PHN

Dduoarlhdor

lcTloepdwpo

woiDyosefOiw

F

T

1d

ediatric Diabetic Ketoacidosis andyperglycemic Hyperosmolar State

icole Glaser, MD

ocptqf

tcdc

PDTmmpaimcpt

PtIiThehpccccd

iabetic ketoacidosis (DKA) is an important complicationof childhood diabetes mellitus and the most frequent

iabetes-related cause of death in children.1,2 In various pop-lation-based studies, reported rates of DKA at presentationf type 1 diabetes have ranged from as low as 15% to as highs 83%,3–7 with most North American and European studieseporting rates of approximately 40%. Although DKA occursess frequently in children with type 2 diabetes, case seriesave documented frequencies of DKA at diagnosis of type 2iabetes in children ranging from 6% to 33%.8–11 A diagnosisf type 2 diabetes cannot be excluded based on the occur-ence of DKA.

Young children with new onset of type 1 diabetes are moreikely to present with DKA,4,6,12 as are children who reside inountries with a low overall prevalence of type 1 diabetes.5

he higher frequency of DKA at presentation in these groupsikely reflects the greater difficulty in recognizing symptomsf diabetes in these populations. In a European study, anducational program directed at parents and primary careediatricians was shown to decrease the frequency of DKA atiagnosis of type 1 diabetes from almost 80% to just 12.5%,hich supported the concept that the frequency of DKA atresentation of diabetes is related to recognition of symptomsf diabetes in the population studied.7

In children who have established diabetes, DKA may occurith episodes of infection or other illnesses or with insulinmission or malfunction of diabetes care equipment, such asnsulin pumps. In children who have established diabetes,KA occurs at a rate of approximately 1% to 8% perear.4,13–15 DKA in patients who have established diabetesccurs more frequently in persons with lower socioeconomictatus, lack of adequate health insurance, higher HbA1c lev-ls, and psychiatric disorders.13 Insulin omission is the mostrequent cause of DKA in children who have known diabetes.ne study investigated the frequency of viral and bacterial

nfections in children who have DKA. Among all childrenho presented with DKA, bacterial infections were present in

rom the Department of Pediatrics, University of California Davis, School ofMedicine, 2516 Stockton Boulevard, Sacramento, CA 95817.

his article is reprinted from Pediatr Clin N Am 52 (2005) 1611-1635.

(E-mail: nsglaser@ucdavis.edu

071-9091/06/$-see front matter © 2006 Elsevier Inc. All rights reserved.oi:10.1016/j.spen.2006.01.003

nly 13% and viral infections in 18%.16 In the subgroup ofhildren who have known diabetes, bacterial infections wereresent in 17% and viral infections in 20%. These data con-rast with data for adult populations, in which higher fre-uencies of infection or other illnesses as precipitating factorsor DKA have been reported.17,18

Although the risk of mortality from childhood DKA is lesshan 0.5%, DKA is still the most frequent diabetes-relatedause of death in children.1,2 Most of these DKA-relatedeaths are caused by cerebral edema (62%–87%), a compli-ation that is discussed in more detail later.

athophysiology ofiabetic Ketoacidosis

he physiologic abnormalities in patients who have DKAay be viewed as an exaggeration of the normal physiologicechanisms responsible for maintaining adequate fuel sup-ly to the brain and other tissues during periods of fastingnd physiologic stress. The relative concentration of insulinn relation to glucagon and other counterregulatory hor-

ones or stress hormones (eg, epinephrine, norepinephrine,ortisol, and growth hormone) primarily mediates thesehysiologic abnormalities rather than the absolute concen-ration of insulin itself.19,20

athophysiologic Abnormalities Early inhe Development of Ddiabetic Ketoacidosisn a child who has new onset of type 1 diabetes, decliningnsulin production lowers the ratio of insulin to glucagon.his decrease in relative insulin concentration leads to excessepatic glucose production (Fig. 1A). Early in the course ofvolving DKA, when levels of epinephrine and other stressormones are normal or minimally elevated, increased he-atic glucose output is mainly caused by stimulation of gly-ogenolysis, with a smaller contribution from increased glu-oneogenesis.21–23 Low serum insulin concentrations alsoontribute to hyperglycemia by decreasing peripheral glu-ose uptake in muscle and adipose tissue. This effect is me-iated by diminished translocation of glucose transporter

GLUT)4 glucose transporters to the cell membrane.24,25 In-

187

ccgmsui

ftdsoictcshthh

krt

Fcgiapstcnfgecmtcnprghcose uptake and augment gluconeogenesis.

FiA(muotpmCCibigis

188 N. Glaser

reased hepatic glucose output and decreased peripheral glu-ose use contribute to hyperglycemia.26 When the serumlucose concentration rises above approximately 180 to 200g/dL, which exceeds the renal threshold for glucose reab-

orption,27,28 osmotic diuresis results, with an increase inrine output. Fluid losses then stimulate compensatory oral

ntake of fluids, which leads to polydipsia.Low insulin concentrations also stimulate the release of

ree fatty acids (FFA) from adipose tissue by allowing activa-ion of hormone-sensitive lipase (Fig. 2). This increase in FFAelivery to the liver is necessary but not sufficient for thetimulation of ketone body formation.29 For ketogenesis toccur, activation of the hepatic �-oxidative enzyme sequences also necessary.20,30,31 It is mainly a further decline in insulinoncentration relative to glucagon that allows this activationo occur. A larger decline in insulin concentration relative toounterregulatory hormones is necessary to promote lipoly-is and ketogenesis, compared with that required to causeyperglycemia.32 These findings in part explain the lesserendency toward the development of DKA in patients whoave type 2 diabetes, despite the occurrence of substantialyperglycemia.Under fasting conditions in a normal individual, modest

etosis occurs, but marked ketoacidosis is prevented by di-ect ketone-induced stimulation of insulin, which limits fur-her release of FFAs from adipose tissue.33 In children who

igure 2 Decreased insulin concentrations result in increased activ-ty of hormone-sensitive lipase in adipose tissue with release of FFA.s concentrations of stress hormones (eg, cortisol, growth hormone

GH), catecholamines) increase later in the course of DKA, hor-one-sensitive lipase activity is further stimulated. FFAs are takenp by the liver, where they are esterified to fatty acyl-CoA. Transportf the CoA ester across the mitochondrial membrane for �-oxida-ion requires transesterification with carnitine, which is accom-lished by carnitine palmityl transferase 1 (CPT-1). Once inside theitochondria, esterification to carnitine is reversed, and fatty acyl-oA undergoes �-oxidation to form ketones (AcAc) and (�-OHB).PT-1 is regulated by the concentration of malonyl CoA, which

nhibits CPT-1 activity. Malonyl CoA is produced from acetyl-CoAy acetyl-CoA carboxylase (ACC), whose activity is increased by

nsulin and decreased by glucagon and �-adrenergic agents. Gluca-on also decreases the concentration of malonyl CoA by diminish-ng the rate of glycolysis and the rate of production of citrate, the

igure 1 (A) Early in the development of DKA, a decrease in theoncentration of insulin relative to glucagon results in stimulation oflycogenolysis by promoting conversion of glycogen synthase � tonactive glycogen synthase � and conversion of phosphorylase � toctive phosphorylase �. Gluconeogenesis is also stimulated butlays a lesser role in the increase in hepatic glucose output at thistage than does glycogenolysis. An increase in the ratio of glucagono insulin stimulates a decrease in fructose 2,6 bisphosphate con-entrations mediated by phosphorylation of 6-phosphofructo-2-ki-ase/ fructose-2,6-bisphosphatase. The decreased concentration of

ructose 2,6 bisphosphate inactivates the rate-limiting enzyme forlycolysis (6-phosphofructo-1-kinase) and stimulates gluconeogen-sis via activation of fructose-2,6-bisphosphatase. Decreased insulinoncentrations also result in a lower peripheral glucose uptake byuscle and adipose tissue with diminished transport of GLUT4 to

he cell membrane. (B) Later in the development of DKA, elevatedoncentrations of other counterregulatory hormones (eg, cortisol,oripinephrine, epinephrine, growth hormone) further increase he-atic glucose output and decrease peripheral glucose uptake. �-Ad-energic agonists enhance glycogenolysis and promote release ofluconeogenic substrate from muscle. Elevated cortisol and growthormone concentrations cause further declines in peripheral glu-

ubstrate for malonyl CoA production.

hir

PtPdh(iftfenndcgclCsdetppstccmalcl

etadcco

aciipctatt

cmrsnrasdTap

CoCwibiiscHascDTssDttmp

FStS

DKA and hyperglycemic hyperosmolar state 189

ave type 1 diabetes, however, this “hormonal brake” is lack-ng and ketone production proceeds unchecked, eventuallyesulting in acidosis with an elevated anion gap.

