Insulin bolus in DKA, is it necessary?

Date First Published:
September 21, 2005
Last Updated:
September 22, 2005
Report by:
Russell A Duncan, Specialist Registrar (Ninewells Hospital, Dundee)
Search checked by:
Elspeth Pitt, Ninewells Hospital, Dundee
Three-Part Question:
In [diabetic ketoacidosis], does [an initial bolus of insulin result in faster, safer restoration of normoglycaemia and pH] when compared [with no bolus prior to commencing continuous insulin infusion]?
Clinical Scenario:
A 27 year old female presents to A&E, she is drowsy and complains of feeling thirsty and having diffuse abdominal pain. She is a known diabetic and has been vomiting for 2 days. She is tachycardic, clinically dehydrated and is breathing very deeply. Initial investigations confirm the presence of diabetic ketoacidosis. You start IV rehydration and an insulin infusion at 6 units per hour. A senior nurse states that another doctor usually gives a bolus of insulin prior to commencing the infusion. You wonder if this is necessary.
Search Strategy:
Medline 1966 to September Week 2 2005 and EMBASE 1980 to 2005 week 38 using the OVID interface. The Cochrane Library was also searched.
Search Details:
(insulin infusion systems/ OR insulin/ OR insulin.mp.) AND (injections, intravenous/ OR bolus.mp.) AND (diabetic ketoacidosis/ OR diabetic ketoacidosis.mp.). Limit to English language.
Outcome:
Altogether 63 papers in Medline and 156 papers in EMBASE were found using the above search. 2 of these were directly relevant. There was a large amount of overlap and both databases yielded the 2 relevant papers. Some other papers were review articles citing these 2 papers when commenting on the above question.
Relevant Paper(s):
Study Title Patient Group Study type (level of evidence) Outcomes Key results Study Weaknesses
Low dose insulin infusion in the treatment of diabetic ketoacidosis: bolus versus no bolus. Fort P, Waters SM, Lifshitz F. 1980 USA 20 episodes DKA in 19 children Case control trial Decline of plasma glucose in first hour. More rapid decline of plasma glucose in 1st hour. (p<0.05 by one-sided t test) No statistical difference in plasma glucose levels at 2 - 5 hours. Small numbers. Performed in children.
Time required to obtain normoglycemia Mean time required to normoglycemia was similar.
The use of insulin bolus in low-dose insulin infusion for paediatric diabetic ketoacidosis. Lindsay R, Bolte RG. 1989 USA 56 episodes DKA in 38 children. Randomised controlled trial Decline in serum glucose and changes in serum osmolality after 1 hour. There were no statistically significant differences in the outcomes of the 2 groups. Small group size. Performed in children.
Time required to reach serum glucose level <250 mg/dl. There were no statistically significant differences in the outcomes of the 2 groups.
Total duration of insulin infusion There were no statistically significant differences in the outcomes of the 2 groups.
Author Commentary:
Although most current texts do not advise an initial bolus of insulin prior to commencing a continuous infusion there are still some advocates of this treatment. [3]
The current evidence suggests that it is unnecessary to give this initial bolus in children and that it does not improve the outcome of DKA. There was no published evidence, directly answering the stated question in the adult population.
Bottom Line:
In DKA a continuous infusion of soluble insulin 0.1mg/kg should be commenced as soon as possible after diagnosis, the addition of an initial bolus is not supported by current published evidence.
References:
  1. Fort P, Waters SM, Lifshitz F.. Low dose insulin infusion in the treatment of diabetic ketoacidosis: bolus versus no bolus.
  2. Lindsay R, Bolte RG.. The use of insulin bolus in low-dose insulin infusion for paediatric diabetic ketoacidosis.
  3. Hardern RD, Quinn ND.. Emergency management of diabetic acidosis in adults.