, auditory and tactile feedback, infusion sets, cassettes,
free-flow prevention mechanisms, tubing, latches, doors or other
components of the physical pump that may be manipulated.
A detailed design description of the software utilized by the
device, if any, including all key elements.⁴
The specifications for the infusion device (e.g. flow rate
accuracy specifications for bolus and basal deliveries, time to
deliver bolus, etc.).

E. Communication Pathway Functional Component

The description of the Communication Pathway functional
component should describe the passage of information between the
functional components, including a description of the hardware and
software that allows the passage of information. The description
should include:

Communication pathway. Applicants should describe all of the
ways each functional component communicates to other functional
components within the system. The Applicant should identify the
flow of communication (e.g., unidirectional or bidirectional)
between the functional components and identify the information that
is passed.
Communication hardware. Applicants should describe how the
information is passed between each functional component and
describe the hardware necessary to communicate this
information.
If the system incorporates or is intended to incorporate
radio-frequency (RF) wireless technology (e.g., IEEE 802.11,
Bluetooth, Zigbee), the description should include information
about the specific RF wireless technology and characteristics, its
use and functions (e.g., remote monitoring or control, software
updates), the data to be transmitted including any alarms by
wireless transmission, quality of service (QoS) needed, wireless
security protocols, and any limitations or restrictions relating to
coexistence with other RF wireless technology or electromagnetic
interference (EMI).
If the device is capable of being remotely controlled or
monitored from a distance, this capability should be identified
with a description of the measures incorporated to assure
safety.
Reliable communication between the various device components of
the APDS is essential to ensure the correct information is passed
to each device component. The Applicant should describe how the
system ensures communication with only the devices approved with
the system.

V. Indications for
Use

The indications for use statement is “ a general
description of the disease or condition the device will diagnose,
treat, prevent, cure, or mitigate, including a description of the
patient population for which the device is
intended.”⁵ The indications for use statement of different APDS may
differ depending on the device design and patient population.
Indications should be proposed based upon the design of the control
algorithm, clinical study design, and the intended patient
population, although we recognize that APDS currently in the
research and development phase may not have a final indication
until they have been studied in depth.

The following statements are provided as examples of appropriate
wording for likely indications for use :

The APDS device system is intended for patients with type 1
diabetes for the subcutaneous infusion of insulin and the
continuous measurement of interstitial glucose to aid in the
management of their disease. The APDS automatically adjusts insulin
delivery in response to CGM values that have exceeded or are
predicted to exceed the bounds of a prespecified blood glucose
range. The APDS is intended to assist the patient in achieving
metabolic goals.
The APDS device system is intended for patients with type 1
diabetes mellitus for the subcutaneous infusion of insulin and the
continuous measurement of interstitial glucose to aid in the
management of their disease. The APDS automatically adjusts insulin
delivery in response to CGM values to maintain a prespecified
target glucose. The APDS is intended assist the patient in
achieving metabolic goals.

FDA recognizes, however, that alternative indication statements
may be appropriate depending on the populations and endpoints
studied. Sponsors who plan to submit a PMA for an APDS that is
specifically intended to improve glycemic control or reduce
hypoglycemia will need to provide valid scientific evidence to
support those claims as required by 21 CFR 814.20(b)(3)(vi), and
should be sure to measure those features as endpoints in the
pivotal clinical study designed to support the PMA submission.
Sponsors seeking to support different indications for use should
discuss appropriate study design and labeling statements with
FDA.

The Agency recommends the following information and performance
characteristics be provided in the PMA.

A. Software

Software documentation is an important aspect of device
validation. The Applicant should provide complete software
documentation in the PMA. Some useful guidance documents for
software considerations are provided below.

The Agency considers the APDS and all of the components of the
system to be a “Major” level of concern for the
purposes of software review in the PMA. The information to provide
in a submission related to software has been delineated in the

Guidance for the Content of Premarket Submissions for Software
Contained in Medical Devices.
If the device includes off-the-shelf software, additional
information should be provided as recommended in the
Guidance for Industry, FDA Reviewers and Compliance on
Off-the-Shelf Software Use in Medical Devices.
Cyber Security – FDA recommends that the
concept of information security be addressed when medical devices
can store, access, and/or transfer information externally.
Information security is the process of preventing the modification,
misuse, and denial of use or the unauthorized use of that
information. Specific concepts are confidentiality, integrity,
availability and accountability (CIAA) :
- Confidentiality assures that no unauthorized
  users have access to the information.
- Integrity is the assurance that the
  information is correct – that is, it has not been improperly
  modified.
- Availability suggests that the information
  will be available when needed.
- Accountability is the application of
  identification and authentication to assure that the prescribed
  access process is being done by an authorized user.
Communication between device components should be secure and
prevent communication from other devices that are not part of the
system. For additional guidance on cybersecurity, please refer to

Guidance for Industry, Cybersecurity for Networked Medical Devices
Containing Off-the-Shelf (OTS) Software.

B. Report of Prior Investigations (If Applicable)

Data from prior investigations – both clinical and nonclinical
— should be included as part of the PMA to the extent applicable.
The Report of Prior Investigations should include a summary of
non-clinical information relied upon to address basic device
safety, characterize catastrophic failure modes and risk mitigation
approaches, and support an expectation that the device will
function as intended. The PMA must contain, to the extent
applicable, a bibliography of published reports and an
identification and discussion of other data and information
relevant to the safety and effectiveness of the APDS. 21 CFR
814.20(b)(6) and (8).

C. Biocompatibility

Biocompatibility testing of the APDS should be performed on the
final, finished, sterilized device for all device components and
accessories. The PMA should include a complete test report of each
biocompatibility test performed. Alternatively, if biocompatibility
information for a test component has been previously evaluated and
found acceptable by FDA (such as during separate premarket review
of the component), the Applicant may provide a summary of the
testing procedures and study. In this case, the Applicant should
reference the FDA document number.

Generally, biocompatibility tests for a PMA device should be
performed keeping in mind the duration and level of contact the
patient is likely to have with the device. We recommend that
Sponsors consider APDS to have prolonged duration of contact with
the patient because of the way the device and its accessories will
be used.

For additional information and detailed instructions on
biocompatibility testing, we recommend following the FDA blue book
memo entitled,
Use of International Standard ISO 10993, ‘Biological
Evaluation of Medical Devices Part 1: Evaluation and Testing’
(Replaces #G87-1 #8294).

D. Sterility

Each of the device components used in the APDS will require
different types of processing or reprocessing based on their
intended use. The intended use will determine whether a device must
be sterile, such as an implant that will be contacting normally
sterile locations within the body, or whether it will require a
lesser degree of microbicidal processing, such as a reusable
component that is intended to contact only intact skin.

For sterile device components, use of FDA recognized consensus
standards for conducting process development and validation testing
is recommended. A searchable list of these standards is available
at
http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfStandards/search.cfm.

Products labeled “sterile” should be processed using
methods that have been comprehensively validated by the Applicant
to provide a sterility assurance level (SAL) of 10 -6.

The product labeling should prominently indicate whether each
component is supplied in a sterile or non-sterile state. For device
components that are not sterile, we recommend that the Applicant
provide a scientifically valid rationale for why sterilization is
unnecessary.

APDS and accessories intended for prolonged use should include
instructions in the labeling for proper cleaning and disinfecting,
as appropriate, between uses. Also, where appropriate, “use
life” information should be provided in the labeling, with
supporting information (see Subsection VI-E, below). This may
include information such as the number of times the device can be
reused, or guidance as to how users can make that determination
(e.g., inspecting for wear and tear). APDS are intended to be used
in the home environment. The Applicant should indicate cleaning
agents/products in the labeling that are readily available to the
average home user along with validated instructions for cleaning
the device in a manner that is consistent with the FDA guidance for
labeling reusable devices.⁶ In addition, reference to relevant Technical
Information Reports (TIR) developed by the Association for the
Advancement of Medical Instrumentation (AAMI) when developing
labeling instructions for reusable medical devices is
recommended.⁷,⁸

If sterility or the cleaning and disinfection of specific system
components has been previously evaluated and found acceptable by
FDA (such as during the premarket review of the component), the
Applicant should provide a summary of testing procedures, the study
results, and the FDA document number where additional information
can be found. This approach may be acceptable if the way in which
the component is used and the way it is packaged has not changed
from the time of the original clearance/approval.

E. Shelf Life

The shelf life of the APDS, including accessories, should be
supported with appropriate data, including both performance-based
testing and package integrity, when applicable.

Performance

If the particular system contains sterile components, materials
or reagents that could degrade over time, a shelf life should be
included on the packaging. Additionally, performance data should be
generated after an appropriate number of complete “use
cycles” which should include cleaning or disinfection per the
labeling.

The Applicant should also provide data demonstrating that the
APDS can maintain the performance specifications throughout the
system’s shelf life. If accelerated test methods are
utilized, the Applicant should demonstrate that the test methods
accurately simulate real-time conditions for the device should be
provided.

Package Integrity

The Applicant should ensure that device package design and
construction are validated to protect the device components from
alteration or damage during shipping and transportation. The
packaging should also be validated to support the labeled shelf
life (e.g., 1 year, 3 years). The validation process should be
designed to assure that packaging will maintain its integrity (no
breaches of the sterile barrier system) after being subjected to
the rigors of the real world (i.e. less-than-ideal shipping and
handling conditions), as well as stability testing (i.e., aging).
This typically requires two validation test pathways: simulated
shipping of packaged product (or accurate surrogate of product)
followed by package integrity testing, and simulated (accelerated)
aging followed by seal strength testing. We recommend that
confirmatory, real-time package shelf life testing be submitted as
part of the PMA. We also recommend that Sponsors use recognized
consensus standards for conducting these various simulations and
validation tests.⁹,¹⁰,¹¹

If the shelf life (or expiration dating) of a system component
has been previously evaluated and found acceptable by FDA (such as
during the premarket review of the component), the Applicant may
submit a summary of testing procedures and study results along with
the FDA document number where additional information can be found.
This approach is only acceptable if the way in which the component
is used and the way it is packaged has not changed from the time of
the original approval.

F. Electrical Safety

A complete test report should be provided in the PMA submission
describing the electrical safety testing used to support approval
of the APDS. Details of the electrical safety can be found in
Appendix A-V-G.

G. Magnetic Resonance (MR) Imaging Safety

Sponsors should clearly identify on the APDS and its label
whether it is MR Safe, MR conditional, or MR unsafe. For
information regarding Magnetic Resonance (MR) Imaging safety
testing and labeling, please see FDA’s guidance document,

Establishing Safety and Compatibility of Passive Implants in the
Magnetic Resonance (MR) Environment. While the subject of
the referenced guidance is passive implants, the information
contained in it is also relevant for active devices like APDS.

H. Quality of Results from Diagnostic Devices Used During the
Clinical Study

Clinical studies often include in vitro diagnostic devices
(IVDs) that provide information used as an endpoint of the study,
e.g., Hemoglobin A1c (HbA1c), urine or blood ketone results, or
blood glucose results from a device other than the BGD used to
calibrate the CGM component of the APDS. Therefore, evidence is
needed to support the quality of those results.

Applicants should provide the following information:

Name of the device, and the associated reagent(s);
An indication of the regulatory status of the device:
- If it has already been granted marketing approval or clearance,
  Applicants should provide the FDA document number where additional
  information can be obtained, if known; or
- If the device has not been granted marketing approval or
  clearance, Applicants should provide data supporting the accuracy
  and reliability of the device¹²;
Any qualifying certifications of the reagent or test system,
e.g., National Glycohemoglobin Standardization Program
certification;
Name and concentration levels of the Quality Control (QC)
materials run to confirm proper performance of the device during
the study;
The frequency or points in time when QC material(s) were
analyzed;
How it was determined that the devices were functioning
properly, e.g., any functional checks performed or criteria applied
to QC material results; and
A statement certifying that a copy of the labeling was provided
to each user who is operating a device at home, or that it was
available to staff who operated a device in an in-patient
setting.

I. Human Factors

The device user interface plays a critical role in the
performance of the APDS and should be considered integral to the
overall performance of these systems.

