|Ciammaichella M. M., Galanti A., Rossi C.
U.O.C. Medicina Interna I per l’Urgenza
(Direttore: Dott. G. Cerqua)
A.C.O. S. Giovanni - Addolorata - Roma, Italia
is characterized by an anion gap acidosis due to high levels of ketoacids.
It occurs exclusively in relation to alcohol abuse but not just in chronic
alcoholics. It has been reported in first-time drinkers whose food intake
The true incidence is unknown, and the frequency is probably directly
related to the incidence of alcoholism in a population.
Several mechanisms have been postulated. In one explanation, ketosis results
from increased mobilization of free fatty acids from adipose tissue coupled
with simultaneous enhancement of the liver's capacity to convert these
substrates into acetoacetate and $-hydroxybutyrate.
During the metabolism of alcohol in the liver the rate of nicotinamide
adenine dinucleotide (NAD) reduction exceeds the rate of mitochondrial
NADH oxidation, causing a decrease in available NAD. This state persists
for a few days in spite of no further alcohol consumption. An NAD-dependent
step in the oxidation of fatty acids in the mitochondria of the hepatocyte
is displaced in favor of ketone body formation.
During alcoholic ketoacidosis insulin levels are low, whereas levels of
cortisol, growth hormone, glucagon, and epinephrine are increased, possibly
as a result of alcohol-induced hypoglycemia. This hormonal milieu promotes
lipolysis, which increases the levels of free fatty acids available for
conversion to ketones.
Additional mechanisms that may contribute to ketosis include the conversion
of acetate, an alcohol breakdown product, to ketones; alcohol-induced
mitochondrial structural changes which enhance the rate of ketosis; and
mitochondrial phosphorus depletion, which inhibits the utilization for
NADH and increases ketone body formation. Finally, vomiting and starvation
superimposed on chronic malnutrition also contribute to ketoacidosis.
The usual history
is one of heavy alcohol consumption or binge drinking with decreased or
absent food intake for several days. Food and alcohol intake are usually
terminated by nausea, protracted vomiting, and abdominal pain occurring
24 to 72 h before presentation. It is during this period that ketoacidosis
Clinically the patient appears acutely ill with dehydration, tachypnea,
tachycardia, and diffuse abdominal pain. Most patients are alert, but
they may be mildly disoriented or occasionally comatose.
There are no specific physical findings. Evidence of dehydration such
as hypotension, orthostatic changes in blood pressure, tachycardia, and
decreased urine output may be present. The temperature varies from hypothermia
to mildly elevated. Abdominal pain due to non-specific causes or due to
gastritis, pancreatitis, or hepatitis is common. Sepsis, meningitis, pyelonephritis,
or pneumonia may be present, and delirium tremens may develop.
are usually low or undetectable, as the alcohol intake is decreased or
discontinued during the period of anorexia and vomiting. Essential to
the diagnosis of alcoholic ketoacidosis is a large anion gap due to high
levels of serum ketones. Most patients have a blood pH reflective of the
underlying metabolic acidosis, but many may present with normal or alkalemic
Fulop and Hoberman compared typical laboratory data from patients with
diabetic ketoacidosis with data from patients with alcoholic ketoacidosis.
The alcoholic patients tended to have a higher blood pH, lower levels
of serum K+ and Cl --, and a higher
level of plasma HCO3-- than the diabetic patients. This
difference is attributed to the severe recurrent vomiting experienced
by the alcoholic patients. Vomiting causes chloride depletion and metabolic
alkalosis. In addition, respiratory alkalosis may occur secondary to fever,
sepsis, or alcohol withdrawal and further increases the blood pH.
The anion gap Na+--Cl--
+ HCO3--) = 12 +/-4
mEq/L] in the patient groups is very similar and is due primarily to high
levels of $-hydroxybutyrate and to
a lesser extent to lactic acid accumulation. The principal ketones are
acetoacetate and $-hydroxybutyrate.
