EVERSENSE CONTINUOUS GLUCOSE MONITORING SYSTEM

P160048 · Senseonics, Incorporated · QCD · Jun 21, 2018 · Clinical Chemistry

Device Facts

• Record ID: P160048
• Device Name: EVERSENSE CONTINUOUS GLUCOSE MONITORING SYSTEM
• Applicant: Senseonics, Incorporated
• Product Code: QCD · Clinical Chemistry
• Decision Date: Jun 21, 2018
• Decision: APPR
• Device Class: Class 3
• Attributes: AI/ML

AI Performance

Output	Acceptance	Observed	Dev DS	Test DS
Continuous Glucose Monitoring (CGM) Accuracy	Not specified	Overall (40-400 mg/dL): 94.3% within 20/20% of comparator (PRECISE II); 92.8% within 20/20% of comparator (PRECISION)	PRECISE study (71-subject European pivotal study) used as a training set for the SW 602 algorithm.	PRECISE II (90 subjects, 15,753 matched pairs) and PRECISION (35 subjects, 15,170 matched pairs) clinical studies.
Hypoglycemic Event Detection (Threshold Alert)	Not specified	Confirmed event detection rate at 70 mg/dL: 87% (PRECISE II); 95% (PRECISION)	Not applicable	PRECISE II (90 subjects) and PRECISION (35 subjects) clinical studies.
Hyperglycemic Event Detection (Threshold Alert)	Not specified	Confirmed event detection rate at 180 mg/dL: 96% (PRECISE II); 98% (PRECISION)	Not applicable	PRECISE II (90 subjects) and PRECISION (35 subjects) clinical studies.

Intended Use

The Eversense CGM System is indicated for continually measuring glucose levels in adults (age 18 and older) with diabetes for up to 90 days. The system is intended to: - Provide real-time glucose readings. - Provide glucose trend information. - Provide alerts for the detection and prediction of episodes of low blood glucose (hypoglycemia) and high blood glucose (hyperglycemia). The system is a prescription device. Historical data from the system can be interpreted to aid in providing therapy adjustments. These adjustments should be based on patterns seen over time. The system is indicated for use as an adjunctive device to complement, not replace, information obtained from standard home blood glucose monitoring devices.

Device Story

Eversense CGM System provides continuous glucose monitoring (CGM) for up to 90 days. System components: subcutaneous fluorescence-based glucose sensor, external smart transmitter, and mobile medical application (MMA). Physician implants sensor in upper arm using provided insertion tools (blunt dissector/insertion tool). Sensor contains dexamethasone acetate-eluting silicone collar to reduce inflammation. Transmitter powers sensor via RF, calculates glucose from fluorescence readings, and transmits data via Bluetooth Low Energy (BLE) to MMA. Transmitter provides vibratory alerts. MMA displays real-time glucose, trends, and alerts; allows calibration via fingerstick blood glucose inputs. Used in home/clinic settings; operated by patient. Healthcare providers use historical data for therapy adjustments. Benefits include long-term wear, reduced insertion frequency, and improved glucose management via trend/predictive alerts. Risks include insertion/removal complications (e.g., infection, sensor breakage), potential for inaccurate readings requiring fingerstick confirmation, and MRI incompatibility.

Clinical Evidence

Pivotal PRECISE II (n=90) and PRECISION (n=35) studies evaluated 90-day accuracy and safety. Primary endpoints: CGM-comparator agreement (YSI 2300 analyzer) and adverse event incidence. Results (SW-602 algorithm): Overall MARD 8.5-9.6%. 94.3% of readings within 20/20% of comparator in PRECISE II. Confirmed hypoglycemic event detection rates 78-98% depending on threshold. Safety: 0% infection rate; rare sensor breakage/removal difficulty. Biocompatibility and human factors studies supported device safety.

Technological Characteristics

Fluorescence-based glucose sensor; subcutaneous implantation. Sensor contains 1.75 mg dexamethasone acetate. RF-powered sensor; BLE connectivity to mobile device. Transmitter rechargeable battery. Materials: silicone collar, biocompatible components (ISO 10993). Sterilization: Ethylene Oxide (EO) to SAL 10^-6. Software: mobile app (Android/iOS) and transmitter firmware. Glucose determination algorithm (SW-602) converts fluorescence to glucose values.

Indications for Use

Indicated for adults (age 18+) with diabetes for continuous glucose monitoring for up to 90 days. Contraindicated for patients requiring MRI, those with dexamethasone/dexamethasone acetate contraindications, or those receiving intravenous mannitol/sorbitol.

Regulatory Classification

Identification

The device is a fully implanted continuous glucose monitoring device intended to detect trends and track patterns in interstitial glucose values. The device is indicated for use as an adjunctive device to complement, not replace, information obtained from standard home blood glucose monitoring devices.

Submission Summary (Full Text)

{0}

SUMMARY OF SAFETY AND EFFECTIVENESS DATA (SSED)

I. GENERAL INFORMATION

Device Generic Name: Continuous glucose monitor (CGM), implanted, adjunctive use
Device Trade Name: Eversense Continuous Glucose Monitoring System
Device Procode: QCD
Applicant’s Name and Address: Senseonics, Incorporated
20451 Seneca Meadows Pkwy
Germantown, MD 20876
Date of Panel Recommendation: March 29, 2018
Premarket Approval Application (PMA) Number: P160048
Date of FDA Notice of Approval: June 21, 2018

II. INDICATIONS FOR USE

The Eversense CGM System is indicated for continually measuring glucose levels in adults (age 18 and older) with diabetes for up to 90 days.

The system is intended to:
- Provide real-time glucose readings.
- Provide glucose trend information.
- Provide alerts for the detection and prediction of episodes of low blood glucose (hypoglycemia) and high blood glucose (hyperglycemia).

The system is a prescription device. Historical data from the system can be interpreted to aid in providing therapy adjustments. These adjustments should be based on patterns seen over time.

The system is indicated for use as an adjunctive device to complement, not replace, information obtained from standard home blood glucose monitoring devices.

III. CONTRAINDICATIONS

The following contraindications are included in the labeling
- The Sensor and Smart Transmitter are incompatible with magnetic resonance imaging (MRI) procedures. DO NOT undergo an MRI procedure while the sensor is inserted

PMA P160048: FDA Summary of Safety and Effectiveness Data

{1}

or when wearing the smart transmitter. Should an MRI be required, please contact your physician to arrange for sensor removal before the procedure.

- The system is contraindicated in people for whom dexamethasone or dexamethasone acetate may be contraindicated.
- Mannitol or sorbitol, when administered intravenously, or as a component of an irrigation solution or peritoneal dialysis solution, may increase blood mannitol or sorbitol concentrations and cause falsely elevated readings of your sensor glucose results. Sorbitol is used in some artificial sweeteners, and concentration levels from typical dietary intake do not impact sensor glucose results.

## IV. WARNINGS AND PRECAUTIONS

The warnings and precautions can be found in the Eversense Continuous Glucose Monitoring System labeling.

## V. DEVICE DESCRIPTION

The Eversense Continuous Glucose Monitoring System (Eversense System, or System) provides continuous glucose measurements over a 40-400 mg/dL range. The system provides real-time glucose values, glucose trends, and alerts for high and low glucose through a mobile application installed on a compatible mobile device platform (e.g., Android or iOS device). The Eversense System consists of a fluorescence-based glucose sensor (Eversense Sensor) that is inserted under the skin by a physician with Insertion Tools; an externally worn Eversense Smart Transmitter (Transmitter); and the Eversense Mobile Medical Application (MMA), which runs on a compatible mobile device. The inserted Sensor is a radiofrequency (RF) powered device that collects readings and sends them to the Transmitter. The Transmitter calculates, stores, and transmits the glucose data via Bluetooth Low Energy (BLE) to the MMA on the mobile device.

The System consists of four principal components:

1. Sensor: The sensor uses a fluorescence sensing mechanism to detect glucose in the interstitial fluid (ISF). The sensor is inserted subcutaneously by a physician, and receives RF-power from the Transmitter to measure interstitial fluid glucose every 5 minutes. The sensor sends fluorescence measurements to the Transmitter for calculation and storage of glucose values. The sensor has a silicone collar component that contains 1.75 mg of an anti-inflammatory steroid drug (dexamethasone acetate) that elutes locally to reduce tissue inflammation around the sensor. The sensor operating life is the lesser of 90 days or until the device's end-of-life is reached. The sensor is provided sterile to the physician, for single use in a sensor holder. The Sensor is inserted by a qualified physician using the provided insertion tools.
2. Transmitter: The transmitter, worn externally over the inserted Sensor, is a device that powers the Sensor, calculates the glucose values from the Sensor-measured

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 2

{2}

fluorescence readings, and using secure BLE wirelessly sends the glucose information to the MMA for display on the handheld device (HHD). An adhesive patch holds the transmitter in place. The transmitter contains a rechargeable battery which is charged with a charging cradle powered by a USB connection. The transmitter also provides vibration signals for alerts and notifications, such as low glucose levels, irrespective of whether the MMA is in the vicinity or not.

3. MMA: The MMA is a software application that runs on a compatible mobile device for display of glucose information provided by the transmitter. The MMA receives and displays the calculated glucose information from the transmitter, including glucose trend information and glucose alerts. The MMA also allows the user to calibrate the CGM System by input of blood glucose measurements. It also communicates with the Senseonics server for a one-time download of calibration parameters specific for each Sensor. The MMA also provides the user an option to upload the data to Senseonics Data Management System (DMS) for historic viewing and storing of glucose data.

4. Insertion Tools: Insertion Tools (a Blunt Dissector and Insertion Tool) are provided to the physician for Sensor implantation. The Blunt Dissector is used to create the subcutaneous space in which the Sensor is placed. The Sensor Holder in which the Sensor is stored during transport and sterilization is used to transfer the Sensor to Insertion Tool. The Insertion Tool is used to place the Sensor into the subcutaneous space.