athophysiologic Abnormalities Later inhe Development of Diabetic Ketoacidosishysiologic stress caused by acidosis and progressive dehy-ration eventually stimulates release of the counterregulatoryormones, cortisol, catecholamines, and growth hormonesee Fig. 1B).26,34,35 Coexisting infection or other illness ornjury likewise can accelerate the development of ketosis viaurther elevations in counterregulatory hormone concentra-ions. Elevated cortisol concentrations augment FFA releaserom adipose tissue to fuel ketogenesis and decrease periph-ral glucose uptake via effects on insulin-dependent mecha-isms of glucose uptake and insulin-independent mecha-isms (Fig. 2).36–38 Increased epinephrine concentrationsirectly increase glycogenolysis and stimulate release of glu-oneogenic precursors from muscle, which allows gluconeo-ensis to make a more substantial contribution to hypergly-emia.22,23,39 Epinephrine and norepinephrine also stimulateipolysis and �-oxidation of FFAs to form ketone bodies.40,41

atecholamines also may inhibit insulin secretion directly viatimulation of �-adrenergic receptors and cause a furtherecline in serum insulin concentrations.42,43 Although thisffect is inconsequential in children who have longstandingype 1 diabetes (and absent or minimal endogenous insulinroduction), it may accelerate the development of DKA inatients with a new diagnosis of type 1 diabetes in whomome insulin-producing capacity remains, and it likely con-ributes more substantially to the development of DKA inhildren who have type 2 diabetes. Elevated growth hormoneoncentrations likewise contribute to worsening hyperglyce-ia, mainly via further decreasing peripheral glucose uptake,

nd enhance ketone production by increasing FFA re-ease.23,44 Growth hormone effects occur over a longer timeourse than those of other counterregulatory hormones thatead to more acute elevations in glucose and FFAs.

With the increase in hepatic glucose production, ketogen-sis, and peripheral insulin resistance stimulated by eleva-ions in counterregulatory hormone concentrations, acidosisnd dehydration worsen. These changes then accelerate theevelopment of DKA by stimulating further increases in theoncentrations of counterregulatory hormones. A vicious cy-le is created and is responsible for the eventual developmentf severe ketoacidosis.Other physiologic processes also contribute to worsening

cidosis and dehydration (Fig. 3). Intestinal ileus occurs as aonsequence of acidosis, potassium depletion, and dimin-shed splanchnic perfusion caused by dehydration. Intestinalleus causes abdominal pain and vomiting, which impairs aatient’s ability to compensate for osmotic diuresis by in-reased intake of fluids. More substantial dehydration even-ually leads to diminished tissue perfusion, which enhancescidosis via accumulation of lactic acid.45,46 Severe dehydra-ion eventually compromises renal function and diminishes

he capacity for clearance of glucose and ketones, which p

auses concentrations of both to rise further. Ongoing os-otic diuresis and ketonuria in the setting of acidosis also

esult in urinary losses of electrolytes, particularly potassium,odium, chloride, calcium, phosphate, and magnesium. Uri-ary losses of sodium and potassium as ketone salts mayesult in excess chloride retention, such that hyperchloremiccidosis is superimposed on the increased anion gap acido-is.47 Elevated aldosterone concentrations that result fromehydration also serve to further enhance potassium loss.46

ypical electrolyte deficits in patients who have DKA includepproximately 5 to 13 mmol/kg of sodium, 3 to 5 mmol/kg ofotassium, and 0.5 to 1.5 mmol/kg of phosphate.48,49

linical Manifestationsf Diabetic Ketoacidosis

lassic symptoms of DKA include polyuria, polydipsia,eight loss, abdominal pain, nausea, and vomiting. Abdom-

nal tenderness, absence of bowel sounds, and guarding maye present and may mimic the acute abdomen.50 Tachycardia

s frequent, and signs of hypoperfusion, such as delayed cap-llary refill time and cool extremities, are also common. Otherigns of dehydration also may be present, including dry mu-ous membranes, absence of tears, and poor skin turgor.ypothermia also has been described.51 Although profound

cidosis may depress myocardial contractility and vascularmooth muscle tone, the occurrence of these effects to alinically relevant degree has not been demonstrated inKA,52 and hypotension in children who have DKA is rare.achypnea occurs in response to metabolic acidosis as a re-ult of stimulation of chemoreceptors in the central nervousystem (CNS). Tachypnea may be extreme and may causeKA to be initially misdiagnosed as respiratory illness. Ace-

one (produced from nonenzymatic decarboxylation of ace-oacetate [AcAc]) typically causes a fruity breath odor, whichay be a helpful initial clue to the diagnosis of DKA. Despiterofound systemic acidosis, most children who have DKA

igure 3 Pathophysiology of diabetic ketoacidosis. (From Glaser NS,tyne DM. Endocrine disorders. In: Behrman R, Kliegman R, edi-ors. Nelson essentials of pediatrics. 3rd edition. Philadelphia: WBaunders; 1997; with permission.)

resent with normal mentation or only minimal depression

oswbh

LiHActtkclpidilipPmlmth5sitd

AC[bbafDpeiptihnticDr

svtmo

wvCatsf(

EHcs1mHss

aRefioDt

FtSs2

190 N. Glaser

f mental status. The lack of substantial neurologic depres-ion reflects the fact that brain pH in patients who presentith DKA is generally preserved within the normal rangeecause of the impermeability of the blood-brain barrier toydrogen ions.53,54

aboratory Abnormalitiesn Diabetic Ketoacidosisyperglycemiadiagnosis of DKA can be made when the serum glucose

oncentration is more than 200 mg/dL and venous pH is lesshan 7.30 (or the serum bicarbonate concentration is lesshan 15 mmol/L) in the presence of elevated urine or serumetone concentrations. DKA with near-normal glucose con-entrations also has been described but occurs infrequent-y.55–57 This euglycemic DKA may occur in pregnancy and inatients who have known diabetes who have administered

nsulin before coming to the emergency department. Chil-ren who have DKA who have prolonged vomiting and min-

mal oral intake before presentation also may present withower initial glucose concentrations. Much of the variabilityn serum glucose concentrations at presentation may be ex-lained by differences in hydration and nutritional status.35

rolonged fasting or poor nutrient intake before the develop-ent of DKA decreases substrate availability and results in

ower serum glucose concentrations at presentation, whereasore severe dehydration favors higher glucose concentra-

ions. In the absence of preexisting renal disease or unusuallyigh carbohydrate intake, blood glucose concentrations of00 to 600 mg/dL imply that dehydration is of sufficienteverity to diminish the glomerular filtration rate by approx-mately 30% to 40%. Blood glucose concentrations morehan 800 mg/dL suggest that the glomerular filtration rate isecreased by 50% or more.58

cidosisoncentrations of ketone bodies (beta-hydroxybutyrate

�OHB] and AcAc) are elevated in DKA. The serum bicar-onate concentration is low because bicarbonate is used as auffer against metabolic acidosis, which results in increasednion gap acidosis. Some degree of hyperchloremic acidosisrequently coexists with increased anion gap acidosis inKA,47 and the anion gap reflects the combination of theserocesses. Although concentrations of �OHB and AcAc arelevated in patients who have DKA, the ratio of �OHB:AcAcs increased during DKA as a result of changes in the redoxotential (NADH/NAD� ratio) in hepatic mitochondria.59 Al-hough the ratio of �OHB:AcAc is typically 1:1 in a normalndividual, this ratio rises to as high as 10:1 in persons whoave DKA. These changes are important mainly because theitroprusside reaction used to test urine ketone concentra-ions detects only AcAc and not �OHB. Although urine test-ng can be relied on to help diagnose DKA, the urine ketoneoncentration should not be relied on as an indication ofKA severity or treatment response, particularly because the

atio of �OHB:AcAc decreases during DKA treatment. Bed- f

ide blood ketone meters recently were developed and pro-ide a rapid means for accurately measuring �OHB ratherhan AcAc in children who have DKA.60 How these measure-ents might best be used to enhance diagnosis and treatment

f DKA, however, remains to be determined.Metabolic acidosis stimulates chemoreceptors in the CNS,

hich results in partial correction of the metabolic acidosisia hyperventilation and a decrease in the partial pressure ofO2. There is a linear relationship between serum bicarbon-te concentration and pCO2, and this relationship suggestshat end-tidal CO2 measurements may be used as a rapidcreen for acidosis in children who have suspected DKA or toollow the course of acidosis in children who have DKAFig. 4).61

lectrolyte Abnormalitiesyperglycemia results in fluid movement from the extravas-

ular to the intravascular space and a decrease in the serumodium concentration. This decrease can be calculated as a.6 mEq/L decrease in sodium concentration for every 100g/dL increase in serum glucose more than 100 mg/dL.62

yperlipidemia caused by lipolysis also may affect serumodium measurements and result in a decrease in measurederum sodium concentrations.63

