Diabetic Ketoacidosis in Adults
Andrew J Krentz and Helen B Holt

Diabetic ketoacidosis has a reported average mortality of approximately five per cent in Western countries. Mortality is generally higher at the extremes of age. Common precipitating causes include infection, insulin management errors, omission of insulin and new cases of diabetes; in many cases no cause is obvious.
Although traditionally considered uncommon in patients with  type 2 diabetes, reports in recent years have drawn attention to diabetic ketoacidosis in non-white patients who are often able to discontinue insulin after recovery.

Ketoacidosis develops when there is an absolute or, more commonly, a relative insulin deficiency, usually in concert with an increase in catabolic hormone concentrations. Hepatic overproduction
of glucose and ketone bodies is compounded by
diminished clearance in peripheral tissues.
Emergencies in Diabetes Edited by Andrew J. Krentz
# 2004 John Wiley & Sons, Ltd ISBN 0-471-49814-9
Definition
The cardinal biochemical features of diabetic ketoacidosis are
hyperketonaemia and metabolic acidosis, in concert with variable
hyperglycaemia. While no univerally agreed criteria for diagnosis
exist, diabetic ketoacidosis may be defined as in Box 1.1.
Box 1.1 Definition of diabetic ketoacidosis
Severe uncontrolled diabetes requiring emergency treatment
with insulin and intravenous fluids
Blood total ketone body concentration, i.e. the sum of
acetoacetate and 3-hydroxybutyrate 5 mmol/L.
Note that biochemical confirmation of the diagnosis is often based
on semi-quantitative urine dipstick methods, a minority of centres
having clinical chemistry laboratories that can measure blood
ketone body concentrations. The diagnostic criteria that we use are
presented in Box 1.2. Note that no threshold for hyperglycaemia is
included in this definition, reflecting the wide variability in blood
glucose concentrations.
Box 1.2 Practical biochemical definition of diabetic ketoacidosis
Blood bicarbonate concentration (capillary or arterial)
15 mmol/L
Significant ketosis, defined as urine Ketostix1 (Bayer
Diagnostics) reaction þþ or plasma Ketostix1 þ or more).
4 DIABETIC KETOACIDOSIS IN ADULTS
Mortality
Diabetic ketoacidosis continues to be an important cause of death
among patients with type 1 diabetes. The average mortality rate for
ketoacidosis today is quoted as between 5 and 10 per cent although
rates vary widely. Experienced centres would expect to report a
mortality rate of less than five per cent. Some deaths are inevitable
consequences of associated medical conditions such as overwhelming
infection. Clearly, the mortality associated with diabetic
ketoacidosis has not been abolished, despite the ready availability
of insulin, at least in Western countries. Mortality rates for diabetic
ketoacidosis vary with prevailing socio-economic factors and the
provision of general medical care.
Developing countries. In Tanzania during the 1980s, for example,
hospital mortality rates from ketoacidosis were reminiscent of
those during the pre-insulin era in the West.
Industrialised countries. Even in countries with well developed
health care systems some deaths associated with ketoacidosis
are potentially preventable, arising from factors such as
delays in presentation or diagnosis
errors in management, either by the patient on the part of
medical attendants.
Differences in (1) the definition of ketoacidosis and (2) patient
selection may also account for a proportion of this variation.
Mortality need not be higher in non-university hospitals than in
teaching centres if appropriate guidelines for management are
implemented. Mortality is generally higher in certain groups of
patients such as the very elderly.
MORTALITY 5
Precipitating factors
Infection. This is the commonest identifiable cause of ketoacidosis
reported in the literature, accounting for approximately
35 per cent of all episodes, pneumonia and urinary tract
infections being the most frequent types.
New cases of diabetes. These account for approximately 10–15 per
cent of episodes.
Management errors. These include inappropriate changes in
insulin dosage initiated either by the patient or sometimes
following medical advice.
Recurrent ketoacidosis. This affects a small subgroup of patients,
the majority being females under the age of 20 years in whom
psychological problems lead to discontinuation of insulin.
Other factors. No precipitating cause is identified in approximately
25–35 per cent of episodes, although this depends on the
rigour with which the search is conducted. Ultimately, it must
be concluded that insufficient insulin has been administered.
Higher rates of ketoacidosis have been observed in some centres
specialising in continuous subcutaneous insulin infusion (CSII). In
the Diabetes Control and Complications Trial (DCCT), a higher rate
of ketoacidosis was observed in patients receiving CSII than
in those on multiple insulin injections. It has been suggested that
the small subcutaneous depot of regular insulin in CSII predisposes
to the rapid development of ketoacidosis if the infusion is
interrupted.
Discontinuation of insulin treatment is a common cause of
diabetic ketoacidosis
6 DIABETIC KETOACIDOSIS IN ADULTS
Pathogenesis
Diabetic ketoacidosis is characterised by marked elevations of
catabolic counter-regulatory hormone concentrations:
glucagon
catecholamines
cortisol
growth hormone.
These increases, often in concert with elevated levels of inflammatory
cytokines, occur in the presence of an absolute or relative
deficiency of insulin. While insulin may be measurable in plasma it
is, by definition, insufficient to maintain a normal metabolic state.
Insulin resistance, a reduced biological action of the hormone, is
acutely increased in diabetic ketoacidosis, compounding any
deficiency in circulating insulin. In patients with type 2 diabetes,
residual endogenous insulin secretion serves to protect against
ketoacidosis (see Chapter 3). However, suppression of b-cell
insulin secretion by catecholamines (via a-adrenergic receptors)
may occasionally precipitate ketoacidosis during acute severe
illness; this may be especially true of patients with type 2 diabetes
of long duration in whom b-cell function may be more compromised.
A practical rule of thumb is to regard any patient with
insulin-treated diabetes as having major insulin deficiency. This
has implications for the management of diabetes during severe
intercurrent illness, trauma and surgery (see Chapter 7).
Severe intercurrent illness may precipitate ketoacidosis in
patients with type 2 diabetes.
Under experimental conditions, withdrawal of insulin from
insulin-dependent patients leads to an early rise in plasma
glucagon. As hyperglycaemia and ketoacidosis develop, progressive
PATHOGENESIS 7
dehydration and acidosis stimulate the release of catecholamines
(adrenaline and nor-adrenaline) and cortisol. A vicious circle
develops in which worsening metabolic decompensation stimulates
further secretion of catabolic hormones. Hepatic overproduction
of glucose and ketone bodies initiates hyperglycaemia and
ketosis, while impaired disposal by peripheral tissues such as
muscle and brain maintains and exacerbates hyperglycaemia and
hyperketonaemia. The rate of glucose production subsequently
decreases towards normal but hyperglycaemia is maintained as the
rates of production and utilisation become equal. Insulin withdrawal
also results in a progressive increase in ketone body
production and utilization. However, the former exceeds the latter,
resulting in a progressive rise in plasma ketone bodies.
Insulin deficiency with elevated counter-regulatory hormones
initiates hyperglycaemia and ketosis
We will now consider the biochemical disturbances in more detail.
Hyperglycaemia. Insulin deficiency and elevated plasma levels
of catabolic hormones, particularly glucagon and catecholamines,
result in increased rates of hepatic glycogenolysis and
gluconeogenesis; in addition, renal gluconeogenesis is enhanced
by acidosis, although it contributes quantitatively less to
hyperglycaemia. Glucose disposal by tissues such as skeletal
muscle and adipocytes is reduced by insulin deficiency;
elevated plasma levels of catabolic hormones and non-esterified
fatty acids with acidosis induce insulin resistance.
Hyperketonaemia. In ketoacidosis, serum ketone body levels are
often raised to 200–300 times their normal fasting values.
Ketone bodies are strong organic acids that are fully dissociated
at physiological pH. In turn, this results in equimolar generation
of hydrogen ions (H
þ
) outstripping the buffering capacity of
fluids and tissues. Metabolic acidosis has a number of serious
detrimental physiological effects that account for many of the
8 DIABETIC KETOACIDOSIS IN ADULTS
serious clinical features of ketoacidosis:
negative inotropic effect on cardiac muscle
exacerbation of systemic hypotension through peripheral
vasodilatation
exacerbation of insulin resistance
increased risk of ventricular arrhythmias
respiratory depression with severe acidosis.
Ketogenesis
Insulin deficiency and catabolic hormone excess, especially of
catecholamines, promote lipolysis within adipocytes, wherein
triglycerides are converted to three fatty acids and one molecule
of glycerol. These effects are mediated via the activity of hormonesensitive
lipase (triacylglycerol lipase), an enzyme normally
regarded as being exquisitely sensitive to inhibition by insulin.
Concurrently, re-esterification, i.e. the formation of new triglycerides,
within adipocytes is impaired. This results in the net release of
long-chain non-esterified fatty acids and glycerol into the circulation
(Figure 1.1).
Glycerol is a gluconeogenic precursor while fatty acids are the
principal substrate for ketone bodies. Hepatic ketogenesis is enhanced
by increased portal delivery of fatty acids liberated from
adipocyte stores. In diabetic ketoacidosis, hepatic re-esterification
of fatty acids is impaired and fatty acids are preferentially partially
oxidised to ketone bodies within mitochondria (Figure 1.2).
1. Fatty acids are converted to coenzyme A (CoA) derivatives
prior to transportation into mitochondria by an active transport
system (the carnitine shuttle). The hormonal imbalance in
diabetic ketoacidosis, i.e. insulin deficiency and excess catabolic
hormones, favours entry of fatty acids into the mitochondria.
This is mediated via a glucagon-mediated decrease in the
PATHOGENESIS 9
cytosolic concentration of malonyl-CoA (via reduced conversion
of pyruvate to acetyl coenzyme A), a potent competitive
inhibitor of carnitine-palmitoyl transferase I. The latter enzyme
is responsible for transport of fatty acyl-CoA derivatives across
the inner mitochondrial membrane.
2. Carnitine-palmitoyl transferase II subsequently liberates fatty
acyl-CoA within the mitochondria. Fatty acyl-CoA undergoes
b-oxidation, forming acetyl-CoA. Carnitine returns to the extramitochondrial
space.
3. Acetyl CoA can then be completely oxidised in the tricarboxylic
(Krebs) acid cycle, utilised in lipid synthesis or partially
oxidised to ketone bodies.
HSL _
triglyceride
+
NEFA
Lipolysis + re-esterification
glycerol
glycerol 3-
phosphate
+
Glucose
+ insulin stimulates
insulin inhibits
NEFA = non-esterified fatty acid
_ HSL = Hormone sensitive lipase
Figure 1.1 Effects of insulin on mobilisation of fatty acids from adipocytes
10 DIABETIC KETOACIDOSIS IN ADULTS
Acetoacetate is in equilibrium with 3-hydroxybutyrate, the balance
being dictated by the hepatic redox state. In ketoacidosis the
plasma ratio of 3-hydroxybutyrate to acetoacetate ratio is typically
elevated to about 3 : 1.
acetoacetate þ NADH þ H
þ ! 3-hydroxybutyrate
dehydrogenase
þNAD
Acetone is formed by the spontaneous decarboxylation of acetoacetate:
acetoacetate ! acetone þ CO2
Elevated acetone concentrations in ketoacidosis do not contribute
to the metabolic acidosis. Acetone is highly fat soluble and is
slowly excreted through the lungs.
Mitochondrial membrane
Inner Outer
CPT-1 CPT-2
CoASH CoASH
ACS
acyl-CoA acyl-carnitine
carnitine
INTRAMITOCHONDRIAL
SPACE
CPT-1 and 2 = carnitine–palmitoyl transferase
ACS = acyl-coenzyme A synthase
NEFA = non-esterified fatty acid
Acyl-CoA = acyl-coenzyme A
CoASH = coenzyme A
Modified with permission from Frayn KN. Metabolic regulation. 2nd ed.
Oxford, Blackwell Publishing, 2003.
carnitine
NEFA acyl-CoA
CYTOSOL
Figure 1.2 Fatty acid transport across the mitochondrial membrane
PATHOGENESIS 11
Ketone body disposal
With the exception of the liver, which lacks 3-oxo-acid CoA
transferase, most tissues have the capacity to utilise ketone bodies.
During treatment of ketoacidosis, oxidation of ketone anions
gradually neutralises the acidosis through the generation of
equimolar quantities of bicarbonate ions. Ketone body excretion
via the kidney and lung are important modes of elimination in
ketoacidosis.
Fluid and electrolyte depletion
Dehydration and electrolyte losses are prominent features of
diabetic ketoacidosis.
Water. When the renal threshold for glucose re-absorption in
the proximal convoluted tubule is exceeded, the resulting
osmotic diuresis leads to dehydration and secondary losses of
electrolytes (Table 1.1). Ketonuria compounds the loss of both
water and electrolytes.
Sodium. Insulin deficiency and glucagon excess exacerbate
sodium depletion via effects on renal sodium reabsorption.
Hyperventilation, fever and sweating due to infection may
Table 1.1 Average deficits of electrolytes in
adults with diabetic ketoacidosis
Sodium 500 mmol
Chloride 350 mmol
Potassium 300–1000 mmol
Calcium 50–100 mmol
Phosphate 50–100 mmol
Magnesium 25–50 mmol
12 DIABETIC KETOACIDOSIS IN ADULTS
further exacerbate fluid and electrolyte depletion, resulting in
average losses of body water in adults of approximately 5 L.
Increasing plasma osmolality leads to intracellular dehydration.
Reduced renal blood flow resulting from extracellular dehydration
impairs a major route of elimination of glucose and ketone
bodies; adequate correction of dehydration is therefore important
during the early phase of therapy.
Potassium. Metabolic acidosis leads to displacement of intracellular
potassium by hydrogen ions; these are subsequently lost in
urine or vomit. Breakdown of cellular protein secondary to
insulin deficiency compounds the loss of intracellular potassium.
However, despite a considerable total body potassium
deficit, serum potassium is usually normal or high at presentation
of diabetic ketoacidosis. Acidosis, insulin deficiency and
renal impairment all contribute to hyperkalaemia. Hypokalaemia
at presentation signifies a marked deficiency of body
potassium that in some patients may be enhanced by antecedent
diuretic therapy.
Phosphate. Total body phosphate deficiency is common in
ketoacidosis and may be exacerbated by co-existing conditions
such as chronic alcoholism. Insulin therapy stimulates cellular
uptake of phosphate and a variable degree of hypophosphataemia
is common during the treatment phase of diabetic
ketoacidosis. Phosphate deficiency is associated with reduced
red cell 2,3-diphosphoglycerate levels, resulting in reduced
oxygen delivery to the tissues. However, the adverse effects on
the oxyhaemoglobin dissociation curve are offset by the acidaemia
of ketoacidosis (the Bo¨hr effect). The benefits of
phosphate supplements on the course and prognosis of diabetic
ketoacidosis have not been substantiated in clinical trials. Large
sample sizes are required to demonstrate statistically significant
clinical benefits in diabetic ketoacidosis. Serum phosphate levels
may be elevated despite a total body deficit. Some authorities
recommend replacement of phosphate if levels fall below
an arbitrary level, e.g. 0.5 mmol/L, although this is not
FLUID AND ELECTROLYTE DEPLETION 13
common practice in the UK. A degree of caution is necessary as
iatrogenic hypocalcaemia may complicate phosphate replacement.
Magnesium. While the clinical significance of hypomagnesaemia
that commonly accompanies diabetic ketoacidosis is also
uncertain, it may exacerbate potassium deficiency.
The development of diabetic ketoacidosis is summarised in
Figure 2.1.
Clinical features
The cardinal symptoms of ketoacidosis include
rapidly increasing polyuria and polydipsia
rapid weight loss – dehydration
nausea and vomiting – hyperketonaemia is emetic
generalised muscular weakness
muscular cramps.
These are followed by serious signs of cerebral dysfunction:
progressive drowsiness and obtundation
coma.
While a decrease in the level of consciousness is common, coma is
encountered in only about 10 per cent of patients. The mechanism
by which ketoacidosis induces coma remains uncertain; impairment
of consciousness correlates with plasma glucose concentration
and osmolarity, coma at presentation being associated with a
worse prognosis. Co-existing causes of coma such as stroke, head
injury or drug overdose should be considered and excluded if
14 DIABETIC KETOACIDOSIS IN ADULTS
serum osmolality is less than approximately 350 mOsmol/kg
(Table 1.2).
Usually symptoms usually require several hours to develop,
often following symptoms of an intercurrent illness. Physical signs
are usually prominent in severe diabetic ketoacidosis.
Dehydration. Variable; approximately 5 L in an average adult.
Hypotension. Supine hypotension, in the absence of confounding
effects of anti-hypertensive drugs, denotes more than 20 per
cent depletion of extracellular fluid volume. Severe hypotension
in ketoacidosis carries an adverse prognosis
Tachycardia. Reflects dehydration, acidosis and sympathetic
activation; drugs with anti-muscarinic effects, e.g. tricyclic
antidepressants used for treatment of symptomatic neuropathy,
may exacerbate tachycardia.
Severe diabetic ketoacidosis may develop within 24 hours.
Table 1.2 Causes of impaired consciousness in patients with
diabetes mellitus
Diabetic ketoacidosis
Hyperosmolar non-ketotic hyperglycaemia
Hypoglycaemia
Lactic acidosis
Other causes:
Stroke (more common in diabetic patients)
Post-ictal (including hypoglycaemia – generalised tonic–clonic
convulsions also cause a self-correcting lactic acidosis; see
Chapter 6)
Cerebral trauma (may follow hypoglycaemia)
Ethanol intoxication (may induce or exacerbate hypoglycaemia in
diabetic patients)
Drug overdose
CLINICAL FEATURES 15
Other clinical and biochemical features include the following.
Air hunger. Acidosis stimulates the respiratory centre within the
medulla oblongata, causing deep rapid respirations (Kussmaul
breathing).
Ketotic fetor. The odour of acetone may be obvious on the breath,
although the capacity to detect acetone varies between individuals.
Hypothermia. Another consequence of acidosis, which may
mask a valuable sign of infection. Rectal temperature should
be taken with a low reading thermometer if hypothermia is
suspected; marked hypothermia carries an adverse prognosis.
Leukocytosis. This is common with hyperketonaemia and does
not necessarily indicate infection.
Gastroparesis. A gastric succusion splash may be evident on
abdominal examination as a consequence of gastric stasis; the
stomach may become distended with several litres of contents,
posing a risk of aspiration in patients with an impaired level of
consciousness.
Abdominal pain. Generalised abdominal pain may occur, particularly
in younger patients with severe acidosis (see Chapter 2).
If abdominal pain does not resolve with resolution of the
acidosis, alternative causes should be suspected. Measurement
of plasma amylase is unhelpful since levels may be raised nonspecifically
in ketoacidosis; ultrasound imaging of the pancreas
may be of assistance in diagnosing pancreatitis.
Diagnosis
Delays in initiating therapy may have disastrous consequences.
Diabetic ketoacidosis should be considered in any unconscious or
16 DIABETIC KETOACIDOSIS IN ADULTS
hyperventilating patient. If there is any doubt about the severity of
the metabolic disturbance in a diabetic patient with ketosis, the
arterial pH should be measured. A brief clinical examination
focuses on
airway protection
cardio-pulmonary status
level of consciousness
precipitating causes.
Bedside blood and urine tests should rapidly confirm the
diagnosis. Treatment should then be commenced without delay.
The initial clinical and biochemical assessment of a patient with
suspected diabetic ketoacidosis is shown in Table 1.3.
Urine. If available, should be tested for the presence of glucose
and, most importantly, for ketones. The presence of protein,
nitrites and leukocytes suggest infection.
Venous blood. Minimum urgent laboratory investigations include
measurement of
glucose (fluoride oxalate tube)
urea (blood urea nitrogen)
creatinine
sodium
potassium
full blood count with differential.
Some of the potential pitfalls in the diagnosis and management of
diabetic ketoacidosis are summarised in Table 1.4.
DIAGNOSIS 17
Table 1.3 Initial assessment of patients with suspected diabetic
ketoacidosis
Clinical history. Initially, brief and relevant (including previous
episodes of ketoacidosis and potential precipitating causes).
Physical examination. Rapid but thorough assessment for signs of
dehydration, level of consciousness, metabolic acidosis (Kussmaul
respiration), hypotension, hypothermia, gastric stasis and any
precipitating condition (e.g. pneumonia, pyelonephritis).
Biochemical assessment. Confirm diagnosis by bedside measurement
of
blood glucose (by glucose-oxidase reagent test strip)
urine ketones (‘Ketostix’).
Venous blood is withdrawn for laboratory measurement of
glucose
urea (BUN)
sodium
potassium
chloride (required for calculation of anion gap)
ketones.
In addition take blood for
full blood count
blood cultures (in all cases)
Inspect plasma for turbidity (hyperlipidaemia).
Capillary or arterial blood gases (corrected for hypothermia) for
pH
bicarbonate
pCO2
arterial pO2.
Repeat laboratory measurement of blood glucose, electrolytes, urea,
gases at 2 and 6 h.
Other investigations. Chest X-ray, culture of urine/sputum/faeces,
electrocardiograph, sickle cell test etc., as indicated.
BUN ¼ Blood urea nitrogen.
18 DIABETIC KETOACIDOSIS IN ADULTS
Practical points to consider include the following.
Glucose. Hyperglycaemia is readily determined on capillary or
venous blood using a glucose-oxidase reagent strip or blood gas
analyser, pending confirmation by the clinical chemistry
laboratory. Diabetic ketoacidosis presenting in the absence of
marked hyperglycaemia is recognised but uncommon; absence
of severe hyperglycaemia does not exclude ketoacidosis.
Ketones. Plasma ketone body concentration should be measured
(semi-quantitatively) with a nitroprusside-based reaction. These
tests are essentially specific for acetoacetate and do not
react with the principal ketoacid in diabetic ketoacidosis,
i.e. 3-hydroxybutyrate; acetone reacts weakly. Experimental
evidence in subjects with type 1 diabetes suggests that fasting
Table 1.4 Potential pitfalls in the diagnosis and management of diabetic
ketoacidosis
Odour of acetone on the patient’s breath. a useful sign but many people
cannot detect acetone
Fever. may be absent in the presence of infection (peripheral vasodilatation
causes cooling)
Blood leukocytosis. neutrophil count may be non-specifically raised
Plasma sodium concentration. may be falsely lowered initially by high lipid
and glucose levels and may appear to rise suddenly after insulin treatment
lowers plasma glucose and lipid levels
Plasma potassium concentration. may be temporarily raised (by acidosis)
despite severe total body potassium depletion
Plasma creatinine concentration. may be falsely elevated (assay interference
by ketone bodies)
‘Ketostix’ testing. may show ‘negative’ or ‘trace’ result when lactic acidosis
or alcoholic ketoacidosis coexist with diabetic ketacidosis (predominance
of 3-hydroxybutyrate). Ketostix reaction may become temporarily stronger
during treatment of diabetic ketoacidosis (conversion of 3-hydroxybutyrate
to acetoacetate)
Plasma transaminases and creatine phosphokinase. may be non-specifically
raised
DIAGNOSIS 19
results in ketoacidosis with smaller increments in blood glucose
concentration compared with the non-fasting state. However,
this should not be confused with mild starvation ketosis in the
non-diabetic subject. In the absence of diabetes, serum glucose
will be normal or slightly reduced by a severely reduced caloric
intake for several days. The resulting mobilisation of fatty acids
from adipocytes, a consequence of appropriately low insulin
concentrations, leads to ketonuria. The brain gradually increases
its utilisation of ketones as an alternative energy source and so
severe ketosis does not develop. This physiological adaptation
contrasts with the situation in diabetic ketoacidosis: the combination
of hyperglycaemia with significant ketosis points to
marked insulin deficiency.
In starvation ketosis, serum glucose is in the low–normal
range; the presence of ketosis with hyperglycaemia indicates
severe insulin deficiency.
A severe metabolic acidosis in the absence of hyperglycaemia, or
other obvious cause of acidosis such as renal failure, raises the
possibility of alternative diagnoses.
1. Alcoholic ketoacidosis. This is encountered in chronic alcoholics,
often following a binge, when carbohydrate intake is reduced
due to intractable vomiting from gastritis or acute pancreatitis.
Elevated circulating concentrations of counter-regulatory hormones
resulting from dehydration and the consequences of
acute alcohol withdrawal stimulate lipolysis and ketogenesis.
Hepatic metabolism of alcohol induces a more reduced
mitochondrial redox state; this increases the ratio of serum
3-hydroxybutyrate to acetoacetate to as high as 7–10 : 1,
compared with 3 : 1 in diabetic ketoacidosis. Under these
circumstances a negative or minor urine or plasma ketone
reaction may give a misleading impression of the degree of
the ketonaemia. Since hyperglycaemia is usually absent, treatment
of acute alcoholic ketoacidosis comprises rehydration
20 DIABETIC KETOACIDOSIS IN ADULTS
with intravenous dextrose and electrolyte replacement. This
condition may be under-diagnosed.
2. Lactic acidosis. A similar diagnostic caveat may occasionally be
encountered when significant lactic acidosis co-exists with
ketoacidosis (see Chapter 6). Causes of an anion gap acidosis
are shown in Table 1.5. The anion gap is elevated when serum
Na
þ ð½Cl
þ ½HCO3 Þ > 10 mmol=L
Potassium is not included in the calculation since the plasma level
of this ion may be altered significantly by acid–base disturbances.
The normal anion gap of approximately 10 mmol/L is accounted
for by proteins, phosphate, sulphate and lactate ions. When the
anion gap is increased, measurement of the plasma concentration
of specific anions, e.g. ketone bodies, lactate, may confirm the
aetiology of the acidosis. Although diabetic ketoacidosis usually
presents as an anion gap acidosis, typically 25–35 mmol/L, a
variety of acid–base disturbances have been reported.
Sodium. Despite a proportionally greater loss of body water,
plasma sodium concentrations are usually normal or low, although
plasma electrolyte concentrations may be falsely depressed by
grossly elevated plasma lipid concentrations in diabetic
ketoacidosis. Plasma should therefore be inspected for turbidity.
Table 1.5 Causes of anion gap acidosis
Ketoacidosis
Diabetic ketoacidosis
Alcoholic ketoacidosis
Lactic acidosis
Chronic renal failure
Drug toxicity
Methanol (metabolised to formic acid)
Ethylene glycol (metabolised to oxalic acid)
Salicylate poisoning
DIAGNOSIS 21
Eruptive xanthomata and lipaemia retinalis are recognised complications
that usually respond rapidly to treatment of the ketoacidosis.
Creatinine. If measured using a colorimetric method, the serum
creatinine concentration may be falsely elevated due to assay
interference by acetoacetate; this may lead to an erroneous
diagnosis of renal failure. Measurements in most modern
laboratories will not be affected.
Enzymes. Serum transaminases and creatine phosphokinase
may be non-specifically elevated in diabetic ketoacidosis and
may be mistaken for evidence of acute myocardial infarction.
Arterial blood gases. The acidosis is quantified by measurement
of pH, pCO2 and bicarbonate concentration. Some gas analysers
measure lactate, glucose and electrolytes.
Other tests. Bacteriological culture of urine, sputum and blood
is mandatory; broad-spectrum antibiotics should be administered
promptly if infection is suspected. Testing for sickle cell
and glucose-6-phosphate dehydrogenase deficiency may be
indicated in selected patients.
Treatment
Investigations should not delay the initiation of treatment or
transfer to a high-dependency or intensive care unit.
Aims of therapy
Treatment comprises rehydration with intravenous fluids, the
administration of insulin and replacement of electrolytes. The
treatment of ketoacidosis in children in considered in Chapter 2.
The importance of general medical care and close supervision by
trained medical and nursing staff deserves emphasis. A treatment
22 DIABETIC KETOACIDOSIS IN ADULTS
flow-chart should be used (see Chapter 3) and updated meticulously.