These ketones are intermediates in the oxidation of fatty acids; they
are normally produced in equal amounts and are not normally detectable
in the serum. Acetoacetate and $-hydroxybutyrate
are a redox pair and are interconverted by an oxidation-reduction reaction
with NAD and NADH as cofactors. In alcoholic ketoacidosis, perhaps because
of lack of NAD, $-hydroxybutyrate accumulates
to levels several times higher than the levels of acetoacetate. Acetone
is a volatile, neutral ketone that is formed from acetoacetate by irreversible
spontaneous decarboxylation. Its presence reflects the level and duration
of acetoacetate elevation and is indicative of a sustained, severe acidosis.
The nitroprusside test is used to detect the presence of ketones in serum
and urine. This is a semiquantitative test that gives a reaction with
acetoacetate, is less sensitive to acetone, and does not detect $-hydroxybutyrate
at all. There is no practical test that measures $-hydroxybutyrate
levels. In most series on alcoholic ketoacidosis, the nitroprusside test
has shown moderate or large ketonemia or ketonuria. But in a significant
minority of patients, the reaction may be weakly positive or negative
even though ketoacidosis, because of high levels of $-hydroxybutyrate,
is pronounced. Reliance on this test alone as a measure of ketoacidosis
may lead to failure to recognize the presence of ketoacidosis or to an
underestimation of the severity of the ketoacidosis.
The blood glucose level in alcoholic ketoacidosis varies from hypoglycemia
to mild elevation. In most series it is normal or slightly increased.
Glucosuria is usually mild or absent. A subset of alcoholic patients in
whom hypoglycemia and ketoacidosis are coexistent has been described.
The pathogenesis of alcohol-induced hypoglycemia includes acute starvation,
depletion of liver glycogen stores because of chronic malnutrition, and
inhibition of gluconeogenesis because of alcohol-induced alteration of
the NAD/NADH ratio. Alcohol also causes decreased peripheral utilization
of glucose, and this acts to balance the glucose-depleting processes.
Devenyi asks if alcoholic hypoglycemia and alcoholic ketoacidosis are
sequential events of the same process. He theorizes that alcohol-induced
hypoglycemia occurs first, causing increased levels of cortisol, growth
hormone, glucagon, and epinephrine; this may correct the hypoglycemia
and mobilize free fatty acids, which are converted to ketones. If this
theory is correct, the diagnosis of alcoholic hypoglycemia or alcoholic
ketoacidosis may depend upon the point in this process at which the disorder
The diagnosis of alcoholic ketoacidosis is easily established in those
patients with an antecedent history of alcohol intake, decreased food
intake, vomiting, and abdominal pain, and laboratory findings of metabolic
acidosis, a positive nitroprusside test, and a low or mildly elevated
Several factors may contribute to the failure to recognize this metabolic
disorder. The blood alcohol level may be zero, and, in the absence of
a history of alcohol intake, this diagnosis may not be considered. The
nitroprusside test may be weakly positive or negative in spite of significant
ketoacidosis. The pH may be mildly acidotic, normal, or even alkalemic
in the face of pronounced metabolic acidosis. There are no specific physical
findings which suggest the diagnosis of alcoholic ketoacidosis. Alcoholic
patients may have a variety of alcohol-induced associated illnesses which
may obscure or distract from this diagnosis. Mental confusion or coma
may be incorrectly attributed to alcoholic intoxication or other causes
if the appropriate laboratory studies are not performed or if they are
Soffer and Hamburger's criteria to define alcoholic ketoacidosis are a
serum glucose level less than 300 mg/dL, a recent history of alcohol intake
with a relative or absolute decline in ethanol consumption 24 to 72 h
before hospitalization, a history of vomiting, and a metabolic acidosis
for which other causes, such as diabetic ketoacidosis, lactic acidosis,
renal failure, or drug ingestion, are excluded by clinical observations
or laboratory studies. A positive serum nitroprusside test, because of
its limitations, is not a criterion for diagnosis.
