Abstract Background: As part of a study to evaluate the accuracy of the GlucoWatch ® G2 TM Biographer and the Continuous Glucose Monitoring System (CGMS TM ) compared with gold standard blood glucose (BG) measurements in children with T1DM, hypoglycemia was induced with intravenous (IV) insulin. This provided the opportunity to study the physiologic response to IV insulin with regard to the rate of development of hypoglycemia. Objective: To determine the BG pattern of insulin-induced hypoglycemia in children with T1DM and assess its relationship to patient factors. Design/Methods: At five clinical centers, 54 subjects with T1DM [26F, 28M, mean age 12+3 yrs (range 7-17)] received an IV insulin bolus (0.05 units/kg when baseline BG mg/dL and 0.1 units/kg when baseline BG >100 mg/dL). After min, if needed, a second bolus ( units/kg) was administered (N=24). Oral or intravenous glucose was given when hypoglycemic levels were reached. Venous blood samples were obtained every 5 minutes for up to 90 minutes and glucose determinations were made at a central laboratory. Results: The BG ranged from 76 to 332 mg/dL at baseline and from 17 to 174 mg/dL at nadir. Among 48 subjects whose baseline BG was >100 mg/dL, BG decreased from baseline by at least 50 mg/dL 92% of the time, and BG decreased by at least 50% of the baseline level 85% of the time. A BG 50% drop from baseline, the mean time to the nadir BG was min for children mg/dL, 94% of pump users had a BG drop >50% compared with 80% of injection-users. There were no serious adverse effects of the protocol. Conclusions: With our protocol, a >50% decrease in BG was achieved by most subjects. The data provide useful information about the physiologic response to intravenous insulin in children with T1DM. In addition, the results will be valuable for the design of future physiological studies which aim to induce hypoglycemia through the administration of intravenous insulin.

INTRODUCTION The Diabetes Research in Children Network (DirecNet) is an NIH-funded collaborative study group that consists of five clinical centers, a coordinating center, a central laboratory and representatives from NICHD and NIDDK. DirecNet conducted a study to evaluate the accuracy of the GlucoWatch® G2™ Biographer (Cygnus, Inc., “GW2B”) and Continuous Glucose Monitoring System (“CGMS” Medtronic MiniMed, Northridge CA) compared with reference venous blood glucose (BG) measurements in children with T1DM ages As part of the study, hypoglycemia was induced with intravenous (IV) insulin. The protocol that was used to induce hypoglycemia and the glucose pattern response provides useful information for future research studies.

Objectives To determine the blood glucose pattern of insulin- induced hypoglycemia in children with T1DM. To provide future researchers a description of the protocol used to induce hypoglycemia. To assess factors related to the successful induction of hypoglycemia.

Design/Methods The purpose of this test was to assess sensor function during an acute fall in blood glucose levels to the mildly hypoglycemic range. Goal of test was to achieve serum glucose <55 mg/dL. Pretest Pretest blood glucose level needed to be >80 mg/dL. If < 80 mg/dL, juice or other carbohydrate given orally to raise glucose above this level before starting test. Insulin Dose units/kg body weight of regular insulin by IV bolus. After 30 min, 2nd dose given if target glucose (<55 mg/dL) not achieved.

Glucose Determinations Venous samples obtained every 5 min for up to 90 min. Samples analyzed at central lab at University of Minn. Bedside glucose monitoring performed for safety End of Test For subjects who did not reach glucose level <80 mg/dL, test ended after 90 min. For subjects whose glucose decreased to <80 mg/dL: Oral or iv glucose given at investigator discretion if blood glucose fell below 55 mg/dL. Test continued until glucose level was >80 mg/dL.