VI. ALTERNATIVE PRACTICES AND PROCEDURES

There are a number of alternative practices used for managing diabetes, and often more than one practice is recommended by health care providers. This includes oral and/or injectable medications, as well as self-monitoring of blood glucose using home blood glucose monitoring devices. Self-monitoring blood glucose meters and test strips provide a blood glucose measurement at a single point in time, whereas CGM provides continuous glucose measurements. Additionally, behavior changes related to physical activity and healthy eating can aid in successful diabetes management.

Each alternative has its own advantages and disadvantages. Patients should thoroughly discuss the alternatives with their physician to choose the method that best suits individual expectations and lifestyles.

VII. MARKETING HISTORY

The Eversense CGM System has not been marketed in the United States.

A different version of the Eversense CGM System has been approved for commercial distribution in the European Union and European Economic Area countries requiring CE Mark since May 2016.

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 3

{3}

The system has not been withdrawn from commercial distribution for any reason related to safety or effectiveness.

## VIII. POTENTIAL ADVERSE EFFECTS OF THE DEVICE ON HEALTH

Below is a list of the potential adverse effects (e.g., complications) associated with use of the device.

Potential adverse effects related to insertion, removal and wear of the sensor include:

- Allergic reaction to adhesives
- Bleeding
- Bruising
- Infection
- Pain or discomfort
- Scarring or skin discoloration
- Sensor fracture during removal
- Skin inflammation, thinning, discoloration or redness

There are risks relating to difficulty with sensor removal, and potential risks associated with subsequent procedures required for sensor removal. Five instances of difficulty with sensor removal, one of which was reported as a serious adverse event, where subjects were referred to a general surgeon for successful sensor removal, were documented in the clinical studies reviewed. Based on postmarket data available with a different version of this device marketed in Europe, and the results observed in these clinical studies, the occurrence of these events is low.

There is a risk of sensor breakage leaving a sensor fragment under the skin. Two instances of sensor breakage were documented in the clinical studies reviewed. Based on postmarket data available with a different version of this device marketed in Europe, and the results observed in these clinical studies, the occurrence and severity of these events is low.

There may be potential risks relating to repeated insertion and removal procedures, including buildup of scar tissue over time at the sensor insertion site, in a small range of locations on the outside surface of the upper arms. Based on postmarket data available with a different version of this device marketed in Europe, and the results observed in these clinical studies, these risks are not expected to occur.

The Eversense CGM System has a drug component, consisting of 1.75 mg of dexamethasone acetate (DXA), contained in a dexamethasone eluting silicone collar to the outside of the Eversense Sensor. Based on information and clinical evaluations performed, the sponsor has demonstrated that risks relating to both local and potential systemic exposure the dexamethasone component of the device, as well as repeated exposure to the dexamethasone component of the device, are not expected to occur. These risks appear to be remote based on the results observed in these clinical studies,

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 4

{4}

although these clinical studies did not include subjects taking dexamethasone (or other glucocorticoid medications).

A minor risk of this device is that users may need to perform unnecessary fingersticks to evaluate their blood glucose when the CGM gives false positive hypoglycemic and hyperglycemic readings or alerts. Inaccurate calculation of the rate of change of glucose could also lead to unnecessary additional blood glucose tests or inappropriate measures to stop a trend of increasing or decreasing glucose level which could result in hyperglycemia or hypoglycemia. There is a minor risk of skin irritation, inflammation, or infection due to either the sensor needle or the adhesive.

There are risks due to missed alerts and false negative hypoglycemia and hyperglycemic readings related to patients not being alerted to the need to perform a fingerstick to detect hypoglycemia or hyperglycemia, particularly since users of this device may rely on these alerts in certain situations to guide their self-treatment strategy (e.g., to alert them to potential nighttime hypoglycemia). There is a risk to false alerts and false positive hypoglycemia and hyperglycemia readings related to the need to perform unnecessary fingersticks to confirm an erroneous low or high reading. Inaccurate calculation of the rate of change of glucose by the CGM could prevent a patient from performing additional blood glucose tests or taking measures to stop a trend of increasing or decreasing glucose levels which could lead to serious hypoglycemia or hyperglycemia if no action is taken to stop these glucose trends. Inaccurate calculation of the rate of change of glucose could also lead to unnecessary additional blood glucose tests or inappropriate measures to stop a trend of increasing or decreasing glucose level which could result in hyperglycemia or hypoglycemia.

There is a risk if patients make decisions on diabetes management based on inaccurate sensor readings alone without confirmation by blood glucose testing. The device labeling states this device is intended to be used to complement, not replace, blood glucose testing.

The body-worn transmitter component of the system provides an alternate means of delivering alerts to users through vibratory feedback. The level of information necessary to understand the safety aspects of the user interface, and how it supports the user and reduces the potential for use error was provided by the sponsor, and found to be adequate. There may be an additional risk that the display, or alerts or alarms related to the CGM device may not be able to override other applications or functions (phone, camera, SMS) within the mobile device. This risk could potentially result in missed alerts or alarms, or temporary loss of access to the display. Missed alerts, alarms, or inability to access the display could result in missed opportunities to detect or prevent hypoglycemia or hyperglycemia, and are discussed above. Human factors studies conducted assessed the safety of the user interface of the mobile app (sole display) for this device, and the ability for users to be receive and understand alerts and notifications via the transmitter vibration feature. The human factors study sufficiently assessed the potential for user error associated with comprehension of the impact of mobile device and app settings on notifications and Bluetooth communications, as well as use of the audio override feature.

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 5

{5}

IX. SUMMARY OF PRECLINICAL STUDIES

A. Laboratory Studies

Pre-clinical testing has been conducted to demonstrate the Eversense CGM System performs as intended and meets its product requirements (see Table 1). The verification and validation tests included compliance with international standards and/or guidance documents where available. The CGM System and its components have various levels of specifications and technological characteristics. Therefore, a combination of full system testing, subsystem and component level testing was performed to demonstrate that the device meets its requirements and is safe for use.

**Device and Electrical Safety:** The Transmitter has undergone testing to demonstrate that the device meets the requirements for medical device safety, including electrical safety, according to the following international standards: IEC 60601-1, 3rd Edition, Medical electrical equipment – General requirements for basic safety and essential performance.

**Electromagnetic Compatibility:** The Transmitter has undergone testing to demonstrate the device meets the following international standard: IEC 60601-1-2, 4th Edition, Medical electrical equipment – Part 1-2, General requirements for basic safety and essential performance – Collateral Standard: Electromagnetic disturbances – Requirements and tests.

**Home Health Care Products:** The Transmitter has undergone testing to demonstrate that the device meets the requirements for medical device safety for home health care products, according to the following international standards: IEC 60601-1-11, 2nd Edition. Medical electrical equipment – General requirements for basic safety and essential performance – Collateral Standard: Requirements for medical electrical equipment and medical electrical systems used in the home healthcare environment.

**Battery Standards:** The Transmitter batteries have undergone testing to demonstrate that the batteries meet the requirements for safety for batteries containing alkaline or other non-acid electrolytes, according to the following international standards: IEC 62133, 2nd Edition. Secondary cells and batteries containing alkaline or non-acid electrolytes – Safety requirements for portable sealed secondary cells, and for batteries made from them for use in portable applications.

**Electrical Testing for Batteries and Bluetooth Function:** Transmitters were subjected to the electrical verification testing summarized in Table 1. Protocols, test reports and acceptance criteria were reviewed and found to be acceptable. The device met the pre-determined acceptance criteria for battery recharge, and communication longevity.

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 6

{6}

Table 1: Summary of Preclinical Testing of the Eversense Smart Transmitter

| Test Name/Description | Test Purpose | Acceptance Criteria |
| --- | --- | --- |
| Device Safety and Electrical Testing | To verify compliance with IEC 60601-1, 3rd Edition | Complies with standard |
| EMC Testing | To verify compliance with IEC 60601-1-2, 4th Edition | Complies with standard |
| Home Health Care Products | To verify compliance with IEC 60601-1-11, 2nd Edition | Complies with standard |
| Battery Standards | To verify compliance with IEC 62133, 2nd Edition | Complies with standard |
| Power 1 – Initial Charge | To verify length of time to fully charge dormant Transmitter | Battery should be fully charged in less than 120 minutes |
| Power 2 - Transmitter Battery Recharge Time | To verify whether the charger can recharge the battery within the specified time | Battery in the fully empty condition should be fully recharged in less than 20 minutes |
| Power 3 – Low Battery Indication | To verify the Transmitter lasts for at least 4 hours after low battery indication | Battery shall last at least 4 hours after 10% battery remaining indication before entering dormant mode |
| Cycled Battery Charge Time | To verify the battery life after 100 charge/discharge cycles | Battery when fully charged should last a minimum of 36 hours after 100 charge/discharge cycles |
| | To verify the battery life after 400 charge/discharge cycles | Battery when fully charged should last a minimum of 8 hours after 400 charge/discharge cycles |
| Bluetooth Range | To verify whether the Transmitter provides reliable communication via Bluetooth within the specified range, and re-establishes communication after moving to and from maximum specified range | Transmitter should communicate with hand-held device within a maximum of 10 meters (32.8 feet) |
| Antenna 1 | To verify peak frequency | 13.56 Mhz ± 7 Khz |
| Antenna 2 - NFC Read performance at 12mm | To verify the Transmitter can communicate with the Sensor from the specified distance | Transmitter shall be able to communicate with the Sensor from the 12 mm maximum distance |
| Charging Cradle Reliability | To verify charging cradle function following 1200 cycles of inserting and detaching the Transmitter to/from the charging cradle | After 1200 cycles, the charging cradle charges the Transmitter, and the Transmitter remains connected to the charging |

PMA P160048: FDA Summary of Safety and Effectiveness Data

{7}

| Test Name/Description | Test Purpose | Acceptance Criteria |
| --- | --- | --- |
| | | cradle |
| Button Reliability | To verify Transmitter button function after actuation (Phoenix Transmitter System High Level Functional Test Procedure) | Verify that after 3000 button presses that the Transmitter's button does not have significant physical damage or wear, and is able to pass all steps of the High Level Functional Test Procedure that involve system responses to button presses |
| Adhesive Patch Operational Test | To verify adhesive patch function following submersion in water for 30 minutes (Phoenix Transmitter System High Level Functional Test Procedure) | Verify that the adhesive patch passes the functionality test |

Transmitter Environmental Exposure and Mechanical Testing: Transmitters were subjected to the following functional and environmental tests. Protocols, test reports and acceptance criteria were reviewed and found to be acceptable. The device met the pre-determined acceptance criteria, as described in Table 2 below.