Accurate recording of fluid balance may necessitate a
urinary catheter if no urine is passed in the first 4 h or so. An initial
treatment plan for diabetic ketoacidosis in adults is shown in Table 1.6.
Table 1.6 Guide to treatment of diabetic ketoacidosis
Fluids and electrolytes
Volumes
1 L/h 2–3, thereafter adjusted according to requirements
Fluids
isotonic saline (0.9%) generally
Hypotonic (0.45%) if serum sodium exceeds 150 mmol/L (no more than
1–2 L – consider 5% dextrose with increased insulin if marked
hypernatraemia)
5% dextrose 1 L every 4–6 h when blood glucose has fallen to 15 mmol/L
(severely dehydrated patients may require simultaneous saline infusion)
Sodium bicarbonate
700 mL of 1.26% or 100 mL of 8.4% (if large vein cannulated) if pH < 7.0
(with extra potassium)
Potassium
No potassium in first 1 L of fluid unless initial plasma potassium < 3.5
mmol/L
Thereafter, add dosages below to each 1 L of fluid. If plasma K
þ
<3.5 mmol/L, add 40 mmol KCl (severe hypokalaemia may require
more aggressive KCl replacement)
3.5–5.5. mmol/L, add 20 mmol KCl
>5.5 mmol/L, add no KCl.
Insulin
Continuous intravenous infusion
5–10 U/h (average 6 U/h) initially until blood glucose has fallen to 15
mmol/L. Thereafter, adjust rate (1–4 U/h usually) during dextrose infusion
(continues overleaf)
TREATMENT 23
Table 1.6 (continued )
to maintain blood glucose 6–11 mmol/L until patient is eating again.
Measure, and record, capillary glucose hourly.
Capillary blood glucose (mmol/L) Soluble insulin infusion rate
0–3.9* 0 see note below*
4–6.9 1 unit per hour
7–9.9 2 units per hour
10–14.9 3 units per hour
Other measures
Search for and treat precipitating cause, e.g. infection.
Hypotension usually responds to adequate fluid replacement.
Central venous pressure monitoring in elderly patients or if cardiac disease
present.
Pass nasogastric tube – with airway protection – if conscious level
impaired.
Pass urinary catheter if conscious level impaired or no urine passed within
4 h of start of therapy.
Continuous electrocardiographic monitoring may warn of hyper- or
hypokalaemia (potassium should be measured at 0, 2 and 6 h – and more
often if indicated by levels outside target range).
Adult respiratory distress syndrome – mechanical ventilation (100% O2,
postive pressure ventilation); avoid fluid overload.
Mannitol (up to 1 g/kg intravenously) if cerebral oedema suspected.
Parenteral dexamethasone as alternative; N.B. induces insulin resistance.
Consider cranial CT scan to exclude alternative pathology (e.g. cerebral
haemorrhage, venous sinus thrombosis).
Treat thrombo-embolic complications if they occur.
Meticulous clinical and biochemical record using a purpose-designed flowchart.
*Note: Intravenous insulin should not be interrupted if at all possible. However, errors
leading to significant hypoglycaemia, e.g. inadvertent interruption of i.v. dextrose, may
necessitate temporary cessation of insulin while corrective action is taken, e.g. increasing
the dextrose infusion rate and/or bolus of 20–30 mL of 50% dextrose into a large vein if
symptomatic (see Chapter 4). Aim to restart i.v. insulin within 15–30 min, at a reduced
rate if indicated, and/or with a higher rate of dextrose infusion; consider 10% dextrose.
Since interruption of i.v. insulin risks relapse of ketoacidosis, some clinicians advocate
continuing insulin at a reduced rate while correcting lesser degrees of hypoglycaemia
with i.v. dextrose. Careful monitoring with attention to infusion apparatus and hourly
checks on the volumes infused will help to minimise the risk of hypoglycaemia.
24 DIABETIC KETOACIDOSIS IN ADULTS
Correction of fluid and electrolyte depletion
Rehydration. Adequate rehydration is an important aspect of
treatment that contributes directly to reductions in hyperglycaemia
and counter-regulatory hormone levels. Considerable
variation in fluid and electrolyte disturbances are observed
between patients and the following recommendations represent
only a guide to therapy.
Rehydration is commenced with isotonic (0.9 per cent,
containing 150 mmol each of Na
þ
and Cl

) saline containing
appropriate potassium supplements (see below). Isotonic
saline is used in preference to hypotonic saline – unless
plasma osmolarity is significantly raised – in order to
minimise the rapid movement of extracellular water into
cells as blood glucose and osmolarity fall with treatment;
such shifts have been implicated in the pathogenesis of the
serious complication of cerebral oedema, discussed below.
Rehydration of the patient must take account of continuing
polyuria and approximately 6–10 L of fluid may be required
during the first 24 h.
In an average adult, 1 L of saline is infused every hour for the
first two to three hours. The rate of infusion is then adjusted
according to the clinical state of the patient. Care is required
in elderly patients or those with cardiac disease; monitoring
of central venous pressure or pulmonary wedge pressure is
recommended in these circumstances.
Occasionally, patients with relatively low admission plasma
glucose concentrations may require a simultaneous infusion
of dextrose to allow administration of sufficient insulin to
suppress lipolysis and ketogenesis without inducing hypoglycaemia.
A rising serum sodium concentration (above 150 mmol/L)
may necessitate the temporary substitution of hypotonic
TREATMENT 25
saline (75 mmol/L each of Na
þ
and Cl

) or even 5 per cent
dextrose (with an appropriate increase in the dose of insulin
if dextrose is used).
When plasma glucose has fallen to 15 mmol/L, saline
is discontinued and replaced immediately by 5 per cent
dextrose, usually at a rate of around 250 mL/h. Undue delay
in commencing dextrose infusion at this point may result in
hypoglycaemia. Intravenous dextrose is given without
interruption until the patient is eating again, since intravenous
insulin must be continued (albeit usually at a lower
dose – see below).
Although the use of hypertonic (10 per cent) dextrose at this
stage of treatment produces a slightly faster fall in total
ketone bodies, this is not reflected in a more rapid resolution
of the acidosis.
Potassium replacement. Cardiac arrhythmias induced by iatrogenic
hypokalaemia represent a major and avoidable cause of
death. Hypokalaemia may also induce life-threatening weakness
of respiratory muscles. Potassium is predominantly (98 per
cent) an intracellular ion. Insulin treatment and rising pH
stimulate the entry of extracellular potassium into cells.
On average 20 mmol of potassium (administered as 1.5 g
potassium chloride) will be required in each litre of fluid
following the start of insulin therapy. Continuous electrocardiographic
monitoring may indicate signs of hyper- or
hypokalaemia, but plasma potassium concentration should
be checked regularly (2 hourly at first) and potassium
supplements adjusted appropriately.
Particular care must be exercised in patients with renal
failure, anuria or oliguria (less than 40 mL/h).
If hypokalaemia is present (plasma potassium < 3.5 mmol/L)
potassium supplements should be doubled to 40 mmol/h; if
26 DIABETIC KETOACIDOSIS IN ADULTS
hyperkalaemia develops potassium should be temporarily
halted, pending the result of further measurements.
Insulin therapy
The successive aims of insulin treatment in ketoacidosis are
inhibition of lipolysis and hence ketogenesis
suppression of hepatic glucose production
enhanced disposal of glucose and ketone bodies by peripheral
tissues.
Soluble (unmodified) insulin only has a plasma half-life of
approximately 5 min, so intermittent i.v. injections lead to
unpredictable and fluctuating plasma insulin concentrations.
Maximal stimulation of potassium transport into cells occurs
with pharmacological plasma insulin concentrations and the risk
of hypokalaemia is therefore greater with large doses of insulin.
With modern insulin regimens, complications of treatment such as
hypokalaemia and late hypoglycaemia are less common than with
the obsolete high-dose intermittent bolus regimens.
Intravenous insulin. Soluble insulin is administered as a
continuous intravenous infusion at a rate of 5–10 (usually 6)
U/h. This produces steady plasma insulin concentrations in the
high physiological (or pharmacological at the higher rates)
range that adequately suppress lipolysis, ketogenesis and
hepatic glucose production, even in the presence of elevated
levels of catabolic hormones. Insulin is diluted to a convenient
concentration (usually 1 U/mL) with isotonic saline in a 50 mL
syringe and delivered by a syringe-driver infusion pump
connected via a Y connector. The infusion apparatus should
be flushed through before connection to the patient. Alternatively,
insulin can be added to a 500 mL bag of isotonic saline
TREATMENT 27
and mixed gently; the insulin must be injected using a needle of
sufficient length to clear the injection port of the bag.
Monitoring response. Capillary blood glucose is checked at
the bedside at hourly intervals and the infusion rate is reduced
to 1–3 U/h, when blood glucose has fallen to 15 mmol/L. The
infusion rate should be adjusted to maintain euglycaemia until
the patient is eating again and subcutaneous insulin is
recommenced (Table 1.6). The rate required at this stage will
vary according to (1) the degree of insulin resistance (see above)
and (2) the rate of dextrose infusion. Intravenous insulin at
6 U/h should produce a steady and predictable fall in plasma
glucose concentrations, averaging 4–6 mmol/h in adults. The
commonest causes of failure to respond are mechanical, i.e.
problems such as the pump being inadvertently switched off or
set at the wrong rate and blockage of the delivery line. It is
sound practice to cross-check (and record on the flow-chart) the
prescribed rate of insulin delivery against the volume infused
each hour during treatment. During treatment of ketoacidosis
there is conversion of 3-hydroxybutyrate to acetoacetate.
Nitroprusside-based tests may therefore give the mistaken
impression that ketosis is either not resolving or is worsening.
A rising plasma bicarbonate will allay such fears.
Transfer to subcutaneous insulin. The first subcutaneous injection
should comprise or include a dose of short- or rapid-acting
insulin. This should be administered 60 min before the i.v.
insulin infusion is terminated to allow time for absorption
of insulin from the subcutaneous depot.
Bicarbonate therapy
The role of bicarbonate in the management of diabetic ketoacidosis
remains controversial. No large clinical trials in severely acidotic
28 DIABETIC KETOACIDOSIS IN ADULTS
patients with diabetic ketoacidosis have been performed. However,
blood pH levels < 7.0 may lead to life-threatening cardiorespiratory
complications. Small doses of bicarbonate (approximately
100 mmol) may be beneficial if the patient is severely acidotic or if
cardio-respiratory collapse appears imminent. However, it is
possible that administration of bicarbonate into the extracellular
space may actually exacerbate intracellular acidosis. Bicarbonate
ions (which cannot diffuse across cell membranes) combine with
H
þ
ions extra-cellularly, producing carbonic acid, which dissociates
into water and CO2. The latter readily enters cells, where the
reverse reaction occurs, generating H
þ
(and bicarbonate ions)
intracellularly. The solution of 8.4 per cent sodium bicarbonate is
hypertonic and extremely irritant and should only be infused into a
large (ideally central) vein; extravasated solution often causes
extensive local necrosis. Bicarbonate should therefore be infused as
an isotonic solution, i.e. 700 mL of 1.26 per cent solution (12.6 g,
each litre containing 150 mmol each of Na
þ
and HCO
3 ) given over
30 min and repeated if necessary to raise the pH to 7.0–7.2. Other
reservations about the use of bicarbonate include the following.
Hypokalaemia. This may be exploited in the treatment of severe
hyperkalaemia. Otherwise, extra potassium (20 mmol potassium
per 100 mmol bicarbonate) should be administered when
bicarbonate is infused and plasma potassium concentration
should be re-checked shortly afterwards.
Paradoxical acidosis of cerebrospinal fluid. The clinical significance
of this complication is uncertain.
Tissue hypoxia. Bicarbonate may have adverse effects on the
oxyhaemoglobin dissociation curve.
Overshoot alkalosis. Complete correction of the acidosis should
not be the objective, since concurrent metabolism of ketone
anions may lead to over-alkalinisation.
TREATMENT 29
Acceleration of ketogenesis. In a controlled clinical study, the fall in
ketone body and lactate concentrations was delayed in patients
with diabetic ketoacidosis who received 150 mmol of sodium
bicarbonate compared with saline.
Cerebral oedema. Treatment with bicarbonate has also been
linked with the development of cerebral oedema in children
with diabetic ketoacidosis (see Chapter 2).
Complications of diabetic
ketoacidosis in adults
Aspiration
The stomach of a patient with diabetic ketoacidosis may contain
1–2 L of fluid, and where consciousness is impaired there is a
possibility of vomiting with inhalation of gastric contents. Nausea
or vomiting in a patient who is semi-conscious should lead to
insertion of a naso-gastric tube under controlled conditions, i.e.
with intubation if necessary.
Adult respiratory distress syndrome
Adult respiratory distress syndrome has been reported occasionally
in patients with ketoacidosis, usually in patients under
50 years. Clinical features include dyspnoea, tachypnoea, central
cyanosis and non-specific chest signs. Arterial hypoxia is characteristic
and chest radiography reveals bilateral pulmonary infiltrates.
Management involves respiratory support with intermittent
positive pressure ventilation and avoidance of fluid overload.
30 DIABETIC KETOACIDOSIS IN ADULTS
Thromboembolism
Thromboembolic complications are important causes of mortality
in patients with hyperglycaemic emergencies arising as a consequence
of dehydration, increased blood viscosity and increased
coaguability. Disseminated intravascular coagulation has also been
reported as a rare complication of diabetic ketoacidosis. The role of
prophylactic anticoagulation has not been clearly established and
routine anticoagulation is not recommended in view of the risks of
haemorrhage. Clinically evident thromboembolism is treated
conventionally.
Rhinocerebral mucormycosis
This aggressive opportunistic fungal infections occasionally develops
in patients with diabetic ketoacidosis or other metabolic
acidoses. The lesion arises in the paransal sinuses and rapidly
invades adjacent tissues (nose, sinuses, orbit and brain). Treatment
comprises correction of acidosis, wide surgical excision of affected
tissue condition and parenteral anti-mycotic agents. The course is
often fulminant; the condition carries a high mortality.
Further reading
Adrogue HJ, Wilson H, Boyd AE et al. Plasma acid–base patterns in
diabetic ketoacidosis. N Engl J Med 1982; 307: 1603–1610.
Alberti KGMM. Low-dose insulin in the treatment of diabetic ketoacidosis.
Arch Intern Med 1977; 137: 1367–1376.
Barrett EJ, DeFronzo RA, Bevilacqua S and Ferrannini E. Insulin resistance
in diabetic ketoacidosis. Diabetes 1982; 31: 923–928.
FURTHER READING 31
Kitabchi AE, Umpierrez GE, Murphy MB, Barrett EJ, Kreisberg RA,
Malone JI et al. Management of hyperglycemic crises in patients with
diabetes. Diabetes Care 2001; 24(1): 131–153.
Kraut JA and Kurtz I. Use of base in the treatment of severe acidemic
states. Am J Kid Dis 2001; 38: 703–727.
Krentz AJ and Nattrass M. Acute metabolic complications of diabetes. In:
Pickup JC, Williams G (Eds). Textbook of Diabetes, 3rd ed. Oxford.
Blackwell 2003 pp. 32.1–24.
McGarry JD and Foster DW. Regulation of hepatic fatty acid oxidation and
ketone body production. Annu Rev Biochem 1980; 49: 395–420.
Miles JM, Rizza RA, Haymond MW and Gerich JE. Effects of acute insulin
deficiency on glucose and ketone body turnover in man. Diabetes 1980;
29: 926–930.
Nattrass M and Hale PJ. Clinical aspects of diabetic ketoacidosis. In:
Nattrass M, Santiago JV (Eds). Recent Advances in Diabetes, 1st ed.
Edinburgh. Churchill Livingstone 1984 pp. 231–238.
Schade DS, Eaton RP, Alberti KGMM and Johnston DG. Diabetic Coma,
Ketoacidotic and Hyperosmolar. Albuquerque, NM. University of New
Mexico Press 1981.
32 DIABETIC KETOACIDOSIS IN ADULTS
2
Diabetic Ketoacidosis
in Childhood
Julie A Edge and David B Dunger
Summary
Approximately 25 per cent of children present in diabetic
ketoacidosis at diagnosis of type 1 diabetes, and this remains
a life-threatening condition.
Guidelines for treatment are necessary, although they must
always be tailored to the individual. Resuscitation is the primary
objective. This includes prevention of aspiration of gastric
contents using a nasogastric tube and adequate – but not
excessive – replacement of circulating volume.
Further management comprises replacement of fluids with 0.9
per cent saline, replacement of potassium losses and the
institution of insulin using a continuous intravenous infusion.
Recovery is usually straightforward, but there is still a significant
mortality and morbidity, largely arising from the unpredictable
Emergencies in Diabetes Edited by Andrew J. Krentz
# 2004 John Wiley & Sons, Ltd ISBN 0-471-49814-9
complication of cerebral oedema. The pathophysiology of this
devastating condition is still unknown, and there is an ongoing
debate as to whether it is related to the treatment received. It is
prudent to ensure that changes in osmolality do not occur too
quickly, and that rehydration is not excessive. However, until
the cause of cerebral oedema is known, no guidelines can be
considered completely infallible. Close supervision from senior
members of staff is essential, and there should be early concern
if progress is not as predicted.
Early assessment of the best place to nurse the child should be
made, and clear instructions given to nursing staff for frequent
monitoring of vital signs and neurological observations. Headache
and behaviour change should be reported at any time to
medical staff. Rapid intervention with intravenous mannitol and
immediate transfer to an intensive care unit for assisted
hyperventilation and additional support is necessary if signs of
cerebral oedema develop.
Introduction
Diabetic ketoacidosis in childhood remains a serious and lifethreatening
condition despite improvements in the management of
the fluid and electrolyte disturbances over the last few decades.
Although increased awareness among primary health care teams
should have led to earlier diagnosis, around a quarter of children
still present in diabetic ketoacidosis at the diagnosis of type 1
diabetes; the proportion is even higher in very young children in
whom diabetes is not recognised early. Forty per cent of children
diagnosed under 4 years of age in an Oxford Regional cohort in
1990 had ketoacidosis at diagnosis and in 25 per cent it was severe,
i.e. arterial pH < 7.1. After the initial diagnosis of type 1 diabetes,
admission rates with diabetic ketoacidosis during childhood are
around 0.2 per patient year. Episodes are usually associated with
intercurrent illness. In the older child, insulin omission and
34 DIABETIC KETOACIDOSIS IN CHILDHOOD
psychological disturbance may result in recurrent hospital admissions.
Sadly, ketoacidosis is still the major cause of death in children
with type 1 diabetes. Approximately one per cent of episodes
among children may be complicated by cerebral oedema, which is
the major cause of type 1 diabetes-related death in children under
the age of 12 years. Ketoacidosis is the most common cause of
death outside hospital in teenagers and young adults with type 1
diabetes.
Diabetic ketoacidosis is the leading cause of death among
children with type 1 diabetes.
Definition
As for adults, the definition of diabetic ketoacidosis is arbitrary. In
practice, the term refers to decompensated type 1 diabetes resulting
in hyperglycaemia, and a metabolic acidosis attributable to
hyperketonaemia. Blood glucose is generally raised, but in some
three per cent of cases it may be less than 15 mmol/L.
Pathophysiology
As discussed in Chapter 1, the primary cause of ketoacidosis is an
absolute or relative insulin deficiency. Briefly, the effects of insulin
deficiency and thus an increase in glucagon/insulin ratio in the
portal circulation together with increases in levels of counterregulatory
hormones (catecholamines, cortisol and growth
hormone) are summarised in Figure 2.1. Elevated levels of ketone
bodies result from mobilisation of fatty acids from adipose tissues
and their preferential b-oxidation within the hepatic mitochondria;
the finite capacity of peripheral tissues to utilise ketone bodies
PATHOPHYSIOLOGY 35
contributes to the hyperketonaemia. Most of the acidaemia in
diabetic ketoacidosis is accounted for by the production and
dissociation of organic ketoacids, but lactic acidosis from tissue
hypoperfusion may also contribute (see Chapter 6).
The regulation of ketosis during childhood differs in some
respects from that observed in adults. In normal children,
circulating levels of ketone bodies are relatively high prior to
puberty. The finding of intermittent ketonuria is not unusual in
non-diabetic young children, particularly after an overnight fast.
During puberty, as plasma insulin levels increase, fasting ketone
Glucagon
Growth hormone
Cortisol
Catecholamines
↓ Peripheral use ↑ Hepatic ↑ Protein Lipolysis
of glucose gluconeogenesis catabolism
Hyperglycaemia ↑ Amino ↑NEFA release
acids from adipose
tissue
Glycosuria
Osmotic diuresis Ketosis
(loss of H2O, sodium, chloride, potassium, phosphate)
Acidosis
Dehydration
Hypovolaemia Coma Hyperpnoea Nausea
Insulin deficiency
Increased counterregulatory hormones
Figure 2.1 Schema of the pathophysiology of diabetic ketoacidosis.
Adapted with permission from Lyen KR, Hale D, Baker L. Endocrine
emergencies. In: Fleisher G, Ludwig S (Eds). Textbook of Pediatric Emergency
Medicine. Baltimore, MD: Williams and Wilkins # 1983. NEFA ¼ nonesterified
fatty acids
36 DIABETIC KETOACIDOSIS IN CHILDHOOD
levels tend to fall. However, in adolescents with type1 diabetes,
even when blood glucose levels are maintained at 5 mmol/L
overnight using an intravenous insulin infusion, there is still a
substantial increase in nocturnal ketogenesis, which appears to be
due mainly to excessive growth hormone secretion. This may help
explain the rapid decompensation that can occur in teenagers
overnight following either a short episode of vomiting or the omission
of bedtime insulin. Teenage girls in particular may present with
severe acidosis despite only modest elevations of blood glucose.
Children develop ketosis more rapidly than adults.
Diabetic ketoacidosis is associated with severe losses of body
fluids. The water deficit is made up of varying combinations from
the osmotic diuresis, vomiting, hyperventilation and, when
present, pyrexia. Sodium losses are also variable, depending on
the predominating route of fluid loss, the duration of polyuria and
the adequacy of renal perfusion. There is always total body
depletion of potassium and phosphate, even though plasma levels
of these ions may be low, normal or high. There will also usually
have been some attempt by the child or its parent to correct the
fluid losses with increased oral consumption of fluids, which can
affect the blood biochemistry at presentation. Much of the
information concerning the specific fluid and electrolyte deficits
in diabetic ketoacidosis has been obtained from experiments
carried out in adults in the 1930s, studies that have not been
repeated in children. Thus there is little direct information
concerning electrolyte and fluid losses in children; this may explain
some of the historical debate concerning optimal management.
Differential diagnosis
The diagnosis is rarely difficult except in younger children, where
the acidotic breathing pattern may easily be confused with an
DIFFERENTIAL DIAGNOSIS 37
upper or lower respiratory tract infection. In a small child, the
characteristic smell of ketones on the breath may be attributed
to fasting, but the diagnosis of diabetic ketoacidosis must be
considered in any sick child, as the consequences of missing
the diagnosis can be devastating. Further differential diagnoses are
considered in Table 2.1.
Hyperosmolar non-ketotic hyperglycaemic coma (see Chapter 3)
is rare in children; serum osmolality usually exceeds 300 mOsmol/L
and ketosis is absent. The mechanism is likely to be prolonged
dehydration with relative insulin sufficiency; it is more common in
children with mental impairment. The management is similar to
that of diabetic ketoacidosis, except that it has been recommended
that insulin and fluids are given more slowly to prevent too rapid a
fall in blood glucose and plasma osmolality.
Maturity onset diabetes of the young (MODY) is inherited in an
autosomal dominant fashion and does not usually present with
major metabolic decompensation. Insulin treatment is not usually
required but may be necessary during intercurrent illness. The
diagnosis should be firmly established using clinical and genetic
criteria before any decision is made to discontinue insulin therapy.
This also applies to the increasingly common situation wherein
type 2 diabetes is diagnosed in childhood or adolescence.
Table 2.1 Differential diagnosis of diabetic ketoacidosis in children
Glycosuria/hyperglycaemia Acidosis predominant
Physical stress, e.g. intercurrent Alcoholic ketoacidosis
infection with transient hyperglycaemia
MODY Severe sepsis
Renal tubular defects Certain inborn errors of
metabolism
Corticosteroid treatment; N.B. may
precipitate metabolic decompensation
Hyperosmolar non-ketotic coma
Type 2 diabetes
MODY ¼ maturity onset diabetes of the young.
38 DIABETIC KETOACIDOSIS IN CHILDHOOD
Management
Diabetic ketoacidosis can be a rapidly changing condition,
particularly in small children. Junior medical staff should routinely
discuss such children with senior colleagues and should feel
adequately supervised while undertaking treatment. Guidelines
that are easy to understand and to use should be readily available
in the hospital emergency department and on the wards, including
high-dependency and intensive care units. The basic principles of
management are as follows:
correct the fluid losses
reverse the acidosis and ketosis
prevent complications such as aspiration of gastric contents,
hypokalaemia, and cerebral oedema.
The guidelines that follow are based on those of the British Society
for Paediatric Endocrinology and Diabetes (BSPED). These can be
found in full on the BSPED website (www.bspe.shef.ac.uk), and a
short algorithm version on the Diabetes UK website (www.
diabetes.org.uk). These are very similar to the guidelines of the
International Society for Paediatric and Adolescent diabetes
(ISPAD; www.ispad.org). All of these are general guidelines for
management. Note that treatment may need to be varied to suit the
individual patient. Guidelines do not remove the need for frequent
detailed reassessments of the individual child’s progress. These
guidelines are intended for the management of the following
children:
more than five per cent dehydrated
and/or vomiting
and/or drowsy
and/or clinically acidotic.
MANAGEMENT 39
Children who are five per cent dehydrated or less and not clinically
unwell usually tolerate oral rehydration and subcutaneous insulin.
Discuss this with the senior doctor on call.
Resuscitation (ABC)
Airway and breathing
The first priority in the treatment of diabetic ketoacidosis, as in the
treatment of any life-threatening illness, is to protect and maintain
the airway. If the childs level of consciousness is impaired, a
nasogastric tube should be inserted immediately, aspirated and left
on free drainage. An oral airway may also be necessary. If
respiration is depressed, or there is accompanying respiratory
pathology, intubation and ventilation may be required; if in doubt,
this is the safest option. Tissue perfusion may be poor and, at least
until the first arterial blood gas results are known, supplemental
oxygen is generally administered by facemask.
Circulation
The next priority is to restore the circulating blood volume.
However, true shock (hypotension and tachycardia) is very rarely
present, and the degree of intravascular dehydration is often overestimated.
If there is truly reduced circulating volume, give 10 mL/kg
0.9 per cent saline over 10 min. This can be repeated to a maximum
of 20–30 mL/kg at this stage, titrated against changes in tissue
perfusion, or in the most severe cases against central venous pressure.
Clinical assessment
The clinical features of diabetic ketoacidosis are shown in Table 2.2.
40 DIABETIC KETOACIDOSIS IN CHILDHOOD
1. During the resuscitation phase a rapid clinical assessment
should have been made of the following.
Airway, breathing and circulation. See above.
Conscious level. If the conscious level is impaired, or there is
any change in neurological status during treatment, the
Glasgow Coma Score should be serially recorded, and
deteriorating conscious level treated as an emergency (see
section on cerebral oedema below).
Degree of dehydration. It has recently been recognised that the
degree of dehydration in children is often over-estimated by
clinical methods, and that signs such as loss of skin turgor or
elasticity occur at around three per cent dehydration, and
not at five per cent as is often quoted. Capillary refill time,
tested by applying digital pressure, may be a useful
technique for the assessment of dehydration in small
children, as long as they are not exposed to a cold
environment.
