Thyroid storm is a life-threatening endocrine emergency for which, remarkably, there are no definitive diagnostic tests, and the management of which is supported by a startlingly weak evidence base.
“What’s tricky is there really are no specific biochemical level cutoffs for thyroid storm, and also no unique laboratory abnormalities. So in the end, it’s a clinical diagnosis and a clinical judgment,” Stephanie B. Mayer, MD, MHSc, observed at HM20 Virtual, hosted by the Society of Hospital Medicine.
Moreover, there are no prospective clinical trials addressing the treatment of thyroid storm, and the 2016 American Thyroid Association clinical practice guidelines on the topic are based upon low-quality evidence from case reports and studies dating back to the 1970s and 1980s. UpToDate reached the same conclusion in 2020, noted Dr. Mayer, an endocrinologist at Virginia Commonwealth University, Richmond.
Thinking that perhaps the guideline writing panel had missed something, she asked a university medical research librarian to custom-build a comprehensive search for studies on thyroid storm management. The search proved unrewarding.
“The evidence is, unfortunately, a little disappointing,” Dr. Mayer said.
Thyroid storm is a rare condition, but one that hospitalists must be ready for. She highlighted current best practices in diagnosis and management.
A high-mortality emergency
Thyroid storm is an extreme manifestation of thyrotoxicosis, which is marked by multiorgan dysfunction and rapid decompensation. In a large, first-of-its-kind, national retrospective U.S. study, the incidence of thyroid storm was 0.57-0.76 cases per 100,000 persons per year. Thyroid storm accounted for 16% of the more than 121,000 hospital discharges featuring a primary diagnosis of thyrotoxicosis. The in-hospital mortality rate for patients with thyroid storm was 1.2%-3.6% during the 10-year study period, a rate 12-fold higher than that among patients with thyrotoxicosis without thyroid storm (Thyroid. 2019 Jan;29[1]:36-43).
Dr. Mayer highlighted a multicenter French study that underscored the current hefty morbidity and mortality associated with thyroid storm. Among 92 patients admitted to the ICU for thyroid storm, the in-ICU mortality rate was 17%, and the mortality rate 6 months after admission was 22%. Independent risk factors for in-ICU mortality were multiorgan failure and the occurrence of cardiogenic shock within the first 48 hours in the ICU (Crit Care Med. 2020 Jan;48[1]:83-90).
How to recognize thyroid storm
The most user-friendly system for assistance in diagnosing thyroid storm is the one put forth by the Japan Thyroid Association and the Japan Endocrine Society, in Dr. Mayer’s view. As a prerequisite to the diagnosis a patient must have thyrotoxicosis as evidenced by elevated free thyroxine (free T4) and free or total triiodothyronine (T3), which in the vast majority of cases, is accompanied by low thyroid stimulating hormone (TSH).
The Japanese diagnostic system for thyroid storm relies on five categories of organ system–based clinical features. This approach places greater weight on disturbances of consciousness – restlessness, delirium, agitation, psychosis, lethargy, coma – than the other four components, which consist of fever of at least 100.4° F, tachycardia of 130 or more beats per minute, heart failure signs and symptoms, and gastrointestinal/hepatic involvement as evidenced by nausea, vomiting, hyperdefecation, and/or a total bilirubin level of 3.0 mg/dL or more.
The Japanese approach offers two paths to a definite diagnosis of thyroid storm. One requires at least one CNS manifestation plus symptoms drawn from any one of the other four categories. The other route, for patients without evident CNS symptoms, requires the presence of symptoms from at least three of the other four categories.
A patient is categorized as having suspected rather than definite thyroid storm if the CNS criterion isn’t met but any two of the others are. A patient also qualifies for suspected thyroid storm when CNS manifestations plus symptoms from at least one other category are present, but thyroid hormone levels aren’t available (Endocr J. 2016 Dec 30;63[12]:1025-64).
