Está en la página 1de 10

Reviews/Commentaries/ADA Statements

C O N S E N S U S S T A T E M E N T

Hyperglycemic Crises in Adult Patients


With Diabetes
A consensus statement from the American Diabetes Association
ABBAS E. KITABCHI,PHD, MD 1 MARY BETH MURPHY, RN, MS, CDE, MBA1 glycemia and hyperosmolality of the ex-
GUILLERMO E. UMPIERREZMD, 2 ROBERT A. KREISBERG, MD3 tracellular space (1,3,10 –17). The com-
bination of insulin deficiency and
increased counterregulatory hormones in
iabetic ketoacidosis (DKA) and hy- PATHOGENESIS — Although the DKA also leads to the release of free fatty

D perosmolar hyperglycemic state


(HHS) are the two most serious
acute metabolic complications of diabe-
pathogenesis of DKA is better understood
than that of HHS, the basic underlying
mechanism for both disorders is a reduc-
tion in the net effective action of circulat-
acids into the circulation from adipose tis-
sue (lipolysis) and to unrestrained hepatic
fatty acid oxidation to ketone bodies (þ-
hydroxybutyrate ([þ-OHB] and acetoace-
tes. Most patients with DKA have autoim-
mune type 1 diabetes; however, patients ing insulin coupled with a concomitant tate), with resulting ketonemia and
with type 2 diabetes are also at risk during elevation of counterregulatory hormones, metabolic acidosis (18). On the other
the catabolic stress of acute illness such as such as glucagon, catecholamines, corti- hand, HHS may be caused by plasma in-
trauma, surgery, or infection. Table 1 out- sol, and growth hormone (1,3,4,8 –13). sulin concentrations that are inadequate
lines the diagnostic criteria and electro- DKA and HHS can fall anywhere along to facilitate glucose utilization by insulin-
lyte and fluid deficits for both disorders. the disease continuum of diabetic meta- sensitive tissues but adequate (as deter-
bolic derangements. At one extreme, pure mined by residual C-peptide) to prevent
The mortality rate in patients with DKA is
lipolysis and subsequent ketogenesis
<5% in experienced centers, whereas the DKA without significant hyperosmolarity
typically indicates the total or relative ab- (15). Both DKA and HHS are associated
mortality rate of patients with HHS still
with glycosuria, leading to osmotic diure-
remains high at ~11% (1– 8). Death in sence of insulin (seen in type 1 diabetes).
sis, with loss of water, sodium, potassium,
these conditions is rarely due to the met- At the other extreme, HHS without keto-
and other electrolytes (6,15–17). The
abolic complications of hyperglycemia or acidosis typically occurs with lesser de-
pathogenic pathways of DKA and HHS
ketoacidosis but rather relates to the un- grees of insulin deficiency, as seen in type are depicted in Fig. 1. The diagnostic cri-
derlying precipitating illness. The prog- 2 diabetes. However, in most circum- teria and typical total deficits of water and
nosis of both conditions is substantially stances, a mixed presentation occurs de- electrolytes in DKA and HHS are summa-
worsened at the extremes of age and in the pending on the duration of symptoms, rized in Table 1. As can be seen, DKA and
presence of coma and hypotension (7,9 – coexisting medical illnesses, or underly- HHS differ in the magnitude of dehydra-
11) ing precipitating cause. In one study (14), tion, ketosis, and acidosis.
This consensus statement will outline 123 DKA laboratory admission profiles DKA is a proinflammatory state pro-
precipitating factors and recommenda- were reviewed, and 37% demonstrated an ducing reactive oxygen species that are
tions for the diagnosis, treatment, and elevated total osmolality. indicative of oxidative stress. A recent
prevention of DKA and HHS in adult sub- Hormonal alterations in DKA and study (19) has shown elevated levels of
jects. It is based on a previous technical HHS lead to increased gluconeogenesis proinflammatory cytokines and lipid per-
review and more recently published peer- and hepatic and renal glucose production oxidation markers, as well as cardiovas-
reviewed articles since 2001, which should and impaired glucose utilization in pe- cular risk factors (plasminogen activator
be consulted for further information. ripheral tissues, which results in hyper- inhibitor-1) and C-reactive protein,
which return to normal levels with insulin
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● therapy and remission of hyperglycemia.
From the 1Division of Endocrinology, Diabetes, and Metabolism, University of Tennessee Health Science
Center, Memphis, Tennessee; the 2Clinical Research Center and Diabetes Unit, Grady Memorial Hospital/
Emory University School of Medicine; and the 3Teaching Faculty, University of South Alabama, Baptist PRECIPITATING FACTORS —
Health System, Birmingham, Alabama. The two most common precipitating fac-
Address correspondence and reprint requests to Dr. Abbas E. Kitabchi, Director, Division of Endocrinol-
ogy, Diabetes and Metabolism, University of Tennessee Health Science Center, 956 Court Ave., Suite D334,
tors in the development of DKA or HHS
Memphis, Tennessee 38163. E-mail: akitabchi@utmem.edu. are inadequate or inappropriate insulin
The initial draft of this position statement was prepared by the authors as listed above. The manuscript was therapy or infection (1,4,8 –12). Other
then peer-reviewed, modified, and approved by the Professional Practice Committee and the Executive precipitating factors include pancreatitis,
Committee, March 2006. myocardial infarction, cerebrovascular
Abbreviations: þ-OHB, þ-hydroxybutyrate; DKA, diabetic ketoacidosis; HHS, hyperosmolar hypergly-
cemic state. accident, and drugs. In addition, new-
The recommendations in this article are based on the evidence reviewed in the following publication: onset type 1 diabetes or discontinuation
Management of hyperglycemic crises in patients with diabetes (Technical Review). Diabetes Care 24:131– of insulin in established type 1 diabetes
153, 2001, as well as subsequent peer-reviewed publications since 2001. commonly leads to the development of
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion
factors for many substances. DKA. Underlying medical illness such as
DOI: 10.2337/dc06-9916 stroke or myocardial infarction that pro-
© 2006 by the American Diabetes Association. vokes the release of counterregulatory

DIABETES CARE, VOLUME 29, NUMBER 12, DECEMBER 2006 2739


Hyperglycemic crises in diabetic adults

Table 1—Diagnostic criteria and typical total body deficits of water and electrolytes in DKA and HHS