Weight of child. This is crucial to the fluid management,
therefore every effort should be made to weigh the child. If
Table 2.2 Clinical features of diabetic ketoacidosis in children
Symptoms Signs
Polyuria Lethargy
Thirst, polydipsia Dehydration
Rapid weight loss Blood pressure normal, rarely low
Abdominal pain Kussmaul respiration, or later depressed
Weakness Smell of ketones on breath
Vomiting Temperature normal
Air hunger 20% disordered consciousness
Confusion, coma 10% unconscious
MANAGEMENT 41
this is not possible because of the clinical condition, use the
most recent clinic weight as a guideline, or an estimated
weight from centile charts.
2. Full clinical assessment can be deferred until the child has been
resuscitated. Attending doctors and nurses should be aware of
the following.
Abdominal pain is a frequent accompaniment of diabetic
ketoacidosis in children; a surgical emergency should not be
assumed until a period of rehydration and insulin and
electrolyte replacement has been allowed.
Pyrexia is not a feature of uncomplicated ketoacidosis, and a
source of infection should be sought if it is present.
Laboratory assessment
Hyperglycaemia can be confirmed quickly by a high capillary
blood glucose measurement, but it is important to ensure that
reagent strips are fresh, that reflectance meters are well maintained
and that staff are trained in their use. Ketone measurements are
also possible on capillary blood, or urine if available. Treatment can
then be started while the results of further tests, such as plasma
electrolytes, are awaited. Suggested laboratory investigations are
listed in Table 2.3. Arterial or capillary blood should ideally be used
for acid–base assessment to confirm the acidosis, but where the
facilities are not available venous pH and bicarbonate may
be reasonable substitutes. Subsequently, acid–base status can be
monitored using venous or non-arterialised capillary blood gases,
since, although these will show a slightly higher CO2 and
bicarbonate, and slightly lower pH, the differences are not usually
clinically significant.
42 DIABETIC KETOACIDOSIS IN CHILDHOOD
Instructions for nursing staff
A decision should be made at an early stage as to where the child
should be nursed. If the child is very young, comatose or shocked,
or if ward staff are exceptionally busy or inexperienced, then
Table 2.3 Suggested laboratory investigations in diabetic ketoacidosis
Investigation Notes
Venous plasma glucose To confirm capillary result
Sodium, potassium, chloride, There may be artefactual lowering of
phosphate, calcium sodium due to hyperlipidaemia
Chloride will help to define type of
acidosis particularly during treatment
Plasma urea (BUN), creatinine Creatinine may be falsely elevated by
hyperketonaemia; not usually a
problem with modern chemical
pathology laboratories
Blood gases, pH, bicarbonate If oxygen saturation is available, a
venous or capillary sample is sufficiently
accurate, otherwise use arterial sample
Urinary ketones Bedside blood ketone measurement may
be helpful but insufficient evidence yet in
children
FBC, PCV PCV may support clinical evidence of
dehydration
Leucocytosis extremely common in
ketoacidosis and does not necessarily
imply infection
Urine culture If clinically indicated
Blood culture, CXR, Only if clinically indicated
throat swab
Plasma amylase If abdominal pain severe and continues
after adequate initial treatment
FBC ¼ full blood count; PCV ¼ packed cell volume; CXR ¼ chest X-ray; BUN ¼ blood
urea nitrogen.
MANAGEMENT 43
admission to an intensive care unit would be appropriate. Urinary
catheterisation may be helpful in the comatose larger child but is
not generally recommended.
Ensure full instructions are given to the senior nursing staff,
emphasising the need for
strict fluid balance of input and output, including oral fluids
and weighing of nappies
urine testing of every sample for ketones (may be superceded by
bedside blood ketone measurements in some centres)
hourly capillary blood glucose measurements
hourly or more frequent neurological observations initially
reporting immediately to the medical staff, even during the night,
symptoms of headache or any change in either conscious level or
behaviour, since these might indicate the development of
cerebral oedema
reporting any changes in the electrocardiograph trace, especially
T wave changes
recording body weight twice daily.
Fluids
1. Choice of intravenous fluid. After restoration of the circulating
volume, the main residual fluid deficit is within the intracellular
compartment. Historically, relatively hypotonic solutions were
recommended, but a failure of plasma sodium concentration to
rise during treatment has been linked with the development of
cerebral oedema (see below). Most authors would therefore
now recommend using isotonic (0.9 per cent) saline for the first
few hours before changing to a more hypotonic solution, such
as 0.45 per cent saline. 0.18 per cent saline is no longer
44 DIABETIC KETOACIDOSIS IN CHILDHOOD
recommended. In practice, the fluid type is usually changed
once the blood glucose concentration has fallen to around 10–
15 mmol/L, but if this occurs very early during treatment, there
is a possibility of giving too little sodium. In this situation,
isotonic saline should be continued simultaneously with sufficient
dextrose to avert hypoglycaemia during the latter stages of
treatment. During such combination therapy, care should be
taken to ensure that overall fluid volume is appropriate.
2. Volume of fluid. The volume of fluid to be replaced is based on
the clinical assessment of the deficit plus maintenance fluid
requirements, with the proviso that ongoing losses are also
replaced. Clinical assessment of the degree of dehydration
cannot be precise, and it is important not to over-estimate the
degree of dehydration. The rate at which the fluids are given is
an area of contention. It has been standard practice to give
initial fluid rapidly initially, and then slow down the rate.
However, since the rapid infusion of large volumes of fluid has
been proposed as one risk factor for the development of cerebral
oedema, once the circulating blood volume has been restored, it
is reasonable to correct the remaining fluid deficit slowly and
evenly over the next 24 or even 36 or 48 h. Furthermore, slow
rehydration has been used successfully in adults with diabetic
ketoacidosis and, paradoxically, it may lead to more rapid
restoration of acid/base balance.
3. Practicalities of fluid administration.
It is essential that all fluids administered are documented
carefully, particularly the fluid that is given in the initial
phase of therapy and during transfer to the ward; this is
where most errors occur.
By this stage, the circulating volume should have been
restored. If not, give a further 10 mL/kg 0.9 per cent saline
or 4.5 per cent albumin over 30 min.
MANAGEMENT 45
Otherwise, once circulating blood volume has been restored,
calculate fluid requirements as follows:
requirement ¼ maintenance þ deficit
½deficit ðlitresÞ ¼ %dehydration body weight ðkgÞ
To avoid over-zealous fluid replacement, which may be a
risk factor for cerebral oedema, never assess dehydration as
more than 10 per cent. Include the volume of fluid that may
have been given during resuscitation.
Age Maintenance values
0–2 yrs 80 mL/kg/24 h
3–5 70 mL/kg/24 h
6–9 60 mL/kg/24 h
10–14 50 mL/kg/24 h
adult (>15) 35 mL/kg/24 h
Add maintenance and deficit and give the total volume
evenly over the next 24 hours. i.e.
hourly rate ¼ maintenance þ deficit
24
Example:
A 20 kg 6 year old boy who is 10 per cent dehydrated will
require 10 per cent 20 kg ¼ 2000 mL deficit
plus 60 mL 20 kg ¼ 1200 mL maintenance
¼ 3200 mL over 24 hours ¼ 133 mL/h.
It may be preferable (although there is no definite evidence)
to lengthen the period of rehydration to 48 h in very young
children or those who are very hyperosmolar with high
plasma sodium levels or very high blood glucose levels.
Discuss this with a senior clinician.
Insulin
Although insulin resistance has been shown to be a feature of
diabetic ketoacidosis, in practice large doses of insulin are
46 DIABETIC KETOACIDOSIS IN CHILDHOOD
not required. A continuous low-dose intravenous infusion of
0.1 U/kg/h of soluble (unmodified) insulin is an effective and
simple method for reversing the metabolic acidosis, and is
associated with a lower incidence of hypoglycaemia and
hypokalaemia than higher doses. A bolus of insulin is not
necessary as large doses may cause a rapid reduction in blood
glucose which may be undesirable.
There are some who believe that younger children (especially
the under 5s) are particularly sensitive to insulin and therefore
require a lower dose of 0.05 U/kg/hour. There is no evidence to
support the lower dose, and only the larger dose has been
shown to correct hyperglycaemia and reverse ketosis.
Insulin should not be added to the replacement fluid bag, but
should be infused using a separate syringe pump, so that
adjustments to fluid and insulin can be made independently.
Potassium
Potassium is mainly an intracellular ion, and there is always
massive depletion of total body potassium in diabetic ketoacidosis,
although initial plasma levels may be low, normal or even
high. Circulating levels will fall once insulin therapy is
commenced.
Potassium replacement should therefore be started immediately
unless anuria is suspected or there are peaked T waves on the
electrocardiogram. The infusion should be altered according to
subsequent plasma electrolyte results to maintain plasma
potassium concentration within the normal range.
Add 20 mmol potassium chloride (KCl) to every 500 mL bag of
fluid if normokalaemia.
A cardiac monitor should be observed frequently for T wave
changes.
MANAGEMENT 47
Bicarbonate
As discussed in Chapter 1, this is rarely, if ever, indicated.
Profound acidosis can theoretically be a cause of poor myocardial
contractility, although there is no evidence to support this in
children. It has been our recent practice only to use bicarbonate at a
blood pH of less than 6.9, or if there is evidence of poor circulation
after adequate administration of resuscitation fluid. The commonest
reason for the failure of acidosis to resolve is inadequate
restoration of the circulating blood volume or late institution of
insulin therapy, which therefore fails to suppress ketogenesis.
Phosphate
Diabetic ketoacidosis is associated with severe phosphate depletion
due to excessive urinary losses, and once insulin treatment is
started levels will fall, because phosphate, like potassium, is taken
up by the cells. Although plasma concentrations may fall in adults to
levels known to have been associated with impaired cardiac
function, respiratory failure and reduced red cell 2,3-diphosphoglycerate
concentrations, these complications are rarely seen, and have
not been reported in children. There is no evidence that replacement
is beneficial; it may precipitate symptomatic hypocalcaemia.
Subsequent management
Check plasma urea and electrolytes 2 h after resuscitation has
begun and then at least 4 hourly during treatment.
Check fluid balance and the clinical state of the child at least
4 hourly, to ensure that positive fluid balance is maintained.
If a massive diuresis continues fluid input may need to be
increased; measurement of urinary electrolytes may be helpful
48 DIABETIC KETOACIDOSIS IN CHILDHOOD
to determine the type of fluid replacement required. If large
volumes of gastric contents are aspirated, these should be
replaced with 0.45 per cent saline containing 10 mmol/L KCl.
If acidosis is not resolving in the first few hours, resuscitation
may have been inadequate; therefore, consider giving
more saline. Sepsis is another possibility that should be
considered.
With initial resuscitation, the blood glucose often falls rapidly.
A fall of more than 5 mmol/L/h has been implicated in the
development of cerebral oedema, although no case–control
studies support this link. However, if it falls rapidly halve the
insulin infusion rate and/or increase the concentration of
dextrose in the replacement fluid.
Once the blood glucose has fallen steadily to 10–15 mmol/L,
fluid replacement should continue with a glucose-containing
solution, usually 5 per cent dextrose. This permits continuing
intravenous insulin therapy whilst avoiding iatrogenic hypoglycaemia.
Subsequently, blood glucose should be maintained by adjusting
the dextrose infusion rather than reducing the insulin infusion
rate lower than 0.05 U/kg/h, since both insulin and glucose
are required for the reversal of ketogenesis and glycogenolysis;
rapid relapse may ensue if insulin is interrupted.
Once the child is rehydrated, and is tolerating food and fluids,
subcutaneous insulin can be substituted for intravenous insulin.
The first dose of subcutaneous insulin should be given an
hour before terminating the intravenous insulin infusion;
this avoids transient insulinopenia. Note that urinary
ketones may be detectable for one or two days, owing to
the conversion of b-hydroxybutyrate (which is not measured
by conventional urine sticks) to acetoacetate. However, there
may still be a degree of insulin resistance at this stage, so larger
doses of insulin than usual may be required to suppress
ketogenesis.
MANAGEMENT 49
Complications
These are listed in Table 2.4. With meticulous management and
observations, severe hypokalaemia or hypoglycaemia and aspiration
pneumonia are now uncommon, and the greatest risk is from
cerebral oedema.
Cerebral oedema
This is almost exclusively a condition of childhood; over 95 per cent
of cases in the largest reported series occurred under the age of
20 years, with one-third under the age of 5 years. Mortality from
cerebral oedema is around 25–30 per cent and around 30 per cent of
survivors are left with major neurological morbidity. It is more
common in children with newly diagnosed type 1 diabetes.
Table 2.4 Complications of diabetic ketoacidosis
Under-treatment:
Unresolved acidosis Try further fluid resuscitation,
increase insulin dose, consider sepsis
Blood glucose not falling Increase insulin dose; ensure adequate
hydration
Recurrence of ketoacidosis Restart protocol from beginning
Over-treatment:
Unrecognised hypokalaemia Regular electrolyte measurements and
ECG monitoring
Hypoglycaemia Reduce but do not stop insulin, increase
dextrose concentration in fluids
Others:
Aspiration of gastric contents Early nasogastric intubation will
prevent this
Cerebral oedema Aetiology not understood
Pulmonary oedema/adult Commoner in adults
respiratory distress syndrome
ECG ¼ electrocardiograph.
50 DIABETIC KETOACIDOSIS IN CHILDHOOD
Subclinical brain swelling appears to be common during the
treatment of diabetic ketoacidosis, and may be present even before
intravenous rehydration is commenced. Whether severe, sudden
clinical cerebral oedema is an extension of this process, or whether
the two are distinct entities, remains to be determined. The clinical
signs of cerebral oedema are variable. Most cases have occurred
between 4 and 12 hours from the start of treatment. Signs and
symptoms of cerebral oedema are presented in Table 2.5.
Cerebral oedema is a feared complication of diabetic ketoacidosis
in children.
If warning features are not recognised, there is commonly a sudden
deterioration, manifest as loss of consciousness, appearance of
fixed dilated pupils or respiratory arrest. Possible contributing
factors include
cerebral anoxia from the reduced blood volume and haemoconcentration
high initial plasma glucose concentration
excessive rates of intravenous fluid administration
a rapid fall in plasma sodium concentration.
Table 2.5 Clinical features of cerebral oedema
complicating diabetic ketoacidosis
headache
confusion
irritability
reduced conscious level
convulsions
small pupils
increasing blood pressure, slowing pulse
papilloedema – not always present acutely
possibly impaired respiratory drive
MANAGEMENT 51
Animal studies have suggested that insulin is required for cerebral
oedema to occur, and hypoxia resulting from rapid bicarbonate
infusion has also been implicated. However, none of these theories
provides a complete explanation, and the fact that the incidence has
remained the same over several decades, despite changes in fluid
regimens, suggests that the fluid regimen may not be a crucial
factor. More recent studies suggest that the most severely
dehydrated children are at greatest risk.
The aetiology and optimal treatment of cerebral oedema
complicating diabetic ketoacidosis remain uncertain.
Only half of patients have a period of neurological deterioration
during which intervention might be effective before respiratory
arrest. Therefore, prevention of this complication remains one of
the most important goals of the management of diabetic ketoacidosis
in children. If cerebral oedema is suspected
exclude hypoglycaemia
inform senior medical staff immediately
give intravenous mannitol 0.5 g/kg stat (¼ 2.5 mL/kg mannitol
20 per cent over 15 min); administer as soon as possible
restrict intravenous fluids to two-third maintenance requirements
and replace deficit over 72 rather than 24 h
transfer child to intensive care unit
if necessary arrange for the child to be intubated and
hyperventilated to reduce blood pCO2
exclude other diagnoses by computed tomography – other
intracerebral events may occur (thrombosis, haemorrhage or
infarction) and present in the same way
52 DIABETIC KETOACIDOSIS IN CHILDHOOD
intracerebral pressure monitoring may be indicated
repeated doses of mannitol (dose as above every 6 h) can be
used to control intracranial pressure (recently it has been
suggested that hypertonic saline may be a more effective
osmotic agent, but insufficient studies have yet been performed).
Further reading
Edge JA. Cerebral oedema: are we any nearer finding a cause? Diabetes
Metab Res Rev 2000; 16: 316–324.
Edge JA, Ford-Adams ME and Dunger DB. Causes of death in children
with insulin dependent diabetes 1990–96. Arch Dis Child 1991; 81: 318–
323.
Glaser N, Barnett P, McCaslin I, Nelson D, Trainor J, Louie J, Kaufman F,
Quayle K, Roback M, Malley R and Kuppermann N. Risk factors for
cerebral oedema in children with diabetic ketoacidosis. N Engl J Med
2001; 344: 264–269.
Krane EJ, Rockoff MA, Wallman JK and Wolfsdorf JI. Subclinical brain
swelling in children during treatment of diabetic ketoacidosis. N Engl J
Med 1985; 312: 1147–1151.
Pinkney JH, Bingley PJ, Sawtell PA, Dunger DB and Gale EAM.
Presentation and progress of childhood diabetes mellitus: a prospective
population-based study. Diabetologia 1994; 37: 70–74.
Sperling MA. Diabetic ketoacidosis. Pediatr Clin N Am 1984; 31(3): 591–610.
FURTHER READING 53
3
Hyperosmolar
Non-ketotic
Hyperglycaemia
Hans J Woerle and John E Gerich
Summary
Hyperosmolar non-ketotic hyperglycaemia is one of the most
serious endocrine emergencies. Prompt and adequate therapy
is critical. Outcome depends on early correction of fluid deficit
and treatment of the underlying precipitating illness. Any
middle aged or elderly person with signs of mental status
deterioration and severe dehydration must be suspected of
hyperosmolar non-ketotic hyperglycaemia.
Diagnosis. Plasma glucose > 30 mmol/L with plasma osmolarity
> 320 mOsm/kg. Severe dehydration with pre-renal uraemia is
the rule. Note lack of ketoacidosis (pH > 7.3; plasma
bicarbonate > 15 mmol/L); cf. diabetic ketoacidosis.
Emergencies in Diabetes Edited by Andrew J. Krentz
# 2004 John Wiley & Sons, Ltd ISBN 0-471-49814-9
Therapy. Fluid replacement (0.9 per cent saline 1–2 L over the
initial 1 h) aiming for gradual restoration of normal osmolarity
(6–12 L over next 12 h). Control hyperglycemia using i.v. insulin
5–10 U/h in adults until plasma glucose has fallen to 15 mmol/L
when insulin infusion rate is reduced to 1–4 U/h. Prevention of
hypokalaemia typically requires 20–40 mmol potassium chloride
per litre, depending on renal function. General medical
support and treatment of precipitating factors, e.g. infection,
are important; patients often have serious comorbidity.
Acute complications. These include thrombo-embolic episodes,
rhabdomyolysis, seizures and transient focal neurological
signs.
Mortality. The case fatality rate is high – up to 50 per cent
depending on co-existing conditions.
Long-term management. A proportion of patients do not
require insulin long term. Avoid precipitating factors in the
future wherever possible.
Pathogenesis
Hyperosmolar non-ketotic hyperglycaemia is a serious and life
threatening condition that carries an average mortality rate of
15 per cent, mortality is increased in the presence of concomitant
illnesses and advancing age to over 50 per cent. The syndrome is
not uncommon, being generally responsible for approximately
1:1000 hospital admissions, more so in patients over the age of
60 years. Hyperosmolar non-ketotic hyperglycaemia is characterised
by
severe hyperglycemia
dehydration with pre-renal uraemia
hyperosmolarity.
56 HYPEROSMOLAR NON-KETOTIC HYPERGLYCAEMIA
Precipitating factors include conditions or medications leading to
severe dehydration and impaired insulin secretion and/or insulin
resistance. The most common are summarised in Table 3.1. The
most frequent and important precipitating factors are those that
cause dehydration. The stress of acute illnesses, e.g. acute
infections, is usually accompanied by insulin resistance manifested
by increased secretion of hormones that antagonise the actions of
insulin, e.g. glucagon, catecholamines, cortisol and growth hormone
together with certain cytokines, which impair insulin
sensitivity and insulin secretion (see Chapter 1). Some conditions
(e.g. acute pancreatitis) and medications (e.g. phenytoin, diazoxide)
directly reduce insulin secretion, while others induce insulin
resistance (e.g. anti-inflammatory doses of corticosteroids).
Hyperosmolar non-ketotic hyperglycaemia generally develops
over several days to weeks. Fever, vomiting and polyuria without
Table 3.1 Common precipitating factors for the development of
hyperosmolar non-ketotic hyperglycemia
Acute illness Drug therapy
Infection (30–60%) Corticosteroids
(most commonly cellulitis, Diuretics
pulmonary and urinary tract) Phenytoin
Sepsis Chlorpromazine
Stroke Cimetidine
Renal failure Diazoxide
Heat stroke b-adrenergic blockers
Hypothermia Antipsychotics – typical and
atypical
Pancreatitis Immunosuppressive agents
Severe thermal burns
Endocrine diseases:
type 2 diabetes
acromegaly
cushing’s syndrome
thyrotoxicosis
PATHOGENESIS 57
appropriate fluid intake, as may occur frequently in elderly people,
lead to severe dehydration. Approximately 80 per cent of affected
patients have been previously diagnosed with type 2 diabetes or
impaired glucose tolerance. Dehydration and hyperosmolarity
promote insulin resistance and impair insulin secretion. Since bcell
function is already impaired in patients with type 2 diabetes
and impaired glucose tolerance, an adequate increase in insulin
secretion does not occur; severe hyperglycaemia develops. An
osmotic diuresis due to hyperglycaemia in combination with
inadequate fluid intake leads to a further fluid loss and decreased
renal perfusion. Ultimately a vicious circle is formed in which
glucose enters plasma much faster than it can be taken up by
tissues or excreted by the kidneys.
Hyperosmolar non-ketotic hyperglycaemia develops gradually
over hours or days.
The typical patient is elderly with a plasma glucose level above
30 mmol/L, plasma hyperosmolarity greater than 320 mOsm/Kg
and a body water deficit of 20–25 per cent, i.e. approximately 10 L.
This water loss causes an increased plasma tonicity, which shifts
water together with potassium out of cells into the extracellular
space. At the same time, hydrogen ions are shifted into the cell.
Consequently, despite marked renal potassium losses, plasma
potassium levels are usually normal or elevated, and the blood pH
is in the normal range at time of admission.
Dehydration is a prominent features of hyperosmolar nonketotic
hyperglycaemia.
Up to 70 per cent of patients present with frank coma, which may
be solely the result of severe dehydration and hyperosmolarity. The
remaining patients show mild to moderate signs of lethargy; partial
58 HYPEROSMOLAR NON-KETOTIC HYPERGLYCAEMIA
seizures and other focal, reversible neurological deficits are
well recognised. A minor degree of ketosis or hyperlactataemia
may be present, the latter reflecting inadequate tissue
perfusion (see Chapter 6). However, by definition plasma ketone
body levels, and levels of their precursors – non-esterified
fatty acids, are much lower than those seen in diabetic
ketoacidosis; accordingly, the arterial pH is generally above 7.3.
Since diabetic ketoacidosis is also a state of hyperosmolarity,
several authors have suggested that hyperosmolar nonketotic
hyperglycaemia should not be differentiated from
diabetic ketoacidosis, but rather be recognised as being at one
end of the spectrum of severe acute metabolic derangements in
diabetes. It has been proposed that a major pathophysiological
differentiation is maintainence of a degree of endogenous insulin
secretion in hyperosmolar non-ketotic hyperglycaemia, in contrast
to the severe relative or absolute insulin deficiency in diabetic
ketoacidosis.
Endogenous insulin secretion. It is suggested that in the hyperosmolar
non-ketotic hyperglycaemia syndrome the presence of
circulating insulin is sufficient to suppress lipolysis and thus
prevent ketosis (see Chapter 1); however, the residual function
of the b-cells may not be enough to prevent hyperglycaemia.
There is, however, surprisingly little evidence to support this
hypothesis. In fact, the only two studies comparing plasma
insulin concentrations in diabetic ketoacidosis and hyperosmolar
non-ketotic hyperglycaemia did not find any significant
difference in plasma insulin concentrations, at least in the
peripheral circulation.
Hyperosmolality. Another theory is that dehydration–hyperosmolarity
and hyperglycaemia play the major role in the absence
of ketoacidosis in this condition: hyperosmolarity, hyperglycaemia
and dehydration reduce lipolysis (and thus availability
of the main precursor availability for ketogenesis, i.e. fatty
acids) and directly impair hepatic ketogenesis and insulin
secretion.
PATHOGENESIS 59
Counter-regulatory hormone levels. In addition, patients with
hyperosmolar non-ketotic hyperglycaemia generally have
lower plasma cortisol and growth hormone levels, which
would provide less stimulation of lipolysis. Regardless of the
pathophysiological mechanism, from a clinical point of view the
distinction between hyperosmolar non-ketotic hyperglycaemia
and diabetic ketoacidosis is of little importance since the main
elements of therapy for both conditions, i.e. fluid replacement
and insulin, are similar. In general, however, patients with
hyperosmolar non-ketotic hyperglycaemia require more fluids
and less insulin.
Significant ketosis and acidosis are not features of hyperosmolar
non-ketotic hyperglycaemia.
Diagnosis
Hyperosmolar non-ketotic hyperglycaemia is a medical
emergency that requires a high degree of suspicion for prompt
recognition and treatment; a fingerstick capillary glucose measurement
will readily indicate the presence of marked hyperglycaemia.
The diagnosis should be considered in any person who exhibits
signs of
dehydration
abnormal mental status or focal neurological signs
hypovolemia or shock.
Although usually regarded as a condition of the elderly, the
syndrome has occasionally been reported in children. The first
60 HYPEROSMOLAR NON-KETOTIC HYPERGLYCAEMIA
approach is a rapid but thorough history and physical examination
with special attention to
prior history of diabetes mellitus – hyperosmolar nonketotic
hyperglycaemia may be the initial manifestation of
diabetes
mental status
signs of broncho-pulmonary infection
airway, cardiovascular and renal status
common precipitating conditions (Table 3.1).
Coma may be present with severe hyperosmolality, but levels less
than 350 mOsm/kg should prompt consideration of alternative
causes, e.g. stroke a sedative drug overdose. Typically, conscious
patients report increasing polydipsia and polyuria and progressive
weight loss from dehydration for several days, sometimes
accompanied by nausea and vomiting, the latter symptoms being
more frequently encountered in diabetic ketoacidosis. Pre-existing
renal impairment may result in more severe degrees of hyperglycaemia
because renal losses are less pronounced; the degree of
dehydration will be reduced. Occasionally abdominal pain may
mimic symptoms of an acute abdomen although, once again, this is
more often encountered in diabetic ketoacidosis, particularly in
children (see Chapter 2). Since acute pancreatitis and other surgical
conditions may precipitate hyperosmolar non-ketotic hyperglycaemia
in predisposed individuals, an intra-abdominal cause should
be excluded. Signs of dehydration include
loss of skin turgor
dry mucous membranes
tachycardia
hypotension
oliguria.
DIAGNOSIS 61
Initial laboratory evaluation:
venous plasma glucose (usually >30 mmol/L)
plasma sodium (normal or increased)
arterial blood pH (normal or slightly reduced)
plasma osmolality (>320 mOsm/kg)
plasma potassium (normal or increased)
plasma creatinine and urea (increased)
blood count with differential – leukocytes (increased).
Plasma sodium concentration is reduced due to movement of water
from the intracellular to extracellular compartment consequent on
hyperglycaemia-associated hyperosmolality. However, countering
this effect, urinary sodium losses, if prolonged, may lead to marked
hypernatraemia. The osmotic diuresis also leads to urinary losses
of phosphate, calcium and magnesium. As for diabetic ketoacidosis,
depletion of total body levels of these ions is not reflected in
admission plasma concentrations.
