Diabetic Ketoacidosis and Hyperosmolar Hyperglycemic Syndrome: Review of Acute Decompensated Diabetes in Adult Patients
Diabetic Ketoacidosis and Hyperosmolar Hyperglycemic Syndrome: Review of Acute Decompensated Diabetes in Adult Patients
Diabetic Ketoacidosis and Hyperosmolar Hyperglycemic Syndrome: Review of Acute Decompensated Diabetes in Adult Patients
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
decompensated diabetes in adult patients
Esra Karslioglu French,1 Amy C Donihi,2 Mary T Korytkowski1
1
Division of Endocrinology and
Metabolism, Department of A B S T RAC T
Medicine, University of Pittsburgh,
Pittsburgh, PA, USA
Diabetic ketoacidosis and hyperosmolar hyperglycemic syndrome (HHS) are life threatening
2
University of Pittsburgh School of complications that occur in patients with diabetes. In addition to timely identification of the
Pharmacy, Pittsburgh, PA, USA
Correspondence to: M Korytkowski precipitating cause, the first step in acute management of these disorders includes aggressive
[email protected]
administration of intravenous fluids with appropriate replacement of electrolytes (primarily
Cite this as: BMJ 2019;365:l1114
doi: 10.1136/bmj.l1114 potassium). In patients with diabetic ketoacidosis, this is always followed by administration
Series explanation: State of the of insulin, usually via an intravenous insulin infusion that is continued until resolution of
Art Reviews are commissioned
on the basis of their relevance to ketonemia, but potentially via the subcutaneous route in mild cases. Careful monitoring
academics and specialists in the US
and internationally. For this reason
by experienced physicians is needed during treatment for diabetic ketoacidosis and HHS.
they are written predominantly by Common pitfalls in management include premature termination of intravenous insulin
US authors
therapy and insufficient timing or dosing of subcutaneous insulin before discontinuation of
intravenous insulin. This review covers recommendations for acute management of diabetic
ketoacidosis and HHS, the complications associated with these disorders, and methods for
preventing recurrence. It also discusses why many patients who present with these disorders
are at high risk for hospital readmissions, early morbidity, and mortality well beyond the acute
presentation.
Table 1 | Diagnostic criteria for diabetic ketoacidosis in adults iture of $2.4bn (£1.9bn; €2.1bn) per year in the US.10 In the
Criteria ADA1 UK2 AACE/ACE5
UK, the average cost for an episode of diabetic ketoacidosis
Year of publication 2009 2013 2016 is estimated to be £2064 ($2682; €2384) per patient.18
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
Plasma glucose concentration, >13.9 (250 mg/dL)* >11 (>200 mg/dL)or NA Diabetic ketoacidosis was a fatal disease before the dis-
mmol/L known diabetes covery of insulin in 1921. Nearing 100 years of insulin
pH Mild: 7.25-7.30; moderate: <7.3 (severe: <7.0) <7.3 availability, mortality related to diabetic ketoacidosis in
7.00-7.24; severe: <7.00
the adult population has progressively declined to less
Bicarbonate concentration, Mild: 15-18; moderate: <15 (severe: <5) NA
mmol/L or mEq/L 10-14.9; severe: <10
than 1%.13 Certain groups of patients have mortality rates
Anion gap: Na+–(Cl–+HCO3–) Mild: >10; moderate: >12; NA (severe: >16) >10 of up to 5%, including patients with significant comor-
severe: >12 bid conditions and those of advanced age.19 HHS, which
Urine acetoacetate Positive Positive Positive occurs more commonly in older adults with underlying
(nitroprusside reaction) comorbidities, has a 10-20% mortality rate.8 20
Blood β-hydroxybutyrate, NA† ≥3 (31 mg/dL) (severe: >6) ≥3.8 (40 mg/dL) Interest has been renewed in euglyemic diabetic
mmol/L
ketoacidosis, which was first described as a case series
Mental status Mild: alert; moderate: alert or NA Drowsy, stupor, or
drowsy; severe: stupor or coma coma of 37 patients presenting with blood glucose below 300
AACE/ACE=American Association of Clinical Endocrinologists/American College of Endocrinology; ADA=American Diabetes mg/dL.9 Euglycemic diabetic ketoacidosis can result in
Association; NA=not included in guideline document. delayed diagnosis and treatment, as well as potential
*2019 ADA guideline provides update to this 2009 guideline and now states “There is considerable variability… ranging from
euglycemia or mild hyperglycemia and acidosis to severe hyperglycemia, dehydration, and coma.”6 for adverse metabolic consequences.17 Since the recent
†β-hydroxybutyrate updated to >3 mmol/L in 2016 updated review cited in 2019 ADA guideline.6 7 introduction of the SGLT2 inhibitors, several case reports
and series describing euglycemic diabetic ketoacidosis
Table 2 | Diagnostic criteria for hyperglycemic hyperosmolar state in adults in patients treated with these agents have been pub-
Criteria ADA1 UK4 lished.21‑23 The entity of euglycemic diabetic ketoacidosis
Year of publication 2009 2015 is not limited to patients using SGLT2 inhibitors, however,
Plasma glucose concentration, mmol/L >33.3 ≥30 as it has also been described in the setting of alcohol use
pH >7.30 >7.30 disorders, pregnancy, and chronic liver disease.9 24
Bicarbonate concentration, mmol/L >18* >15
Anion gap: Na+–(Cl–+HCO3–) NA NA Pathophysiology
Urine acetoacetate (nitroprusside reaction) Negative or low positive NA Both diabetic ketoacidosis and HHS result from relative or
Blood β-hydroxybutyrate, mmol/L NA <3 absolute insulin deficiency together with an increase in
Osmolality, mmol/kg >320† ≥320‡ circulating concentrations of counterregulatory hormones
Presentation Stupor or coma Severe dehydration and feeling unwell (figure).1 25
ADA=American Diabetes Association; NA=not included in guideline document.
*Updated to >15 mmol/L in 2016 updated review cited in 2019 ADA guideline.6 7
†ADA guideline calculates effective plasma osmolality using equation 2×Na+glucose (mmol/L) or 2×Na+glucose (mg/ Pathophysiology of diabetic ketoacidosis
dL)/18. In the usual clinical situation, a rise in counterregula-
‡UK guideline calculates osmolality using equation 2×Na+glucose (mmol/L)+(blood urea nitrogen (mmol/L) or
2×Na+glucose (mg/dL)/18+blood urea nitrogen (mg/dL))/2.8. tory hormones contributes to accelerated gluconeogen-
esis, glycogenolysis, and impaired glucose utilization by
peripheral tissues and leads to diabetic ketoacidosis.25
both diabetic ketoacidosis and HHS that were published Glucagon is the primary counterregulatory hormone
between 1980 and 1987, as it is unlikely that these stud- responsible for development of diabetic ketoacidosis,
ies will be repeated. We included randomized controlled as increases in other counterregulatory hormones (cat-
trials (RCTs) and original articles published in the past echolamines, cortisol, and growth hormone) are not
40 years, as well as consensus statements, guidelines, necessarily observed.1 However, even glucagon is not
and systematic reviews published in the past 10 years. A absolutely essential, as diabetic ketoacidosis has also
supplementary table summarizing the objectives and key been described in patients after pancreatectomy.25 26
findings from several studies of the epidemiology, treat- In the liver, the gluconeogenic enzymes fructose 1,6
ment strategies, and outcomes is available. bisphosphatase, phosphoenolpyruvate carboxykinase
(PEPCK), glucose-6-phosphatase, and pyruvate carbox-
Epidemiology ylase are stimulated by an increase in the glucagon to
The frequency of diabetic ketoacidosis has increased insulin ratio and by an increase in circulating cortisol
during the past decade, with more than 160 000 hospi- concentrations.27 28 Hepatic gluconeogenesis is the main
tal admissions in 2017 in the US.12 13 A recent analysis mechanism for hyperglycemia in ketoacidosis, but renal
found that hospital admissions for diabetic ketoacidosis gluconeogenesis also contributes.29‑31
had increased in the UK for patients with both type 1 and In adipose tissue, the combination of severe insulin
type 2 diabetes.14 HHS is less common and accounts for deficiency with elevated counterregulatory hormone
less than 1% of all diabetes related admissions.15 concentrations activates hormone sensitive lipase lead-
The observed increase in hospital admissions for dia- ing to an increase in circulating free fatty acids.31‑34
betic emergencies has several potential explanations. Excess free fatty acids are oxidized to acetoacetate and
These include the rising prevalence of diabetes, as well β-hydroxybutyrate in hepatic mitochondria, result-
as psychosocial, cultural, and economic factors that often ing in ketonemia and acidosis.35 Glucagon accelerates
limit access to insulin and outpatient medical care.13 16 17 the generation of ketonemia and hyperglycemia in the
Diabetic ketoacidosis causes an estimated medical expend- insulin deficient state, but as previously mentioned it is
$EVROXWHLQVXOLQGHƟFLHQF\
',$%(7,&.(72$&,'26,6
&RXQWHUUHJXODWRU\KRUPRQHV
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
*OXFRQHRJHQHVLVB
*OXFRQHRJHQLF
*O\FRJHQRO\VLVB +\SHUJO\FHPLD 2VPRWLFGLXUHVLV
HQ]\PHVB
*OXFRVHXWLOL]DWLRQ?
(OHFWURO\WH
+RUPRQH ))$WROLYHUB
$&,'26,6 DEQRUPDOLWLHV
VHQVLWLYHOLSDVH .HWRJHQHVLVB
'HK\GUDWLRQ
5HODWLYHLQVXOLQGHƟFLHQF\
+<3(52602/$5+<3(5*/<&(0,$6<1'520(
&RXQWHUUHJXODWRU\KRUPRQHV
*OXFRQHRJHQHVLVB
*OXFRQHRJHQLF
*O\FRJHQRO\VLVB +\SHUJO\FHPLD 2VPRWLFGLXUHVLV
HQ]\PHVB
*OXFRVHXWLOL]DWLRQ?
(OHFWURO\WH
+RUPRQHVHQVLWLYHB $EVHQWRUPLQLPDO
+\SHURVPRODULW\ DEQRUPDOLWLHV
OLSDVH NHWRJHQHVLV
'HK\GUDWLRQ
Pathogenesis of diabetic ketoacidosis and hyperosmolar hyperglycemic syndrome. FFA=free fatty acids
not essential for development of diabetic ketoacidosis. infection, followed by new onset of diabetes and other
In addition to ketone body overproduction, clearance is metabolic stressors.37‑39 Certain drugs including gluco-
also decreased in diabetic ketoacidosis.25 36 corticoids, excess diuretics, atypical antipsychotics, and
Both hyperglycemia and high circulating concentra- others can predispose to severe hyperglycemia, diabetic
tions of ketone bodies result in an osmotic diuresis, which ketoacidosis, and HHS.20 40
leads to hypovolemia and subsequent decrease in glomer-
ular filtration rate. Osmotic diuresis promotes net loss of Precipitating causes of diabetic ketoacidosis
sodium, potassium, calcium, magnesium, chloride, and In a large survey of 283 patients admitted to one of 72
phosphate, further contributing to electrolyte abnormali- hospitals in the UK, infection was identified as the most
ties associated with diabetic ketoacidosis.25 Progressive common precipitating factor for diabetic ketoacidosis
volume depletion leads to decreased glomerular filtra- (45%), followed by insulin omission (20%); other causes
tion rate, with further decreases in clearance of glucose included newly diagnosed diabetes and alcohol or drug
and ketone bodies further contributing to hyperglycemia, related problems.41 42 Insulin omission can be seen in
hyperosmolality, and metabolic acidosis.25 patients of all ages and is more commonly observed in
those with eating disorders, psychological distress, fear
Pathophysiology of HHS of hypoglycemia, and fear of weight gain.39 Other factors
The pathogenesis of HHS differs from that of diabetic that are associated with insulin omission include inability
ketoacidosis in that a more severe degree of dehydration to pay for insulin, the idea that insulin should be with-
is present owing to osmotic diuresis and an absence of held when illness interferes with eating, inadvertent
significant ketosis/ketonemia.1 3 4 7 A higher concentra- omission of an insulin dose, and, rarely, pump malfunc-
tion of circulating hepatic insulin could partly account for tion.17 38 39
the absence of significant ketosis in patients with HHS.3 It is not only patients who stop their insulin therapy.
