Management of diabetic ketoacidosis

Although the mortality of diabetic ketoacidosis (DKA) has decreased substantially in the developed world, high mortality rates still prevail in South Africa, thus making this an important condition to recognise early and manage well. This review discusses the treatment of DKA, with emphasis on the controversial aspect of initial fluid replacement therapy. Current guidelines recommend the use of normal saline. The concern is that normal saline, when used in large volumes, leads to the development of a hyperchloraemic metabolic acidosis which is of uncertain clinical significance. This hyperchloraemic acidosis is better quantified using Stewart’s model, as opposed to the “traditional” Henderson-Hasselbalch equation. Ringer’s lactate is an alternative choice for initial fluid resuscitation, but may exacerbate the high lactate to pyruvate ratio in patients in DKA, and may cause hyperkaleamia. Insulin therapy, prevention of electrolyte abnormalities, and the replacement of bicarbonate and phosphate, are other important considerations in the management of the patient with DKA. Peer reviewed. (Submitted: 2011-02-14, Accepted: 2011-03-22) JEMDSA 2011;16(1):10-14 Review Article: Management of diabetic ketoacidosis Review Article: Management of diabetic ketoacidosis 11 2011 Volume 16 No 1 JEMDSA Isotonic versus hypotonic fluid Since the fluid lost in DKA is hypotonic, should our initial choice of fluid replacement be isotonic or hypotonic? Hypotonic solutions do not remain intravascular, and for that reason are not ideal for the purposes of initial resuscitation. However, they do manage to restore total body fluid losses with distribution to all three compartments. On the other hand, isotonic fluids are more efficient at restoring circulatory volume.6 For every litre of infused isotonic saline, a quarter normally remains in the circulatory volume.6 Because isotonic saline remains largely confined to the extracellular compartment, it does not provide free water to replace intracellular losses. Isotonic fluids are mainly distributed to the interstitial space, and if administered in excess, may lead to peripheral and pulmonary oedema when the interstitial compartment becomes overexpanded.6 Normal saline versus Ringer’s lactate Current guidelines from the USA and the UK recommend intravenous isotonic fluid replacement in the management of DKA. A recent consensus statement from the American Diabetes Association9 advocates the use of isotonic saline as the initial fluid therapy in the absence of cardiac compromise. The Association of British Clinical Diabetologists’ guideline10 agrees. In fact, early studies on DKA in the 1970s used normal saline,11 and its use continues to be advocated in modern textbooks.12 Over the past decade, a number of articles have compared two different approaches to interpreting acid-base disorders.13,14 The “traditional” approach uses the Henderson-Hasselbalch equation to describe and classify metabolic acidosis. A shortcoming of the traditional approach is that it doesn’t allow a distinction between the various possible causes of metabolic acidosis. The “modern” approach is Stewart’s physical chemical approach: a model more useful for quantifying acid-base disorders. In Stewart’s analysis, there are only three independent variables that determine pH: • Partial CO2 tension (PCO2). • Total concentration of weak acid (ATOT). In plasma, these consist of albumin and inorganic phosphate. • Strong ion difference (SID). The SID describes the difference between the concentrations of the strong cations (Na+, K+, Mg2+, and Ca2+) and strong anions (Cl-, lactate, ketoacids, sulphate and others) in a fluid compartment, and may be calculated using the complex equation below. Any change in pH must be because of a change in one or more of these independent variables, or in the dependent variables, such as hydrogen ion concentration (H+) and bicarbonate concentration (HCO3 -). Calculation of strong ion gap (SIG) SIG = [(Na+ + K+ + Ca2+ + Mg2+) – (Cl+ lactate-)] – (2.46 x 10-8 x PCO2/10 –pH + [albumin (g/dl)] x (0.123 x pH – 0.631) + [PO4(mmol/l) x (pH – 0.469)] The concern is that administration of large volumes of normal saline is associated with the development of a hyperchloraemic metabolic acidosis in the majority of patients.15 This acidosis may be more accurately described as a strong ion acidosis.16 A problem arises when clinicians use the base deficit in preference to the anion gap to document an improvement in the DKA. The base deficit, although an accurate measure of the total metabolic component of an acidosis, cannot differentiate between coexistent causes like ketosis and hyperchloraemia. Hyperchloraemia, if not explicitly recognised as giving rise to acidosis, may mask resolution of the ketoacidosis. The unwary clinician may erroneously interpret the low bicarbonate as being due to ongoing ketosis or hypovolaemia, and this may prompt an unnecessary alteration in therapy: either a change in insulin dose and/or increased fluid administration. This is where Stewart’s theory comes into play. The “traditional” HendersonHasselbalch model assumes that bicarbonate is an independent variable and is not influenced by chloride, hence it cannot satisfactorily explain the mechanism of the hyperchloraemic acidosis. However, Stewart’s strong ion approach provides a mechanistic explanation for hyperchloraemic acidosis by using a set of equilibria that describes the chemistry of plasma.17 Besides the development of hyperchloraemic acidosis, the administration of large volumes of normal saline may theoretically cause hyperkalaemia through an extracellular shift of K+ ions caused by acute changes in blood H+ ion concentration, which occurs secondary to the hyperchloraemic acidosis.15 A few small studies have suggested that normal saline administration may be detrimental to renal function.15 Table I: Composition of intravenous fluids, based on one-litre bags Normal saline Ringer’s lactate Na+ mmol/l 154 130 K+ mmol/l 0 4 Clmmol/l 154 109 Ca2+ mmol/l 0 2.7 HCO3 mmol/l 0 0 Lactate mmol/l 0 28 pH 5.0 6.5 Osmolality 308 273 Review Article: Management of diabetic ketoacidosis Review Article: Management of diabetic ketoacidosis 12 2011 Volume 16 No 1 JEMDSA What about the use of Ringer’s lactate in DKA? There are arguments against its use for several reasons. Patients in DKA have a high lactate to pyruvate ratio, and the 28 mmol/l of lactate in Ringer’s lactate could exacerbate this.18 Secondly, a litre of Ringer’s lactate contains 4 mmol of potassium, which may be life-threatening for a patient who is initially hyperkalaemic. Another consideration is the cost. One litre of Ringer’s lactate costs R33.72, as opposed to R12.77 for a litre of normal saline.19 Therefore, the question of which fluid replacement is optimal in DKA still remains unanswered. No randomised controlled trials are currently available to support the superiority of one fluid over another. Endocrinologists advocate normal saline, whereas critical care specialists argue against it due to the likelihood of saline causing a hyperchloraemic acidosis. Yet, there is no evidence in the literature that this is clinically significant or dangerous to patients. More prospective studies in this area are needed. How much fluid to replace? The average fluid deficit in an adult presenting in DKA is five-to-ten litres,20 or 100 ml/kg. The fluid deficit may be calculated using the following formula: Total body water deficit = (0.6 x body weight in kg) x (1-140/serum Na+)21 Corrected Na+ = serum Na+ + 1,6 x [(plasma glucose in mmol/l – 5.551)/5.551] Patients should receive 1-1.5 l of fluid in the first hour,9 and thereafter 250-500 ml per hour. The aim is to replace 50% of the fluid deficit during the first 12 hours of presentation, and the remainder within the next 12-16 hours.22 Since hyperglycaemia is corrected faster than ketoacidosis,23 dextrose-containing fluids should be used once the glucose falls to < 14 mmol/l to prevent hypoglycaemia.