Additionally,
bacterial cultures of blood, urine and sputum should be
obtained if ongoing infection is suspected
a chest radiograph should be obtained
electrocardiography should be considered.
Treatment
As in the management of diabetic ketoacidosis (see Chapters 1
and 2), a flow sheet monitoring treatment given, responses to
treatment and the results of serial laboratory determinations is
extremely useful (Figure 3.1).
62 HYPEROSMOLAR NON-KETOTIC HYPERGLYCAEMIA
Time Elapsed (h)
Glucose
Na
K
Cl
CO2
Urea (BUN)
Creatinine
pH
Serum osmolality
Insulin given
Fluids given
Urine output
Net fluid balance
Mental status*
BP
Pulse
Temp
CVP
0
Laboratory
* N, normal; D, drowsy; S, stuporous; C, comatose.
Notes: BUN, blood urea nitrogen; CVP, central venous pressure; BP, blood pressure.
Glasgow coma scale is used widely in UK.
Figure 3.1 Flow sheet for monitoring treatment
Management in an intensive care setting might be required,
depending on the clinical condition of the patient and presence
of co-morbidities. Therapy should be initiated without delay.
Diligent monitoring of clinical and biochemical variables is
required.
Capillary blood glucose is measured hourly at the bedside
during the initial phase of treatment, i.e. until blood glucose
is normalised and stable.
Serum electrolytes and fluid balance should be determined
hourly at least for the first three hours and if satisfactory
responses occur subsequently these can be measured at 2–3 h
intervals.
Other responses to treatment, e.g. fluid balance, mental
status, etc. should be assessed at similar intervals. It is
recommended that the chest X-ray be repeated after 24 h of
treatment because pulmonary infiltrates not present initially
may become evident after rehydration.
By 12 h or so, appropriate correction of hyperglycemia and
hyperosmolarity have usually been obtained and 60–80 per cent
of the fluid deficit restored; subsequent correction of residual fluid
deficits can be achieved orally since patients should be ready to
resume eating and drinking at this point. Although treatment
needs to be individualised, a generally applicable algorithm for
adults is given in Figure 3.2.
Fluid replacement
Severe dehydration is the leading element in the development of
hyperosmolar non-ketotic hyperglycaemia and is a major determinant
64 HYPEROSMOLAR NON-KETOTIC HYPERGLYCAEMIA
of morbidity and mortality. Consequently, rapid and adequate
fluid replacement is crucial. In all patients, an intravenous line
should be placed, and fluid therapy should be initiated immediately
to prevent avoidable complications due to dehydration
(stroke, myocardial infarction, acute renal failure, etc.).
Regardless of the initial plasma sodium concentration, it is
recommended that 1 L of 0.9 per cent saline be infused during
the initial 30 min. If the initial plasma potassium is below
3.5 mmol/L, add 40 mmol/L of saline and continue to add
potassium until the plasma potassium is above 3.5 mmol/L (see
below).
If the patient is hypotensive or has a plasma sodium level below
130 mmol/L, another litre of 0.9 per cent saline should be
infused over the next 30–60 min.
Subsequently, and in the absence of hypotension, hydration
may be continued with either 0.9 or 0.45 per cent sodium
chloride. No data are available from randomised studies to
FLUIDS
1 L 0.9% NaCl/ 30 60 min
INSULIN
10 15 U, IV bolus
~5 U/h, IV
POTASSIUM
Hypotensive or
plasma Na < 130 mmol/L
Normotensive
plasma Na > 130 mmol/L
1 L 0.9% NaCl/30 60 min
Continue until
normotensive, plasma
Na > 130 mmol/L
0.45% NaCl
0.5 1.0 L/h
At 1 h if plasma glucose* has
decreased 10 15%, continue
and monitor at 1 2 h intervals
Reduce insulin when
plasma glucose ≈ 15 mmol/L
Adjust insulin infusion to
maintain plasma glucose
6 11 mmol/L
If plasma K+ < 3.3 mmol/L
add 40 mmol/L to initial
saline infusion and continue
until plasma K+ > 3.3 while
delaying start of insulin
therapy.
If plasma K+ < 5.5 mmol/L
add 20 40 mmol/L to each L
of NaCl/dextrose
If plasma K+ > 5.5 mmol/L
withhold potassium until it
is < 5.5 mmol/L
*If plasma glucose does not decrease by 10%, double the insulin infusion rate hourly until a decrease of 10 15% /h is achieved.
Change to 5% dextrose
(0.5 1.0 L/h) when
plasma glucose ≈ 15 mmol/L
_ _
_
_
_
_
_
_
_
_
Figure 3.2 Suggested algorithm for the management of HNKH
TREATMENT 65
guide selection of initial fluid replacement; the use of 0.9 per
cent saline is widely accepted although some authorities
recommend 0.45 per cent saline. During the first 2 h the average
adult patient should receive 2–4 L of fluid. Thereafter the
infusion rate can be reduced to 0.5–1.0 L/h depending on
urinary volumes and cardiac status.
The severity of dehydration may ultimately require up to 12–15 L
within the first 12–24 h, taking continued diuresis into account.
Some authorities recommend calculating fluid deficits. However,
these consider only extracellular dehydration and have not
been shown to be superior to clinical assessment.
No controlled clinical studies are available to guide selection
of crystalloids for rehydration.
A thorough attention to urine output, signs of fluid overload such
as pulmonary congestion and jugular venous distension must
guide the rate of fluid administration; early bladder catheterisation
is helpful, although it should be removed promptly after recovery.
Central venous pressure and continuous urine output should be
monitored in patients with
a history of congestive heart failure
a history of renal insufficiency
acute renal failure
Accurate ascertainment of fluid balance is an important
aspect of management.
Fluid repletion itself will have a major impact on lowering plasma
glucose concentrations. Correction of volume contraction will
enhance renal glucose excretion and reduce overactivity of
the sympathetic nervous system, leading to enhanced hepatic
66 HYPEROSMOLAR NON-KETOTIC HYPERGLYCAEMIA
and peripheral insulin sensitivity. Too rapid a fall of plasma
glucose should be avoided to prevent dangerous fluid and
electrolyte shifts. Fluid may shift too rapidly from extra- to
intracellular space with the risk of cerebral oedema and compromised
perfusion of the brain and other vital organs.
Insulin therapy
Insulin therapy is perhaps less important than fluid replacement
and may be withheld – temporarily – until plasma potassium is
above 3.5 mmol/L in the unusual situation when patients present
with severe hypokalaemia.
After an initial i.v. bolus (10–15 U), which some clinicians
regard as optional, insulin should initially be infused at a low
rate, i.e. 5–10 U/h – typically 6 U/h in an adult. Note that the
short half-life of insulin in the circulation requires an uninterrupted
intravenous infusion.
The plasma glucose should not be lowered by more than 15 per
cent per hour. When the plasma glucose reaches 15 mmol/L,
an infusion of 5 per cent dextrose (plus potassium as necessary)
should be started; 100–150 mL/h is usually appropriate. In most
cases, dextrose will replace saline at this point; saline can be
infused simultaneously if necessary. Since the objective is to
maintain blood glucose on a plateau, the insulin infusion rate
should be lowered and adjusted so as to gradually achieve
values between approximately 5 and 10 mmol/L. This will
usually be achieved at rate of 0.5–4 U/h, with concomitant
infusion of dextrose.
If the plasma glucose initially fails to decrease, the volume
replacement regimen should be reassessed and the integrity of
the insulin infusion should be checked, e.g. was it made up
correctly? Is it functioning correctly? etc.
TREATMENT 67
Suggested infusion regimen
5 mL of U-100 human soluble insulin in 1 L of saline; at an infusion
rate of 10 mL/h this delivers 5 U/h insulin.
If plasma glucose has not decreased by 15 per cent within 2 h of
insulin and volume replacement, the insulin infusion rate should
be doubled and the response reassessed at hourly intervals.
Potassium replacement
If the initial plasma potassium concentration is 3.5–5.5 mmol/L,
co-administer 20 mmol potassium chloride per hour, added to
the infusate, be it saline or dextrose. Adequate fluid and insulin
administration will rapidly lower the plasma potassium as
potassium re-enters the intracellular compartment.
No potassium should be infused if hyperkalaemia (>5.5 mmol/L)
is present. Care must be exercised in patients with pre-existing
renal impairment or oliguria. Electrocardiographic monitoring
is recommended in all patients receiving higher potassium
doses for hypokalaemia, or showing any abnormal rhythm
(including tachycardia).
Some clinicians recommend that one-third of the potassium might
be given as potassium phosphate if plasma phosphate is low (i.e.
<0.5 mmol/L), since phosphate shifts along with potassium back
into the intracellular compartment. Potential complications of
severe hypophosphataemia, i.e. <0.3 mmol/L, are
haemolytic anaemia
muscle weakness
depressed systolic cardiac and respiratory performance.
On the other hand, controlled clinical trials have not demonstrated
any benefit from routine phosphate therapy in diabetic ketoacidosis
68 HYPEROSMOLAR NON-KETOTIC HYPERGLYCAEMIA
(see Chapter 1). It should also be borne in mind that excessive
administration may cause hypocalcaemia with tetany, soft tissue
calcification and renal failure. Therefore, care is required and it is
prudent to monitor plasma phosphate levels in patients being
given potassium phosphate. Most patients appear to recover fully
without the need for intravenous phosphate replacement.
The role of intravenous phosphate replacement in the
treatment of hyperosmolar non-ketotic hyperglycaemia is
uncertain.
Other aspects of management
After 2–4 h, when plasma glucose levels have decreased appreciably
and several litres of fluid have been administered, it is crucial
at this point to monitor the mental status of the patient. Any
reversion of initial improvement may be a sign of too vigorous
rehydration and/or too rapid reduction in plasma osmolarity,
causing the development of pulmonary or, occasionally, cerebral
oedema. The best evidence for an appropriate fluid management is
a constant improvement in mental status. Furthermore, hourly
glucose and electrolyte checks are recommended to avoid
hypoglycaemia and hypokalaemia.
Complications
Thrombo-embolic complications
Despite the high frequency of thrombo-embolic complications in
patients with the hyperosmolar syndrome, the role of prophylactic
anticoagulation is unclear. Anti-coagulation in an acutely
sick patient carries risks of gastrointestinal haemorrhage – an
TREATMENT 69
occasional cause of death. The alternative approach is to treat
clinically overt thrombo-embolic events as they arise; this approach
will lead to occasional failure to recognise the insidious development
of serious thrombosis. The risk–benefit equation will differ
between individual patients and this presents the clinician with a
dilemma.
Rhabdomyolysis
Non-traumatic rhabdomyolysis in patients with greater degrees of
hyperosmolarity occasionally precipitates acute renal failure. This
complication has been associated with a poor prognosis in some
reports. The diagnosis is suggested by a greatly elevated serum
creatinine kinase concentration (usually >1000 IU/L) in the
absence of alternative causes such as myocardial infarction, stroke
or pre-existing end-stage renal failure.
Discharge planning
It is important to emphasise the need to continue some insulin to
avoid relapses of hyperglycaemia and/or hyperosmolarity. Subcutaneous
insulin treatment should not be initiated too early;
insulin may be absorbed poorly and erratically from subcutaneous
tissue before effective perfusion is re-established.
By 12–24 h, once the patient is able to eat and drink,
subcutaneous insulin can be commenced, either as a multipledaily
(basal–bolus) regimen or as twice-daily injections of
biphasic insulin. Avoid using a sliding scale regimen for
subcutaneous insulin; this treats hyperglycaemia after it occurs
when the objective is to prevent hyperglycaemia.
Approximately 0.5–1.0 U/kg/day will usually be needed, in
divided doses: half is given as insulin glargine or isophane at
70 HYPEROSMOLAR NON-KETOTIC HYPERGLYCAEMIA
bedtime and the remaining half as short-acting insulin before
meals in proportion to the carbohydrate content, with somewhat
more given with breakfast; start with an injection of
soluble (or rapid-acting) insulin. Pre-prandial and bedtime
blood glucose levels are measured to guide changes in insulin
doses. In general, we do not recommended discontinuation of
insulin for several days, if at all. Insulin requirements are still
increased above what patients may require under normal life
conditions due to acute insulin resistance.
As insulin requirements decrease, other choices of therapy may
be considered; the decision to withdraw insulin should be made
by an experienced clinician.
A proportion of patients admitted with hyperosmolar non-ketotic
hyperglycaemia will be able to discontinue insulin and maintain
good glycaemic control with oral anti-diabetic agents or even
dietary measures. Precipitating factors, e.g. drugs, should be
avoided to prevent a recurrence.
Some patients with hyperosmolar non-ketotic hyperglycaemia
are able to discontinue insulin therapy after recovery.
Further reading
American Diabetes Association. Hyperglycemic crisis in patients with
diabetes mellitus. Diabetes Care 2003; 26: S109–S117.
Fishbein HA and Palumbo PJ. Acute metabolic complications in diabetes.
In: National Diabetes Data Group, ed. Diabetes in America 1995 pp. 283–
291. Bethesda, MD: National Institutes of Health.
Gerich JE. Hyperosmolar nonketotic coma. In: Kassirer J, ed. Current
Therapy in Internal Medicine, 3rd ed. 1991 pp. 1278–1281. Philadelphia,
PA: Decker.
FURTHER READING 71
Kitabchi AE, Umpierrez GE, Murphy MB, Barrett EJ, Kreisberg RA,
Malone JI and Wall BM. Management of hyperglycemic crises in
patients with diabetes. Diabetes Care 2001; 24: 131–53.
Matz R. Hyperosmolar nonacidotic diabetes. In: Porte D Jr and Sherwin
RS, eds. Diabetes Mellitus: Theory and Practice, 5th ed. 1997 pp. 845–860.
Amsterdam: Elsevier.
72 HYPEROSMOLAR NON-KETOTIC HYPERGLYCAEMIA
4
Insulin-induced
Hypoglycaemia
Simon R Heller
Summary
Hypoglycaemia may occur following ingestion of drugs such as
aspirin or alcohol and in those who develop an insulinoma, but
by far the commonest cause is insulin treatment in people with
insulin-treated diabetes. Most episodes are due to the limitations
of current subcutaneous insulin delivery, which is too
crude to prevent inappropriately raised plasma insulin levels in
between meals.
The brain is vulnerable to hypoglycaemia since cerebral tissue
cannot store carbohydrate and needs a continuous glucose
supply. Severe hypoglycaemia leads to impaired cognition,
confusion and coma. Very low levels (<1 mmol/L) that are
prolonged for hours can cause death or permanent cerebral
damage.
While many patients experience multiple episodes without
apparent ill effects, repeated severe hypoglycaemia may
Emergencies in Diabetes Edited by Andrew J. Krentz
# 2004 John Wiley & Sons, Ltd ISBN 0-471-49814-9
cause significant cognitive impairment. At diagnosis and for
some years after, patients with type 1 diabetes can mount a
powerful endocrine response to hypoglycaemia during which
glucagon and adrenaline resist the glucose-lowering effect of
insulin. Activation of the autonomic nervous system produces
symptoms that alert patients, giving them time to treat
themselves.
With increasing duration of either type 1 or type 2 diabetes,
both the hormonal and symptomatic responses to hypoglycaemia
often become impaired, a change also observed during
intensive insulin therapy. This makes it more difficult for
patients to identify the onset of hypoglycaemia and, although
not a major problem if mild, poses a considerable risk to those
severely affected. It is caused in part by periods of hypoglycaemia,
which cause maladaptive changes within the brain.
Recent work has shown that programmes of hypoglycaemia
avoidance can restore symptoms and awareness, at least in
part. The crucial issue in managing severe hypoglycaemia is to
reverse the threat of cerebral dysfunction and damage.
Intramuscular glucagon is the treatment of choice; its administration
needs no expert training or equipment and can be
given rapidly by family members or paramedics. Intravenous
glucose is only needed in the few who fail to respond
to glucagon. Failure to recover conciousness after 1 h is
indicative of possible cerebral damage; full recovery can follow
coma of some hours.
Epidemiology and frequency
Standard therapy with twice daily insulin leads to severe
hypoglycaemic episodes in around 10 per cent of patients with
type 1 diabetes per annum. This risk generally rises threefold
among those attempting tight glucose control with either multiple
74 INSULIN-INDUCED HYPOGLYCAEMIA
daily injections or continuous subcutaneous insulin infusion. The
risk of severe hypoglycaemia is considerably less in type 2 diabetes,
affecting around 0.5 per cent in a year of those taking sulphonylureas
(see Chapter 5) and 2–3 per cent in patients taking insulin.
Nevertheless, since type 2 diabetes is ten times more prevalent, as
many patients with insulin-treated type 2 experience a severe
episode as those with type 1 diabetes.
The risk of severe insulin-induced hypoglycaemia is considerably
lower in patients with type 2 diabetes in comparison
with patients with type 1 diabetes.
As the duration of type 2 diabetes increases, so the risk of severe
hypoglycaemia rises. This is probably due to a combination of
failing counter-regulatory responses (see below)
symptomatic unawareness and an increasing dependence on
the vagaries of exogenous insulin.
Vulnerability of the brain to
hypoglycaemia
During prolonged starvation the brain utilises ketone bodies as
alternative fuels and has also been shown to metabolise lactate.
However, under normal conditions, the brain depends entirely
upon glucose to sustain its metabolism. Unlike other tissues, the
brain has only limited stores of carbohydrate such as glycogen,
which can be rapidly converted to glucose. If the supply of glucose
to the brain is reduced below a critical value for even a few
minutes, then cerebral function becomes impaired. In humans,
this occurs at a glucose concentration of just above 3 mmol/L and
is initially manifest as a lengthening of reaction time, which can
be measured experimentally. If blood glucose continues to fall
VULNERABILITY OF THE BRAIN TO HYPOGLYCAEMIA 75
cerebral function progressively deteriorates, leading to increasingly
impaired cognition, confusion and eventually loss of consciousness.
The glucose concentration at which these changes occur
varies within and between individuals on different occasions. Some
people with diabetes can sometimes appear to behave completely
normally at a glucose concentration around 1 mmol/L; at other
times the same blood glucose concentration will render the patient
unconscious. The cause of this variability is unclear but perhaps
relates to the ability of the brain to adapt to hypoglycaemia through
mechanisms that maintain intra-neuronal glucose or via utilisation
of alternative metabolic fuels (see above).
Brain function is dependent on a continuous supply of
glucose.
Morbidity and mortality
It is not possible to perform experimental studies in humans to
establish the precise levels of hypoglycaemia that produce
permanent brain injury. Animal experiments have shown that at
glucose levels of below 1 mmol/L the electroencephalogram
tracing becomes flat; if sustained this causes severe and permanent
brain damage. However, under certain circumstances, the human
brain may withstand even this level of glucose without permanent
effects.
In the 1940s and 1950s, severe insulin-induced hypoglycaemia
was used to treat severe psychiatric conditions such as
schizophrenia. Although it was eventually shown to be an
ineffective treatment, some patients were subjected to repeated
episodes of insulin shock at glucose levels well below 1 mmol/L.
These produced coma and seizures, and although some patients
died or were left with permanent brain damage others
apparently suffered surprisingly few permanent effects.
76 INSULIN-INDUCED HYPOGLYCAEMIA
Severe hypoglycaemia (blood glucose below 1 mmol/L) of
longer duration, i.e. hours, can cause permanent brain injury.
On other occasions, however, even short-lived episodes can also
result in sustained cognitive impairment.
Nonetheless, severe episodes of hypoglycaemia caused by
accidental or suicidal overdose of insulin do not inevitably
lead to irreversible brain damage. When patients recover
consciousness they often exhibit severe memory loss, which then
recovers over the following weeks or months. Even prolonged
coma lasting 1–2 days can be followed by complete recovery.
Severe, recurrent or prolonged hypoglycaemia may cause
permanent brain damage.
Cerebral damage secondary to hypoglycaemia seems to be a
relatively rare cause of death in diabetic patients treated with
insulin; however, insulin-induced death may result from other
mechanisms. In a seminal paper, Tattersall and Gill (1991)
investigated all unexpected sudden deaths in patients with type 1
diabetes in the UK under the age of 40 during a single year. They
identified only two deaths that could be attributed to hypoglycaemic
brain damage with certainty. This study highlighted the
cases of 22 young people who were found dead in an undisturbed
bed. There was strong circumstantial evidence implicating hypoglycaemia;
many of the 22 had been experiencing intermittent
nocturnal hypoglycaemia in the months before their death.
Tattersall and Gill speculated that a cardiac arrhythmia might
have been involved. Recent studies have shown that experimental
hypoglycaemia in humans causes lengthening of the electrocardiographic
QT interval, a known cause of fatal dysrhythmias in some
other conditions. These conduction defects appear to be caused by
a combination of falls in plasma potassium, secondary to
sympatho-adrenal activation
direct effects of circulating adrenaline on the myocardium.
VULNERABILITY OF THE BRAIN TO HYPOGLYCAEMIA 77
What remains unclear is whether certain individuals with diabetes
are particularly vulnerable to QT lengthening during hypoglycaemia.
Since pharmacological selective b-adrenergic blockade prevents
these changes, if it proves possible to identify those most
vulnerable to QT lengthening then they might be protected by the
use of such agents; this hypothesis remains to be tested.
The dead-in-bed syndrome in patients with type 1 diabetes is
thought to be a result of severe nocturnal hypoglycaemia.
Why is hypoglycaemia common
in diabetes?
Unphysiological insulin delivery
Insulin treatment is designed to mimic the physiology of the islet
b-cell, delivering substantial and precise amounts of insulin to
cover the hyperglycaemia that follows meals, yet ensuring much
lower but stable basal concentrations in between. However, current
subcutaneous insulin preparations are inadequate to this task even
when administered in multiple small doses. The use of rapid-acting
insulin analogues delivered via continuous subcutaneous infusion
using an external electromechanical pump perhaps provides the
closest approximation to physiological insulin replacement. However,
even this form of insulin delivery produces inadequate
insulin concentrations during meals and inappropriately raised
plasma insulin concentrations when absorption from the gastrointestinal
tract is complete. This leads to a combination of
high post-prandial glucose concentrations
vulnerability to hypoglycaemia between meals. This is a
particular problem at night when a period of >12 h can
78 INSULIN-INDUCED HYPOGLYCAEMIA
separate the evening meal and the breakfast that follows the
morning after. Even if patients take a bedtime snack they
remain susceptible to nocturnal hypoglycaemia during the
second half of the night.
Current insulin regimens fall short of normal physiology,
thereby presenting the risk of hypoglycaemia.
The limitations of intermittent subcutaneous insulin delivery arise
partly because insulin enters the systemic rather than the portal
circulation. The inability to deliver the insulin directly to the liver,
as happens when insulin is secreted from the b-cells, causes higher
insulin levels in the peripheral circulation. In addition, short-acting
insulin preparations tend to self-associate into hexamers, delaying
the rate of absorption of insulin into the bloodstream from
subcutaneous depots. Different approaches have been tried to
develop a basal insulin preparation, i.e. that mimics normal
background low-level insulin secretion, that has stable and
consistent characteristics. However, lente and isophane preparations
not only produce an undesirable peak of insulin but have
considerable inter- and intra-subject variability.
Recently introduced rapid-acting and so-called basal insulin
analogues have improved pharmacokinetics that may prove more
useful than conventional insulin preparations. To date, clinical trial
data indicate that these novel analogues provide a relatively
modest albeit useful advance with reduced rates of hypoglycaemia
at similar levels of glycaemic control. Clinicians – and patients –
need to gain more experience with these analogues in order to
determine their optimal application. As for conventional insulin,
the particular circumstances of the patient will require careful
individualisation of therapy to attain glycaemic targets safely.
Insulin analogues with improved pharmacokinetic profiles
are associated with a reduced risk of hypoglycaemia.
WHY IS HYPOGLYCAEMIA COMMON IN DIABETES? 79
Pathophysiology of hypoglycaemia in diabetes
Recovery from insulin-induced hypoglycaemia would take many
hours if dissipation of insulin were the sole mechanism (Figure 4.1).
Additional physiological mechanisms help to resist the glucose
lowering effect of insulin and restore blood glucose after an episode
of hypoglycaemia. Secretion of counter-regulatory hormones
promotes glucose release from the liver
opposes glucose uptake in peripheral tissues such as fat and
muscle.
The most important of these hormones, in terms of recovery from
hypoglycaemia, are glucagon and adrenaline, although others such
as growth hormone and cortisol make aminor contribution (Table 4.1).
These are precisely the same hormones that,when present in excess in
Table 4.1 Counter-regulatory hormones
*Glucagon
*Adrenaline
Cortisol
Growth hormone
*Major effects on recovery from acute hypoglycaemia
Release of counter-regulatory hormones 3.5 mmol/L
Prolonged reaction time 3.2 mmol/L
Symptoms 3.0 2.8 mmol/L
Confusion 2.0 mmol/L
Coma 1.0 mmol/L
_
Figure 4.1 Sequence of events during hypoglycaemia
80 INSULIN-INDUCED HYPOGLYCAEMIA
concert with insulin deficiency, promote major metabolic decompensation
such as diabetic ketoacidosis (see Chapter 1).
Glucagon. This hormone ranks high in the hierarchy of hormonal
defences against hypoglycaemia, causing a prompt and substantial
release of glucose from the liver. However, within a few
years of diagnosis of type 1 diabetes the glucagon response to
hypoglycaemia becomes progressively impaired. While the
cause of this maladaptation remains unknown, increasing
evidence suggests that b-cell destruction prevents the appropriate
fall in local insulin concentration within the islet, which
stimulates the a-cells to release glucagon. As the glucagon
response fails, patients with diabetes become increasingly
dependent upon the release of adrenaline to help to restore
circulating glucose concentrations.
Adrenaline. Release of adrenaline reflects activation of the
sympatho-adrenal system, which generates a set of characteristic
symptoms that alert patients to a falling plasma glucose
concentration. Some individuals, particularly those with diabetes
of long duration, also fail to release adrenaline during
hypoglycaemia; this combined hormonal failure leads to a
major increase in the risk of hypoglycaemia. As described
below, this occurs because
1. those affected not only lack the hormonal mechanisms to
raise blood glucose,
2. they can no longer reliably identify an impending episode of
hypoglycaemia because warning symptoms are blunted or
absent.
Symptoms of hypoglycaemia
Patients with insulin-treated diabetes rely on the physiological
responses to hypoglycaemia to alert them to a falling glucose that
WHY IS HYPOGLYCAEMIA COMMON IN DIABETES? 81
prompts them to take action by taking refined carbohydrate
(Table 4.2). Symptoms are generated through a combination of
1. sympathoadrenal activation (often termed autonomic symptoms)
and
2. cerebral dysfunction caused by a failing glucose supply to the
brain (termed neuroglycopenic symptoms).
Each patient learns to recognise his or her own pattern of symptoms,
although these can vary over time, even day to day.
Autonomic symptoms. In the early years after diagnosis,
autonomic symptoms are usually more prominent and patients
have sufficient time to correct impending hypoglycaemia.
However, those individuals who lose their sympatho-adrenal
response note a reduction in autonomic symptoms such as
sweating and tremor. Increasingly, the patient has to depend
upon neuroglycopenic symptoms such as loss of concentration.
Neuroglycopenic symptoms. These reflect deteriorating cerebral
function, and so by the time these develop the cognitive ability
of those affected is already diminished. Unless patients can take
steps to raise their blood glucose within a few minutes any
further decline will, through neuroglycopenia, render them
incapable of responding appropriately; without prompt external
assistance severe hypoglycaemia may result, leading to coma or
convulsions.