DKA
Mild Moderate Severe HHS
Diagnostic criteria and classification
Plasma glucose (mg/dl) >250 mg/dl >250 mg/dl >250 mg/dl >600 mg/dl
Arterial pH 7.25–7.30 7.00 to <7.25 <7.00 >7.30
Serum bicarbonate (mEq/l) 15–18 10 to <15 <10 >15
Urine ketone* Positive Positive Positive Small
Serum ketone* Positive Positive Positive Small
Effective serum osmolality† Variable Variable Variable >320 mOsm/kg
Anion gap‡ >10 >12 >12 <12
Mental status Alert Alert/drowsy Stupor/coma Stupor/coma
Typical deficits
Total water (l) 6 9
Water (ml/kg)§ 100 100–200
Na+ (mEq/kg) 7–10 5–13
Cl— (mEq/kg) 3–5 5–15
K+ (mEq/kg) 3–5 4–6
PO4 (mmol/kg) 5–7 3–7
Mg++ (mEq/kg) 1–2 1–2
Ca++ (mEq/kg) 1–2 1–2
*Nitroprusside reaction method. †Calculation of effective serum osmolality: 2[measured Na+ (mEq/l)]+ [glucose (mg/dl)]/18. ‡Calculation of anion gap: (Na+) —
[Cl— + HC03 — (mEq/l)]. §Per kg body wt. Data adapted from refs. 1, 4, and 7.

hormones and/or compromises the access derly individuals with new-onset diabetes second-generation antipsychotics agents
to water is likely to result in severe dehy- (particularly residents of chronic care fa- (20) may precipitate the development of
dration and HHS. In most patients, re- cilities) or individuals with known diabe- HHS or DKA. In young patients with type
stricted water intake is due to the patient tes who become hyperglycemic and are 1 diabetes, psychological problems com-
being bedridden or restrained and is ex- unaware of it or are unable to take fluids plicated by eating disorders may be a con-
acerbated by the altered thirst response of when necessary are at risk for HHS tributing factor in 20% of recurrent
the elderly. Because 20% of these patients (8,11,12). Drugs that affect carbohydrate ketoacidosis (21). Factors that may lead
have no history of diabetes, delayed rec- metabolism, such as corticosteroids, thia- to insulin omission in younger patients
ognition of hyperglycemic symptoms zides, and sympathomimetic agents (e.g., include fear of weight gain with improved
may have led to severe dehydration. El- dobutamine and terbutaline) (10) and metabolic control, fear of hypoglycemia,
rebellion from authority, and the stress of
chronic disease. Before 1993, the use of
continuous subcutaneous insulin infu-
sion devices had also been associated with
an increased frequency of DKA (22), but
with improvement in technology and bet-
ter education of patients, the incidence of
DKA appears to have reduced in pump
users. However, additional prospective
studies are needed to document reduc-
tion of DKA incidence with the use of con-
tinuous subcutaneous insulin infusion
devices (23).
During the past decade, an increasing
number of DKA cases without precipitat-
ing cause have been reported in children,
adolescents, and adult subjects with type
2 diabetes. Observational and prospective
studies indicate that over half of newly
diagnosed adult African-American and
Hispanic subjects with unprovoked DKA
have type 2 diabetes (24 –27). In such pa-
tients, clinical and metabolic features of
Figure 1— Pathogenesis of DKA and HHS, stress, infection, and/or insufficient insulin. ++Ac- type 2 diabetes include a high rate of obe-
celerated pathway (ref. 10). sity, a strong family history of diabetes, a

2740 DIABETES CARE, VOLUME 29, NUMBER 12, DECEMBER 2006


Kitabchi and Asssociates

measurable pancreatic insulin reserve, icking an acute abdomen, is present in and the presence of altered mental status
low prevalence of autoimmune markers 50 –75% of DKA cases (33,34). The ab- (1). Significant overlap between DKA and
of þ-cell destruction, and the ability to dominal pain usually resolves with cor- HHS has been reported in more than one-
discontinue insulin therapy during fol- rection of hyperglycemia and metabolic third of patients (1,2,14). Although most
low-up (28,29). This variant of type 2 di- acidosis. The most common clinical pre- patients with HHS have an admission pH
abetes has been referred to in the sentation in patients with HHS is altered >7.30, a bicarbonate level >20 mEq/l,
literature as idiopathic type 1 diabetes, sensorium (4,8,11,12). Physical examina- mild ketonemia may be present.
atypical diabetes, Flatbush diabetes, type tion reveals signs of dehydration with loss The majority of patients with hyper-
1.5 diabetes, and more recently as keto- of skin turgor, weakness, tachycardia, glycemic emergencies present with leuko-
sis-prone type 2 diabetes (24,30). At pre- and hypotension. Fever due to underly- cytosis proportional to blood ketone body
sentation, they have markedly impaired ing infection is common, and signs of ac- concentration (2,10). However, leukocy-
insulin secretion and insulin action idosis (Kussmaul breathing, acetone tosis >25,000 may designate infection
(25,26,29), but aggressive management breath) are usually absent. In some pa- and require further evaluation (36). The
with insulin significantly improves þ-cell tients, focal neurologic signs (hemipare- admission serum sodium is usually low
function, allowing discontinuation of in- sis, hemianopsia) and seizures (partial because of the osmotic flux of water from
sulin therapy within a few months of fol- motor seizures more common than gen- the intracellular to the extracellular space
low-up (25–27). Recently, it was reported eralized) may be the dominant clinical in the presence of hyperglycemia. An in-
that the near-normoglycemic remission is features (1,4,6,8). crease in serum sodium concentration in
associated with a greater recovery of basal the presence of hyperglycemia indicates a
and stimulated insulin secretion and that Laboratory ftndings rather profound degree of water loss. Un-
10 years after diabetes onset, 40% of pa- The initial laboratory evaluation of pa- less the plasma is cleared of chylomi-
tients with ketosis-prone type 2 diabetes tients with suspected DKA or HHS should crons, pseudonormoglycemia and
are still non–insulin dependent (24 –27). include determination of plasma glucose, pseudohyponatremia may occur in DKA
Furthermore, a novel genetic mecha- blood urea nitrogen, creatinine, serum (37,38). Serum potassium concentration
nism related to the high prevalence of glu- ketones, electrolytes (with calculated an- may be elevated because of an extracellu-
co se -6 -p ho sp hate dehyd ro genase ion gap), osmolality, urinalysis, urine ke- lar shift of potassium caused by insulin
deficiency has been linked with ketosis- tones by dipstick, as well as initial arterial deficiency, hypertonicity, and acidemia
prone diabetes (31). blood gases and complete blood count (3,10,39).
with differential. An electrocardiogram, Patients with low normal or low se-
DIAGNOSIS chest X-ray, and urine, sputum, or blood rum potassium concentration on admis-
cultures should also be obtained, if clini- sion have severe total-body potassium
History and physical examination cally indicated. HbA1c may be useful in deficiency and require very careful car-
The process of HHS usually evolves over determining whether this acute episode is diac monitoring and more vigorous po-
several days to weeks, whereas the evolu- the culmination of an evolutionary pro- tassium replacement, because treatment
tion of the acute DKA episode in type 1 cess in previously undiagnosed or poorly lowers potassium further and can pro-
diabetes or even in type 2 diabetes tends controlled diabetes or a truly acute epi- voke cardiac dysrhythmia. The classic
to be much shorter. Although the symp- sode in an otherwise well-controlled pa- work of Atchley et al. (40) established that
toms of poorly controlled diabetes may be tient. The diagnostic criteria for DKA and the total body deficit of sodium and po-
present for several days, the metabolic al- HHS are shown in Table 1. tassium might be as high as 500 –700
terations typical of ketoacidosis usually DKA consists of the biochemical triad mEq (39,40).
evolve within a short time frame (typically of hyperglycemia, ketonemia, and meta- Studies on serum osmolality and
<24 h). The classic clinical picture of pa- bolic acidosis. Accumulation of ketoacids mental alteration have established a posi-
tients with DKA includes a history of results in an increased anion gap meta- tive linear relationship between osmolal-
polyuria, polydipsia, weight loss, vomit- bolic acidosis. The anion gap is calculated ity and mental obtundation (14). The
ing, abdominal pain, dehydration, weak- by subtracting the sum of chloride and occurrence of stupor or coma in diabetic
ness, mental status change, and coma. bicarbonate concentration from the so- patients in the absence of definitive eleva-
Physical findings may include poor skin dium concentration [Na+ — (Cl— + tion of effective osmolality (320 mOsm/
turgor, Kussmaul respirations, tachycar- HCO3—)]. The normal anion gap has been kg) demands immediate consideration of
dia, hypotension, alteration in mental sta- historically reported to be <12 ± 2 other causes of mental status change. In
tus, shock, and ultimately coma. Up to mEq/l. Most laboratories, however, cur- the calculation of effective osmolality
25% of DKA patients have emesis, which rently measure sodium and chloride con- {2[measured Na (mEq/l)] + [glucose
may be coffee-ground in appearance and centrations using ion-specific electrodes, (mg/dl)]/18}, the urea concentration is
guaiac positive. Mental status can vary which measure plasma chloride concen- not taken into account because it is freely
from full alertness to profound lethargy or tration 2– 6 mEq/l higher than with prior permeable and its accumulation does not
coma, with the latter more frequent in methods (35). Thus, the normal anion induce major changes in intracellular vol-
HHS. Although infection is a common gap using the current methodology is be- ume or osmotic gradient across the cell
precipitating factor for both DKA and tween 7 and 9 mEq/l, and an anion gap membrane (4).
HHS, patients can be normothermic or >10 –12 mEq/l indicates the presence of Amylase levels are elevated in the ma-
even hypothermic primarily because of increased anion gap acidosis. The severity jority of patients with DKA, but this may
peripheral vasodilation (32). Severe hy- of DKA is classified as mild, moderate, or be due to nonpancreatic sources, such as
pothermia, if present, is a poor prognostic severe based on the severity of metabolic the parotid gland (41). A serum lipase de-
sign. Abdominal pain, sometimes mim- acidosis (blood pH, bicarbonate, ketones) termination may be beneficial in the dif-

