Making the Most of Continuous Glucose Monitoring Gary Scheiner MS, CDE Owner & Clinical Director Integrated Diabetes Services LLC Wynnewood, PA “Dean” Type-1 University www.type1university.com

Making the Most of Continuous Glucose Monitoring What Information Is Available How to Use Immediate Data How to Use Intermediate Data What Can Be Learned from Retrospective Analysis Optimizing CGM System Performance

Medtronic Carelink Personal Report Options: Medtronic Carelink Personal Sensor Daily Overlay Sensor Overlay By Meal

Medtronic Carelink Personal Report Options: Medtronic Carelink Personal Daily Summary Layered Report: Sensor tracing & BG entries Basal & bolus delivery Carbohydrate, exercise & logbook entries

Medtronic Carelink Personal Report Options: Medtronic Carelink Personal Statistical Summary Avg, SD, Hi/Low # Hi/Low Excursions, AUC % Time above, below, within target range

Report Options: DexCom DM3* Hourly Stats w/data table for each hour Glucose Trend Includes event entries * Dexcom 7+ System not FDA approved for use in children under age 18 in the U.S.

Report Options: DexCom DM3 BG Distribution % high, low, normal for each segment of the day Modal Day Customizable date range

Report Options: DexCom DM3 16 Jan - 15 Apr 10 16 Apr - 15 Jul 10 Change A1c % 0.0 % N/A Mean Glucose 191 164 -14 % Standard Deviation 65 64 -2 % % in Hypoglycemia (39-55 mg/dL) 1 % in Low (55-70 mg/dL) 3 200 % % in Target (70-160 mg/dL) 33 48 45 % % in High (160-240 mg/dL) 46 36 -22 % % in Hyperglycemia (240-401 mg/dL) 20 12 -40 % Days Sensor Used 22 91 314 % Success Report Changes in control: week-to-week, month-to-month, or quarter-to-quarter Breakdown by hour, with averages 12 am 1 am 2 am 3 am 4 am 5 am 6 am 7 am 8 am 9 am 10 am 11 am 16 Jan - 15 Apr 10 206 199 204 208 198 179 169 166 172 181 196 16 Apr - 15 Jul 10 165 167 170 168 163 156 148 147 152 164 175 176

Report Options: Freestyle Navigator/Copilot Modal Day Report Customizable by date range, day of week Glucose Line Report Stats Report Broken down by phase of the day

Report Options: Freestyle Navigator/Copilot Logbook Report (Sensor BG q10 minutes)

What Do We Get in Real Time? Trends Alerts Numbers

Decision-Making Based on Trend Information Self-Care Choices To snack? To check again soon? To exercise? To adjust insulin? Key Situations Driving Sports Tests Bedtime Third, REAL-Time Trend Arrows. One arrow up or down indicates a change of 1-2 mg/dl/min in the last 20 minutes. Two arrows up or down indicate a change of more than 2 mg/dl/min in the last 20 minutes. Arrows are only displayed if there has been at least a 1 mg/dl/min change in the last 20 minutes. Patients should think about where glucose may be in 20 minutes if it continues at the same rate.

Bolus Adjustment Based on Trend Information BG Stable: Usual Bolus Dose BG Rising Gradually:  bolus slightly* BG Rising Sharply:   bolus modestly** BG Dropping Gradually:  bolus slightly* BG Dropping Sharply:   bolus modestly** * Enough to offset 25 mg/dl (1.5 mmol/l) ** Enough to offset 50 mg/dl (3 mmol/l) Third, REAL-Time Trend Arrows. One arrow up or down indicates a change of 1-2 mg/dl/min in the last 20 minutes. Two arrows up or down indicate a change of more than 2 mg/dl/min in the last 20 minutes. Arrows are only displayed if there has been at least a 1 mg/dl/min change in the last 20 minutes. Patients should think about where glucose may be in 20 minutes if it continues at the same rate. 13