Patients presenting with HHS have lower concentrations In the UK survey described above, more than 7% of cases
of free fatty acids, cortisol, growth hormone, and gluca- of diabetic ketoacidosis occurred in an inpatient popula-
gon than do those presenting with diabetic ketoacidosis. tion.41 We (and patients who reviewed this manuscript)
Patients with HHS may have mild metabolic acidosis due have observed situations in which insulin therapy is
to renal failure and dehydration.20 erroneously withheld in patients admitted to hospital
with type 1 diabetes who are placed in a fasting state for
Precipitating causes surgical or other procedures. Some healthcare providers
The most common precipitating factors in diabetic also make erroneous assumptions that patients over 50
ketoacidosis and HHS are inadequate insulin therapy and years of age have type 2 diabetes and can tolerate periods
of insulin omission when admitted to hospital for acute A direct relation exists between hyperosmolality and
illness or surgical procedures (based on personal obser- depressed sensorium in patients presenting with blood
vations and experiences). Additionally, insulin therapy glucose above 33.3 mmol/L (600 mg/dL) and insignifi-
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
can be erroneously withheld or discontinued in patients cant concentrations of ketones.20 52 Patients presenting
admitted to the hospital with an insulin pump device with diabetic ketoacidosis, including euglycemic diabetic
owing to a lack of familiarity with these devices on the ketoacidosis, and HHS need immediate referral for emer-
part of hospital personnel.43‑45 gency evaluation and treatment.
SGLT2 inhibitors have been identified as causal agents
in several reported cases of euglycemic diabetic ketoaci- Acute management
dosis.23 46 A discussion of how these agents contribute to The management of both diabetic ketoacidosis and HHS
diabetic ketoacidosis is beyond the scope of this manu- includes fluids (usually administered intravenously),
script, and readers are referred to several publications in electrolytes, and insulin. Identifying the cause of acute
which this is discussed.47‑49 Risk factors for euglycemic decompensated diabetes is important, but this should not
diabetic ketoacidosis with SGLT2 inhibitors include latent cause any delay in treatment. Patients (and their family
autoimmune diabetes of adulthood, surgery, low carbo- members), especially those who present with more severe
hydrate diets, insulin withdrawal or dose reduction, and symptoms of diabetic ketoacidosis or HHS, often describe
acute medical illness. anxiety as part of their acute presentation. An explana-
Other clinical scenarios associated with euglycemic tion of what is happening, and the anticipated course of
diabetic ketoacidosis include pregnancy, decreased management, can alleviate some of these concerns and
caloric intake, heavy alcohol use, and chronic liver dis- open discussions as to how this can be prevented in the
ease.21 50 An awareness of these factors is important so as future (see prevention section below).
to avoid missing a diagnosis of diabetic ketoacidosis and Use of standardized paper based and computerized
delaying treatment.50 guidelines and protocols for management of diabetic
ketoacidosis has been shown to decrease the time to
Precipitating causes of HHS anion gap closure, reduce length of stay in hospital, and
HHS occurs most commonly, but not exclusively, in older minimize complications during treatment for diabetic
people with type 2 diabetes and accompanying comor- ketoacidosis.53‑57 However, even when protocols are avail-
bidities. Precipitating factors include pneumonia (40– able and healthcare providers are aware of them, adher-
60%) and urinary tract infections (5–16%) or other acute ence is often poor.41 42 58 Treatment protocols must be
conditions such as cerebrovascular disease, myocardial designed to be flexible and simple to use by physicians,
infarction, or trauma.3 20 51 The risk of HHS increases in nurses, and other healthcare providers, as the manage-
settings of inadequate fluid intake due to altered thirst ment of diabetic ketoacidosis and HHS is complicated,
mechanisms with aging or inability to access fluids.51 requiring close monitoring and modification over the
duration of treatment. It is especially important to rec-
Clinical presentation and diagnosis ognize that guidelines and order sets that are written or
Table 1 and table 2 outline diagnostic criteria for diabetic embedded in an electronic health record do not substitute
ketoacidosis and HHS as recommended by the Ameri- for sound clinical judgment.59
can Diabetes Association (ADA), Joint British Diabetes
Societies for Inpatient Care, and American Association Management of diabetic ketoacidosis
of Clinical Endocrinologists. Diagnostic criteria for dia- The goals of management of diabetic ketoacidosis include
betic ketoacidosis can include an elevation of urine restoration of intravascular volume, prevention and/or cor-
acetoacetate or blood β-hydroxybutyrate. Point of care rection of electrolyte abnormalities, correction of acidosis,
blood ketone meters and test strips for measurement of and correction of hyperglycemia (if present) (box 1).1 7 63
β-hydroxybutyrate are costly and not readily available in Patients presenting with mild diabetic ketoacidosis who are
many institutions, but they are likely to become standard alert and able to tolerate oral fluids may be able to receive
of care over time as they also provide accurate information treatment in the emergency department, potentially with
for guiding treatment (see management section below).36 oral fluids and subcutaneous insulin, and without need for
Patients with diabetic ketoacidosis can present with hospital admission.64 Patients presenting with more severe
some or all of the following symptoms: polyuria, poly- degrees of metabolic derangement need to be admitted to
dipsia, nausea, vomiting, abdominal pain, visual dis- a hospital unit with trained personnel and resources for
turbance, lethargy, altered sensorium, tachycardia, intensive monitoring and administration of intravenous flu-
tachypnea, and Kussmaul respirations, with a fruity ids, potassium, and insulin. The ADA and the Joint British
odor to the breath. Patients are usually severely volume Diabetes Societies for Inpatient Care have both published
depleted with orthostatic hypotension. Patients with guidelines for management of diabetic ketoacidosis, but
euglycemic diabetic ketoacidosis secondary to treat- these guidelines have several differences. These differences
ment with a SGLT2 inhibitor may have less polyuria and are primarily due to lack of published evidence to guide
polydipsia owing to the milder degree of hyperglycemia treatment in many areas.63 Despite this, the ADA guideline
and may instead present with malaise, anorexia, tachy- has been widely distributed, is used internationally, and has
cardia, or tachypnea with or without fever.24 Patients with been cited more than 600 times.63 In addition, a survey of
HHS often present with an altered level of consciousness, 118 UK clinical diabetes teams found that the UK guideline
which may mask the usual symptoms of hyperglycemia. has been adopted by more than 90% of those surveyed.10 65
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
saline) for initial fluid replacement. deficits that must be replaced after adequate dL), adjust dextrose or intravenous insulin
• ADA: 1000-1500 mL of normal saline during renal function (urine output) is assessed rate to maintain blood glucose in the 8-11
the first hour • ADA: 20-30 mmol (20-30 mEq) potassium mmol/L (150-200 mg/dL) range until DKA has
in each liter of infusion fluid when serum resolved
• UK: 1000 mL of normal saline during the first
hour potassium is <5.2 mmol/L (<5.2 mEq/L) • UK: Increase intravenous insulin rate
• UK: 40 mmol (40 mEq) in each liter of normal hourly using direct measurement of
2 After the first hour, the rate of intravenous fluids
saline when serum potassium is <5.5 mmol/L β-hydroxybutyrate and increasing the insulin
should be adjusted on the basis of the patient’s
(<5.5 mEq/L) rate by 1 unit/h increments to achieve blood
hemodynamic and electrolyte status and
ketone reduction of at least 0.5 mmol/L/h
generally maintained between 250 and 500 mL/h • Evidence is lacking to support either
(5.2 mg/dL/h); if blood β-hydroxybutyrate
• ADA: Patients with a normal or high corrected recommendation over the other
cannot be measured, increase insulin
sodium concentration can be switched to • Note: Because insulin therapy promotes
infusion rates by 1 unit/h to achieve
0.45% sodium chloride after the first hour of an intracellular shift of potassium, it is
increases of bicarbonate concentrations
fluid replacement recommended that insulin should not be
at a rate of ≥3.0 mmol/L/h (>3 mEq/L/h) or
• UK: Normal saline should be continued started if the serum potassium is <3 mmol/L
decreases in blood glucose by ≥3 mmol/L/h
throughout the management of DKA (<3 mEq/L) to avoid worsening of hypokalemia
(>50 mg/dL/h)
• Insufficient evidence exists to support the 2 Routine replacement of phosphate or
• Evidence is lacking to support either
hypothesis that balanced electrolyte solutions bicarbonate is not recommended (see text for
recommendation over the other. Note:
improve time to DKA resolution or prevent details)
bicarbonate concentrations may not be
major adverse kidney events in this population Insulin* reliable after the first 6 hours owing to the
Note: Continued use of normal saline after the 1 Intravenous insulin should not be started hyperchloremia resulting from continued use
initial resuscitation may result in hyperchloremic until after initiation of fluid resuscitation and of normal saline
metabolic acidosis and the inability to use correction of any hypokalemia 3 UK guidelines recommend continuing
plasma bicarbonate as a marker for DKA • ADA: Give intravenous insulin at either a fixed patients’ usual basal insulin dose or starting
resolution weight based dose of 0.14 units/kg/h or at weight based basal insulin during acute
3 Add dextrose to the intravenous fluid if/when a fixed weight based dose of 0.1 units/kg/h DKA management. In patients not previously
blood glucose approaches normal to allow following a 0.1 units/kg bolus treated with insulin, start basal insulin at a
continued insulin infusion at a rate sufficient to • UK: Give intravenous regular insulin at a fixed dose of 0.25-0.3 units/kg74 75
resolve DKA while avoiding hypoglycemia weight based dose of 0.1 units/kg/h Identify/treat reason for DKA and prevent
• ADA: Add dextrose 5% when the blood • ADA recommendation is supported by a small recurrence
glucose falls below 11 mmol/L (200 mg/dL) randomized controlled trial71 See box 4
• UK: Add dextrose 10% when the blood 2 Adjustment of intravenous insulin rate is ADA=recommendations in the American
glucose falls below 14 mmol/L (250 mg/dL) needed to ensure resolution of DKA Diabetes Association guideline1;
• A small randomized trial found no difference • ADA: If plasma glucose does not decrease UK=recommendation in the Joint British
between 5% and 10% dextrose in DKA by 3-4 mmol/L/h (50-75 mg/dL) from the Diabetes Societies for Inpatient Care guideline2
outcomes, but 10% dextrose caused more initial value in the first hour, the insulin *See box 2 for information about use of subcutaneous rapid acting
hyperglycemia66 infusion should be increased every hour insulin instead of intravenous insulin in selected patients with DKA
Intravenous fluids After the initial hour, the rate of intravenous fluids should
Administration of fluid is the first line of treatment. Appro- be adjusted on the basis of the patient’s hemodynamic and
priate fluid administration not only restores intravascular electrolyte status and generally maintained between 250
volume but also lowers blood glucose, raises blood pres- and 500 mL/h in adult patients without cardiac or renal
sure, ensures perfusion of peripheral tissues, and facili- compromise, advanced liver disease, or other states of fluid
tates resolution of metabolic acidosis. Current ADA and overload. No RCTs are available to guide the appropriate
UK guidelines for management of diabetic ketoacidosis in rate of intravenous fluids. The ADA guideline recommends
adults both recommend 0.9% sodium chloride solution that patients with a normal or high corrected sodium con-
(normal saline) for initial fluid replacement.1 2 Specifically, centration can be switched to 0.45% sodium chloride
the ADA recommends 1000-1500 mL and the UK guide- after the first hour of fluid replacement.1 However, the UK
line recommends 1000 mL of normal saline during the first guideline recommends the continuation of normal saline
hour. Although normal saline contains a supraphysiologic throughout the management of diabetic ketoacidosis.2
concentration of chloride potentially leading to hyperchlo- No published studies support or refute a switch to 0.45%
remic metabolic acidosis, as discussed in the complica- sodium chloride; however, continued use of normal saline
tions section of this review, insufficient evidence exists to after the initial resuscitation may result in hyperchloremic
support the hypothesis that balanced electrolyte solutions metabolic acidosis and the inability to use plasma bicar-
improve time to resolution of diabetic ketoacidosis or pre- bonate as a marker for resolution of diabetic ketoacidosis.