DIABETES CARE, VOLUME 29, NUMBER 12, DECEMBER 2006 2741


Hyperglycemic crises in diabetic adults

ferential diagnosis of pancreatitis; wt · h—1 or 1–1.5 l during the first hour. insulin analogs in the management of pa-
however, lipase could also be elevated in The subsequent choice for fluid replace- tients with uncomplicated DKA. Patients
DKA. Finally, abnormal acetoacetate lev- ment depends on the state of hydration, treated with subcutaneous rapid-acting
els may falsely elevate serum creatinine if serum electrolyte levels, and urinary out- insulin received an initial injection of 0.2
the clinical laboratory uses a colorometric put. In general, 0.45% NaCl infused at units/kg followed by 0.1 unit/kg every
method for the creatinine assay (42). 4 –14 ml · kg—1 body wt · h—1 is appro- hour or an initial dose of 0.3 units/kg fol-
priate if the corrected serum sodium is lowed by 0.2 units/kg every 2 h until
Differential diagnosis normal or elevated; 0.9% NaCl at a similar blood glucose was <250 mg/dl, then the
Not all patients with ketoacidosis have rate is appropriate if corrected serum so- insulin dose was decreased by half to 0.05
DKA. Starvation ketosis and alcoholic ke- dium is low (Fig. 2). Successful progress or 0.1 unit/kg, respectively, and adminis-
toacidosis are distinguished by clinical with fluid replacement is judged by he- tered every 1 or 2 h until resolution of
history and by plasma glucose concentra- modynamic monitoring (improvement in DKA (52,53). There were no differences
tions that range from mildly elevated blood pressure), measurement of fluid in- in length of hospital stay, total amount of
(rarely >200 mg/dl) to hypoglycemia. In put and output, laboratory values, and insulin administration until resolution of
addition, although alcoholic ketoacidosis clinical examination. Fluid replacement hyperglycemia or ketoacidosis, or num-
can result in profound acidosis, the serum should correct estimated deficits within ber of hypoglycemic events among treat-
bicarbonate concentration in starvation the first 24 h. In patients with renal or ment groups. In addition, the use of
ketosis is usually not <18 mEq/l. DKA cardiac compromise, monitoring of se- insulin analogs allowed treatment of DKA
must also be distinguished from other rum osmolality and frequent assessment in general wards or in the emergency de-
causes of high anion gap metabolic acido- of cardiac, renal, and mental status must partment, avoiding admission to an inten-
sis, including lactic acidosis; ingestion of be performed during fluid resuscitation to sive care unit. By avoiding intensive care
drugs such as salicylate, methanol, ethyl- a voi d i a t r o g e n i c fluid o v e r l o a d admissions, these investigators reported a
ene glycol, and paraldehyde; and chronic (1,3,4,10,12,16,17). Adequate rehydra- reduction of 30% in the cost of hospital-
renal failure. tion with subsequent correction of the hy- ization (51–54).
A clinical history of previous drug perosmolar state has been shown to result Ketonemia typically takes longer to
abuse or metformin use should be sought. in a more robust response to low-dose in- clear than hyperglycemia. Direct mea-
Measurement of blood lactate, serum sa- sulin therapy (49). surement of þ-OHB in the blood is the
licylate, and blood methanol level can be preferred method for monitoring DKA
helpful in these situations. Ethylene gly- Insulin therapy and has become more convenient with
col (antifreeze) is suggested by the pres- Unless the episode of DKA is uncompli- the recent development of bedside meters
ence of calcium oxalate and hippurate cated and mild/moderate (Table 1), regu- capable of measuring whole-blood
crystals in the urine. Paraldehyde inges- lar insulin by continuous intravenous þ-OHB (55). The nitroprusside method,
tion is indicated by its characteristic infusion is the treatment of choice. In which is used in clinical chemistry labo-
strong odor on the breath. Because these adult patients, once hypokalemia (K+ < ratories, measures acetoacetic acid and
intoxicants are low–molecular-weight or- 3.3 mEq/l) is excluded, an intravenous acetone; however, þ-OHB, the strongest
ganic compounds, they can produce an bolus of regular insulin at 0.1 unit/kg and most prevalent acid in DKA, is not
osmolar gap in addition to the anion gap body wt, followed by a continuous infu- measured by the nitroprusside method.
acidosis (10,43). A recent report (44) sug- sion of regular insulin at a dose of 0.1 unit During therapy, þ-OHB is converted to
gested a relationship between low carbo- · kg—1 · h—1 should be administered. This acetoacetic acid, which may lead the cli-
hydrate dietary intake and metabolic low dose of insulin usually decreases nician to believe that ketosis has wors-
acidosis. plasma glucose concentration at a rate of ened (42). Therefore, assessments of
Finally, four case reports have shown 50 –75 mg · dl—1 · h—1, similar to a higher- urinary or serum ketone levels by the ni-
that patients with undiagnosed acromeg- dose insulin regimen (50,51). If plasma troprusside method should not be used as
aly may present with DKA as the primary glucose does not decrease by 50 –75 mg an indicator of response to therapy. Dur-
manifestation of their disease (45– 48). from the initial value in the first hour, the ing therapy for DKA or HHS, blood
insulin infusion may be doubled every should be drawn every 2– 4 h for deter-
TREATMENT — Successful treatment hour until a steady glucose decline is mination of serum electrolytes, glucose,
of DKA and HHS requires correction of achieved. When the plasma glucose blood urea nitrogen, creatinine, osmolal-
dehydration, hyperglycemia, and electro- reaches 200 mg/dl in DKA or 300 mg/dl ity, and venous pH (for DKA). Generally,
lyte imbalances; identification of comor- in HHS, it may be possible to decrease the repeat arterial blood gases are unneces-
bid precipitating events; and above all, insulin infusion rate to 0.05– 0.1 unit · sary during the treatment of DKA in he-
frequent patient monitoring. Protocols kg—1 · h—1, at which time dextrose may be modynamically stable patients. Since
for the management of patients with DKA added to the intravenous fluids (1,4,10). venous pH is only 0.02– 0.03 units lower
and HHS are summarized in Figs. 2 and 3. Thereafter, the rate of insulin administra- than arterial pH (56), it is adequate to as-
tion or the concentration of dextrose may sess venous pH response to therapy, thus
Fluid therapy need to be adjusted to maintain the avoiding the pain and potential complica-
Initial fluid therapy is directed toward ex- above-glucose values until acidosis in tions associated with repeated arterial
pansion of the intravascular and extra vas- DKA or mental obtundation and hyperos- punctures.
cular volume and restoration of renal molality in HHS are resolved. Criteria for resolution of DKA include
perfusion. In the absence of cardiac com- Prospective and randomized studies glucose <200 mg/dl, serum bicarbonate
promise, isotonic saline (0.9% NaCl) is have reported on the efficacy and cost ef- Š18 mEq/l, and venous pH >7.3. When
infused at a rate of 15–20 ml · kg—1 body fectiveness of subcutaneous rapid-acting the patient is able to eat, a multiple-dose