Typically, serum potassium concentrations at presentationre in the high-normal range or even above the normal range.edistribution of potassium ions from the intracellular to thextracellular space in DKA results from a combination ofactors, including direct effects of low insulin concentrations,ntracellular protein and phosphate depletion, and bufferingf hydrogen ions in the intracellular fluid compartment.64

espite normal or elevated initial potassium concentrations,otal body potassium concentrations are depleted, often pro-

igure 4 End-tidal CO2 levels versus serum bicarbonate concentra-ions in children with diabetic ketoacidosis. (From Fearon DM,teele DW. End-tidal carbon dioxide predicts the presence andeverity of acidosis in children with diabetes. Acad Emerg Med002;9:1373–8; with permission.)

oundly, and serum potassium concentrations usually drop

rcbsc

OWwpntasaccaedobpa

4wvclft

TDFIsqflwflttahhmflutcd7ap

0crpuvs

ttcca

IIsotriargiidhbvtak

tWmissfsgbeaatslc2gl

EW

DKA and hyperglycemic hyperosmolar state 191

apidly with insulin treatment. The initial serum potassiumoncentration should not be taken as an indication of totalody potassium stores. Serum phosphate concentrations areimilarly elevated or normal at presentation but tend to de-rease during treatment.

ther Biochemical Abnormalitieshite blood cell counts are frequently elevated in childrenho have DKA, and the differential may be left shifted. Therecise mechanism responsible for leukocytosis in DKA isot fully understood, but elevated catecholamine concentra-ions may play a role.65,66 Another contributing factor may ben elevation in proinflammatory cytokines (eg, tumor necro-is factor-�, interleukin-6, interleukin-8, interleukin-1�)nd C-reactive protein caused by DKA.67–69 Cytokine con-entrations are substantially increased during DKA and de-rease promptly with the initiation of insulin therapy. C-re-ctive protein concentrations, although also frequentlylevated in patients who have DKA, show a less consistentecrease with treatment.68 Infection is infrequently the causef DKA in children,16 and an elevated or left-shifted whitelood cell count need not prompt a search for an infectiousrocess unless fever or other symptoms or signs of infectionre present.

Serum amylase or lipase concentrations are elevated in0% of children who have DKA and in 40% to 80% of adultsho have DKA.70–72 The cause and significance of these ele-ations, however, are not known. Clinical pancreatitis inhildren who have DKA is rare, and elevated amylase oripase concentrations need not prompt further investigationor pancreatitis unless abdominal pain persists after resolu-ion of ketosis.

reatment ofiabetic Ketoacidosis

luidsntravenous fluids (0.9% saline or other isotonic fluids)hould be administered as soon as possible to restore ade-uate perfusion and hemodynamic stability. An intravenousuid bolus of 10 to 20 mL/kg is often required. In patientsho are well perfused and hemodynamically stable, an initialuid bolus may not be necessary. A recent study indicatedhat physicians’ clinical assessments of the degree of dehydra-ion in children who have DKA correlate poorly with thectual percentage dehydration and often underestimate de-ydration severity.73 Difficulties in clinical estimation of de-ydration may result in part from osmotically mediated waterovement from the tissues to the intravascular space. Thisuid movement results in preservation of intravascular vol-me and may obscure some of the clinical signs of dehydra-ion. Because severity of dehydration is difficult to estimatelinically, it may be most appropriate to assume an averageegree of dehydration for most patients (approximately%–9% of body weight73,74). This estimated fluid deficit,long with maintenance fluid requirements, should be re-

laced evenly over a 36- to 48-hour period using 0.45% to s

.9% saline. Because the serum glucose concentration typi-ally decreases to levels near the renal threshold for glucoseeabsorption within a few hours of initiating treatment, re-lacement of ongoing fluid losses from osmotic diuresis issually unnecessary. Ongoing fluid losses caused by profuseomiting or diarrhea may need to be replaced on rare occa-ion.

The serum glucose concentration often decreases substan-ially with rehydration alone as a result of improvements inhe glomerular filtration rate and decreased concentrations ofounterregulatory hormones.46,75 This decline in glucoseoncentration early in treatment should not be interpreted asn indication of excessive insulin administration.

nsulin and Dextrosensulin is required to resolve acidosis and hyperglycemia viauppression of ketogenesis, gluconeogenesis, and glycogen-lysis and promotion of peripheral glucose uptake and me-abolism. Insulin should be administered intravenously at aate of 0.1 U/kg/h.75 An initial bolus or loading dose of insulins unnecessary because maximal reductions in ketogenesisnd lipolysis are achieved rapidly with the insulin infusionate specified previously.26,32 More rapid declines in serumlucose concentration may be achieved with insulin admin-stered at rates in excess of 0.1 U/kg/h, but these highernsulin dosages may increase the frequency of hypoglycemiauring therapy.32,75 The risk of hypokalemia also is greater atigher insulin infusion rates.32,75 Thus, there seems to be noenefit to higher insulin dosages, and the potential for ad-erse effects may increase. The use of insulin dosages lesshan 0.1 U/kg/h have not been studied extensively, but avail-ble data suggest that these lower dosages may not suppressetogenesis adequately.76

With insulin treatment, serum glucose concentrations of-en normalize before ketosis and acidosis have resolved.

hen the serum glucose concentration declines to approxi-ately 250 to 300 mg/dL, dextrose should be added to the

ntravenous fluids to avoid hypoglycemia as the insulin infu-ion is continued to promote resolution of ketosis and acido-is. The two-bag system is an effective and efficient methodor administering dextrose in children who have DKA. Thisystem allows a more rapid response to changes in serumlucose concentration and is more cost effective than single-ag methods.77 Two bags of intravenous fluids with identicallectrolyte content but varying dextrose concentrations (usu-lly 0% and 10%) are administered simultaneously. The rel-tive rates of administration of the two fluids can be adjustedo vary the dextrose concentration while maintaining a con-tant overall rate of administration of fluid and other electro-ytes (Fig. 5). Once this system is established, the blood glu-ose concentration should be maintained between 150 and50 mg/dL to strike a balance between avoidance of hypo-lycemia during treatment and prevention of ongoing fluidosses from osmotic diuresis.

lectrolytesith insulin treatment and resolution of acidosis, there is

ubstantial movement of potassium from the extracellular

sci4rpbori

rplwfatatt

ttcmeoorddchoptohrfrcp

mhctda

BBwwt

Fbfleg0dcUmliG(mpiGtb134:377; with permission.)

192 N. Glaser

pace to the intracellular space, and serum potassium con-entrations may decrease precipitously. Intravenous admin-stration of potassium is essential, and concentrations of 30 to0 mEq/L intravenous fluids are usually required. Adequateenal function should be ensured before administration ofotassium. Potassium chloride may be used alone or in com-ination with potassium phosphate or potassium acetate. Usef combinations of potassium salts may help to diminish theisk of development of hyperchloremic acidosis by decreas-ng the chloride load.

Studies have demonstrated that some degree of hyperchlo-emic acidosis develops during treatment of DKA in mostatients, and the severity of hyperchloremic acidosis corre-

ates with serum urea nitrogen concentrations.47 Patientsho are less dehydrated and have better preservation of renal

unction have a greater tendency to develop hyperchloremiccidosis during treatment. This tendency is likely caused byhe increased urinary loss of bicarbonate precursors (keto-cid and lactic acid anions) and diminished conversion ofhese precursors to bicarbonate with insulin administra-ion.47,78

Whether phosphate replacement should be given rou-inely in children who have DKA is controversial. It is knownhat 2,3-diphosphoglycerate levels in red blood cells are de-reased in patients who have DKA, and hypophosphatemiaay result in persistence of low 2,3-diphosphoglycerate lev-

ls. This situation theoretically may lead to reduced tissuexygen delivery, particularly during therapy when correctionf acidosis increases the affinity of hemoglobin for oxygen,eversing the Bohr effect.79,80 Occurrence of this effect to aegree that would be clinically relevant, however, has beenifficult to demonstrate.80,81 Conversely, although hypocal-emia can result from phosphate replacement, symptomaticypocalcemia has been documented mainly with aggressiver rapid phosphate replacement and is uncommon whenhosphate is administered slowly in more modest concentra-ions.81,82 It is difficult to make a strong case either in favor ofr against phosphate replacement. Case reports, however,ave documented rhabdomyolysis and hemolytic anemia asesults of severe hypophosphatemia during DKA.83,84 There-ore, regardless of whether phosphate replacement is givenoutinely, it is necessary to monitor serum phosphate con-entrations during treatment and administer phosphate re-lacement if severe hypophosphatemia develops.Hypomagnesemia is common during DKA treatment anday contribute to hypocalcemia by inhibition of parathyroidormone secretion.85,86 Although monitoring of serum cal-ium and magnesium concentrations is recommended to de-ect rare cases of severe hypomagnesemia or hypocalcemia,ecreases in the concentrations of these electrolytes are usu-lly mild and asymptomatic and rarely require treatment.