Reports of device-related incidents and recalls for diabetes
devices have shown that patterns of use errors resulting from
deficiencies in the design of the user interface have led to
patient harm. Human factors testing can help identify and mitigate
these deficiencies. For this reason, FDA recommends that PMAs
include comprehensive application of human factors in the design
and evaluation of the user interface components of the entire
APDS.

The term user interface denotes all components of the
device system with which the user interacts; for example:

Control mechanisms (e.g., key pads, touch screens, slide
controls)
Feedback mechanisms (e.g., auditory alarms, visual alarms,
status indicators, and other messages to users)
Graphical user interface, including representations of
responses to user actions (including visual feedback related to
changes in device operation or status)
Labeling (including directions for use, user manuals,
quick-start guides, package inserts, information on packaging,
etc.)

Hazards associated with use of functional components of the APDS
are unique in that they exist even if a device operates within its
specifications. These hazards often do not involve failures due to
faulty mechanical, electrical or software components that are
previously known or reasonably anticipated but rather, arise
specifically from interaction with a human operator.

Analyses of use-related hazards should consider the
following:

The safety of interactions between the user and all parts of
the user interface components of the system:
- Adequacy and convenience of the arrangement of user interface
  components for users’ physical interactions with the
  device
- Potential errors associated with atypical user actions or
  techniques
- Legibility of visual information, including device labels and
  displays
- Audibility and distinctiveness of auditory information,
  including different alarm tones and logic of alarm activation
Potential use errors or difficulties associated with:
- Each possible setting or input available to operators
- Input, selection or modification of critical treatment
  parameters
- Non-standard or unusual parameter settings or default
  values
- Non-standard, unfamiliar or ambiguous conventions or
  abbreviations
- Non-standard, ambiguous, or inadequate alarm condition or
  informational messages
- Improper storage conditions (e.g., test strip/reagent storage
  temperature and humidity, etc.)
- The user’s inability to understand the indications for
  use of the device and limitations of the device.
Potential errors associated with use of the CGM component,
including:
- Incorrect data entry during CGM calibration
- Improper timing of CGM calibration (e.g., when conditions are
  not optimal)
- Failure to calibrate the CGM at the recommended frequency
- Failure to discontinue CGM use at the end of the sensor wear
  period when CGM results may be compromised (e.g., when there is no
  hard stop on CGM results generation)
- Improper anatomical placement of CGM
- Use of an expired sensor
- Use of the CGM under inappropriate conditions
Potential errors associated with use of the BGD used to
calibrate the CGM component, including:
- Improper fingerstick sampling technique (e.g., “milking” the
  finger)
- Failure to take a fingerstick sample to confirm questionable
  CGM readings (e.g., that do not correspond with user’s
  clinical symptoms or user’s expectations of what glucose
  should be at that time)
- Inadequate volume of blood sample collected
- Failure to follow recommended quality control procedures
- Improperly performing quality control procedures
- Use of expired reagents or test strips
- Improper storage of BGD reagents or test strips
  - Note: This error is extremely important for Applicants to
    address. Improper test strip storage (e.g., in car glove
    compartments) is identified as the most common source of error by
    manufacturers when consumers report an improperly functioning
    BGD.
- Improper handling of reagents or test strips, such as leaving
  the reagent bottle cap off for longer than is recommended
- Inadequate cleaning or maintenance of the BGD
- Patients using a BGD to calibrate the CGM during conditions
  which are contrary to use of the BGD. For example, measurements
  from many BGDs are affected by conditions such as ketoacidosis,
  hypoglycemia or ingestion of large doses of vitamin C.

The APDS development process should include human
factors/usability testing to ensure that the device system will be
safe and effective in the hands of the intended users. This testing
should be conducted with people who are representative of the
intended users and under conditions that are comparable to actual
conditions of use. The intended users should be defined precisely
but they might include health care providers, patients, and lay
caregivers (e.g., elderly spouses or parents of children), and the
users’ ages and functional capabilities could span a wide
range. The conditions of use should include provision of labeling,
such as instructions for use, and training that is comparable to
the training that actual users will receive. The testing should
assess not only the user interface components of the devices in the
system, but also the adequacy of the labeling and training to
support users to use the system safely and effectively.¹³

J. Glucose Monitoring Functional Component

1. CGM Component

Applicants should provide appropriate information regarding the
safety and effectiveness of the CGM functional component when used
as part of an APDS. Applicants may find it helpful to review the
FDA-recognized Clinical Laboratory Standards Institute (CLSI) POCT
05-A guideline¹⁴ and applicable FDA guidance documents.

Applicants should provide protocols and test reports for the
following performance characteristics established during the
pivotal CGM trial:

Accuracy of CGM results. Applicants should characterize
accuracy by summarizing the point-to-point agreement between blood
glucose reference readings and paired CGM results. Applicants
should present the total and cumulative percentage (and numbers) of
CGM values presented as various differences from the paired blood
glucose reference result (e.g., within 10, 20, 30, 40 or >40
mg/dL). FDA recommends that blood glucose reference values be
stratified according to glucose concentration (as determined by the
blood glucose reference values) in various glucose concentration
bins (e.g., <40, 41-50, 51-60, 61-70, 71-80, 81-120, 121-180,
180-250, 250-325, and 326-400 mg/dL).
Bias across the reportable range of the CGM. Applicants should
calculate bias at various glucose concentrations (according to the
blood glucose reference values) of 60, 80, 120, 180, 250, 325, and
400 mg/dL and should include 95% confidence intervals.
Threshold alarm performance (Detection rates and false alarm
rates). In addition to point-to-point alarm detection rates,
detection rates should also be characterized according to whether
the CGM detected the hypoglycemic and hyperglycemic event within 15
and 30 minutes of the event. False alarms should be similarly
characterized in a point-to-point analysis and also in an analysis
which does not consider an alarm a false alarm if the event
actually occurred within plus or minus 15 and 30 minutes of the
alarm.
Prediction alarm performance. Performance of representative
prediction alarms that are utilized in the APDS should be
summarized. Sponsors should characterize detection rates according
to whether the CGM detected the predicted hypoglycemic and
hyperglycemic glucose level within the horizon setting(s). A
similar analysis should be performed for the false alarm rate.
Imprecision observed when sensors are inserted into the same
anatomical site and when sensors are inserted into different
anatomical sites.
Analytical specificity, including:
- Cross-reactivity with molecular compounds similar to
  glucose;
- Interference (both endogenous and exogenous
  compounds/conditions, as well as both prescription and
  over-the-counter medications).
- Environmental interference (e.g., from temperature or water
  exposure, such as bathing or swimming, etc.).
Study protocols should minimally include, as applicable: number of
patients, number of samples tested, number of replicates of each
sample tested, number of devices tested, matrix and concentration
of the sample tested, how CGM and blood glucose reference readings
were paired and the statistical analysis used.

Applicants should summarize important user functions as
characterized during the CGM pivotal clinical trial, including:

Length of sensor wear period. Applicants should present the
distribution of the number of hours that sensors remained
functional.
The number and percentage of CGM results that can be expected
to be generated during a wear period. Applicants should present the
distribution of the number of CGM results which were generated
during each individual wear period for all patients enrolled in the
trial. This analysis should include data from all sensors that were
calibrated and able to generate data for at least one hour. The
percentages of results generated should also be calculated using
the total number of values that were possible during each wear
period as the denominator, i.e., the sensor had not missed any data
in between the calibration and when the wear period ended.

2. BGD Component

Applicants should provide appropriate safety and effectiveness
information for the Blood Glucose Device (BGD) component of the
CGM. FDA recommends that Applicants refer to recent and applicable
study guidelines, such as those published by the Clinical
and Laboratory Standards Institute (CLSI), and applicable
FDA guidance documents to assist with the design of th e analytical
and clinical evaluation studies and the data analysis.¹⁵,¹⁶,¹⁷,¹⁸ Applicants should identify all applicable standards or
FDA guidance documents they followed as they evaluated the
device.

The performance specifications of the BGD and the study
protocols and data generated to verify them, including, for
example:
- Bias
- Imprecision
- Linearity
- Measuring range
- Traceability to reference materials or methods
- Stability of device components
- Expected values, as appropriate;
- Detection limit (e.g., limit of blank, limit of detection, and
  limit of quantification), as appropriate;
- Analytical specificity, as appropriate, including:
  - Cross-reactivity with compounds that have similar molecular
    structures, such as maltose;
  - Interference from endogenous compounds such as ascorbic acid;
    exogenous compounds such as prescription and over-the-counter
    medications; or medical conditions such as ketoacidosis or abnormal
    hematocrit concentrations; and.
  - Environmental interference (e.g., from temperature, humidity
    and altitude).
- User studies (where intended users collect (e.g., performing
  the fingerstick) and analyze the sample, and where results from the
  BGD are compared to results obtained with a traceable reference
  method.
- Matrix comparison, if more than one sample type may be
  analyzed.
- Lot release criteria used during the manufacturing of the BGD
  reagent or test strips.

Study protocols should minimally include, as applicable: number
of patients, number of samples tested, matrix and concentration of
the sample tested, number of replicates tested, number of meters,
number of test strip lots and the statistical analysis used.

If, as indicated in Section IV, the CGM is an FDA approved
device and the device has not been modified in a way which would
affect device performance then a copy of the device labeling may be
sufficient in lieu of the information in this section.

K. Control Algorithm/Signal Processing Functional
Component

The complete description of the control algorithm should be
provided as discussed in Section IV-C. In general, Sponsors should
submit the control algorithm for an APDS for review as part of the
PMA, consistent with FDA practice for other devices that use
similarly complex algorithms. In addition to the description, there
are critical elements of the control algorithm that should be
provided to support its safe use:

Control Algorithm Verification – prior to clinical use, the
sponsor should test the algorithm to assure that it has been
properly programmed into software and provide verification. Details
surrounding this verification testing can be found in Appendix
A-V-C. In addition, we recommend that Applicants describe how they
have assured the correct algorithm has been properly coded into
their final finished device as part of their PMA applications.
Parameter Sensitivity Analysis – the control algorithm in an
APDS may contain parameters that are adjustable by the healthcare
provider or patient. These adjustable parameters should be
identified in the device description. Although a limited
sensitivity analysis is expected prior to an IDE approval (Appendix
A-V-A-5), Applicants should provide a comprehensive parameter
sensitivity analysis as part of their PMA applications. This
analysis should evaluate the likelihood of an improper/unsafe
insulin dose or insulin pump shutoff for all combinations of
adjustable parameter values using representative CGM tracings. The
CGM tracings should be representative tracings of the selected
patient population.

If using theoretical or computer modeling to test the algorithm,
the Applicant should provide tracings that demonstrate device
behavior when values outside the bounds of expected use are
encountered so as to describe device behavior under worst-case
scenarios. A summary of the complete test report, justification of
how the CGM tracings used are representative of the intended
patient population, and reference to the full test report in the
software documentation set should be provided by the Applicant.

L. Infusion Pump Functional Component

FDA recommends Applicants provide appropriate information
regarding the safety of the infusion pump.

Drug Stability and Compatibility

The Applicant should demonstrate that the system does not
adversely affect the drug product being delivered by the infusion
pump and that these products do not adversely affect the infusion
pump.

If the infusion pump includes a reservoir or container, the
Applicant should provide stability and compatibility data, which
assesses the stability and compatibility for the recommended use
period and conditions included in the product labeling. The
assessment should include an assay to demonstrate that the drug
retains its specifications. The Applicant should also provide a
safety evaluation of any extractables, leachables, impurities and
degradants.

Some infusion pumps use syringes as the “drug reservoir” to
contain the product that is being infused. If the infusion pump is
set up in this configuration, the Applicant should adapt the
stability and compatibility testing stated in the paragraph above
to include combinations of drugs and syringes that are intended to
be used with the pump functional component.

The Applicant should verify that the mechanical structure and
function of any drug contacting components are not degraded to the
point that harm could occur to the patient or infusion pump
user.

As noted in Section VIII, Labeling, the Applicant should
identify the particular drugs that have been evaluated for use with
the infusion pump functional component. For pumps that utilize a
syringe as the “drug reservoir,” the labeling should identify the
specific syringes that are approved for use with the pump.

Catheter Occlusion Bench Testing

If the APDS is designed to completely turn off insulin delivery,
Applicants should provide a complete test report in the PMA
submission describing the bench testing performed to show catheter
occlusion does not occur when the pump is turned off. Details of
the Catheter Occlusion Bench Testing can be found in Appendix
A-V-K.

