Therapeutic challenges in the management of diabetic ketoacidosis

Diabetic ketoacidosis (DKA) is a severe and frequent life-threatening condition of diabetes. In March 2010, the Joint British Diabetes Societies (JBDS) published national UK guidance on the use of the fixed rate intravenous insulin infusion (FRIII) to manage DKA. ¹ This guidance was updated in September 2013. ² A summary of the guidelines pertinent to intensivists was published in 2014. ³ The national guidance was produced in an effort to prevent the pitfalls associated with recognition, resuscitation, and subsequent management of patients with DKA, and also to reduce the length of hospital stay. Among other things, the guidance recommends the use of bedside blood ketone meters to promote prompt recognition; the use of the FRIII to promote ketone clearance; glucose containing solutions to be administered once blood glucose is below 14 mmol/L; administration of potassium after initial resuscitation; and continuation of basal insulin.^1–3

A national survey conducted in the UK following the publication of the revised JBDS guidance on the management of DKA revealed that the incidence of biochemical abnormalities remains high. ⁴

The results of the national survey demonstrated that hospital acquired DKA was the third most common cause of DKA and accounted for 7.8% of all DKA episodes. In non-hospital acquired DKA, median time from hospital admission to a diagnosis of DKA was 35 min, 0.9% sodium chloride was commenced at a median time of 41 min, and the FRIII was commenced at a median time of 60 min; 67% of patients were hypokalaemic (serum potassium < 4 mmol/L) at 24 h; 28% of patients developed hypoglycaemia (blood glucose < 4 mmol/L) at some point during their treatment.

Furthermore, these results are in line with a respective study performed at one of largest teaching hospitals in the UK, in which 40% of patients suffered at least one episode of hypoglycaemia and 46% suffered hypokalaemia during the initial 24 h after presentation with DKA. ⁵

In 2012, the critical care unit of the West Suffolk Hospital (WSH) protocoled the management of DKA. The protocol essentially followed the JBDS guidelines including the administration of 10% glucose solution at blood glucose <14 mmol/L, but the protocol also demanded the administration of 20% glucose solution should the blood glucose fall below 6 mmol/L. In order to prevent hyperchloraemic metabolic acidosis (HCMA), the critical care protocol recommends the use of Hartmann’s solution with additional potassium chloride, rather than 0.9% saline with pre-mixed potassium chloride. This is possible as the critical care is exempt from NPSA guidance on storing strong potassium. ⁶

Retrospective case note review of the patients admitted to the critical care unit at the WSH was performed for patients admitted from October 2013 to April 2015. Twenty patients were identified that fulfilled the diagnostic criteria for DKA (capillary ketones > 3 mmol/L; blood glucose > 11 mmol/L and pH < 7.3).

In our cohort none of the patients had hospital-acquired DKA, with 19 being diagnosed promptly in the emergency department. In the emergency department, the median time to initiate fluid therapy was 41 min and the time to commence intravenous insulin therapy was 60 min, which was similar to the reported results of the national survey. The median time from hospital admission to critical care admission was 3.5 h. The time to resolution of DKA from diagnosis was less than 24 h in 85% of patients; 15% (3/20) patients developed hypoglycaemia during the treatment; 20% (4/20) had pre-existing hypokalaemia before admission to critical care; 24 h after both critical care and hospital admission 65% (13/20) were hypokalaemic.

Episodes of hypoglycaemia are independent predictors of increased length of hospital stay, and mortality.^7–9 The JBDS DKA guidelines recommend the introduction of 10% dextrose solution when the blood glucose is less than 14 mmol/L, but stop short of recommending the rate of its administration or introduction of higher concentration of dextrose. The consequences were evident in the national survey with 28% patients developing hypoglycaemia during the treatment, while at the teaching hospital, it was 40%. In comparison, the guidelines in our critical care recommend the introduction of 10% dextrose infusion at 100 ml/h when the blood glucose is less than 14 mmol/L and introduction of 20% dextrose if the blood glucose inadvertently falls below 6 mmol/L. Consequently, only 15% of our DKA patients experienced any episodes of hypoglycaemia during their treatment for DKA. We therefore argue that the JBDS guidelines should be modified and suggest protocoling management of hypoglycaemia associated with the FRIII (Figure 1). Furthermore, the national guidance suggests continuation of the FRIII until ketone levels are < 0.6 mmol/L, while the level for diagnosis is >3 mmol/L. As it is the high dose FRIII that is driving the hypoglycaemia, we now suggest that the FRIII should be converted to the variable rate intravenous insulin infusion (VRIII) at the ketone level of 2 mmol/L.

Figure 1.

West Suffolk Hospital’s modified critical care guidance for treatment of DKA.

Hypokalaemia, like hypoglycaemia, is a direct consequence of the insulin treatment; 65% of our patients were hypokalaemic at 24 h, which is similar to the national survey data (67%). This is not surprising as we used the same potassium replacement rate as the national guidance, but instead of administering the potassium in premixed bags of saline, 40 mmol of KCl was added to 1 l bags of Hartmann’s solution. Subsequently, we will now administer 3% KCl independently of the resuscitation fluid, and titrate it accordingly (Figure 1). This will also facilitate the critical care unit to be compliant with NPSA advice on storage of strong potassium, even though critical care units are excluded. ⁶

In summary, our findings add further evidence that the use of the FRIII is associated with hypoglycaemia and it is now necessary to review mechanisms to improve the safety of the FRIII. Furthermore, current potassium replacement strategies are insufficient to prevent hypokalaemia. We therefore suggest that the next edition of the JBDS guidelines on DKA consider these findings, and in the meantime local policies should be updated.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: NL was the lead anaesthetic author of the JBDS guidelines on The Management of Diabetic Ketoacidosis in Adults. Second edition. London.

JS, CP, AS: declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.