Table 2: Mechanical Testing of the Eversense Smart Transmitter

| Test Name/Description | Test Purpose | Acceptance Criteria |
| --- | --- | --- |
| Shipping | To verify devices as packaged can meet functional requirements after simulated shipping conditions, including conditioning based upon ISTA 3A and Shipping Simulation testing according to ASTM D4169-16 Cycle 13, Assurance Level I | Devices must pass visual inspection and Phoenix Transmitter System High Level Function Test Procedure |
| Thermal Shock | To verify devices function following thermal shock | Devices must pass Phoenix Transmitter System High Level Function Test Procedure |
| Storage Conditions | To verify devices function following storage at low and high temperatures (0 and 35°C) | Devices must pass visual inspection and Phoenix Transmitter System High Level Function Test Procedure |
| Operating Conditions Test – Temperature | To verify devices function following exposure to extreme temperatures and humidity (5 to 40°C and relative | Devices must pass Phoenix Transmitter System High Level Function Test Procedure |

PMA P160048: FDA Summary of Safety and Effectiveness Data

{8}

| Test Name/Description | Test Purpose | Acceptance Criteria |
| --- | --- | --- |
| and Humidity | humidity 15 to 90%) | |
| Mechanical Shock | To verify devices function following mechanical shock conditions as specified in IEC 60601-1-11 | Devices must pass visual inspection and Phoenix Transmitter System High Level Function Test Procedure |
| Vibration | To verify devices function following vibration conditions | Devices must pass visual inspection and Phoenix Transmitter System High Level Function Test Procedure |
| Drop | To verify devices function as intended following repeated drops from a height of 1 meter unto a hardwood board | Devices must pass visual inspection and Phoenix Transmitter System High Level Function Test Procedure |
| Push | To verify devices function following application of a steady force of 250 N ± 10 N (56.2 lb ± 2.2 lb) for a period of 5 seconds, using a test tool which provides contact over a circular plane surface 30mm | Devices must pass visual inspection and Phoenix Transmitter System High Level Function Test Procedure |
| Operational Life Test | To verify devices ability to function over a 1 year life | Devices must pass functional requirements |
| Water Ingress Test | To evaluate transmitter compliance with IP67 rating and charging cradle compliance with IP22 rating of IEC 60529 | Transmitter must demonstrate no water ingress. Transmitter and charging cradle must pass a comprehensive functional test procedure following exposure to the water ingress stress conditions |

Insertion Tools Environmental Exposure and Mechanical Testing: Insertion Tools (Insertion Tool and Blunt Dissector) were subjected to the following functional and environmental tests described in Table 3. Protocols, test reports and acceptance criteria were reviewed and found to be acceptable. The device met the pre-determined acceptance criteria.

PMA P160048: FDA Summary of Safety and Effectiveness Data

{9}

Table 3: Mechanical and Environmental Testing of the Eversense Insertion Tools

| Test Name/ Description | Test Purpose | Acceptance Criteria |
| --- | --- | --- |
| Actuation Mechanism Test | To evaluate the mechanism of actuation of the insertion tool by locking and unlocking | Verification of lock and unlocked positions |
| Push and Pull Test | To evaluate the mechanical strength of the cannula of the insertion tool and metal portion of the blunt dissector following compression and tension | Withstand minimum push or pull force of 44.5 N |
| Actuation Force Test | To evaluate the force needed for the actuation mechanism of the insertion tool | Actuate with less than 2.2 lbf |
| Marking Durability | To evaluate the markings on the tool remains visible | Marks remain visible and do not degrade |
| Shipping and Handling Extremes | To evaluate whether the devices within their packaging can withstand exposure to extreme temperatures and humidity | Verification of package integrity and device function |

Sensor Environmental Exposure and Electrical Testing: Sensor verification testing was performed to evaluate the Sensor electronics and glucose indicator to verify the design meets the essential performance described in Table 4. Sensors were subjected to testing to evaluate label marking durability through shipping tests, dimensional, and maintaining electrical essential performance. Protocols, test reports and acceptance criteria were reviewed and found to be acceptable. The device met the pre-determined acceptance criteria.

Table 4: Environmental and Electrical Testing of the Eversense Sensor

| Test Name/ Description | Test Purpose | Acceptance Criteria |
| --- | --- | --- |
| Sensor Electro-optical Interface Circuit Testing | To evaluate functionality of the near field communication and electro-optical circuitry | Sensor electronic can communicate via the ISO 15693 protocol, and are able to excite the fluorescent glucose indicator and detect its emitted fluorescent light according to Specification limits |
| Sensor Glucose Indicator Test | To evaluate the glucose responsivity of the fluorescent glucose indicator | Sensor must meet specification limit for fluorescent signal strength and sensitivity to glucose levels |
| Marking Durability | To evaluate that the Sensor package marking is protected against the effects of temperature and humidity. | The marking on the sensor packaging shall not visibly deteriorate upon humidity exposure. |
| Shipping and | To evaluate whether the devices | Following the shipping exposure, |

PMA P160048: FDA Summary of Safety and Effectiveness Data

{10}

| Test Name/Description | Test Purpose | Acceptance Criteria |
| --- | --- | --- |
| Handling Extremes | within their packaging can withstand exposure to extreme temperatures and humidity | the samples shall meet the essential performance requirement |

Biocompatibility Testing: Biocompatibility studies were selected and performed in consultation with international recognized safety standards (ISO 10993-1, Biological Evaluation of Medical Devices - Part 1: Evaluation and Testing) and in accordance with the FDA guidance document entitled "Use of International standard ISO 10993-1, Biological Evaluation of Medical Devices Part 1: Evaluation and testing within a risk management process" dated June 16, 2016. All studies cited in this section were conducted in compliance with 21 CFR Part 58 - Good Laboratory Practice for Nonclinical Laboratory Studies (GLPs). All studies had passing results. Results of the biocompatibility studies are summarized in Table 5, Table 6, and Table 7.

Table 5: Summary of the Biocompatibility Tests and Results for the Eversense Sensor

| Biocompatibility Test | ISO Standard | Test Method | Results |
| --- | --- | --- | --- |
| Cytotoxicity | ISO 10993-5 | MEM Elution | Pass – Not cytotoxic |
| Sensitization | ISO 10993-10 | Maximization Sensitization | Pass - Not Sensitizing |
| Irritation | ISO 10993-10 | Intracutaneous Reactivity | Pass – Nonirritant |
| Systemic Toxicity | ISO 10993-11 | Acute Systemic Toxicity | Pass - Not toxic |
| Systemic Toxicity | ISO 10993-11 | Material Mediated Pyrogen | Pass – Not pyrogenic |
| Subchronic Toxicity and Implantation | ISO 10993-6 | 4 and 13 Week Systemic Toxicity in Rats-Subcutaneous Implant | Pass - Not systemically toxic |
| Chronic Toxicity and Implantation | ISO 10993-6 | 26 Week Systemic Toxicity in Rats-Subcutaneous Implant | Pass - Not systemically toxic |
| Genotoxicity/Carcinogenicity | ISO 10993-3 | Bacterial Reverse Mutation | Pass - Non-mutagenic |
| Genotoxicity/Carcinogenicity | ISO 10993-3 | Mouse Lymphoma | Pass - Non-mutagenic |
| Genotoxicity/Carcinogenicity | ISO 10993-3 | Peripheral Blood Micronucleus Test | Pass - No damage to chromosomes |
| Chemical Characterization | ISO 10993-17 ISO 10993-18 | Exhaustive Extraction | Pass - no leachables/extractables from the Sensor are |

PMA P160048: FDA Summary of Safety and Effectiveness Data

{11}

Table 6: Summary of the Biocompatibility Tests and Results for the Eversense Transmitter and Adhesive Patch

| Biocompatibility Test | ISO Standard | Test Method | Results |
| --- | --- | --- | --- |
| Cytotoxicity | ISO 10993-5 | Transmitter: MEM Elution
Adhesive Patch: Agarose Overlay Method | Pass – Not cytotoxic |
| Sensitization | ISO 10993-10 | Transmitter: Maximization Sensitization
Adhesive Patch: Maximization Sensitization | Pass - Not Sensitizing |
| Irritation | ISO 10993-10 | Transmitter: Primary Skin Irritation
Adhesive Patch: Primary Skin Irritation | Pass – Nonirritant |

PMA P160048: FDA Summary of Safety and Effectiveness Data

{12}

Table 7: Summary of the Biocompatibility Tests and Results for the Eversense Insertion Tools

| Biocompatibility Test | ISO Standard | Test Method | Results |
| --- | --- | --- | --- |
| Cytotoxicity | ISO 10993-5 | MEM Elution | Pass – Not cytotoxic |
| Sensitization | ISO 10993-10 | Maximization Sensitization | Pass - Not Sensitizing |
| Irritation | ISO 10993-10 | Intracutaneous Reactivity | Pass – Nonirritant |
| Systemic Toxicity | ISO 10993-11 | Acute Systemic Toxicity | Pass - Not toxic |
| Systemic Toxicity | ISO 10993-11 | Material Mediated Pyrogen | Pass - Non-pyrogenic |

Interference Testing: Interference in the Eversense CGM system was assessed using in vitro testing. During in vitro testing, sensors were placed into glucose solutions to which potentially interfering substances were then added. The sponsor based the selection of concentrations of potential interferents on recommendations from interference testing standards/guidelines (e.g. Clinical & Laboratory Standards Institute (CLSI) EP7A2), FDA guidance documents for other glucose measurement devices (e.g. "Self-Monitoring blood glucose test systems for over-the-counter use" issued October, 2016), or based on information available in literature. In some cases, information on ISF concentration of potential interferants was not available. In these situations, plasma concentrations were used to assess interference; this approach represents a worst-case scenario, as ISF concentrations are unlikely to be higher than plasma concentrations. Most tested substances occur in ISF due to diffusion of the substance into ISF from the bloodstream.