icarbonateicarbonate should not be administered routinely in childrenho have DKA because acidosis usually can be correctedith insulin and fluids alone, and hemodynamic instability

igure 5 Two-bag system and illustrative typical course. (A) Two-ag system allows independent manipulation of glucose and totaluid volume, because electrolyte content of two bags is identicalxcept for dextrose. (B) Differential rates of two bags modulatelucose delivery, which can be any concentration ranging from% to 10%. Total fluid volume is based on a patient’s degree ofehydration and ongoing fluid requirement. (C) In this typicalourse, insulin therapy is instituted as continuous infusion of 0.1/kg/h, and total fluid rate is set at 200 mL/h. Because patient isarkedly hyperglycemic, no dextrose is given initially. As insu-

in action lowers patient’s glucose level, dextrose is titrated intontravenous fluid without changing administered fluid volume.lucose titration aims to control rate of blood glucose decline

possible risk factor for cerebral edema) and prevent hypoglyce-ia in the face of continued insulin requirement. Later, when a

atient’s dehydration and ketosis become partially corrected,nsulin and total fluid can be independently adjusted. (Fromrimberg A, Cerri RW, Satin-Smith M, et al. The “two bag sys-

em” for variable intravenous dextrose and fluid administration:enefits in diabetic ketoacidosis management. J Pediatr 1999;

hat results from acidosis is rare.52 Most studies have found

mabbbtfm

hihCbattcbmotwsib

MShfcMimttmcsmaif

aCrRiD2

CTp

sctilevliumidcppwr

lnRnrhwslawdcvsD

mbccQictttQ

esciasDm

DKA and hyperglycemic hyperosmolar state 193

inimal or no differences in the rapidity of correction ofcidosis in patients who have DKA treated with or withouticarbonate.87–89 One reason for the apparent lack of effect oficarbonate on rapidity of resolution of acidosis is that bicar-onate administration may cause an increase in hepatic ke-one production.90 It is believed that this increase resultsrom pH-dependent stimulation of ketogenesis via increased

itochondrial uptake of fatty acyl-CoA.Bicarbonate administration also increases the likelihood of

ypokalemia during DKA treatment.91 and theoretically mayncrease tissue hypoxia as a result of leftward shifts in theemoglobin-oxygen dissociation curve.79 Bicarbonate B andtreatment also may lead to paradoxic acidosis of the cere-

rospinal fluid.53,92 This phenomenon likely occurs becausedministration of bicarbonate results in diminished respira-ory drive and a rise in the partial pressure of CO2. Althoughhe blood-brain barrier is impermeable to bicarbonate, CO2

rosses the blood-brain barrier readily and generates car-onic acid and cerebrospinal fluid acidosis. Bicarbonate ad-inistration also has been associated with an increased risk

f cerebral edema in childhood DKA.93 Routine administra-ion of bicarbonate is not recommended. In rare cases inhich hemodynamic instability is believed to be caused by

evere acidosis and does not respond to standard measures orn rare cases of symptomatic hyperkalemia, however, bicar-onate administration should be considered.

onitoringpecific recommendations for monitoring of children whoave DKA are out- lined in the report of the European Societyor Pediatric Endocrinology/Lawson Wilkins Pediatric Endo-rine Society international DKA consensus conference.94,95

ost patients who have DKA should be treated in a pediatricntensive care unit or other unit with similar capacities for

anaging children who have DKA. Blood glucose concentra-ions should be measured hourly and electrolyte concentra-ions should be monitored every 2 to 4 hours. Venous pHeasurements are helpful because serum bicarbonate con-

entrations may not increase over the first several hours de-pite improvements in acidosis. Arterial blood gas measure-ents, however, are generally unnecessary. Lack of

ppropriate improvement in acidosis with treatment suggestsnadequate insulin infusion, inadequate rehydration, renalailure, sepsis, or other intercurrent condition.

Vital signs and mental status should be monitored hourly,nd fluid intake and output should be recorded accurately.ardiac monitoring is recommended because cardiac ar-

hythmias may occur during treatment, albeit infrequently.ecent data demonstrated a high frequency of prolonged QT

nterval corrected for heart rate (QTc) in children who haveKA (N. Kuppermann, MD, personal communication,005).

omplicationshe most frequent complications of DKA treatment are hy-

oglycemia and hypokalemia. With adequate monitoring of s

erum glucose and potassium concentrations, however, theseomplications are usually detected at an early stage, are easilyreated, and rarely result in permanent morbidity or mortal-ty. More serious complications of DKA are rare but may beife threatening, including cerebral edema,93,96 pulmonarydema,97–99 CNS hemorrhage or thrombosis,100 other largeessel thromboses,101 cardiac arrhythmias caused by electro-yte disturbances,93,102,103 pancreatitis,104 renal failure,105 andntestinal necrosis.106–108 Patients who have DKA are alsoniquely susceptible to rhinocerebral and pulmonary mucor-ycosis, a rare fungal infection.109 Acidosis interferes with an

mportant host defense mechanism against this fungus byisrupting the capacity of transferrin to bind iron. Mucormy-osis occurs most frequently in children with longstandingoor blood glucose control. This infection carries a poorrognosis with high mortality rates. Aggressive treatmentith antifungal agents and early resection of involved tissue is

ecommended.110

Although severe dehydration and electrolyte depletionikely cause some of the complications of DKA, the mecha-isms responsible for several others are not well understood.ecent studies have suggested that �-OHB may cause pulmo-ary vascular endothelial dysfunction and that perfusion ofabbit lungs with either �-OHB or AcAc results in edema andemorrhage.111 DKA also may cause a prothrombotic state,hich may predispose children to CNS and other thrombo-

es.101,112 Studies have reported increased levels of von Wil-ibrand factor and decreased free protein S and protein Cctivity in DKA and enhanced platelet aggregation associatedith hyperglycemia.112–114 Case series have suggested thateep venous thromboses may develop in as many as 50% ofhildren with femoral central venous catheters.101,115 Centralenous catheters, particularly femoral venous catheters,hould therefore be used with caution in children who haveKA.Cardiac arrhythmias occur infrequently during DKA treat-ent and generally have been attributed to electrolyte distur-

ances. Recent data, however, documented a consistent in-rease in the QT interval corrected for heart rate (QTc) inhildren during acute DKA, with 47% of children having aTc above 450 msec, the threshold generally considered to

ndicate prolongation of QTc.96 In the recent study, the in-rease in QTc did not correlate with electrolyte concentra-ions, and the frequency of abnormal electrolyte concentra-ions in the study group was low, which raised the possibilityhat ketosis per se might have an effect on the myocardium.Tc intervals normalized after treatment of DKA.The most frequent serious complication of DKA is cerebral

dema, which occurs in 0.3% to 1% of pediatric DKA epi-odes.93,96,116,117 Symptoms and signs of cerebral edema in-lude headache, altered mental status, recurrence of vomit-ng, hypertension, inappropriate slowing of the heart rate,nd other signs of increased intracranial pressure. Recenttudies have documented a 21% to 24% mortality rate forKA-related cerebral edema and a 21% to 26% rate of per-anent neurologic morbidity.93,96

Although less than 1% of children who have DKA develop

ymptomatic cerebral edema, studies that used sequential CT

sDlTdmbvpisttrecDr

ccgbDbi

rbcAdaicm

ecacbbdawcitte

DIadcstcdcho

wihdtome1

FaDabN

194 N. Glaser

cans or other imaging technologies in children who haveKA showed that mild, asymptomatic cerebral edema is

ikely present in most children who have DKA (Fig. 6).118–120

he pathophysiologic mechanisms that cause cerebral edemauring DKA remain unclear and have been the source ofuch controversy. Many investigators have attributed cere-

ral edema to rapid changes in serum osmolality or overlyigorous fluid resuscitation during DKA treatment. This hy-othesis, however, has not been supported by data from clin-

cal studies. In studies that used appropriate multivariatetatistical techniques to adjust for DKA severity, an associa-ion between the rate of change in serum glucose concentra-ion or the volume or sodium content of fluid infusions andisk for cerebral edema was not demonstrated.93,121,122 Sev-ral case reports also described symptomatic and even fatalerebral edema that occurred before hospital treatment forKA.1,123,124 This information suggests that DKA-related ce-

igure 6 CT scans of the same patient during DKA treatment (A) andfter recovery from DKA (B). Narrowing of the ventricles duringKA indicates cerebral edema, although the patient was asymptom-

tic. (From Krane EJ, Rockoff MA, Wallman JK, et al. Subclinicalrain swelling in children during treatment of diabetic ketoacidosis.Engl J Med 1985;312:1147–51; with permission.)

