Article Text


Physiologic management of DKA
  1. G D Harris,
  2. I Fiordalisi
  1. Brody School of Medicine at East Carolina University, Greenville, North Carolina, USA
  1. Correspondence to
    Dr Harris;

Statistics from

Inward and Chambers provide a provocative description and discussion of the continuing confusion regarding the issues surrounding rehydration and treatment of the pediatric patient with diabetic ketoacidosis (DKA).1 They review some of the key issues that link fluid therapy to complications from brain swelling, and question the appropriateness of using a volume of fluid calculated by “maintenance plus deficit”, calling for a second revolution in the management of DKA. In the accompanying commentary, Edge makes several statements concerning fluid therapy in DKA, including that “DKA is associated with severe fluid losses”, that “any guidelines for fluid and electrolyte management must be simple to calculate”, that administration of base is a risk factor for intracranial complications, and that despite published data and “changes in protocols”, there is no evidence that the “incidence of cerebral oedema has changed over the past 20 years”.1

It is our opinion that the problem in the rehydration of the pediatric patient with DKA does not lie in assigning a maintenance fluid allotment. Rather, the source of error lies largely with failure to accurately estimate the volume of deficit and the tendency to automatically assume a severe degree of dehydration. From our experience with over 450 consecutive episodes of moderate and severe DKA, and our weight gain data,2 severe DKA (ie severe ketoacidemia) does not necessarily mean severe dehydration; the converse is also true.2,3 The degree of dehydration ranges from negligible (<1 %) to extreme (>20 %).3 Severe ketoacidemia, however, does cause vasoconstriction which may be manifested peripherally by cool, mottled skin, and Kussmaul breathing which leads to very dry oral mucosa. The striking appearance of a parched mouth and the presence of cool, even mottled skin without a critical assessment of vital signs and examination of distal (foot) pulses often results in an erroneous impression of shock and “severe dehydration”:3 A method for estimation of the volume of deficit was described in 19902 and we continue to use this approach successfully. Successful therapy requires not only gradual deficit replacement (evenly over 48 hours) but an accurate estimation of the volume of deficit along with critical monitoring of the clinical and biochemical response.3 If the deficit is assumed to be 10–15% but is actually only 3%, that patient will receive excess water independent of the more gradual timeframe and independent of the type of fluid given. Guidelines that have proposed “safe” limits to fluid volumes administered such as 4 litres/m2/day4 or 50 ml/kg body weight/4 hours5 violate the concept of the individualised assessment of the degree of dehydration and will invariably overhydrate the mild to moderately dehydrated child; the problem is compounded when actual body weight is used instead of ideal body weight in fluid calculations for the obese patient. On the other hand, certain patients, particularly those with complicating illness—for example, septic shock, pancreatitis—may require more than 20 ml/kg of fluid resuscitation in the first treatment hour and more than 50 ml/kg in the first four hours. Setting arbitrary fluid volume limits per hour or per day endanger particularly those patients at the mild and severe ends of the dehydration spectrum. Although the insult would be greater with hypotonic fluid, overhydration occurs readily with isotonic fluid as well when water requirements are overestimated.

DKA represents the effects of a complex disruption of normal metabolism, which leads to metabolic death if left untreated. Shock (decreased peripheral pulses, with or without hypotension), if present, should be corrected rapidly. Insulin should be given preferably by continuous, low dose, intravenous infusion, as soon as possible to begin correction of ketoacidemia/ketoacidosis. Regardless of the serum concentration of glucose, insulin is required to inhibit the hepatic fatty acyl carnitine cycle leading to ketoacid formation.6 A delay in insulin administration only serves to enhance and prolong ketoacidemia, thereby extending the period of time during which the patient remains vulnerable to central nervous system and other complications.

Our proposed management strategy may not satisfy the call for simplicity but it is an easily learned approach. It requires an understanding of relevant, known pathophysiology, the monitoring of serial physical examinations and laboratory studies with special attention to correction of acidemia and osmolality, and the anticipatory care that is inherent in the care of the critically ill.2,3,6–10,12 Physiologic management was first described between 198810 and 1990,2 and set forth with additional detail and data in 1994.3 It is rarely described in its complete form when referenced in texts; mere portions of our recommendations do not constitute what we have called physiologic management. Not only is it unlikely that large numbers of patients outside our own institution have been managed using our guidelines in their entirety, but the recommendations simply are not old enough to be reflected in data over the past 20 years. We suspect that physiologic management is significantly underrepresented in the multicentre studies conducted thus far, all of which compare variations of traditional therapy (empiric volume resuscitation whether or not shock is present, assumption of a large volume of deficit, planned rehydration in less than 48 hours with either 0.45% or 0.9% NaCl, with or without urinary output replacement). In a retrospective portion of our study in 19902 we compared these same therapies and also found that no form of traditional therapy minimised the risk of brain herniation during treatment.

Comments regarding the administration of base should be better defined. Rapid administration or “pushes” of hypertonic sodium bicarbonate should not be given. On the other hand, there is no evidence that administration of physiologic concentrations of base in the rehydration solution are either harmful or undesirable. In our experience, this practice mitigates the development of hyperchloremic acidosis during treatment.

As ours is a referral centre, most of our patients have some form of therapy initiated in outlying hospitals, sometimes in keeping with our recommended approach, and sometimes with our recommendations instituted only after initial telephone contact. In this setting, we have managed certain patients with severe DKA who received resuscitation fluids in excess of what their physical examination and laboratory data would dictate. It is not unusual for such patients to require as little as a typical maintenance allotment (without a deficit replacement component) for the remainder of therapy; some patients required fluid restriction to as little as two thirds the usual maintenance volume.

Our approach has been criticised because of the incidence of mannitol administration in our series.1 In our mannitol recipients, several of whom did not receive their initial management by us, there was no central nervous system morbidity or mortality. In another large series of patients there was a 50% failure rate of mannitol to reverse a deteriorating neurologic status, even when mannitol was given before respiratory arrest, with a near 100% failure rate when mannitol was given after respiratory arrest.11 It is possible that not all of our mannitol recipients actually had raised intracranial pressure. We believe, however, that the key to our good outcome is that the fluid and electrolyte therapy on which mannitol is superimposed is relevant to its success. It is erroneous to assume that the 100% success rate among our mannitol recipients would be reproducible in the setting of a therapy that violates the fundamental principles of rehydrating the hypertonic state of DKA.

Drs Inward and Chambers ask “do we have it right yet?” and convey concern that certain recommendations do not, as of yet, “have it right”. We agree.

Our work regarding the management of the pediatric patient in moderate to severe DKA has spanned 14 years1,12 and nearly 500 consecutive prospectively managed episodes. We remain available to participate in any endeavour to continue to improve the care of the paediatric patient in DKA.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.