- © 2013 by American Society of Clinical Oncology
Case of Sorafenib-Induced thyroid storm
- Sigurdis Haraldsdottir⇑,
- Quan Li,
- Miguel A. Villalona-Calero,
- Thomas E. Olencki,
- Kari Kendra and
- Steven W. Ing
- The Ohio State University Comprehensive Cancer Center, Columbus, OH
- The Ohio State University Medical Center, Columbus, OH
- Corresponding author: Sigurdis Haraldsdottir, MD, Division of Medical Oncology, 365 M Starling Loving Hall, 320 West 10th Ave, Columbus, OH 43210; e-mail: Sigurdis.Haraldsdottir{at}osumc.edu.
Case Report
We present the case of a 72-year-old man with a past medical history of atrial fibrillation who was diagnosed with stage II renal clear cell carcinoma in 2002 and had a radical nephrectomy. The patient had no evidence of recurrence until 9 years later, when he presented with shoulder pain and was found to have metastasis to the deltoid as well as to bone, adrenal gland, and lung. He had been receiving amiodarone for atrial fibrillation for 1 year before his metastatic cancer diagnosis, but this was stopped preemptively to prevent drug interactions before the initiation of tyrosine kinase inhibitor (TKI) treatment. The patient received pazopanib for 6 months when, on disease progression, he started receiving sorafenib 400 mg twice per day. Deterioration was noted after 2 to 3 weeks of therapy and progressed over 5 weeks with anorexia, diaphoresis, fatigue, tremors, diarrhea, and intermittent confusion.
After 8 weeks of sorafenib therapy, the patient was hospitalized with worsening symptoms and atrial fibrillation. He was agitated and had a pulse of 137 beats/min. He had a nonenlarged, non-nodular, nontender thyroid gland and a fine resting tremor. He had markedly abnormal thyroid function tests (TFTs) with an undetectable thyroid-stimulating hormone (TSH) of < 0.008 μIU/mL, elevated free T4 (fT4) of 4.06 ng/dL (normal range, 0.89 to 1.76 ng/dL), T4 of 19 μg/dL (normal range, 4.5 to 10.9 μg/dL), and mildly elevated fT3 of 4.4 pg/mL (normal range, 2.3 to 4.2 pg/mL). Figure 1 depicts the changes in TSH and fT4 since the diagnosis of metastatic disease; the initiation of TKI therapy is marked with arrows. Thyroglobulin was elevated at 115 ng/mL (normal range, 1.6 to 59.9 ng/mL), whereas antithyroglobulin, antithyroid peroxidase, and thyroid-stimulating immunoglobulin were undetectable. Ultrasonography revealed a heterogeneous and atrophic right thyroid lobe with no focal lesions; the left lobe was not visualized. thyroid storm was diagnosed and treated with high-doses of propylthiouracil, hydrocortisone, and propranolol. The patient's confusion persisted, and atrial fibrillation remained poorly controlled on intravenous diltiazem and a beta blocker. Three days after admission the patient had a cardiac arrest in the setting of aspiration and died.
Discussion
TKI-induced thyroid dysfunction was first reported by Desai et al1 with the observation of two index cases of hypothyroidism in patients receiving sunitinib for imatinib-resistant GI stromal tumors. This led to a prospective observational study of 42 patients receiving sunitinib that revealed the development of persistent, primary hypothyroidism in 15 (36%) of them. Six of these patients had asymptomatic suppression of TSH concentrations before developing hypothyroidism, which suggested a process of thyroiditis that led to hypothyroidism.1
Four TKIs—sunitinib, sorafenib, pazopanib, and axitinib—have been approved by the US Food and Drug Administration for use in metastatic renal cell carcinoma. They all have antiangiogenic, antiproliferative, and proapoptotic effects by targeting vascular endothelial growth factor receptor (VEGFR) -1, VEGFR-2, VEGFR-3, platelet-derived growth factor receptor α/β, and (weakly) FLT3, as well as c-Kit, RAF, and RET receptor tyrosine kinase.2,3 They have all been associated with thyroid dysfunction,4,5 mainly hypothyroidism with sunitinib, which was the first reported, with incidence ranging from 27% to 85% of patients6,7; 30% of patients require thyroid hormone replacement therapy.7 Sorafenib has been reported to cause TFT abnormalities in 8% to 67% of patients6,8,9; less than 30% of patients are given thyroid hormone replacement therapy. Pazopanib has the lowest reported incidence, with less than 10% of patients developing hypothyroidism (grade 3 to 4 events in < 1%) in a phase III trial.4 There are currently no available data on TSH trends in patients receiving pazopanib. Early reports on axitinib have shown an even higher incidence: 89% of patients had TSH elevations in a recently published phase I study in which proactive thyroid hormone replacement therapy prevented the development of grade 3 to 4 fatigue.10 A recent cohort study in Germany11 revealed that 13.7% of patients receiving sunitinib and 6.3% of patients receiving sorafenib required thyroid hormone replacement therapy during TKI therapy.
Our patient had normal baseline TFTs (Fig 1) and presented with thyroid storm 8 weeks after initiating sorafenib. Although a thyroid uptake scan to assess the cause of thyrotoxicosis was unobtainable because of propylthiouracil administration, thyroiditis was supported by an elevated serum thyroglobulin concentration that was out of proportion to the size of the atrophied thyroid gland, a lack of thyroid autoantibodies, and disproportionally low fT3 and total T3 levels relative to the degree of T4 elevation. TFTs improved after therapy initiation, which suggests some biochemical response to treatment.