Clinical Characteristics of Test Subjects (N=54) Gender (male/female)52%/48% Age (mean ± SD)12.3 ± 3.4 years Race:White 85%, Hispanic/Latino 6%, African American 6%, Other 4% Duration of Diabetes (years) 5.6 ± 3.8 Insulin Treatment Modality Pump/Injections44%/56% HbA1c (mean ± SD)7.7% ± 1.3% BMI Percentile (mean ± SD)67% ± 23%

Bolus Injections BaselineInsulin# with 2 nd 2 nd Dose** Glucose*Dose**N Injection<.06 ≥.06 < 100 < ≥ <200< ≥ ≥ 200 < ≥ * mg/dL **units/kg

Baseline and Nadir Glucose Levels by Age and Treatment Modality Baseline Nadir Overall 7-<12 12-<18 CSII MDI Overall 7-<12 12-<18 CSII MDI AgeInsulin AgeInsulin

Change in Glucose from Baseline to Nadir % Meeting NChange (mg/dL) % ChangeTarget* Median 25 th -75 th %-iles Overall5496 (68, 164)66% (59%, 73%)57% Subject Age (years) 7-< (69, 159)65% (58%, 72%)57% 12-< (59, 173)67% (59%, 73%)58% Insulin Route Pump2490 (60, 140)66% (60%, 74%) 79% ** Injections3096 (75, 167)64% (55%, 72%)40% # Bolus Injections One3071 (58, 94)63% (58%, 67%)77% Two24164 (137, 187)71% (60%, 76%)33% *Target glucose was ≤ 55 mg/dL. **P-Value= 0.02, controlling for baseline glucose.

Nadir Glucose Value by Baseline Baseline Blood Glucose (mg/dL) Nadir Glucose (mg/dL) <100 N=6 100-<200 N=31 ≥200 N=17 ≤150100%88% ≤10097%65% ≤8597%59% ≤7094%47% ≤55100%71%18% ≤4083%26%6%

Comparison of Characteristics of Patients Meeting and Not Meeting Target of 55 mg/dL Did Not Met Target Meet Target (N=31)(N=23) Age (median) Pump User 61%22% Baseline Glucose (median)(mg/dL) 1 st Insulin Dose45%52% ≥ 0.06 units/Kg 2 nd Insulin Dose 26%70%

Overall One Bolus Injection Two Bolus Injections =200 =200 N=49 N=6 N=20 N=1 N=0 N=8 N=14 Rate of Change from Baseline to Nadir by Baseline Glucose (mg/dL) Restricted to subjects in whom hypoglycemia was successfully induced (decrease of 60mg/dL or 50% from baseline)

Adverse Effects Symptoms Pallor,sweating, shakiness 37% Confusion 4% Seizure, loss of consciousness 0

Representative tracings of sensor glycemic patterns during Insulin induced hypoglycemia

Representative tracings of sensor glycemic patterns during insulin induced hypoglycemia

Sensor Accuracy According to Rate of Glucose Change During Test *Calculated from the previous reference glucose value (drop in glucose divided by minutes between values). Rate of Change* (mg/dL per min)

Fulfillment of ISO Criteria* According to Rate of Glucose Change During Test * ISO criteria: for reference BG ≤ 75 mg/dL, SG within ± 15 mg/dL; for reference BG > 75 mg/dL, SG within 20%. Rate of Change (mg/dL per min)

Conclusions During the IV insulin test, a ≥50% decrease in BG was achieved by most subjects. However, only 57% of subjects achieved the target glucose level of 55 mg/dL. Target was reached by only 18% of subjects with baseline glucose >200 mg/dL. Target was reached more often by subjects using pumps than those using injections. Future protocols to induce hypoglycemia should pretreat with insulin to obtain a blood glucose <200 mg/dL prior to starting the test. GW2B and CGMS sensor glucose values were generally within 20% of the serum glucose and accuracy was not affected by rapid decreases in the blood glucose.

Barbara Davis Center H. Peter Chase Rosanna Fiallo-Scharer Jennifer Fisher University of Iowa Eva Tsalikian Michael Tansey Linda Larson Nemours Children’s Clinic Tim Wysocki Nelly Mauras Kristen Gagnon Stanford University Bruce Buckingham Darrell Wilson Jennifer Block Yale University William Tamborlane Stuart Weinzimer Elizabeth Boland Jaeb Center for Health Research Roy Beck Katrina Ruedy Craig Kollman Andrea Booth Gladys Bernett