Thoughts on the Japanese and American Perspectives on Thyroid Storm

Thyroid storm is a rare, acute, life-threatening exacerbation of the clinical consequences of thyrotoxicosis, with reported mortality rates ranging from 2% to 75% in hospitalized patients. With the more widespread awareness of thyroid diseases, and improved diagnostic instruments and approaches to treatment of thyrotoxicosis, death from thyroid storm or from its treatment should be rare, and thyroid storm appears to be occurring less frequently today than in the past (1). Nevertheless, it may occur in 1% to 2% of hospital admissions for thyrotoxicosis (2), and therefore it must still be considered one of the more important and prevalent endocrine emergencies.

In 2012, we commented in Thyroid (1) on a landmark study by Akamizu et al. (3) describing two national Japanese surveys of thyroid storm that formulated diagnostic criteria for this entity and evaluated its incidence, clinical features, mortality, and the likelihood of survival with complications. The authors first developed diagnostic criteria based on 99 patients in the literature and seven of their own patients. Thyrotoxicosis was a prerequisite for both sets of diagnostic criteria. Additional criteria were a combination of classical features such as central nervous system (CNS) manifestations, fever, tachycardia, congestive heart failure, and gastrointestinal or hepatic disturbances.

We commended (1) the study by Akamizu et al. (3) as probably the best surrogate for a prospective study, and our hope was that clinical thyroidologists and epidemiologists in other countries around the world would perform similar studies to better understand the presentation of thyroid storm in genetically different populations, those with different levels of iodine intake, and those with other underlying risk factors. Thyroid storm in Japan was estimated to occur in 0.2% of all thyrotoxic patients and 5% of thyrotoxic inpatients (3). The primary triggers of thyroid storm were inconsistent use or discontinuation of antithyroid drug therapy for Graves' disease and infection.

There was subsequently a new storm of ∼250 case report publications on thyroid storm in PubMed. However, there are still no prospective trials or cohort or epidemiological studies. The Japanese group has recently developed their national survey results into Japanese guidelines, and a review of these has appeared in the January 2018 issue of Thyroid (4).

More recently, in the January 2019 issue of Thyroid, Galindo et al. (5) determined the incidence of thyrotoxicosis with and without thyroid storm and clinical outcomes among hospitalized patients from 2004 to 2013 in the United States. They used data from the National Inpatient Sample database, the largest public inpatient database, which includes a representative sample of non-federal U.S. hospitals. Among 121,384 patients admitted with thyrotoxicosis, 78% were female, half were Caucasian, and the mean age was around 50 years. A total of 19,723 patients (16.2%, or 1 of every 6 admissions with thyrotoxicosis) were diagnosed with thyroid storm. Thus, during the past decade, the incidence of thyroid storm ranged between 0.57–0.76 cases/100,000 U.S. persons per year, and 4.8–5.6/100,000 hospitalized patients per year, and were thus higher than the Japanese figures.

This being mentioned, it should be emphasized that there is no global “gold standard” diagnostic test for thyroid storm. The diagnosis is based on clinical assessment and in the United States, the most common clinical score is based on the Burch Wartofsky scale. The manuscript by Galindo et al. (5) used the International Classification of Diseases discharge codes for the diagnosis of thyroid storm, thus it is not known how those diagnoses were made. The Japanese, on the other hand, applied their own separate scoring system. This was developed on the basis of an original two-phase nationwide analysis of clinical and laboratory features of thyroid storm. This involved thyrotoxic patients without storm as well as 99 patients in the initial phase with a final dataset of 356 patients whose features formed the basis for their scoring system. Their score differs somewhat from the Burch–Wartofsky score.

In the study by Galindo et al. (5) thyroid storm was associated with a 12-fold increase in mortality compared to patients who had thyrotoxicosis without storm. In this study, the largest on the topic to date, the authors concluded that the incidence and mortality of thyroid storm has not substantially changed in the past decade, although hospitalization costs related to treatment of thyroid storm have significantly increased. Readers of the study by Galindo et al. (5) should, however, note that the data for costs were based on “the amount needed to produce the service,” not actual charges. Moreover, physician fees were not included in the cost estimates.

Another very recent study from Texas (6), covering the period from 2011 to 2017, found that patients with thyrotoxicosis and lack of health insurance had 12 times higher odds of being hospitalized for complicated thyrotoxicosis compared to patients with commercial insurance. Conversely, the odds of hospitalization for complicated thyrotoxicosis were reduced by 63% in patients with higher median income, and by 33% in those with college or university education versus a high school education. Thirty-two percent of patients hospitalized for complicated thyrotoxicosis presented with thyroid storm, which accounted for 7% of the studied cohort. These figures are very much in line with a publication by Sherman et al. published in 1996, where similar socioeconomic and insurance status predispositions to complicated thyrotoxicosis were described (7).