Substances were tested at 2 glucose concentrations – a low concentration of 72 mg/dL and a high concentration of 324 mg/dL. The glucose level measured by the sensors was recorded before and after the addition of the potential interferant, and the degree of bias was calculated.

Senseonics defined significant interference as a bias greater than 10 mg/dL for glucose levels below 100 mg/dL, or greater than 10% for glucose levels above 100 mg/dL.

Based on the results of this testing, the following statements are included in the product labeling:

- Mannitol or sorbitol, when administered intravenously, or as a component of an irrigation solution or peritoneal dialysis solution, may increase blood mannitol or sorbitol concentrations and cause falsely elevated readings of your sensor glucose results. Sorbitol is used in some artificial sweeteners, and concentration levels

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 13

{13}

from typical dietary intake do not impact sensor glucose results.

- Antibiotics of the tetracycline class may falsely lower sensor glucose readings. You should not rely on sensor glucose readings while taking tetracyclines.

Software: The applicant performed software verification and validation testing in accordance with the FDA guidance document entitled "Guidance for the Contents of Premarket Submissions for Software Contained in Medical Devices," dated May 11, 2005. Verification and validation testing included units test, system level verification tests (which included functional testing to demonstrate the device meet its requirements), code review, traceability linking and validation testing to ensure the software conforms to user needs and intended uses.

Human Factors/Usability: Human factors Validation testing was conducted per the FDA guidance entitled "Applying Human Factors and Usability Engineering to Medical Devices" dated February 3, 2016. The Human Factors Validation testing considered the intended users, uses and use environments in the design of the simulated use testing.

Human factors studies were conducted to evaluate patient users as well as physician users who perform the sensor insertions.

Patient user human factors studies evaluated adult users with a variety of smartphone experience and an approximately equal distribution of iPhone and Android users. The initial human factors study evaluated usability tasks such as performing initial setup and pairing, setting alerts, calibrating sensor, and responding to alerts. A supplemental human factors study evaluated critical use-related tasks for the Eversense mobile app such as notification setup and responding to alert notifications including low and high glucose alerts, transmitter disconnect alert and battery low alert. In one scenario, study participants were asked to use their mobile device for a distracting task with the Eversense app in the background. While the app was in the background, study staff triggered low battery alerts for the system. Two study participants using the Android version of the app said that they noticed the transmitter vibration and correctly understood that this indicated an alert, but they did not want to stop what they were doing to check the Eversense app immediately. This failure was not observed for users of the iOS app, as the iOS app had implemented banner type notifications for instances when the Eversense app was in the background. In response to this observation, the applicant implemented banner type notifications for the Android version of the app.

A physician human factors study evaluated physician's ability to successfully perform sensor insertion procedures after receiving training. Sensor insertions were performed using simulated tissue products designed to mimic real arm tissue where insertions would normally be done. Each participant completed a sensor insertion scenario encompassing the full procedure from patient and equipment preparation to wound closure. All 16 physicians successfully used the blunt dissector tool to create the

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 14

{14}

subcutaneous pocket for the sensor. One physician participant (out of 16) failed to successfully insert the sensor into the subcutaneous pocket. During debriefing, it was determined that the physician did not successfully load the sensor into the insertion tool. During real use scenarios, for their first insertions physicians are monitored by Senseonics training staff. Training staff were not included in this roll during the human factors assessment. As this type of use error has not been observed in any clinical studies to date, the risk appears to be well mitigated in practice.

The human factors validation evaluation and testing demonstrates that the device can be used by the intended users without serious use errors or problems, for the intended uses and under the expected use conditions.

**Sterility:** The Sensor with its holder is a provided sterile for single-use and is sterilized using ethylene oxide (EO). The sterilization process was validated in accordance with ISO 11135-1, Sterilization of Health Care Products – Ethylene oxide – Part 1: Requirements for development, validation and routine control of a sterilization process for medical devices and in consideration of ISO 11135-2, Sterilization of Health Care Products – Ethylene oxide – Part 2: Guidance on the application of ISO 11135-1. The device is sterilized to a sterility assurance level (SAL) of 10⁻⁶. EO and ethylene chlorohydrin (ECH) residuals are monitored and meet the limits specified in ISO 10993-7, Biological evaluation of medical devices – Part 7: Ethylene oxide sterilization residuals. The Sensor is provided pyrogen free.

The Insertion Tools are provided sterile for single-use, and are sterilized using EO. The sterilization process was validated in accordance with ISO 11135-1, Sterilization of Health Care Products – Ethylene oxide – Part 1: Requirements for development, validation and routine control of a sterilization process for medical devices, and in consideration of ISO 11135-2, Sterilization of Health Care Products – Ethylene oxide – Part 2: Guidance on the application of ISO 11135-1. The device is sterilized to a SAL of 10⁻⁶. EO and EC residuals are monitored and meet the limits specified in ISO 10993-7, Biological evaluation of medical devices – Part 7: Ethylene oxide sterilization residuals.

**Shelf Life and Packaging:** The Sensor with the sensor holder is provided sterile for single use with recommended storage between 2°C and 8°C (36°F and 46°F) and a labeled expiration date set at 1 month. Shelf life studies of the Sensor are ongoing under an approved protocol and the shelf life will be updated upon successful completion of each subsequent test time point. The Insertion Tools are provided in a single package, sterile for single use with recommended storage at room temperature and a labeled shelf life of 6 months.

**B. Animal Studies**

A separate animal study was conducted to compare the biocompatibility of the Sensor with steroid eluting collar to a steroid eluting pacing lead, (an approved medical device) that elutes the same drug (dexamethasone acetate) from a silicone carrier. The

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 15

{15}

Sensor and the pacing lead were implanted subcutaneously in Sprague Dawley rats (one device per animal) in a 90-day implantation study and local tissue histology analyzed after 30 and 90 days of implantation. No adverse tissue reactions were observed after 30 or 90 days with either the Sensor or the pacing lead.

# C. Sensor Insertion Tools

The applicant changed the design of one of the sensor insertion tools after the completion of the clinical studies.

The insertion tools, as pictured below, were used during the Eversense CGM clinical studies.

Figure 1 - Sensor Insertion tools that were used for the PRECISE II and PRECISION clinical studies

Senseonics has developed a new version of the Blunt Dissector tool (Figure 2 below). Senseonics states that this re-design is being made to mitigate the risk of physicians inserting sensors too deeply. This was observed once during the PRECISE II study, and the result was that exploratory surgery with the patient under general anesthesia was required to remove the sensor; this was categorized as a serious adverse event. This event happened three times during the PRECISION study, and a surgeon was able to remove the sensor in each case using local anesthesia.

The new blunt dissector design has not been used in clinical studies. The design of the blunt dissector has been updated to add two guides (indicated by an orange arrow in figure 15 below). Also, the metal dissector portion is now shorter, and the user inserts it fully into the subdermal space. Previously, there were two lines etched onto the metal portion of the dissector to indicate how deep it should be inserted (see Figure 1 above).

PMA P160048: FDA Summary of Safety and Effectiveness Data

{16}

Figure 2 - Updated design of the Blunt Dissector tool, and examples of how it would be used (pictured with synthetic tissue samples).



To validate the new blunt dissector tool, Senseonics performed a human factors study. The human factors study participants included 16 healthcare providers who treat patients with diabetes. Participants completed sessions that included a system overview, watching a training video, a discussion of the package insert, and product training using simulated skin and the insertion tools. The synthetic tissue used for this process was a commercially available product. This was followed by a decay period of at least one hour before participants completed a usability testing scenario. This usability test involved participants performing a complete sensor insertion procedure on simulated skin installed in a model human arm (to mimic realistic arm position). Participants had an assistant available to assist with ancillary tasks (i.e. handling materials so sterility could be maintained). Successful use of the blunt dissector was judged based on the final insertion depth of the sensor in the simulated skin. Correct sensor depth was judged based on whether the sensor could be palpated after implantation. A selection of these synthetic tissue specimens (four of the fifteen) were dissected later and the actual sensor depth was measured and found to be within the intended insertion depth of $3 - 5\mathrm{mm}$ (actual depths ranged from 3.3 to $3.9\mathrm{mm}$ ).

Senseonics concluded that all participants were able to use the tool successfully to create a satisfactory sensor pocket in synthetic tissue. The one error scenario reported was when a participant failed to load a sensor into the insertion tool before inserting the tool into the sensor pocket in the artificial tissue.

Clinical use of the blunt dissector tool will be further assessed in the post-approval study phase.

# D. Sensor Design

The applicant changed the design of the sensor after the completion of the clinical studies.

The Eversense sensor includes a plastic end cap that is attached using epoxy after the electronics assembly is installed. The function of the cap is to seal the end of the

PMA P160048: FDA Summary of Safety and Effectiveness Data

{17}

sensor and to provide a smooth, uniform surface.

During sensor removal procedures in the clinical studies, there were several instances where the end cap of the sensor was broken off or missing after sensor removal. In some cases, the broken end caps were located, and in other cases the end caps were not located. A root-cause analysis into this failure concluded that the cause was most likely physicians grasping the end cap with the forceps during removal, instead of grabbing the sensor body. To reduce the potential for this failure, Senseonics redesigned the sensor end cap (see Figure 3 below) to be flush with the end of the sensor.

Figure 3 - (A) Sensor design used in PRECISE II and PRECISION studies, and (B) the proposed new sensor design with modified end cap. This design has not been used in any clinical studies to date.

