This study will test the hypothesis that a wearable automated bionic pancreas system that automatically delivers both insulin and glucagon can improved glycemic control vs. usual in the outpatient environment.

The study population will be volunteers with type 1 diabetes who are 21 years of age or older. The setting will be an outpatient environment in a three square-mile area on the Boston Peninsula (see Appendix A). Volunteers will stay in a hotel adjacent to the MGH campus at night and will have free activity during the day within the specified geographic area. They will determine the timing and nature of their meals from local restaurants, with food brought from home, or food kept in their hotel room, which will have a small refrigerator. They will have the opportunity to exercise as they wish in their choice of two gyms. There will be minimal scheduling constraints, limited only by an 11:00 PM curfew and morning departure time from the hotel of no earlier than 7:00 AM. They will be able to work if they wish as long as that can be done within the geographical constraints (e.g. if they work in the downtown Boston area or can "work from home" or have meetings at a conference room in the hotel or at a restaurant). During the entire experiment they will be closely monitored by study staff (RN, NP, or MD) around the clock. They will remain within direct line of sight and no more than a short distance away from study staff during the daytime for safety. During the night they will be continuously monitored via BG telemetry from a nearby hotel room and study staff will be able to enter their rooms quickly, should that become necessary. During the night, when volunteers will remain in their rooms, one study staff member will monitor up to two volunteers at a time.

Capillary BG will be tested every two hours during the day using a highly accurate, laboratory equivalent meter (HemoCue, selected for maximum data integrity) and venous BG will be tested every 30 minutes overnight using an autosampling device (GlucoScout). Continuous glucose monitoring (Dexcom G4) will be done throughout the study period. Photos and menu information, if available, will be documented for each meal and snack by the escort and estimates of carbohydrate intake will be estimated later from this information by a nutritionist. The type and level of activity being performed by the volunteers (e.g. lying, sitting, standing, walking, running) will be documented in 15 minute intervals by the study staff escort. Additional data will be collected using an accelerometer. During exercise, the type and duration of exercise, and, depending on the kind of activity, the heart rate (recorded using a Polar heart rate monitor) will be documented every 15 minutes and point-of-care blood glucose will be documented every 30 minutes.

Inclusion Criteria:

• Age 21 years or older with type 1 diabetes for at least one year
• Stimulated C-peptide < 0.1 nmol/L at 90 minutes after liquid mixed meal by the DCCT protocol
• Diabetes managed using an insulin infusion pump and rapid- or very-rapid-acting insulins including insulin aspart (NovoLog), insulin lispro (Humalog), and insulin glulisine (Apidra) for at least three months prior to enrollment
• Otherwise healthy (mild chronic disease such as asthma, hypertension, and depression will be allowed if well controlled)

Exclusion Criteria:

• Unable to provide informed consent
• Unable to comply with study procedures
• Total daily dose (TDD) of insulin that is > 1.5 U/kg
• Pregnancy (positive urine HCG), breast feeding, plan to become pregnant in the immediate future, or sexually active without use of contraception.
• Hypoglycemia unawareness (self-reported lack of hypoglycemia symptoms when BG is < 50 mg/dl)
• End stage renal disease on dialysis (hemodialysis or peritoneal dialysis).
• Any known history of coronary artery disease (CAD)
• Abnormal EKG suggestive of coronary artery disease or increased risk of malignant arrhythmia
• Congestive heart failure (established history of CHF, paroxysmal nocturnal dyspnea, or orthopnea).
• History of TIA or stroke.
• History of pheochromocytoma. Fractionated metanephrines will be tested in patients with history increasing the risk for a catecholamine secreting tumor
• Untreated or inadequately treated mental illness
• Current alcohol abuse or substance abuse
• Electrically powered implants (e.g. cochlear implants, neurostimulators) that might be susceptible to RF interference.
• Use non-insulin, injectable anti-diabetic medications or oral anti-diabetic medications
• History of adverse reaction to glucagon (including allergy) besides nausea and vomiting
• Unwilling or unable to completely avoid acetaminophen
• ALT > 3-fold upper limit of normal
• Albumin < 3 g/dl
• Body mass index less than18 or greater than 35

12 days

Adverse events were collected for the Bionic Pancreas period, and were not collected during the usual care period.