vent major adverse kidney events in this population.60 The As glucose normalizes with treatment, dextrose must
UK guideline acknowledges this and includes a statement be added to the intravenous fluid to allow continued insu-
that balanced electrolyte solutions can be used, but as they lin infusion at a rate sufficient to resolve ketonemia while
are not commercially available with premixed potassium avoiding hypoglycemia. The UK guideline recommends
chloride, they must be used in conjunction with additional adding dextrose 10% when the blood glucose falls below
potassium replacement.2 13.9 mmol/L (250 mg/dL).2 The ADA recommends adding
dextrose 5% when the blood glucose falls below 11 mmol/L designed for diabetic ketoacidosis management should
(200 mg/dL).1 A small randomized trial of 17 patients not be used. Use of protocols that target pre-specified goal
found no difference in capillary blood pH or bicarbonate blood glucose ranges such as 6.1-7.8 mmol/L (110-140
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
concentration when 5% and 10% dextrose were compared, mg/dL) or 7.8-10 mmol/L (140-180 mg/dL) for patients
although 10% dextrose caused more hyperglycemia, with in critical care settings can lead to inappropriately low
a mean blood glucose of 15.7 mmol/L at six hours com- doses or even premature discontinuation of intravenous
pared with 11.5 mmol/L in the group receiving 5% dex- insulin, which delays resolution of diabetic ketoacido-
trose despite concurrent use of more insulin (P<0.05).66 One sis.67 Even patients with euglycemic diabetic ketoacido-
example diabetic ketoacidosis protocol instructs nurses to sis need adequate insulin therapy to resolve ketonemia,
add 5% dextrose at the same rate using the same concentra- albeit with early addition of fluids containing dextrose to
tions of sodium chloride and potassium when blood glu- prevent hypoglycemia.24
cose falls below 13.9 mmol/L (250 mg/dL).67 For example, Insulin should not be started until after initiation of
0.45% sodium chloride with 20 mEq potassium chloride fluid resuscitation and correction of any hypokalemia to
(KCl) in each liter would change to dextrose 5% with 0.45% avoid worsening volume or potassium deficits with shift-
sodium chloride and 20mEqKCl/L at the same infusion rate. ing of potassium, glucose, and water from the extracellu-
Another strategy that has recently emerged is use of a lar to intracellular fluids compartment. The UK guideline
“two bag method” for fluid replacement.68 69 This method recommends starting intravenous regular insulin at
consists of two bags of 0.45% sodium chloride, one with a fixed weight based dose of 0.1 units/kg/h.2 The ADA
and one without 10% dextrose, that are adjusted on the guideline recommends starting intravenous regular insu-
basis of hourly blood glucose monitoring to maintain an lin at either a fixed weight based dose of 0.14 units/kg/h
intravenous fluid rate of 250 mL/h. Two retrospective stud- or a fixed weight based dose of 0.1 units/kg/h after a 0.1
ies involving more than 500 patients found that the two bag units/kg bolus of intravenous insulin.1 The ADA recom-
method was associated with earlier correction of acidosis mendation is based on a small RCT of 37 patients rand-
and shorter duration of intravenous insulin compared with omized to receive a loading dose of 0.07 unit/kg of regular
conventional delivery of intravenous fluids.68 69 When used insulin followed by 0.07 unit/kg/h, regular insulin 0.07
in the emergency department, this method may reduce the unit/kg/h without a loading dose, or regular insulin 0.14
need for hospital admission, and it may be associated with unit/kg/h without a loading dose.71 Most patients in the
less hypoglycemia compared with conventional treatment. 0.07 unit/kg/h group without the loading dose needed
Of note, vulnerabilities in the intravenous fluid sup- supplemental doses of insulin to obtain desired changes
ply in the US have existed for several years, prompting in glucose; with these supplemental doses, no difference
some hospitals to develop protocols for oral rehydration was seen in the time to resolution of diabetic ketoacidosis
in patients with dehydration.61 70 Although these protocols between the three groups.71 Although a rate of 0.1 unit/
have not been specifically tested in patients with diabetic kg/h may not be sufficient to suppress hepatic glucose
ketoacidosis, this strategy can be considered in selected production and stimulate peripheral glucose uptake, it
patients with mild-moderate diabetic ketoacidosis when is likely sufficient to suppress lipolysis and ketogenesis.72
accompanied by appropriate monitoring for therapeutic The ADA guideline states that reducing insulin infusion
efficacy.61 rates to 0.02-0.05 units/kg/h at the same time that dex-
trose 5% is added to the intravenous fluids may be pos-
Potassium replacement sible when blood glucose declines to below 11 mmol/L
Patients with diabetic ketoacidosis have total body potas- (200 mg/dL).1 However, this rate has not been proven to
sium deficit despite measured serum potassium concen- be sufficient to suppress lipolysis and ketosis.72 Addi-
trations that may be normal or even high at presentation. tional studies are needed to assess whether the potential
The ADA recommends adding 20-30 mEq potassium benefits of reduction of the insulin rate (reduced hypo-
in each liter of infusion fluid when serum potassium is glycemia and hypokalemia) justify the potential delay in
below 5.2 mEq/L.1 By contrast, the UK guideline recom- resolution of diabetic ketoacidosis. The UK guideline does
mends 40 mmol/L in each liter of normal saline when not recommend any adjustment in the insulin rate when
serum potassium is below 5.5 mmol/L and the patient dextrose 10% is added after the blood glucose falls below
is passing urine.2 As existing published trials evaluating 14 mmol/L (252 mg/dL).2 65
potassium replacement rates were not designed to deter- After insulin has been started, the ADA recommends
mine optimal potassium replacement rates in patients increasing insulin infusion rates hourly to reduce blood
with diabetic ketoacidosis, evidence is lacking to support glucose at a rate of 3-4 mmol/L/h (50-75 mg/dL/h)
either recommendation over the other.60 Because insulin until concentrations of 8-11 mmol/L (150-200 mg/dL)
therapy promotes an intracellular shift of potassium, it are achieved.1 On the other hand, the more recent UK
is recommended that insulin should not be started if the guideline recommends using direct measurement of
serum potassium is below 3 mmol/L to avoid worsening β-hydroxybutyrate (a hydroxy acid) and increasing the
of hypokalemia. insulin rate by 1 unit/h increments to achieve blood
ketone reduction of at least 0.5 mmol/L/h (5.2 mg/dL/h).2
Intravenous insulin The rate of glucose lowering is not directly correlated with
The primary purpose of insulin in diabetic ketoacidosis the degree of ketonemia suppression and reversal of aci-
management is to halt lipolysis and ketogenesis. Titration dosis; however, point of care blood ketone meters are not
protocols for intravenous insulin infusion that are not available in all hospitals, and in those hospitals that do
Box 2 | Option for diabetic ketoacidosis (DKA) management Use of subcutaneous rapid acting insulin
using subcutaneous rapid acting insulin* An alternative to intravenous insulin therapy for acute
• Potential candidates for DKA management with management of mild-moderate diabetic ketoacidosis is the
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
subcutaneous insulin therapy are those who meet the use of subcutaneous rapid acting insulin (box 2).76 Poten-
following criteria: tial candidate patients include those who are alert and do
––Patient is alert not otherwise need admission to a critical care area, have a
––Patient is able to tolerate oral fluid intake (no nausea/ pH above 7.0, and bicarbonate of at least 10 mmol/L (≥10
vomiting) mEq/L).77 Four prospective randomized studies in adult
––pH >7.0 patients with diabetic ketoacidosis compared subcutaneous
––Bicarbonate ≥10 mmol/L (≥10 mEq/L) rapid acting insulin (initial bolus of 0.3 units/kg followed by
• Doses of rapid acting insulin (eg, lispro, aspart, glulisine): 0.1-0.2 units/kg every 1-2 hours) with conventional diabetic
––Initial bolus: 0.3 units/kg bolus ketoacidosis treatment and found no difference in patient
––Maintenance: 0.2 units/kg every 2 hours outcomes.77‑80 With subcutaneous insulin regimens, admin-
––When blood glucose is <13.9 mmol/L (250 mg/dL): istration of intravenous fluids and electrolytes, as well as
administer 0.05-0.1 units/kg every 2 hours until monitoring, is identical to what is recommended with intra-
resolution of DKA (see table 3 for criteria) venous insulin therapy. Factors that may affect the efficacy of
• Treatment with intravenous fluids and potassium subcutaneous insulin regimens include potential absorption
replacement strategies are identical to those used with problems in patients who are severely dehydrated, possibil-
intravenous regular insulin ity of missed insulin doses, and altered pharmacokinetics of
• Continue the patient’s home basal insulin regimen. Give subcutaneous compared with intravenous insulin.
a dose (either the patient’s home basal insulin dose
or a weight based dose of 0.2 units/kg) at the time of
presentation with DKA, if the patient has not received a Phosphate replacement
dose within the previous 24 hours or a dose is due Patients with diabetic ketoacidosis also have total body
• If the anion gap is not closed within 12 hours using this phosphate deficits, but no prospective studies have
protocol, the patient should be switched to an intravenous shown that phosphate replacement improves outcome.