Table 4.2 Symptoms of hypoglycaemia
Neuroglycopenic Autonomic
Confusion Tremor
Irritability and bad temper Sweating
Aggression Palpitations
Lack of concentration
Diminished conscious level
Coma
82 INSULIN-INDUCED HYPOGLYCAEMIA
Hypoglycaemia unawareness
Overall, around 25 per cent of patients with type 1 diabetes
experience difficulties in identifying when their blood glucose
concentration is low; this situation is probably the rule in those with
diabetes of long duration. The majority of patients have some
degree of unawareness after 20 years or more of diabetes.
However, hypoglycaemia unawareness can become a major clinical
problem for a few individuals with type 1 diabetes regardless of
duration. This situation increases the risk of severe hypoglycaemia
sevenfold and can devastate the lives of both the patient and
their families. Those severely affected are unable to drive motor
vehicles and many find that their jobs are impossible to hold down.
Loss of the warning symptoms of hypoglycaemia –
hypoglycaemia unawareness – increases the risk of severe
recurrent hypoglycaemia.
As outlined above, symptomatic awareness of hypoglycaemia
largely depends upon an intact sympatho-adrenal response; it is
the failure of this response which causes the clinical problem of
hypoglycaemia unawareness. Although a sympatho-adrenal response
still occurs, it is only activated at a low glucose concentration, i.e.
approximately 2.5 mmol/L, the normal level for activation being
around 3.5 mmol/L. It is clearly hazardous for those affected to try
to maintain blood glucose levels close to normal by embarking on
intensive insulin therapy. Frequent blood glucose monitoring will
reduce the risk of a severe episode of hypoglycaemia, but affected
patients should be encouraged to keep their blood glucose levels
slightly above normal. For many this proves a difficult prospect, as
they are often more worried about the risks of long-term
complications resulting from hyperglycaemia.
Hypoglycaemia unawareness may be reversed with meticulous
avoidance of low blood glucose concentrations.
WHY IS HYPOGLYCAEMIA COMMON IN DIABETES? 83
Causes of hypoglycaemia in
insulin-treated patients
The traditional causes include
excessive insulin doses
missed meals
inappropriate physical exercise.
In blaming the patient, invoking these factors fails to recognise
what is increasingly appreciated as the main issue, i.e. ineffective
insulin delivery. Many severe episodes of hypoglycaemia have no
discernible explanation and probably, at least in part, reflect the
variability of absorption of conventional insulin preparations.
Anyone with insulin-treated diabetes, even with relatively poor
overall glycaemic control, is at risk of the occasional severe episode,
particularly during the night. Unsurprisingly, attacks are more
likely in those whose aim is to maintain blood glucose concentrations
close to normal (Table 4.3). This is in part because of a greater
probability of low blood glucose values in those striving for near
normoglycaemia but also because recurrent hypoglycaemia itself
induces a failure of the physiological defences that oppose
hypoglycaemia.
Table 4.3 Factors associated with reduced awareness of hypoglycaemia
and an increased risk of hypoglycaemia
Duration of diabetes
Intensive insulin therapy
Drugs, including alcohol
Extremes of age
84 INSULIN-INDUCED HYPOGLYCAEMIA
Duration of diabetes
Increasing duration of disease leads to progressive failure of
insulin secretion in both type 1 and type 2 diabetes. As patients
become increasingly dependent on exogenous injected insulin, they
develop more erratic plasma insulin profiles. This makes them
prone to hypoglycaemic episodes, particularly during the night,
where the use of insulin preparations such as isophane, which has
a peak of action in the early hours, has its greatest effect. Both of the
main hormonal defences against hypoglycaemia are compromised
as a function of the duration of disease.
As discussed above, diminishing endogenous insulin secretion
probably contributes to a failure of the glucagon response.
The failure of the sympatho-adrenal response to hypoglycaemia
in some patients is also related to increasing disease duration,
although its cause is unclear. The belief that it is due to classic
autonomic neuropathy is probably incorrect since some patients
with severe autonomic neuropathy on cardiovascular reflex
testing can mount a brisk sympatho-adrenal response. Furthermore,
many individuals with impaired counter-regulatory
responses have normal autonomic function tests.
Tight glycaemic control and the effects of
antecedent hypoglycaemia
Intensive insulin therapy is now known to be an important cause of
reduced physiological defences to hypoglycaemia and unawareness.
Despite efforts to exclude high-risk patients, the incidence of
severe hypoglycaemia was increased several-fold among patients
randomised to the intensive treatment arm of the US Diabetes
Control and Complications Trial compared with the conventional
treatment group; the latter had a significantly higher mean glycated
haemoglobin concentration.
CAUSES OF HYPOGLYCAEMIA IN INSULIN-TREATED PATIENTS 85
The increased risk of hypoglycaemia appears to be due to
periods of antecedent hypoglycaemia that usually accompanies
intensification of treatment. The changes appear similar to those
observed with increased disease duration with a resetting of the
threshold for activation of the autonomic response and symptoms
to a glucose level below rather than above that for cognitive
dysfunction. The site and precise nature of the abnormality are
unknown but presumably the cerebral pathways responsible for
sensing and activating the autonomic response are disrupted. This
might be due to adaptation to hypoglycaemia within the central
nervous system or possibly modulation by one of the counterregulatory
hormones. Cortisol is known to have powerful effects
on neuronal function and infusing cortisol to levels seen during
hypoglycaemia can produce impaired hormonal responses and
symptoms in response to subsequent episodes. Thus, recurrent
hypoglycaemia can produce a vicious circle of reduced physiological
protection leading to a greater risk of subsequent hypoglycaemia
and eventually a state of hypoglycaemia unawareness.
Prior episodes of hypoglycaemia alter the glycaemic threshold
for secretion of counter-regulatory hormones.
Since short-lived hypoglycaemic episodes can cause severe functional
defects without any apparent structural change, it follows
that avoidance of hypoglycaemia might reverse the defect. This has
now been demonstrated in clinical studies in which meticulous
avoidance of hypoglycaemia led to recovery of both symptoms and
the hormonal response to hypoglycaemia. In these studies it was
possible to restore awareness of hypoglycaemia, at least in part,
without significantly worsening blood glucose control, although
HbA1c concentration tended to rise. Furthermore, hypoglycaemia
reversal programmes, although requiring intensive input from
diabetes health care professionals, is within the scope of most
diabetes units, particularly as those severely affected are relatively
86 INSULIN-INDUCED HYPOGLYCAEMIA
rare. The essential principle is avoidance of any hypoglycaemia,
however mild, through a combination of frequent self-monitoring
of blood glucose by the patient allied to a willingness to accept less
stringent glucose targets.
Alcohol
Alcohol, while not directly lowering blood glucose, prevents
recovery from hypoglycaemia via inhibition of hepatic gluconeogenesis.
This results from the altered redox state generated by the
metabolism of ethanol. Alcohol can therefore turn a mild
hypoglycaemic attack into a severe and prolonged episode. It
also suppresses some of the symptoms of hypoglycaemia such as
tremor and this, combined with impaired cognition, can induce a
temporary state of hypoglycaemia unawareness. The danger of
unrecognised hypoglycaemia is increased by the similarity of
symptoms of hypoglycaemia to intoxication with alcohol. Thus,
friends and relatives may assume that a patient exhibiting odd
behaviour due to hypoglycaemia is drunk and leave them to sleep
it off, with potentially disastrous consequences.
Those who take insulin need to be warned of the dangers of
hypoglycaemia when consuming alcohol. People around them
should be aware of their diabetes.
It is prudent to maintain blood glucose levels a little higher
if they plan to drink a potentially intoxicating volume of
alcohol; carbohydrate snacks should be eaten as well in such
situations.
Alcohol has a well recognised propensity to impair recovery
from hypoglycaemia.
CAUSES OF HYPOGLYCAEMIA IN INSULIN-TREATED PATIENTS 87
Patients particularly prone to hypoglycaemia
Those with diabetes at the extreme of age are at particularly high
risk of hypoglycaemia.
Infants and young children. These groups are vulnerable due to
the irregularity of meals and unpredictable exercise. Those in
this age group are prone to nocturnal episodes partly because
over 12 h often separates their evening meal from breakfast. One
study reported rates of hypoglycaemia of up to 70 per cent in an
unselected group of children attending a hospital clinic, who
remained asleep despite profound nocturnal hypoglycaemia.
The developing brain may be especially vulnerable to the
damaging effects of hypoglycaemia. There is evidence that
severe and repeated hypoglycaemia in early childhood can
result in developmental delay and measurable defects in
cognitive function. Both parents and health care professionals
responsible for the care of young children with diabetes face
considerable difficulties in reconciling the desire for tight
metabolic control with the threat of hypoglycaemia. Rapid
acting insulin analogues have been used post-prandially
with some success and there is a growing trend to multiple
injections or insulin pumps, specifically to reduce the risk of
hypoglycaemia.
Elderly patients. The elderly exhibit diminished sensitivity to
catecholamines as well as a less prominent sympatho-adrenal
response. Symptoms of hypoglycaemia are often less specific
and coupled with the tendency of others to attribute confusion
and abnormal behaviour to cerebrovascular disease increases
their vulnerability to hypoglycaemic episodes. Since near
normoglycaemia has uncertain long-term benefit in this age
group, it seems sensible to aim for less stringent levels of
glycaemic control to prevent unnecessary hypoglycaemia.
88 INSULIN-INDUCED HYPOGLYCAEMIA
Patients at the extremes of age are more vulnerable to the
adverse effects of hypoglycaemia.
Does the species of insulin affect the
risk of hypoglycaemia?
The question of whether human insulin might contribute to
hypoglycaemia unawareness was raised during the 1980s. The
development of recombinant insulin of human structure resulted in
its widespread introduction to many patients who previously been
using animal insulin without problems. A minority complained
vociferously of different problems including a major reduction in
hypoglycaemic warning signs, which improved when they were
transferred back to animal insulin. However, repeated studies have
failed either to confirm a consistent reduction in physiological
responses and symptoms in those on human insulin or identify any
convincing mechanisms. Furthermore, the concerns were confined
to only a few countries such as the UK and Switzerland. The
introduction of human insulin in others such as the USA and
Germany produced few problems. To some, the most likely
explanation is that the time of transfer coincided with attempts to
tighten glycaemic control, and it was this that led to a loss of
hypoglycaemic warning. Others have suggested that an increased
rate of asymptomatic nocturnal hypoglycaemia due to differences
in insulin kinetics could have caused a reduction in autonomic
symptoms. Many years later, the question remains unresolved and
it now seems unlikely that a study with sufficient power will ever
be mounted. It is ironic that recombinant technology has now
resulted in the development of both short- and long-acting insulin
analogues whose chief benefit is to reduce the incidence of
hypoglycaemia.
CAUSES OF HYPOGLYCAEMIA IN INSULIN-TREATED PATIENTS 89
The controversial issue of insulin species and risk of
hypoglycaemia remains unresolved.
Diagnosis and management
In a patient known to have insulin-treated diabetes with features
suggestive of hypoglycaemia,
rapidly measure capillary blood glucose at the bedside.
Take venous blood (in a fluoride oxalate tube) for confirmation
of the diagnosis by an accredited laboratory.
In the uncommon situation wherein the patient is not known to
have diabetes, take 10 mL plasma simultaneously and ask the
laboratory to freeze at 20 C; this allows plasma insulin and/or
sulphonylurea concentrations to be measured subsequently if
an insulinoma needs to be excluded or sulphonylurea-induced
hypoglycaemia (see Chapter 5) is suspected; certain tumours are
associated with hypoglycaemia with appropriately suppressed
insulin concentrations. Other potential causes of spontaneous
hypoglycaemia include
excessive alcohol consumption, especially in children or in
the absence of carbohydrate intake
severe liver disease
hypoadrenalism – primary or secondary e.g. hypopituitarism.
If the patient is able to maintain an airway give oral glucose ideally
in liquid form, e.g. a proprietary glucose drink. Preparations in
gelform, e.g. Hypostop1, may also be useful. If patients are unable
to take anything orally then give either i.m. glucagon or i.v.
dextrose.
90 INSULIN-INDUCED HYPOGLYCAEMIA
Glucagon. This is the initial treatment of choice as it can be given by
both paramedics and nursing staff. Give 0.5–1.0 mg i.m.; it can
cause vomiting, particularly in children and at higher doses in
adults. Give i.v. dextrose if there has been no response after 10 min.
It is important to ensure that oral carbohydrate is taken after the
patient regains consciousness to prevent early relapse.
Intravenous dextrose. Give as 20–30 mL of 50 per cent dextrose into a
large forearm vein, which should be flushed afterwards with saline
to reduce the chances of thrombophlebitis. Do not overtreat. Start
an i.v. infusion of 10 per cent dextrose in those who fail to maintain
their blood glucose at normal levels or who do not regain
consciousness.
Other measures. Prolonged coma, i.e. >60 min, raises the possibility
of cerebral damage. Maintain blood glucose at around 10 mmol/L.
Dexamethasone is often given to reduce cerebral oedema, although
there are no published data establishing benefit. Even prolonged
coma lasting for many hours has been followed by an apparently
full recovery. Those who make a rapid recovery and whose blood
glucose remains stable over one hour can be discharged from
hospital, ideally in the care of a relative or friend. Consider the
precipitating cause. If one can be identified, take steps to avoid
further episodes, e.g. by reducing insulin dose, if appropriate. Inform
the clinician supervising the usual care of the patient.
What do patients need to know about
hypoglycaemia?
The symptoms of hypoglycaemia should be explained to all
patients starting insulin together with the appropriate action they
should take in response to such symptoms. It is still part of the
educational policy in some centres to induce a hypoglycaemic
episode in hospital shortly after diagnosis. This is difficult to
achieve reliably and the symptoms produced in this artificial
DIAGNOSIS AND MANAGEMENT 91
situation are often different from those experienced at home. Since
poorly supervised attempts have occasionally resulted in permanent
brain damage and death such a policy seems of little value
and should be abandoned.
Many patients fear the risks and results of hypoglycaemia more
than microvascular complications and a discussion of these issues
should be part of any structured education programme. Individuals
and their families should be told that mild hypoglycaemia is
an inevitable part of life for anyone maintaining tight glucose
control and that even severe episodes rarely cause permanent
harm. The families of those with insulin-treated diabetes also need
to know how to treat a severe episode, including the use of
glucagon.
The problem of hypoglycaemia in insulin-treated patients will
only be solved by the still distant prospect of new methods of
insulin delivery. However, the present situation may be alleviated
by the imminent development of reliable continuous glucose sensors
incorporating an alarm, e.g. to wake patients before they develop
severe nocturnal hypoglycaemia. Until then, we need to ensure that
patients are realistically informed about the risks and that they
possess the skills to manage their insulin treatment effectively.
Steps should be taken to avoid recurrence of severe insulininduced
hypoglycaemia wherever possible.
Further reading
Cranston I, Lomas J, Maran A, Macdonald IA and Amiel SA. Restoration
of hypoglycaemia unawareness in patients with long-duration insulindependent
diabetes. Lancet 1994; 344: 283–287.
Cryer PE. Hierarchy of physiological responses to hypoglycemia:
relevance to clinical hypoglycemia in type I (insulin dependent)
diabetes mellitus. Hormone Metab Res 1997; 29: 92–96.
92 INSULIN-INDUCED HYPOGLYCAEMIA
Davis SN, Shavers C, Costa F and Mosqueda-Garcia R. Role of cortisol in
the pathogenesis of deficient counterregulation after antecedent
hypoglycemia in normal humans. J Clin Invest 1996; 98: 680–691.
Diabetes Control and Complications Trial Research Group. The effect of
intensive treatment of diabetes on the development and progression of
long-term complications in insulin-dependent diabetes mellitus. N Engl
J Med 1993; 329: 683–689.
Heller SR. How should hypoglycaemia unawareness be managed? In: Gill
G, Williams G, Pickup J, eds. Difficult Diabetes – Current Management
Challenges. Oxford: Blackwell; 2001.
Heller SR, Macdonald IA, Herbert M and Tattersall RB. Influence of
sympathetic nervous system on hypoglycaemic warning symptoms.
Lancet 1987; ii: 359–363.
Jorgensen LN, Dejgaard A and Pramming SK. Human insulin and
hypoglycaemia: a literature survey. Diabet Med 1994; 11: 925–934.
Matyka KA, Wigg L, Pramming S, Stores G and Dunger DB. Cognitive
function and mood after profound nocturnal hypoglycaemia in
prepubertal children with conventional insulin treatment for diabetes.
Arch Dis Child 1999; 81: 138–142.
Meneilly GS, Cheung E and Tuokko H. Altered responses to hypoglycaemia
of healthy elderly people. J Clin Endocrinol Metab 1994; 78:
1341–1348.
Pramming S, Thorsteinsson B, Bendtson I and Binder C. Symptomatic
hypoglycaemia in 411 type 1 diabetic patients. Diabet Med 1991; 8:
217–222.
Rovet JF and Ehrlich RM. The effect of hypoglycemic seizures on cognitive
function in children with diabetes: a 7-year prospective study. J Pediatr
1999; 134: 503–506.
Tattersall RB and Gill GV. Unexplained deaths of type 1 diabetic patients.
Diabet Med 1991; 8: 49–58.
FURTHER READING 93
5
Hypoglycaemia
Caused by Insulin
Secretagogues
Kathleen M Colleran, Andrew J Krentz
and Mark R Burge
Summary
The sulphonylureas were introduced into clinical practice in the
1950s, their availability representing an early milestone in the
treatment of type 2 diabetes. Sulphonylureas remain a mainstay
of treatment. There are a wide variety of sulphonylurea
agents with differing pharmacokinetic properties that confer
differences in risk of hypoglycaemia.
Sulphonylureas work primarily by stimulating islet b-cells
to secrete insulin; the sulphonylureas are therefore insulin
secretogogues. As a result sulphonylureas, like insulin, carry a
risk of hypoglycaemia, this being the main unwanted effect of
these agents. Due to the wide availability of the sulphonylureas,
intentional overdose and accidental ingestion and are not
uncommon, particularly among children.
Emergencies in Diabetes Edited by Andrew J. Krentz
# 2004 John Wiley & Sons, Ltd ISBN 0-471-49814-9
Sulphonylureas are a causative factor in up to two-thirds of all
hypoglycaemic events. Actual or potential hypoglycaemia is
often the limiting factor in attaining target glucose goals in
subjects with type 2 diabetes. In fact, retrospective studies
report an incidence of up to 20 per cent for hypoglycaemia
associated with use of these agents.
Severe sulphonylurea induced hypoglycaemia is a medical
emergency requiring hospital admission. The elderly are at
highest risk, particularly if food intake is compromised or
hepatic or renal disease is present; care must also be taken to
avoid drug interactions. Treatment with intravenous dextrose
may be needed for several days, particularly for agents with a
long duration of action. The somatostatin analogue, octreotide,
may be a useful adjunct to dextrose. Mortality related to
sulphonylurea-induced hypoglycaemia is not insignificant,
with case-fatality rates approaching 10 per cent.
Pathophysiology
All of the sulphonylureas are associated with a risk of hypoglycaemia.
The mechanism of hypoglycaemia is related to the
mechanism of action and the pharmacodynamic properties of the
drugs. As shown in Figure 5.1, sulphonylureas work by increasing
insulin availability through a multi-step process. They initially bind
to the sulphonylurea receptor located on the plasma membrane of
the islet b-cell. Upon binding to the receptor, an adenosinephosphate-
dependent potassium channel is inhibited. This in turn
leads to depolarisation of the cell membrane. Subsequently,
calcium channels open and changes in calcium flux result in the
release of preformed insulin from the cell.
Sulphonylureas stimulate insulin secretion from islet b-cells;
all sulphonylureas have the potential to cause hypoglycaemia.
96 HYPOGLYCAEMIA CAUSED BY INSULIN SECRETAGOGUES
In patients with type 2 diabetes, sulphonylureas stimulate insulin
release. The effects of the sulphonylureas are more pronounced in
the setting of hyperglycaemia. In fact, sulphonylurea agents and
glucose have a synergistic effect on insulin release; this can become
a limiting factor in treatment. In the normal pancreas, insulin
secretion is terminated once plasma glucose levels have normalised.
In the presence of sulphonylureas, however, insulin release
persists despite euglycaemia. Hypoglycaemia can thus develop if
insulin release exceeds glucose availability. This occurrence can be
observed in many settings, including excessive sulphonylurea
dosage or lack of concomitant carbohydrate ingestion.
Another contributing factor to sulphonylurea induced hypoglycaemia
is the presence of biologically active metabolites of
the drugs. While the sulphonylureas are principally metabolised
in the liver, many of the agents have active metabolites that are
eliminated by the kidney (see Table 5.1). In patients receiving
sulphonylurea therapy, the development of hepatic dysfunction or
renal insufficiency can lead to reduced metabolism and delayed
clearance of the drug. This is associated with prolonged exposure
Figure 5.1 Schematic representation of the mechanism of action of the
sulphonylurea agents
PATHOPHYSIOLOGY 97
of the islet b-cells to sulphonylurea, and the resultant hyperinsulinaemia
may be followed by hypoglycaemia. Sulphonylureas with
longer half-lives and/or active metabolites are more likely to cause
hypoglycaemia compared with agents that do not have these properties.
Drugs with these properties include chlorpropramide (no
longer used in the UK) and the most popular sulphonylurea in the
USA, glibenclamide (glyburide). The recently introduced oncedaily
sulphonylureas, glimepiride and gliclazide MR (not available
in the USA), carry a relatively low risk, recent head-to-head data
suggesting a lower risk with gliclazide MR. Tolbutamide, a lowpotency
first generation agent, has a short half-life with a correspondingly
low risk of hypoglycaemia.
The pharmacokinetics of sulphonylureas is an important
determinant of the risk and severity of hypoglycaemia
associated with their use.
Table 5.1 Pharmacokinetics of the sulphonylureas
Trade T1=2 Active Renal excretion
Drug name (hours) metabolite of metabolite (%)
Tolbutamide Orinase 6–10 þ 100
Rastinon
Acetohexamide Dymelor 12–18 þþ 100
Tolazamide Tolinase 16–24 þ 100
Chlorpropamide Diabinase 24–72 þ 100
Glibenclamide Micronase 16–24 þ/ 50
(glyburide) Daonil
Diabeta
Glynase
Glipizide Glucotrol 12–16 85
Glibenese
Glipizide GITS Glucotrol XL 12–16 85
Glimepiride Amaryl 24 þ 60
Gliclazide Diamicron 10–20 60–70
Gliclazide MR Diamicron MR 30–120 60–70
Trade names may differ between countries, as may the availabity of certain drugs
or preparations.
98 HYPOGLYCAEMIA CAUSED BY INSULIN SECRETAGOGUES
Other factors include the following.
Advanced age. This appears to be an additional risk factor for
sulphonylurea induced hypoglycaemia. Retrospective studies
suggest that elderly diabetic subjects do not metabolise or
clear sulphonylureas as readily as do younger individuals. Drug
may accumulate in this population, leading to sulphonylurea
induced hyperinsulinaemia and hypoglycaemia. However, one
prospective study in otherwise healthy elderly individuals with
type 2 diabetes did not find such an association. It may be that
additional or concomitantly occurring risk factors, such as acute
illness, polypharmacy, medication error, or ethanol ingestion,
coincide to increase the incidence of sulphonylurea induced
hypoglycaemia in elderly subjects. Risk factors for sulphonylurea
induced hypoglycaemia are listed in Table 5.2.
Drug interactions. Several drugs can interfere with the metabolism
of sulphonylureas. This may prolong their half-life and
result in increased insulin secretion, thereby increasing the risk
of hypoglycaemia. Table 5.3 lists these drugs.
Patients with renal or hepatic disease are at increased risk of
sulphonylurea induced hypoglycaemia.
Table 5.2 Risk factors for hypoglycaemia with sulphonylureas
Renal insufficiency
Hepatic insufficiency
Prolonged fasting, e.g. peri-operatively
Acute or chronic intercurrent illness
Long acting sulphonylureas
(chlorpropramide, glibenclamide)
Drugs that interfere with the metabolism of sulphonylureas and
prolong their bioactivity (see Table 5.3)
Short duration of use
Elderly patients
Polypharmacy
Alcohol consumption
PATHOPHYSIOLOGY 99
Sulphonylurea induced hypoglycaemia is not only a problem in
individuals with diabetes. It is also occurs in subjects who do not
have diabetes, and these agents have been used in suicide attempts
by both diabetic and non-diabetic individuals. This is in part due to
the wide availability of, and access to, these potentially dangerous
drugs. In fact, people without diabetes may be more sensitive to the
hypoglycemic effects of sulphonylureas than are people with
underlying insulin resistance and impaired b-cell function.
Additionally, prescribing or dispensing errors have occasionally
resulted in the inadvertent administration of sulphonylureas to
subjects without diabetes, causing hypoglycaemia that often
requires a thorough medical investigation to uncover (see Chapter
4). Sulphonylurea induced hypoglycaemia in children is not
uncommon. In fact, approximately half of all cases of sulphonylurea
ingestion reported to US poison control centres occur in the
paediatric age group. This is mainly attributable to the large
number of sulphonylurea prescriptions written each year and
the wide availability of these agents. A single sulphonylurea tablet
ingested by a child can lead to life threatening hypoglycaemia and
requires substantial evaluation, observation and treatment as
described below.
Sulphonylureas may be ingested accidentally, especially by
children, or used in deliberate self-poisoning by adults.
Table 5.3 Drugs that delay metabolism of
sulphonylureas
Warfarin
H2 receptor blockers
Sulphonamides
Salicylates
Ciprofloxacin
100 HYPOGLYCAEMIA CAUSED BY INSULIN SECRETAGOGUES
Meglitinide analogues
The risk of hypoglycaemia associated with the recently introduced
meglitinide analogue class of rapid-acting secretagogues appears to
be lower than that observed with some sulphonylureas; this is
particularly relevant among patients with erratic meal patterns.
Repaglinide is a benzamido derivative that is taken with meals. It
has a short duration of action and does not stimulate insulin release
in the absence of glucose. If a meal is not taken, the corresponding
dose of repaglinide should be omitted. Repaglinide appears to be
safe in patients with mild to moderate renal impairment, although
caution is required. Nateglinide is an amino acid derivative that,
like repaglinide, is marketed as a prandial glucose regulator. The
rapid and relatively short-lived insulin secretion that these drugs
produce, in the presence of adequate b-cell reserve, reduces postprandial
glucose excursions; however, the meglitinides also lower
fasting plasma glucose concentrations, repaglinide being more
effective than nateglinide in this respect. Clinical experience with
these drugs is as yet limited. An interaction between repaglinide
and gemfibrozil has been reported with enhancement and
prolongation of repaglinide’s hypoglycaemic effect; this is thought
to be mediated through interference with the drug’s metabolism by
cytochrome P450 2C8. While the risk of hypoglycaemia with the
meglitinides appears to be relatively low, care is required to ensure
appropriate use in order to minimise this possibility, e.g. when
carbohydrate intake is reduced due to anorexia or vomiting.
Combining a meglitinide with a drug from a different class of antidiabetic
agents, e.g. metformin or a-glucosidase inhibitors, will
increase the risk of hypoglycaemia.
Repaginide and nateglinide are rapid-acting secretagogues
that may carry a relatively low risk of severe hypoglycaemia.
PATHOPHYSIOLOGY 101
Diagnosis
Clinical hypoglycaemia is characterised by Whipple’s triad (Table 5.4):
symptoms of hypoglycaemia
low blood glucose concentration
resolution of symptoms with the administration of carbohydrates.