2742 DIABETES CARE, VOLUME 29, NUMBER 12, DECEMBER 2006


Kitabchi and Asssociates

2743
DIABETES CARE, VOLUME 29, NUMBER 12, DECEMBER 2006
Figure 2— Protocol for the management of adult patients with DKA. *DKA diagnostic criteria: serum glucose >250 mg/dl, arterial pH <7.3, serum bicarbonate <18 mEq/l,
and moderate ketonuria or ketonemia. Normal laboratory values vary; check local lab normal ranges for all electrolytes. †After history and physical exam, obtain capillary
glucose and serum or urine ketones (nitroprusside method). Begin 1 liter of 0.9% NaCl over 1 h and draw arterial blood gases, complete blood count with differential, urinalysis,
serum glucose, BUN, electrolytes, chemistry profile, and creatinine levels STAT. Obtain electrocardiogram, chest X-ray, and specimens for bacterial cultures, as needed. *Serum
Na+ should be corrected for hyperglycemia (for each 100 mg/dl glucose >100 mg/dl, add 1.6 mEq to sodium value for corrected serum sodium value). Adapted from ref. 1.
Hyperglycemic crises in diabetic adults

Figure 3— Protocol for the management of adult patients with HHS. HHS diagnostic criteria: serum glucose >600 mg/dl, arterial pH >7.3, serum
bicarbonate >15 mEq/l, and minimal ketonuria and ketonemia. Normal laboratory values vary; check local lab normal ranges for all electrolytes.
†After history and physical exam, obtain capillary glucose and serum or urine ketones (nitroprusside method). Begin 1 liter of 0.9% NaCl over 1 h
and draw arterial blood gases, complete blood count with differential, urinalysis, serum glucose, BUN, electrolytes, chemistry profile and creatinine
levels STAT. Obtain electrocardiogram, chest X-ray, and specimens for bacterial cultures, as needed. Adapted from ref. 1. *Serum Na+ should be
corrected for hyperglycemia (for each 100 mg/dl glucose >100 mg/dl, add 1.6 mEq to sodium value for corrected serum sodium value).

insulin schedule should be started that should be started at a dose of 0.5– 0.8 ml/h). Generally, 20 –30 mEq potassium
uses a combination of short- or rapid- units · kg—1 · day—1, including regular in each liter of infusion fluid is sufficient
acting and intermediate- or long-acting or rapid-acting and basal insulin until an to maintain a serum potassium concentra-
insulin as needed to control plasma glu- op- timal dose is established. However, tion within the normal range of 4 –5
cose. Intravenous insulin infusion should good clinical judgment and frequent mEq/l. Rarely, DKA patients may present
be continued for 1–2 h after the subcuta- glucose assessment are vital in initiating with significant hypokalemia. In such
neous insulin is given to ensure adequate a new in- sulin regimen in insulin-naı¨ve cases, potassium replacement should be-
plasma insulin levels. An abrupt discon- patients. gin with fluid therapy, and insulin treat-
tinuation of intravenous insulin coupled ment should be delayed until potassium
with a delayed onset of a subcutaneous Potassium concentration is restored to >3.3 mEq/l
insulin regimen may lead to hyperglyce- Despite total-body potassium depletion to avoid arrhythmias or cardiac arrest and
mia or recurrence of ketoacidosis. If the (40,57), mild to moderate hyperkalemia respiratory muscle weakness (57–58).
patient is to remain n.p.o., it is preferable is not uncommon in patients with hyper-
to continue the intravenous insulin infu- glycemic crises. Insulin therapy, correc- Bicarbonate
sion and fluid replacement. Patients with tion of acidosis, and volume expansion Bicarbonate use in DKA remains contro-
known diabetes may be given insulin at decrease serum potassium concentration. versial (58). At a pH >7.0, administration
the dose they were receiving before the To prevent hypokalemia, potassium re- of insulin blocks lipolysis and resolves ke-
onset of DKA or HHS. In insulin-naı¨ve placement is initiated after serum levels toacidosis without any added bicarbonate
patients, a multidose insulin regimen decrease to <5.3 mEq/l, assuming the (3,18,40). However, the administration
presence of adequate urine output at 50
2744 DIABETES CARE, VOLUME 29, NUMBER 12, DECEMBER 2006
Kitabchi and Asssociates