insulin infusion (see box 1). Insulin therapy will decrease phosphate, but overcor-
*Adapted from Ersoz et al, Umpierrez et al (×2), and Karoli et al77‑80 rection with phosphate can cause hypocalcemia. The
UK guideline recommends against routine phosphate
have them they are often not available 24 hours a day.41 replacement; checking and replacing phosphate should
In addition, blood ketones above 3 mmol/L (>31 mg/dL) occur only if the patient has symptoms of respiratory and
cannot be reliably measured using currently available skeletal muscle weakness.2 The ADA recommends that
devices, and measurements of decrements of 0.5 mmol/ 20-30 mmol of phosphate may be indicated in patients
L/h (5.2 mg/dL/h) may be beyond the coefficient of vari- with cardiac dysfunction, anemia, respiratory depres-
ation of existing meters.62 73 Given this, the UK guideline sion, or a phosphate concentration below 0.32 mmol/L
provides alternatives for situations in which bedside (<1 mg/dL).1
measurement of ketones is not possible. Specifically, the
UK guideline recommends increasing insulin infusion Use of sodium bicarbonate
rates by 1 unit/hour to achieve increases of bicarbonate A systematic review found that the available evidence,
concentrations at a rate of at least 3 mmol/L/h (>3 mEq/ which includes three RCTs, does not support replacing
L/h) or decreases in blood glucose by at least 3 mmol/L/h bicarbonate in adult diabetic ketoacidosis patients with a
(>50 mg/dL/h).2 Of note, bicarbonate concentrations may pH of 6.9 or higher.81 The review also found retrospective
not be reliable after the first six hours owing to the hyper- evidence of transient paradoxical worsening of ketosis
chloremia resulting from continued use of normal saline. and an increased need for potassium supplementation
in patients who received bicarbonate.81 Owing to possi-
Concurrent basal insulin ble harm and lack of sustained benefits, the UK guide-
The UK guideline recommends continuing a patient’s line does not recommend the use of bicarbonate in any
usual basal insulin dose or starting weight based basal patients with diabetic ketoacidosis.2 On the other hand,
insulin during acute diabetic ketoacidosis management.2 as the use of bicarbonate has not been prospectively stud-
In patients not previously treated with insulin, the UK ied in patients with a pH below 6.9, and these patients
guideline recommends starting basal insulin at a dose are at very high risk for poor outcome, the ADA guideline
of 0.25 units/kg.2 This approach is supported by a small recommends that slow administration of 100 mmol (100
RCT that compared 31 patients treated with intravenous mEq) NaHCO3 over two hours be considered when the pH
insulin and 30 patients treated with intravenous insu- is below 6.9.1
lin plus 0.25 units/kg of insulin glargine given within
12 hours of starting the intravenous insulin.74 Patients Management of HHS
receiving basal insulin had less rebound hyperglycemia The goals of HHS treatment include correction of volume defi-
without increased hypoglycemia in the 12 hours after dis- cits while reducing and normalizing plasma hyperosmolality,
continuation of intravenous insulin. Another randomized which will correct hyperglycemia (box 3). Patients with HHS
controlled pilot study found trends toward a reduction in always need to be admitted to hospital, not only to achieve
time to anion gap closure and length of stay in hospital metabolic stability, which often requires more than 24 hours
with concurrent insulin glargine.75 of treatment, but also to identify and treat the precipitating
Box 3 |Principles of hyperosmolar hyperglycemic syndrome (HHS) management in adult cause. The ADA has not published a separate guideline for
patients the management of HHS and includes recommendations for
both diabetic ketoacidosis and HHS in the same document.1
Intravenous fluids
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
The Joint British Diabetes Societies for Inpatient Care have a
1 Use 1000 mL of 0.9% sodium chloride solution (normal saline) over 1 hour for initial fluid
dedicated document for the management of HHS.4 Of note,
replacement
virtually no published studies exist to guide the optimal treat-
2 After the first hour, the rate of intravenous fluids should be adjusted on the basis of the
ment of HHS, which has resulted in differences in the ADA
patient’s hemodynamic and electrolyte status and generally maintained between 250
and 500 mL/h and UK guidelines.63
• ADA: Patients with a normal or high corrected sodium can be switched to 0.45%
sodium chloride after the first hour of fluid replacement Intravenous fluids
• UK: Adjust hourly rate of normal saline to achieve a decline in blood glucose of 4-6 Intravenous fluids are the first line of treatment in HHS to
mmol/L/h (70-100 mg/dL/h) and plasma osmolality of 3-8 mOsmol/kg/h. Switch normalize osmolality, restore intravascular volume, and
from normal saline to 0.45% sodium chloride solution only when the fall in plasma lower blood glucose. ADA and UK guidelines for HHS
osmolality and plasma glucose plateaus in the setting of adequate positive fluid management both recommend at least 1 L of normal
balance saline during the first hour, followed by adjustment of the
3 Add dextrose to the intravenous fluid if/when blood glucose approaches normal to allow infusion rate on the basis of the patient’s hemodynamic
continued insulin infusion at a rate sufficient to resolve diabetic ketoacidosis while and electrolyte status and the achievement of a positive
avoiding hypoglycemia fluid balance.
• ADA: Add dextrose 5% when the blood glucose falls below 16.7 mmol/L (300 mg/dL) The ADA generally recommends 250-500 mL/h (after
• UK: Add dextrose 5% or dextrose 10% when the blood glucose falls below 14 mmol/L the first hour) in adult patients without cardiac or renal
(250 mg/dL) compromise, advanced liver disease, and other states of
Electrolytes fluid overload. The ADA recommends a switch to 0.45%
1 Patients with HHS have total body potassium deficits that must be replaced after sodium chloride if sodium concentrations are high and
adequate renal function (urine output) is assessed the addition of dextrose 5% when blood glucose reaches
• ADA: 20-30 mmol (20-30 mEq) potassium in each liter of infusion fluid when serum 16.7 mmol/L (300 mg/dL).1 The UK guideline recom-
potassium is <5.2 mmol/L (<5.2 mEq/L) mends adjusting the hourly rate of normal saline to
• UK: 40 mmol/L (40 mEq) in each liter of normal saline when serum potassium is <5.5 achieve a decline in blood glucose of 4-6 mmol/L/h (70-
mmol/L (<5.5 mEq/L) and the patient is passing urine 100 mg/dL/h) and plasma osmolality of 3-8 mOsmol/
• Evidence is lacking to support either recommendation over the other kg/h.4 The UK guideline recommends switching from
• Note: Because insulin therapy promotes an intracellular shift of potassium, it is normal saline to 0.45% sodium chloride solution only
recommended that insulin (if used) should not be started if the serum potassium is <3 when the fall in plasma osmolality and plasma glucose
mmol/L (<3 mEq/L) to avoid worsening of hypokalemia plateaus in the setting of adequate positive fluid balance.
2 Routine replacement of phosphate is not recommended The UK guideline recommends adding dextrose 5-10%
Insulin to intravenous fluids when the blood glucose falls below
1 The optimal time to start intravenous insulin in the management of HHS has not been 13.9 mmol/L (250 mg/dL).4
determined
• ADA: Start intravenous insulin after initiation of fluid resuscitation and correction of Potassium and phosphate replacement
any hypokalemia No studies have investigated the optimal approach to
• UK: Start intravenous insulin if patient has 3-β-hydroxybutyrate concentrations >1.0 treatment of hypokalemia or hypophosphatemia in HHS
mmol/L (>10.4 mg/dL) (or >1.5 mmol/L (>15.6 mg/dL)82). Otherwise, intravenous patients; however, patients with HHS typically have more
insulin can be started once blood glucose is falling below 5 mmol/L/h (<90 mg/dL/h) in severe total body depletion than diabetic ketoacidosis
the setting of adequate positive fluid balance. patients, and close monitoring is advised. The ADA rec-
2 The optimal starting dose of intravenous insulin in the management of HHS has not been ommends adding 20-30 mmol (20-30 mEq) potassium
determined in each liter of infusion fluid when serum potassium is
• ADA: Start intravenous insulin at either a fixed weight based dose of 0.14 units/kg/h or below 5.2 mmol/L (5.2 mEq/L).1 By contrast, the UK
at a fixed weight based dose of 0.1 units/kg/h after a 0.1 units/kg bolus guideline recommends 40 mmol/L (40 mEq/L) in each
• UK: Start low dose intravenous insulin (0.05 units/kg/h) liter of normal saline when serum potassium is below 5.5
3 The rate of intravenous insulin should be adjusted to achieve an adequate drop in blood mmol/L (5.5 mEq/L) and the patient is passing urine.4
glucose Because insulin therapy promotes an intracellular shift
• ADA: If plasma glucose does not decrease by 3-4 mmol/L/h (50-75 mg/dL/h) from the in potassium, it is recommended that insulin should not
initial value in the first hour, the insulin infusion should be increased every hour until a be started if the serum potassium is below 3 mmol/L
steady glucose decline is achieved. When blood glucose is below 16 mmol/L (<300 mg/ (3 mEq/L) to avoid worsening of hypokalemia. The UK
dL), adjust dextrose or intravenous insulin rate to maintain concentrations in the 14-16 guideline recommends replacing phosphate if hypophos-
mmol/L (250-300 mg/dL) range until HHS has resolved
phatemia persists beyond the acute phase of treatment
• UK: Adjust insulin infusion rate hourly by 1 unit/h to maintain blood glucose decline of HHS.4
of <5 mmol/L/h (<90 mg/dL/h), noting that blood glucose should not fall below 10-15
mmol/L (180-270 mg/dL) during the first 24 hours of management
Insulin
Identify/treat reason for HHS and prevent recurrence The question of when to start insulin administration in
See box 4 the management of HHS has not been formally studied.
ADA=recommendations in the American Diabetes Association guideline1; The ADA guideline recommends starting intravenous reg-
UK=recommendation in the Joint British Diabetes Societies for Inpatient Care guideline4 ular insulin in the same way as during diabetic ketoaci-
Table 3 | Criteria for resolution of diabetic ketoacidosis (DKA) and hyperglycemic hyperosmolar Box 4 | Identification and treatment of underlying causes
state (HHS) and prevention of recurrence of diabetic ketoacidosis or
Criteria ADA1 UK2 4 hyperosmolar hyperglycemic syndrome
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
Resolution of Blood glucose <11 mmol/L pH >7.3 AND blood ketone concentration <0.6 mmol/L (<6.2 mg/ 1 Assess and manage reasons for insulin non-adherence
DKA (200 mg/dL) PLUS any two of dL). Bicarbonate not recommended because hyperchloremic (if applicable)
bicarbonate ≥15, pH >7.3, and acidosis is associated with large volumes of 0.9% sodium
anion gap ≤12 chloride solution • Consider alternative (human) insulin regimens and/
Resolution of Normal osmolality and regain of Intravenous insulin (if administered) can usually be discontinued or social work consultation if the patient is unable to
HHS normal mental status once patient is eating and drinking, but intravenous fluids may be afford the prescribed insulin
needed for longer if intake is inadequate • Facilitate treatment or social services for patients with
untreated or poorly controlled psychiatric disorders
dosis management. That is, starting intravenous regular • Refer patients with substance or alcohol use disorders
insulin at either a fixed weight based dose of 0.14 units/ to treatment programs
kg/h or at a fixed weight based dose of 0.1 units/kg/h 2 Provide education on appropriate sick day management
(if applicable) and refer the patient to a formal diabetes
followed by a 0.1 units/kg bolus of intravenous insulin
education program after hospital discharge
after initiation of fluid resuscitation and correction of any
3 Treat treatable precipitating causes such as infection,
hypokalemia.1 The ADA recommends reducing insulin
cerebrovascular accident, myocardial infarction, or
infusion rates to 0.02-0.05 units/kg/h at the same time trauma
that dextrose 5% is added to the intravenous fluids when
blood glucose declines to below 16.7 mmol/L (300 mg/ between pre-breakfast, pre-lunch, and pre-supper doses of
dL) in patients with HHS.1 rapid acting insulin.1 2 7
On the other hand, the UK guideline recommends Patients who meet metabolic criteria for resolution
delaying intravenous insulin therapy unless the patient of the diabetic crisis (table 3), but who remain critically
has 3-β-hydroxybutyrate concentrations above 1.0 ill and unable to tolerate oral intake, will need ongoing
mmol/L (10.4 mg/dL) or above 1.5 mmol/L (15.6 mg/ management with intravenous fluids and insulin, the lat-
dL).4 82 The rationale for this recommendation is that ter titrated using an insulin infusion protocol with pre-
early initiation of insulin in the setting of inadequate specified glycemic targets for critical illness.83 84
fluid replacement can aggravate hypoperfusion with an
increase in risk for circulatory compromise and throm- Prevention of diabetic ketoacidosis and HHS
bosis, and intravenous fluid alone will reduce hypergly- Recurrent episodes of diabetic ketoacidosis and HHS can
cemia.4 The ideal decline in blood glucose is less than be prevented in many cases.85 The cause of every episode
5 mmol/L/h (<90 mg/dL/h), and concentrations should of diabetic ketoacidosis and HHS should be determined so
not fall below 10-15 mmol/L (180-270 mg/dL) during the that tailored education and intervention can be provided
first 24 hours of management. If the blood glucose fails to to the patient to prevent recurrence (box 4).85 86 To prevent
decline despite adequate fluids, then the recommended readmission to hospital and visits to the emergency depart-
fixed rate of intravenous insulin is lower (0.05 units/kg) ment, education based on the needs of the individual patient
than is recommended for diabetic ketoacidosis.4 should be provided before discharge.86 This education, some-
times called “survival skills” education, can include a review
Transition from acute management of the causes, signs, and symptoms of impending diabetic
After resolution of diabetic ketoacidosis and HHS (table ketoacidosis, as well as what to do and who and when to
3), all patients need to be transitioned from intravenous to call when symptoms develop. Providing patients and their
subcutaneous insulin.1 67 This includes patients with eug- family members with information on how to manage their
lycemic diabetic ketoacidosis secondary to treatment with diabetes drugs during periods of acute illness can reduce the
an SGLT2 inhibitor and those with ketosis prone diabetes metabolic decompensation that can occur with inappropri-
who present with diabetic ketoacidosis. To prevent rebound ate reductions or omissions of insulin doses.87 In addition,
ketoacidosis or hyperglycemia, administration of a long act- ensuring that patients have access to their diabetes drugs,
ing basal insulin (if this has not already been given in the including the ability to pay for these drugs, is important in
previous 24 hours) with or without a short or rapid acting preventing readmissions.88 89
insulin is needed at least two hours before the intravenous All patients with diabetes should receive initial and
insulin infusion is stopped.67 74 75 This overlap is needed ongoing education about diabetes in the community set-
mainly because of the short half life of intravenous insulin ting, as this can help to prevent diabetes related hospital
(approximately 10 minutes).7 When this transition is timed admissions (and readmissions). This education is best
before a meal, administration of a prandial dose of short provided or directed by a trained diabetes educator.86 89
90
or rapid acting insulin together with the basal insulin may Patient specific instructions for sick day management,
allow discontinuation of intravenous insulin in an hour.82 including reminders that intermediate or long acting
The ADA and UK guidelines recommend that patients previ- insulin should never be stopped and when to call for
ously treated with subcutaneous insulin can be restarted on assistance, can help to prevent future episodes of dia-
their pre-admission insulin doses if these are determined to betic ketoacidosis.91 92 All patients with type 1 diabetes
be appropriate.1 2 Otherwise, a weight based subcutaneous and many insulin treated patients with type 2 diabetes
insulin regimen can be started by calculating a total dose should have a means to measure ketones at home and be
of 0.5-0.7 units/kg/d and giving 50% of the total dose as instructed to monitor both glucose and ketones at least
once daily basal insulin and dividing the other 50% equally every four hours during illness.