Typically, symptoms of hypoglycaemia begin when plasma
glucose concentrations fall below 3.3 mmol/L in healthy nondiabetic
individuals. Central nervous system impairment occurs at
approximately 2.8 mmol/L. The symptom complex of hypoglycaemia
can be divided into neurogenic (or autonomic) and
neuroglycopenic symptoms (Table 5.5). The neurogenic symptoms
develop in response to counter-regulatory hormones that are
secreted in response to declining plasma glucose concentrations,
namely epinephrine, cortisol and glucagon. Neuroglycopenic
symptoms reflect an absolute central nervous system deficiency
of glucose substrate. Neurogenic symptoms typically arise earlier
and with relatively higher concentrations of glucose, while
neuroglycopenic symptoms occur with lower levels of glucose
(see Chapter 4).
Individuals with type 2 diabetes can have altered thresholds
for development of symptoms of hypoglycaemia (see Chapter 4).
Specifically, those individuals with chronically elevated blood
glucose may experience hypoglycemic symptoms at normal or
even elevated blood sugar levels compared with individuals with
Table 5.4 Whipple’s triad criteria for the diagnosis of hypoglycaemia
1. Symptoms of hypoglycaemia.
2. Low plasma glucose at the time of symptoms – measured using reliable
methodology.
3. Prompt resolution of symptoms with administration of carbohydrate.
102 HYPOGLYCAEMIA CAUSED BY INSULIN SECRETAGOGUES
normal glucose tolerance. On the other hand, individuals with
recurrent hypoglycaemia secondary to tight glucose control may
not experience symptoms of hypoglycaemia until blood glucose is
significantly decreased. This occurrence is probably a result of
adaptation by the brain to ambient glucose levels (see Chapter 4).
Thus, it might be more useful to think of hypoglycaemic symptoms
in terms of glycaemic threshold rather than absolute glucose
concentration. The glycaemic threshold differs between individuals
and is lowered by the occurrence of recent hypoglycaemic
episodes. As such, it is common for individuals with newly
diagnosed diabetes to experience symptoms of hypoglycaemia
shortly after they are started on pharmacotherapy for diabetes. In
other words, because of the acute lowering of blood glucose that
occurs with the institution of therapy, patients may develop
symptoms without overt hypoglycaemia in this setting. Other
patients with type 2 diabetes report hypoglycaemic symptoms
despite normal blood glucose concentrations that improve with the
ingestion of food. These symptoms are usually neurogenic in
nature and represent examples of subjective hypoglycaemia that do
not fully fulfil Whipple’s triad.
Clinically significant hypoglycaemia is poorly defined. Based on
the above information, we propose the following definition of
Table 5.5 Symptoms of hypoglycaemia
Neurogenic (autonomic) Neuroglycopenic
symptoms symptoms
Sweating Dizziness
Tremor Confusion
Anxiety Fatigue
Palpitations Weakness
Hunger Warmth
Tingling Headache
Difficulty speaking
Difficulty concentrating
Loss of consciousness
Seizure
DIAGNOSIS 103
hypoglycaemia in patients with type 2 diabetes:
1. plasma glucose levels < 3.3 mmol/L in conjunction with the
occurrence of neurogenic or neuroglycopenic symptoms of
hypoglycaemia or
2. any plasma glucose concentration below 2.8 mmol/L regardless
of symptoms.
Because the symptom complex of hypoglycaemia can be nonspecific,
confirmatory plasma glucose determination – using
reliable methodology – is essential to verify that symptoms are
truly attributable to hypoglycaemia and not to some other cause,
e.g. anxiety. Once a diagnosis of hypoglycaemia is established, the
aetiology of the hypoglycaemia should be ascertained. As shown in
Table 5.6, the causes of hypoglycaemia can be divided into two
main categories with differing aetiologies:
1. hyperinsulinaemic, e.g. sulphonylurea induced hypoglycaemia
Table 5.6 Classification of causes of hypoglycaemia
Hyperinsulinaemic hypoglycaemia Hypoinsulinaemic hypoglycaemia
Excessive insulin dose; mismatch to
requirements
Insulin antibodies; rare with modern
monocomponent and recombinant
insulin preparations
Sulphonylureas Reactive hypoglycaemia; uncommon
Insulinoma (may be inappropriate
non-suppression of insulin
secretion rather than overt
hyperinsulinaemia; proinsulin
excess
Short bowel syndrome
Mesothelial tumours; primary
hepatocellular carcinoma
Severe sepsis
Adrenal insufficiency
Wasting syndrome
Due to insulin-like growth factor-2 production.
104 HYPOGLYCAEMIA CAUSED BY INSULIN SECRETAGOGUES
2. hypoinsulinaemic, e.g. alcohol induced hypoglycaemia, Addisonian
crisis.
Most of these can be ruled out by a thorough medical history and
physical examination. It can sometimes be difficult, however, to
differentiate among the hyperinsulinaemic causes of hypoglycaemia,
and further laboratory testing is often required. In this
situation, appropriate laboratory testing includes measurement of
plasma concentrations of
insulin
C-peptide
sulphonylurea.
These need to be obtained during the occurrence of symptoms of
hypoglycaemia, wherever possible. C-peptide is a marker of
endogenous insulin secretion. It is produced from the islet b-cells
on an equimolar basis with insulin but is not cleared on first pass
metabolism through the liver, c.f. insulin, more than 50 per cent of
which does not reach the peripheral circulation. Plasma levels of Cpeptide
are increased by insulin secretagogues but are suppressed
by the administration of exogenous insulin; clearance is reduced in
renal impairment. Table 5.7 demonstrates the laboratory findings
on the various causes of hyperinsulinaemic hypoglycaemia.
Unfortunately, results of these tests are rarely immediately
available and clinical judgment is frequently required in emergency
management. Insulin auto-antibodies are common among
Table 5.7 Laboratory diagnosis of hyperinsulinaemic hypoglycaemia
Insulin Insulin Sulphonylurea
levels C-peptide antibodies levels
Exogenous insulin þþþ þ
Insulinoma þþþ þþþ
Sulphonylureas þþþ þþþ þþþ
N.B. Plasma insulin and/or C-peptide concentrations must be interpreted in
conjunction with the simultaneous plasma glucose concentration.
DIAGNOSIS 105
insulin treated patients; interpretation of their clinical relevance
may be problematic.
Because pharmacy error and accidental ingestion of sulphonylureas
are not uncommon causes of hypoglycaemia, it is prudent
that all medications that a patient is taking be checked for accuracy.
Additionally, it is imperative to review all medications in the home
in cases of hypoglycaemia in children, patients with limited
cognitive capacity, or suicide attempts.
Management
The goals of treatment of all forms of sulphonylurea induced
hypoglycaemia are
prompt restoration of euglycaemia
prevention of future episodes of hypoglycaemia.
There are several therapeutic modalities available to accomplish
these goals, including
elimination of drug from the gastrointestinal tract,
administration of glucose – oral or parenteral – to re-establish
euglycaemia
administration of agents that will attenuate the release of insulin
from the sensitized b-cells.
If a patient presents after a known intentional overdose of a large
number of sulphonylurea tablets in a suicide attempt, activated
charcoal should be administered. The role of activated charcoal in
accidental ingestion of one to two tablets (such as usually occurs in
children) is uncertain. In this setting, activated charcoal would be
expected to be useful if the ingestion occurred within one hour of
106 HYPOGLYCAEMIA CAUSED BY INSULIN SECRETAGOGUES
presentation. Induction of emesis is not recommended due to the
risk of central nervous system depression associated with
hypoglycaemia and the subsequent risk of aspiration of gastric
contents.
If hypoglycaemia is not initially present, serial monitoring of
blood glucose is essential. Blood glucose levels should be checked
every 1–2 h. Two factors must be considered in determining the
duration of observation needed following accidental or intentional
overdose:
1. onset of action of the drug
2. duration of effect.
Hypoglycaemia will typically occur within 8 h of ingestion in
accidental or intentional overdoses. Thus, patients in these
situations will require monitoring and observation for at least 8 h
after ingestion. If no hypoglycaemia occurs during this time period,
they may be released. Individuals with sulphonylurea induced
hypoglycaemia due to hepatic or renal failure may require longer
periods of observation, as the duration of drug effect may be
prolonged.
If patients remain euglycaemic, they should be allowed free
access to food and will not necessarily require intravenous
administration of glucose. A desirable outcome can be anticipated as
long as the blood glucose concentrations remain above 3.5 mmol/L. If
hypoglycaemia develops at any time following sulphonylurea
overdose, glucose administration is required.
Dextrose. An intravenous infusion of dextrose (5 or 10 per cent
solution) should be initiated and titrated to keep glucose
concentrations greater that 3.5 mmol/L. Supplemental boluses
of 25 or 50 per cent dextrose (20–30 mL, repeated as necessary)
may be required intermittently (flush cannula to reduce risk of
thrombophlebitis; Chapter 4). The patient should be admitted
for overnight observation and glucose administration. Caution
must be taken to avoid hyperglycemia with the administration
MANAGEMENT 107
of dextrose. As previously discussed, sulphonylureas have a
synergistic effect with hyperglycaemia on the islet b-cell and
may stimulate further insulin release and lead to persistent or
rebound hypoglycaemia. The goal is to maintain plasma glucose
concentrations between 3.5 and 4.5 mmol/L.
Glucagon. If intravenous glucose is not readily available, 1 mg
glucagon can be administered as a subcutaneous or intramuscular
injection (see Chapter 4). Glucagon can also be used in the
setting of hypoglycemic seizures. Glucagon usually results in
symptomatic improvement within 10–20 min. The duration of
benefit of glucagon is relatively short, lasting only 60–120 min.
Because sulphonylurea induced hypoglycaemia can be prolonged
in the setting of drugs with prolonged half-lives,
decreased drug clearance (i.e. renal or hepatic impairment) or
drug interactions, glucagon by itself is inadequate for the
treatment of hypoglycaemia; dextrose administration is also
required. Glucagon may be ineffective when hepatic glycogen
stores are depleted, e.g. in the alcoholic patient. Some
authorities caution against use of glucagon in sulphonylurea
induced hypoglycaemia, since further insulin secretion may be
stimulated.
Diazoxide. This is a non-diuretic vasodilator, has been used in
the treatment of sulphonylurea induced hypoglycaemia. This
drug is an adenosine trisphosphate-sensitive potassium channel
opener, which counteracts the effects of sulphonylureas on islet
b-cells and will theoretically prevent insulin release. Onset of
action is immediate after intravenous administration, and the
effect lasts approximately 8 h. The drug must be administered
in a separate intravenous line because it can precipitate in the
presence of glucose. Unfortunately, side effects are common
and include hypotension, tachycardia, nausea, vomiting and
dizziness. Due to the side-effect profile, which is particularly
undesirable in elderly patients with overt or subclinical
cardiovascular disease and/or impaired autonomic reflexes,
108 HYPOGLYCAEMIA CAUSED BY INSULIN SECRETAGOGUES
this agent is considered second line for the treatment of
sulphonylurea overdose.
Octreotide. This, as an adjunct to intravenous dextrose, is
the drug of choice for the treatment of sulphonylurea
induced hypoglycaemia. Octreotide is a somatostatin analogue
that potently inhibits the secretion of several hormones
including insulin, glucagon, growth hormone and gastrin.
Octreotide binds to the somatostatin sub-type 2 receptor on
the islet b-cells, inhibiting calcium channel opening, which
decreases calcium flux; this results in greatly reduced insulin
release, even in the presence of insulin secretogogues. Octreotide
appears to be relatively safe in the acute setting. Side effects
are mild, including nausea, headache, diarrhoea, fat malabsorption
and discomfort at the injection site. Octreotide can be
administered subcutaneously or intravenously. The onset of
action is rapid, with a tissue distribution of 12 min; the half-life
of octreotide is 1.5 h. Octreotide effectively ameliorates
hypoglycaemia associated with sulphonylurea overdose
and often decreases the need for exogenous dextrose infusion.
These attributes may especially benefit patients with
congestive heart failure in whom fluid volume overload is a
concern. In the clinical setting, octreotide administration
effectively reduces the occurrence of rebound hypoglycaemia
following glucose infusion and may shorten hospital stays.
Octreotide may be administered in doses of 50 mg subcutaneously
or intravenously every 6–8 h as needed until
hypoglycaemia resolves. An algorithm for the evaluation and
treatment of sulphonylurea induced hypoglycaemia is presented
in Figure 5.2.
All accidental ingestions of sulphonylureas in children require
evaluation and observation in the emergency room. Studies suggest
that if hypoglycaemia does not develop within 8 h of ingestion,
then the risk of subsequent hypoglycaemia is low and the
child may be released. If, however, hypoglycaemia occurs, the
child should be admitted for continued observation and
MANAGEMENT 109
treatment for at least 24 h to avoid relapse due to rebound
hypoglycaemia.
Further reading
Burge MR, Schmitz-Fiorentino K, Fischette C, Qualls CR, and Schade DS.
A prospective trial of risk factors for sulphonylurea-induced hypoglycaemia
in type 2 diabetes mellitus. JAMA 1998; 279: 137–143.
Burge MR, Sobhy TA, Qualls CR, and Schade DS. Effect of short term
glucose control on glycemic thresholds for epinephrine and hypoglycemic
symptoms. J Clin Endocrinol Metab 2001; 86: 5471–5478.
Figure 5.2 Clinical care algorithm for the evaluation and treatment of
sulphonylurea-induced hypoglycaemia
110 HYPOGLYCAEMIA CAUSED BY INSULIN SECRETAGOGUES
Boyle, PJ, Justice, K, Krentz, AJ, Nagy RJ, and Schade DS. Octreotide
reverses hyperinsulinemia and prevents hypoglycemia induced by
sulphonylurea overdoses. J Clin Endocrinol Metab 1993; 76: 752–756.
Herbel G and Boyle, PJ. Hypoglycaemia pathophysiology and treatment.
Endocrinol Metab Clin North Am 2000; 29: 725–743.
Jennings AM, Wilson RM, and Ward JD. Symptomatic hypoglycaemia in
NIDDM patients treated with oral hypoglycaemic agents. Diabetes Care
1989; 12: 203–208.
Seltzer HS. Drug induced hypoglycaemia: a review of 1418 cases.
Endocrinol Metab Clin North Am 1989; 18: 163–183.
Shorr RI, Ray WA, Daugherty JR, and Griffin MR. Incidence and risk
factors of serious hypoglycaemia in older persons using insulin or
sulphonylureas. Arch Intern Med 1997; 57: 1681–1686.
Stahl M and Berger W. Higher incidence of severe hypoglycaemia leading
to hospital admission in type 2 diabetic patients treated with longacting
versus short-acting sulphonylureas. Diabet Med 1999; 16: 586–
590.
FURTHER READING 111
6
Lactic Acidosis
in Diabetes
Jean-Daniel Lalau
Summary
The classical view of lactic acid is that (1) it may be responsible
for metabolic acidosis, mainly due to anoxia or ischaemia, to
which diabetic subjects are particularly prone, and (2) such
lactic acidosis is associated with poor prognosis. Indeed, blood
lactate concentration is one of the best predictors of fatal
outcome in critical illness. This observation contrasts with the
fact that lactate is both a gluconeogenic substrate and easily
oxidised, an apparent contradiction raising the question of
whether excess lactate is deleterious, or possibly beneficial.
Current data are consistent with the notion that lactate
production and related metabolic acidosis due to the stimulation
of the anaerobic metabolism might be an adapted protective
response. Diabetic subjects can also develop lactic acidosis due
to accumulation of biguanides such as metformin. In clinical
practice, this is typically referred to as ‘metformin associated
Emergencies in Diabetes Edited by Andrew J. Krentz
# 2004 John Wiley & Sons, Ltd ISBN 0-471-49814-9
lactic acidosis’. This term is, however, confusing, because
metformin use may be either the cause of lactic acidosis, for
example, when renal failure leads to accumulation, but also,
and more frequently, coincidental.
This raises two more questions. Regarding not only metabolic
disorders, but also, and more importantly, clinical outcome, is
metformin per se toxic? What role might metformin play in
patients who develop lactic acidosis independently of, but
coincidental to, administration of the drug? A recent insight is
that no mortality is attributable to metformin alone, and that in
the true metformin associated – as distinct from metformin
induced – lactic acidosis, i.e., when both metformin and the
associated disorder contribute to lactic acidosis, the metabolic
and vascular effects of metformin may even confer some
protection.
Lactic acidosis in clinical practice
Lactic acidosis is the most frequent cause of metabolic acidosis,
with a prevalence of about one per cent among adult hospitalised
patients. Lactic acidosis is of relevance to diabetic patients in two
respects, one aspect being specific to diabetes, whereas the other is
not. In the former category, causes of lactic acidosis include
biguanide therapy, i.e. metformin accumulation. In the latter
category, disorders of acid–base equilibrium are common in
patients with critical illness, and critical illnesses are more common
in diabetic patients than in non-diabetic subjects. Consequently,
one should consider whether there is a difference between
1. lactic acidosis not related causally related to biguanide therapy
2. lactic acidosis induced by biguanides.
This distinction is relevant not only in terms of incidence but also,
and more importantly, in terms of clinical outcome. Clarification of
114 LACTIC ACIDOSIS IN DIABETES
this question is relevant the management of overweight diabetic
patients, in whom metformin has been recommended as first-line
pharmacological therapy.
Lactic acidosis is the most common cause of metabolic
acidosis among hospitalised patients.
Following the classification of lactic acidosis by Cohen and Woods
(1976) according to the presence or absence of adequate tissue
oxygenation (Box 6.1), several excellent reviews have provided
background knowledge on the diagnosis, clinical presentation,
pathogenesis and management of lactic acidosis. We will therefore
limit reconsideration of classical information, and focus on new
directions.
Box 6.1 Distinctions among types and related causes of
lactic acidosis
Classical distinction.
Lactic acidosis has been classically divided into either type A
(anaerobic) or type B (aerobic).
Type A is due to tissue hypoperfusion with reduced
arterial oxygen content.
Type B is due to a defect in energy metabolism independent
of hypoxia. It includes common disorders (such as
sepsis, hepatic failure, renal failure, and cancer), drugs or
toxins (such as biguanides, ethanol, salicylates, methanol,
ethylene glycol and niacin) and other conditions
(such as strenuous muscular exercise, grand mal seizures
and D-lactic acidosis).
Criticism of this classical distinction.
The above distinction is considered obsolete, given that
restricted oxygen supply and metabolic factors often operate
simultaneously.
LACTIC ACIDOSIS IN CLINICAL PRACTICE 115
Lactic acidosis independent of biguanides
Diagnosis
The definition of lactic acidosis is arbitrary (Table 6.1). In our
opinion, reports in the literature and cluster analyses of laboratory
findings in patients admitted to hospital for internal diseases
support the use of this definition of lactic acidosis.
Pathogenesis
While hyperlactataemia refers to an increase in blood lactate, lactic
acidosis indicates accumulation of both lactate and hydrogen (H
þ
)
ions (protons). In fact, the formation of lactate from glucose neither
consumes nor generates H
þ
. Although the metabolism of one
molecule of glucose leads to the production of two protons, both go
into the formation of lactate, such that:
glucose þ 2 ADP þ 2 Pi ! 2 lactate þ 2 ATP
ADP ¼ adenosine diphosphate
ATP ¼ adenosine triphosphate
Pi
¼ inorganic phosphate.
It is actually the degradation of the ATP obtained that may cause
excess formation of H
þ
when tissue hypoxia hampers the recycling
Table 6.1 Definition of lactic acidosis
Arterial whole blood lactate concentration > 5 mmol/L
Arterial pH 7.35
Note that lactic acidosis may co-exist with other causes of metabolic
acidosis, e.g. diabetic ketoacidosis.
116 LACTIC ACIDOSIS IN DIABETES
of ATP from its metabolites:
ATP ! ADP þ Pi þ H
þ þ energy
Lactate production may also lead to metabolic acidosis in another
manner, via the lactate/hydroxide (OH

) exchange mechanism.
The formation of OH

from extracellular H2O leads to entry of
OH

into the cell, where it prevents lowering of pH. Hydrogen
ions are released in the extracellular space at the same time.
However, pH does not fall if compensatory hyperventilation is
sufficiently effective.
Theoretically, intracellular acidosis should exacerbate overproduction
of lactate because the lactate/pyruvate ratio depends on
the ratio of reduced to oxidized nicotinamide adenine dinucleotides:
½NADH ½H
þ =½NAD
This influence is, however, less marked than the strong negative
effect of intracellular acidosis on phosphofructokinase activity, an
action that acts as a protective mechanism during hypoxia by
sparing glucose and, thereby, preventing the overproduction of
protons originating from ATP hydrolysis. In addition, acidosis may
improve tissue oxygen extraction, reflected in a shift of the
oxyhaemoglobin curve to the right.
Prognosis
Lactic acidosis is still considered to have a high mortality rate.
Indeed, blood lactate concentration is well recognised as one of the
best predictors of a fatal issue in critically ill patients. Studies
carried out in the 1970s showed a mortality rate of approximately
80 per cent for patients with lactate levels 5 mmol/L. This
relationship has changed. In the 1990s, we observed the same
mortality rate of 80 per cent for a fourfold higher lactate level.
However, more important than estimating the isolated prognostic
value of blood lactate concentration is determining what has
LACTIC ACIDOSIS INDEPENDENT OF BIGUANIDES 117
caused lactate to increase in specific clinical situations. Because
lactate may be metabolised and oxidised by most cells, the essential
questions regarding the relationship between lactate and a patient’s
critical status are, in fact, whether
1. a high blood lactate level is the cause or the consequence of the
severity of the metabolic disorder
2. the increase in lactate is a rescue mechanism that should be
preserved.
Instead of being toxic, lactate may act as a preferred substrate for
aerobic energy production during the initial stages of recovery
from cerebral ischaemia or hypoxia (Box 6.2).
Box 6.2 Hypothesis for an adaptive role of the increase in
blood lactate
A shuttle for energy metabolism between tissues. When
lactate is excreted by one tissue but oxidised in another,
the latter tissue can be considered to carry out respiration
for the former. When lactate is produced in excess, the
higher the lactate turnover, the higher the energy
metabolism of these producing tissues supported by the
other tissues.
A role of sparing glucose metabolism. The higher the
tissue lactate level, the more glucose is spared for tissues
with particular priorities (e.g. heart tissue).
A substrate for specific cellular function. Lactate is an
important determinant of cellular ATPase enzyme activities.
This idea is supported by the observation that lactate administration
can prevent cerebral dysfunction during hypoglycaemia. The
118 LACTIC ACIDOSIS IN DIABETES
following mechanism has been proposed to explain these observations.
During hypoxia, the ATP/ADP ratio is very low. When
oxygen is restored to normal levels, lactate oxidation into pyruvate
predominates over glucose oxidation, which necessitates preliminary
phosphorylation to glucose-6-phosphate, an energy requiring
process. In addition, for patients in critical condition, lactate can be
viewed as a major metabolic substrate, per se or via its impact on
glucose metabolism.
The view of lactate as an exclusively toxic molecule can be
challenged.
Management
Vasodilators. Drugs such as nitroprusside, which improve tissue
perfusion, have been reported both to ameliorate and to
precipitate lactic acidosis; their use has never been evaluated
in a controlled clinical trial.
Vasoconstrictors. The use of vasoconstrictive drugs, such as
norepinephrine, is complicated by the potential of these agents
to exacerbate ischaemia in tissues, such as skeletal muscle and
liver, that are important in maintaining lactate homeostasis.
Drugs that act principally by increasing cardiac contractility
have not been rigorously evaluated in patients with lactic
acidosis, but their utility may be mitigated by the negative
inotropic action of acidemia per se. Despite decades of use,
intravenous sodium bicarbonate has never been demonstrated
to reduce morbidity or mortality in lactic acidosis
and it may even be deleterious in some patients. Sodium
bicarbonate is associated with several potential negative effects
including
LACTIC ACIDOSIS INDEPENDENT OF BIGUANIDES 119
1. decline in intracellular and cerebrospinal fluid pH,
2. exacerbation of tissue hypoxia,
3. circulatory congestion,
4. hypernatraemia and,
5. hyperosmolarity.
Why this paradox? The explanation proposed – perhaps overly
simplistic – is that CO2 originating from sodium bicarbonate
(NaHCO3) diffuses more rapidly into the cells than its precursor.
Pyruvate dehydrogenase activation. Finally, there has been much
interest in the use of sodium dichloroacetate, which is an
activator of pyruvate dehydrogenase. Consequently, dichloroacetate
tends to eliminate lactate through the oxidative metabolic
pathway. Nonetheless, although dichloroacetate has been
shown to significantly decrease arterial lactate concentration
and to significantly increase arterial pH in patients with
severe lactic acidosis in a placebo controlled trial, this was
not accompanied by improvement in haemodynamics or
survival.
Other measures. The efficacy of Carbicarb (an equimolar mixture
of sodium bicarbonate and sodium carbonate) has not been
confirmed in humans with lactic acidosis.
The above findings prompt us to concentrate on treating the
precipitating cause of lactic acidosis and to optimize ventilation to
compensate for metabolic acidosis. If lactate is a good metabolic
substrate, hyperlactataemia should, in fact, be preserved (Box 6.3).
Following this line of reasoning, the infusion of sodium lactate has
been even suggested. Because lactate is both a strong and a
metabolizing anion, sodium lactate provides a means of infusing
sodium and decreasing the concentration of protons, which are
removed along with lactate.
120 LACTIC ACIDOSIS IN DIABETES
Box 6.3 Lactic acidosis: conclusions for new insights
The classical view.
Lactic acidosis is
the cause of metabolic acidosis,
primarily related to anoxia or ischaemia,
and consequently associated with a poor prognosis.
From a biochemical point of view:
rather than being toxic, lactate is considered a favourable
metabolic substrate,
lactate overproduction and acidosis may be a protective
adapted response,
management should, therefore, focus on treatment of the
precipitating cause.
Lactic acidosis and biguanide therapy
Biguanides have enjoyed many years of use as oral antihyperglycaemic
agents. Phenformin, however, was withdrawn in
the late 1970s in most countries because of its association with a
high incidence of lactic acidosis. Metformin has also been linked
with this metabolic disorder, although less frequently (Box 6.4).
This difference between the two biguanides warrants reflection. Is
it simply limited to the incidence of associated lactic acidosis, or is
the different frequency attributable to a fundamental difference in
the nature of these agents?
LACTIC ACIDOSIS AND BIGUANIDE THERAPY 121
Box 6.4 Incidence of lactic acidosis in metformin therapy
The estimated rate is 1–9 cases per 100 000 person
years, i.e. 10–20 times lower than that seen with
phenformin.
However, as metformin may either cause lactic acidosis or
occur concomitantly with it, rates of lactic acidosis should
be compared between users and non-users of metformin.
This comparison was performed in a study of the US
market before and after the introduction of metformin.
This study found no distinguishable difference.
The above finding suggests a coincidental rather than causal
relationship between metformin and lactic acidosis. However,
given that metformin accumulation may lead to lactic acidosis, it
would seem important to determine whether concurrent use of
metformin, with or without accumulation, contributes to the course
of coincidental lactic acidosis, especially with regard to outcome.
Relationship between metformin and
lactic acidosis
The study of this relationship requires knowing whether or not
metformin has accumulated. The best way of determining this is to
measure the drug concentration in plasma, a measurement rarely
performed in the literature (Box 6.5).
As a consequence of the high clearance of metformin under
normal circumstances, and with the exception of intoxication with
metformin, high plasma metformin concentrations imply both
defective elimination and continuation of therapy. However,
because the assay of metformin is not readily available, a fortiori
in an emergency context, one may estimate the risk and extent of
metformin accumulation based upon its pharmacokinetic characteristics
(Table 6.2), the status and course of renal function, and
122 LACTIC ACIDOSIS IN DIABETES
dosage and time of last metformin administration. It is also
possible to measure metformin concentration in red blood cells,
where the elimination is far slower than that in plasma, thereby
providing better retrospective information.