Immediate Info: Hypoglycemia Alerts Predictive Hypo Alert or Hypo Alert & recovering: Subtle Treatment 50% of usual carbs Med-High G.I. food Hypo Alert & Dropping: Aggressive Treatment Full or increased carbs High G.I. food vs Third, REAL-Time Trend Arrows. One arrow up or down indicates a change of 1-2 mg/dl/min in the last 20 minutes. Two arrows up or down indicate a change of more than 2 mg/dl/min in the last 20 minutes. Arrows are only displayed if there has been at least a 1 mg/dl/min change in the last 20 minutes. Patients should think about where glucose may be in 20 minutes if it continues at the same rate. 14

Types of Alerts Hi/Low Alert: Cross specified high or low thresholds Predictive Alert: Anticipated crossing of high or low thresholds Rate of Change: Rapid rise or fall

The Value of Alerts: Minimizing the DURATION and MAGNITUDE of BG Excursions

CGM Turns Mountains into Molehills

Uniform Response is Key! Fingerstick Act on the Fingerstick

Setting Alerts Hi/Low alert thresholds are not BG target ranges Balance need for alerts against “nuisance factor” Predictive alerts lose value the further the advance warning (keep below 10 min) Rate of FALL alerts helpful for long-term hypo prevention (>3 mg/dl/min) (.17)

Initial Hi/Low Alert Settings LOW: 80 mg/dl (4.5 mmol) (90/5+ if hypo unaware) HIGH: 300 mg/dL (18 mmol) (lower progressively toward 180/10) It is Medtronic’s recommendation that the starting alert setting on initiation should be a low of 80 mg/dL and a high of 240 mg/dL. It is important that patients do not confuse these alert settings with blood glucose targets. For instance, the low threshold of 70 might mean that the patient would miss hypoglycemia in the 60 or 50 range, or a high setting of 140 would mean that the patient would unnecessarily get alarms when glucose was in the normal range of 120 or multiple alarms when the glucose was in the 160 range. NOT RECOMMENDED: Low 70 (4) NOT RECOMMENDED: High 140 (8)

Special Alert Settings Young children (higher, wider range) Hypoglycemia unawareness, high- risk professions (higher hypo setting) Pregnancy (lower, narrower range) HbA1c of 11.0% (higher initially)

The Numbers: Ballpark Estimates +/- 20% if >80 (4.4) +/- 20 mg/dl if <80 (+/- 1 mmol/l if < 4.4) Third, REAL-Time Trend Arrows. One arrow up or down indicates a change of 1-2 mg/dl/min in the last 20 minutes. Two arrows up or down indicate a change of more than 2 mg/dl/min in the last 20 minutes. Arrows are only displayed if there has been at least a 1 mg/dl/min change in the last 20 minutes. Patients should think about where glucose may be in 20 minutes if it continues at the same rate. 22

Can The Numbers Be Trusted? Not during first 1-2 cycles of using the system Not during the first 12-24 hrs after sensor insertion If BG Stable If Recent calibrations in-line If No recent alarms Third, REAL-Time Trend Arrows. One arrow up or down indicates a change of 1-2 mg/dl/min in the last 20 minutes. Two arrows up or down indicate a change of more than 2 mg/dl/min in the last 20 minutes. Arrows are only displayed if there has been at least a 1 mg/dl/min change in the last 20 minutes. Patients should think about where glucose may be in 20 minutes if it continues at the same rate. 23

Specific Insights to Derive (a purely retrospective journey)

CGM Data Analysis Tools Hardware/Software Medtronic: Internet Access to Carelink Carelink USB Adapter Dexcom*: PC, DM3 Software Connector Cable Color Printer * Dexcom 7+ System not FDA approved for use in children under age 18 in the U.S.

Before You Analyze, Qualify. Were sufficient calibrations performed? Did the calibrations match the CGM data reasonably well? Was the data mostly continuous? Was the time/date set correctly?

Before You Analyze, Qualify. MAD = 28% gaps & inaccuracy Abnormal Artifact

Objectives-Based Analysis Are bolus amounts appropriate? Meal doses Correction doses How long do boluses work? What is the magnitude of postprandial spikes? Is basal insulin holding BG steady?