A positive nitroprusside test and a very low plasma bicarbonate concentration
suggest ketosis with a high level of $-hydroxybutyrate.
The combination of a barely positive nitroprusside test and a low plasma
bicarbonate concentration signifies either a very reduced state with high
concentrations of $-hydroxybutyrate
or else a coincidental lactic acidosis. The measurement of serum lactate
levels aids in this differential diagnosis.
The entity with which alcoholic ketoacidosis is most often confused is
diabetic ketoacidosis. The magnitude of ketoacidosis is equal in these
two disorders. It is important to make the proper distinction, as the
treatment of each entity is different. In diabetic ketoacidosis, hyperglycemia
and glucosuria are present. The serum glucose level in alcoholic ketoacidosis
varies from hypoglycemia to mild elevation, and glucosuria is usually
mild or absent. This differential diagnosis can be made in the emergency
Therapy of alcoholic
ketoacidosis is simple and effective and consists of the intravenous administration
of a glucose and saline solution. Patients given only saline improve,
but not as rapidly as those who are also given glucose. Thiamine, 50 to
100 mg intravenously, should be given before the glucose to prevent precipitation
of Wernicke's disease. Reversal of ketoacidosis usually occurs in 12 to
Restoration of intravascular volume is best accomplished by alternating
infusions of glucose-containing normal and half-normal saline. Volume
repletion is necessary to correct insulin-release inhibition by adrenergic
nerve endings in the islets of Langerhans as well as by circulating catecholamines.
Glucose infusion stimulates insulin release, and insulin inhibits lipolysis
and terminates ketoacid production. Glucose may inhibit further ketoacid
production by increasing oxidation of accumulated NADH via glucose-induced
uptake of phosphorus by the hepatic mitochondria.
Exogenous administration of insulin is not indicated in treatment of alcoholic
ketoacidosis; this aspect of therapy differs from therapy of diabetic
ketoacidosis. Inappropriate administration of insulin to a patient with
a normal or low glucose level could be dangerous.
Administration of sodium bicarbonate is usually not required. As ketoacid
levels fall, plasma bicarbonate levels increase, and the pH returns to
normal. A small amount of bicarbonate may be indicated if the pH is less
than 7.1 or if the patient is clinically deteriorating as evidenced by
a weak, rapid pulse, hypotension, or inability to compensate by hyperventilation
because of weakness. The role of phosphorus replenishment in therapy of
alcoholic ketoacidosis is not clear.
With recovery and reversal of the acidosis, $-hydroxybutyrate
is converted to acetoacetate. As this process occurs, the nitroprusside
test becomes more positive because of higher levels of acetoacetate. This
factitious hyperketonemia may cause the uninformed clinician unnecessary
concern, as it appears that the ketoacidosis is worsening. Clinical improvement
of the patient and increasing blood pH values are more reliable parameters
of recovery than the nitroprusside test.
The survival rates of patients with alcoholic ketoacidosis are good. Those
patients that die usually do so because of other complications of chronic
alcoholism. A thorough search for and treatment of associated alcoholic
disorders is essential. Recurrent episodes of alcoholic ketoacidosis after
subsequent alcoholic debauch are not uncommon.
1)Fulop M, Ben-Ezra J, Bock J: Alcoholic ketosis. Alcoholism: Clin Exp
Res 10:610, 1986.
2)Fulop M, Hoberman HD: Alcoholic ketosis. Diabetes 24:785, 1975.
3)Levy LJ, Duga J, Girgis M, et al: Ketoacidosis associated with alcoholism
in nondiabetic subjects. Ann Intern Med 78:213, 1973.
4)Miller PD, Heinig RE, Waterhouse C: Treatment of alcoholic acidosis—The
role of dextrose and phosphorus. Arch Intern Med 138:67, 1978.
5)Soffer A, Hamburger S: Alcoholic ketoacidosis: A review of 30 cases.
J Am Med Wom Assoc 37:106, 1982.