Although a low blood metformin concentration can rule out the
drug as the cause of lactic acidosis in specific cases, the measurement
of metformin concentration does not provide a reliable
threshold of metformin accumulation and related disorders. This is
because high plasma metformin concentrations are not necessarily
Table 6.2 Pharmacokinetic characteristics of metformin
Plasma half-life: from 1.5 to 4.9 h
Metabolism: not detectable – excreted unchanged
Elimination: rapid renal elimination involving glomerular filtration and
tubular secretion with a clearance of four to five times that of creatinine
Approximately 90% of the ingested dose is eliminated in 12 h
Box 6.5 An illustration of the importance of measuring
plasma metformin concentration
This is the case of a patient with
anuria (serum creatinine level of 350 mmol/L)
lactic acidosis (lactate 16.3 mmol/L, pH 7.09)
no interruption of metformin therapy.
A high plasma metformin concentration would have been
expected, with the conclusion that the occurrence of lactic
acidosis was secondary to drug accumulation. In fact, the
plasma metformin concentration was within the therapeutic
range. What else could explain the genesis of lactic acidosis?
The occurrence of renal failure was recent and secondary to
cardiogenic shock, which was the actual cause of lactic
acidosis in this patient.
LACTIC ACIDOSIS AND BIGUANIDE THERAPY 123
complicated by lactic acidosis. This means that associated factors
necessarily play a role, which may be evident in the case of patent
disease processes, or less obvious, as in cases of a latent defect in
energy metabolism. This ultimately calls for careful analysis of the
history as well as the clinical and laboratory picture (Box 6.6).
Box 6.6 Questions to address when searching for a link
between metformin and lactic acidosis
Has metformin accumulated?
Are relevant associated disease processes present?
If present, is organ failure primary, or secondary to a
shock syndrome (e.g. renal failure)?
Are metformin accumulation and associated disorders
underlying factors, precipitating factors, or both?
Metformin accumulation may be either
a precipitating factor, as in metformin overdose or acute renal
failure (when there has been no discontinuation of metformin
therapy)
an underlying factor, as in chronic renal failure.
Similarly, system failures may be either precipitating factors, as in
sepsis or in haemorrhage, or underlying conditions, as in chronic
organ failure. Taking such factors into consideration should help in
estimating the prognosis.
Prognosis
Lactic acidosis associated with metformin therapy is typically
associated with a mortality rate of around 50 per cent, similar to
that reported with phenformin. However, since lactic acidosis in
critically ill patients also carries a poor prognosis, it is important to
124 LACTIC ACIDOSIS IN DIABETES
clarify whether poor outcome is due to metformin alone, to the
associated disorders, or both. To clarify this issue, the outcome of
lactic acidosis in metformin treated patients in whom plasma
metformin concentrations were available has been compared with
the outcome in patients who did not receive metformin. The
prognosis was better in patients treated with metformin, even
though the observations were earlier and, on average, the patients
were older, were selected on the ground of lactic acidosis and not
only hyperlactaemia and, more importantly, had a median lactate
level almost twice as high as the metformin untreated patients. In
the metformin-treated patients, there was no relevant difference in
median lactate level between those who survived and those who
died. There was even unexpected survival in many who had severe
lactic acidosis (with a lactate level up to 35.5 mmol/L) and
circulatory shock. It was concluded that prognosis of lactic acidosis
is independent of lactate level in metformin treated patients.
In the same manner as for lactate, if metformin were toxic,
plasma metformin concentrations would have been higher in the
patients with poor prognosis. It appeared, on the contrary, that the
majority of patients with plasma metformin at the therapeutic level
or even lower had the poorest prognosis while the majority of
patients with high plasma metformin levels survived (Figure 6.1). If
0
20
40
60
80
100
<1 mg/L 1 _ 5 mg/L >5 mg/L
Figure 6.1 Lactic acidosis in 49 metformin treated patients with plasma
metformin concentration available: mortality (%) according to plasma metformin
concentrations (<1 mg/L, therapeutic or low; 1–5 mg/L, moderately
increased; >5 mg/L, markedly increased) (Lalau, unpublished data)
LACTIC ACIDOSIS AND BIGUANIDE THERAPY 125
not lactate or metformin, what might account for a fatal outcome?
The observation that most patients had at least one additional risk
factor for lactic acidosis supports the hypothesis that concurrent
diseases are likely to determine the outcome. Another point to
consider is the difference between metformin, which has favourable
metabolic and vascular effects, and phenformin. Although
these two biguanides belong to the same family, they are
structurally distinct, and the structural differences lead to
differences in effects (Table 6.3).
The relatively good prognosis of metformin treated patients with
high lactate levels may not be surprising. Indeed, as already stated,
lactate may be metabolically advantageous. Bearing in mind that
metformin treated patients with lactic acidosis may have very high
lactate levels, the proportion of hyperlactataemia related to
metformin with respect to that of an associated disease process
may be relevant. Another reason may lie in various vascular
properties of metformin (Table 6.4), many of which are unique,
Table 6.3 Differences between phenformin and metformin
Structural differences:
phenformin has a long lipophilic side chain that mediates its binding to
mitochondrial membranes, where it inhibits the major pathways of lactate
disposal: gluconeogenesis and oxidation
metformin has two small moieties that confer much less lipophilicity and,
consequently, no marked inhibiting effect on oxidative phosphorylation.
Pharmacological and metabolic differences:
phenformin is metabolised in the liver, and is associated with a well
defined hyperlactataemic effect involving increased release from skeletal
muscle and inhibition of oxidation
metformin does not undergo metabolic transformation, or influence lactate
turnover or oxidation. Its hyperlactataemic effect – actually minimal at
the recommended dosage – originates from the splanchnic bed via lactate
production by the small intestine after meals and/or defective lactate
uptake by liver cells.
126 LACTIC ACIDOSIS IN DIABETES
accounting for protection under ischaemic conditions. These
vascular effects sharply contrast with those of phenformin, which
has been shown to decrease cardiac output. While these differences
provide a theoretical basis for the hypothesis that metformin may
have beneficial effects in patients with lactic acidosis, this
hypothesis has not been tested in a clinical trial.
Role of dialysis
Haemodialysis is classically considered to be the most efficient
method, providing both symptomatic and aetiological treatment by
eliminating lactate and metformin. This is actually a misconception.
Lactate elimination cannot participate in recovery of acid–base
balance since lactate per se is not an acid generating substance.
Instead, the excess protons from hydrolysis of ATP during
anaerobic glycolysis tend to be removed by endogenous buffers,
which are regenerated through lactate metabolism. In addition, as
already stated, lactate can be viewed as an energetic substance, and
metformin is held to have beneficial metabolic and vascular effects.
Thus, even if metformin can be readily dialysed, haemodialysis
should only be considered for correction of blood volume and
osmolarity in patients with anuria.
Table 6.4 Vascular effects of metformin
Macroangiopathy: reduction in atherosclerosis
reduction in thrombosis
Microangiopathy: improvement in haemorheology
increase in nutritive blood flow
reduction in vessel permeability
Angiogenesis: reduction in neovascularisation
Haemostasis: increase in fibrinolysis
Oxidative stress: reduction in oxidative stress
Protein glycation: reduction in protein glycation.
LACTIC ACIDOSIS AND BIGUANIDE THERAPY 127
A critical analysis of the relationship between
metformin and lactic acidosis
Because of the pitfalls listed in Box 6.7, considerations of the link
between metformin and lactic acidosis in the literature are sometimes
flawed. This was reflected in an analysis of 26 consecutive
case-reports of so-called ‘metformin associated lactic acidosis’, in
which lactic acidosis was in fact absent in four cases, not
precipitated by metformin in another eight and of uncertain origin
in two, leaving only 12 that were, in our judgment, precipitated by
metformin.
Box 6.7 Summary of the pitfalls in determining the
relationship between metformin and lactic acidosis.
Overestimation of the prognostic significance of hyperlactataemia
independent of metformin.
Overestimation of the prognostic significance of metformin
induced hyperlactataemia.
Failure to measure metformin concentration in plasma.
Failure to consider plasma metformin concentrations
generating hyperlactataemia.
Failure to consider prerequisites for the development of
lactic acidosis.
Failure to consider the prognostic significance of associated
disorders.
Failure to distinguish the effects of metformin from those
of phenformin.
Failure to account for different scenarios with distinct
prognoses.
128 LACTIC ACIDOSIS IN DIABETES
Conclusions
Use of the term ‘metformin associated lactic acidosis’, which
commonly refers to all situations of lactic acidosis in patients
receiving metformin therapy, is potentially confusing with regard
to both pathophysiology and prognosis. Strictly speaking, this term
should refer to metformin and concurrent disease processes as coprecipitating
factors of lactic acidosis. Study of the link between
metformin and lactic acidosis should instead lead to the distinction
of different scenarios (Figure 6.2; Box 6.8). In true metformin
associated lactic acidosis, for a given lactate level, the higher the
Box 6.8 Lactic acidosis during metformin therapy: new
insights
The classical view of ‘metformin associated lactic acidosis’ is
based on purely epidemiological data, which indicates an
overall mean mortality rate of approximately 50 per cent.
It is more pertinent to replace this overall viewpoint by
distinct clinical scenarios.
It is no longer acceptable to consider the association of lactic
acidosis and metformin use in terms of mean mortality rates.
Newly coined terms corresponding to entities of differing
prognoses have greater clinical relevance:
metformin unrelated lactic acidosis (without metformin
accumulation): the prognosis is that of a serious,
frequently life threatening underlying condition
metformin induced lactic acidosis: the prognosis is excellent
because metformin per se is not a toxic substance
metformin associated lactic acidosis, with metformin
accumulation and concurrent factors: the prognosis is
intermediate, depending upon the severity of the underlying
and precipitating factors.
LACTIC ACIDOSIS AND BIGUANIDE THERAPY 129
No metformin therapy Metformin accumulation Metformin accumulation
System failure, sepsis, etc. No associated pathology Associated pathology(ies)
Metformin unrelated
lactic acidosis
Metformin induced
lactic acidosis
Metformin associated
lactic acidosis
Precipitating factor:
acute metformin
accumulation
Precipitating factor:
associated
pathology(ies)
Metformin induced
lactic acidosis with
associated factor
Metformin unrelated
lactic acidosis with
metformin accumulation
Mortality + Mortality
Mortality
unless severe
associated factor
Mortality ±
(lower than without
metformin treatment?)
_
_
Figure 6.2 Lactic acidosis in metformin therapy: different scenarios and their prognosis
degree of metformin accumulation, the lower the influence of the
concurrent disease, possibly the higher the protective effect of
metformin, and, ultimately, the better the prognosis.
It is clear from hospital and community based surveys that many
patients receive metformin uneventfully even in the presence of
traditional contraindications. Until such time that the contraindications
to metformin are revised, the conditions listed in
Table 6.5 provide a framework within which the drug should be
used. The main contraindications were rigorously reaffirmed when
metformin was introduced in the US in the mid-1990s. It remains
possible that the low incidence of lactic acidosis among metformintreated
patients reflects, at least to some extent, observance of the
main contraindications. There is scope for controlled clinical trials
to clarify the contentious issues concerning the traditionally
perceived safety record of this drug.
Further reading
Cohen R and Woods HF. The clinical presentations and classifications of
lactic acidosis. In: Clinical and Biochemical Aspects of Lactic Acidosis,
R Cohen and HF Woods, (Eds.) 1976, Blackwell: Boston. pp. 40–52.
Table 6.5 Suggested revised contraindications and guidelines for
withdrawing metformin
Discontinue if plasma creatinine becomes elevated
Withdraw during periods of tissue hypoxia, e.g. acute myocardial
infarction, sepsis
Withdraw for 3 days after i.v. contrast media is administered, and when
normal renal function has been confirmed
Withdraw 2 days before major surgery – use insulin if necessary – and
reinstate only when renal function is normal and hypoxia, hypovolaemia,
hypotension and sepsis have resolved.
Modified from Jones CG, Macklin JP and Alexander WD. Contraindications to the use of
metformin. Br Med J 2003; 326: 4–5.
FURTHER READING 131
Frayn KN. Metabolic Regulation. A human perspective. 2nd edition. 2003,
Blackwell: Oxford. p. 339.
Fulop M and Hoberman H. Phenformin-associated metabolic acidosis.
Diabetes 1976; 25: 292–296.
Lalau J, Lacroix C, Compagnon P et al. Role of metformin accumulation in
metformin-associated lactic acidosis. Diabetes Care 1995. 18: 779–784.
Lalau J, Race J. Metformin and lactic acidosis in diabetic humans. Diabetes
Obes Metab 2000. 2: 131–137.
Lalau J, Race J. Lactic acidosis in metformin therapy: searching for a link
with metformin in reports of ‘metformin-associated lactic acidosis’.
Diabetes Obes Metab 2001. 3: 195–201.
Leverve X. Energy metabolism in critically ill patients: lactate is a major
oxidizable substrate. Curr Opin Clin Nutr Metab Care 1999; 2: 165–169.
Scheen A. Clinical pharmacokinetics of metformin. Clin Pharmacol 1996;
30: 359–371.
Stacpoole P. Lactic acidosis. Endocrinol Metabol Clin North Am 1993. 22:
221–245.
Vincent J. Lactate levels in critically ill patients. Acta Anaesthesiol Scand
1995. 39 (suppl. 107): 261–266.
Wiernsperger NF. Metformin: intrinsic vasculoprotective properties.
Diabetes Technol Ther 2000. 2(2): 259–272.
132 LACTIC ACIDOSIS IN DIABETES
7
Management of
Diabetes during
Surgery, Myocardial
Infarction and Labour
Aftab M Ahmad and Jiten P Vora
Summary
Surgery
Patients with diabetes are at increased risk of elective and
emergency surgery, often as a consequence of complications
such as advanced foot disease and the manifestations of
atherosclerosis. Surgery stimulates the release of hormones
and cytokines that antagonise insulin action and suppress
insulin secretion. In patients with diabetes, this may result in
hyperglycaemia and ketosis, depending on the degree of
surgical trauma; even laparoscopic abdominal surgery can elicit
a significant hormonal response. Metabolic decompensation
increases morbidity in the diabetic patient via electrolyte losses,
impaired wound healing and less effective cellular responses to
Emergencies in Diabetes Edited by Andrew J. Krentz
# 2004 John Wiley & Sons, Ltd ISBN 0-471-49814-9
sepsis. Inappropriate treatment with insulin or sulphonylureas
may cause hypoglycaemia in diabetic patients fasted for
surgery.
Intensive control of hyperglycaemia is necessary in high-risk
patients in the post-operative period. Simple guidelines and
good pre-operative planning are essential for successful management
of surgery in diabetes. Patients with either type 1
diabetes, type 2 diabetes treated with insulin or patients with
poor antecedent glycaemic control undergoing major surgery
should receive a dextrose and insulin infusion; potassium
should be added as required. This can be delivered as a
combined infusion (glucose–insulin–potassium; GKI). More
flexibility is derived by infusing insulin and dextrose
(þpotassium) via separate lines; the insulin infusion rate is
adjusted according to frequent near-patient blood glucose
monitoring.
For patients with diet or tablet treated type 2 diabetes, longacting
sulphonylureas such as glibenclamide should be changed
to shorter-acting agents – or insulin if indicated – a few days
prior to surgery; this will reduce the risk of hypoglycaemia, the
clinical features of which are obscured by general anaesthesia.
Patients with well controlled type 2 diabetes undergoing minor
surgery should have capillary blood glucose concentrations
monitored every 2 hours pre-operatively and during surgery;
treatment is re-instituted with the first post-operative meal.
Metformin should be avoided for all but minor surgical
procedures because of the concerns about lactic acidosis.
Patients undergoing open-heart surgery with cardio-pulmonary
bypass require higher doses of insulin, partly because of the use
of high volumes of glucose containing fluids.
Myocardial infarction
Coronary artery disease is the most common cause of death in
patients with type 2 diabetes and is an important cause in
134 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
premature mortality in type 1 diabetes. Diabetic patients tend
to have more severe and widespread atherosclerosis than age
matched non-diabetic controls. Acute myocardial infarction has
a higher immediate and delayed mortality than in non-diabetic
individuals, cardiac failure and re-infarction being the main
causes of death. The excess mortality among patients with
diabetes has persisted with the introduction of effective
treatment such as thrombolysis. However, diabetic patients
derive benefits from such therapeutic interventions that are
similar to or greater than those observed in non-diabetic
patients; this reflects the higher absolute risk conferred by
diabetes. Prevention of coronary heart disease, and other sequelae
of atherosclerosis, is a high priority in patients with diabetes.
In a randomised clinical trial, the Diabetes Mellitus Insulin,
Glucose Infusion in Acute Myocardial Infarction (DIGAMI)
investigators used an insulin þ dextrose infusion to achieve
tight glycaemic control for the initial 24 hours on the coronary
care unit in patients with an admission blood glucose
>11.1 mmol/L; this was followed by multiple daily injections
of subcutaneous insulin for at least 3 months. A control group
received insulin according to clinical indications, a significant
difference between the groups in the improvement in glycated
haemoglobin levels becoming evident. Use of other treatment
such as thrombolysis, aspirin and cardio-selective b-blockers
was similar between the groups. Mortality was significantly
reduced by 28 per cent ( p < 0.01) in the intensive treatment
group, with a reduction in absolute risk of 11 per cent at
3.5 years; this translates into 11 patients treated with
the intensive insulin regimen to save one life. Intriguingly,
benefit was most apparent in a pre-defined subgroup of insulinnaı
¨ve patients, perceived as being at lower risk of mortality.
Further evidence concerning the relative contributions of
the early insulin–dextrose infusion versus subsequent insulin
therapy is awaited from the DIGAMI 2 study. The latter trial
may also determine whether twice-daily insulin is an effective
alternative to multiple daily injections.
The cardiovascular safety of sulphonylureas, particularly agents
that bind to sulphonylurea (SUR) receptors in cardiac and
SUMMARY 135
vascular tissues, has been an issue of controversy for decades.
While no firm conclusion has been reached, some authorities
recommend avoidance of drugs such as glibenclamide that can
impair the phenomenonof ischaemic preconditioning(seebelow).
Labour
Diabetic women have a higher incidence of spontaneous
premature delivery than non-diabetic women. Labour and
delivery are potentially hazardous events for both mother and
infant.
Insulin resistance increases during the second and third
trimesters, necessitating an increase in insulin doses. All
women with tablet treated type 2 diabetes should be treated
with insulin during pregnancy. Blood glucose must be monitored
carefully during labour. An intravenous infusion of soluble
insulin should be commenced and insulin rate adjusted (usually
2–4 U/h) to maintain blood glucose levels of 6–8 mmol/L in all
insulin treated women; to prevent hypoglycaemia, a 10 per cent
dextrose solution should be co-infused at 125 mL/h.
Women with gestational diabetes who have well controlled diet
treated diabetes do not usually require insulin treatment;
nonetheless, blood glucose is closely monitored during labour.
After delivery, insulin and dextrose infusion rates should
immediately be halved. The mother’s pre-pregnancy insulin
regimen should be restarted once she resumes eating.
Diabetes and surgery
The diabetic patient is at increased risk of requiring surgery
for complications such as advanced foot disease and the
136 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
manifestations of atherosclerosis. Often, such surgery is performed
on emergency operating lists out of regular working hours when
medical and nursing staffing levels are lower; in some countries,
limited capacity may mean that surgery is frequently postponed,
posing additional difficulties for patients who are nil-by-mouth.
Even elective surgery is associated with hazards, being associated
with higher rates of morbidity and mortality than in the nondiabetic
patient. Major metabolic decompensation, myocardial
infarction and infection are the main underlying causes for the
high mortality during and after surgery. Diabetes is commonly
encountered on general surgical wards; duration of hospital stay is
often prolonged in patients with diabetes. Important contributory –
but potentially modifiable – factors include
sub-optimal peri-operative metabolic control
imperfect monitoring of metabolic control during surgery and
post-operatively
presence of chronic complications, e.g. coronary artery disease
or autonomic neuropathy, that render the patient with diabetes
more vulnerable to adverse surgical outcomes.
Morbidity and mortality among patients with diabetes undergoing
surgery are higher than those among non-diabetic
patients.
The use of uncomplicated management protocols, well planned
pre-operative assessment and improved surgical methods has led
to a reduction in mortality of diabetic patients undergoing surgery.
Of note, recent data have demonstrated the benefits of intensive
glycaemic control using i.v. insulin regimens in critically ill
patients.
DIABETES AND SURGERY 137
Intensive insulin therapy can improve prognosis in critically
ill patients with hyperglycaemia.
Factors adversely affecting metabolic control during surgery
include
diseases underlying the need for surgery, e.g. sepsis, atherosclerosis
hormonal and metabolic responses to trauma
nosocomial infection – more common in diabetic patients
sub-optimal timing of meal delivery on wards
starvation – may accelerate the development of ketosis
certain drugs, e.g. anaesthetic drugs, corticosteroids etc.
Pathophysiology
As reviewed in Chapter 1, insulin is a powerful anti-catabolic
hormone which promotes tissue glucose uptake, glycogen formation
in liver and muscle, protein synthesis and lipogenesis. Catabolic
hormones such as cortisol, catecholamines, growth hormone and
glucagon oppose the actions of insulin by stimulating glycogen
breakdown, gluconeogenesis, lipolysis and inhibition of
protein synthesis. In non-diabetic subjects, surgical trauma results
in increased catabolic hormone secretion, a relative decrease in
insulin and increased insulin resistance in major target tissues such
as muscle, liver and adipocytes; elevated levels of cytokines may
exacerbate insulin resistance. The metabolic results of these
changes can be intense catabolism, depending on the extent of
surgical trauma, with increased glucose release and metabolic
138 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
decompensation. Patients with type 1 diabetes requires higher
insulin doses to counter catabolism. Those with glucose intolerance
or type 2 diabetes are unable to increase endogenous insulin secretion
sufficiently; insulin treatment may be required temporarily.
Management
Aims
The main management aims for diabetic patients undergoing
surgery should be to prevent the following.
Metabolic decompensation. This includes hyperglycaemia, ketoacidosis
and hyperosmolar non-ketotic hyperglycaemia.
Hypoglycaemia. Note that the diabetic patient under a general
anaesthetic is unable to recognise or report warning symptoms
of hypoglycaemia (see Chapters 4 and 5). Changes such as
tachycardia that accompany the adrenergic response to hypoglycaemia
may be misinterpreted as being attributable to blood
loss. Thus, careful and frequent monitoring of capillary blood
glucose is the only means of detecting hypoglycaemia.
Delayed wound healing and sepsis. Hyperglycaemia impairs
cellular response to injury and infection.
Pre-operative management
One of the most important steps in achieving a good perioperative
glycaemic control is a carefully planned preoperative assessment in
diabetic patients undergoing surgery. This is best initiated well in
advance of planned surgery, although in practice can be difficult to
ensure.
DIABETES AND SURGERY 139
A careful general medical assessment should be performed,
with particular attention paid to systems affected by diabetes.
These include the cardiovascular system, renal function,
autonomic system and blood pressure. Investigations should
usually include
12-lead electrocardiograph – even in the absence of
symptoms of cardiac disease; exercise testing and coronary
angiography may be indicated; note that clinically silent
myocardial ischaemia is more common in the diabetic
patient
chest X-ray, if any suspicion of cardio-pulmonary disease
biochemical assessment of renal function, i.e. plasma
creatinine and electrolyte concentrations. Note that reliance
on plasma creatinine concentrations may underestimate
renal impairment, particularly in the elderly. Hyporeninaemic
hypoaldosteronism associated with diabetic nephropathy
predisposes to hyperkalaemia.
It is important to liaise with the anaesthetist, who should be
made aware of all diabetic patients for whom surgery is
planned.
Further assessment depends on whether these patients are
undergoing minor or major surgery. The former includes daycase
procedures such as upper gastrointestinal endoscopy.
Minor Surgery and Day-case Procedures
Ideally patients should be admitted to the hospital the night
before the operation, but if not possible than they should be
dealt with as day cases and admitted early on the morning of
the operation.
Aim to ensure good pre-operative glycaemic control; since this
may take some time to organise, the earliest opportunity should
140 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
be taken to initiate changes in therapy. Liaise with the diabetes
care team. Check haemoglobin A1c and use the patient’s blood
glucose self-monitoring results to optimise anti-diabetic therapy.
Targets of 7.5 per cent for glycated haemoglobin and preprandial
capillary blood glucose concentrations 6 mmol/L
should be sought, where feasible. In patients with co-morbidities,
e.g. renal or cardiac failure, it may prove difficult to attain
such targets safely because of the high risk of hypoglycaemia
(see Chapter 4). This may necessitate adjustment of present
doses of oral anti-diabetic agents and/or insulin; some patients
will require the addition of insulin to oral therapy, or complete
substitution of insulin for oral agents. While excellent glycaemic
control is particularly important in the patient heading for
major surgery (see below), minor procedures, e.g. angiography
for aorto-femoral atherosclerosis, or digit amputation, sometimes
require more invasive interventions because of complications.
Major Surgery
All diabetic patients undergoing major surgery should ideally
be admitted to hospital 2–3 days prior to surgery. If this is not
possible then admit at least 24 hours beforehand.
Replace long-acting blood glucose lowering agents such as
glibenclamide with shorter-acting agents such as gliclazide or
glipizide to reduce the risk of hypoglycaemia (see Chapter 5).
Hypoglycaemia with agents such as chlorpropamide, no longer
used in the UK, can sometimes occur.
Avoid the use of metformin as this may predispose to lactic
acidosis, particularly in patients with renal insufficiency, sepsis
or hypotension (see Chapter 6).
Monitor capillary blood glucose levels during the stay and
optimise glycaemic control. Change from oral agents to insulin
DIABETES AND SURGERY 141
if control is inadequate (see above). For some patients already
on insulin therapy, a change in regimen or insulin preparation
may be required. For example, patients with inadequate control
using twice-daily insulin may be usefully changed to multiple
daily injections using short- or rapid-acting insulin before main
meals, using a longer-acting preparation (isophane, insulin
glargine, insulin detemir) at bedtime. In general, it is usually
better to use smaller doses given more often in such
circumstances.
If possible, operate during the morning to avoid prolonged
fasting. Although this is not essential, it is helpful to have the
diabetes care team available for consultation in the postoperative
period. Insulin withdrawal studies suggest that the
development of ketosis (see Chapter 1) is accelerated in patients
with type 1 diabetes who are fasted.
Operative management
Management of diabetes during surgery depends on factors such
as
estimating whether the patient can secrete adequate amounts of
insulin
duration and type of surgery
whether post-operative ileus will be present.
Patients treated with insulin – whether considered to have type 1 or
type 2 diabetes – should be assumed to have negligible insulin
reserves; this is always the case in the former group and in the
latter group is often the case for patients with diabetes of long
duration. The implication is that these patients are at high risk of
developing ketoacidosis if not treated with exogenous insulin in
adequate doses. The clinical team, in collaboration with the patient
where possible, must therefore assume responsibility for ensuring
142 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
that sufficient insulin is provided to cover the hormonal response
that accompanies major surgery. It should be noted that so-called
minimally invasive surgery, e.g. laparoscopic cholecystectomy, is
reportedly accompanied by a catabolic hormone response comparable
to that observed with conventional surgery.
Minor Surgery and Day-case Procedures It is generally agreed that
a conservative approach is reasonable in these patients.
Operate early in the morning, if possible.
Omit breakfast.
Omit the morning dose of oral anti-diabetic agents in patients
with type 2 diabetes.
Omit morning dose of insulin in type 1 and type 2 diabetic
patients treated with insulin.
Measure blood glucose every 1–2 h using reagent test strips in
conjunction with a reflectance meter.
Dextrose containing i.v. solutions should be avoided, unless
hypoglycaemia develops.