of bicarbonate may be associated with serum phosphate concentration <1.0 rebral water diffusion and cerebral vascu-
several deleterious effects including an in- mg/dl (66,67). When needed, 20 –30 lar perfusion during the treatment of 14
creased risk of hypokalemia (59), de- mEq/l potassium phosphate can be added children with DKA found that the cere-
creased tissue oxygen uptake, and to replacement fluids. bral edema was not a function of cerebral
cerebral edema (60,61). A prospective tissue edema but rather a function of in-
randomized study in 21 patients failed to COMPLICATIONS — The most creased cerebral perfusion. There is a lack
show either beneficial or deleterious common complications of DKA and HHS of information on the morbidity associ-
changes in morbidity or mortality with include hypoglycemia and hypokalemia ated with cerebral edema in adult pa-
bicarbonate therapy in DKA patients with due to overzealous treatment with insulin. tients; therefore, any recommendations
an admission arterial pH between 6.9 and Low potassium may also occur as a result of for adult patients are based on clinical
7.1 (62). This study was small and limited treatment of acidosis with bicarbonate. Hy- judgment rather than scientific evidence.
to those patients with an admission arte- perglycemia may occur secondary to inter- Preventive measures that might decrease
rial pH of >6.9. The average pH in the ruption/discontinuance of intravenous the risk of cerebral edema in high-risk pa-
bicarbonate group was 7.03 ± 0.1 and for insulin therapy after recovery from DKA but tients are gradual replacement of sodium
the nonbicarbonate group was 7.0 ± without subsequent coverage with subcuta- and water deficits in patients who are hy-
0.02. Therefore, if the pH is 6.9 –7.0, it neous insulin. Commonly, patients recov- perosmolar and the addition of dextrose
seems prudent to administer 50 mmol bi- ering from DKA develop a transient to the hydrating solution once blood glu-
carbonate in 200 ml of sterile water with hyperchloremic non–anion gap acidosis cose reaches 200 mg/dl in DKA and 300
10 mEq KCL over 1 h until the pH is (68 –70). The hyperchloremic acidosis is mg/dl in HHS. In HHS, a glucose level of
>7.0. No prospective randomized stud- caused by the loss of large quantities of ke- 250 –300 mg/dl should be maintained
ies concerning the use of bicarbonate in toanions that occur during the develop- until hyperosmolarity and mental status
DKA with pH values <6.9 have been re- ment of DKA. Because ketoanions are improves and the patient becomes clini-
ported. Given that severe acidosis may metabolized with regeneration of bicarbon- cally stable. Hypoxemia and, rarely, non-
lead to a myriad of adverse vascular ef- ate, the prior loss of ketoacid anions in the cardiogenic pulmonary edema may
fects, adult patients with a pH <6.9 urine hinders regeneration of bicarbonate complicate the treatment of DKA. Hypox-
should receive 100 mmol sodium bicar- during treatment (71). Other mechanisms emia is attributed to a reduction in colloid
bonate (two ampules) in 400 ml sterile include the administration of intravenous osmotic pressure that results in increased
water (an isotonic solution) with 20 mEq fluids containing chloride that exceeds the lung water content and decreased lung
KCl administered at a rate of 200 ml/h for plasma chloride concentration and the in- compliance (10). Patients with DKA who
2 h until the venous pH is >7.0. Bicar- tracellular shifts of NaHCO3 during correc- have a widened alveolo-arteriolar oxygen
bonate as well as insulin therapy lowers tion of DKA (70). gradient noted on initial blood gas mea-
serum potassium; therefore, potassium Cerebral edema is a rare but fre- surement or with pulmonary rales on
supplementation should be maintained in quently fatal complication of DKA, occur- physical examination appear to be at
the intravenous fluid as described above ring in 0.7–1.0% of children with DKA. It higher risk for the development of pulmo-
and carefully monitored. (See Fig. 2 for is most common in children with newly nary edema.
guidelines.) Thereafter, venous pH diagnosed diabetes, but it has been re-
should be assessed every 2 h until the pH ported in children with known diabetes PREVENTION — Many cases of DKA
rises to 7.0, and treatment should be re- and in young people in their twenties and HHS can be prevented by better ac-
peated every 2 h if necessary. See refer- (72–74). Fatal cases of cerebral edema cess to medical care, proper education,
ence 1 for further review. have also been reported with HHS. Clin- and effective communication with a health
ically, cerebral edema is characterized by care provider during an intercurrent illness.
Phosphate deterioration in the level of conscious- The observation that stopping insulin for
Despite whole-body phosphate deficits in ness, lethargy, decreased arousal, and economic reasons is a common precipi-
DKA that average 1.0 mmol · kg—1 · headache. Neurological deterioration tant of DKA in urban African Americans
body wt—1, serum phosphate is often may be rapid, with seizures, inconti- and Hispanics (2,76,77) underscores the
normal or increased at presentation. nence, pupillary changes, bradycardia, need for our health care delivery systems
Phosphate con- centration decreases with and respiratory arrest. These symptoms to address this problem, which is costly
insulin therapy. Prospective randomized progress as brain stem herniation occurs. and clinically serious (78). Sick-day man-
studies (63,64) have failed to show any The progression may be so rapid that pap- agement should be reviewed periodically
beneficial effect of phosphate illedema is not found. Once the clinical with all patients. It should include spe-
replacement on the clinical outcome in symptoms other than lethargy and behav- cific information on 1) when to contact
DKA, and overzealous phos- phate ioral changes occur, mortality is high the health care provider, 2) blood glucose
therapy can cause severe hypocal- cemia (>70%), with only 7–14% of patients re- goals and the use of supplemental short-
(63,65). Therefore, the routine use of covering without permanent morbidity. or rapid-acting insulin during illness, 3)
phosphate in the treatment of DKA or Although the mechanism of cerebral means to suppress fever and treat infec-
HHS has resulted in no clinical benefit to edema is not known, it may result from tion, and 4) initiation of an easily digest-
the patient (63). However, to avoid car- osmotically driven movement of water ible liquid diet containing carbohydrates
diac and skeletal muscle weakness and into the central nervous system when and salt. Most importantly, the patient
respiratory depression due to hypophos- plasma osmolality declines too rapidly should be advised to never discontinue
phatemia, careful phosphate replacement with the treatment of DKA or HHS (72– insulin and to seek professional advice
may sometimes be indicated in patients 74). However, a recent study (75) using early in the course of the illness. Success-
with cardiac dysfunction, anemia, or re- magnetic resonance imaging to assess ce- ful sick-day management depends on in-
spiratory depression and in those with a
DIABETES CARE, VOLUME 29, NUMBER 12, DECEMBER 2006 2745
Hyperglycemic crises in diabetic adults