ebral edema likely cannot be explained simply by osmoti- i

ally mediated fluid shifts. More recent data suggested thaterebral edema during DKA may be predominantly vaso-enic and may result from activation of ion transporters in thelood-brain barrier.118,125 Cerebral hypoperfusion duringKA or direct effects of ketosis or inflammatory cytokines onlood-brain barrier endothelial cell function might play a role

n stimulating this process.69,118,126

Epidemiologic studies have shown that children at greatestisk for symptomatic cerebral edema are children with highlood urea nitrogen concentrations.93 at presentation andhildren who present with more profound hypocapnia.93,122

lesser rise in the measured serum sodium concentrationuring treatment (as the serum glucose concentration falls)lso indicates increased risk for cerebral edema.93,127 Morentensive monitoring of neurologic state and vital signs forhildren who present with these risk factors is recom-ended.Clinical studies have not demonstrated a clear beneficial

ffect of any pharmacologic agent in treating DKA-relatederebral edema. Case reports, however, suggest that promptdministration of mannitol (0.25–1 g/kg) may be benefi-ial.128,129 Intubation with associated hyperventilation haseen correlated with poorer outcomes of DKA-related cere-ral edema.130 Therapeutic hyperventilation that attempts toecrease pCO2 below a patient’s own compensation for met-bolic acidosis likely should be avoided in intubated childrenho have DKA except when absolutely necessary to treat

linically overt elevated intracranial pressure. CNS imagingn patients with suspected cerebral edema is recommendedo rule out other causes of altered mental status, such as CNShromboses; however, treatment for suspected cerebraldema should not be delayed while awaiting imaging studies.

ifferential Diagnosisn children, findings of hyperglycemia, increased anion gapcidosis, and ketonuria or ketonemia generally indicate aiagnosis of DKA, and other disorders that result in thisonstellation of biochemical abnormalities are rare. Occa-ionally, however, other disorders may have a similar presen-ation. Rare metabolic defects may cause ketoacidosis, in-luding succinyl-CoA: 3-ketoacid coenzyme A transferaseeficiency, a defect in ketolysis, and beta-ketothiolase defi-iency, a defect in L-isoleucine catabolism. These conditions,owever, are most frequently associated with hypoglycemiar normoglycemia rather than hyperglycemia.30,131–133

In the setting of gastroenteritis, hyperglycemia may occurhen stress hormone concentrations are markedly elevated

n response to dehydration. Lactic acidosis results from de-ydration, and the combination of hyperglycemia with aci-osis initially may suggest a diagnosis of DKA. FFA concen-rations also may be elevated, and modest ketonemiaccasionally occurs.134–138 These findings have been docu-ented most frequently in infants and toddlers. In rare cases,

xtreme elevations in serum glucose concentration (� 800–000 mg/dL) have been reported in infants with gastroenter-

tis without diabetes mellitus.139 Rapid resolution of hyper-

gd

HSEohmqwswHihfHmvicotcbca

ecswprttwpuilhmHhsetm

ctsgr

shaoaiicanb

iaigtiidlrptmempmctiwu

qcrH1riosumawCbc

SDtt

DKA and hyperglycemic hyperosmolar state 195

lycemia with hydration alone can be helpful inifferentiating this situation from DKA.134

yperglycemic Hyperosmolartate Without Ketosis

xtreme hyperglycemia and hyperosmolality can occur with-ut ketosis in patients who have diabetes (hyperglycemicyperosmolar state [HHS]). This condition occurs muchore frequently in adults than in children and more fre-

uently in patients who have type 2 diabetes than in personsho have type 1 diabetes. Among pediatric patients, case

eries have suggested that obese African-American childrenho have type 2 diabetes may be at greatest risk forHS.140,141 HHS also has been documented to occur with

ncreased frequency in patients who are predisposed to de-ydration because of limited access to fluids, including in-ants and children with cognitive deficits.139,142 AlthoughHS has been viewed as a condition separate from DKA, itay be more appropriate to view HHS as one extreme in the

arious presentations of altered glucose and fat metabolismn patients who have diabetes. DKA with near-normal glu-ose concentrations (euglycemic DKA) may be viewed as thepposite extreme on this continuum. Where a particular pa-ient falls in this spectrum is determined by the relative con-entrations of insulin and counterregulatory hormones andy the states of hydration and nutrition of the patient. Thelinical picture in many patients may have elements of DKAnd HHS.143

The pathogenesis of HHS is similar to that of DKA; how-ver, some important differences should be noted. Hypergly-emia without ketosis generally occurs in patients who retainome ability to produce insulin, most commonly personsho have type 2 diabetes. Ketogenesis and lipolysis are sup-ressed at lower serum insulin concentrations than the levelsequired to suppress hepatic glucose production, and pa-ients develop hyperglycemia without ketosis.32,144 In pa-ients who do not develop ketoacidosis, osmotic diuresisith electrolyte and water loss may persist for prolongederiods and result in profound dehydration. Without ketosis,rinary cation excretion is not necessary to balance ketoan-

on excretion, and less electrolyte loss relative to free wateross occurs in HHS than in DKA, which contributes to theyperosmolar state. Nonetheless, because the duration of os-otic diuresis in HHS may be lengthy, patients who haveHS may have greater electrolyte deficits than patients whoave DKA.145 Diminished renal function that results fromevere dehydration is particularly important in the pathogen-sis of HHS because diminished capacity for glucose excre-ion is necessary for the development of marked hyperglyce-ia.The criteria for diagnosis of HHS include blood glucose

oncentration more than 600 mg/dL, serum osmolality morehan 330 mOsm/kg, and lack of significant ketosis.146 Theerum sodium concentration, when corrected for the bloodlucose concentration, is generally above the normal

ange.146,147 The clinical presentation of HHS is otherwise D

imilar to that of DKA, with some exceptions. Children whoave HHS often have a more prolonged history of polyuriand polydipsia than children who have DKA.140,142 Becausef the absence of ketosis, fruity breath odor is not present,nd tachypnea is not a prominent feature, except in patientsn whom substantial lactic acidosis occurs. In adults, approx-mately 10% to 20% of patients who have HHS present inoma, and other mental status abnormalities at presentationre more frequent than in DKA. Seizures may occur, and focaleurologic deficits (eg, hemiparesis, hemianopsia, chorea-allismus) also have been described with HHS.139,142,146,148

Because HHS occurs infrequently in children, data regard-ng the optimal approach to treatment are lacking. Someuthors have suggested that it may be preferable to delaynsulin therapy in patients who have HHS because the serumlucose concentration decreases considerably with rehydra-ion alone.142 Patients who have HHS are not ketotic, andnsulin is not needed for resolution of acidosis.142 Delayingnsulin treatment in these patients may result in more gradualeclines in serum glucose concentration and serum osmola-

ity. Use of 0.9% saline for intravenous fluid replacementather than hypotonic saline also has been recommended toromote a more gradual decline in serum sodium concentra-ion. Because patients with HHS may have had ongoing os-otic diuresis for prolonged periods before presentation,

lectrolyte deficits may be particularly pronounced. Closeonitoring of serum electrolyte concentrations (particularlyotassium and phosphate) is recommended.146 Hypernatre-ia frequently develops during therapy as the serum glucose

oncentration declines and water returns to the extravascularissues. Hypernatremia occasionally may be difficult to treatn patients who have HHS, in part because of ongoing freeater losses caused by osmotic diuresis and persistent stim-lation of sodium retention by aldosterone.149

In contrast to DKA, in which complications occur infre-uently and the mortality rate is less than 1%, HHS is asso-iated with more frequent complications and a high mortalityate. Although limited epidemiologic data are available onHS in children, one report documented a mortality rate of4%,150 similar to the approximately 15% to 20% mortalityate of HHS in adults.143 Thromboembolic complications,ncluding pulmonary emboli and deep venous thromboses,ccur frequently in patients who have HHS as a result ofevere dehydration and increased blood viscosity.151 Routinese of anticoagulant therapy, however, is controversial. Aalignant hyperthermia-like syndrome with hyperpyrexia

nd rhabdomyolysis also was described in several childrenho had HHS.141 The cause of this syndrome is unclear.ardiac arrhythmias caused by severe electrolyte distur-ances, cerebral edema, and pulmonary edema also may oc-ur.140,141

ummaryKA occurs frequently in children who have diabetes, par-

icularly at the time of diagnosis. Greater efforts are necessaryo promote earlier recognition of new onset of diabetes so that

KA can be prevented and to avoid subsequent occurrences

orerwm

R

196 N. Glaser

f DKA in children who have established diabetes. Furtheresearch is also necessary to understand and prevent cerebraldema, the most serious complication of DKA. Internationalecommendations for DKA treatment in children recentlyere published and will be helpful in standardizing the treat-ent of this condition.94,95

eferences1. Edge J, Ford-Adams M, Dunger D: Causes of death in children with

insulin-dependent diabetes 1990–96. Arch Dis Child 81:318-323,1999

2. Scibilia J, Finegold D, Dorman J, et al: Why do children with diabetesdie? Acta Endocrinol Suppl (Copenh) 279:326-333, 1986

3. Faich G, Fishbein H, Ellis E. The epidemiology of diabetic acidosis: apopulation-based study. Am J Epidemiol 117:551-8, 1983

4. Pinkney J, Bingley P, Sawtell P: Presentation and progress of child-hood diabetes mellitus: a prospective population-based study. Diabe-tologia 37:70-74, 1994

5. Levy-Marchal C, Patterson C, Green A: Geographical variation of pre-sentation at diagnosis of type 1 diabetes in children: the EURODIABstudy. Diabetologia 44(Suppl 3):B75-80, 2001