Dose Accuracy

Applicants should provide a complete test report in the PMA
submission describing the dose accuracy of the pump. The testing
should focus on the ability of the pump to accurately deliver the
recommended dose of the control algorithm. The purpose of this
testing is to understand how well the infusion pump can deliver the
wide range of recommended doses. Details of the Dose Accuracy Bench
Testing can be found in Appendix A-V-L.

VII. Clinical Study
Progression

The guidelines in this section are intended to facilitate timely
progression from an Early Feasibility study to a full pivotal
investigation while assuring adequate patient protections. In
developing these recommendations, we have considered the least
burdensome approach. Each step is designed to test specific aspects
of the APDS functionality and performance. Sponsors who believe
they already have sufficient, valid scientific evidence to fulfill
the purpose of a particular study phase and justify moving to the
next study phase are encouraged to discuss the evidence with FDA
staff. Such evidence may be clinical or non-clinical and may be
obtained from studies performed outside the US that comply with 21
CFR 814.15.

FDA recommends that the APDS be studied in three phases:
Early Feasibility Study, Transitional Study, and Pivotal
Study. The Early Feasibility Study is intended to
demonstrate that the APDS functions as expected and has no obvious,
unexpected safety concerns. The Transitional Study evaluates the
APDS under supervised, real-world conditions while the Pivotal
Study is conducted in an outpatient, unsupervised setting.

The size and duration of each study phase is dependent on the
design and features of the APDS and its proposed indication. We
recognize that each APDS will likely have unique features that
affect study design. Therefore, the guidelines below do not set
specific requirements for size and duration, but rather build in
maximum flexibility for completion of each study phase, while also
aiming to take the least burdensome approach. In addition, because
some APDS may be composed of parts that have already been approved
or cleared by FDA, we encourage Sponsors to leverage what is
already known about the safety and effectiveness of the individual
components to streamline the clinical testing of such a system.

A. Early Feasibility Study

The Early Feasibility Study (sometimes referred to as the Pilot
Study) is intended to demonstrate that the device system functions
as expected, does not have any obvious unexpected safety concerns,
and can address the hazards associated with errors in the
individual system components and the system as a whole. We expect
that these studies will be of short duration involving few
patients. The precise number of patients required will depend on
the device hazards identified and the success of the implemented
mitigating factors. Upon completion of the Early Feasibility Study,
the APDS should be ready for testing in a real-world setting with
close monitoring in a Transitional Study. The clinical protocol for
the Early Feasibility Study should be developed with an
identifiable goal and pre-specified success criteria.

The objective of the Early Feasibility Study might be to:

provide proof of a system concept, i.e., the components and an
algorithm;
examine the effects of specific modifications to the
algorithm;
validate performance of the APDS across the operating range of
the CGM;
isolate and examine how an algorithm performs when exposed to
conditions known to challenge the system, e.g., meal challenges,
exercise, obstruction in IV tubing and errors in the system’s
components; or
combine stress conditions in order to more realistically
capture home-use conditions.

Information gleaned from each study could be used to make
changes to the system or adjust the algorithm. It also might serve
to validate the algorithm. When each known or reasonably expected
hazard to the system is demonstrated to be adequately mitigated by
the APDS, then it may be appropriate for the device to enter the
Transitional Phase.

As investigators conduct their Early Fesibility Studies it is
not known whether the APDS will adequately mitigate the risks to
patients. For this reason we recommend that the Early Feasibility
studies be performed in a hospital setting, such as a clinical
research center (CRC). The study should be performed under the
close supervision of a medical team that can intervene to prevent
the occurrence of severe hypoglycemia or hyperglycemia. FDA
recommends that Sponsors demonstrate that their APDS can adequately
identify and compensate for CGM errors prior to moving to a less
supervised outpatient setting.

One condition that will undoubtedly stress the system is
significant errors in CGM readings. Currently marketed CGMs
experience periods when they generate incorrect data, e.g.,
indicating that glucose levels are significantly above or below the
true blood glucose value, or the CGM indicates that glucose levels
are rising when they are actually falling. When incorrect
information of this magnitude is fed into the APDS it can be
life-threatening to the patient. An additional consideration is
that CGMs may stop providing data, e.g., they may fail to provide
data for 2 hours or they might stop functioning altogether. This
latter condition might have serious consequences if the patient
were sleeping at this time and fails to respond to an alarm.

To provide safety monitoring and for purposes of assessing the
accuracy of CGM values during the study, reference blood glucose
measurements (i.e., those measured with a traceable laboratory
reference method) should be collected and checked frequently. The
reference blood glucose measurements will allow the Sponsor to
detect when a CGM error may be occurring and anticipate the
severity of the error and its effect on the patient.

The Sponsor should design the Early Feasibility Study to verify
that the APDS can adequately mitigate CGM errors. The following is
a current list of common types of CGM errors. One method by which
Sponsors could simulate these errors might be to manually enter
false CGM information into the APDS. The effects of these errors
should be evaluated at both low and high glucose levels as the
effects and potential impacts vary at different glucose levels.

Erroneously high CGM values;
Erroneously low CGM values;
Erroneous CGM trending information:
- CGM indicates that glucose concentrations are rising when they
  are actually falling; and
- CGM indicates that glucose concentrations are falling when they
  are actually rising;
The sensor stops functioning:
- For a short and intermediate length of time;
- Completely (such as what might occur when the sensor fails or
  when it reaches the end of the wear period);
- Any other hazards that occur during the Early Feasibility Study
  should be recorded as an incident for data analysis;
The pump stops functioning; and
The pump does not deliver the appropriate dose.

We recognize that, as CGM technology improves, these factors may
change. We encourage Sponsors to discuss with us any other
potential errors they wish to measure during a pre-IDE meeting.

B. Transitional Clinical Study

The Transitional Study evaluates the APDS to see if it functions
as expected under real-world conditions, while allowing for close
medical supervision. In other words, once the hazards associated
with the system have been shown to be sufficiently mitigated in the
Early Feasibility Study, we recommend that a supervised outpatient
study be performed. It is likely that these studies would be
conducted at settings such as a diabetes “camp” or
“dormitory”, which would allow subjects to undertake
more of their day-to-day activities while being closely monitored
by on-site personnel.

Transitional studies can provide the opportunity to gain
additional feasibility information for the development of the APDS.
Specifically, the study should evaluate APDS performance over the
life of its disposable components. The study should be conducted
for a duration that assures the safe and effective continuous use
of system components such as the CGM, catheter, and reservoir. For
disposable components, the study should be conducted for a period
of time that allows the Sponsor to measure device performance
before, during and after the disposable components are changed.
Sensors and pump reservoirs should be replaced as recommended.

Ideally, the Sponsor will use the same version of the APDS that
will be used in the Pivotal Study to ensure safe and effective
continuous use of system components. If modifications to the APDS
were made between the Early Feasibility/Transitional and Pivotal
study, bridging studies may be appropriate; however, it will depend
upon the type and extent of the change(s). We recommend the
Applicant/Sponsor seek FDA input via the pre-IDE process when
intending to make modifications to any of the device components
included as part of the system.

The Transitional Study should stress the APDS to identify any
potential limitations that should be addressed prior to the Pivotal
Study. When appropriate, additional blood glucose results can be
obtained.

Sponsors should provide their proposed outcome measures for the
Transitional Study in advance. We anticipate that the Transitional
Study will be reasonably small and of short duration; however, the
Transitional Study will only be considered complete, and the APDS
ready for use in a Pivotal Study, when the Sponsor can confirm that
the pre-specified clinical outcome measures have been met and the
device is safe for use in more patients in an outpatient
setting.

In an effort to minimize any delay between Transitional Study
completion and Pivotal Study initiation, the Sponsor may choose to
submit the Transitional and Pivotal Study plans for review
concurrently and predefine the Transitional Study success criteria.
FDA may approve the design of the Pivotal Study contingent upon
successful completion of the agreed upon Transitional Study plan
and submission of the study results.

C. Pivotal Clinical Study

The pivotal study should be performed in the actual use,
real-world environment and by the intended use population. It can
be conducted with subjects in their homes going about their normal
activities of daily living or in children not only at home but also
at school and participating in sports. The pivotal study should be
conducted with the finished APDS for which approval will be sought
and should be designed to demonstrate the safety and effectiveness
of the complete device system in the intended use population. A
description of an example of a pivotal study design can be found in
Section VII-D.

To demonstrate the safety and effectiveness of the APDS, we
recommend comparisons be made between patients using the current
standard of care, for example sensor-augmented pump control, and
patients using the APDS. The Agency recommends a 6-month minimum
study duration to allow for multiple periods of sensor wear,
changes in subject needs over time, and repeated HbA1c
measurements. We anticipate that, over time, designs of acceptable
pivotal trials will evolve as the devices to which a comparison
should be made change and as the technologies continue to
improve.

A robust trial design to validate an APDS could include a
randomized cross-over design or a randomized parallel design for
evaluating the safety and effectiveness of the APDS in an
outpatient setting. Patients in the control group should manage
their diabetes according to the standard of care, e.g, by
responding to alarms, performing finger stick blood glucose tests,
and determining therapy according to these results. Patients in the
test group should monitor their glycemic control by responding to
alarms, performing finger stick blood glucose tests, and adjusting
their treatment according to these results as directed by the
instructions for using the APDS.

The following considerations are important for the
Sponsor/Applicant to take into account when designing a pivotal
study.

1. Patient Population

The overall goal of an APDS clinical study is to determine the
safety and effectiveness of the APDS in maintaining glucose values
within range or near target while minimizing adverse events such as
hypoglycemia and hyperglycemia. Applicants may choose their study
population but should recognize that the population they select to
study may influence the study design, sample size, duration of
follow-up, and final approved device indications. FDA recommends
the following criteria be considered for enrolling patients with DM
into initial studies for the APDS:

Initial subject population:

Experienced with pump > 6 months
willing to perform ≥ 4 finger stick blood glucose
measurements daily
willing to perform required sensor calibrations
willing to wear the system ≥ 6 days per week
willing to keep a minimum log of:
- sick days
- days with exercise
- symptoms of low and high blood glucose episodes

The Agency recommends that initial studies for a novel APDS be
performed in subjects age >18 years to ensure adequate ability
of patients to respond to device problems. FDA is very interested
in promoting the development of a safe and effective APDS for
subjects < 18 years. Younger subjects may be enrolled after a
sufficient number of adults have demonstrated reasonable safety and
prospect of benefit (21 CFR 50 Subpart D).

To improve study efficiency by reducing sample size and study
duration, Applicants may wish to enrich the patient population with
subjects more likely to reach a study clinical endpoint. Some
examples of potential populations are as follows.

Patients with a high % HbA1c and frequent hypoglycemia despite
aggressive attempts at improved glycemic control;
Patients who have purposely maintained a high HbA1c to avoid
any hypoglycemia;
Patients who have frequent hypoglycemia.

Patients with Sensor-Augmented Pumps

Literature has indicated that the conversion from multiple daily
injection (MDI) to continuous subcutaneous insulin infusion (CSII)
and the addition of sensor guided therapy improves glycemic
control.¹⁹,²⁰,²¹,²² Therefore, it is important that any study designed to
examine the APDS should specifically test the effects of the APDS
function and not simply the effects of newly implemented
sensor-augmented pump control.

While the ideal patient population would consist of patients who
have already used sensor-augmented pump control for >3-6 months,
we recognize it may be difficult to identify and enroll patients
who are experienced with sensor-augmented pumps. As an alternative,
patients who have successfully used pumps without sensors can
undergo a training period with sensors for 4-6 weeks. This learning
period will screen out subjects who cannot optimally use
sensor-augmented pumps and reduce the likelihood that novel
sensor-augmented pump control would confound any effect observed in
the study. FDA recommends that final enrollment and randomization
occur after this training period.

Applicants who desire to develop their APDS for a specific
patient population are encouraged to seek the advice of the Agency
through a pre-IDE submission to determine how they can most
efficiently assess the safety and effectiveness of the device for
use in that population.