This updated sensor design has not been studied in any clinical study. Senseonics has provided the results of manufacturing validation studies to demonstrate that the new sensors are being manufactured to the correct specifications. Part of this testing includes simulating the forces involved during sensor removal to demonstrate that the new end cap design can withstand greater forces than the previous design. This design change is not expected to affect clinical performance of the Eversense system. The effectiveness of this change in reducing the frequency of sensor fragmentation will be monitored during a post-approval study.

# E. Additional Studies

The Eversense Sensor was exposed to X-ray and ultrasonic energy test conditions stated in EN 45502-1. Essential performance was verified on the samples after the completion of exposure.

PMA P160048: FDA Summary of Safety and Effectiveness Data

{18}

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 19

# X. SUMMARY OF PRIMARY CLINICAL STUDIES

The applicant performed a pivotal clinical study (PRECISE II) to evaluate the performance of the Eversense CGM System to support 90 days of use. An additional clinical study (PRECISION) was performed to collect additional system accuracy data. Both clinical studies were conducted under IDE # G150165. Data from these two clinical studies served as the primary clinical studies for this premarket approval application.

## Glucose Determination Algorithm

One key element of the system that is responsible for Sensor performance and accuracy is the glucose determination algorithm (which includes the finger-stick calibration algorithm). The glucose determination algorithm is pre-programmed in the transmitter firmware, and it converts the raw data collected by the Sensor into glucose readings.

After the two clinical studies were completed, the applicant implemented a modified glucose determination algorithm. The applicant stated that the purpose of this algorithm change was to improve system accuracy, particularly in the early sensor wear period and in the hypoglycemic range.

The version of this algorithm that was used during the US clinical studies is referred to as the "study software" and is abbreviated "study SW." The new version of the algorithm is referred to as "software version 602" and is abbreviated "SW 602."

The algorithm changes within the SW 602 algorithm version targeted accuracy improvement in: 1) the early Sensor life, and 2) the hypoglycemic range throughout the Sensor life. The clinical accuracy data from a 71-subject European pivotal study, PRECISE (Kropff, Choudhary, Neupane, & Barnard, 2017), was used for as a training set for this new algorithm. Data from the PRECISE II study and PRECISION study were not used to develop the new algorithm (SW 602).

The applicant has not studied this new algorithm (SW 602) in real-time in a clinical trial. Rather, they have post hoc processed the raw sensor data from the PRECISE II and PRECISION studies using the new algorithm. The applicant stated that the raw sensor data is independent of this algorithm, so performing this processing post-hoc yields the same final glucose values as if the algorithm had been used during the study.

All data below and in the approved labeling for this device are from the SW 602 analysis.

## A. Study Design

The PRECISE II study was a non-randomized, blinded, prospective, single-arm, multi-center study, evaluating 90 adult subjects with diabetes mellitus in the United States at 8 sites. The investigation included both clinic visits and home use of the Eversense CGM System. 75 subjects had one sensor inserted in the upper arm by trained investigators. A subset of 15 subjects, at one clinical site, had two Sensors inserted. The accuracy of the CGM System was evaluated during clinic visits on days

{19}

1, 30, 60 and 90 by comparing Sensor glucose values and plasma glucose values drawn every 5 to 15 minutes for a period of approximately $4 \frac{1}{2}$ to $12 \frac{1}{2}$ hours and measured on a bedside glucose analyzer. During Sensor accuracy clinic visits, qualifying subjects participated in hyperglycemia and hypoglycemia challenges, as well as upper arm exercise sessions and separate compression sessions for a duration of 30 minutes each.

The first subject was enrolled on January 15, 2016. The last subject was completed on July 26, 2016. Eighty-seven (87) subjects completed the study; 2 subjects withdrew consent and 1 subject was lost to follow-up. Eighty-two (82) subjects completed the day 90 visit with accuracy data collection.

The CGM glucose values and all glucose-related alerts were blinded to both the subjects and the investigators for the duration of the study. All diabetes care decisions were based on SMBG blood glucose values and clinical standard of care, rather than CGM System results. The subjects did use the device for non-glucose related notifications such as calibration reminders and battery levels.

The subject visit schedule, which included 7 visits over a period of approximately 5 months, is summarized in Figure 4 below:

Figure 4: Primary Clinical Study Visit Schedule for PRECISE II

In the PRECISION study, the first subject was enrolled on July 25, 2017. The final subject was completed on February 1, 2018. The study evaluated 35 subjects, all of whom completed the study through the day 90 accuracy evaluation. Eight subjects were inserted with one Sensor (left arm) and 27 subjects were inserted with two Sensors (one in each arm). The PRECISION study shared the same design as the PRECISE II with the following exceptions (see Figure 5). Additional accuracy assessments were added on Day 7 and Day 14 to characterize Sensor accuracy during this period of wear, and patients underwent sleep assessments to evaluate accuracy and system performance during sleep. In addition, patients were not blinded to the glucose values and alerts during the

PMA P160048: FDA Summary of Safety and Effectiveness Data

{20}

PRECISION study.

Figure 5: Primary Clinical Study Visit Schedule for PRECISION

1. Inclusion/Exclusion Criteria for the Studies

Inclusion Criteria:

Male and Female Subjects meeting the following inclusion criteria were included in these studies:

1. Adult subjects, age $\geq 18$ years
2. Clinically confirmed diagnosis of diabetes mellitus for $\geq 1$ year
3. Subject has signed an informed consent form and is willing to comply with protocol requirements

Exclusion Criteria:

Subjects meeting any of the following exclusion criteria at the time of screening were excluded from these studies:

1. History of severe hypoglycemia in the previous 6 months. Severe hypoglycemia is defined as hypoglycemia resulting in loss of consciousness or seizure
2. History of diabetic ketoacidosis requiring emergency room visit or hospitalization in the previous 6 months
3. Female subjects of childbearing capacity (defined as not surgically sterile or not menopausal for $\geq 1$ year) who are lactating or pregnant, intending to become pregnant, or not practicing birth control during the course of the study.
4. A condition preventing or complicating the placement, operation, or removal of the Sensor or wearing of transmitter, including upper extremity

PMA P160048: FDA Summary of Safety and Effectiveness Data

{21}

deformities or skin condition.

5. Symptomatic coronary artery disease; unstable angina; myocardial infarction, transient ischemic attack or stroke in the past 6 months; uncontrolled hypertension (systolic >160 mm Hg or diastolic >100 mm Hg at time of screening); current congestive heart failure; history of cardiac arrhythmia (benign PACs and PVCs allowed). Subjects with asymptomatic coronary artery disease (e.g., CABG, stent placement or angioplasty) may participate if negative stress test within 1 year prior to screening and written clearance from Cardiologist documented.

6. Hematocrit <30% or >55%

7. History of hepatitis B, hepatitis C, or HIV

8. Current treatment for a seizure disorder unless written clearance by neurologist to participate in study

9. History of adrenal insufficiency

10. Currently receiving (or likely to need during the study period): immunosuppressant therapy; chemotherapy; anticoagulant/antithrombotic therapy (excluding aspirin); glucocorticoids (excluding ophthalmic or nasal). This exclusion does include the use of inhaled glucocorticoids and the use of topical glucocorticoids (over sensor site only); antibiotic for chronic infection (e.g. osteomyelitis, endocarditis)

11. A condition requiring or likely to require magnetic resonance imaging (MRI)

12. Known topical or local anesthetic allergy

13. Known allergy to glucocorticoids

14. Any condition that in the investigator's opinion would make the subject unable to complete the study or would make it not in the subject's best interest to participate in the study. Conditions include but are not limited to psychiatric conditions, known current or recent alcohol abuse or drug abuse by subject history, a condition that may increase the risk of induced hypoglycemia or risk related to repeated blood testing. Investigator will supply rationale for exclusion

15. Participation in another clinical investigation (drug or device) within 2 weeks prior to screening or intent to participate during the study period

16. The presence of any other active implanted device (as defined further in protocol)

17. The presence of any other CGM sensor or transmitter located in upper arm (other location is acceptable)

2. Follow-up Schedule

At the end of the Day 90 Clinic Visit, the Sensor was removed per the Eversense Physician Insertion & Removal Instructions; all the Sensor insertion sites were examined and evaluated by the study staff. A follow-up visit was scheduled 10 days later for evaluation of the Sensor site and close out. All used and unused Systems and sub-components, except for used insertion tools, were returned by study staff to Senseonics for examination. Study investigators documented any Adverse Device Effects and evaluated safety issues related to system use during the study.

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 22

{22}

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 23

3. Clinical Endpoint
The study characterized the performance of the System in comparison with the laboratory reference venous plasma sample measurements and assessed the system-reference matched pairs obtained in the in-clinic sessions.

Safety data for the Eversense System were also collected and characterized by incidence, severity and relatedness. Device incidents and malfunctions were also collected.

B. Accountability of Study Subjects and Time of Exposure
In the PRECISE II study, 90 subjects were inserted with the Sensor and 87 (97%) completed the study. The mean duration of Sensor use was 92.2 days and the median duration was 93.0 days, resulting in 9,773 in vivo days of Sensor use in 90 subjects to assess safety. A total of 106 sensors were inserted, including 75 subjects with 1 Sensor and 15 with 2 Sensors, and 1 Sensor replacement during the study. Two subjects withdrew consent and had Sensors removed on Days 62 and 92. One subject was lost to follow up, but subsequently returned to the site and had the sensor removed 196 days after insertion.

In the PRECISION study, 36 subjects were enrolled and 35 were inserted with Sensors with 8 receiving one (1) Sensor and 27 receiving two (2) Sensors. All 35 subjects completed all visits at Day 1, 7, 14, 30, 60 and 90.

C. Study Population Demographics and Baseline Parameters
A summary of demographic characteristics is presented in Table 8 and Table 9.