Objectives-Based Analysis Are asymptomatic lows occurring? Are there rebounds from lows? Are lows being over/under treated? How does exercise affect BG? Immediate Delayed effects Is amylin/GLP-1 doing the job?

Objectives-Based Analysis How do various lifestyle events affect BG? Hi-Fat meals Unusual foods Stress Illness Work/School Sex Alcohol

Reports to Focus on Summary Statistics Modal Day / Overlay Graphs Individual Day Details                                                 

These Are a Few of My Favorite Stats… Mean (avg) glucose % Of Time Above, Below, Within Target Range Standard Deviation # Of High & Low Excursions Per Week

(the “retrospective journey”) Case Examples (the “retrospective journey”)

Case Study 1: The “Dark Side of the Moon” Type 2; using glargine and metformin Fasting readings OK; HbA1c elevated Glucose (mg/dL) 400 300 200 100 3 AM 6 AM 9 AM 12 PM 3 PM 6 PM 9 PM BG rising & staying high after meals. Consider meglitinide, exenatide, mealtime bolus insulin

Case Study 2a: Fine-Tuning Meal/Correction Boluses 34-y.o. pump user Glucose (mg/dL) 400 300 200 100 3 AM 6 AM 9 AM 12 PM 3 PM 6 PM 9 PM Breakfast and lunch doses may be too low Dinner dose appears OK Night-snack dose clearly insufficient

Case Study 2b: Fine-Tuning Meal/Correction Boluses 5-year-old on MDI; levemir BID. Dropping low 2-3 hours after dinner. Consider decreasing dinner bolus.

Case Study 2c: Fine-Tuning Meal/Correction Boluses Teenager on a pump; stays up late. BG Rising 9pm-1am. Consider structured night snacks with increased bolus amount.

Case Study 2d: Fine-Tuning Meal/Correction Boluses Pumper, dropping low after correcting for highs during the night Corr. Bolus  Consider increasing nighttime correction factor / insulin sensitivity

Case Study 3a: Postprandial Analysis Young adult on MDI. HbA1c are higher than expected based on SMBG Tired and lethargic after meals Significant postprandial spikes (300s). Consider pramlintide before meals. Glucose (mg/dL) 400 300 200 100 Meal

Case Study 3b: Postprandial Analysis Pump user, usually bolusing right before eating. Potatoes w/dinner most nights. Spiking primarily after dinner. Consider lower g.i. food or pre-bolusing.

Case Study 3c: Postprandial Analysis Pump user, 6 months pregnant Pre-bolusing (15-20 min) at most meals. Spiking primarily after breakfast. Consider “splitting” breakfast or walking post-bkfst.

Case Study 4a: Basal Insulin Regulation Pump user, 6 months pregnant Generally not eating (or bolusing) after 8pm. BG rising 1am-6am. Consider raising basal insulin 12am-5am.

Case Study 4b: Basal Insulin Regulation Type 1 diabetes; using insulin glargine & MDI History of morning lows Snacking at night and not “covering” w/bolus Glucose (mg/dL) 400 300 200 100 3 AM 6 AM 9 AM 12 PM 3 PM 6 PM 9 PM Basal dose is likely too high. Consider reducing.

Case Study 4c: Basal Insulin Regulation Pump user, frequent lows before breakfast and dinner. Glucose (mg/dL) 400 300 200 100 BG dropping after bolus action completed. Consider reducing basal rates early morning & late afternoon.

Case Study 5: Determination of Insulin Action Curve 12am 3am 6am 3-Hour Duration 4-Hour Duration 5-Hour Duration

Case Study 6: Detection of Silent Hypoglycemia Type1 college student; on pump Frequent fasting highs (9-10 AM). Wanted to raise overnight basal rates. Dropping & rebounding during the night. Consider decreasing basal in early part of night.