volvement by the patient and/or a family that active cocaine use is an independent diabetic ketoacidosis: physiologic versus
member. The patient/family member risk factor for recurrent DKA (81). pharmacologic doses of insulin and their
must be able to accurately measure and routes of administration. In Handbook of
record blood glucose, urine, or blood ke- Diabetes Mellitus. Brownlee M, Ed. New
tone determination when blood glucose is Acknowledgments — Studies cited by the au- York, Garland ATPM, 1981, p. 95–149
thors were supported in part by USPHS grants 15. Chupin M. Charbonnel B, Chupin F: C-
>300 mg/dl; insulin administered; tem- RR00211 (to the General Clinical Research peptide blood levels in ketoacidosis and
perature; respiratory and pulse rates; and Center) and AM 21099, training grant AM in hyperosmolar non-ketotic diabetic
body weight, and must be able to commu- 07088 of the National Institutes of Health, and coma. ACTA Diabetol 18:123–128, 1981
nicate all of this to a health care profes- grants from Novo-Nordisk, Eli Lilly, the 16. Hillman K: Fluid resuscitation in diabetic
sional. Adequate supervision and help American Diabetes Association, and the Abe emergencies: a reappraisal. Intensive Care
from staff or family may prevent many of Goodman Fund. Med 13:4 – 8, 1987
the admissions for HHS due to dehydra- 17. Delaney MF, Zisman A, Kettyle WM: dia-
tion among elderly individuals who are betic ketoacidosis and hyperglycemic
unable to recognize or treat this evolving References hyperosmolar nonketotic syndrome. Endo-
1. Kitabchi AE, Umpierrez GE, Murphy MB, crinol Metab Clin North Am 29:683–705,
condition. Better education of caregivers
Barrett EJ, Kreisberg RA, Malone JI, Wall 2000
as well as patients regarding signs and BM: Management of hyperglycemic crises 18. McGarry JD, Woeltje KF, Kuwajima M,
symptoms of new-onset diabetes; condi- in patients with diabetes (Technical Re- Foster DW: Regulation of ketogenesis and
tions, procedures, and medications that view). Diabetes Care 24:131–153, 2001 the renaissance of carnitine palmitoyl-
worsen diabetes control; and the use of 2. Umpierrez GE, Kelly JP, Navarrete JE, transferase. Diabete Metab Rev 5:271–284,
glucose monitoring could potentially de- Casals MM, Kitabchi AE: Hyperglycemic 1989
crease the incidence and severity of HHS. crises in urban blacks. Arch Intern Med 19. Stentz FB, Umpierrez GE, Cuervo R,
The annual incidence rate for DKA 157:669 – 675, 1997 Kitabchi AE: Proinflammatory cytokines,
from population-based studies ranges 3. DeFronzo RA, Matzuda M, Barret E: Dia- markers of cardiovascular risks, oxidative
from 4.6 to 8 episodes per 1,000 patients betic ketoacidosis: a combined metabolic- stress, and lipid peroxidation in patients
nephrologic approach to therapy. Diabetes with hyperglycemic crises. Diabetes
with diabetes, with a trend toward an in- Rev 2:209 –238, 1994 53:2079 –2086, 2004
creased hospitalization rate in the past 2 4. Ennis ED, Stahl EJVB, Kreisberg RA: The 20. Newcomer JW: Second generation (atyp-
decades. The incidence of HHS accounts hyperosmolar hyperglycemic syndrome. ical) antipsycotics and metabolic effects: a
for <1% of all primary diabetic admis- Diabetes Rev 2:115–126, 1994 comprehensive literature review. CNS
sions. Significant resources are spent on 5. Fishbein HA, Fishbein HA, Palumbo PJ: Drugs 19 (Suppl. 1):1–93, 2005
the cost of hospitalization. DKA episodes Acute metabolic complications in diabe- 21. Polonsky WH, Anderson BJ, Lohrer PA,
represent more than $1 of every $4 spent tes. In Diabetes in America. National Dia- Aponte JE, Jacobson AM, Cole CF: Insulin
on direct medical care for adult patients betes Data Group, National Institute of omission in women with IDDM. Diabetes
with type 1 diabetes and $1 of every $2 in Health, 1995, p. 283–291 (NIH publ. no. Care 17:1178 –1185, 1994
95-1468). 22. Peden NR, Broatan JT, McKenry JB: Dia-
those patients experiencing multiple epi-
6. Lorber D: Nonketotic hypertonicity in di- betic ketoacidosis during long-term treat-
sodes. Based on an annual average of abetes mellitus. Med Clin North Am 79: ment with continuous subcutaneous
100,000 hospitalizations for DKA in the 39 –52, 1995 insulin infusion. Diabetes Care 7:1–5,
U.S., with an average cost of $13,000 per 7. Kreisberg RA: Diabetic ketoacidosis: an 1984
patient (79), the annual hospital cost for update. Crit Care Clin 3:817– 834, 1987 23. Weissberg-Benchell J, Antisdel-Lomaglio
patients with DKA may exceed $1 billion 8. Wachtel TJ, Tctu-Mouradjian LM, Gold- J, Seshadri, R: Insulin pump therapy: a
per year. Many of these hospitalizations man DL, Ellis SA, O’Sullivan PS: Hyper- meta-analysis. Diabetes Care 26:1079 –
could be avoided by devoting adequate osmolarity and acidosis in diabetes 1087, 2003
resources to apply the measures de- mobility. J Gen Int Met 6:495–502, 1991 24. Umpierrez GE, Smiley D, Kitabchi AE:
scribed above. Because repeated admis- 9. Malone ML, Gennis V, Goodwin JS: Char- Ketosis-prone type 2 diabetes mellitus.
acteristics of diabetic ketoacidosis in older Annals Int Med 144:350 –357, 2006
sions for DKA are estimated to drain versus younger adults. J Am Geriatr Soc 25. Maldonado M, Hampe CS, Gaur LK,
approximately one of every two health 40:1100 –1104, 1992 D’Amico S, Iyer D, Hammerle LP, Bolgiano
care dollars spent on adult patients with 10. Kitabchi AE, Umpierrez GE, Murphy MB: D, Rodriguez L, Rajan A, Lernmark A, Bala-
type 1 diabetes, resources need to be re- Diabetic ketoacidosis and hyperglycemic subramanyam A: Ketosis-prone diabetes:
directed toward prevention by funding hyperosmolar state. In International Text- dissection of a heterogeneous syndrome us-
better access to care and educational pro- book of Diabetes Mellitus. 3rd ed. De- ing an immunogenetic and beta-cell func-
grams tailored to individual needs, in- Fronzo RA, Ferrannini E, Keen H and tional classification, prospective analysis,
cluding ethnic and personal health care Zimmet P, Eds. John Wiley & Sons, and clinical outcomes. J Clin Endocrinol
beliefs. In addition, resources should be Chichester, U.K., 2004, p. 1101–1119 Metab 88:5090 –5098, 2003
directed toward the education of primary 11. Wachtel TJ, Silliman RA, Lamberton P: 26. Mauvais-Jarvis F, Sobngwi E, Porcher R,
Prognostic factors in the diabetic hyper- Riveline JP, Kevorkian JP, Vaisse C, Char-
care providers and school personnel so osmolar state. J Am Geriatr Soc 35:737– pentier G, Guillausseau PJ, Vexiau P,
that they can identify signs and symptoms 741, 1978 Gautier JF: Ketosis-prone type 2 diabetes
of uncontrolled diabetes and new-onset 12. Wachtel TJ: The diabetic hyperosmolar in patients of sub-Saharan African origin:
diabetes can be diagnosed earlier. This has state. Clin Geriatr Med 6:797– 806, 1990 clinical pathophysiology and natural his-
been shown to decrease the incidence of 13. Gerich JE, Martin MM, Recant L: Clinical tory of beta-cell dysfunction and insulin
DKA at the onset of diabetes (80). and metabolic characteristics of hyperos- resistance. Diabetes 53:645– 653, 2004
molar nonketotic coma. Diabetes 20:228 – 27. Umpierrez GE, Casals MM, Gebhart SP,
NOTE ADDED IN PROOF — A 238, 1971 Mixon PS, Clark WS, Phillips LS: Diabetic
recent study from a city hospital reports 14. Kitabchi AE, Fisher JN: Insulin therapy of ketoacidosis in obese African-Americans.