Table 8: Demographic Information
| Demographic | PRECISE II (n=90) | PRECISION (n=35) |
| --- | --- | --- |
| Gender [n (%)] | | |
| Male | 54 (60) | 18 (51) |
| Female | 36 (40) | 17 (49) |
| Age (years) [mean (SD)] | 45(16) | 52 (16) |
| Min, Max | 18, 77 | 18, 75 |
| Race n (%) | | |
| Caucasian | 77 (86) | 32 (91) |
| Black or African American | 7 (8) | 1 (3) |
| Asian | 3 (3) | 2 (6) |
| American Indian or Alaska Native | 2 (2) | 0 (0) |
| Native Hawaiian or Other Pacific Islander | 1 (1) | 0 (0) |
| Dominant Hand [n (%)] | | |
| Right | 78 (87) | 33 (94) |
| Left | 12 (13) | 2 (6) |

{23}

| Demographic | PRECISE II (n=90) | PRECISION (n=35) |
| --- | --- | --- |
| Body Mass Index Class [n (%)] [mean (SD)] | 29 (6) | 28 (5) |
| Min, Max | 19, 50 | 19, 44 |
| Normal (<25 kg/m²) | 22 (24) | 9 (26) |
| Overweight (≥25 and <30) | 27 (30) | 11 (31) |
| Obese (≥30) | 41 (46) | 15 (43) |

Table 9: Diabetic History

| Diabetic History | PRECISE II (n=90) | PRECISION (n=35) |
| --- | --- | --- |
| Years since diabetes diagnosis (years) | 26.0 (14.3) | 20.1 (13.7) |
| [mean(SD)] | 4, 57 | 1, 53 |
| Min, Max | | |
| Diabetes type (n/%) | | |
| Type I | 25 (71.4) | 61 (67.8%) |
| Type II | 10 (28.6) | 29 (32.2%) |
| Type of insulin therapy (n/%) | | |
| None | 5 (14.3) | 20 (22.2%) |
| Multiple daily injections | 11 (31.4) | 24 (26.7%) |
| Continuous insulin infusion pump | 19 (54.3) | 43 (47.8%) |
| Other | | 3 (3.3%) |
| History of ketoacidosis (n/%) | 0 (0.0) | 0 (0%) |
| History of hypoglycemia (n/%) | 0 (0.0) | 1 (1.1%) |

Of the 125 subjects in both studies, 86 had Type I diabetes (61 in PRECISE II study and 25 in the PRECISION study). A total of 62 subjects had continuous insulin infusion pump (43 in the PRECISE II study and 19 in the PRECISION study).

## A. Safety and Effectiveness Results

### 1. Safety Results

The safety endpoints and evaluations performed in the PRECISE II study and the PRECISION study were the same. At each study visit a safety evaluation was performed. Sensor sites were evaluated and assessed for any signs of irritation or infection, including increased temperature, pain, redness, warmth, swelling or purulence. In addition, subjects were queried at each visit for Sensor site assessment between visits, as well as other adverse events. Subjects were asked at the beginning of each visit if anything had changed medically since their last visit. All adverse events identified, regardless of relatedness to the device or insertion/removal procedure, were documented.

The primary safety analysis was based upon all subjects in the investigation who were not screen failures or withdrawals prior to a first insertion attempt. Ninety

PMA P160048: FDA Summary of Safety and Effectiveness Data

{24}

(90) subjects were successfully inserted with a Sensor in the PRECISE II Study and 35 in the PRECISION Study, forming the basis of the safety populations. In the PRECISE II study, 15 subjects had two (2) Sensors inserted (one in each arm) and 75 subjects had one (1) Sensor inserted. One subject had a replacement Sensor inserted after the primary Sensor had a suspected electrical or mechanical failure. In the PRECISION study, 8 subjects had one (1) Sensor inserted and 27 had two (2) Sensors inserted (one in each arm).

The primary safety endpoint was the incidence of device-related or Sensor insertion/removal procedure-related serious adverse events (SAE) through 90 days post insertion or Sensor removal and follow-up. An adverse event relationship was considered non-related, possibly related, related or unknown based upon review and categorization by the independent medical monitor. An analysis was provided through Sensor removal as shown in Table 10. The proportion of subjects experiencing a serious adverse event is presented together with the associated $95\%$ confidence interval.

Table 10: Safety Endpoints in the PRECISE II and PRECISION Studies

| SAEs by Relationship to Study | PRECISE II (N=90) | PRECISION (N=35) |
| --- | --- | --- |
| | Number of Subjects with SAEs (%) | Number of Subjects with SAEs (%) |
| All SAEs | 1 (1.1%) | 3 (8.6%) |
| Device-Related SAEs | 0 (0.0%) | 0 (0.0%) |
| Sensor Insertion/Removal Procedure-Related SAEs | 1 (1.1%) | 0 (0.0%) |
| Study Procedure-Related SAEs | 0 (0.0%) | 0 (0.0%) |
| Unrelated to Study SAEs | 0 (0.0%) | 3 (8.5%) |

The secondary safety endpoints included:

- Incidence of device-related or insertion/removal procedure-related serious adverse events over the operating life of the Sensor.
- Incidence of insertion/removal procedure or device-related adverse events in the clinic and during home use.
- Incidence of all adverse events in the clinic and during home use.
- Incidence of hospitalizations due to hypoglycemia, hyperglycemia or ketoacidosis occurring during home use.
- Incidence of reported hypoglycemic and hyperglycemic events occurring during home use.

PMA P160048: FDA Summary of Safety and Effectiveness Data

{25}

Table 11, below, shows the safety data from each study. Fourteen (14) adverse events that were determined to be device- and/or insertion/removal procedure-related or possibly related, including the one (1) SAE (in which the initial attempt to remove a sensor was unsuccessful and a surgeon later removed the sensor using general anesthesia), occurred in the PRECISE II study among 7 (7.7%) subjects. All events adjudicated as related or possibly related to the device and/or insertion/removal procedures had complete resolution by study completion with exception of one subject. One subject had a delayed report of intermittent pain adjudicated as possibly related. Eight (8) adverse events occurred in 5 subjects (14.3%) in the PRECISION study, and all device-related adverse events were mild or moderate in severity and resolved within 2 weeks of Sensor removal. Importantly, most subjects received two Sensors in the PRECISION study, which resulted in higher device-related adverse events rate when compared to PRECISE II study. There were no unanticipated adverse events and no UADEs. There were no infections observed in either study, resulting in an infection rate of 0.0%.

Table 11: Adverse Events Related or Possibly Related to the Study Device or Insertion/removal Procedure

| | PRECISE II (n=90) | | PRECISION (n=35) | |
| --- | --- | --- | --- | --- |
| | Number of Events | Number of Subjects (%) | Number of Events | Number of Subjects (%) |
| Event Physiologic System | 14 | 7 (7.7%) | 8 | 5 (14.3%) |
| Dermatological | 8 | 4 (4.4%) | 6 | 4 (11.4%) |
| Bruising | 2 | | 0 | |
| Erythema | 2 | | 0 | |
| Pain/Discomfort | 4 | | 2 | |
| Dermatitis | 0 | | 2 | |
| Hyperpigmentation of skin | 0 | | 2 | |
| Musculoskeletal/Rheumatologic | 1 | 1 (1.1%) | 0 | 0 (0.0%) |
| Pain/Discomfort | 1 | | 0 | |
| Neurological | 2 | 2 (2.2%) | 0 | 0 (0.0%) |
| Paresthesia | 1 | | 0 | |
| Syncope-vasovagal | 1 | | 0 | |
| Other | 3 | 3 (3.3%) | 2 | 1 (2.9%) |
| Device fragment not recovered | 2 | | 0 | |
| Additional procedure to remove Sensor | 1 | | 2 | |

# Assessment of Dexamethasone Exposure

During the PRECISION Study, a subset of 8 subjects with one Sensor inserted into the left arm had blood samples drawn at 30 minutes, 2 hours and 4 hours post-insertion and then daily for at least the first 8 days of Sensor wear for additional DXA evaluation and to determine blood draw time points during the

PMA P160048: FDA Summary of Safety and Effectiveness Data

{26}

first week of Sensor wear for the remaining subjects. The assay used to measure DXA in plasma had a limit of quantitation (LoQ) of $50~\mathrm{pg / mL}$ . The analysis showed that DXA greater than $50~\mathrm{pg / ml}$ was not detected in any subject during the first 8 days nor at subsequent visits through 90 days in this group.

The remaining subjects had 2 Sensors inserted, one in each arm, and underwent blood draws for DXA evaluation 2 hours after insertion and at every clinic visit including 2 draws on clinic visits that spanned two calendar days. There were 18 of 27 (66%) subjects with two Sensors that had detectable levels above $50~\mathrm{pg / ml}$ in the first 8 days. In all cases in these 18 subjects, the plasma DXA was below the LoQ of $50~\mathrm{pg / mL}$ by day 8 post-insertion. The maximum level detected was $114~\mathrm{pg / ml}$ at day 2 in one subject which fell below the detection limit by day 7. DXA greater than $50~\mathrm{pg / ml}$ was not detected at subsequent visits through 90 days in this group.

The results from the DXA measurements taken throughout this study are provided below in Table 12 below.

Table 12: Plasma DXA measurements in subjects during the PRECISION clinical study. Subjects with 1 sensor implanted had serial measurements collected on the day of sensor insertion, all of which showed DXA values below the assay LoQ of $0.050\mathrm{ng / mL}$ .

| Time After Insertion | Subjects with 1 Sensor (n=8) | Subjects with 2 Sensors (n=27) |
| --- | --- | --- |
| | Detections >0.050 ng/mL | Detections >0.050 ng/mL |
| Day 0 (Immediate Post Insertion) | 0 | 0 |
| Day 1 | 0 | 18 |
| Day 2* | 0 | 10 |
| Day 3 | 0 | 0 |
| Day 4 | 0 | 0 |
| Day 5 | 0 | 0 |
| Day 6 | 0 | 0 |
| Day 7 | 0 | 1 |
| Day 8 | 0 | 1 |
| Day 9 | 0 | 0 |
| Day 14 | 0 | 0 |
| Day 15 | 0 | 0 |
| Day 30 | 0 | 0 |

* Highest detected amount was ${114}\mathrm{\;{pg}}/\mathrm{{mL}}$ on Day 2

Sensors removed from patients in clinical studies were returned to the applicant for evaluation of residual DXA content. Explanted Sensors retained approximately 80-90% of their original DXA content at 90 days. This corresponds to approximately 0.18 - 0.35 mg of DXA being released into the body from a single Sensor over the

PMA P160048: FDA Summary of Safety and Effectiveness Data

{27}

course of 90 days.