This approach is widely practiced. There are no data from
randomised studies to guide clinicians as to whether it is better
to (1) omit medication or (2) to give the morning dose of
sulphonylurea and to infuse 5 per cent or 10 per cent dextrose
during the operation until meals resume again. Omission of
medication has the advantage of an added degree of safety, since
risk of hypoglycaemia is lower. For the majority of patients, the
approach suggested will usually suffice.
Major Surgery
Omit breakfast.
Omit regular oral anti-diabetic drugs and breakfast insulin dose.
DIABETES AND SURGERY 143
At 0800–0900 h, an i.v. infusion of insulin þ dextrose is initiated
as follows.
Short-acting (soluble) insulin (50 U) is added to 50 mL saline
(0.9 per cent) in a 50 mL syringe and delivered via a variable
rate electromechanical pump (with built-in battery supply).
Commence with 1–2 U/h, increasing or reducing the insulin
infusion rate according to hourly blood glucose measurements;
aim to maintain blood glucose concentration between
6 and 11 mmol/L. An example of a variable rate i.v. insulin
infusion regimen is presented in Chapter 1. Note that postoperative
complications may greatly increase insulin
requirements – 10 U/h or more may be required in the
presence of severe sepsis. Thus, therapy must be individualised
according to the clinical circumstances.
Dextrose (10 per cent) is co-administered via a Y-connector
using a drip-counter at a rate of 100 mL/h, with appropriate
potassium chloride (usually 20 mmol/L if the patient is
normokalaemic with satisfactory renal function).
Advantages of this regimen include rapid and precise
adjustment of the ratio of insulin to dextrose; disadvantages
are risk of hypo- or hyperglycaemia if the insulin or dextrose
infusion rate is incorrect or delivery is interrupted. It can be
helpful to infuse insulin þ dextrose for a few hours before
surgery to permit stabilisation of blood glucose concentrations
before the start of surgery.
An alternative approach is to use a fixed ratio of insulin to
dextrose. When potassium chloride is added, this approach is
usually known as the glucose–potassium–insulin–GKI regimen.
500 mL 10 per cent dextrose solution containing 10 mmol of
potassium chloride (KCl) and 15 U of soluble insulin is
infused at 100 mL/h.
Capillary blood glucose is measured every 1–2 hours reducing
frequency if necessary, i.e. unable to achieve satisfactory control.
144 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
Aim to maintain glucose levels between 6 and 11 mmol/L.
If >11 mmol/L, change to an infusion containing 20 U
soluble insulin, i.e. increase the insulin by 5 U in the new
bag.
If <6 mmol/L, change to an infusion with 10 U of insulin, i.e.
reduce insulin dose by 5 U.
In elderly patients and patients with congestive cardiac failure,
the standard regimen might produce fluid overload. In these
patients a double strength infusion is prepared by adding 30 U
soluble insulin and 20 mmol of potassium chloride to 20 per
cent dextrose solution, infused at a rate of 50 mL/h.
Caesarean section is usually an elective procedure and an
infusion of dextrose þ insulin in fixed proportions is reasonable
(see below). However, insulin requirements can be expected to
be higher in these patients due to pregnancy associated insulin
resistance. Therefore, it is sensible to use 20 U of soluble insulin
instead of the 15 U in the standard infusion. The pregnant
diabetic woman already in labour who requires emergency
caesarean section would already be on insulin þ dextrose
infusion; the latter should be continued during the procedure.
A fixed dextrose þ insulin infusion is also acceptable in
emergency surgery. However, in cases where the last time of
insulin injection is not known, care needs to be taken to account
for the possible continued absorption of the preceding
subcutaneous injection. In these circumstances, insulin þ
dextrose may be better delivered via separate infusion lines to
permit rapid adjustments. Take care to correct any electrolyte
disturbances.
Patients undergoing open-heart surgery with cardio-pulmonary
bypass require higher doses of insulin, partly because of the use
of high volumes of glucose containing fluids.
DIABETES AND SURGERY 145
Post-operative management
Minor Surgery/Day Case
Patients should be re-started on oral hypoglycaemic agents with
the first post-operative meal.
Major Surgery
Continue i.v. insulin þ dextrose until patient starts to eat when
the patient’s usual treatment can be re-started. Intensive insulin
therapy in critically ill intensive care patients with blood
glucose concentrations > 12 mmol/L reduced serious morbidity
and mortality rate by 40 per cent in a recent clinical trial.
Note that so-called ‘sliding scales’ of subcutaneous insulin are
not recommended, since they use retrospective information to
rigidly guide therapy. This is not to imply that pre-determined
regimens should be unthinkingly enforced either. Rather,
capillary glucose concentrations should be measured before
meals and insulin prescribed according to food intake, knowledge
of previous insulin requirements, likely insulin resistance etc. This
requires experience and skill and is often a clinical challenge.
Plasma urea and electrolytes should be measured 4–6 h postoperatively
and then daily if the patients are receiving an
insulin þ dextrose infusion for more than 24 h; hyponatraemia
is well recognised and can be largely averted by using 20 per
cent dextrose in reduced volume (see above) or, where fluid
overload is not a risk, by co-infusion of saline, e.g. 1 L every 24 h.
Care should be taken to maintain normokalaemia.
To summarise, surgery in the diabetic patient needs careful
planning and organisation to reduce complications and to improve
mortality and morbidity rates. Using a variable dose i.v. insulin þ
dextrose infusion is a safe and widely accepted method to maintain
satisfactory glycaemic control during surgical procedures. However,
careful monitoring by trained staff and a rapid response if things do
not go to plan are essential to ensure the patient’s safety. The approach
to managing diabetes during surgery is outlined in Figure 7.1.
146 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
Pre-operative
Not insulin treated Insulin treated
During surgery
Post-operative
• Plan date of surgery
• Liaise with anaesthetists & diabetes
care team
• Replace long-acting with shortacting
secretagogues
• Major surgery & poor
glycaemic control.
Admit 2 3 days before
surgery
Optimise control
• Minor surgery & good
glycaemic control.
Admit 1 day before surgery
Monitor blood glucose
• Admit 2 3 days before
surgery
• Optimise metabolic control
• Omit breakfast
• Omit oral agents & insulin on the morning
• Operate in the morning
• Check blood glucose 2 hourly
• Major surgery & poor glycaemic
control.
Insulin + dextrose infusion
• Minor surgery & good glycaemic
control.
Avoid dextrose infusion
2-hourly blood glucose
measurements
• Insulin + dextrose infusion
• Monitor blood glucose 2 hourly
• Re-start oral agents once blood
glucose control satisfactory; use
insulin if necessary at this stage
• Re-start s.c. insulin (short or
rapid acting) before 1st meal.
• Terminate i.v. infusion 30 min later
_
_
Figure 7.1 Protocol for management of diabetes in surgical patients
DIABETES AND SURGERY 147
Diabetes and myocardial infarction
Diabetes is associated with a two- to fourfold increase in risk of
cardiovascular disease relative to the general population. Cardiovascular
mortality is doubled in diabetic men and the relative risk
is even higher in women with diabetes. Data from Finland have
suggested that mortality rates are comparable to those of nondiabetic
people who have previously suffered a myocardial
infarction. Acute myocardial infarction accounts for 30 per cent
of all deaths in the whole diabetic population. More than 50 per
cent of all patients admitted to coronary care units with acute
myocardial infarction have some impairment of glucose tolerance.
Epidemiological studies demonstrate an increased risk of early and
late mortality in diabetic patients.
Cardiac failure is the main cause of death following myocardial
infarction in patients with diabetes.
Re-infarction is also more common than in non-diabetic patients.
Accordingly, the prevention of atherosclerotic complications
through aggressive management of modifiable risk factors including
hypertension and dyslipidaemia is a major objective.
In patients with diabetes, myocardial ischaemia may present
without pain – so-called ‘silent ischaemia’, which is thought to
result, at least in part, from autonomic neuropathy. Atypical
presentations of myocardial infarction include
breathlessness due to worsening heart failure
acute deterioration in glycaemic control
vomiting or collapse
acute confusion in the elderly.
A high index of suspicion is required to diagnose silent myocardial
infarction. As mentioned above, subclinical ischaemia is more
148 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
common among diabetic patients; this may be identified by a
standard 12-lead electrocardiograph or by an exercise tolerance
test.
Pathophysiology
It has long been recognised that patients with diabetes have severe
and more widespread atherosclerosis than their non-diabetic
counterparts. Traditional risk factors for atherosclerosis do not
explain the high rate of coronary artery disease in diabetic patients.
It is now widely thought that clustering of atherogenic risk factors
is important. The metabolic syndrome of insulin resistance is
commonly found in patients with glucose intolerance or type 2
diabetes. In 2001, the US National Cholesterol Education Program
suggested clinical and biochemical criteria for the metabolic
syndrome (Table 7.1). Other abnormalities, e.g. chronic inflammation
Table 7.1 National Cholesterol Education Program Adult Treatment
Panel III criteria for diagnosis of the metabolic syndrome
Fasting hyperglycaemia
glucose > 6.0 mmol/L
Central obesity
men >102 cm
women >88 cm
Hypertension
130/85 mm Hg
Dyslipidaemia
hypertriglyceridaemia >1.7 mmol/L
low high-density lipoprotein cholesterol
& <1.0 mmol/L for men
& <1.3 mmol/L for women
The presence of three or more constitutes the metabolic syndrome
DIABETES AND MYOCARDIAL INFARCTION 149
or impaired fibrinolysis (see below), are also associated with
insulin resistance. Also of note, the 2001 NCEP report included
diabetes as a ‘coronary risk equivalent’ in recognition of the high
risk conferred by diabetes. The World Health Organisation has also
produced a slightly different set of criteria for the metabolic
syndrome that is not as readily employed in routine clinical
practice. The metabolic syndrome is highly prevalent in the USA
and many other industrial and developing countries, driven by
adverse trends in body weight and lifestyle changes.
The prevalence of the metabolic syndrome of cardiovascular
risk factors is increasing in many parts of the world.
Atheromatous changes lead to impaired arterial relaxation due to
reduced production of nitric oxide, a potent vasodilator; insulin
resistance per se may be associated with endothelial dysfunction.
Diabetes mellitus is also associated with hypercoagulability, the
procoagulant changes on the endothelial surface favouring
thrombosis. Platelet-rich thrombus in the coronary arteries is
unstable and likely to rupture, causing acute coronary occlusion.
Plaque in patients with diabetes may be particularly vulnerable to
rupture due to a high inflammatory cell content and other adverse
components.
The explanation for the continuing poor prognosis in the diabetic
patient may lie, in part, in the secretion of counter-regulatory
hormones that ensue after acute myocardial infarction; these result
in adverse changes in cellular metabolism that are exacerbated by
diabetes (see Chapter 1). Hyperglycaemia – secondary to acutely
exacerbated insulin resistance and insulin deficiency – is accompanied
by acceleration of adipocyte lipolysis, the latter resulting in
release of non-esterified fatty acids (NEFAs). Myocardial glucose
uptake and metabolism are reduced by insulin deficiency. Under
these circumstances, the oxygen consumption of the ischaemic
myocardium is increased by reliance on NEFA oxidation; this
results in myocardial dysfunction that can be reduced if cellular
glucose uptake and metabolism can be improved (see below). A
150 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
chronic diabetic cardiomyopathy has also been described, which
may contribute to the excess risk of heart failure after myocardial
infarction in diabetic patients; the literature suggests that the size of
infarcts is no greater among patients with diabetes.
Management
Once myocardial infarction is diagnosed in a patient with known
diabetes, or indeed any patient with a blood glucose concentration >
11.0 mmol/L, the immediate steps in the management, in addition
to analgesia and supplemental oxygen, are the following.
Aspirin. This decreases the risk of reinfarction to a degree
comparable to that of non-diabetic subjects. A dose of 300 mg
should be chewed.
Thrombolysis. This has been shown to decrease mortality in
diabetic as well as non-diabetic patients and should be
administered according to current criteria. However, the
discrepancy in mortality persists in diabetic patients even
with thrombolysis. Note that the presence of diabetic retinopathy
is not a contra-indication to thrombolysis, the risk of
intra-ocular pressure being minimal and far outweighed by the
benefits of reperfusion.
Percutaneous transluminal coronary angioplasty (PTCA). Symptoms
can be relieved immediately and at-risk myocardium can
be preserved using this procedure; however, current practice in
the UK is focused on revascularisation using thrombolysis as
first-line therapy. The immediate outcomes of PTCA in diabetic
patients are similar to those in non-diabetic patients, but reocclusion
has tended to be commoner in diabetic patients. A
follow-up study of the BARI study (Bypass Angioplasty
Revascularization Investigation) showed that diabetic patients
who had previously undergone coronary artery bypass grafting
DIABETES AND MYOCARDIAL INFARCTION 151
(CABG) had a better prognosis after subsequent acute myocardial
infarction as compared with those who had received
PTCA. Use of new anti-platelet agents and drug-eluting stents
has reduced re-stenosis rates in diabetic patients. Specialist
interventional cardiology services are a crucial component of
high-quality care for high-risk patients.
Management of hyperglycaemia
It has become apparent that use of insulin to rigorously control
hyperglycaemia after acute myocardial infarction is associated with
major improvements in survival. The Diabetes Mellitus Insulin
Glucose Infusion in Acute Myocardial Infarction (DIGAMI) study
group treated patients with known diabetes, and those with a
blood glucose values 11 mmol/L, intensively with an i.v.
insulin þ dextrose infusion for 24 h after admission and thereafter
with multiple daily injections of insulin for a minimum period of
3 months. Patients in the intensively treated group received an
infusion of 80 U soluble insulin mixed in 500 mL 5 per cent
dextrose solution, initially at 30 mL/h for 24 h after admission and
adjusted to achieve a target blood glucose range of 7–10 mmol/L
(Figure 7.2). Careful monitoring by trained staff is required for safe
implementation of the protocol, which carries risks of hypoglycaemia.
The intravenous infusion was followed by pre-meal soluble
insulin injections given subcutaneously for a minimum of 3 months,
wherever possible.
In this randomised trial, the intensively treated patients were
compared with patients treated with insulin only if this was
considered indicated on clinical grounds. The 12 month
mortality was reduced by 28 per cent in the intensively treated
group; this improvement extended out to 3.4 years, the absolute
reduction in mortality being 11 per cent (Figure 7.3). Interestingly,
most of the benefit was observed in a subgroup of
patients who had not previously been treated with insulin. For
the total cohort, this translates into one life saved for 11 patients
152 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
Immediate reperfusion
• Aspirin.
• Thrombolysis.
Management of hyperglycaemia
•
•
•
•
β-blockers.
ACE inhibitiors.
Statins – regardless of serum cholesterol concentration.
Aggressive treatment of hypertension.
Advise against smoking.
Cardiac rehabilitation programme.
• Opiates.
• Nitrates.
Pain relief
As required.
500 mL 5% dextrose with 80 U soluble insulin.
Start infusion at 30 mL/h.
Check blood glucose hourly.
Adjust infusion rate:
Blood glucose (mmol/L)
>15
Adjustment
8 U insulin as i.v. bolus
↑infusion by 6 mL/h
↑infusion by 3 mL/h
maintain current rate
↓infusion by 6 mL/h
stop infusion until blood
glucose > 7 mmol/L
give 20 mL 30% glusose I.V.
if symptomatic
hypoglycaemia. Re-start
infusion with rate decreased
by 6 mL/h
11 14.9
7 10.9
4 6.9
<4
_
_
_
••••
•
•
Further management
Figure 7.2 Management of myocardial infarction in patients with diabetes
DIABETES AND MYOCARDIAL INFARCTION 153
treated intensively; this compares very favourably with the
effects of other measures, e.g. thrombolysis.
Theoretically, some of the benefit of the intensive approach may
have been due to a reduction in plasma NEFA concentrations,
thereby reducing myocardial injury or improving myocardial
uptake and metabolism of glucose. Another contributory factor
may be the reduction in the sensitivity to circulating catecholamines
and inhibition of the inappropriate neuroendocrine
activation after infarction brought about by insulin infusion,
thus lowering the risk of heart failure. A follow-up study,
DIGAMI-2, aims to determine whether the benefits observed in
the intensive treatment group were primarily a consequence of
the early i.v. insulin þ dextrose infusion or subsequent subcutaneous
insulin therapy; it will also examine the efficacy of
twice-daily insulin versus multiple insulin injections in this
context.
Infusion
Control
Control 314 232 187 116 58 14
Infusion 306 248 202 128 50 13
0 1 2 3 4 5
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Death rate
(no of deaths/no originally in group)
(n = 620)
Years in study
No of patients at risk
Figure 7.3 Actuarial mortality curves for patients receiving insulin–
dextrose infusions followed by multiple daily insulin injections and controls.
See the text for details. Reproduced with permission from Malmberg K et al.
Br Med J 1997; 314: 1512–1515
154 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
Intensive treatment with insulin reduces mortality in hyperglycaemic
patients following acute myocardial infarction.
The cardiovascular safety of sulphonylureas, particularly agents
that bind to sulphonylurea (SUR2) receptors in cardiac and
vascular tissues, has been an issue of controversy for decades.
While no firm conclusion has been reached, some authorities
recommend avoidance of drugs such as glibenclamide that can
impair the phenomenon of ischaemic preconditioning, i.e.
reduced myocardial damage following episodes of prior
ischaemic. Sulphonylureas release insulin by binding to SUR1
receptors on the membrane of the islet b-cells, thereby closing
potassium channels. Potentially disadvantageous interactions
have been described between the anti-anginal agent nicorandil –
which opens potassium channels in cardiac myocytes as a
protective action – and glibenclamide; this is not observed with
some other sulphonylureas, e.g. gliclazide. However, the
clinical relevance of these observations remains uncertain. No
adverse effect of sulphonylureas was observed in the United
Kingdom Prospective Diabetes Study.
The relationship between sulphonylureas and outcomes
after myocardial infarction remain controversial.
Longer-term management
All patients should be commenced on a cardio-selective
b-blocker and an angiotensin converting enzyme (ACE)
inhibitor, unless either is contraindicated; benefits of these
agents are at least as great, and in the case of b-blockers even
greater, than in non-diabetic patients. The small risks of
reducing warning symptoms and recovery from hypoglycaemia
(see Chapter 4) should not deter the use of cardio-selective bblockers.
Evidence is mounting that angiotensin II1 receptor
DIABETES AND MYOCARDIAL INFARCTION 155
antagonists are useful alternatives to ACE inhibitors for the
minority of patients who develop a cough with the former class.
Aspirin, 75 mg daily, should be continued, the optimal dose
being uncertain.
Serum lipids should be measured. Note that the very high risk
of further coronary events in diabetic patients in effect means
that all patients who survive a myocardial infarction should
receive a statin, unless there are good reasons to omit this
therapy. This view is supported by a considerable volume of
evidence from controlled clinical trials and is endorsed by
expert groups in the US and Europe. The primary focus on
lowering low-density lipoprotein (LDL) cholesterol using
statins is emphasised in the National Cholesterol Education
Panel 2001. Doses of drugs that have been shown to be effective
in clinical trials should be employed. The plasma lipid profile
can be altered temporarily, with a risk that hypercholesterolaemia
may be underestimated. Further assessment of a complete
lipid profile, i.e. total, LDL and HDL cholesterol together with
triglycerides after an overnight fast, should be performed at
outpatient follow-up. More complex dyslipidaemias may
require additional therapy; specialist advice may be needed. A
mixed dyslipidaemia characterised the abnormalities in
Table 7.1, is common among patients with type 2 diabetes.
Other drugs, such as fibric acid derivatives, may be helpful,
although evidence for the efficacy of this class lags behind the
firm evidence base for the statins.
Hypertension should be aggressively managed in diabetic
patients; this often means therapy with drugs from two or three
different classes. The target blood pressure is <130/80 mm Hg.
Patients should be strongly advised against smoking; support
should be offered, including pharmacological measures.
All patients should, if possible, be enrolled in a cardiac
rehabilitation programme; the benefit of supervised exercise
156 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
programmes has probably been underestimated by many
clinicians. Weight reduction, where indicated, and regular
physical exercise as a long-term lifestyle modification, are
strongly recommended, where feasible.
Secondary prevention measures after myocardial infarction
include anti-platelet drugs, statins, tight blood pressure
control, avoiding smoking and regular physical exercise.
Diabetes and labour
While this section focuses on management of diabetes during labour,
some key points concerning diabetes and pregnancy deserve mention.
Pregnancy should be planned wherever possible to ensure
excellent glycaemic control at conception; this helps to reduce
the rate of congenital malformations.
Pregnancies in mothers with diabetes are associated with a high
rate of late stillbirths; this remains incompletely unexplained.
Late stillbirths are associated with maternal hyperglycaemia
and fetal macrosomia. Macrosomia has implications for
delivery, e.g. shoulder dystocia.
Pregnancy is a state in which insulin resistance is temporarily
exacerbated. This may precipitate glucose intolerance in
susceptible women and has predictable implications for
glycaemic control in women with pre-existing diabetes. Insulin
resistance increases during the second and third trimesters due
to hormonal changes associated with pregnancy.
Women with type 2 diabetes or gestational diabetes that have
sub-optimal glycaemic control should be treated with insulin
DIABETES AND LABOUR 157
during pregnancy. Women with well controlled gestational
diabetes should measure their blood glucose six times per day
on three days per week.
Pre-eclampsia is at least two- to fourfold more common in
diabetic than in non-diabetic pregnancies. Diabetic pre-eclampsia
is associated with 60 per 1000 deaths compared to 3.3 per
1000 deaths in normotensive diabetic pregnancies. Vascular
disease, pre-existing microalbuminuria, diabetic nephropathy,
long duration of diabetes and poor glycaemic control are
independent risk factors for pre-eclampsia. Pre-eclampsia is one
of the causes of spontaneous pre-term labour and delivery,
which has an increased peri-natal morbidity and mortality.
Women with insulin treated diabetes (which includes both type 1
and type 2 diabetes) that antedates pregnancy have up to 25 per
cent incidence of spontaneous premature labour and delivery.
Spontaneous pre-term delivery is associated with antecedent
poor glycaemic control and urogenital infection.
The incidence of gestational diabetes, i.e. diabetes diagnosed
during pregnancy, reflects factors such as age, obesity, ethnicity
and methods of ascertainment.
Pregnancy is regarded as high risk in women with diabetes,
posing hazards for both mother and fetus.
Pathophysiology
Maternal diabetes may affect the structure and function of the
placenta. Fetal hypoxia, acidosis, placental dysfunction and
hypokalaemia leading to cardiac dysrhythmias are the possible
underlying causes for the higher incidence of late stillbirths.
Excessive oxidative stress has been reported at delivery, which
158 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
may also contribute. Impaired oxygen delivery to the foetus, due to
a higher affinity of glycated haemoglobin for oxygen than nonglycosylated
haemoglobin, may be another contributory factor. The
contribution of intra-uterine malnutrition to chronic disease including
diabetes and coronary artery disease in later life – the foetal
origins hypothesis – is supported by epidemiological and experimental
evidence. Tight glycaemic control is the aim in women with
glucose intolerance during pregnancy.
Management
Fetal monitoring
In view of the association between maternal hyperglycaemia
and fetal hypoxia, foetal monitoring is important. However, the
intensity of monitoring is still a subject of debate.
In cases where labour is progressing normally, the woman has
no diabetic complications and wishes to mobilise, intermittent
foetal monitoring may be acceptable if the initial foetal trace has
been satisfactory. Continuous cardiotocography (GTG) may not
be essential but may be reassuring. Some obstetric units now
employ telemetric CTG.
Time and mode of delivery
Premature induction of labour for the fear of sudden death in
late pregnancy is no longer practised.
Delivery can safely be delayed until term or 39 weeks in diabetic
women with good glycaemic control.
The mode and time of delivery need to be individualised in
diabetic women with pregnancy complications or coexisting
severe diabetic angiopathy.
DIABETES AND LABOUR 159
In the case of planned labour the following may be necessary.
Intra-cervical application of prostaglandin jelly to dilate the
cervix.
Oxytocin.
Amniotomy.
Fetal heart rate should be monitored continuously during
vaginal delivery and pH measurements should be performed
when indicated.
In cases of severe exacerbations of retinal changes and general
obstetric indications, caesarean section should be performed.
Management of diabetes during labour
Management of diabetes during labour and postpartum period
should be planned and discussed with the patient; the outline
protocol should be recorded in the case notes.
Insulin treated women are at risk of developing hypoglycaemia
during labour; capillary blood glucose should be carefully
monitored.
In cases where labour is induced, women should be continued
on their usual diet and insulin regimen until labour becomes
imminent.
The full range of obstetric analgesia should be available.
Once the woman is in labour, the following are required.
Short-acting insulin should be commenced and given through
a syringe-driver infusion pump containing 1 U/mL of soluble
insulin added in isotonic (0.9 per cent) saline (Figure 7.4).
160 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
10 per cent dextrose solution should be constantly infused
intravenously at a rate of 125 mL/h, which provides 50 kcal
of energy per hour during labour to prevent hypoglycaemia.
Capillary blood glucose should be monitored hourly.
Plan delivery
• Discuss time & mode of delivery.
• Discuss management of diabetes.
• Record outlined protocol in patients’ notes.
Management of diabetes during labour
• Continue diet and regular insulin until labour imminent.
• Capillary blood glucose monitoring every hour.
• Prepare a syringe-driver pump containing 1 U/mL of
soluble insulin in isotonic saline.
• Start 2 8 U/h of soluble insulin.
• Commence i.v. 10% dextrose infusion.
• Adjust insulin infusion rate to maintain blood glucose
between 6 8 mmol/L.
• Less insulin may be needed during prolonged labour.
Postpartum management
• Halve the insulin pump and dextrose infusion rate.
• Discontinue both as soon as the mother starts eating.
• Re-start pre-pregnancy insulin regimen.
• Paediatrician should be present at delivery to monitor
neonate blood glucose levels.
_
_
Figure 7.4 Management of diabetes during labour
DIABETES AND LABOUR 161
Insulin infusion rate should be adjusted to maintain blood
glucose levels between 6 and 8 mmol/L.
Intravenous insulin requirements are typically 2–4 U/h in
uncomplicated deliveries; insulin requirements may decline
during prolonged labour. Note that use of b-adrenergic
agonists (to retard pre-term labour) and parenteral dexamethasone
(for fetal lung immaturity) can dramatically
increase maternal insulin requirements; severe hyperglycaemia
and even ketoacidosis may be precipitated.
After delivery and in the immediate postpartum period, we
have the following.
Insulin requirement normally decreases promptly to the prepregnancy
levels after the third stage of labour.
Both the i.v. insulin infusion and i.v. dextrose infusion rates
should be reduced to half of that required during labour.
Once the mother starts eating, insulin and dextrose infusions
should be stopped; pre-pregnancy insulin dosage are
recommenced with the first meal. Women treated successfully
with oral anti-diabetic agents prior to pregnancy may be
able to re-start their medication, if not breast feeding. Note
that oral anti-diabetic agents are always replaced by insulin
during pregnancy because of concerns of teratogenicity and
inadequate metabolic control.
A paediatrician should be present at every delivery of a
diabetic mother as neonates of diabetic mothers frequently
have hyperinsulinaemia; up to 50 per cent develop hypoglycaemia
and the neonates’ capillary blood glucose should
be monitored every 2–4 hours. Hypoglycaemia usually
responds to frequent early feeds, i.v. therapy being largely
avoidable.
For women with gestational diabetes, a follow-up oral glucose
tolerance test is recommended at 6 weeks. Such women are at high
162 SURGERY, MYOCARDIAL INFARCTION AND LABOUR
risk of developing type 2 diabetes and 6–12 monthly follow-up
thereafter is recommended. Contraception and planning of future
pregnancies should be discussed.