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
before planned surgical procedures or during periods of non-anion gap hyperchloremic metabolic acidosis.41 56 67
acute illness.5 47 50 The half life of this class of agents is Hypokalemia is reported more frequently than hyper-
approximately 13 hours, but the clinical effect can last kalemia and usually results from delays in administration
for several days after discontinuation.47-50 93 Patients of or insufficient potassium containing supplementation.
need to be educated about situations that predispose Hyperkalemia can result from overly aggressive potas-
them to developing euglycemic diabetic ketoacidosis, sium replacement, particularly in patients with underly-
including dehydration or ingestion of excessive amounts ing renal dysfunction.55 67 Hypoglycemia can result from
of alcohol.5 47 50 Patients taking SGLT2 inhibitors should overly aggressive insulin infusions, insufficient frequency
be instructed to check for ketones during times of ill- of blood glucose monitoring, or failure to add dextrose to
ness and metabolic stress. When high ketone concen- intravenous fluids when blood glucose concentrations
trations are detected, a proposed strategy to prevent approach 13.9 mmol/L (250 mg/dL).55 56
progression to diabetic ketoacidosis is to stop the SGLT2 In a retrospective review comparing 8550 adult
inhibitor, inject bolus insulin, consume 30 g of carbo- patients with diabetic ketoacidosis admitted to an inten-
hydrate, hydrate with water, and continue to monitor sive care unit where blood glucose was corrected to 10
ketones every three to four hours.50 Patients should be mmol/L (180 mg/dL) or lower or to above 10 mmol/L (180
educated on potential symptoms heralding the onset of mg/dL) within 24 hours, those in the latter group had less
diabetic ketoacidosis and be instructed to seek care if hypoglycemia, hypokalemia, and hypo-osmolality, with
they develop these symptoms, regardless of their blood lower mortality.97 However, the fact that these compli-
glucose concentration. cations still occurred in approximately 25% of patients
treated with standardized protocols suggests the need to
Complications related to management of diabetic monitor electrolytes and blood glucose concentrations no
ketoacidosis and HHS less often than every two hours while intravenous insulin
Several complications can be associated with diabetic infusions are continued. In a national survey from the
crises and their management. UK, suboptimal adherence to hospital based protocols
and national guidelines was identified as a major con-
Cerebral edema tributor to these frequently observed complications.41 42
Of these complication, the development of cerebral Although efforts are under way in some institutions to
edema is the most serious.11 94 This has been described mitigate these complications through modification of
most commonly in young children and adolescents pre- existing protocols, attention to staff education and ongo-
senting with diabetic ketoacidosis as the initial manifes- ing monitoring or protocol adherence may also help to
tation of new onset type 1 diabetes, but it has also been reduce these risks.67
described in young adults up to age 28.94 95 The development of non-anion gap hyperchloremic
Rare cases of cerebral edema occur in adults over age metabolic acidosis often follows the acute phase of diabetic
28, but current recommendations suggest maintaining ketoacidosis management.1 34 98 This has been attributed to
blood glucose concentration no lower than 13.9-16.6 administration of large volumes of intravenous fluids con-
mmol/L (250-300 mg/dL) for several hours during the taining normal saline during acute diabetic ketoacidosis
course of treatment for patients with either diabetic management, as well as to urinary losses of keto-anions
ketoacidosis or HHS as a potential method for avoiding that are needed for regeneration of bicarbonate. This is
this devastating complication.1 2 7 not considered to be a serious complication and usually
Early recognition of potential neurologic deterioration resolves spontaneously in a few days.1 98 No evidence sup-
such as new onset or intensifying headache, a decline in ports the use of colloid solutions, balanced electrolyte solu-
level of consciousness, recurrent vomiting, incontinence, tions (such as Plasma-Lyte), or Ringer’s lactate in place of
irritability, abnormal respirations, a delayed rise in serum 0.9% sodium chloride solution in the management of dia-
sodium with treatment, or evidence of cranial nerve betic ketoacidosis.60 In one RCT conducted at two institu-
dysfunction provide suggestive evidence of onset of cer- tions, 57 patients admitted with diabetic ketoacidosis were
ebral edema. Prompt administration of mannitol therapy randomly assigned to receive fluid resuscitation with 0.9%
administered at a dose of 0.5-1 g/kg over 20 minutes can sodium chloride or Ringer’s lactate. It found no difference in
help to abort further neurologic deterioration.96 Delay- time to resolution of diabetic ketoacidosis, with a delay in
ing potential lifesaving treatment while awaiting results achieving blood glucose concentrations below 14 mmol/L
of computed tomography imaging is not recommended. (252 mg/dL) in the group receiving Ringer’s lactate.99 In
Modification of the rate of intravenous fluid admin- another RCT, 45 adult patients with type 1 diabetes admit-
istration has previously been suggested as a potential ted with diabetic ketoacidosis were randomly assigned to
method for ameliorating this risk. A recent multicenter standardized resuscitation with normal saline or a balanced
trial that included 1255 children with diabetic ketoaci- electrolyte solution with measurement of serum chloride
dosis randomly assigned to one of four treatment groups and bicarbonate every two hours for 12 hours. Those receiv-
receiving 0.45% or 0.9% sodium chloride containing ing the balanced electrolyte solution had lower chloride
intravenous fluids administered at a slow or rapid rate and higher bicarbonate concentrations, although the clini-
showed no difference in neurologic outcomes.95 cal significance remains to be determined.100
Cardiac, respiratory, and muscle complications necessary as the subtle differences in rate of intravenous
Other complications that occur less frequently but for fluid administration, electrolyte replacement, or dosing
which monitoring is important include myocardial infarc- of intravenous insulin are not likely to result in major dif-
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
tion, potential for pulmonary edema in patients with ferences in outcomes. One difference in the guidelines for
underlying congestive heart failure, and rhabdomyoly- diabetic ketoacidosis that may need additional investiga-
sis in patients who present with more severe degrees of tion is continuation of the home insulin regimen during
dehydration.101 Awareness is also increasing that com- the acute management of diabetic ketoacidosis. The UK
plications of diabetic ketoacidosis may continue beyond guideline recommends continuation of basal insulin;
the acute hospital admission. In one recent study of 3572 however, administration of basal insulin remains an area
patients with type 2 diabetes and 7144 controls matched of investigation in the US, as it may be associated with
for age, sex, and baseline diabetes complications and increased rates of hypoglycemia when used concurrently
comorbidities, patients with diabetic ketoacidosis were with the recommended rates of intravenous insulin. An
1.55 times more like to experience a stroke within six additional area that needs further investigation is the use
months than were those without diabetic ketoacido- of blood ketones to adjust the rate of intravenous insulin
sis.102 In another study examining long term outcomes in in the management of diabetic ketoacidosis, as is recom-
patients admitted to an intensive care unit with diabetic mended in the UK guidelines but not widely adopted in
ketoacidosis, one in 10 patients died within one year of the US owing to limitations and availability of point of
hospital discharge.103 The average age of patients in this care blood ketone meters. A third area needing further
report was 38 years, suggesting that these patients rep- investigation is when intravenous insulin is indicated in
resent a group for whom a heightened degree of surveil- the acute management of pure HHS without any evidence
lance with associated interventions is needed to offset of elevated ketones. The UK guidelines do not recommend
the mortality risk. routine use of intravenous insulin in the initial manage-
ment of HHS.
Recurrent diabetic ketoacidosis The weakness of any guideline is in the implemen-
Patients admitted to hospital with diabetic ketoacido- tation. Evidence suggests that guidelines for diabetic
sis represent a group at high risk for all cause mortality ketoacidosis and HHS are not consistently followed by
and hospital readmissions including recurrent diabetic many institutions in the US as well as the UK. Efforts
ketoacidosis.13 103-105 Those with a history of psychiatric to improve adherence to these guidelines may result in
illness or alcohol or substance misuse are at even higher improved outcomes for patients and reductions in costs
risk for readmission, which contributes to long term of care.41 67
risk for death.106 107 Interventions designed to improve a Also included in this manuscript is a discussion of
patient’s compliance with diabetes treatment and help to euglycemic diabetic ketoacidosis with use of the SGLT2
reduce the risk for recurrent diabetic ketoacidosis, such inhibitor class of drugs. The only published statement for
as those described in the prevention section of this paper, euglycemic diabetic ketoacidosis with this class of drugs
are essential. was published as a position statement in 2016 by the
American Association of Clinical Endocrinologists.5For
Summary of guidelines this disorder, education of patients and healthcare pro-
The ADA published a consensus statement guiding viders about recognition that this may be occurring is
treatment for adult patients with hyperglycemic cri- important to avoid significant patient morbidity.
ses in 2009.1 Although not specifically designated as a
guideline, it has been used for this purpose with several Emerging treatments
more recent reviews and commentaries that outline both Several studies are investigating optimal treatment
the strengths and weaknesses of this widely followed strategies with insulin and intravenous fluids, as well
report, as well as areas that are in need of revision or as alternative treatment strategies and previously unex-
future study.7 10 60 63 The most recent 2019 ADA clinical plored adverse outcomes in adult patients presenting
practice recommendations for management of diabetic with decompensated diabetes. Some of these trials are
ketoacidosis include several updates to the 2009 consen- listed below.
sus statement, as pointed out in table 1 and table 2.6 The Clinicaltrials.gov NCT02930044—The purpose of this
UK guidelines for management of diabetic ketoacidosis study is to determine whether adult patients with dia-
and HHS, written by the Joint British Diabetes Societies betic ketoacidosis who present to the emergency depart-
Inpatient Care Group and most recently updated in 2013 ment and are treated with glargine insulin (0.3 units/
and 2016, respectively, have also been subject to similar kg with a maximum dose of 30 units) within two hours
scrutiny to those of the ADA.2 4 after starting the intravenous insulin infusion need a
The sections on treatment of diabetic ketoacidosis and shorter duration of intravenous insulin administration
HHS included in this manuscript provide direct compari- to resolve diabetic ketoacidosis compared with standard
sons of recommendations of the ADA and UK guidelines care (insulin glargine administered two to three hours
for fluid resuscitation, electrolyte repletion, insulin ther- before stopping intravenous insulin). This study will
apy, and transitioning patients back to a home regimen provide additional information on use of long acting
of glycemic management where this is appropriate. No insulin preparations early in the course of treatment for
studies have compared strategies recommended by the diabetic ketoacidosis.74 75 It is scheduled to be completed
ADA and UK, and we believe that this type of study is not in November 2019.