# 2. Effectiveness Results

The analysis of effectiveness was based on the observed accuracy of sensors in 90 evaluable patients in the PRECISE II study who contributed 15,753 CGM-comparator matched glucose data pairs, and 35 evaluable patients in the PRECISION study who contributed 15,170 CGM-comparator matched glucose data pairs.

All effectiveness analyses presented in this document were performed using the approved glucose determination algorithm, referred to as the "SW-602" algorithm.

The following tables show the rate at which the Eversense CGM System agreed with a laboratory comparator method (CM). The tables are organized by CM system glucose ranges, and they tabulate the percent of CGM system measurements that were within a given range of paired comparator measurements. The ranges included below are 15, 20, 30, 40, and greater than 40. For comparator values below $80\mathrm{mg / dL}$ , the units of the range value are $\mathrm{mg / dL}$ . For CGM values above $80\mathrm{mg / dL}$ , the units of the range value are percent.

The data which are tabulated in Table 13 below is a combination of data collected on four different days of the PRECISE II study: days 1, 30, 60, and 90 of sensor wear.

Table 13: CGM System and Comparator Agreement in Different Comparator Glucose Ranges, PRECISE II Study, SW-602 data

| CM Glucose Range (mg/dL) | Number of Paired Eversense CGM and CM | Percent of CGM System Readings Within | | | | |
| --- | --- | --- | --- | --- | --- | --- |
| | | Percent 15/15% of CM | Percent 20/20% of CM | Percent 30/30% of CM | Percent 40/40% of CM | Percent Greater than 40/40% of CM |
| Overall (40-400) | 15753 | 86.8 | 94.3 | 98.6 | 99.6 | 0.4 |
| < 40 | 7 | 71.4 | 71.4 | 100.0 | 100.0 | 0.0 |
| 40 - 60 | 488 | 89.5 | 95.1 | 98.8 | 99.8 | 0.2 |
| 61 - 80 | 1159 | 84.5 | 92.0 | 97.7 | 99.1 | 0.9 |
| 81 - 180 | 7540 | 85.6 | 93.0 | 98.0 | 99.4 | 0.6 |
| 181 - 300 | 5378 | 88.4 | 95.9 | 99.4 | 99.9 | 0.1 |
| 301 - 350 | 820 | 88.4 | 97.4 | 99.8 | 100.0 | 0.0 |
| 351 - 400 | 326 | 86.5 | 96.6 | 98.5 | 100.0 | 0.0 |
| > 400 | 35 | 91.4 | 100.0 | 100.0 | 100.0 | 0.0 |

The data which are tabulated in the following table are a combination of data collected on six different days of the PRECISION study: days 1, 7, 14, 30, 60, and 90 of sensor wear.

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 28

{28}

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 29

{29}

Table 14: CGM System and Comparator Agreement in Different Comparator Glucose Ranges, PRECISION Study, SW-602 data

| CM Glucose Range (mg/dL) | Number of Paired Eversense CGM and CM | Percent of CGM System Readings Within | | | | |
| --- | --- | --- | --- | --- | --- | --- |
| | | Percent 15/15% of CM | Percent 20/20% of CM | Percent 30/30% of CM | Percent 40/40% of CM | Percent Greater than 40/40% of CM |
| Overall (40-400) | 15170 | 85.4 | 92.8 | 98.1 | 99.3 | 0.7 |
| < 40 | 15 | 60.0 | 73.3 | 86.7 | 100.0 | 0.0 |
| 40 - 60 | 1267 | 86.8 | 92.6 | 98.1 | 99.1 | 0.9 |
| 61 - 80 | 2212 | 85.8 | 93.0 | 98.5 | 99.3 | 0.7 |
| 81 - 180 | 5685 | 80.6 | 89.4 | 96.7 | 98.8 | 1.2 |
| 181 - 300 | 3210 | 87.4 | 94.9 | 98.6 | 99.8 | 0.2 |
| 301 - 350 | 1527 | 91.4 | 97.8 | 100.0 | 100.0 | 0.0 |
| 351 - 400 | 1174 | 93.4 | 97.5 | 99.7 | 100.0 | 0.0 |
| > 400 | 80 | 81.3 | 93.8 | 97.5 | 100.0 | 0.0 |

The following two tables provide the number and percentage of CM measurements collected while the continuous glucose monitor read 'low' (&lt;40 mg/dL) or 'high' (&gt;400 mg/dL) during the PRECISE II and PRECISION clinical studies.

Table 15: The number and percentage of Comparator values collected when CGM readings displayed 'Low' (less than $40\mathrm{mg / dL}$ ) or 'High' (greater than $400\mathrm{mg / dL}$ ); PRECISE II Study, SW-602 algorithm

| Comparator mg/dL | | | | | | | |
| --- | --- | --- | --- | --- | --- | --- | --- |
| CGM Readings | CGM-Ref pairs | <55 | <60 | <70 | <80 | >80 | Total |
| ‘LOW’ | n | 0 | 1 | 2 | 2 | 0 | 2 |
| | % | 0% | 50% | 100% | 100% | 0% | 100% |
| | | | | | | | |
| Comparator mg/dL | | | | | | | |
| CGM Readings | CGM-Ref pairs | >340 | >320 | >280 | >240 | <240 | Total |
| ‘HIGH’ | n | 67 | 68 | 68 | 68 | 0 | 68 |
| | % | 98.5% | 100% | 100% | 100% | 0% | 100% |

PMA P160048: FDA Summary of Safety and Effectiveness Data

{30}

Table 16: The number and percentage of Comparator values collected when CGM readings displayed 'Low' (less than $40\mathrm{mg / dL}$ ) or 'High' (greater than $400\mathrm{mg / dL}$ ); PRECISION Study, SW-602 algorithm

| Comparator mg/dL | | | | | | | |
| --- | --- | --- | --- | --- | --- | --- | --- |
| CGM Readings | CGM-Ref pairs | <55 | <60 | <70 | <80 | >80 | Total |
| ‘LOW’ | n | 6 | 7 | 9 | 9 | 0 | 9 |
| | % | 66.7% | 77.8% | 100% | 100% | 0% | 100% |
| | | | | | | | |
| Comparator mg/dL | | | | | | | |
| CGM Readings | CGM-Ref pairs | >340 | >320 | >280 | >240 | <240 | Total |
| ‘HIGH’ | n | 359 | 383 | 399 | 404 | 1 | 405 |
| | % | 88.6% | 94.6% | 98.5% | 99.8% | 0.2% | 100% |

The following tables show the rate of concurrence between the Eversense CGM System and a laboratory comparator method. The tables are organized by CGM system glucose ranges, and they tabulate the percent of paired CM measurements that were in the identical range (shaded diagonal), as well as those CM measurements that were in glucose ranges above and below the paired CGM readings. Cells with dashes $-$ indicate zero percent $(0\%)$ .

Table 17: CGM System concurrence to Comparator organized by CGM glucose ranges; data pooled from accuracy assessments on days 1, 30, 60, and 90 combined of the PRECISE II clinical study, analyzed using SW-602 algorithm

| CGM (mg/dL) | Percent of Matched Pairs in Each CM Glucose Range for Each CGM Range (mg/dL) | | | | | | | | | | | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | Number of Paired CGM-CM | <40 | 40-60 | 61-80 | 81-120 | 121-160 | 161-200 | 201-250 | 251-300 | 301-350 | 351-400 | >400 |
| 40-60 | 480 | 1% | 63% | 34% | 3% | -- | -- | -- | -- | -- | -- | -- |
| 61-80 | 1111 | -- | 16% | 63% | 20% | 1% | -- | -- | -- | -- | -- | -- |
| 81-120 | 3066 | -- | -- | 9% | 76% | 14% | -- | -- | -- | -- | -- | -- |
| 121-160 | 3245 | -- | -- | -- | 11% | 73% | 15% | -- | -- | -- | -- | -- |
| 161-200 | 2812 | -- | -- | -- | -- | 15% | 64% | 21% | -- | -- | -- | -- |
| 201-250 | 2614 | -- | -- | -- | -- | -- | 13% | 68% | 18% | -- | -- | -- |
| 251-300 | 1484 | -- | -- | -- | -- | -- | 1% | 17% | 58% | 23% | 1% | -- |
| 301-350 | 692 | -- | -- | -- | -- | -- | -- | 1% | 19% | 59% | 20% | -- |
| 351-400 | 249 | -- | -- | -- | -- | -- | -- | -- | -- | 20% | 66% | 13% |

PMA P160048: FDA Summary of Safety and Effectiveness Data

{31}

Table 18: CGM System concurrence to Comparator, organized by CGM glucose ranges; data pooled from accuracy assessments on days 1, 7, 14, 30, 60, and 90 combined of the PRECISION clinical study, analyzed using SW-602 algorithm

| CGM (mg/dL) | Percent of Matched Pairs in Each CM Glucose Range for Each CGM Range (mg/dL) | | | | | | | | | | | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | Number of Paired CGM-CM | <40 | 40-60 | 61-80 | 81-120 | 121-160 | 161-200 | 201-250 | 251-300 | 301-350 | 351-400 | >400 |
| 40-60 | 1236 | 1% | 63% | 34% | 2% | -- | -- | -- | -- | -- | -- | -- |
| 61-80 | 2003 | -- | 22% | 67% | 10% | -- | -- | -- | -- | -- | -- | -- |
| 81-120 | 2524 | -- | 2% | 17% | 71% | 10% | -- | -- | -- | -- | -- | -- |
| 121-160 | 2342 | -- | -- | -- | 18% | 71% | 11% | -- | -- | -- | -- | -- |
| 161-200 | 1727 | -- | -- | -- | 1% | 24% | 59% | 16% | -- | -- | -- | -- |
| 201-250 | 1502 | -- | -- | -- | -- | 1% | 19% | 65% | 14% | 1% | -- | -- |
| 251-300 | 1257 | -- | -- | -- | -- | -- | 1% | 18% | 51% | 27% | 3% | -- |
| 301-350 | 1628 | -- | -- | -- | -- | -- | -- | 1% | 10% | 57% | 32% | 1% |
| 351-400 | 951 | -- | -- | -- | -- | -- | -- | -- | 2% | 26% | 65% | 7% |

The following tables show the consistency of sensor clinical performance during the sensor wear period by comparing the CM values to their paired sensor points collected on days 1, 30, 60, and 90 of the PRECISE II study, and on days 1, 7, 14, 30, 60, and 90 of the PRECISION study.