Broadening Patient Population

Sponsors may want to widen the criteria for the enrollment of
subjects (e.g., patients who are younger, who have insulin
resistance, or have co-morbidities that may increase their risk
during the study) and should consider how the inclusion of
different subject groups may affect the study design, endpoints,
and analysis of study outcomes. For example, if the Applicant
chooses to pursue a pediatric-specific indication²³, the pediatric inclusion and exclusion criteria should
be identified, as should any protocol-specific issues (such as
exercise or the daily volume allowed for blood draws, etc.).

2. Study Endpoints

Clinical study endpoints should represent objective
characteristics or variables that reflect how a patient feels,
functions, or survives. Surrogate endpoints should predict
meaningful clinical outcomes and be based on valid scientific
evidence.

Primary Endpoints

The primary endpoints for the pivotal trial should provide a
meaningful assessment of the APDS safety and effectiveness and
reflect the anticipated device indications for use. Secondary
endpoints can then be used to bolster additional claims or intended
uses. We recommend that Sponsors use HbA1c as their primary
endpoint.

HbA1c

HbA1c estimates the average glycemic exposure of red blood cells
over a 90-day period. It is the primary efficacy measure used in
the majority of trials assessing the effectiveness of a treatment
or intervention on glycemic control. Additionally, HbA1c has been
used to inform our understanding of the association of long-term
glycemia and the development of complications associated with
diabetes.

It is suggested that Sponsors using HbA1c as a primary endpoint
design a study to demonstrate a decrease (or, minimally, no
increase) in % HbA1c. We recommend a margin of at least 0.4% to
demonstrate a detectable difference.²⁴ We note that Applicants may propose that in certain
populations, acceptable increases in %HbA1c may be offset by
benefit in another endpoint (such as a reduction in hypoglycemic
events).

Co-primary endpoints may be acceptable and could be used by a
Sponsor to support a successful PMA. We encourage Sponsors to
discuss their choice of primary endpoint and study design with us.
Examples of co-primary endpoints may include:

CGM-Based Events

Most endpoints that will be used to evaluate APDS performance
require measurement or estimation of blood glucose levels. The
characteristics of an APDS necessitate frequent and long-term
measurement of blood glucose, and designing a study to achieve this
can be challenging. The American Diabetes Association (ADA)
Workgroup on Hypoglycemia acknowledged the limitations for
obtaining plasma glucose values noting that “although a
precise laboratory-based plasma glucose measurement would be ideal,
monitor-based estimates (or those with a validated glucose sensor)
are the only practical method.”²⁵

FDA proposes the use of a CGM-based correlate in evaluating the
APDS. To assist in the challenges associated with appropriate
endpoints for these innovative systems, FDA held a joint workshop
in collaboration with the National Institutes of Health,²⁶ which discussed the clinical expectations of clinical
studies for artificial pancreas device systems. As a result of this
workshop and continued collaboration between FDA and the
investigators of these device systems, FDA proposes the use of a
CGM-based correlate in evaluating the APDS.

The purpose of this section is to propose a potential method for
defining CGM-based events (CGM-BE) so that CGM-based data may be
used to evaluate device performance at targets or thresholds within
the system. The use of CGM-BE will permit comparison of the rates
of these events between treatment groups and further describe the
duration and magnitude of the CGM-BE.

Analysis of CGM data should be filtered to avoid erroneous
signals. For example, events should not be immediately preceded by
a decrease in glucose concentrations of ≥7 mg/dL/min as these
rates are not likely to be physiological. In addition, there may be
periods when the sensor either fails to report values or has
“noise”. Filters should be applied in a consistent and
pre-specified manner to exclude erroneous signals for the
definition of a CGM-BE. Sponsors should use experience gained in
the Early Feasibility Study to develop the appropriate signal
processing algorithms specific to their systems. Additional
filtering can also be provided with a justification.

Sponsors may propose alternate clinically-meaningful methods of
defining the CGM-BE beyond those outlined below.

CGM-BE for Hypoglycemia

The following is a proposed description of a CGM-BE for
detection of hypoglycemia:

A CGM-BE correlate for hypoglycemia may be defined as:
- CGM value < hypoglycemia threshold (e.g., 60, or 70
  mg/dL).
- A CGM value below the hypoglycemic threshold for at least 10
  continuous minutes.
- There is no patient intervention for 30 minutes after the
  activation of the alarm/suspend threshold. A period of 30 minutes,
  although arbitrary, is proposed because any changes in CGM values
  during this period of time are more likely to reflect the
  patient’s actions such as eating or restarting the pump (due
  to false positive alarm/suspend). Additionally, 30 minutes is
  within the expected duration of action for insulin infused prior to
  pump suspension. Other times can be used/proposed with
  justification.
- Each event should be described for duration and severity (e.g.,
  area under the curve).

We believe that a similar endpoint could be developed to detect
hyperglycemia. If a Sponsor wishes to develop such an endpoint, we
encourage them to discuss it with us during a pre-IDE meeting.

A Note about Time In Range – Time in Range (TIR) is an
endpoint that measures how successfully an APDS is able to
normalize or improve glycemic control (increased time in range)
without increasing hypoglycemia. Although studies have used TIR as
a primary endpoint, there is some uncertainty about whether it is a
good surrogate for determination of safety and effectiveness.
Currently, FDA does not believe that pivotal studies for APDS
should be based on a primary endpoint of TIR; however, as more data
is developed on this endpoint, we may be willing to accept it in
the future. Sponsors who wish to focus on TIR should talk to us
during a pre-IDE meeting to determine if it is an appropriate
choice for them.

Other effectiveness endpoints may also be considered based on
the intended use of the device.

Safety

There are many different endpoints that can be used to determine
the safety of an APDS. However, any safety study of an APDS should
determine that the APDS does not increase the incidence of severe
hypoglycemia (e.g., seizure or need for third party assistance),
severe hyperglycemia or DKA. In designing their safety study, the
Sponsor should propose specific safety endpoints that address these
concerns. Some examples of the types of endpoints Sponsors may
consider in developing a safety study include the following
metrics, which the APDS should be shown not to increase:

Severe hyperglycemia blood glucose above 240 mg and elevated
ketones;
DKA;
Number of CGM-defined hyperglycemic events;
Mean area under the curve (AUC) above 240 mg/dL as calculated
from CGM readings;
HgbA1C above a predefined accepted increase that may occur as a
result of reduction of hypoglycemia;
Percentage of CGM readings in the higher hyperglycemic
ranges;
Severe hypoglycemia (e,g, seizure or need for third party
assistance);
Number of CGM-defined hypoglycemic events;
Mean AUC below 60 or 70 mg/dL as calculated from CGM readings;
or
Percentage of CGM readings in the hypoglycemic range (< 60
or 70 mg/dL).

We are open to considering other safety endpoints based on the
intended use of the device.

Secondary Endpoints

Additional potential endpoints, as appropriate for the intended
use of the device (whether as primary or secondary), include, but
are not limited to:

Incidence of Diabetic Ketoacidosis (DKA) or severe
hyperglycemia within each group
Incidence of catheter blockage within each group
Capillary blood glucose values < 70 mg/dL and > 240
mg/dL.
Fasting whole blood ketone concentrations within each group,
evaluating elevated beta-hydroxybutyrate concentrations.
Time spent (hours/week) in hypoglycemic events <70 mg/dL and
hyperglycemic events > 240 mg/dL, including both day and
night
Average duration for all hypoglycemic events <70 mg/dL and
hyperglycemic events > 240 mg/dL within each group
Glycemic variability (such as coefficient of variation and
standard of deviation within each group)
Incidence and timing of CGM-BE for hypoglycemia and
hyperglycemia (e.g., timing during the day and night)
Safety and efficacy sub-group analysis, such as pediatric
subjects
Quality of Life

Exploratory Endpoints

Other exploratory endpoints may also be considered provided the
submission identifies those endpoints as exploratory, justifies the
use of these exploratory endpoints, and proposes a clinical study
that would allow further validation of these endpoint(s).

3. Statistical Analysis

Study Populations

The safety population should include all randomized subjects.
For effectiveness endpoints, two widely used populations are the
Intention to Treat (ITT) Population and the Per Protocol (PP)
Population. The Intention to Treat (ITT) population should include
all randomized subjects. The Per Protocol (PP) population should
include all randomized subjects who finish both treatment periods
successfully without major protocol deviations. The ITT population
is preferred for the analysis of primary endpoints. FDA recommends
the Sponsor/Applicant provide details on defining the major
protocol deviation in the PP population.

Primary Hypothesis

The Sponsor/Applicant should clearly state the hypothesis for
each primary endpoint and define the overall success criterion of
the study.

Superiority vs. Non-Inferiority

The statistical plan for the pivotal trial should be defined in
advance and may be designed to assess either non-inferiority or
superiority between the APDS and control group endpoints. The
Agency recognizes that a non-inferiority trial is more likely to be
successful. However, the Sponsor may choose to pursue a superiority
study to justify specific APDS labeled indications and claims (such
as superiority over other DM treatments). We have proposed certain
criteria for primary endpoints that Sponsors may apply in the
conduct of a superiority study, although the Agency remains open to
considering others if sufficiently scientifically justified.

a) HbA1c – Superiority

For Sponsors wishing to demonstrate APDS superiority for the
HbA1c endpoint, it is recommended that the study demonstrate a
superiority margin of 0.4% (absolute difference). The hypothesis is
mathematically expressed as:

H O: μ APDS ≥ μ CONTROL – 0.4%

H A: μ APDS < μ CONTROL – 0.4%

Where μ APDS is the mean of HbA1c (%) of the APDS group, and
μ CONTROL is the mean of HbA1c (%) of the Control group. The
Null hypothesis is rejected if the two-sided 95% upper boundary of
the difference between the two treatments, μ APDS – μ
CONTROL, is less than -0.4%

A goal for superiority in HbA1c should be 0.4%. However,
somewhat lesser improvements may be deemed acceptable based on the
effect on the co-primary endpoints and safety profile of the device
system.

b) CGM-BE for Hypoglcyemia – Superiority

For sponsors wishing to demonstrate APDS superiority for the
primary effectiveness endpoints in terms of reduction of severe
hypoglycemia or CGM-based hypoglycemic events, it is recommended
that superiority be demonstrated with a margin of 30% (relative
difference) for either severe hypoglycemia or CGM-based Event Rate
(prevention), or Event AUC (mitigation). The hypothesis is
mathematically expressed as:

H O: μ APDS ≥ 70%×μ CONTROL

H A: μ APDS < 70%×μ CONTROL

Where μ APDS and μ CONTROL are the endpoints of the APDS
and control groups, respectively. Specifically, for Event Rate,
μ ON and μ OFF are the proportions of hypoglycemic event as
previously defined. For AUC, μ ON and μ OFF are the means of
AUC per event. If AUC is not normally distributed, an appropriate
nonparametric test should be used to compare the distributions of
AUC within patients and between groups.

There is limited literature describing the clinical benefit for
a reduction in a CGM-BE outcome measure.
²⁷ Based upon this limited information, we recommend
a minimum 30% superiority margin be used to ensure that a
clinically significant reduction is observed. Smaller success
criteria may be appropriate with a justification based upon system
design, patient population, and risk profile. Please note, use of
smaller success criteria may affect the indication or claims for
the APDS.

Sample Size
Considerations

For a cross-over or parallel study design, sample size estimates
should be calculated. FDA recommends the overall significance level
be two-sided 5% and the overall power be no less than 80%. The
Applicant should make reasonable assumptions of important
parameters, including the means and standard deviations of % HbA1c
and other primary endpoints, the correlation between groups and
within subjects, the loss-to-follow-up (LTFU) rate, and provide
justifications for these assumptions. An appropriate statistical
method should be provided to calculate the overall sample size
while controlling the overall type I error rate under 5% and
maintaining the overall power above 80%. If necessary, simulation
might be needed to calculate the sample size.

If an interim analysis is planned, the sample size should be
further adjusted using appropriate methods to control the overall
false positive rate.

Handling of Missing
Data

Starting at the study design stage and throughout the clinical
trial, every effort should be made to minimize patient withdrawals
and loss to follow-ups. Premature discontinuation should be
summarized by reason for discontinuation and treatment group. For
an ITT population, an appropriate imputation method should be
specified to impute missing HbA1c and other primary endpoints in
the primary analysis. It is recommended that the Sponsor/Applicant
plan a sensitivity analysis in the protocol to evaluate the impact
of missing data using different methods, which may include but is
not limited to per protocol, Last Observation Carry Forward (LOCF),
multiple imputation, all missing as failures or success, worst case
scenario, best case scenario, tipping point, etc.