6. Mallare J, Cordice C, Ryan B, et al: Identifying risk factors for thedevelopment of diabetic ketoacidosis in new onset type 1 diabetesmellitus. Clin Pediatr 42:591-597, 2003

7. Vanelli M, Chiari G, Ghizzoni L, et al: Effectiveness of a preventionprogram for diabetic ketoacidosis in children. Diabetes Care 22:7-9,1999

8. Hathout E, Thomas W, El-Shahawy M, et al: Diabetic autoimmunemarkers in children and adolescents with type 2 diabetes. Pediatrics107:e102, 2001

9. Glaser N, Jones K: Non-insulin dependent diabetes mellitus in Mexi-can-American children and adolescents. West J Med 168:11-16, 1998

10. Grinstein G, Muzumdar R, Aponte L, et al: Presentation and 5-yearfollow-up of type 2 diabetes mellitus in African-American and Carib-bean-Hispanic adolescents. Horm Res 60:121-126, 2003

11. Zdravkovic V, Daneman D, Hamilton J: Presentation and course oftype 2 diabetes in youth in a large multi-ethnic city. Diabet Med21:1144-1148, 2004

12. Blanc N, Lucidarme N, Tubiana-Rufi N: Factors associated with child-hood diabetes manifesting as ketoacidosis and its severity. Arch Pedi-atr 10:320-325, 2003

13. Rewers A, Chase H, Mackenzie T, et al: Predictors of acute complica-tions in children with type 1 diabetes. JAMA 287:2511-2518, 2002

14. Rosilio M, Cotton J, Wieliczko M, et al: Factors associated with glyce-mic control: a cross-sectional nationwide study in 2,579 French chil-dren with type 1 diabetes. The French Pediatric Diabetes Group.Diabetes Care 21:1146-1153, 1998

15. Norfeldt S, Ludvigsson J: Adverse events in intensively treated chil-dren and adolescents with type 1 diabetes. Acta Paediatr 88:1184-1193, 1999

16. Flood R, Chiang V: Rate and prediction of infection in children withdiabetic ketoacidosis, Am J Emerg Med 19:270-273, 2001

17. Umpierrez G, Murphy M, Kitabchi A: Diabetic ketoacidosis and hy-perglycemic hyperosmolar syndrome. Diabetes Spectrum 15:28-36,2002

18. Azoulay E, Chevret S, Didier J, et al: Infection as a trigger of diabeticketoacidosis in intensive care-unit patients. Clin Infect Dis 32:30-35,2001

19. Gerich J, Lorenzi M, Bier D, et al: Prevention of human diabetic keto-acidosis by somatostatin: evidence for an essential role of glucagon.N Engl J Med 292:985-989, 1975

20. Schade D, Eaton R: Glucagon regulation of plasma ketone body con-centration in human diabetes. J Clin Invest 56:1340-1344, 1975

21. Boden G, Cheung P, Homko C: Effects of acute insulin excess anddeficiency on gluconeogenesis and glycogenolysis in type 1 diabetes.Diabetes 52:133-137, 2003

22. Gustavson S, Chu C, Nishizawa M, et al: Interaction of glucagon and

epinephrine in the control of hepatic glucose production in theconscious dog. Am J Physiol Endocrinol Metab 284:E695-707, 2002

23. Gustavson S, Chu C, Nishizawa M, et al: Glucagon’s actions are mod-ified by the combination of epinephrine and gluconeogenic precursorinfusion. Am J Physiol Endocrinol Metab 285:E534-544, 2003

24. Cheatham B, Volchuk A, Kahn R, et al: Insulin-stimulated transloca-tion of GLUT4 glucose transporters requires SNARE-complex pro-teins. Proc Natl Acad Sci U S A 93:15169-15173, 1996

25. Minokoshi Y, Kahn R, Kahn B. Tissue-specific ablation of the GLUT4glucose transporter or the insulin receptor challenges assumptionsabout insulin action and glucose homeostasis. J Biol Chem 278:33609-33612, 2004

26. Luzi L, Barrett E, Groop L, et al: Metabolic effects of low-dose insulintherapy on glucose metabolism in diabetic ketoacidosis. Diabetes 37:1470-1477, 1988

27. Menzel R, Kaisaki P, Rjasanowski I, et al: A low renal threshold forglucose in diabetic patients with a mutation in the hepatocyte nuclearfactor-1alpha (HNF-1alpha) gene. Diabet Med 15:816-820, 1998

28. Ruhnau G, Faber O, Borch-Johnsen K, et al: Renal threshold for glu-cose in non-insulin-dependent diabetic patients. Diabetes Res ClinPract 36:27-33, 1997

29. McGarry J, Foster D: Regulation of ketogenesis and clinical aspects ofthe ketotic state. Metabolism 21:471-489, 1972

30. Fukao T, Lopaschuk G, Mitchell G: Pathways and control of ketonebody metabolism: on the fringe of lipid biochemistry. ProstaglandinsLeukot Essent Fatty Acids 70:243-251, 2004

31. McGarry J, Wright P, Foster D: Rapid activation of hepatic ketogeniccapacity in fed rats by anti-insulin serum and glucagon. J Clin Invest55:1202-1209, 1975

32. Schade D, Eaton R: Dose response to insulin in man; differentialeffects on glycose and ketone body regulation. J Clin EndocrinolMetab 44:1038-1053, 1977

33. Madison L, Mebane D, Unger R, et al: The hypoglycemic action ofketones. II. Evidence for a stimulatory feedback of ketones on thepancreatic beta cells. J Clin Invest 43:408-415, 1964

34. Umpierrez G, DiGirolamo M, Tuvlin J, et al: Differences in metabolicand hormonal milieu in diabetic and alcohol-induced ketoacidosis. JCrit Care 15:52-59, 2000

35. Burge M, Garcia N, Qualls CS: Differential effects of fasting and dehy-dration in the pathogenesis of diabetic ketoacidosis. Metabolism 50:171-177, 2001

36. Djurhuus C, Gravholt C, Nielsen S, et al: Additive effects of cortisoland growth hormone on regional and systemic lipolysis in humans.Am J Physiol Endocrinol Metab 286:E488-E494, 2003

37. Nolte L, Rincon J, Wahlstrom E, et al: Hyperglycemia activates glucosetransport in rat skeletal muscle via a Ca(2 �)- dependent mechanism.Diabetes 44:1345-1348, 1995

38. Jonhston D, Gill A, Orskov H, et al: Metabolic effects of cortisol inman: studies with somatostatin. Metabolism 31:312-317, 1982

39. Sacca L, Vigorito C, Cicala M, et al: Role of glyconeogenesis in epi-nephrine-stimulated hepatic glucose production in humans. Am JPhysiol 245:E294-302, 1983

40. Krentz A, Freedman D, Greene R, et al: Differential effects of physio-logical versus pathophysiological plasma concentrations of epineph-rine and norepinephrine on ketone body metabolism and hepaticportal blood flow in man. Metabolism 45:1214-1220, 1996

41. Avogaro A, Valerio A, Gnudi L, et al: The effects of different plasmainsulin concentrations on lipolytic and ketogenic responses to epi-nephrine in normal and type 1 (insulin-dependent) diabetic humans.Diabetologia 35:129-138, 1992

42. Porte D: Sympathetic regulation of insulin secretion. Arch Intern Med123:252-260, 1969

43. Nonogaki K.: New insights into sympathetic regulation of glucose andfat metabolism. Diabetologia 43:533-549, 2000

44. Butler P, Kryshak E, Rizza R: Mechanism of growth hormone-inducedpostprandial carbohydrate intolerance in humans. Am J Physiol 260:E513-520, 1991

45. Zimmet P, Taft P, Ennis G, et al: Acid production in diabetic acidosis:

a more rational approach to alkali replacement. BMJ 3:610-612, 1970

DKA and hyperglycemic hyperosmolar state 197

46. Waldhausl W, Kleinberger G, Korn A, et al: Severe hyperglycemia:effects of rehydration on endocrine derangements and blood glucoseconcentration. Diabetes 28:577-584, 1979

47. Adrogue H, Wilson H, Boyd A, et al: Plasma acid-base patterns indiabetic ketoacidosis. N Engl J Med 307:1603-1610, 1982

48. Foster D, McGarry J: The metabolic derangements and treatment ofdiabetic ketoacidosis. N Engl J Med 309:159-169, 1983

49. Kreisberg R: Diabetic ketoacidosis: new concepts and trends in patho-genesis and treatment. Ann Intern Med 88:681-695, 1978

50. Valerio D: Acute diabetic abdomen in children. Lancet 1(7950):66-68, 1976

51. Matz R: Hypothermia in diabetic acidosis. Hormones 3:36-41, 197252. Maury E, Vassal T, Offenstadt G: Cardiac contractility during severe

ketoacidosis. N Engl J Med 341:1938, 199953. Assal J, Aoki T, Manzano F, et al: Metabolic effects of sodium bicar-

bonate in management of diabetic ketoacidosis. Diabetes 23:405-411,1973

54. Ohman J, Marliss E, Aoki T, et al: The cerebrospinal fluid in diabeticketoacidosis. N Engl J Med 284:283-290, 1971

55. Cullen M, Reece E, Homko C, et al: The changing presentations ofdiabetic ketoacidosis during pregnancy. Am J Perinatol 13:449-451,1996

