A 45 year old woman with a past medical history of hyperthyroidism is brought into the emergency department by her partner. She has had urinary tract symptoms for the last 2 days, but now she looks really sick. She has been having diarrhea and crampy abdominal pain. She is having a hard time concentrating, is tremulous, and drenched in sweat. Her vital signs at triage include an irregular heart rate of 142 and a temperature of 39.5 degrees celsius. Do you have an approach to thyroid storm?
My approach to thyroid storm
The diagnosis of thyroid storm is clinical. Laboratory testing, although helpful, does not distinguish between thyrotoxicosis and thyroid storm.1 Treatment should be started immediately based on clinical findings.2 The key clinical features are: fever, altered mental status, and signs of sympathetic over-activity.
Support airway and respiration as necessary.
Manage agitation. Significant agitation can interfere with the other urgent interventions that are required. If the patient is significantly agitated, treat with a benzodiazepine.3
- Midazolam 5-10mg IV q5min as needed
Take control of the body temperature. Start cool IV fluids and external cooling as necessary.2,3
Begin fluid resuscitation. These patients will essentially all require fluids.1 Some will be in heart failure as part of their presentation, but it is high output heart failure, and even these patients will probably need fluids. The patients will also have low glycogen reserves, so adding sugar to your fluids makes sense.1
Start a beta-blocker. This will help control cardiac arrhythmias and will also make the patient more comfortable by controlling tremor. Propranolol is the agent typically used, because it inhibits the peripheral conversion of T4 to T3 and is non-cardioselective so it will also control symptoms such as agitation, fever, and psychosis.1 Esmolol is also a reasonable choice.2 The target should be a heart rate less than 90.3 Note: EMCrit suggests a target heart rate of 100, because there are some case reports of cardiovascular collapse after propranolol that might be the result of these patients requiring a rapid heart rate as a compensatory mechanism.
- Propranolol: start with 0.5-1 mg slow IV push, then titrate 1-3 mg IV q 10 min1
- Esmolol 500 mcg/kg load over 1 min, then 50-200 mcg/kg/min1
- If the patient has a significant contraindication to beta-blockers, you can use reserpine 2.5-5 mg IM q4h1
Start a thyrostatic agent (thionamide) to lower thyroid hormone levels. Propylthiouracil is the preferred agent, because it also inhibits the peripheral conversion of T4 to T3.3 Both of these agents can be given through a NG tube or rectally if needed.1
- Propylthiouracil 600-1000 mg PO (loading dose)1
- Methimazole 20-25 mg PO (loading dose). An IV version is available in Europe, but not in North America.1
Give a dose of glucocorticoid. Thyroid storm typically causes depression of the hypothalamic-pituitary axis. Glucocorticoids also inhibit the peripheral conversion of T4 to T3. 2
- Hydrocortisone 300 mg IV2
Search for the underlying cause. For hypothyroidism to progress to thyroid storm, there generally has to be a precipitating event.1,3 The most common cause is going to be infection, and as these patients are going to present hot and altered, I would routinely culture and start empiric antibiotics on any patient I am treating for thyroid storm.1 Other important causes to consider are: myocardial infarction, DKA, pregnancy, and trauma.3 Amiodarone is another potential trigger (because of the iodine) and the diagnosis might be partially masked because of the beta-blocking properties of amiodarone.
In refractory cases, or if there are absolute contraindications to the above medications, other therapeutic options are lithium, hemodialysis, charcoal hemoperfusion, and plasmapheresis.1,3
- Lithium 300mg every 8 hours3
Iodine administration is not part of the urgent management of these patients because it must be given at least 1 hour after the thyrostatic medication. These patients will all get iodine, but there is not a rush as long as you have started the other therapies, so it can be left the the admitting team.