General Considerations for Data
Analysis

FDA recommends that patient demographics, medical history, and
other important baseline characteristics (e.g., HbA1c, body mass
index, average daily insulin requirements, education levels, etc.)
are summarized using descriptive statistics and frequency tables as
appropriate. Patient accountability and withdrawals from the
treatment phase of the study should be reported. Summaries (number
and percent) of the reasons for withdrawals should be presented by
treatment group. The effects of carryover, sequence, site, baseline
variables and prognostic variables should be tested using
appropriate models (usually, a linear model for a continuous
variable and a logistic regression for a binomial variable).

Analysis of Primary
Endpoints

The primary effectiveness analysis is the between-group
comparison of all primary endpoints. Appropriate statistical models
should be specified to evaluate the impact of covariates. If some
covariates are found to confound a primary endpoint, their effects
should be adjusted through appropriate models.

Analysis of Secondary
Endpoints

For all secondary endpoints, descriptive statistics are
recommended. If the Sponsor/Applicant intends to make claims for
any of the endpoints in the labeling then the hypotheses,
statistical analysis, and success criteria should be clearly
specified in advance in the study protocol. An appropriate
multiplicity adjustment strategy to control the type I error rate
may also needed.

Adaptive Study Design

Adaptive study design provides flexibility in modifying some
aspects of the clinical study during the clinical trial. If an
adaptive study design is desired, the Agency recommends that the
Sponsor/Applicant prespecify details such as the number of interim
analyses, the time at which these analyses will be performed, the
stopping rules, and the criteria to control the type I error rate,
etc. Due to the complexities of an adaptive study design, FDA
recommends the Sponsor/Applicant include their proposal in their
pre-IDE submission to discuss their design.

Safety Analyses

Descriptive statistics of all adverse events should be presented
for the safety population. The descriptive statistics of the
following subgroups should also be summarized. This includes, but
is not limited to, the following information:

All adverse events

Serious Adverse Events (SAE)

Adverse events and other reasons that lead to patient withdrawal
from the trial

Unanticipated Adverse Device Effects (UADE);
Severe Hypoglycemia (prospectively defined in the
protocol);
Severe Hyperglycemia (prospectively defined in the
protocol);
Diabetic Ketoacidosis (prospectively defined in the
protocol);
Ketone testing: Urine ketones should be measured every morning
to screen for preceding nocturnal ketosis. Capillary blood ketone
levels (betahydroxybutyrate) should be evaluated any time the blood
glucose is above 300 or if the subject is experiencing nausea,
abdominal pain, or vomiting. First morning urine ketones may be
positive even if the fasting blood is negative for
betahydroxybutyrate if transient nocturnal ketonemia occurred
earlier during the night as a result of insulin suspension, but
subsequently resolved with resumption of insulin infusion; and
Skin infection.

The premarket application must include labeling in sufficient
detail to satisfy the requirement of 21 CFR 814.20(b)(10), which
states that copies of all proposed labeling for a device must be
submitted in a PMA. Labeling must also satisfy the requirements of
21 CFR Parts 801 & 809.

In general, labeling for the APDS should include:

a user manual for the patient, written at an 8 th grade reading
level;
all training materials;
professional labeling for the prescribing physician;
Package inserts for the APDS and all components packaged
separately from the system (e.g., BGD reagents or test strips,
quality control materials, catheters, inserters, reservoirs, etc.);
and
Box and container labels for the APDS and each component that
is packaged separately from the system.

Applicants may refer to the following documents for information
on important principles for developing clear and complete labeling
for the CTR/CTT system.

Guidance on Medical Device Patient Labeling; Final Guidance for
Industry and FDA (2001)
Labeling of Home-Use In Vitro Testing Products; Approved
Guideline, CLSI GP-14 (1996)
Device Advice website titled
Labeling Requirements – In Vitro Diagnostic Devices
IEC 60601-1-11 General Requirements for the basic safety and
essential performance – Collateral standard: Requirements for
medical electrical equipment and medical electrical systems used in
the home healthcare environment

The patient instructions for use should be as simple and concise
as possible and be easy to understand. Labeling for lay users
should be written at an 8 th grade reading level. Applicants should
consider the use of graphics such as line drawings, illustrations,
photographs, tables and graphs. Applicants should ensure that terms
are used consistently throughout the labeling and should avoid
using synonyms or alternate phrases. Comprehensive directions for
preparation and use of all functions of the APDS and the
accessories should be provided. The Applicant should provide
labeling that contains examples of expected performance and
addresses issues that may occur in the home environment.
²⁸

The professional labeling for the prescribing physician should
describe in sufficient detail the clinical testing performed for
APDS approval. The purpose of this information is to allow the
physician to make an informed decision on whether to prescribe the
APDS to a particular patient. Information such as indications,
warnings, precautions, contraindications should be provided. In
addition, critical bench testing for the infusion pump (e.g., MR
testing and drug stability testing) and CGM (e.g., analytical
specificity, accuracy, etc.) should be described.

Appendix A: IDE Content
for APDS Studies

This Appendix provides suggested content for IDE submissions for
evaluation of an APDS. This Appendix is structured as an outline of
the IDE submission and identifies the elements of an IDE review.
FDA recommends that Sponsors follow this outline and address each
section heading as part of their IDE submission. When different
information is needed between Early Feasibility, Transitional, or
Pivotal (unsupervised outpatient setting) Studies, the section is
divided.

The Sponsor should provide background information related to the
development of the APDS that it intends to studied. The Sponsor
should identify whether there has been previous communication with
the Agency regarding this device within a pre-IDE submission (the
Sponsor should identify the pre-IDE #) or previous US or Outside
the US clinical studies performed using this device system (the
Sponsor should identify the IDE #).

II. Device
Description

This section should include a device description of the
APDS.

If the Sponsor is using previously approved/cleared devices,
please include following information for each device:

the name of the device
model number
PMA or 510(k) number for the referenced devices
Identify if the functional component has been modified from its
approved/cleared form. If so, the Sponsor should describe how the
device has been modified. This would include, for example, whether
the run-in time or calibration frequency have been modified. For
all components that have been modified, Sponsors should provide a
rationale for the change and an analysis of the likely impact it
will have on the performance.

If the Sponsor is not using previously approved/cleared devices,
FDA recommends the Sponsor include a complete description of all
functional components of the system (i.e., BGD, CGM, control
algorithm, and pump) as described in
Section IV of the guidance.

The Sponsor should also identify the insulins and/or other drugs
that are intended to be used with the APDS in the clinical
study.

III. Letters of
Authorization

Letters of Authorization (LOA) are needed if the Sponsor intends
to reference safety/effectiveness information from another
manufacturer that has been included with a device Master
File²⁹ or another regulatory submission. Some examples are
identified below.

If the Sponsor intends to use a medical device from a different
manufacturer that has been modified and there is a document such as
a device master file describing the changes and additional testing
for this modification.
If the Sponsor intends to use a device from a different
manufacturer contained within a document such as a device master
file that allows the interconnection of various device components
into one system.

IV. Indication for
Use

Describe the indication for use. Please refer to
Section V of the guidance .

V. Nonclinical
studies/Prior Investigations

Per 812.27, a report of prior investigations shall include
reports of all prior clinical, animal, and laboratory testing of
the device and shall be comprehensive and adequate to justify the
proposed investigation.

The Agency recommends the Sponsor also provide the following
information as part of the IDE.

A. Algorithm

The Sponsor should provide information regarding the regulatory
status of the algorithm to be used in the IDE study. If the device
in question utilizes software that has not been previously reviewed
and cleared or approved by the Agency, a description of the
algorithm should be provided as part of the IDE submission. If the
algorithm is identical to that used in a cleared or approved
product, then the IDE submission should contain the name of the
product and the FDA document number under which it was cleared or
approved (if known).

If the Sponsor does not have access to the algorithm and cannot
provide it, a rationale for why it is not being provided should be
included as part of the IDE.

1. Definition of Algorithm(s)

FDA recommends the Sponsor define the algorithm in symbolic form
and briefly define the purpose for each equation in the control
algorithm.

2. Definition of Algorithm Symbols/Parameters

FDA recommends the Sponsor define each symbol (i.e., parameter)
in the algorithm. This can be in table format.

3. Identification of Fixed Parameters

FDA recommends the Sponsor identify each fixed parameter and the
value of this parameter. FDA defines a fixed parameter as a
parameter value that will not be changed during the course of the
clinical study. This can be provided in a table format.

Symbol/Parameter	Value

4. Identification of Adjustable Parameters that May be
Modified During the Study

The Sponsor should identify each parameter and parameter value
range that may be adjusted during the course of the study. This can
be provided in table format.

Adjustable parameters
Parameter	Symbol	Typical Starting Value	Minimum Value	Maximum Value

Please note, that once approval of the IDE is obtained, if the
Sponsor modifies the adjustable parameter within the predefined
ranged, the Sponsor can continue the study without Agency
notification.

5. Parameter Sensitivity Analysis

For each parameter that is defined as adjustable, the Sponsor
should provide a parameter sensitivity analysis to show the
equation does not result in unsafe dosing adjustments. For a PMA,
FDA recommends the Sponsor evaluate combinations across the entire
range of parameter values and the effects on the system as
described in Section VI-K of the guidance. Such an analysis should
evaluate combinations of adjustable parameters using the minimum,
maximum and typical starting value for each adjustable parameter.
The analysis should identify if any unsafe dosing adjustments have
occurred. This type of analysis can be evaluated using CGM glucose
tracings that would approximate the expected tracings observed in
the study. For Early Feasibility studies (Section VII-A of the
guidance) where patient safety has been significantly mitigated due
to physician monitoring, a limited sensitivity analysis is
acceptable.

6. Summary of the Verification Activities for the Control
Algorithm

The Sponsor should provide a summary of the testing (i.e.,
verification activities) they have performed to show that the
algorithm has been properly programmed into the software to support
the safe and effective use of the device in any IDE for a pivotal
study. This summary should identify the test method used to verify
the algorithm and reference where the detailed test reports can be
found in the software documentation set.

B. Software Documentation

Software documentation should be provided for the APDS prior to
major clinical studies. Full software documentation is not
necessary for Early Feasibility Studies. However, documentation
should be provided to demonstrate that the Sponsor can trace the
development history of all components of their software and
identify any unresolved anomalies (i.e., “bugs”) that
may affect the safety of their software for the purpose of
providing complete software documentation at a later time. For
devices that have been modified from their previously
approved/cleared form, the Sponsor should highlight any differences
between the modified and approved/cleared versions. For assistance
in developing the appropriate documentation set, Sponsors should
refer to the FDA’s 2005 software guidance document.³⁰ All APDS are identified as a major level of concern
for purposes of the guidance. The software documentation set can be
included as an Appendix to the IDE.

We encourage all Sponsors whose APDS is comprised of previously
approved/cleared devices that they did not manufacture to pursue
obtaining an LOA from the manufacturer of such devices to gain
access to the software Master File. If the Sponsor is not able to
gain access to the software documentation in this way, the Sponsor
should provide evidence of his or her attempts to obtain the
documentation and an attestation of the manufacturers’
refusal to provide it as part of the IDE.

C. Summary of System Communication

If the APDS connects a CGM to a control algorithm and/or a
control algorithm to a pump in which information is passed
automatically (without user acceptance) and this is not a
previously approved device system, a summary of the system level
testing is needed prior to a pivotal study. This summary should
address how the Sponsor has ensured the correct passage of
information such as CGM values and or insulin dosing
recommendations. This summary should identify the test method used
to verify the algorithm and reference where the detailed test
reports can be found in the software documentation set.

D. Safety Measures for Dosing

The Sponsor should identify if there are any hard-limits coded
into the software of the APDS that would restrict the minimum and
maximum dose recommended by the algorithm. The Sponsor should
identify the frequency of dosing recommendations and the time
needed to deliver the minimum and maximum dose.

