• Time in the glycaemic target range (5.5-7.0 mmol/l) during the period of observation of 48 hours (expected to be longer in the intensive insulin group) [ Time Frame: 48 h ] [ Designated as safety issue: No ]
  

• Time to reach the target range (expected to be shorter in the intensive insulin group). [ Time Frame: 24 h ] [ Designated as safety issue: No ]
  
• Frequency of hypoglycaemia (plasma glucose < 3.8 mmol/l) (safety endpoint, expected to be similar in the two groups) [ Time Frame: during observation of 48 hours ] [ Designated as safety issue: Yes ]
  
• Frequency of severe hypoglycaemia (plasma glucose < 2.5 mmol/l) (safety endpoint, expected to be similar in the two groups) [ Time Frame: during observation of 48 hours ] [ Designated as safety issue: Yes ]
  
• Frequency of hypokalaemia (safety endpoint, expected to be similar in the two groups) [ Time Frame: during observation of 48 hours ] [ Designated as safety issue: Yes ]
  

BACKGROUND: Prospective randomized trials have shown that near-normoglycemic blood glucose control using insulin infusions achieves a significant reduction in mortality of severely ill patients in intensive care units, of patients with acute myocardial infarction and with stroke. This implies that most severely ill patients with hyperglycemia should be treated with insulin to reach near-normoglycemia. However, this is not common practice today in emergency room admissions outside the intensive care unit, and strategies to achieve near-normoglycemia safely outside the ICU setting with s.c. injections (insulin infusions are too risky outside the ICU) have not been established.

AIM: To evaluate an insulin therapy algorithm using s.c. injections which permits effective and safe glycemic management of emergency room patients with hyperglycemia.

DESIGN: Randomized, controlled trial with an open intervention. Patients presenting with hyperglycemia on admission to the emergency room are randomized 1:1 either to conventional treatment (conventional insulin group) or to intensive treatment (intensive insulin group).

METHODS: 140 patients admitted to the medical emergency rooms of the University Hospital Basel and the Regional Hospital of Solothurn will be included and randomized as described above. All patients with plasma glucose levels exceeding 8.0 mmol/l will be included.

Exclusion criteria include severely immunocompromised patients, patients in shock, patients with terminal illnesses on palliative care, type 1 diabetes with or without ketoacidosis and patients which require intensive care unit (ICU) or cardial care unit (CCU) therapy.

PRIMARY ENDPOINT: Time in the glycaemic target range (5.5-7.0 mmol/l) during the period of observation of 48 hours (expected to be longer in the intensive insulin group)

SECONDARY ENDPOINTS: Time to reach the target range. Frequency of hypoglycaemia (plasma glucose < 3.8 mmol/l). Frequency of severe hypoglycaemia (plasma glucose < 2.5 mmol/l. Frequency of hypokalaemia.