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
identified before these drugs are administered? completed in 2020.
2 What are the optimal methods for continuation of Clinicaltrials.gov NCT02443415—The purpose of this
intravenous insulin therapy that minimize risk for study, which has been completed (2017) but not yet pub-
hypoglycemia? lished, was to evaluate changes in brain and memory
3 What type of fluid replacement strategies would function using cognitive testing and magnetic resonance
allow for correction of dehydration without resultant imaging in patients with diabetes who have no history of
hyperchloremic metabolic acidosis? diabetic ketoacidosis or who have one episode or more
4 Does a therapeutic approach exist that would further than three episodes of diabetic ketoacidosis. Healthy
minimize the risk of cerebral edema as a complication of patients without diabetes were also studied.
diabetic ketoacidosis in young adults?
5 What interventions effectively reduce the risk of recurrent Conclusions
diabetic ketoacidosis in patients who have frequent
Diabetic ketoacidosis and HHS are serious metabolic emer-
hospital readmissions?
gencies that have potential for adverse outcomes when not
promptly recognized and treated. A worrisome increase in
HOW PATIENTS WERE INVOLVED IN THE CREATION OF THIS the incidence of admissions for diabetic ketoacidosis has
MANUSCRIPT
occurred over the past two decades, with insulin omission
We invited three people with type 1 diabetes and a personal
and infection remaining the two predominant precipitat-
previous history of diabetic ketoacidosis who were known
to us to review advanced drafts of this manuscript. We
ing causes.12 13 15 The introduction of the SGLT2 inhibitors
asked them for their comments about what they felt were has increased awareness of euglycemic diabetic ketoaci-
areas that were missed and what may have been correctly dosis, which can occur with use of these agents as well as
or incorrectly emphasized. The consensus among these in other clinical situations.24 Educating patients to check
reviewers was that an emphasis was needed on the patient urine or blood ketones in the presence of symptoms such
and family experience of decompensated diabetes as well as nausea, vomiting, and/or fatigue, even if blood glucose
as on prevention of recurrent episodes. All reported either concentrations are below 11.1 mmol/L (200 mg/dL), is
personal or observed episodes of insulin doses being
important as these symptoms may represent early warning
erroneously withheld during a hospital admission. Another
suggestion was to emphasize the need to recognize diabetic signs for diabetic ketoacidosis.50
ketoacidosis in both older and younger adult patients. The Despite well written published guidelines for management
patient reviewers suggested emphasizing education on sick of diabetic ketoacidosis and HHS, these guidelines are not
day management strategies such as checking for urine or always followed.41 42 67 This can result in rebound hyperglyce-
blood ketones and preventing deterioration from ketosis mia, hypoglycemia, electrolyte abnormalities, and recurrence
or severe hyperglycemia due to diabetic ketoacidosis or of diabetic ketoacidosis.67 Additionally, common pitfalls in
hyperosmolar hyperglycemic syndrome (HHS), respectively. management include premature termination of intravenous
In addition, they recommended an emphasis on providing
insulin therapy and insufficient timing or dosing of subcuta-
information on what to expect over the course of treatment
for diabetic ketoacidosis and HHS as a way of decreasing neous insulin before discontinuation of intravenous insulin.
fear and anxiety in patients and their families. As a result of There is increasing recognition that a large percentage
their input, we clarified several areas of the manuscript and of patients who are admitted with diabetic ketoacidosis
gave additional emphasis to other areas. are a group at high risk for other life threatening events,
mortality, and hospital readmission.103 107 Strategies to pre-
Clinicaltrials.gov NCT03717896—The purposes of this vent hospital readmissions, as well as long term morbidity
randomized, double blind, placebo controlled trial are to and mortality, include engagement with diabetes educa-
determine whether administration of intravenous thia- tors who can educate patients on strategies for managing
mine (200 mg in normal saline twice daily for two days) blood glucose concentrations during periods of illness and
will lead to quicker resolution of acidosis in patients with instruct patients on when to call their physician or diabetes
diabetic ketoacidosis and to investigate whether thiamine healthcare provider for additional information.86 A greater
improves cellular oxygen consumption, shortens length focus is needed to educate physicians and patients and
of stay in hospital, or decreases resource use. This novel their families about early recognition, causes, and strate-
study is based on preliminary studies from the investiga- gies for prevention of acute decompensated diabetes.
tors showing that thiamine concentrations, which are defi- We thank Paul Strumph, Linda F Fried, and Heba Ismail for reviewing an
cient in up to 37% of patients with diabetic ketoacidosis, earlier version of this manuscript and for providing commentary that
was helpful in making this a patient based review of an important topic.
are inversely associated with the severity of acidosis. It is We also thank Kellie Antinori-Lent for her expert suggestions for the
scheduled to be completed in 2023. section on patient education.
Clinicaltrials.gov NCT02172092—The purpose of this Contributors: EKF did the primary literature review for this manuscript and
study is to assess shifts in blood and urine concentra- wrote the sections on the epidemiology and pathophysiology of diabetic
ketoacidosis and hyperosmolar hyperglycemic syndrome (HHS). ACD
tions of sodium, chloride, and hemoglobin during fluid wrote the sections on acute management of diabetic ketoacidosis and
resuscitation in patients with diabetic ketoacidosis. It is HHS and carefully compared and contrasted published guidelines from
scheduled to be completed in December 2019. the UK and US. MTK guided the writing of the full manuscript and assumed
primary responsibility for the sections on clinical presentation, prevention,
Clinicaltrials.gov NCT02864914—This study aims to and complications. All authors reviewed all sections of the manuscript,
assess the risk of acute liver or kidney injury, urinary providing suggestions for included content and references.
Competing interests: We have read and understood the BMJ policy on 25 Nyenwe EA, Kitabchi AE. The evolution of diabetic ketoacidosis: An update
declaration of interests and declare the following interests: none. of its etiology, pathogenesis and management. Metabolism 2016;65:507-
Provenance and peer review: Commissioned; externally peer reviewed. 21. 10.1016/j.metabol.2015.12.007 pmid:26975543.
1 Kitabchi AE, Umpierrez GE, Miles JM, Fisher JN. Hyperglycemic crises 26 Miles JM, Rizza RA, Haymond MW, Gerich JE. Effects of acute insulin
deficiency on glucose and ketone body turnover in man: evidence for
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
in adult patients with diabetes. Diabetes Care 2009;32:1335-43.
10.2337/dc09-9032 pmid:19564476. the primacy of overproduction of glucose and ketone bodies in the
2 Joint British Diabetes Societies Inpatient Care Group. The management genesis of diabetic ketoacidosis. Diabetes 1980;29:926-30. 10.2337/
of diabetic ketoacidosis in adults. 2nd edition. 2013. http://www. diab.29.11.926 pmid:6775996.
diabetologists-abcd.org.uk/JBDS/JBDS_IP_DKA_Adults_Revised.pdf. 27 DeFronzo RA, Ferrannini E. Regulation of hepatic glucose metabolism
3 Pasquel FJ, Umpierrez GE. Hyperosmolar hyperglycemic state: a historic in humans. Diabetes Metab Rev 1987;3:415-59. 10.1002/
review of the clinical presentation, diagnosis, and treatment. Diabetes dmr.5610030204 pmid:3552529.
Care 2014;37:3124-31. 10.2337/dc14-0984 pmid:25342831. 28 Exton JH. Mechanisms of hormonal regulation of hepatic glucose
4 Scott AR. Joint British Diabetes Societies (JBDS) for Inpatient Care JBDS metabolism. Diabetes Metab Rev 1987;3:163-83. 10.1002/
hyperosmolar hyperglycaemic guidelines group. Management of dmr.5610030108 pmid:3032541.
hyperosmolar hyperglycaemic state in adults with diabetes. Diabet Med 29 Meyer C, Stumvoll M, Dostou J, Welle S, Haymond M, Gerich J. Renal
2015;32:714-24. 10.1111/dme.12757 pmid:25980647. substrate exchange and gluconeogenesis in normal postabsorptive
5 Handelsman Y, Henry RR, Bloomgarden ZT, et al. American Association humans. Am J Physiol Endocrinol Metab 2002;282:E428-34. 10.1152/
of Clinical Endocrinologists and American College of Endocrinology ajpendo.00116.2001 pmid:11788376.
Position Statement On The Association Of SGLT-2 Inhibitors and Diabetic 30 Meyer C, Stumvoll M, Welle S, Woerle HJ, Haymond M, Gerich J. Relative
Ketoacidosis. Endocr Pract 2016;22:753-62. 10.4158/EP161292. importance of liver, kidney, and substrates in epinephrine-induced
PS pmid:27082665. increased gluconeogenesis in humans. Am J Physiol Endocrinol Metab
6 American Diabetes Association. 15. Diabetes Care in the Hospital: 2003;285:E819-26. 10.1152/ajpendo.00145.2003 pmid:12959936.
Standards of Medical Care in Diabetes-2019. Diabetes Care 31 Foster DW, McGarry JD. The metabolic derangements and treatment
2019;42(Suppl 1):S173-81. 10.2337/dc19-S015 pmid:30559241. of diabetic ketoacidosis. N Engl J Med 1983;309:159-69. 10.1056/
7 Umpierrez G, Korytkowski M. Diabetic emergencies - NEJM198307213090307 pmid:6408476.
ketoacidosis, hyperglycaemic hyperosmolar state and
32 Miles JM, Haymond MW, Nissen SL, Gerich JE. Effects of free fatty acid
hypoglycaemia. Nat Rev Endocrinol 2016;12:222-32. 10.1038/
availability, glucagon excess, and insulin deficiency on ketone body
nrendo.2016.15 pmid:26893262.
production in postabsorptive man. J Clin Invest 1983;71:1554-61.
8 Wachtel TJ, Tetu-Mouradjian LM, Goldman DL, Ellis SE, O’Sullivan PS.
10.1172/JCI110911 pmid:6134753.