E. Biocompatibility Testing

FDA recommends biocompatibility testing of the device in
accordance with FDA blue book memo,
Use of International Standard ISO 10993, ‘Biological
Evaluation of Medical Devices Part 1: Evaluation and Testing’
(Replaces #G87-1 #8294). The Sponsor should provide appropriate
biocompatibility testing for duration and level of contact. FDA
recognizes that early studies may use device components previously
approved/cleared. If this is the case, the Sponsor should provide
the appropriate cross-reference (or an LOA) to reference the
appropriate PMA or 510(k) documents. If the Sponsor has modified
the approved device, it may be possible to reference the
biocompatibility of the approved/cleared devices if the Sponsor can
justify how the modifications do not affect the biocompatibility.
If the Sponsor uses a new device, then complete biocompatibility
documentation is needed as described in Section VI- of the
guidance. FDA notes that the biocompatibility testing provided in
the IDE may be limited due to the short duration of contact of the
APDS in the proposed clinical study design.

F. Electrical Safety

If applicable, the following electrical safety information
should be addressed in any IDE submission for major clinical
studies. This information may not be necessary for Early
Feasibility studies.

Electromagnetic Compatibility

The IDE submission should include a complete description of the
Electromagnetic Compatibility (EMC) characteristics of the device,
and the information to verify those characteristics.
Electromagnetic compatibility is the ability of a device to operate
properly in its intended environment of use without introducing
harmful electromagnetic disturbances into that environment.

The Agency recommends that the APDS system meet the EMC
requirements of IEC 60601-1-2. IEC 60601-1-2 describes EMC testing
and includes both tests for immunity of the device to outside noise
and emissions from the device to the outside. In addition to
evidence of compliance with this standard, complete testing
information describing what was done, how the device functions,
modes that were tested, pass/fail criteria, reference standards,
any deviations or allowances that were taken, and any device
modifications needed to pass the testing should be provided with
appropriate labeling.

Applicable Standards

The Sponsor should identify if the device meets the electrical
safety requirements of IEC 60601-1. Complete test reports
demonstrating that the device meets the electrical safety
requirements should be provided.

Radio Frequency

If the submission includes radio frequency (RF) technologies,
the IDE submission should include a complete description of the RF
use. While applications of RF wireless technologies might comply
with applicable technology standards and Federal Communications
Commission rules, medical device safety and effectiveness concerns
may remain. For detailed information about possible hazards,
reference should be made to the draft guidance,
Radio-Frequency Wireless Technology in Medical
Devices.³¹

Particular points that should be addressed in the IDE include:
quality of service needed, data integrity, coexistence, security,
and EMC. Due to the increased use of RF wireless technology that
operates in the same frequency range, RF wireless coexistence via
testing with other common applications of RF wireless technology
that can be expected to be in the environment of use should be
carefully addressed. The testing should also address the ability of
two or more of the systems to co-operate wirelessly in
proximity.

If the Sponsor is using previously approved or cleared products,
the electrical safety may have been addressed in another regulatory
submission. The Sponsor should evaluate any differences in the test
environment from the proposed clinical study and the
approved/cleared devices. Differences in test environments (e.g.,
home vs. hospital use) may require additional electrical safety
testing. The Sponsor should justify these differences are minimal
or provide additional testing.

G. Animal/In-Silico Testing

The Sponsor should provide evidence of safety for the APDS
intended to be studied. The Agency has accepted different types of
nonclinical studies to support IDE approval. These types are
briefly described below.

Animal testing

Animal testing should employ a device system similar to that
intended for use in the clinical study. If there are any
differences in the system or study timeline of the animal study
versus clinical study, these differences should be identified. The
Sponsor should justify that the differences would not affect the
safe use of the device in humans. The animal model should best
represent the intended patient population, and a justification
should be provided. Prior to performing animal studies, the Agency
recommends that the Sponsor seek FDA input on the animal study
protocol via pre-IDE. FDA recommends the nonclinical laboratory
studies be conducted in accordance with 21 CFR 58, Good Laboratory
Practice for Nonclinical Laboratory Studies. Please note that all
real-time traces of the animal study should be provided in an
appendix.

In-Silico Testing

As part of the Artificial Pancreas Critical Path Initiative, the
Agency has accepted in-silico (i.e., software-based theoretical
models) modeling as a reasonable nonclinical assessment tool. An
in-silico model is a method to test the control algorithm in a
theoretical human model of insulin and glucose metabolism using a
sophisticated computer model rather than expensive and
time-consuming animal experiments. This tool can be used to justify
and support initiation and expansion of human clinical trials.
Prior to using an in-silico model, the Agency recommends submission
of the model for FDA review under a pre-IDE. This model should
minimally include the variability in human glucose metabolism,
performance characteristics of the CGM and insulin pump, the
pharmacokinetics of insulin, and diffusion of glucose between the
blood and interstitial fluid. A complete test report for the
in-silico testing of the control algorithm should be included in
the IDE submission. Due to the flexibility of a theoretical model,
the Sponsor should design the in-silico model similar to the
proposed clinical study. All real-time traces should be provided in
an appendix in the IDE.

I. Human Studies

FDA recommends the Sponsor provide all reasonably known clinical
data applicable to the safe use of the APDS in humans. This may be
clinical data to support device components of the system (e.g., CGM
clinical studies), studies conducted previously in another IDE or
studies conducted outside the US. FDA recommends a complete test
report be provided.

J. Human Factors/Usability Testing

Early Feasibility (Pilot) Study

The Early Feasibility Study will typically involve proof of
concept for the technology; thus, user interaction with the device
will not be the focus of the study. Human Factors/Usability Testing
are generally not expected to be part of the Early Feasibility
Study. Use errors should be collected and described in the case
report forms.

Transitional Study

Risks associated with use error are present for any operator of
an APDS. FDA recommends the Sponsor evaluate the design of the
system, including its labeling and user training, during the
transitional study phase to establish that the design of the system
supports safe and effective use by the intended users under
conditions of simulated use prior to the initiation of an
unsupervised outpatient clinical study. Alternately, Sponsors
should design parallel human factors evaluations to assure that
human error is controlled and use-related risks are mitigated as
much as possible prior to the pivotal study. Sponsors should record
all human factors test protocols and results in the Case Report
Forms.

Pivotal Study/Unsupervised Outpatient Study

In an unsupervised outpatient study, data collection is limited.
Self-report data (e.g., as recorded in patient diaries) and calls
to telephone help lines can provide useful information; however
these types of data should be supplemented with essential user
performance data, for example collected through a data logger
incorporated into the devices used in the study. Depending on the
results obtained and consequent design modifications implemented
following analysis of results of the evaluation, it might be
necessary to perform a simulated-use test to assess the
effectiveness of the modifications and overall use safety of the
system. Soliciting comments from the participants would provide
essential additional information regarding users’ perceptions
of the system, potential use-related problems, and ways in which
the system might be improved.

An overview of human factors/usability testing processes is
described in Section VI-I of this document. FDA recommends the
Sponsor conduct, describe, and provide a rationale for the human
factors/usability testing they conducted to support the safe use of
the system in humans in the outpatient setting.

K. Catheter Occlusion Bench Testing

APDS are in part intended to improve glycemic control by
modulating insulin infusion, including, in certain instances,
shutting the pump off for finite periods of time. Insulin
crystallization is a chemical process that occurs with or without
flow, but the likelihood of crystallization is increased in the
absence of flow. Such crystallization raises the risk of catheter
blockage and the inability of the pump to deliver the appropriate
insulin dosage when the system returns insulin delivery. Although
the incidence of catheter blockage due to insulin crystallization
can be further evaluated in a clinical study, FDA recommends this
risk be assessed via appropriate bench testing prior to an
unsupervised clinical study. The testing of this system should
mimic the conditions of the clinical study as closely as possible.
Temperature should reflect the use environment and to ensure
safety, the duration of time evaluated should be double the maximum
time allowable for pump shutoff in the system. FDA recommends the
Sponsor report the incidence of crystallization and the incidence
of catheter blockage due to crystallization.

L. Dose Accuracy Bench Testing

APDS control algorithms currently recommend periodic insulin
dosing with frequencies ranging from 1-30 minutes. These frequent
bolus doses are extremely small and ask the pump to perform
accurately near the lowest doses available in the pump. In order to
understand how well the APDS can deliver the recommended insulin or
drug, bolus dose accuracy testing should be performed. This testing
should evaluate the APDS using the most frequent dosing rate (i.e.,
the shortest time between dosing adjustments) and accuracy
measurements should be tested using the minimum dose, maximum dose
and incremental doses between the min and max. The testing of this
system should mimic the conditions of the clinical study and the
measurement technique should account for evaporation of small doses
during the testing. Temperature should be controlled and reflect
the use environment.

M. Diagnostic Devices To Be Used During the Clinical Study

In addition to the diagnostic device components of the APDS,
other diagnostic devices are commonly used during clinical studies,
(e.g., those that measure blood glucose for purposes of evaluating
the APDS, or ketones).

To ensure patient safety and the accuracy of these devices
Sponsors should provide the following information for each
diagnostic device that will be used onsite in the clinical
study:

Name of the device, including model numbers, as
applicable.
Description of the function performed by the device during the
study (e.g., monitoring patient glucose or ketone concentrations as
a secondary endpoint in the study or calibrating the CGM).
Regulatory status of the device (including the FDA document
number, if known).
List of all device components and accessories. In addition to
the instrument, reagents and quality control materials, accessories
might include standards (calibrators), data transmitting equipment
or software.
For labeling recommendations of device components that are part
of the APDS, please refer to Appendix A-XIII.
For diagnostic devices used in the clinical study that are not
part of the APDS system.
- Unless a justification can be provided, Sponsors should provide
  patients who are operating any device all labeling associated with
  the device.
- Sponsors should determine whether it is necessary to provide
  clinical investigators with the labeling of diagnostic devices used
  at the study site. If Sponsors believe that this is not necessary
  they should:
  - Describe what functions the investigator will be performing
    with the device and explain why it is not necessary to provide them
    with the labeling. (For example, the operator may have extensive
    experience operating the device.)
  - Certify that labeling will be available at each clinical site
    should it be needed.
If the device was previously cleared or approved, Sponsors
should describe any physical or labeling modifications that were
made to the device for purposes of conducting the study. If
modifications were made, Sponsors should:
- Describe the modification, provide a rationale for the change
  and description of how the modification might affect device
  use/performance.
- Sponsors should also address how they will ensure that the
  instructions for use properly communicate any changes in how the
  device is to be operated, if applicable.
Sponsors should describe how performance of the device will be
monitored to ensure accurate results. This information should
include, where applicable:
- Quality Control (QC) materials to be analyzed.
- Number and concentration of QC materials.
- Frequency and timing of analysis of the QC materials.
- Criteria for determining acceptability of QC results.
Sponsors should describe how individuals using the device
during the study will be trained to operate it.

N. Drugs Used During the Study

Please identify the name of the drugs (e.g., insulin, glucagon,
etc.) intended to be used in the APDS system and provide the drug
labeling. Sponsors should also indicate any drugs, such as
acetaminophen, that are given to patients during the study, as they
may affect CGM performance.

The Sponsor should provide a bibliography of all relevant
publications. Copies of critical publications needed to support the
proposed study should be included as an appendix.

VII. Clinical
Study

A. Purpose/Objective(s)

The Sponsor should briefly describe the purpose/objective of the
study.

B. Study Design

The Sponsor should briefly describe the study design. For
example:

A nonrandomized double center study with X subjects who have
Type I Diabetes will participate in one X hour inpatient
experiment. The study will compare the treatment arm to a control
arm. The arms are defined as:

Treatment Arm
Control Arm

C. Sample Size and Investigational Sites

The Sponsor should define the number of subjects that are
intended to participate in the study, the proportion of male to
female, age range, Type of diabetes, etc. The Sponsor should
identify the investigational site(s) and include the address for
each site.

D. Study Duration

The Sponsor should define the study duration for each subject
(e.g., subject will participate in two 24-hour experiments). The
Sponsor should also define how long they plan entire study will
take to complete.

E. Inclusion Criteria

The Sponsor should provide a listing of the inclusion
criteria.

F. Exclusion Criteria

The Sponsor should provide a listing of the exclusion
criteria.

G. Study Timeline

The Sponsor should provide a detailed description of how the
study will be performed. For example:

Enrollment Visit:

Informed Consent is obtained from eligible subjects,
etc.