56. Jenkins D, Close C, Krentz A, et al: Euglycemic diabetic ketoacidosis:does it exist? Acta Diabetol 30:251-253, 1993

57. Burge M, Hardy K, Schade D: Short term fasting is a mechanism for thedevelopment of euglycemic ketoacidosis during periods of insulindeficiency. J Clin Endocrinol Metab 76:1192-1198, 1993

58. Halperin M, Goldstein M, Richardson R, et al: Quantitative aspects ofhyperglycemia in the diabetic: a theoretical approach. Clin Invest Med2(4):127-130, 1980

59. Laffel L: Ketone bodies: a review of physiology, pathophysiology andapplication of monitoring to diabetes. Diabetes Metab Res Rev 15:412-426, 1999

60. Ham M, Okada P, White P: Bedside ketone determination in diabeticchildren with hyperglycemia and ketosis in the acute care setting.Pediatr Diabetes 5:39-43, 2004

61. Fearon D, Steele D: End-tidal carbons dioxide predicts the presenceand severity of acidosis in children with diabetes. Acad Emerg Med9:1373-1378, 2002

62. Katz M: Hyperglycemia-induced hyponatremia: calculation of ex-pected serum sodium depression. N Engl J Med 289:843-844, 1973

63. Kaminska E, Pourmotabbed G: Spurious laboratory values in diabeticketoacidosis and hyperlipidemia. Am J Emerg Med 11:77-80, 1993

64. Halperin M, Bear R, Goldstein M, et al: Interpretation of the serumpotassium concentration in metabolic acidosis. Clin Invest Med 2(1):55-57, 1979

65. Yu D, Clements P: Human lymphocyte subpopulations effect of epi-nephrine. Clin Exp Immunol 25:472-479, 1976

66. Bessey P, Watters J, Aoki T, et al: Combined hormonal infusion sim-ulates the metabolic response to injury. Ann Surg 200:264-281, 1984

67. Stentz F, Umpierrez G, Cuervo R, et al: Proinflammatory cytokines,markers of cardiovascular risk, oxidative stress, and lipid peroxidationin patients with hyperglycemic crises. Diabetes 53:2079-2086, 2004

68. Dalton R, Hoffman W, Passmore G, et al: Plasma C-reactive proteinlevels in severe diabetic ketoacidosis. Ann Clin Lab Sci 33:435-442,2003

69. Hoffman W, Burek C, Waller J, et al: Cytokine response to diabeticketoacidosis and its treatment. Clin Immunol 108:175-181, 2003

70. Kaergaard J, Salling N, Magid E, et al: Serum amylase during recoveryfrom diabetic ketoacidosis. Diabetes Metab 10:25-30, 1984

71. Moller-Petersen J, Andersen P, Hjorne N, et al: Hyperamylasemia,specific pancreatic enzymes and hypoxanthine during recovery fromdiabetic ketoacidosis. Clin Chem 31:2001-2004, 1985

72. Haddad N, Croffie J, Eugster E: Pancreatic enzyme elevations in chil-dren with diabetic ketoacidosis. J Pediatr 145:122-124, 2004

73. Koves I, Neutze J, Donath S, et al. The accuracy of clinical assessmentof dehydration during diabetic ketoacidosis in childhood. DiabetesCare 27:2485-2487, 2004.

74. Smith L, Rotta A. Accuracy of clinical estimates of dehydration in

pediatric patients with diabetic ketoacidosis. Pediatr Emerg Care18:395-396, 2002.

75. Burghen G, Etteldorf J, Fisher J, et al. Comparison of high-dose andlow-dose insulin by continuous intravenous infusion in the treatmentof diabetic ketoacidosis in children. Diabetes Care 3:15-20, 1980.

76. Miles J, Gerich J. Glucose and ketone body kinetics in diabetic keto-acidosis. Clin Endocrinol Metab 12:303-319, 1983.

77. Grimberg A, Cerri R, Satin-Smith M, et al. The “two bag system” forvariable intravenous dextrose and fluid administration: benefits indiabetic ketoacidosis management. J Pediatr 134:376-378, 1999

78. Oh M, Banerji M, Carroll H. The mechanism of hypercholoremicacidosis during the recovery phase of diabetic ketoacidosis. Diabetes30:310-313, 1981

79. Alberti G, Darley J, Emerson P, et al. 2,3-Diphosphoglycerate andtissue oxygenation in uncontrolled diabetes mellitus. Lancet 2(7774):391-395, 1972

80. Fisher J, Kitabchi A. A randomized study of phosphate therapy in thetreatment of diabetic ketoacidosis. J Clin Endocrinol Metab 57:177-180, 1983

81. Gibby O, Veale K, Hayes T, et al. Oxygen availability from the bloodand the effect of phosphate replacement on erythrocyte 2,3-dipospho-glycerate and haemoglobin-oxygen affinity in diabetic ketoacidosis.Diabetologia 15:381-385, 1978.

82. Becker D, Brown D, Steranka B, et al. Phosphate replacement duringtreatment of diabetic ketoacidosis. Am J Dis Child 137:241-246, 1983

83. Lord G, Scott J, Pusey C, et al. Diabetes and rhabdomyolysis: a rarecomplication of a common disease. BMJ 307:1126-1128, 1993.

84. Shilo S, Werner D, Hershko C. Acute hemolytic anemia caused bysevere hypophosphatemia in diabetic ketoacidosis. Acta Haematol73:55-57, 1985

85. Zipf W, Bacon G, Spencer M, et al. Hypocalcemia, hypomagnesemia,and transient hypoparathyroidism during therapy with potassiumphosphate in diabetic ketoacidosis. Diabetes Care 2:265-268, 1979.

86. Ionescu-Tirgoviste C, Bruckner I, Mihalache N, et al. Plasma phos-phorus and magnesium values during treatment of severe diabeticketoacidosis. Med Intern 19:63-68, 1981.

87. Green S, Rothrock S, Ho J, et al. Failure of adjunctive bicarbonate toimprove outcome in severe pediatric diabetic ketoacidosis. AnnEmerg Med 31:41-8, 1998.

88. Lever E, Jaspan J. Sodium bicarbonate therapy in severe diabetic ke-toacidosis. Am J Med 75:263-268, 1983.

89. Morris L, Murphy M, Kitabchi A. Bicarbonate therapy in severe dia-betic ketoacidosis. Ann Intern Med 105:836-840, 1986.

90. Okuda Y, Adrogue H, Field J, et al. Counterproductive effects ofsodium bicarbonate in diabetic ketoacidosis. J Clin Endocrinol Metab81:314-320, 1996.

91. Soler N, Bennet M, Dixon K, et al. Potassium balance during treatmentof diabetic ketoacidosis with special reference to the use of bicarbon-ate. Lancet 30:665-667, 1972.

92. Bureau M, Begin R, Berthiaume Y, et al. Cerebral hypoxia from bicar-bonate infusion in diabetic acidosis. J Pediatr 96:968-97, 1983.

93. Glaser N, Barnett P, McCaslin I, et al. Risk factors for cerebral edemain children with diabetic ketoacidosis. N Engl J Med 344:264-269,2001.

94. Sperling M, Dunger D, Acerini C, et al. ESPE / LWPES consensusstatement on diabetic ketoacidosis in children and adolescents. Pedi-atrics 113:e133-e1340, 2003.

95. Dunger D, Sperling M, Acerini C, et al. ESPE / LWPES consensusstatement on diabetic ketoacidosis in children and adolescents. ArchDis Child 89:188-194, 2003.

96. Edge J, Hawkins M, Winter D, et al. The risk and outcome of cerebraloedema developing during diabetic ketoacidosis. Arch Dis Child 85:16-22, 2001.

97. Young M. Simultaneous acute cerebral and pulmonary edema com-plicating diabetic ketoacidosis. Diabetes Care 18:1288-1290, 1995.

98. Hoffman W, Locksmith J, Burton E, et al. Interstitial pulmonaryedema in children and adolescents with diabetic ketoacidosis. J Dia-betes Complications 12:314-320, 1998.

99. Breidbart S, Singer L, St. Louis Y, et al. Adult respiratory distress

1

1

1

11

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

198 N. Glaser

syndrome in an adolescent with diabetic ketoacidosis. J Pediatr111:736-737, 1987.

00. Rosenbloom A. Intracerebral crises during treatment of diabetic keto-acidosis. Diabetes Care 13:22-33, 1990.

01. Gutierrez J, Bagatell R, Sampson M, et al. Femoral central venouscatheter-associated deep venous thrombosis in children with diabeticketoacidosis. Crit Care Med 31:80-83, 2003.

02. Malone J, Brodsky S. The value of electrocardiogram monitoring indiabetic ketoacidosis. Diabetes Care 3:543-547, 1980.

