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
Thyroid storm is a rare complication of hyperthyroidism
in which the manifestations of thyrotoxicosis are exaggerated to life-threatening
proportions. Thyroid storm is most often seen in a patient with moderate
to severe antecedent Graves’ disease and is usually precipitated by a
stressful event. It must be suspected and treated based upon a clinical
impression, as there are no pathognomonic findings or confirmatory tests.
Hyperthyroid patients who are undiagnosed or undertreated are at risk
for this complication. The duration of uncomplicated thyrotoxicosis preceding
the onset of storm varies from months to years. A majority of the patients
have had symptoms of hyperthyroidism for fewer than 24 months. It is not
possible to predict accurately which thyrotoxic patient will develop storm
as there is no predisposition by age, sex, or race.
A wide variety of factors have been reported as precipitating
events. Thyroid surgery for treatment of hyperthyroidism used to be the
most common cause of storm. Medically identified causes of thyroid storm
are numerous and now predominate over surgical ones. Infection, especially
pulmonary infection, is the most common precipitating event. In diabetic
patients, ketoacidosis, hyperosmolar coma, and insulin-induced hypoglycemia
have provoked storm. Events known to increase the levels of circulating
thyroid hormones and initiate storm in susceptible persons, include premature
withdrawal of antithyroid drugs, administration of radioactive iodide,
use of an iodinated contrast medium during x-ray study, thyroid hormone
overdose, administration of a saturated solution of potassium iodide to
patients with nontoxic goiters, and vigorous palpation of the thyroid
gland in thyrotoxic patients. Additional events implicated as causes of
storm include vascular accidents, pulmonary emboli, toxemia of pregnancy,
and emotional stress. Finally, hospitalization may lead to storm because
of the rigors of diagnostic procedures.
The exact pathogenesis of thyroid storm has not been defined.
It is attractive to attribute storm to excess thyroid hormone production
or secretion. The results of thyroid function studies are elevated in
the vast majority of patients during storm, but the values are not significantly
different from those found in uncomplicated thyrotoxicosis. An increase
in free triiodothyronine (T3) or free thyroxine (T4) levels has been suggested
as causative of storm. But storm has occurred in the absence of elevated
free T3 or T4 levels. Something in addition to excess amounts or forms
of thyroid hormones must occur during storm.
Adrenergic hyperreactivity due to either patient sensitization by thyroid
hormones or altered interaction between thyroid hormones and catecholamines
has been suggested. Plasma levels of epinephrine and norepinephrine are
not increased during thyroid storm. The exact role of catecholamines in
storm awaits further study.
Altered peripheral response to thyroid hormone, causing increased lipolysis
and overproduction of heat, is another theory. This theory maintains that
excessive lipolysis due to catecholamine-thyroid hormone interaction results
in excessive thermal energy and fever. Finally, exhaustion of the body's
tolerance to the action of thyroid hormones, leading to decompensated
thyrotoxicosis, is a long-standing viewpoint. This implies an altered
response to thyroid hormones rather than a sudden increase in their concentration.
Thyroid storm is a clinical diagnosis as there are no laboratory
studies that distinguish it from thyrotoxicosis. Although the clinical
presentation is extremely variable, there are clues to diagnosis. A history
of hyperthyroidism, eye signs of Graves’ disease, widened pulse pressure,
and a palpable goiter are present in most patients who develop thyroid
storm. However, the history may be unobtainable and the usual features
of Graves’ disease absent, including no obvious goiter in up to 9 percent
of patients with Graves’ disease.
The generally accepted diagnostic criteria for thyroid storm are a temperature
higher than 37.8?C (100?F); marked tachycardia out of proportion to the
fever; dysfunction of the CNS, cardiovascular system or gastrointestinal
(GI) system; and exaggerated peripheral manifestations of thyrotoxicosis.
The signs and symptoms of storm usually occur suddenly, but there may
be a prodromal period with subtle increases in the manifestations of thyrotoxicosis.
Signs and Symptoms
The earliest signs are fever, tachycardia, diaphoresis, increased CNS
activity, and emotional lability. If the condition is untreated, a hyperkinetic
toxic state ensues in which the symptoms are intensified. Progression
to congestive heart failure, refractory pulmonary edema, circulatory collapse,
coma, and death may occur within 72 h.