Activities performed prior to CRC or Study
Admission:

Sensor placement, etc.

CRC Admission:

Detailed description of the CRC timeline

Follow-Up

Describe the criteria used to determine when a subject can
safely be discharged from the CRC.
Describe when and how often a health care provider will
follow-up with the subject after discharge.

H. Safety Monitoring/Risk Analysis

Describe the Safety Monitoring that will be performed during the
study. For example:

Glucose Monitoring Risk – FDA recommends that performance of
the APDS be assessed, in part, by evaluating blood glucose
measurements taken from the subject while they are enrolled in the
clinical study. It is therefore important to collect the most
accurate glucose information possible.
- Early Feasibility Study (In-hospital): For studies
  taking place in CRC settings, Sponsors should use the most accurate
  method available for measuring subject glucose concentrations,
  i.e., traceable reference methods. Reliable laboratory tests, such
  as those utilizing a hexokinase method, are most appropriate.
- Pivotal Study (Outpatient): The need for accurate
  glucose information also exists for studies taking place in the
  home setting. The Sponsor should carefully consider the BGD that
  they intend to use and assess the risk for measurement error.
Hypoglycemic/Hyperglycemic Risk – To decrease the risk of
severe hypoglycemia and hyperglycemia, the Sponsor should construct
a schedule for monitoring blood glucose concentrations. The Sponsor
should address how the interval of sampling and method of
determination may be affected by the subject’s current blood
glucose value or period of the trial, such as during hypoglycemia
induction. This information can be provided in tabular format.

Blood Glucose (mg/dL) Frequency of BG measurement

0-XX X min

XX-YY Y min

Please Note: The Sponsor should describe how they will intervene
for hypoglycemic and hyperglycemic episodes. This description
should include time and glucose concentration. The Sponsor should
describe how each defined episode will be treated.
Calibration of CGM risk – When an erroneous glucose value is
used to calibrate a CGM, the bias is carried through until the next
opportunity to re-calibrate the CGM. This can result in an
incorrect bias that lasts for 12 hours. Sponsors are encouraged to
mitigate the risks posed by BGDs as much as possible when designing
studies because they are used to calibrate the CGMs and could
result in inappropriate insulin dosing.
Sterilization Risk – The Sponsor should identify and describe
if all of the devices are sterilized, where required. If not, the
Sponsor should assess this risk.
Reuse Risk – The Sponsor should describe if components of the
system can be reused for other patients within the study. If
applicable, the Sponsor should describe if the reusable devices are
patient contacting. If they are patient contacting, the Sponsor
should describe the reprocessing (cleaning, disinfection,
re-sterilization) of the reusable devices. Please note, validation
may be needed to ensure reusable devices have been adequately
cleaned, disinfected and re-sterilized, as applicable.
Hb1Ac risk – Please refer to Section VII-C-2 of the guidance
regarding the risk of variability in HbA1c measurements.
Misuse Risk – Sponsors should provide a detailed description of
how training will take place regarding the operation of the APDS
and all of the functional components during the study. As
applicable, this should include training for clinical staff and/or
the study subject. If the study is being conducted for the purposes
of supporting a marketing application, all training of staff and
users should mimic that which will take place when the system is
marketed. This includes written materials, videos and or
checklists.
Risks of blood sample collection, contamination from sampling
techniques. Sample collection procedures in hospitals are
responsible for a significant number of errors when patients are in
hospitals. This is particularly true when samples are taken from an
intravenous (IV) line, irrespective of the fluids being
administered. Ideally, the technique used to obtain the sample
should limit the amount of blood taken so as not to harm the
patient. The technique should ensure mitigation of the risk of
contamination.

Blood Glucose (mg/dL)	Frequency of BG measurement
0-XX	X min
XX-YY	Y min

I. Stopping Rules

The Sponsor should describe stopping rules for the subject and
study.

The Sponsor should describe under what subject conditions the
patient study would be halted.
The Sponsor should describe under what study conditions the
entire study would be halted. For example, if 3 subjects were
consecutively stopped.

J. Endpoints

The Sponsor should define the primary and secondary endpoints
for safety and effectiveness.

K. Success Criteria/Goal

The Sponsor should define how the study will be determined a
success.

Early Feasibility & Transitional Studies

In Early Feasibility studies, the success criteria can be general.
FDA recommends the Sponsor identify criteria that would allow the
Sponsor to progress to the next study.

Pivotal Study

FDA recommends the Sponsor provide success criteria in accordance
with the statistical plan.

L. Statistical Analysis Plan

Early Feasibility & Transitional Studies

These studies typically do not have sufficient sample size to allow
for a statistical analysis. The Sponsor should describe the
analysis that will be used to determine progression to the next
phase of the study.

Pivotal Study

The Sponsor should describe the complete statistical analysis plan
to support the study objective(s).

VIII. Informed
Consent

The Sponsor should provide a statement that all forms and
informational materials to be presented to the subject were
submitted and included in the IDE application. A copy of the
informed consent and any informational or recruiting materials
should be provided as an Appendix. All Informed Consent documents
must adhere to the requirements described in 21 CFR Part 50 –
Protection of Human Subjects and must contain the information
described in 21 CFR 50.25(a). If the Sponsor chooses to pursue a
pediatric-specific indication they must be aware that the pediatric
population represents a vulnerable subgroup and special measures
should be taken to protect the rights, safety, and welfare of
pediatric study subjects. The regulations at
21 CFR Part 50 – Subpart D Additional Safeguards for Children
in Clinical Investigations further describe specific requirements
for pediatric study subjects.

FDA does not recommend that the consent process include only a
“short form” written consent (see section 50.27(b)(2)).

IX. Patient Case Report
Form(s)

The Sponsor should provide a draft copy of the case report
forms.

X. Investigator
Agreement Forms

If the investigators are determined prior to the IDE submission,
the Sponsor should identify the name and address of each
investigator that will participate in the study. The Sponsor should
provide an Investigator Agreement Form and this form should
minimally have the information contained within 21CFR 812.43(c)(4).
In addition to this form, the Sponsor should certify that no
investigator will participate in this study prior to signing the
investigator agreement form.

Financial disclosure of clinical investigators participating in
a clinical study is a requirement which applies to any clinical
study submitted in a marketing application to the FDA. Financial
interests can be a potential source of bias in the outcome of a
clinical study; therefore, any financial arrangements must be
disclosed. As per 21 CFR Part 54 – Financial Disclosure by Clinical
Investigators, Sponsors must certify the absence of certain
financial interests of clinical investigators on Financial Interest
Form: Certification:
Financial Interests and Arrangements of Clinical Investigations FDA
Form 3454, or disclose those financial interests on Financial
Interest Forms: Disclosure:
Financial Interests and Arrangements of Clinical Investigators FDA
Form 3455.

XI. Monitoring
Information

The following is recommended for adequate monitoring
information.

Written procedures for monitoring and the name and address of
any monitor (21 CFR 812.25(e)).
Monitor will report to the Sponsor any noncompliance with the
signed Investigator’s Agreement, conditions imposed by the IRB or
FDA, and the requirements of the IDE. Sponsor shall then
either secure compliance, or discontinue shipments of the device to
the investigator and terminate the investigator’s participation in
the investigation (21 CFR 812.46(a)).
A Sponsor shall select monitors qualified by training and
experience to monitor the investigational study in accordance with
FDA regulations (21 CFR 812.43(d)).
Monitor will conduct a pre-investigational visit. Monitor will
ensure that the study protocol is thoroughly understood.
A Sponsor shall immediately conduct an evaluation of any
unanticipated adverse device affects (21 CFR 812.46(b)(1)) and
report the findings to the FDA.
A Sponsor who determines that an unanticipated adverse device
effect presents an unreasonable risk to subjects shall terminate
all investigations or parts of investigations presenting that risk
as soon as possible. Termination shall occur not later than 5
working days after the Sponsor makes this determination and not
later than 15 working days after the Sponsor first received notice
of the effect (21 CFR 812.46(b)(2)).
A Sponsor may not resume a terminated investigation without IRB
and FDA approval (21 CFR 812.46(c)).

XII. Institutional
Review Board (IRB) Information

The Sponsor should provide the following IRB information.

Identification of the IRB or IRBs.
Name, address and chairperson of each IRB.
Action taken by IRB,(i.e., approval).
Identification of how many IRBs have approved the
investigation.
Identification of how many IRBs are currently reviewing the
investigation or will review it in the future.

FDA recommends the Sponsor provide the following product
labeling information.

Early Feasibility & Transitional
Study

Sponsors should provide labeling for the investigational APDS,
including the instructions for operating each of its functional
components, as necessary. The purpose of product labeling during an
Early Feasibility or Transitional study is to ensure that operators
have adequate instructions for safely operating a device during the
study. Operators include clinical investigators (when studies or a
portion of a study are conducted in a clinical setting) or patients
(when a study or portion of the study takes place at home).

The amount of labeling necessary to ensure safe operation of the
system or functional components is dependent on the study design.
For example, some studies involve having a CGM inserted into
patients prior to them reporting to a clinic where the APDS will be
evaluated and they are not expected to operate it while they are at
home. In this example, it may not be necessary to provide the
patient with labeling for the CGM. It may only be necessary to
provide patients with instructions on what to do if they
experienced an adverse event involving the CGM, such as a reaction
at the insertion site. However, a patient might be expected to
operate the CGM, BGD, control algorithm, or pump, in which case
they should be provided with labeling which provides complete
instructions for performing each of the functions they are expected
to carry out.

Sponsors should identify each operator involved in the study,
and list each of the functions they are expected to carry out.

Sponsors should provide a copy of the draft labeling that
includes:

Adequate instructions that enable each operator to safely
perform all of the functions they are expected to carry out during
the study.
A caution statement, “Caution – Investigational
Device. Limited by Federal (or United States) law to
investigational use” on the APDS labeling.
Unless a justification can be provided, Sponsors should provide
patients who are operating the device with all labeling associated
with functions they are to perform with the device.
Sponsors should determine whether it is necessary to provide
clinical investigators with the labeling. If Sponsors believe that
this is not necessary they should:
- Describe what functions the investigator will be performing
  with the device and explain why it is not necessary to provide them
  with the labeling. (For example, the operator may have extensive
  experience operating the device.)
- Certify that labeling will be available at each clinical site
  should it be needed.

Pivotal Study (unsupervised outpatient
study)

The purpose of the product labeling should allow the subject to
safely operate the APDS. FDA recommends the Sponsor provide a
complete set of product labeling (
Section VIII of the guidance). In addition, the product
labeling should contain the following statement, “Caution
– Investigational Device. Limited by Federal (or United States) law
to investigational use”.

FDA recommends the instructions for use, such as user guides or
any written materials that will be provided to individuals during
this study should be the same as what will be provided with the
system when it is marketed.

XIV. Anticipated
Changes

The Sponsor should describe any changes that are anticipated
during the clinical study. For example, if the Sponsor intends to
modify their adjustable parameters during their study within the
predefined value range.

XV.
Manufacturing

Early Feasibility Study

This information is not generally needed for Early Feasibility
Studies that use devices that have already been approved or
cleared. The Sponsor should describe the devices used in the study
and provide the appropriate PMA and/or 510(k) number for completion
of this section.

Pivotal Study

The following information should be provided to support a
pivotal study design.

Certification that device will be manufactured in accordance
with Good Manufacturing Practices (21 CFR 812.20).
A description of the methods, facilities, and controls used for
the manufacture, processing, packing, and storage as required by 21
CFR 812.20(b)(3).
The QA program should be described. The Sponsor can
provide quantitative tests along with pass/fail criterion.
QA/QC tests monitor processing methods and can be used in lieu of
more detailed descriptions.
Procedures for specification control measures are established
to assure that the design basis for the device is correctly
translated into approved specifications (21 CFR 820.100(a)(1).
A description of the processes in accordance with 21 CFR
820.

Appendix B: Glossary

Analytical specificity – How well an assay
detects only a specific analyte (e.g., glucose) and does not detect
closely related substances.

Bias – The difference between the expectation
of test results and an accepted reference value. (CLSI EP21-A)

Blood Glucose Device (BGD) – A device which
measures blood glucose concentrations.