Case Study 7: Effectiveness of Pramlintide/Exenatide 15 mcg pramlintide 60 mcg pramlintide

Case Study 8: Response Curve to Different Food Types Cereal Oatmeal Yogurt Postprandial peak: cereal > oatmeal > yogurt

Case Study 9: Immediate Responses to Unusual Events Type 1 diabetes; pump user 40 years old; athletic Handsome, excellent speaker Glucose (mg/dL) 400 300 200 100 Late for meeting Gets flat tire; eats 15g carbs to prepare for tire change Spare is flat too!! 9 AM 12 PM 3 PM 6 PM 9 PM STRESS CAN RAISE BLOOD GLUCOSE… A LOT!!!

Case Study 10a: Delayed Effects Pump user Basal rates confirmed overnight “yellow” night: light cardio workout prior evening “Red” night: Lifting & cardio workout prior evening Experiencing delayed-onset hypoglycemia from heavy workouts. Consider temp basal reduction.

Case Study 10b: Delayed Effects Pump user Normal fasting readings during the week, but high on weekends Saturday Nights, Dinner Out Delayed rise from high-fat meals. Consider using temp basal increase.

Your Turn! What conclusions might you draw? What recommendations would you give? 52

Your Turn! What conclusions might you draw? In clinics that have utilized the MiniMed Paradigm REAL-Time System in studies, certain experiences can be gained. An analysis of patients who did well experienced benefits from using it continuously or at least 90 percent of the time. Patients who have experience with insulin pump therapy seemed to benefit more than those who were new to pump therapy and therefore still learning how to optimize their insulin pump usage. Patients who looked at it frequently, almost 10-20 times per day, but did not overreact to the data. Patients who had their CareLink data examined by the healthcare team, and with the help of the healthcare team, looked at trends and patterns and responses of glucose to insulin and made changes in carbohydrate to insulin ratios and basal rates. Patients who used the Bolus Wizard calculator did best and avoided hypoglycemia. It can also be said to the contrary that a lot has been learned from patients who do not benefit as much of the time: patients who did not wear the sensor at least 50 percent of the time; patients who did not trust the readings, and therefore did not look at the sensor; and patients who became extremely frustrated with nuisance alarms, causing them to stop using the sensor. So patient behavior during wearing CGM is also a key factor to achieve a better outcome. What conclusions might you draw? What recommendations would you give? 53

Optimizing CGM System Performance Calibration Sensor & Site Care Signal reception Ingredients for success 54

Optimal Calibration Calibrate at times when blood glucose (BG) is stable (fasting, pre-meals)* Avoid calibrations during times of rapid glucose change* Post meal UP or DOWN arrows are displayed In the period following a correction with food or insulin During exercise * Not required w/Dexcom system The best time to calibrate is when the blood glucose (BG) is stable — before meals or at bedtime—for then it is less likely that rapid changes in BG are going to take place. NOTE: When the BG is rapidly changing and a patient attempts to calibrate with a “single point in time” BG fingerstick reading, there may be enough of a change taking place to cause a calibration error to occur. If 2 calibration errors occur in a row, the glucose sensor can no longer be used and must be replaced. For this reason, patients should try not to calibrate within 2 hours after a meal or when 2 UP or DOWN arrows are displayed beside the current glucose reading on the insulin pump screen. The continuous glucose monitoring system must be calibrated a minimum of every 12 hours. However, 3 to 4 times a day is preferred. If calibration is not done every 12 hours, the insulin pump screen will stop displaying sensor glucose readings until a fingerstick BG value is entered Remind your patient to calibrate before going to sleep to prevent a METER BG NOW alarm from sounding during sleep hours

Calibration Calibrate before bedtime to avoid alarms during the night Use good SMBG technique Proper coding Clean hands Sufficient blood sample Fresh strips USE FINGERSTICKS Enter the calibration immediately after the fingerstick (Dexcom, Medtronic systems) The best time to calibrate is when the blood glucose (BG) is stable — before meals or at bedtime—for then it is less likely that rapid changes in BG are going to take place. NOTE: When the BG is rapidly changing and a patient attempts to calibrate with a “single point in time” BG fingerstick reading, there may be enough of a change taking place to cause a calibration error to occur. If 2 calibration errors occur in a row, the glucose sensor can no longer be used and must be replaced. For this reason, patients should try not to calibrate within 2 hours after a meal or when 2 UP or DOWN arrows are displayed beside the current glucose reading on the insulin pump screen. The continuous glucose monitoring system must be calibrated a minimum of every 12 hours. However, 3 to 4 times a day is preferred. If calibration is not done every 12 hours, the insulin pump screen will stop displaying sensor glucose readings until a fingerstick BG value is entered Remind your patient to calibrate before going to sleep to prevent a METER BG NOW alarm from sounding during sleep hours 56