2746 DIABETES CARE, VOLUME 29, NUMBER 12, DECEMBER 2006


Kitabchi and Asssociates

Diabetes 44:790 –795, 1995 tern Med 91:200 –204, 1979 curacy of an electrochemical sensor for
28. Banerji MA, Chaiken RL, Huey H, Tuomi 42. Gerard SK, Rhayam-Bashi H: Character- measuring capillary blood ketones by fin-
T, Norin AJ, Mackay IR, Rowley MJ, Zim- ization of creatinine error in ketotic pa- gerstick samples during metabolic deteri-
met PZ, Lebovitz HE: GAD antibody neg- tients: a prospective comparison of oration after continuous subcutaneous
ative NIDDM in adult black subjects with alkaline picrate methods with an enzy- insulin infusion interruption in type 1 di-
diabetic ketoacidosis and increased fre- matic method. Am J Clin Pathol 84:659 – abetic patients. Diabetes Care 26:1137–
quency of human leukocyte antigen DR3 661, 1985 1141, 2003
and DR4: Flatbush diabetes. Diabetes 43: 43. Kitabchi AE, Fisher JN, Murphy MB, 56. Kelly AM: The case for venuous rather
741–745, 1994 Rumbak MJ: Diabetic ketoacidosis and than arterial blood gages in diabetic keto-
29. Umpierrez GE, Woo W, Hagopian WA, the hyperglycemic hyperosmolar non-ke- acidosis. Emerg Med Australia 18:64 – 67,
Isaacs SD, Palmer JP, Gaur LK, Nepom totic state. In Joslin’s Diabetes Mellitus. 2006
GT, Clark WS, Mixon PS, Kitabchi AE: Im- 13th ed. Kahn CR, Weir GC, Eds. Phila- 57. Beigelman PM: Potassium in severe dia-
munogenetic analysis suggests different delphia, Lea & Febiger, 1994, p. 738 – betic ketoacidosis (Editorial). Am J Med
pathogenesis for obese and lean African- 770 54:419 – 420, 1973
Americans with diabetic ketoacidosis. Dia- 44. Shah P, Isley WL: Ketoacidosis during a 58. Abramson E, Arky R: Diabetic acidosis
betes Care 22:1517–1523, 1999 low carbohydrate diet. N Engl J Med 354: with initial hypokalemia: therapeutic im-
30. Kitabchi AE: Editorial: Ketosis-prone dia- 97–98, 2006 plications. JAMA 196:401– 403, 1966
betes: a new subgroup of patients with 45. Soveid M, Ranjbar-Omrani G: Ketoacido- 59. Viallon A, Zeni F, Lafond P, Venet C,
atypical type 1 an type 2 diabetes? J Clin sis as the primary manifestation of acro- Tardy B, Page Y, Bertrand JC: Does bicar-
Endocrinol Metab 88:5087–5089, 2003 megaly. Arch Iranian Med 8:326 –328, bonate therapy improve the management
31. Sobngwi E, Gautier JF, Kevorkian JP, Vil- 2005 of severe diabetic ketoacidosis? Crit Care
lette JM, Riveline JP, Zhang S, Vezian P, 46. Katz JR, Edwards R, Kahn M, Conway GS: Med 27:2690 –2693, 1999
Led, SM, Vaisse C, Mauvis-Jarris F: High Acromegaly presenting with diabetic ke- 60. Glaser NS, Wooten-Gorges SL, Marcin JP,
prevalence of glucose 6-phosphate dehy- toacidosis. Postgrad Med J 72:682– 683, Buonocore MH, Dicarlo J, Neely EK, Bar-
drogenase deficiency without gene muta- 1996 nes P, Bottomly J, Kuppermann N: Mech-
tion suggests a novel genetic mechanism 47. Vidal-Cortada J, Conget-Donlo JI, Na- anism of cerebral edema in children with
predisposed to ketosis-prone diabetes varro-Tellex MP, Halperin Rabinovic I, diabetic ketoacidosis. J Pediar 145:149 –
J Clin End Metab 90:4446 – 4451, 2005 Vilardell Latorre E: Diabetic ketoacidosis 150, 2004
32. Matz R: Hypothermia in diabetic acidosis. as the first manifestation of acromegaly. 61. Glaser NS, Wooten-Gorges SL, Buono-
Hormones 3:36 – 41, 1972 An Med Interna 12:76 –78, 1995 core MH, Marcin JP, Rewers A, Strain J,
33. Umpierrez G, Freire AX: Abdominal pain 48. Szeto CC, Li KY, Ko GT, Szeto CC, Li KY, Dicarlo J, Neely EK, Barnes P, Kupper-
in patients with hyperglycemic crises. J Ko GT, Chow CC, Yeung VT, Chan JC, mann N: Frequency of subclinical cere-
Crit Care 17:63– 67, 2002 Cockram CS: Acromegaly presenting in a bral edema in children with diabetic
34. Campbell IW, Duncan LJ, Innes JA, Mac- woman with diabetic ketoacidosis and in- ketoacidosis. Pediatr Diabetes 7:75– 80,
Cuish AC, Munro JF: Abdominal pain in sulin resistance. Int J Clin Pract 51:476 – 2006
diabetic metabolic decompensation: clin- 477, 1997 62. Morris LR, Murphy MB, Kitabchi AE: Bi-
ical significance. JAMA 233:166 –168, 49. Bratusch-Marrain PR, Komajati M, Wald- carbonate therapy in severe diabetic keto-
1975 hausal W: The effect of hyperosmolarity acidosis. Ann Intern Med 105:836 – 840,
35. Winter SD, Pearson JR, Gabow PA, on glucose metabolism. Pract Cardiol 11: 1986
Schultz AL, Lepoff RB The fall of the se- 153–163, 1985 63. Fisher JN, Kitabchi AE: A randomized
rum anion gap. Arch Intern Med 150: 311– 50. Kitabchi AE, Ayyagari V, Guerra SM: The study of phosphate therapy in the treat-
313 efficacy of low-dose versus conventional ment of diabetic ketoacidosis. J Clin Endo-
36. Slovis CM, Mark VG, Slovis RJ, Bain RP: therapy of insulin for treatment of dia- crinol Metab 57:177–180, 1983
Diabetic ketoacidosis & infection leuko- betic ketoacidosis. Ann Intern Med 84: 64. Barsotti MM: Potassium phosphate and
cyte count and differential as early predic- 633– 638, 1976 potassium chloride in the treatment of di-
tors of infection. Am J Emeg Med 5:1–5, 51. Kitabchi AE: Low dose insulin therapy in abetic ketoacidosis. Diabetes Care 3:569,
1987 diabetic ketoacidosis: fact or fiction. Dia- 1980
37. Kaminska ES, Pourmoabbed G: Spurious bete Metab Rev 5:337–363, 1989 65. Winter RJ, Harris CJ, Phillips LS, Green
laboratory values in diabetic ketoacidosis 52. Umpierrez GE, Latif K, Stoever J, Cuervo OC: Diabetic ketoacidosis: induction of
and hyperlipidaemia. Am J Emerg Med 11: R, Park L, Freire AX, A EK: Efficacy of hypocalcemia and hypomagnesemia by
77– 80, 1993 subcutaneous insulin lispro versus con- phosphate therapy. Am J Med 67:897–
38. Rumbak MJ, Hughes TA, Kitabchi AE: tinuous intravenous regular insulin for 900, 1979
Pseudonormoglycaemia in diabetic keto- the treatment of patients with diabetic ke- 66. Keller V, Berger W: Prevention of hy-
acidosis with elevated triglycerides. Am J toacidosis. Am J Med 117:291–296, 2004 pophosphalemia by phosphate infusion
Emerg Med 9:61– 63, 1991 53. Umpierrez GE, Cuervo R, Karabell A, La- during treatment of diabetic ketoacidosis
39. Adrogue HJ, Lederer ED, Suki WN, tif K, Freire AX, Kitabchi AE: Treatment of and hyperosmolar coma. Diabetes 29:87–
Eknoyan G: Determinants of plasma po- diabetic ketoacidosis with subcutaneous 95, 1980
tassium levels in diabetic ketoacidosis. insulin aspart. Diabetes Care 27:1873– 67. Kreisberg RA: Phosphorus deficiency and
Medicine 65:163–171, 1986 1878, 2004 hypophosphatemia. Hosp Pract 12:121–
40. Atchley DW, Loeb RF, Richards DW, 54. Della Manna T, Steinmetz L, Campos PR, 128, 1977
Benedict EM, Driscoll ME: A detailed Farhat SC, Schvartsman C, Kuperman H, 68. Adrogue HJ, Wilson H, Boyd AE 3rd, Suki
study of electrolyte balance following Setian N, Damiani D: Subcutaneous use of WN, Eknoyan G: Plasma acid-base pat-
withdrawal and reestablishment of insu- a fast-acting insulin analog: an alternative terns in diabetic ketoacidosis. N Engl
lin therapy. J Clin Invest 12:297–321, treatment for pediatric patients with dia- J Med 307:1603–1610, 1982
1933 betic ketoacidosis. Diabetes Care 28: 69. Oh MS, Carroll HJ, Goldstein DA, Fein
41. Vinicor F, Lehrner LM, Karn RC, Merritt 1856 –1861, 2005 IA: Hyperchloremic acidosis during the
AD: Hyperamylasemia in diabetic ketoac- 55. Guerci B, Benichou M, Floriot M, Bohme recovery phase of diabetic ketosis. Ann In-
idosis: sources and significance. Ann In- P, Fougnot S, Franck P, Drouin P: Ac- tern Med 89:925–927, 1978

