Hypoglycemia is the most common complication of Type 1 diabetes particularly in those who strive for good glycemic control. In some patients there is a loss of awareness of hypoglycaemia so that the first manifestation of hypoglycaemia may be confusion or coma. Such a situation may carry risks that the patient may harm him or her self. Medtronic has recently developed a continuous glucose sensor that determines the glucose level every 3 minutes using a subcutaneous sensor and transmits the information to a remote device (Paradigm RT) that may be worn or left close by eg on a bedside table. The device may be programmed to alarm for a rapidly falling glucose or for low glucose levels.

We propose to use this in 16 patients with severe hypoglycemia as evidenced by a high HYPO score and see if we can decrease the number of hypoglycemic reactions and document this improvement with a better HYPO score.

Sixteen patients with Type 1 diabetes (as defined by: onset under the age of 25, lean at time of onset, continuous insulin use, and/or history of ketoacidosis and C-peptide negative), with a HYPO score over the seventy-fifth percentile (>423), will be approached and offered the study.

The study will be conducted over a three month time-frame. The first month will be a run-in period for participants to collect four weeks of glucose readings and information about their hypoglycemic events. These records will be used to calculate the baseline modified HYPO score. Patients will be screened during the first month for thyroid disease, celiac disease and Addison's disease.

During the second month they will meet with the study nurse for an intensive instruction period on the use of the Paradigm RT sensor. They will spend this month practicing to use the sensor and confirming they can use it adequately.

During the final month they will wear the sensor and collect glucose and hypoglycemia records for calculation of the final modified HYPO score.

The end point will be the change from the baseline modified HYPO score to the final four week HYPO score and this would be used to identify any improvement.

Inclusion Criteria:

• Type 1 Diabetes Mellitus confirmed by C-peptide <0.10 nmol/L
• age 18 - 70 yrs.
• HYPO score >423
• normal TSH, serum cortisol and anti-transglutaminase (ATTG)

Exclusion Criteria:

• current diagnosis of cancer
• planning a pregnancy
• Inability to give informed consent