The Sensors Storage Site Selection Refrigeration preferred (but not required) OK to use 1-2 months past expiration Site Selection “Fleshy” areas At least 2” Away from insulin infusion Avoid tight clothing areas, scars, bruises, lipoatrophy Rotate sites

The Sensors Timing Bleeding/Irritation Allow adequate “wetting” time (Medtronic) Put sensor in the night before connecting the transmitter (Medtronic) Bleeding/Irritation Slight bleeding OK Profuse bleeding: remove Remove introducer needle at proper angle

The Sensors Adhesive Site Irritation Completely cover the Transmitter & Sensor (Navigator & Medtronic systems) Check sensor daily for loose tape Apply extra tape over sensor & transmitter if tape patch begins to “curl” around edges Site Irritation Watch for redness, swelling, tenderness Remove sensor with prolonged irritation (>1 hour)

Signal Reception Heed transmitter ranges Medtronic: 6 ft. Dexcom: 5 ft. Navigator: 10 ft. Signals do not travel well through water Wear receiver on same side of body as sensor Keep receiver very close while charging (Dexcom) Charge transmitter fully every 6 days (Medtronic)                     In clinics that have utilized the MiniMed Paradigm REAL-Time System in studies, certain experiences can be gained. An analysis of patients who did well experienced benefits from using it continuously or at least 90 percent of the time. Patients who have experience with insulin pump therapy seemed to benefit more than those who were new to pump therapy and therefore still learning how to optimize their insulin pump usage. Patients who looked at it frequently, almost 10-20 times per day, but did not overreact to the data. Patients who had their CareLink data examined by the healthcare team, and with the help of the healthcare team, looked at trends and patterns and responses of glucose to insulin and made changes in carbohydrate to insulin ratios and basal rates. Patients who used the Bolus Wizard calculator did best and avoided hypoglycemia. It can also be said to the contrary that a lot has been learned from patients who do not benefit as much of the time: patients who did not wear the sensor at least 50 percent of the time; patients who did not trust the readings, and therefore did not look at the sensor; and patients who became extremely frustrated with nuisance alarms, causing them to stop using the sensor. So patient behavior during wearing CGM is also a key factor to achieve a better outcome.

Ingredients For Success Have the right expectations Wear the CGM at least 90% of the time Look at the monitor 10-20 times per day Do not over-react to the data; take IOB into account Adjust your therapy based on trends/patterns Calibrate appropriately Minimize “nuisance” alarms In clinics that have utilized the MiniMed Paradigm REAL-Time System in studies, certain experiences can be gained. An analysis of patients who did well experienced benefits from using it continuously or at least 90 percent of the time. Patients who have experience with insulin pump therapy seemed to benefit more than those who were new to pump therapy and therefore still learning how to optimize their insulin pump usage. Patients who looked at it frequently, almost 10-20 times per day, but did not overreact to the data. Patients who had their CareLink data examined by the healthcare team, and with the help of the healthcare team, looked at trends and patterns and responses of glucose to insulin and made changes in carbohydrate to insulin ratios and basal rates. Patients who used the Bolus Wizard calculator did best and avoided hypoglycemia. It can also be said to the contrary that a lot has been learned from patients who do not benefit as much of the time: patients who did not wear the sensor at least 50 percent of the time; patients who did not trust the readings, and therefore did not look at the sensor; and patients who became extremely frustrated with nuisance alarms, causing them to stop using the sensor. So patient behavior during wearing CGM is also a key factor to achieve a better outcome. 61

Questions?