Hyperosmolarity and acidosis in diabetes mellitus: a three-year
experience in Rhode Island. J Gen Intern Med 1991;6:495-502. 33 Miles JM, Gerich JE. Glucose and ketone body kinetics in diabetic
10.1007/BF02598216 pmid:1765864. ketoacidosis. Clin Endocrinol Metab 1983;12:303-19. 10.1016/S0300-
9 Munro JF, Campbell IW, McCuish AC, Duncan LJ. Euglycaemic 595X(83)80043-7 pmid:6409465.
diabetic ketoacidosis. Br Med J 1973;2:578-80. 10.1136/ 34 Palmer BF, Clegg DJ. Electrolyte and Acid-Base Disturbances in Patients
bmj.2.5866.578 pmid:4197425. with Diabetes Mellitus. N Engl J Med 2015;373:548-59. 10.1056/
10 Dhatariya KK, Umpierrez GE. Guidelines for management of diabetic NEJMra1503102 pmid:26244308.
ketoacidosis: time to revise?Lancet Diabetes Endocrinol 2017;5:321-3. 35 McGarry JD, Foster DW. Regulation of hepatic fatty acid oxidation and
10.1016/S2213-8587(17)30093-1 pmid:28372975. ketone body production. Annu Rev Biochem 1980;49:395-420. 10.1146/
11 Troy PJ, Clark RP, Kakarala SG, Burns J, Silverman IE, Shore E. Cerebral annurev.bi.49.070180.002143 pmid:6157353.
edema during treatment of diabetic ketoacidosis in an adult with 36 Dhatariya K. Blood Ketones: Measurement, Interpretation, Limitations,
new onset diabetes. Neurocrit Care 2005;2:55-8. 10.1385/ and Utility in the Management of Diabetic Ketoacidosis. Rev Diabet Stud
NCC:2:1:055 pmid:16174971. 2016;13:217-25. 10.1900/RDS.2016.13.217 pmid:28278308.
12 Benoit SR, Zhang Y, Geiss LS, Gregg EW, Albright A. Trends in Diabetic 37 Umpierrez GE, Kelly JP, Navarrete JE, Casals MM, Kitabchi AE. Hyperglycemic
Ketoacidosis Hospitalizations and In-Hospital Mortality - United States, crises in urban blacks. Arch Intern Med 1997;157:669-75. 10.1001/
2000-2014. MMWR Morb Mortal Wkly Rep 2018;67:362-5. 10.15585/ archinte.1997.00440270117011 pmid:9080921.
mmwr.mm6712a3 pmid:29596400. 38 Randall L, Begovic J, Hudson M, et al. Recurrent diabetic ketoacidosis in inner-
13 Zhong VW, Juhaeri J, Mayer-Davis EJ. Trends in Hospital Admission for city minority patients: behavioral, socioeconomic, and psychosocial factors.
Diabetic Ketoacidosis in Adults With Type 1 and Type 2 Diabetes in Diabetes Care 2011;34:1891-6. 10.2337/dc11-0701 pmid:21775761.
England, 1998-2013: A Retrospective Cohort Study. Diabetes Care
39 Lohiya S, Kreisberg R, Lohiya V. Recurrent diabetic ketoacidosis in two
2018;41:1870-7. 10.2337/dc17-1583 pmid:29386248.
community teaching hospitals. Endocr Pract 2013;19:829-33. 10.4158/
14 Javor KA, Kotsanos JG, McDonald RC, Baron AD, Kesterson JG, Tierney WM.
EP13057.RA pmid:23757621.
Diabetic ketoacidosis charges relative to medical charges of adult
patients with type I diabetes. Diabetes Care 1997;20:349-54. 10.2337/ 40 Ananth J, Parameswaran S, Gunatilake S. Side effects of atypical
diacare.20.3.349 pmid:9051386. antipsychotic drugs. Curr Pharm Des 2004;10:2219-29.
15 Centers for Disease Control and Prevention. National Diabetes Statistics 10.2174/1381612043384088 pmid:15281897.
Report. 2017. https://www.cdc.gov/diabetes/data/statistics/statistics- 41 Dhatariya K, Nunney I, Iceton G. Institutional factors in the management
report.html. of adults with diabetic ketoacidosis in the UK: results of a national survey.
16 Vellanki P, Umpierrez GE. Diabetic ketoacidosis: A common debut of Diabet Med 2016;33:269-70. 10.1111/dme.12877 pmid:26235104.
diabetes among african americans with type 2 diabetes. Endocr Pract 42 Dhatariya KK, Nunney I, Higgins K, Sampson MJ, Iceton G. National survey
2017;23:971-8. 10.4158/EP161679.RA pmid:28534682. of the management of Diabetic Ketoacidosis (DKA) in the UK in 2014.
17 Cefalu WT, Dawes DE, Gavlak G, et al. Insulin Access and Affordability Diabet Med 2016;33:252-60. 10.1111/dme.12875 pmid:26286235.
Working Group: Conclusions and Recommendations. Diabetes Care 43 Polonsky WH, Anderson BJ, Lohrer PA, Aponte JE, Jacobson AM, Cole CF.
2018;41:1299-311. 10.2337/dci18-0019 pmid:29739814. Insulin omission in women with IDDM. Diabetes Care 1994;17:1178-85.
18 Dhatariya KK, Skedgel C, Fordham R. The cost of treating diabetic 10.2337/diacare.17.10.1178 pmid:7821139.
ketoacidosis in the UK: a national survey of hospital resource use. Diabet 44 Giessmann LC, Kann PH. Risk and Relevance of Insulin Pump Therapy in
Med 2017;34:1361-6. 10.1111/dme.13427 pmid:28727175. the Aetiology of Ketoacidosis in People with Type 1 Diabetes. Exp Clin
19 Delaney MF, Zisman A, Kettyle WM. Diabetic ketoacidosis and Endocrinol Diabetes 2018.pmid:30049002.
hyperglycemic hyperosmolar nonketotic syndrome. Endocrinol Metab 45 Yogi-Morren D, Lansang MC. Management of patients with type 1 diabetes
Clin North Am 2000;29:683-705, V. 10.1016/S0889-8529(05)70159- in the hospital. Curr Diab Rep 2014;14:458. 10.1007/s11892-013-
6 pmid:11149157. 0458-9 pmid:24398661.
20 Fadini GP, de Kreutzenberg SV, Rigato M, et al. Characteristics and
46 Blau JE, Tella SH, Taylor SI, Rother KI. Ketoacidosis associated with SGLT2
outcomes of the hyperglycemic hyperosmolar non-ketotic syndrome in a
inhibitor treatment: Analysis of FAERS data. Diabetes Metab Res Rev
cohort of 51 consecutive cases at a single center. Diabetes Res Clin Pract
2017;33. 10.1002/dmrr.2924 pmid:28736981.
2011;94:172-9. 10.1016/j.diabres.2011.06.018 pmid:21752485.
21 Peters AL, Buschur EO, Buse JB, Cohan P, Diner JC, Hirsch IB. Euglycemic 47 Rosenstock J, Ferrannini E. Euglycemic Diabetic Ketoacidosis: A
Diabetic Ketoacidosis: A Potential Complication of Treatment Predictable, Detectable, and Preventable Safety Concern With SGLT2
With Sodium-Glucose Cotransporter 2 Inhibition. Diabetes Care Inhibitors. Diabetes Care 2015;38:1638-42. 10.2337/dc15-
2015;38:1687-93. 10.2337/dc15-0843 pmid:26078479. 1380 pmid:26294774.
22 Meyer EJ, Gabb G, Jesudason D. SGLT2 Inhibitor-Associated Euglycemic 48 Taylor SI, Blau JE, Rother KI. SGLT2 Inhibitors May Predispose to
Diabetic Ketoacidosis: A South Australian Clinical Case Series and Ketoacidosis. J Clin Endocrinol Metab 2015;100:2849-52. 10.1210/
Australian Spontaneous Adverse Event Notifications. Diabetes Care jc.2015-1884 pmid:26086329.
2018;41:e47-9. 10.2337/dc17-1721 pmid:29440112. 49 Hattersley AT, Thorens B. Type 2 Diabetes, SGLT2 Inhibitors, and
23 Ueda P, Svanström H, Melbye M, et al. Sodium glucose cotransporter 2 Glucose Secretion. N Engl J Med 2015;373:974-6. 10.1056/
inhibitors and risk of serious adverse events: nationwide register based cohort NEJMcibr1506573 pmid:26332554.
study. BMJ 2018;363:k4365. 10.1136/bmj.k4365 pmid:30429124. 50 Garg SK, Peters AL, Buse JB, Danne T. Strategy for Mitigating DKA Risk
24 Modi A, Agrawal A, Morgan F. Euglycemic Diabetic in Patients with Type 1 Diabetes on Adjunctive Treatment with SGLT
Ketoacidosis: A Review. Curr Diabetes Rev 2017;13:315-21. Inhibitors: A STICH Protocol. Diabetes Technol Ther 2018;20:571-5.
10.2174/1573399812666160421121307 pmid:27097605. 10.1089/dia.2018.0246 pmid:30129772.
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
diab.23.6.525 pmid:4834291. 76 Andrade-Castellanos CA, Colunga-Lozano LE, Delgado-Figueroa N, Gonzalez-
53 Hara JS, Rahbar AJ, Jeffres MN, Izuora KE. Impact of a hyperglycemic Padilla DA. Subcutaneous rapid-acting insulin analogues for diabetic
crises protocol. Endocr Pract 2013;19:953-62. 10.4158/EP13077. ketoacidosis. Cochrane Database Syst Rev 2016;(1):CD011281.pm
OR pmid:23807521. id:26798030.
54 Bull SV, Douglas IS, Foster M, Albert RK. Mandatory protocol for treating 77 Ersöz HO, Ukinc K, Köse M, et al. Subcutaneous lispro and intravenous
adult patients with diabetic ketoacidosis decreases intensive care regular insulin treatments are equally effective and safe for the
unit and hospital lengths of stay: results of a nonrandomized trial. Crit treatment of mild and moderate diabetic ketoacidosis in adult
Care Med 2007;35:41-6. 10.1097/01.CCM.0000249825.18677. patients. Int J Clin Pract 2006;60:429-33. 10.1111/j.1368-
D2 pmid:17095944. 5031.2006.00786.x pmid:16620355.
55 Thuzar M, Malabu UH, Tisdell B, Sangla KS. Use of a standardised 78 Umpierrez GE, Cuervo R, Karabell A, Latif K, Freire AX, Kitabchi AE. Treatment
diabetic ketoacidosis management protocol improved clinical of diabetic ketoacidosis with subcutaneous insulin aspart. Diabetes Care
outcomes. Diabetes Res Clin Pract 2014;104:e8-11. 10.1016/j. 2004;27:1873-8. 10.2337/diacare.27.8.1873 pmid:15277410.
diabres.2014.01.016 pmid:24507867. 79 Umpierrez GE, Latif K, Stoever J, et al. Efficacy of subcutaneous insulin
56 Ullal J, Aloi JA, Reyes-Umpierrez D, et al. Comparison of Computer- lispro versus continuous intravenous regular insulin for the treatment of
Guided Versus Standard Insulin Infusion Regimens in Patients With patients with diabetic ketoacidosis. Am J Med 2004;117:291-6. 10.1016/j.
Diabetic Ketoacidosis. J Diabetes Sci Technol 2018;12:39-46. amjmed.2004.05.010 pmid:15336577.
10.1177/1932296817750899 pmid:29291648. 80 Karoli R, Fatima J, Salman T, Sandhu S, Shankar R. Managing diabetic
57 Laliberte B, Yeung SYA, Gonzales JP. Impact of diabetic ketoacidosis ketoacidosis in non-intensive care unit setting: Role of insulin
management in the medical intensive care unit after order set analogs. Indian J Pharmacol 2011;43:398-401. 10.4103/0253-
implementation. Int J Pharm Pract 2017;25:238-43. 10.1111/ 7613.83109 pmid:21844993.
ijpp.12346 pmid:28338247. 81 Chua HR, Schneider A, Bellomo R. Bicarbonate in diabetic ketoacidosis - a
58 Martin C, McKinney ZJ, Hoody D, Fish L. Diabetic ketoacidosis critical care systematic review. Ann Intensive Care 2011;1:23. 10.1186/2110-5820-1-
pathway implementation: Incorporation into emr significantly decreases 23 pmid:21906367.
length of stay. Endocr Pract 2016;22:673-8. 10.4158/EP151073. 82 Frontino G, Bonfanti R, Rigamonti A, et al. Management of hyperosmolar
OR pmid:27176140. hyperglycaemic state in adults with diabetes. Diabet Med 2016;33:552.
