Glycemic variability has been associated with mortality in hospitalized patients with hyperglycemia. However, it is unknown how modulation of glycemic variability would impact outcomes. One possibility is that glycemic variability could impact autonomic tone. In particular, heart rate variability (HRV) measurement is a sensitive marker for measuring autonomic tone, and aberrations in HRV have been associated with mortality. The current randomized pilot study will compare the effects of continuous intravenous (IV) insulin and subcutaneous basal bolus insulin on glycemic variability and autonomic tone in hospitalized non-critically ill patients with diabetes. Non-critically ill patients who are hyperglycemic or are requiring at least 20 units of insulin per day will be included. Patients with conditions that preclude accurate HRV readings (such as atrial fibrillation or paced rhythms) will be excluded. Patients randomized to intravenous insulin will receive the therapy for 24 hours according to our standard hospital guideline. Patients randomized to subcutaneous (SQ) insulin will receive basal bolus therapy using insulin analogues. All therapies will begin between 8 and 10 AM. Patients will undergo repeated heart rate variability recordings during the 24 hour period. Blood draws will be collected at baseline and at 24 hours for measurement of catecholamines, insulin, and c-peptide. Glycemic variability will be measured using a continuous subcutaneous glucose monitor and reported as coefficient of variation. The primary outcome measure is low frequency-to-high frequency power spectrum ratio of heart rate variability.

1. Glycemic variability is associated with unfavorable changes in autonomic tone, as assessed by low frequency (LF)/high frequency (HF) HRV ratio, independent of changes in overall glycemia.
2. Short-term increases in glycemic variability, followed by more prolonged glycemic stability are observed in generalized hospitalized patients treated with intravenous insulin compared to standardized basal bolus therapy. LF/HF HRV differs among subjects receiving intravenous compared to subcutaneous insulin.
3. Glycemic variability differs among subjects receiving intravenous compared to subcutaneous insulin

Inclusion Criteria:

• Age 18 and above
• Admitted to a general medicine or medicine subspecialty service
• Insulin use (>20 units/day) or hyperglycemia. Hyperglycemia is defined as glucose greater than 180 mg/dL on at least 2 occasions separated by at least 4 hours apart.

Exclusion Criteria:

• Type 1 diabetes
• Hospital stay expected less than 48 hours
• Inability to consent
• Pregnancy
• Prisoners
• Previous participation
• Autonomic neuropathy
• Conditions that preclude accurate heart rate variability monitoring: atrial fibrillation, frequent ectopy, congestive heart failure, paced heart rhythms
• Conditions which require lower dose insulin algorithms: end stage renal or liver disease

Since this was a pilot study no formal power analysis was possible. It was hypothesized that IV insulin would be associated with lower LF/HF HRV.