DIABETES CARE, VOLUME 29, NUMBER 12, DECEMBER 2006 2747


Hyperglycemic crises in diabetic adults

70. Oh MS, Carroll HJ, Uribarri J: Mechanism Sterns RH: Pathogenesis of cerebral 78. Kitabchi AE: Hyperglycemic crises: im-
of normochloremic and hyperchloremic edema after treatment of diabetic ketoac- proving prevention and management.
acidosis in diabetic ketoacidosis. Nephron idosis. Kidney Int 51:1237–1244, 1997 Fam Physician 71:1659 –1660, 2005
54:1– 6, 1990 75. Glaser NS, Wootton-Gorges SL, Marcin 79. Javor KA, Kotsanos JG, McDonald RC,
71. Fleckman AM: Diabetic ketoacidosis. En- JP, Buonocore MH, DiCarlo J, Neely K, Baron AD, Kesterson JG, Tierney WM: Di-
docrinol Metabol Clin North Am 22:181– Barnes P, Bottomly J, Kuppermann N: abetic ketoacidosis charges relative to
207, 1993 Mecahanism of cerebral edema in chil- medical charges of adult patients with
72. Duck SC, Wyatt DT: Factors associated dren with diabetic ketoacidosis. J Pediatr type 1 diabetes. Diabetes Care 20:349 –
with brain herniation in the treatment of 145:164 –171, 2004 354, 1997
diabetic ketoacidosis. J Pediatr 113:10 – 76. Musey VC, Lee JK, Crawford R, Klatka 80. Vanelli M, Chiari G, Ghizzoni L, Costi G,
14, 1988 MA, McAdams D, Phillips LS: Diabetes in Giacalone T, Chiarelli F: Effectiveness of a
73. Glaser N, Barnett P, McCaslin I, Nelson D, urban African-Americans. I. Cessation of prevention program for diabetic ketoaci-
Trainor J, Louie J, Kaufman F, Quayle K, insulin therapy is the major precipitating dosis in children: an 8-year study in
Roback M, Malley R, Kuppermann N: cause of diabetic ketoacidosis. Diabetes schools and private practices. Diabetes
Risk factors for cerebral edema in children Care 18:483– 489, 1995 Care 22:7–9, 1999
with diabetic ketoacidosis: the Pediatric 77. Maldonado MR, Chong ER, Oehl MA, 81. Nyenwe E, Loganathan R, Blum S, Ezuteh
Emergency Medicine Collaborative Re- Balasubramanyam A: Economic impact of D, Erani D, Wan J, Palace M, Kitabchi A:
search Committee of the American Acad- diabetic ketoacidosis in a multiethnic in- Active use of cocaine: an independent risk
emy of Pediatrics. N Engl J Med 344:264 – digent population: analysis of costs based factor for recurrent diabetic ketoacidosis
269, 2001 on the precipitating cause. Diabetes Care in a city hospital. Endocrine Practice. In
74. Silver SM, Clark EC, Schroeder BM, 26:1265–1269, 2003 press

2748 DIABETES CARE, VOLUME 29, NUMBER 12, DECEMBER 2006