Table 19: Sensor stability (accuracy over time); PRECISE II Study, SW-602

| Day | Number of Readings | Mean Absolute Relative Difference (%) | Median Absolute Relative Difference (%) | Percent of CGM System Readings Within | | | | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | | | Percent within 15/15% CM | Percent within 20/20% CM | Percent within 30/30% CM | Percent within 40/40% CM | Percent greater than 40/40% CM |
| 1 | 1708 | 10.7 | 8.2 | 76.8% | 87.1% | 96.3% | 98.5% | 1.5% |
| 30 | 5081 | 7.4 | 5.5 | 90.7% | 96.0% | 99.3% | 99.8% | 0.2% |
| 60 | 4725 | 8.2 | 6.3 | 87.3% | 94.7% | 98.8% | 99.8% | 0.2% |
| 90 | 4239 | 9.1 | 7.3 | 85.4% | 94.7% | 98.6% | 99.6% | 0.4% |

PMA P160048: FDA Summary of Safety and Effectiveness Data

{32}

Table 20: Sensor stability (accuracy over time); PRECISION study, SW-602

| Day | Number of Readings | Mean Absolute Relative Difference (%) | Median Absolute Relative Difference (%) | Percent of CGM System Readings Within | | | | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | | | Percent within 15/15% CM | Percent within 20/20% CM | Percent within 30/30% CM | Percent within 40/40% CM | Percent greater than 40/40% CM |
| 1 | 2665 | 11.6 | 8.5 | 79.1% | 88.9% | 95.8% | 98.5% | 1.5% |
| 7 | 2926 | 9.8 | 6.8 | 86.1% | 93.3% | 98.1% | 99.0% | 1.0% |
| 14 | 2997 | 9.0 | 6.6 | 88.1% | 94.6% | 98.8% | 99.6% | 0.4% |
| 30 | 2284 | 8.9 | 6.8 | 88.0% | 94.3% | 98.9% | 100.0% | 0.0% |
| 60 | 2133 | 8.7 | 6.9 | 86.9% | 93.7% | 98.5% | 99.6% | 0.4% |
| 90 | 2165 | 9.7 | 7.8 | 83.9% | 92.2% | 98.5% | 99.3% | 0.7% |

The tables below provide the percent agreement of the Eversense CGM system and comparator method (CM) within a specific time range after calibration for the PRECISE II and PRECISION studies.

Table 21: Agreement rates for every 2-hour period post calibration; PRECISE II, SW-602

| Time from Calibration | Number of Paired CGM and CM | Percent of CGM System Readings Within | | | | |
| --- | --- | --- | --- | --- | --- | --- |
| | | Percent within 15/15% CM | Percent within 20/20% CM | Percent within 30/30% CM | Percent within 40/40% CM | Percent greater than 40/40% CM |
| (0, 2) Hours | 4347 | 85.0% | 92.2% | 97.8% | 99.3% | 0.7% |
| [2, 4) Hours | 2800 | 87.5% | 94.8% | 98.9% | 99.7% | 0.3% |
| [4, 6) Hours | 2396 | 85.5% | 93.8% | 98.5% | 99.7% | 0.3% |
| [6, 8) Hours | 2115 | 87.6% | 95.6% | 99.1% | 99.6% | 0.4% |
| [8, 10) Hours | 2019 | 87.8% | 95.9% | 99.3% | 100.0% | 0.0% |
| [10, 12) Hours | 1815 | 88.9% | 95.8% | 98.8% | 99.6% | 0.4% |

Table 22: Agreement rates for every 2-hour period post calibration; PRECISION, SW-602

| Time from Calibration | Number of Paired CGM and CM | Percent of CGM System Readings Within | | | | |
| --- | --- | --- | --- | --- | --- | --- |
| | | Percent within 15/15% CM | Percent within 20/20% CM | Percent within 30/30% CM | Percent within 40/40% CM | Percent greater than 40/40% CM |
| (0, 2) Hours | 4034 | 86.0% | 93.6% | 98.0% | 99.3% | 0.7% |
| [2, 4) Hours | 3979 | 85.6% | 92.8% | 98.4% | 99.5% | 0.5% |
| [4, 6) Hours | 2308 | 84.3% | 92.2% | 97.7% | 99.0% | 1.0% |
| [6, 8) Hours | 1614 | 84.3% | 92.7% | 97.8% | 99.4% | 0.6% |
| [8, 10) Hours | 1372 | 88.0% | 94.4% | 98.5% | 99.6% | 0.4% |
| [10, 12) Hours | 1295 | 86.1% | 92.9% | 98.1% | 99.2% | 0.8% |

PMA P160048: FDA Summary of Safety and Effectiveness Data

{33}

The following tables provide data to present sensor accuracy at detecting specific glucose rates of change. These concurrence tables provide the percent of matched CM pairs to CGM values over specific glucose rates of change as observed during the PRECISE II and PRECISION studies.

Table 23: Concurrence of CGM and Comparator Method (CM) rate of change stratified by different CGM rate ranges; PRECISE II, SW-602

| CGM Trend (mg/dL/Min) | Comparator Rate of Change (mg/dL/Min) Percent of Matched Pairs in Each Comparator Trend Range for Each CGM Trend Range | | | | | Total |
| --- | --- | --- | --- | --- | --- | --- |
| | < -2 | [-2, -1) | [-1, 1] | (1, 2] | > 2 | |
| < -2 | 33% | 44% | 23% | 0% | 0% | 212 |
| [-2, -1) | 4% | 39% | 56% | 1% | 0% | 1202 |
| [-1,1] | 0% | 5% | 88% | 6% | 1% | 11546 |
| (1,2] | 0% | 1% | 49% | 38% | 12% | 1101 |
| > 2 | 0% | 1% | 17% | 34% | 48% | 503 |
| Total | 159 | 1161 | 11471 | 1262 | 511 | 14564 |

Table 24: Concurrence of CGM and Comparator Method (CM) rate of change stratified by difference CGM rate ranges; PRECISION, SW-602

| CGM Trend (mg/dL/Min) | Comparator Rate of Change (mg/dL/Min) Percent of Matched Pairs in Each Comparator Trend Range for Each CGM Trend Range | | | | | Total |
| --- | --- | --- | --- | --- | --- | --- |
| | < -2 | [-2, -1) | [-1, 1] | (1, 2] | > 2 | |
| < -2 | 22% | 36% | 41% | 0% | 1% | 473 |
| [-2, -1) | 4% | 24% | 71% | 1% | 0% | 1115 |
| [-1,1] | 1% | 5% | 89% | 5% | 1% | 10655 |
| (1,2] | 0% | 1% | 55% | 35% | 9% | 997 |
| > 2 | 0% | 1% | 32% | 39% | 27% | 529 |
| Total | 212 | 990 | 11137 | 1085 | 345 | 13769 |

Table 25 and Table 26 below provide the Eversense sensor percent difference with respect to comparator method (CM) values. The comparator method used during this study was the Yellow Springs Instruments 2300 glucose analyzer.

PMA P160048: FDA Summary of Safety and Effectiveness Data

{34}

Table 25: Difference measures between Eversense CGM System and Comparator Readings (CM), PRECISE II study using SW-602

| CM Glucose Range (mg/dL) | Number of Paired CGM- CM | Mean Absolute Relative Difference (%) | Median Absolute Relative Difference (%) |
| --- | --- | --- | --- |
| Overall | 15753 | 8.5 | 6.5 |
| < 40* | 7 | 13.0 | 8.0 |
| 40-60* | 488 | 7.8 | 6.0 |
| 61-80* | 1159 | 8.7 | 6.0 |
| 81-180 | 7540 | 8.2 | 6.1 |
| 181-300 | 5378 | 7.6 | 6.2 |
| 301-350 | 820 | 7.9 | 6.9 |
| 351-400 | 326 | 7.5 | 6.1 |
| > 400 | 35 | 8.3 | 7.4 |

*For CM ≤ 80 mg/dL, the differences in mg/dL are included instead of percent difference (%).

Table 26: Difference measures between Eversense CGM System and Comparator Readings, PRECISION study using SW-602

| CM Glucose Ranges (mg/dL) | Number of Paired CGM- CM | Mean Absolute Relative Difference (%) | Median Absolute Relative Difference (%) |
| --- | --- | --- | --- |
| Overall | 15170 | 9.6 | 7.1 |
| <40* | 15 | 16.2 | 14 |
| 40-60* | 1267 | 8.1 | 6 |
| 61-80* | 2212 | 8.6 | 7 |
| 81-180 | 5685 | 9.7 | 7.5 |
| 181-300 | 3210 | 7.7 | 5.9 |
| 301-350 | 1527 | 6.8 | 5.7 |
| 351-400 | 1174 | 6.5 | 5.5 |
| >400 | 80 | 11.2 | 10.5 |

*For CM ≤ 80 mg/dL, the differences in mg/dL are included instead of percent difference (%).

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 35

{35}

PMA P160048: FDA Summary of Safety and Effectiveness Data
Page 36

# Precision Analysis

Precision of the System was evaluated by comparing the results from two separate sensors worn on the same subject at the same time. During the PRECISE II study, a total of 15 subjects contributed 10,371 between-sensor matched pairs. During the PRECISION study, a total of 27 subjects contributed 3…