Fever ranges from 38?C (100.4?F) to 41?C (105.8?F). The pulse rate may
range between 120 and 200 beats per minute but has been reported as high
as 300 beats per minute. Sweating may be profuse, leading to dehydration
from insensible fluid loss.
Central nervous system disturbance occurs in 90 percent of patients with
thyroid storm. Symptoms vary from restlessness, anxiety, and emotional
lability, manic behavior, agitation, and psychosis, to mental confusion,
obtundation, and coma. Extreme muscle weakness can occur. Thyrotoxic myopathy
can occur and usually involves the proximal muscles. In severe forms,
muscles of the more distal extremities and muscles of the trunk and face
may be involved. About 1 percent of patients with Graves’ disease develop
myasthenia gravis, producing an occasional confusing clinical situation.
The response of thyrotoxic myopathy to edrophonium (Tensilon test) is
incomplete, unlike the complete response that occurs with myasthenia gravis.
Hypokalemic periodic paralysis may also occur in patients with thyrotoxicosis.
Cardiovascular abnormalities are present in 50 percent of the patients
regardless of underlying heart disease. Sinus tachycardia is usual. Arrhythmias,
especially atrial fibrillation, but also including premature ventricular
contractions and, rarely, complete heart block, may be present. In addition
to increased heart rate, there is increased stroke volume, cardiac output,
and myocardial oxygen consumption. Pulse pressure is characteristically
widened. Congestive heart failure, pulmonary edema, and circulatory collapse
may be terminal events.
Gastrointestinal symptoms develop in most patients in storm. Before the
onset of storm, a history of severe weight loss is usual. Diarrhea and
hyperdefecation seem to herald impending storm and can be severe, contributing
to dehydration. During storm, anorexia, nausea, vomiting, and crampy abdominal
pain may occur. Jaundice and tender hepatomegaly due to passive congestion
of the liver, or even hepatic necrosis, have been reported. Jaundice is
a poor prognostic sign.
There are no laboratory tests that confirm thyroid storm.
The combination of free T4 assay and a sensitive TSH assay are the primary
means of assessing thyroid status in all patients. Some authors feel a
single TSH assay is an appropriate screening tool. Both tests are needed
in an emergency setting. When the clinical impression is thyrotoxicosis,
an elevated free T4 and a suppressed unmeasurable TSH confirms the diagnosis.
In an emergency setting rarely would additional thyroid tests be necessary
to establish a diagnosis of thyrotoxicosis. The uncommon T3 thyrotoxicosis
would present with a suppressed TSH and a normal or low free T4. With
this entity, the total T3 or free T3 would be required to accurately make
Routine laboratory data during thyroid storm show wide variation. Nonspecific
abnormalities in the complete blood cell count, electrolyte levels, and
liver function studies may be found. Bacterial infection may be reflected
only by a leftward shift in the differential white count without elevation
in the total white cell count. Hyperglycemia (?120 mg/dL) is common, and
hypercalcemia is occasionally present. One series reported all patients
to have low cholesterol levels with a mean value of 117 mg/dL. Plasma
cortisol levels have been observed to be inappropriately low for the degree
of stress, suggesting a lack of adrenal reserve.
Apathetic thyrotoxicosis is a rare, distinct form that usually
occurs in the elderly and is frequently misdiagnosed. These patients may
develop thyroid storm without the usual hyperkinetic manifestations and
may quietly lapse into coma and die. There are some salient clinical characteristics
which are helpful in establishing this diagnosis. The patient is generally
in the seventh decade or older, with lethargy, slowed mentation, and placid
apathetic facies. Goiter is usually present but may be small and multinodular.
The usual eye signs of exophthalmos, stare, and lid lag are absent, but
blepharoptosis (drooping upper eye lid) is common. Excessive weight loss
and proximal muscle weakness are usual. These patients generally have
had symptoms longer than nonapathetic thyrotoxic patients.
“Masked” thyrotoxicosis occurs when signs and symptoms referable to dysfunction
of one organ system dominate and obscure the underlying thyrotoxicosis.