In Japan, multiple organ failure was the most common cause of death in the patients with thyroid storm (3), followed by congestive heart failure, respiratory failure, arrhythmia, disseminated intravascular coagulation, gastrointestinal perforation, hypoxic brain injury, and sepsis. Higher Glasgow Coma Scale scores and serum urea nitrogen concentrations were associated with irreversible brain or other damage. Interestingly, among the patients with thyroid storm, the occurrence of any one of the five major groups of clinical manifestations (CNS, fever, tachycardia, congestive heart failure, gastrointestinal/hepatic) did not positively correlate with the occurrence of any other group of clinical manifestations (1).

Another interesting observation in the Japanese study was the finding of similar values for serum free thyroxine and free triiodothyronine in patients with and without thyroid storm. This implies that the concentrations of thyroid hormones do not predict the progression of thyrotoxicosis to thyroid storm. The occurrence of multiorgan failure during the development of thyroid storm will add an overlay of nonthyroidal illness to the thyroid function tests, thereby lowering first serum triiodothyronine (T3) followed by serum thyroxine (T4) (8). It was noted in the Japanese recommendations (4) that there is an inverse relationship between the T3/T4 ratio and thyroid storm severity.

Based on the seriousness of the condition and its relatively high prevalence in Japan, the Japanese group created a treatment strategy for thyroid storm which was intended to be detailed, concrete, useful in clinical practice, evidence-based, and possibly internationally applicable (4). Based on their studies and literature review, there are three treatment recommendations that differ from those made in the 2016 American Thyroid Association guidelines (9).

Thus, the Japanese recommendations include using methimazole instead of propylthiouracil (which is usually recommended in European and American guidelines to obtain the double benefit of blocking thyroidal hormone production and blocking peripheral conversion of T4 to T3). Their argument was that methimazole has fewer side effects and is superior in both thyroid storm and non-storm Graves' patients (4,10), although we would note that the propylthiouracil dose equivalent was much lower than that of methimazole in that study, considering the higher potency of methimazole (11).

A strong argument for the superiority of propylthiouracil in the guidelines of the American and European Thyroid Associations was the more effective lowering of T3 concentrations to 45% after 24 hours compared to only 10–15% lower after methimazole (11,12). In contrast to this, the Japanese showed that methimazole normalized T3 more rapidly than propylthiouracil in non-storm patients. In their hands the outcomes were identical with propylthiouracil or methimazole in patients with storm, and furthermore, in general, side effects less common with methimazole. Whether these discrepancies might be related to differences in iodine intake is unclear.

While in the United States, the recommendation is to delay the administration of iodine for one hour after starting antithyroid drug treatment (9), the Japanese recommend starting both agents concomitantly. In both Europe and the United States, concomitant use is considered relatively contraindicated, since iodine would be expected to exacerbate hyperthyroidism due to toxic nodular goiter (but not Graves' disease). Prior administration of the antithyroid drug (if possible by the intravenous or rectal route due to questionable gastrointestinal absorption in these patients) allows the antithyroid drug to enter into the thyroid gland before it is overloaded by inorganic iodide, hence, minimizing the organification of iodide.

The Japanese recommendations are all based on Japanese studies, and therefore derived from a population with much higher iodine intake than in most other regions. It is very likely that similar studies performed in other populations and regions might result in different results. In many areas of Europe T3 thyrotoxicosis is still prevalent, and higher T3/T4 ratios are likely to be far more common in areas with historically low iodine intakes than is the case in Japan. Because of the much higher iodine consumption in Japan the prevalence of toxic nodular goiters is lower and, therefore, the risk of worsening of the thyrotoxicosis is lower.

One of the reasons the American Thyroid Association recommends waiting an hour is due to the uncertainty as to the etiology of the thyrotoxicosis. Strangely enough, the very recent European Thyroid Association guidelines completely fail to mention iodine as part of the acute therapy for thyroid storm (12), and do not discuss why.

Further recommended treatment strategies summarized in the Japanese guidelines, such as supplementing with intravenous glucocorticoids, fluid resuscitation, therapy to treat the precipitating cause, use of beta-blockers to control tachycardia, and supportive treatment for the multiorgan failure were similar to other recommendations. There is, however, a third difference in the guidelines, since the American and European Thyroid Associations specifically suggest propranolol because it also blocks conversion of T4 to T3 (esmolol is mentioned as an alternative in the table). The Japanese reported excessive deaths in patients with chronic heart failure given propranolol and therefore suggest beta-selective agents such as esmolol (4), which might be optimal also considering the very short half-life of this drug. Any major cardiac adverse reaction can rapidly be ameliorated by discontinuing the drug.

The Japanese group headed by Dr. Akamizu (4) should be highly commended for bringing the serious and clinically-relevant issue of thyroid storm into focus and for providing guidance. The recent U.S. register study is another important piece of the puzzle, demonstrating the still very high (probably higher than anticipated by most) incidence and prevalence of thyroid storm among U.S. inpatients with thyrotoxicosis (5).

We hope that the Japanese recommendations and the U.S. registry study will serve as inspiration for others to perform similar studies in other regions with different iodine intakes and ethnicities to further clarify the global epidemiology and clinical management of thyroid storm. Prospective studies may not be feasible given the overall rarity of thyroid storm. The introduction of truly universal guidelines, as suggested by the Japanese group, also may not be feasible due to regional differences in epidemiology, diagnostic measures, iodine intake, and management practices.