Sorafenib is extensively metabolized in the liver through CYP4503A4 and UGT1A9. There were no signs of liver dysfunction in this patient, nor did he have known liver metastasis, so it is unlikely that impaired metabolism resulted in the toxicities. No significant drug-drug interactions between sorafenib and other medications that the patient was taking were detected, and the patient had not received any iodine-containing contrast after initiating sorafenib. The Naranjo score12 for all concurrent medications was either 0 (doubtful) or 1 (possible), making it unlikely that they caused thyroid storm, whereas sorafenib scored 7 (probable). In theory, thyroid storm secondary to the remote use of pazopanib cannot be completely eliminated in this case, although the likelihood is minimal compared with sorafenib. This is supported by the half-time of pazopanib being 30.9 hours and a lack of reports in the literature.
Several mechanisms have been suggested to explain TKI-induced thyroid dysfunction. In animal models, VEGF inhibition causes capillary regression to several organs, with the greatest regression to the thyroid gland, suggesting that the thyroid gland is particularly sensitive to this inhibition.13 Sunitinib has been shown to have anti–thyroid peroxidase activity in vitro14; furthermore, blocking of iodine uptake by TKIs has also been proposed.15 Clinically, TKI-induced thyroid dysfunction does not seem to be mediated by autoimmune mechanisms, given that anti–thyroid peroxidase activity and antithyroglobulin antibodies have been found to be negative in most patients with thyroid dysfunction.1,7,15 It is possible that more than one mechanism causes TKI-induced thyroid dysfunction, given that some patients do develop thyroiditis with or without symptoms of hyperthyroidism, whereas others do not. It is also likely that the incidence of thyrotoxicosis is underestimated, given the fact that the symptoms are nonspecific and overlap with other TKI-induced adverse effects.
Four different reports8,9,16,17 have described sorafenib-induced thyrotoxicosis with associated thyroiditis in the literature, but none of the cases were severe or required aggressive therapy. In a prospective observational study of TFT abnormalities in 69 patients treated with sorafenib, 68% developed biochemical hypothyroidism (6% required therapy). Eleven of these patients (24%) first showed a suppressed TSH before progressing to hypothyroidism, and the transient period of thyrotoxicosis was apparently asymptomatic.8
Grossman et al18 described a fatal case of sunitinib-induced thyrotoxicosis that occurred 9 weeks after initiation of therapy; the patient was treated with propylthiouracil and IV glucocorticoids but died 3 days after admission with worsening encephalopathy and multiorgan failure. Optimal therapy for TKI-induced thyrotoxicosis is unknown. Our patient did demonstrate evidence of a biochemical improvement of TFTs after therapy was initiated to treat the thyroid storm, but whether he would have shown overall clinical improvement is unknown because of the fatal aspiration event that occurred a few days after hospitalization. If thyroiditis is the cause for most cases of TKI-induced thyrotoxicosis, initiating antithyroid medication is unlikely to be helpful. Although one previous case report demonstrated benefit with glucocorticoids,16 our patient did not benefit.
The development of hypothyroidism has been associated with clinical efficacy. The two largest studies to date showed significant increases in progression-free survival19 (16 v 6 months; P = .032 in patients with and without hypothyroidism) and objective remission rates6 (28.3% v 3.3%; P < .001) in patients with metastatic renal cell carcinoma receiving sunitinib or sorafenib. It is unclear whether this reflects higher drug concentrations or TSH effects on tumor growth, and it is not known if treatment of hypothyroidism affects the clinical efficacy.
Risk factors for thyroid dysfunction development on TKI therapy have not been elucidated, but increasing age did seem to be a risk factor in the prospective study.8 Schmidinger et al6 did not observe an association with sex or age. It is possible that previous amiodarone therapy, which is well known to cause thyroid dysfunction, made this patient more susceptible to developing sorafenib-induced thyroiditis. Previous TKI therapy with sunitinib/sorafenib has not been associated with increased incidence of thyroid dysfunction when another TKI therapy is initiated,6 but previous treatment with pazopanib, which our patient was receiving before sorafenib, has not been investigated.
Guidelines for thyroid function monitoring have not been developed, but several authors have suggested monitoring TSH for the first 4 months of therapy, given that thyroid dysfunction tends to occur early. Wolter et al20 suggested monitoring TSH on days 1 and 28 of the first four cycles of sunitinib therapy and, if normal, then on day 28 every 3 cycles. Patients with preexisting thyroid dysfunction may require closer monitoring, and patients who discontinue therapy may also require monitoring, given that thyroid dysfunction can resolve once therapy is stopped.
To our knowledge, this is the first report of sorafenib-induced thyroid storm. It can occur within a few weeks of starting TKI therapy and may be under-recognized because symptoms of thyroid storm can overlap with symptoms of other TKI-related toxicities. Although optimal therapy is not known, treatment should include therapy for severe hyperthyroidism and general supportive care. thyroid storm requires prompt recognition and treatment because of high mortality, and requires a high level of suspicion for diagnosis.
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The author(s) indicated no potential conflicts of interest.