Signs and symptoms referable to the cardiovascular system tend to mask
thyrotoxicosis in apathetic patients. These patients frequently present
with atrial fibrillation and congestive heart failure. In one series of
nine patients, the diagnosis of hyperthyroidism was unsuspected in each
case because of the predominance of cardiovascular symptoms. Congestive
heart failure in this setting may be refractory to the usual therapy unless
the underlying hyperthyroidism is diagnosed and treated.
The pathogenesis of an apathetic response to thyrotoxicosis is not understood.
Age alone is not the determining factor, as apathetic thyrotoxicosis has
been described in the pediatric age group. A high index of suspicion must
be maintained for this diagnosis.
The importance of early treatment of thyroid storm based
upon the clinical impression cannot be overemphasized. Before the therapy
is begun, blood should be drawn for sensitive TSH assay, free T4 index
or free T4 assay and free T3 and cortisol levels, a complete blood cell
count, and routine chemistries. Appropriate cultures in search of infection
Specific therapeutic goals can be divided into five areas; general supportive
care, inhibition of thyroid hormone synthesis, retardation of thyroid
hormone release, blockade of peripheral thyroid hormone effects, and identification
and treatment of precipitating events. Each of these goals must be pursued
General Supportive Care
Adequate hydration with intravenous fluids and electrolytes to replace
insensible and GI losses is indicated. Supplemental oxygen is needed because
of increased oxygen consumption. Hyperglycemia and hypercalcemia usually
improve with fluid administration but occasionally require specific therapy.
Fever should be controlled through the use of antipyretics and a cooling
blanket. Aspirin should be used with caution or not at all during storm
because salicylates increase free T3 and T4 levels because of decreased
protein binding. This objection to the use of aspirin is theoretical,
as no untoward clinical effect from aspirin use has been demonstrated.
Caution should also be exercised in the use of sedatives during thyroid
storm. Sedation depresses the level of consciousness and reduces the value
of this parameter as an indicator of clinical improvement. Sedation may
also cause hypoventilation.
Congestive heart failure should be treated with digitalis and diuretics
even though congestive failure due to hyperthyroidism may be refractory
to digitalis. Cardiac arrhythmias are treated with the usual antiarrhythmic
agents. Atropine should be avoided, as its parasympatholytic effect may
accelerate the heart rate. Atropine also may counteract the effect of
Intravenous glucocorticoids equivalent to 300 mg of hydrocortisone per
day should be given. The role of the adrenal glands in the pathogenesis
of thyroid storm is uncertain, but the use of hydrocortisone has been
reported to increase the rate of survival. Dexamethasone offers an advantage
over other glucocorticoids as it decreases the peripheral conversion of
T4 to T3.
Inhibition of Thyroid Hormone Synthesis
The antithyroid drugs propylthiouracil (PTU) and methimazole act to block
the synthesis of thyroid hormone by inhibiting the organification of tyrosine
residues. This action begins within 1 h after administration, but a full
therapeutic effect is not achieved for weeks. An initial loading dose
of PTU, 900 to 1200 mg, should be given, followed by 300 to 600 mg daily
for 3 to 6 weeks or until the thyrotoxicosis comes under control. Methimazole,
90 to 120 mg initially followed by 30 to 60 mg daily, is an acceptable
alternative. Both preparations must be given orally or via a nasogastric
tube, as no parenteral form is available. The PTU has an advantage over
methimazole because it inhibits the peripheral conversion of T4 to T3
and produces a more rapid clinical response. Although these drugs inhibit
the synthesis of new thyroid hormone, they do not affect the release of
Retardation of Thyroid Hormone Release
Iodide administration promptly retards thyroidal release of stored hormones.
Iodide can be provided by various preparations. It can be given as strong
iodine solution, 1 mL 3 times daily, as potassium iodide, 10 drops of
a solution containing 1 g/mL every 4 to 6 h, or as sodium iodide, 1 g
every 8 to 12 h by slow intravenous infusion. Caution should be used in
those patients taking potassium-sparing diuretics or potassium-containing
drugs as potassium iodide preparations will add to the potassium load.
Concomitant use of iodides and lithium salts may result in an additive
hypothyroid effect. Iodide should be administered 1 h after the loading
dose of antithyroid medication to prevent utilization of the iodide by
the thyroid in the synthesis of new hormone.