Without treatment, thyroid storm is universally fatal. With treatment, the mortality is still as high as 20-50%.1
Salicylates should be avoided. They can make the thyrotoxicosis worse by decreasing binding to thyroid binding protein and therefore increasing levels of free T4 and T3.1
Most sources seem to advise against giving empiric antibiotics.1 However, we are generally working with limited information in the emergency department. If the patient is febrile and altered, I think antibiotics should be given up front and then sorted out later.
There are clinical criteria for thyroid storm. A score of 45 or more is highly suggestive of thyroid storm. A score of less than 25 makes thyroid storm unlikely. The grey area in between is exactly that, grey area, but could represent impending thyroid storm.2,3
Other FOAMed Resources
Thyroid Storm in the Internet Book of Critical Care
Thyroid Storm on the Life in the Fastlane CCC
Diagnosing hyperthyroidism: Answers to 7 common questions and Thyroid storm treatment strategies on ALiEM
You can find more First10EM resuscitation plans here.
- McKeown NJ, Tews MC, Gossain VV, Shah SM. Hyperthyroidism. Emergency medicine clinics of North America. 23(3):669-85, viii. 2005. [pubmed]
- Chiha M, Samarasinghe S, Kabaker AS. Thyroid storm: an updated review. Journal of intensive care medicine. 30(3):131-40. 2015. [pubmed]
- Sharp CS, Wilson MP, Nordstrom K. Psychiatric Emergencies for Clinicians: The Emergency Department Management of Thyroid Storm. The Journal of emergency medicine. 51(2):155-8. 2016. [pubmed]
Morgenstern, J. Thyroid storm, First10EM, November 21, 2016. Available at:
10 thoughts on “Thyroid storm”
Wouldn’t you withhold beta blockers if they were in cardiogenic shock?
Thanks for the comment Derek.
If I had a patient with thyroid storm and cardiogenic shock, I would definitely be getting expert advice as soon as possible. My sense is that although congestive heart failure is relatively common, cardiogenic shock is rare. This is high output heart failure, and beta blockers are definitely indicated.
It is all on the continuum.
The natural history of untreated cardiac failure (low or high output) is eventually cardiogenic shock. Shock being defined by inadequate flow to meet tissue metabolic demands. In the case of hyperthyrodism, the pathophysiology is increased cellular metabolism and autoregulative vasodilation. This initially leads to a hyperdynamic circulation. At this point the patient will not have significant evidence of pulmonary oedema or decreased perfusion. But as the conditions remains untreated, cardiac function cannot continue to meet this demand (partly due to tachycardia cardiomyopathy which is addressed by beta blockade) resulting in reduced renal perfusion and fluid retention leading to the vicious spiral of cardiac failure.
My question was more subtle. We need to make some initial evaluation of cardiac function and filling pressures in case further fluid or the sole addition of a negative inotrope effect will cause further haemodynamic deterioration in the patient. In rare situations a pressor or mechanical assist might be needed to support the patient through the period until metabolic activity normalises through the treatments you have already mentioned.
It all depends how sick the patient is, but a proportion with true storm will require invasive monitoring in the ICU.
Great post: well-written and very easy to follow. I like how you have broken down the key concepts and a couple of nice little nuggets such as the mechanism of action of propranolol. I have a question: in your hypothetical patient at the beginning, could she not also just have severe sepsis? How can you differentiate, since as you rightly pointed out TFTs aren’t necessarily going to diagnose thyroid storm in someone with a history of hyperthyroidism? Should we be screening our ‘septic’ patients for thyroid storm? In my – albeit limited – experience of thyroid storm the ‘altered mental status’ may perhaps be represented by increased psychomotor activity rather than confusion and drowsiness; and I guess BP may be elevated. And I’m not sure the Burch & Wartofsky scale would necessarily help either. What do you think?
Thanks for the comment Anu. I am not sure I have a perfect answer right now. There are a number of important diagnoses to consider in the hot an altered patient. It is a topic I am planning to cover in a future blog post, but it will require some research and thought. Currently I focus on neurologic features atypical to sepsis (tremor, clonus, agitation), tachycardia out of proportion, and most importantly reassessment and failure to respond to usual therapy.