Is the Risk of Thyroid Storm Higher When Patient-Related Factors Delay Treatment of Hyperthyroidism,Or Is It Higher When Patients Undergo Thyroidectomy While Still Thyrotoxic?

In Japan, 0.22% of hospitalized thyrotoxic patients are diagnosed with thyroid storm; the mortality rate is 10.7%. ¹ In the United States, utilizing an inpatient database, 16.2% of hospitalized thyrotoxic patients are diagnosed with thyroid storm; the mortality is up to 3.6%. ² The difference in these two reports may be partly due to the more stringent requirements for the diagnosis of thyroid storm using the Japanese criteria versus the Burch–Wartofsky criteria, ³ or the overzealous use of the term “storm” when using hospital discharge codes in the United States. Regardless, thyroid storm is associated with significant morbidity and mortality. Thyroidectomy as well as nonthyroidal surgery and anesthesia have been reported to be precipitants of severe thyrotoxicosis. ³

The 2016 American Thyroid Association (ATA) guidelines state that hyperthyroid patients should be rendered euthyroid before surgery “whenever possible,” to prevent thyroid storm. ³

In this issue of Thyroid, Fazendin et al. from the University of Alabama, Birmingham, report that surgery for hyperthyroidism can be done safely before a euthyroid state is obtained. ⁴ While they suggest that achieving a euthyroid state is ideal, they indicate that patient-related factors may make it difficult to achieve a euthyroid state preoperatively. They compared 134 euthyroid patients who had a thyroidectomy after attaining a euthyroid state with medical treatment with 141 patients who were hyperthyroid at the time of surgery.

Of note, only 69 patients had a free thyroxine level that exceeded 3.12 ng/dL (the upper quartile of their hyperthyroid patients). None of the patients had postoperative thyroid storm, although 11 patients had been previously admitted to the hospital with thyroid storm. The only differences between the two groups were longer operative time (>1 hour in 19.8% of hyperthyroid patients vs. 7.3% of euthyroid patients), more blood loss (20 vs. 15 mL), and a higher incidence of transient hypocalcemia (13.4% vs. 4.7%). The authors conclude that thyroidectomy when hyperthyroid is safe.

They cite several other studies in which surgery while hyperthyroid was reported to be safe, including one in which the hyperthyroid patients required a higher use of intraoperative beta-blockers. ⁵ They did not cite articles from the 1980s in which 1 of 22 patients given beta-blockers alone, ⁶ and 4 of 966 patients treated with iodine and steroids ⁷ developed thyroid storm.

Allergy or intolerance to antithyroid drugs is a potential indication for thyroidectomy in a hyperthyroid patient, and several series have suggested alternative preoperative regimens using beta-blockers, iodine-containing medications, and steroids. ^8,9 However, only 12% of the patients from the Alabama study were intolerant to antithyroid drugs. Rather, the authors cite the high number of patients living below the poverty line in Alabama, a medically underserved population due to financial limitations and poor access. Alabama ranks 7th in the United States with 14.96% of the population living in poverty versus Massachusetts that ranks 43rd with 9.11% of the population living in poverty. ¹⁰ Thus, even in parts of the United States rich in medical resources, there are still patients in whom successful preoperative treatment of hyperthyroidism may be an elusive goal, and these data from Alabama are widely applicable.

However, I am uncertain whether the message from this study should be as simple as the title of the article suggests. Since they used the Japanese criteria for thyroid storm, and were likely avoiding overzealous use of the term, their study was probably underpowered to detect a significant increase in thyroid storm given the low 0.22% incidence reported in Japan. ¹ In addition, based on my clinical experience, I suspect the authors have seen some patients whose hyperthyroidism was significantly exacerbated perioperatively, even though they did not meet the criteria for a diagnosis of thyroid storm.

I have referred many hyperthyroid patients for surgery for several decades either because they were allergic to antithyroid drugs, or because of poor compliance or other patient-related factors prevented me from rendering them euthyroid. Most had no perioperative issues, but a few required very high doses of beta-blockers for tachycardia and arrhythmias, and ultimately did well utilizing the expertise available in an academic medical center. I am concerned that outcomes might not be as good if this study were to encourage thyroid surgery in hyperthyroid patients in a community setting.

The important message that this study suggests to me is a message that parallels our treatment paradigms for amiodarone-associated hyperthyroidism. We now accept the fact that many patients, some with severe underlying heart disease, whose thyrotoxicosis is not responding to antithyroid drugs or steroids, may be at higher risk for death or morbidity from weeks of attempted outpatient management than they are from an hour or two in the operating room for a definitive thyroidectomy. ³ Similarly, this study from Alabama suggests that there is a group of patients, seen more commonly in areas with low socioeconomic status, who, for whatever reason, remain hyperthyroid despite initial attempts to treat them medically. Such patients may be better served by early surgery when hyperthyroid, rather than waiting until they present to the emergency department with severe hyperthyroidism or thyroid storm.

The ATA guidelines, which recommends that patients be rendered euthyroid before surgery whenever possible, are still valid. ³ While the risk of an adverse outcome from operating on a thyrotoxic patient may be quite low, the risk should not be considered negligible. But this report also illustrates that perhaps the risk of not operating on a noncompliant thyrotoxic patient is also not negligible; 11 of their patients were hospitalized with thyroid storm before surgery, none after surgery. Most patients have a satisfactory symptomatic response to antithyroid drugs within 4–6 weeks of initiation of treatment. But some patients do not respond, or are not expected to respond, due to patient-related or societal factors. The study from Alabama suggests that we need to balance the risks of surgery while still thyrotoxic with the risk that the thyrotoxicosis will worsen during prolonged, and frequently futile, attempts to control the hyperthyroidism medically. For some patients, surgery while thyrotoxic may be the better choice.