Blockade of Peripheral Thyroid Hormone Effects
Adrenergic blockade is a maintstay of therapy for thyroid storm. Currently,
the ?-adrenergic blocking agent propranolol is the drug of choice. In
addition to reducing sympathetic hyperactivity, propranolol also partially
blocks the peripheral conversion of T4 to T3.
Propranolol can be given intravenously at a rate of 1 mg/min with cautious
incremental increases of 1 mg every 10 to 15 min to a total dose of 10
mg. The effects of the drug in controlling cardiac and psychomotor manifestations
of storm should be seen in 10 min. The lowest possible dose required to
control thyrotoxic symptoms should be used, and this dose can be repeated
every 3 to 4 h as needed. The oral dose of propranolol is 20 to 120 mg
every 4 to 6 h. When given by mouth, propranolol is effective in about
1 h. Propranolol has been used successfully in treatment of thyroid storm
in childhood. Younger patients may require a dosage as high as 240 to
320 mg/day orally.
The usual precaution of avoiding propranolol in patients with bronchospastic
disease and heart block should be observed. Electrocardiographic evaluation
for conduction disturbance should occur before the administration of propranolol.
In patients with congestive heart failure, the benefit of slowing the
heart rate and controlling certain arrhythmias must be weighed against
the risk of depressing myocardial contractility with ?-adrenergic blockade.
Urbanic believes the benefit outweighs the risk in this situation but
recommends administration of digitalis before propranolol.
Guanethidine and reserpine also provide effective autonomic blockade and
are alternatives to propranolol. Guanethidine depletes catecholamine stores
and blocks their release. When given 1 to 2 mg/kg per day orally (50 to
150 mg), it is effective in 24 h, but it may not have its maximum effect
for several days. Toxic reactions are cumulative and include postural
hypotension, myocardial decompensation and diarrhea. It has an advantage
over reserpine because it does not cause the pronounced sedation observed
with that drug.
Reserpine acts to deplete catecholamine stores. As the initial dose, 1
to 5 mg is given intramuscularly, followed by 1 to 2.5 mg every 4 to 6
h. Improvement may be seen within 4 to 8 h. Side effects include sedation;
psychic depression, which can be severe; abdominal cramping; and diarrhea.
Finally, a thorough evaluation for a precipitating cause of thyroid storm
should be made. The treatment of thyroid storm should not be delayed by
this evaluation, which may have to wait until the patient is at least
Following the initiation of therapy, symptomatic improvement should occur
within a few hours, primarily due to adrenergic blockade. Resolution of
thyroid storm requires degradation of the already-circulating thyroid
hormones, whose biological half-life is 6 days for T4 and 22 h for T3.
Storm may last from 1 to 8 days, with an average duration of 3 days. If
conventional therapy is not successful in controlling storm, alternative
therapeutic modalities include peritoneal dialysis, plasmapheresis, and
charcoal hemoperfusion to remove circulating thyroid hormone. Following
recovery from thyroid storm, radioactive iodine therapy is the treatment
of choice for hyperthyroidism. The mortality in untreated thyroid storm
approaches 100 percent. Decreased mortality has occurred with the use
of antithyroid drugs. Underlying illness is the cause of death in many
cases. Prevention of thyroid storm is the ultimate solution to reduce
1)Cooper DS: Antithyroid drugs. N Engl J Med 311:1353, 1984.
2)Davis PJ, Davis FB: Hyperthyroidism in patients over the age of 60 years:
Clinical features in 85 patients. Medicine 53:161, 1974.
3)Hay ID, Bayer MF, Kaplan MM, et al: American thyroid association assessment
of current free thyroid hormone and thyrotropin measurements and guidelines
for future clinical assays. Clin Chem 37:2002, 1991.
4)Mazzaferri EL, Reynolds JC, Young RL, et al: Propranolol as primary
therapy for thyrotoxicosis: Results of a long-term prospective study.
Arch Intern Med 136:50, 1976.
5)Thomas FB, Mazzaferri EL, Skillman TG: Apathetic thyrotoxicosis: A distinctive
clinical and laboratory entity. Ann Intern Med 72:679, 1970.
6)Woeber KA: Thyrotoxicosis and the heart. N Engl J Med 327:94, 1992.