59 Devalia B. Adherance to protocol during the acute management of 10.1111/dme.12866 pmid:26206227.
diabetic ketoacidosis: would specialist involvement lead to better 83 Magaji V, Nayak S, Donihi AC, et al. Comparison of insulin infusion protocols
outcomes?Int J Clin Pract 2010;64:1580-2. 10.1111/j.1742- targeting 110-140 mg/dL in patients after cardiac surgery. Diabetes Technol
1241.2010.02348.x pmid:20846206. Ther 2012;14:1013-7. 10.1089/dia.2012.0114 pmid:23046399.
60 Tran TTT, Pease A, Wood AJ, et al. Review of Evidence for Adult Diabetic 84 George S, Dale J, Stanisstreet D. Joint British Diabetes Societies (JBDS) for
Ketoacidosis Management Protocols. Front Endocrinol (Lausanne) Inpatient Care JBDS Medical VRIII Writing Group. A guideline for the use of
2017;8:106. 10.3389/fendo.2017.00106 pmid:28659865. variable rate intravenous insulin infusion in medical inpatients. Diabet Med
61 Patiño AM, Marsh RH, Nilles EJ, Baugh CW, Rouhani SA, Kayden S. 2015;32:706-13. 10.1111/dme.12756 pmid:25980646.
Facing the Shortage of IV Fluids - A Hospital-Based Oral 85 Mays JA, Jackson KL, Derby TA, et al. An Evaluation of Recurrent Diabetic
Rehydration Strategy. N Engl J Med 2018;378:1475-7. 10.1056/ Ketoacidosis, Fragmentation of Care, and Mortality Across Chicago, Illinois.
NEJMp1801772 pmid:29561701. Diabetes Care 2016;39:1671-6. 10.2337/dc16-0668 pmid:27422579.
62 Misra S, Oliver NS. Utility of ketone measurement in the prevention, 86 Donihi AC. Practical Recommendations for Transitioning Patients with Type
diagnosis and management of diabetic ketoacidosis. Diabet Med 2 Diabetes from Hospital to Home. Curr Diab Rep 2017;17:52. 10.1007/
2015;32:14-23. 10.1111/dme.12604 pmid:25307274. s11892-017-0876-1 pmid:28573408.
63 Dhatariya KK, Vellanki P. Treatment of Diabetic Ketoacidosis (DKA)/ 87 Vellanki P, Umpierrez GE. Increasing Hospitalizations for DKA: A Need for
Hyperglycemic Hyperosmolar State (HHS): Novel Advances in the Prevention Programs. Diabetes Care 2018;41:1839-41. 10.2337/dci18-
Management of Hyperglycemic Crises (UK Versus USA). Curr Diab Rep 0004 pmid:30135197.
2017;17:33. 10.1007/s11892-017-0857-4 pmid:28364357. 88 Kimmel B, Sullivan MM, Rushakoff RJ. Survey on transition from inpatient to
64 Ginde AA, Pelletier AJ, Camargo CA Jr. National study of U.S. emergency outpatient for patients on insulin: what really goes on at home?Endocr Pract
department visits with diabetic ketoacidosis, 1993-2003. Diabetes Care 2010;16:785-91. 10.4158/EP10013.OR pmid:20350914.
2006;29:2117-9. 10.2337/dc06-0627 pmid:16936163. 89 Krall JS, Donihi AC, Hatam M, Koshinsky J, Siminerio L. The Nurse Education
65 Sampson M, Jones C. Joint British Diabetes Societies (JBDS) for Inpatient and Transition (NEAT) model: educating the hospitalized patient with
Care. Joint British Diabetes Societies for Inpatient Care: clinical guidelines diabetes. Clin Diabetes Endocrinol 2016;2:1. 10.1186/s40842-016-
and improving inpatient diabetes care. Diabet Med 2018;35:988-91. 0020-1 pmid:28702237.
10.1111/dme.13672 pmid:30152584. 90 Evans NR, Richardson LS, Dhatariya KK, Sampson MJ. Diabetes specialist
66 Krentz AJ, Hale PJ, Singh BM, Nattrass M. The effect of glucose and nurse telemedicine: admissions avoidance, costs and casemix. Eur Diabetes
insulin infusion on the fall of ketone bodies during treatment of diabetic Nurs 2012;9:17-2110.1002/edn.198.
ketoacidosis. Diabet Med 1989;6:31-6. 10.1111/j.1464-5491.1989. 91 Berg EG. Sick-day rules: What to do when you are feeling under the weather.
tb01135.x pmid:2522371. Diabetes Forecast 2013;66:32-3.pmid:24224354.
67 Karajgikar ND, Manroa P, Acharya R, et al. Addressing pitfalls in 92 Auchterlonie A, Okosieme OE. Preventing diabetic ketoacidosis: do
management of diabetic ketoacidosis (DKA) with a standardized protocol. patients adhere to sick-day rules?Clin Med (Lond) 2013;13:120. 10.7861/
Endocr Pract 2019. 10.4158/EP-2018-0398 pmid:30657360. clinmedicine.13-1-120 pmid:23472520.
68 Haas NL, Gianchandani RY, Gunnerson KJ, et al. The Two-Bag Method for 93 Scheen AJ. Pharmacokinetic Characteristics and Clinical Efficacy of an SGLT2
Treatment of Diabetic Ketoacidosis in Adults. J Emerg Med 2018;54:593- Inhibitor Plus DPP-4 Inhibitor Combination Therapy in Type 2 Diabetes.
9. 10.1016/j.jemermed.2018.01.004 pmid:29628184. Clin Pharmacokinet 2017;56:703-18. 10.1007/s40262-016-0498-
69 Munir I, Fargo R, Garrison R, et al. Comparison of a ‘two-bag system’ versus 9 pmid:28039605.
conventional treatment protocol (‘one-bag system’) in the management 94 Rosenbloom AL. Intracerebral crises during treatment of diabetic
of diabetic ketoacidosis. BMJ Open Diabetes Res Care 2017;5:e000395. ketoacidosis. Diabetes Care 1990;13:22-33. 10.2337/
10.1136/bmjdrc-2017-000395 pmid:28878933. diacare.13.1.22 pmid:2105195.
70 Mazer-Amirshahi M, Fox ER. Saline Shortages - Many Causes, No 95 Kuppermann N, Ghetti S, Schunk JE, et al. PECARN DKA FLUID Study Group.
Simple Solution. N Engl J Med 2018;378:1472-4. 10.1056/ Clinical Trial of Fluid Infusion Rates for Pediatric Diabetic Ketoacidosis. N Engl
NEJMp1800347 pmid:29561694. J Med 2018;378:2275-87. 10.1056/NEJMoa1716816 pmid:29897851.
71 Kitabchi AE, Murphy MB, Spencer J, Matteri R, Karas J. Is a priming dose 96 Wolfsdorf JI, Glaser N, Agus M, et al. ISPAD Clinical Practice Consensus
of insulin necessary in a low-dose insulin protocol for the treatment of Guidelines 2018: Diabetic ketoacidosis and the hyperglycemic
diabetic ketoacidosis?Diabetes Care 2008;31:2081-5. 10.2337/dc08- hyperosmolar state. Pediatr Diabetes 2018;19(Suppl 27):155-77.
0509 pmid:18694978. 10.1111/pedi.12701 pmid:29900641.
72 Cardoso L, Vicente N, Rodrigues D, Gomes L, Carrilho F. 97 Mårtensson J, Bailey M, Venkatesh B, et al. Intensity of early correction
Controversies in the management of hyperglycaemic emergencies of hyperglycaemia and outcome of critically ill patients with diabetic
in adults with diabetes. Metabolism 2017;68:43-54. 10.1016/j. ketoacidosis. Crit Care Resusc 2017;19:266-73.pmid:28866977.
metabol.2016.11.010 pmid:28183452. 98 Dhatariya KK. Diabetic ketoacidosis. BMJ 2007;334:1284-5. 10.1136/
73 Loh TP, Saw S, Sethi SK. Bedside monitoring of blood bmj.39237.661111.80 pmid:17585123.
ketone for management of diabetic ketoacidosis: proceed 99 Van Zyl DG, Rheeder P, Delport E. Fluid management in diabetic-acidosis-
with care. Diabet Med 2012;29:827-8. 10.1111/j.1464- -Ringer’s lactate versus normal saline: a randomized controlled trial. QJM
5491.2011.03490.x pmid:21999575. 2012;105:337-43. 10.1093/qjmed/hcr226 pmid:22109683.
74 Hsia E, Seggelke S, Gibbs J, et al. Subcutaneous administration of 100 Mahler SA, Conrad SA, Wang H, Arnold TC. Resuscitation with balanced
glargine to diabetic patients receiving insulin infusion prevents rebound electrolyte solution prevents hyperchloremic metabolic acidosis in
hyperglycemia. J Clin Endocrinol Metab 2012;97:3132-7. 10.1210/ patients with diabetic ketoacidosis. Am J Emerg Med 2011;29:670-4.
jc.2012-1244 pmid:22685233. 10.1016/j.ajem.2010.02.004 pmid:20825879.
101 Issa M, Alqahtani F, Ziada KM, et al. Incidence and Outcomes of 104 Bradford AL, Crider CC, Xu X, Naqvi SH. Predictors of Recurrent Hospital
Non-ST Elevation Myocardial Infarction in Patients Hospitalized Admission for Patients Presenting With Diabetic Ketoacidosis and
with Decompensated Diabetes. Am J Cardiol 2018;122:1297-302. Hyperglycemic Hyperosmolar State. J Clin Med Res 2017;9:35-9.
10.1016/j.amjcard.2018.07.004 pmid:30131108. 10.14740/jocmr2792w pmid:27924173.
102 Chen YL, Weng SF, Yang CY, Wang JJ, Tien KJ. Long-term risk of 105 Gibb FW, Teoh WL, Graham J, Lockman KA. Risk of death following
BMJ: first published as 10.1136/bmj.l1114 on 29 May 2019. Downloaded from http://www.bmj.com/ on 29 May 2019 by guest. Protected by copyright.
stroke in type 2 diabetes patients with diabetic ketoacidosis: admission to a UK hospital with diabetic ketoacidosis. Diabetologia
A population-based, propensity score-matched, longitudinal 2016;59:2082-7. 10.1007/s00125-016-4034-0 pmid:27397023.
follow-up study. Diabetes Metab 2017;43:223-8. 10.1016/j. 106 Isidro ML, Jorge S. Recreational drug abuse in patients hospitalized for
diabet.2016.11.003 pmid:28129999. diabetic ketosis or diabetic ketoacidosis. Acta Diabetol 2013;50:183-7.
103 Azevedo LC, Choi H, Simmonds K, Davidow J, Bagshaw SM. Incidence 10.1007/s00592-010-0243-z pmid:21136122.
and long-term outcomes of critically ill adult patients with moderate-to- 107 McCoy RG, Herrin J, Lipska KJ, Shah ND. Recurrent hospitalizations
severe diabetic ketoacidosis: retrospective matched cohort study. for severe hypoglycemia and hyperglycemia among U.S. adults with
J Crit Care 2014;29:971-7. 10.1016/j. diabetes. J Diabetes Complications 2018;32:693-701. 10.1016/j.
jcrc.2014.07.034 pmid:25220529. jdiacomp.2018.04.007 pmid:29751961.