Continuous Glucose Monitor (CGM) – A sensor
placed under the patient’s skin (subcutaneously), which measures
the glucose in the fluid around the cells (interstitial fluid). A
small transmitter continually sends information to a receiver,
which converts the information to an estimate of blood glucose.

Control algorithm – A control algorithm is
software embedded in a computer that receives information from the
CGM and performs a series of mathematical calculations. Based on
these calculations, the controller sends instructions to alter the
insulin infusion of the pump.

Enriched population – For this guidance, an
enriched population is to study a patient population that is likely
to have a physiological phenomenon with an event frequency that is
sufficient to detect treatment-related differences in
occurrence.

Imprecision – An uncertainty of measurement
parameter, associated with the result of measurement, that
characterizes the dispersion of the values that could reasonably be
attributed to the measurand (the quantity intended to be measured).
It is expressed numerically as standard deviation (SD) or
coefficient of variation (CV). (POCT05)

Insulin infusion pump – A pump for delivering
insulin into the subcutaneous tissue to achieve glycemic control.
The pump is composed of a pump reservoir similar to that of an
insulin cartridge, a battery-operated pump, and a computer chip
that allows the user to control the amount of insulin being
delivered.

Interference – The act of hindering,
obstructing, or impeding the performance of a device.

In-silico model – a method to test the control
algorithm in a theoretical human model of insulin and glucose
metabolism using a sophisticated computer model rather than
expensive animal experiments.

Linearity – The ability (within a given range)
to provide results that are directly proportional to the
concentration (amount) of analyte in the test sample.(CLSI
EP6-A)

Measuring Range – The range of values (in units
appropriate for the analyte) over which the acceptability criteria
for the method have been met; that is where errors due to
nonlinearity, imprecision or other sources are within defined
limits. (CLSI EP6-A)

Pediatric – Of or relating to the medical care
of children. CDRH defines the pediatric age range from birth to 21
years of age.

¹ For purposes of this guidance,
Sponsor refers to any person who takes the responsibility
for and initiates a clinical investigation; Applicant
refers to any person who submits an application, amendment, or
supplement to obtain FDA approval of a new medical product or any
other person who owns an approved application. Sponsor is
used primarily in relation to investigational device exemption
(IDE) applications and Applicant is used primarily in
relation to premarket approval (PMA) submissions.

² 21 CFR 3.2(e): Combination product
includes: (1) a product comprised of two or more regulated
components, i.e., drug/device, biologic/device, drug/biologic, or
drug/device/biologic, that is physically, chemically, or otherwise
combined or mixed and produced as a single entity: (2) two or more
separate products packaged together in a single package or as a
unit and comprised of drug and device products, device and
biological products, or biological and drug products; (3) a drug,
device, or biological product packaged separately that according to
its investigational plan or proposed labeling is intended for use
only with an approved individually specified drug, device, or
biological product where both are required to achieve the intended
use, indication, or effect and where upon approval of the proposed
product the labeling of the approved product would need to be
changed, e.g., to reflect a change in intended use, dosage form,
strength, route of administration, or significant change in dose;
or (4) any investigational drug, device, or biological product
packaged separately that according to its proposed labeling is for
use only with another individually specified investigational drug,
device, or biological product where both are required to achieve
the intended use, indication, or effect.

³ Currently, to get the most accurate
readings possible from a CGM, the patient needs to periodically
calibrate the CGM using a blood glucose measurement from a BGD;
therefore, the BGD still plays a critical role in the proper
management of patients with an APDS. However, over time, we
anticipate that improved CGM performance may obviate the need for
periodic blood glucose checks with a BGD.

⁴ See
Guidance for the Content of Premarket Submissions for Software
Contained in Medical Devices for more information.

⁵ See 21 CFR 814.20(b)(3)(i).

⁶ See
Draft Guidance for Industry and FDA Staff – Processing/Reprocessing
Medical Devices in Health Care Settings: Validation Methods and
Labeling.

⁷ AAMI TIR 12:2010, Designing, testing
and labeling reusable medical devices for reprocessing in health
care facilities: A guide for medical device manufacturer

⁸ AAMI TIR 30:2003, A compendium of
processes, materials, test methods, and acceptance criteria for
cleaning reusable medical devices

⁹ AAMI / ANSI/ ISO 11607-1:2006,
Packaging for terminally sterilized devices – Part 1: Requirements
for materials, sterile barrier systems and packaging systems.

¹⁰ AAMI / ANSI/ ISO 11607-2:2006,
Packaging for terminally sterilized devices – Part 2: Validation
requirements for forming, sealing, and assembly processes.

¹¹ ASTM D4169-09, Standard Practice for
Performing Testing of Shipping Containers and Systems.

¹² See 21 CFR 809.10(b)(8).

¹³ FDA has published draft guidance on
the use of human factors in optimizing medical device design,

Draft Guidance for Industry and Food and Drug Administration Staff
– Applying Human Factors and Usability Engineering to Optimize
Medical Device Design. Although the recommendations
contained in this guidance are not in effect at this time, general
information contained in the guidance about human factors
considerations in medical device design is relevant to an
understanding of this topic.

¹⁴ CLSI POCT 05-A, Performance Metrics
for Continuous Interstitial Glucose Monitoring

¹⁵ CLSI EP5-A2 Protocol (Evaluation of
Precision Performance of Quantitative Measurement Methods; Approved
Guideline—Second Edition)

¹⁶ CLSI EP7-A2 Protocol (Interference
Testing in Clinical Chemistry; Approved Guideline- Second
Edition)

¹⁷ CLSI EP6-A document (Evaluation of
the Linearity of Quantitative Measurement Procedures, A Statistical
Approach; Approved Guideline, 2003)

¹⁸ CLSI EP9-A3 protocol (Method
Comparison and Bias Estimation Using Patient Samples; Approved
Guideline- Third Edition)

¹⁹ Bergenstal RM, et al. (2010)
Effectiveness of sensor-augmented insulin-pump therapy in Type 1
Diabetes, NEJM:363:311-320.

²⁰ Hermanides, J, et al. (2011)
Sensor-augmented pump therapy lowers HbA1c in suboptimally
controlled Type 1 Diabetes; a randomized controlled trial. Diabetic
Medicine (Accepted Article)

²¹ Junvenille Diabetes Research
Foundation Continuing Glucose Monitoring Study Group (2009) The
effect of continuous glucose monitoring in well-controlled Type 1
Diabetes. Diabetes Care 32:1378-1383

²² Garg, SK et al. (2007) Continuous
Home Monitoring of Glucose – Improved glycemic control with
real-life use of continuous glucose sensors in adult subjects with
Type 1 Diabetes. Diabetes Care 30:3023-3025

²³ The pediatric population is defined
as birth to 21 years of age. For details surrounding this
definition and recommended pediatric subpopulations, please refer
to
Guidance for Industry and Staff: Pediatric Expertise for Advisory
Panels. For the purposes of the LGS system, FDA recommends
the subpopulation of 18-21 be considered transitional adolescents
enabling this pediatric subpopulation to be studied with
adults.

²⁴ There is significant variability in
performance among HbA1c assays and point-of-care HbA1c test systems
may not be as accurate as assays performed in central laboratories.
Therefore, Sponsors/Applicants should minimize potential variables
in the study by having all study subjects’ HbA1c values
determined at one central laboratory location using only a National
Glycohemoglobin Standardization Program (NGSP) certified laboratory
method. The Sponsor/Applicant should provide the name of the HbA1c
test system that was used to obtain the HbA1c values and indicate
whether it is a NGSP certified method.

²⁵ American Diabetes Association
Workgroup on Hypoglycemia. Defining and Reproing Hypoglycemia in
Diabetes (2005), Diabetes Care 28: 1245-1249.

²⁶ November 10, 2010, Innovations in
Technology for the Treatment of Diabetes: Clinical Development of
the Artificial Pancrease (and Autonomous System).

²⁷ American Diabetes Association
Workgroup on Hypoglycemia. Defining and Reporting Hypoglycemia in
Diabetes (2005) Diabetes Care, 28:1245-1249.

²⁸ the study by having all study
subjects’ HbA1c values determined at one central laboratory
location using only a National Glycohemoglobin Standardization
Program (NGSP) certified laboratory method. The Sponsor/Applicant
should provide the name of the HbA1c test system that was used to
obtain the HbA1c values and indicate whether it is a NGSP certified
method.

CDRH Home Use Website

²⁹ See 21 CFR 814.20(c). Master
files are described on Device Advice. See link:
Premarket Approval (PMA) Master Files

³⁰
Guidance for the Content of Premarket Submissions for Software
Contained in Medical Devices.

³¹ Note that this guidance is in draft
form, but when final, this guidance will represent the Agency’s
thinking on this topic.

SOURCE

Preface

Table of Contents

The Content of Investigational Device Exemption (IDE) and Premarket Applications for Artificial Pancreas Device Systems

A. Overview

B. Currently Marketed Devices to Treat DM

C. Basic Design of an APDS

D. Different APDS Types

IV. Device Description

A. APDS System Level Description

B. Glucose Monitoring Functional Components

C. Control Algorithm & Signal Processing Functional Component

D. Infusion Pump Functional Component

E. Communication Pathway Functional Component

V. Indications for Use

A. Software

B. Report of Prior Investigations (If Applicable)

C. Biocompatibility

D. Sterility

E. Shelf Life

F. Electrical Safety

G. Magnetic Resonance (MR) Imaging Safety

H. Quality of Results from Diagnostic Devices Used During the Clinical Study

I. Human Factors

J. Glucose Monitoring Functional Component

1. CGM Component

2. BGD Component

K. Control Algorithm/Signal Processing Functional Component

L. Infusion Pump Functional Component

Drug Stability and Compatibility

Catheter Occlusion Bench Testing

Dose Accuracy

VII. Clinical Study Progression

A. Early Feasibility Study

B. Transitional Clinical Study

C. Pivotal Clinical Study

1. Patient Population

2. Study Endpoints

3. Statistical Analysis

Appendix A: IDE Content for APDS Studies

II. Device Description

III. Letters of Authorization

IV. Indication for Use

V. Nonclinical studies/Prior Investigations

A. Algorithm

1. Definition of Algorithm(s)

2. Definition of Algorithm Symbols/Parameters

3. Identification of Fixed Parameters

4. Identification of Adjustable Parameters that May be Modified During the Study

5. Parameter Sensitivity Analysis

6. Summary of the Verification Activities for the Control Algorithm

B. Software Documentation

C. Summary of System Communication

D. Safety Measures for Dosing

E. Biocompatibility Testing

F. Electrical Safety

Electromagnetic Compatibility

Applicable Standards

Radio Frequency

G. Animal/In-Silico Testing

Animal testing

In-Silico Testing

I. Human Studies

J. Human Factors/Usability Testing

Early Feasibility (Pilot) Study

Transitional Study

Pivotal Study/Unsupervised Outpatient Study

K. Catheter Occlusion Bench Testing

L. Dose Accuracy Bench Testing

M. Diagnostic Devices To Be Used During the Clinical Study

N. Drugs Used During the Study

VII. Clinical Study

A. Purpose/Objective(s)

B. Study Design

C. Sample Size and Investigational Sites

D. Study Duration

E. Inclusion Criteria

F. Exclusion Criteria

G. Study Timeline

H. Safety Monitoring/Risk Analysis

Draft Guidance for Industry and Food and Drug
Administration Staff

The Content of Investigational Device Exemption
(IDE) and Premarket Applications for Artificial Pancreas Device
Systems

IV. Device
Description

C. Control Algorithm & Signal Processing Functional
Component

V. Indications for
Use

H. Quality of Results from Diagnostic Devices Used During the
Clinical Study

K. Control Algorithm/Signal Processing Functional
Component

VII. Clinical Study
Progression

Appendix A: IDE Content
for APDS Studies

II. Device
Description

III. Letters of
Authorization

IV. Indication for
Use

V. Nonclinical
studies/Prior Investigations

4. Identification of Adjustable Parameters that May be
Modified During the Study

6. Summary of the Verification Activities for the Control
Algorithm

VII. Clinical
Study

VIII. Informed
Consent

IX. Patient Case Report
Form(s)

X. Investigator
Agreement Forms

XI. Monitoring
Information

XII. Institutional
Review Board (IRB) Information

XIV. Anticipated
Changes

XV.
Manufacturing