03. Tanel R. ECGs in the ED. Pediatr Emerg Care 20:849-851, 2004.04. Slyper A, Wyatt D, Brown C. Clinical and/or biochemical pancreatitis

in diabetic ketoacidosis. J Pediatr Endocrinol 7:261-264, 1994.05. Murdoch I, Pryor D, Haycock G, et al. Acute renal failure complicating

diabetic ketoacidosis. Acta Paediatr 82:498-500, 1993.06. Todani T, Sato Y, Watanabe Y, et al. Ischemic jejunal stricture devel-

oping after diabetic coma in a girl: a case report. Eur J Pediatr Surg3(2):115-117, 1993.

07. Chan-Cua S, Jones K, Lynch F, et al. Necrosis of the ileum in a diabeticadolescent. J Pediatr Surg 27:1236-1238, 1992.

08. Dimeglio L, Chaet M, Quigley C, et al. Massive ischemic intestinalnecrosis at the onset of diabetes mellitus with ketoacidosis in a three-year old girl. J Pediatr Surg 38:1537-9, 2003.

09. Khanna S, Soumekh B, Bradley J, et al. A case of fatal rhinocerebralmucormycosis with new onset diabetic ketoacidosis. J Diabetes Com-plications 12:224-227, 1998.

10. Dokmetas H, Canbay E, Yilmaz S, et al. Diabetic ketoacidosis andrhino-orbital mucormycosis. Diabetes Res Clin Pract 57:139-142,2002.

11. McCloud L, Parkerson J, Freant L, et al. Beta-hydroxybutyrate inducespulmonary endothelial dysfunction in rabbits. Exp Lung Res 30:193-206, 2004.

12. Carl G, Hoffman W, Passmore G, et al. Diabetic ketoacidosis promotesa prothrombotic state. Endocr Res 29:73-82, 2003.

13. Mustard J, Packham M. Platelets and diabetes mellitus. N Engl J Med311:665-667, 1984.

14. Timperley W, Preston F, Ward J. Cerebral intravascular coagulation indiabetic ketoacidosis. Lancet 18:952-956, 1974.

15. Worly J, Fortenberry J, Hansen I, et al. Deep venous thrombosis inchildren with diabetic ketoacidosis and femoral central venous cath-eters. Pediatrics 113:e57-60, 2004.

16. Bello F, Sotos J. Cerebral oedema in diabetic ketoacidosis in children.Lancet 336(8706):64, 1990.

17. Cummings E, Lawrence S, Daneman D. Cerebral edema (CE) in pe-diatric diabetic ketoacidosis (DKA) in Canada. Diabetes 52:A400,2003.

18. Glaser N, Gorges S, Marcin J, et al. Mechanism of cerebral edema inchildren with diabetic ketoacidosis. J Pediatr 145:164-171, 2004.

19. Hoffman W, Steinhart C, El Gammal T, et al. Cranial CT in childrenand adolescents with diabetic ketoacidosis. AJNR Am J Neuroradiol9:733-739, 1988.

20. Krane E, Rockoff M, Wallman J, et al. Subclinical brain swelling inchildren during treatment of diabetic ketoacidosis. N Engl J Med312:1147-1151, 1985.

21. Hale P, Rezvani I, Braunstein A, et al. Factors predicting cerebraledema in young children with diabetic ketoacidosis and new onsettype I diabetes. Acta Paediatr 86:626-631, 1997.

22. Mahoney C, Vlcek B, Del Aguila M. Risk factors for developing brainherniation during diabetic ketoacidosis. Pediatr Neurol 21:721-772,1999.

23. Glasgow A. Devastating cerebral edema in diabetic ketoacidosis beforetherapy. Diabetes Care 14(1):77-78, 1991.

24. Couch R, Acott P, Wong G. Early onset of fatal cerebral edema indiabetic ketoacidosis. Diabetes Care 14:78-79, 1991.

25. Lam T, Anderson S, Glaser N, et al. Bumetanide reduces cerebraledema formation in rats with diabetic ketoacidosis. Diabetes 54:510-516, 2005.

26. Isales C, Min L, Hoffman W. Acetoacetate and B-hydroxybutyrate

differentially regulate endothelin-1 and vascular endothelial growth

factor in mouse brain microvascular endothelial cells. J DiabetesComplications 13(2):91-97, 1999.

27. Harris G, Fiordalisi I, Harris W, et al. Minimizing the risk of brainherniation during treatment of diabetic ketoacidemia: a retrospectiveand prospective study. J Pediatr 117:22-31, 1990.

28. Roberts M, Slover R, Chase H. Diabetic ketoacidosis with intracerebralcomplications. Pediatr Diabetes 2:103-114, 2001.

29. Franklin B, Liu J, Ginsberg-Fellner F. Cerebral edema and ophthal-moplegia reversed by mannitol in a new case of insulin-dependentdiabetes mellitus. Pediatrics 69:87-90, 1982.

30. Marcin J, Glaser N, Barnett P, et al. Clinical and therapeutic factorsassociated with adverse outcomes in children with DKA-related cere-bral edema. J Pediatr 141:793-797, 2003.

31. Gibson K, Feigenbaum A. Phenotypically mild presentation in a pa-tient with 2-methylacetoacetylcoenzyme A (beta-keto) thiolase defi-ciency. J Inherit Metab Dis 20:712-713, 1997.

32. Longo N, Fukao T, Singh R, et al. Succinyl-CoA:3-ketoacid transferase(SCOT) deficiency in a new patient homozygous for an R217X muta-tion. J Inherit Metab Dis 27:691-692, 2004

33. Fukao T, Scriver C, Kondo N. The clinical phenotype and outcome ofmitochondrial acetoacetyl-CoA thiolase deficiency (beta-ketothiolaseor T2 deficiency) in 26 enzymatically proved and mutation-definedpatients. Mol Genet Metab 72:109-114, 2001.

34. Boulware S, Tamborlane W. Not all severe hyperglycemia is diabetes.Pediatrics 89:330-332, 1992.

35. Stevenson R, Bowyer F. Hyperglycemia with hyperosmolal dehydra-tion in nondiabetic infants. J Pediatr 77:818-823, 1970.

36. Mandell F, Fellers F. Hyperglycemia in hypernatremic dehydration.Clin Pediatr 13:367-369, 1974.

37. Perkins K. Hyperglycemia without ketosis or hypernatremia. Am J DisChild 128:885, 1974.

38. Rabinowitz L, Joffe B, Abkiewicz C, et al. Hyperglycemia in infantilegastroenteritis. Arch Dis Child 59:771-715, 1984.

39. Rother K, Schwenk W. An unusual case of the nonketotic hypergly-cemic syndrome during childhood. Mayo Clin Proc 70:62-65, 1995.

40. Morales A, Rosenbloom A. Death caused by hyperglycemic hyperos-molar state at the onset of type 2 diabetes. J Pediatr 144:270-273,2004.

41. Hollander A, Olney R, Blackett P, et al. Fatal malignant hyperthermia-like syndrome with rhabdomyolysis complicating the presentation ofdiabetes mellitus in adolescent males. Pediatrics 111:1447-1452,2003.

42. Gottschalk M, Ros S, Zeller W. The emergency management of hyper-glycemic-hyperosmolar nonketotic coma in the pediatric patient. Pe-diatr Emerg Care 12:48-51, 1996.

43. Maclsaac R, Lee L, McNeil K, et al. Influence of age on the presentationand outcome of acidotic and hyperosmolar diabetic emergencies. IntMed J 32:379-385, 2002.

44. Chupin M, Charbonnel B, Chupin F. C-peptide levels in keto-acidosisand in hyperosmolar non-ketotic diabetic coma. Acta Diabetol 18:123-128, 1981.

45. Wachtel T, Silliman R, Lamberton P. Predisposing factors of the dia-betic hyperosmolar state. Arch Intern Med 147:499-501, 1987.

46. Kitabchi A, Umpierrez G, Murphy M, et al. Management of hypergly-cemic crises in patients with diabetes. Diabetes Care 24:131-153,2001.

47. Gerich J, Martin M, Recant L. Clinical and metabolic characteristics ofhyperosmolar nonketotic coma. Diabetes 20:228-238, 1971.

48. Lai P, Tien R, Chang M, et al. Chorea-ballismus with nonketotichyperglycemia in primary diabetes mellitus. AJNR Am J Neuroradiol17:1957-1964, 1996.

49. Tanaka S, Kobayashi T, Kawanami D, et al. Paradoxical glucose infu-sion for hypermatremia in diabetic hyperglycaemic hyperosmolarsyndrome. J Int Med 248:166-167, 2000.

50. Fourtner SH, Weinzimer SA, Levitt-Katz LE. Hyperglycemic hyperos-molar nonketotic syndrome in children with type 2 diabetes. PediatrDiabetes 6:129-135, 2005.

51. Kian K, Eiger G. Anticoagulant therapy in hyperosmolar non-ketotic

diabetic coma. Diabetes Med 20:603, 2003.