Thyrotoxicosis due to 1000-fold error in compounded liothyronine: A case elucidated by mass spectrometry

Highlights

• •Thyrotoxicosis due to liothyronine overdose may present with headache and gastrointestinal symptoms.
• •Tandem mass spectrometry can be used to rapidly investigate ingestions and overdoses.
• •Prescription of compounded pharmaceutical agents caries inherent risks.
• •The regulatory landscape for compounding pharmacies is in flux.

Abstract

Background

Thyrotoxicosis attributable exclusively to triiodothyronine (T3) is, by necessity, caused by accidental or intentional ingestion of pharmaceutical preparations. The clinical presentation of T3 overdose appears to differ from classic thyroid storm.

Case

A 30-year-old female patient presented serially to the emergency department with headache, nausea and vomiting. Neurological work-up was negative and she was treated for presumed viral gastroenteritis. Eventually she developed confusion and was admitted. Laboratory investigations showed a suppressed TSH and a free T3 above the linear range (>30 pmol/L), estimated by dilution in normal serum to be 330 pmol/L. She was diagnosed with thyrotoxicosis secondary to recently prescribed compounded liothyronine and was treated with seven rounds of plasmapheresis. Using a rapidly developed mass spectrometric method for T3, it was determined that compounding pharmacy had dispensed liothyronine at a concentration $\simeq 1000$-times the prescribed dosage.

Conclusion

The clinical and mass spectrometry laboratories played an essential role in the diagnosis of thyroid storm in this case of T3 overdose as the expected clinical features of hyperpyrexia, tachycardia and hypertension were initially absent.

1. Introduction

While there are occasions where manual compounding of pharmaceuticals is necessary [1], the practice has inherent risks [2], [3], [4] and has been associated with a number of overdoses and deaths. In British Columbia Canada, naturopathic practitioners are free to prescribe thyroid hormone, including manually compounded forms. Herein, we describe a case of iatrogenic thyrotoxicosis where large doses of liothyronine were accidentally administered due to a 1000-fold pharmaceutical compounding error.

2. Case

A 30-year-old female had been diagnosed with “hypothyroidism” by her family physician on the basis of a TSH of 4.41 mU/L (N: 0.5–5.0 mU/L) and a free thyroxine of 12 pmol/L (N: 10–22 pmol/L). While initially treated with levothyroxine, she had subsequently been on liothyronine 25 μg/day for two years. Approximately one month prior to presentation the daily dose increased to 30 μg/day (15 μg twice daily) in a manually compounded extended release form as prescribed by a naturopath with whom she also consulted.

She presented to the emergency department (ED) of St. Paul’s Hospital with a two-day history of squeezing bilateral frontal and occipital headache associated with abdominal pain, nausea, and vomiting. She was presumptively diagnosed with viral gastroenteritis, treated with ketorolac, morphine and dimenhydrinate and discharged. Her symptoms quickly recurred, prompting a return to the ED the following day. She was again treated for pain and nausea, given IV fluids, and advised to return should symptoms persist.

She returned the next day complaining of diplopia. An oral dose of hydromorphone relieved the headache and because a computed tomography (CT) scan of the head showed only mild cerebral atrophy unrelated to the headache, she was discharged. At home, her condition worsened as she developed confusion and disorientation over the ensuing two days. She returned to the ED for a fourth time when she was assessed for acute onset of confusion and decreased level of consciousness (Glasgow Coma Score 11–14). She was noted to have a mild fever (37.7 °C) and tachycardia (110–130 bpm) and was admitted. A lumbar puncture was performed to rule out encephalitis and additional laboratory testing including thyroid function were arranged.

The CSF was clear and showing normal protein and glucose (Table 1), grew no organisms and was negative for HSV and enterovirus. A repeat CT scan was non-contributory. The thyroid panel showed TSH <0.02 mU/L, free thyroxine (fT4) of 10.9 pmol/L (N: 10–22), and free triiodothyronine (fT3) >30 pmol/L (N: 3.5–6.5), above the analytical measurement range of the immunoassay (Siemens Centaur, Tarrytown, NY).

Table 1.

Admission laboratory results.

Electrolytes
Na+	145 mmol/L	135–148 mmol/L
K+	3.8 mmol/L	3.6–4.7 mmol/L
Cl−	111 mmol/L	100–112 mmol/L
${\text{HCO}}_3^{-}$	29 mmol/L	22–31 mmol/L
Complete Blood Count
Hemoglobin	102 g/L	120–155 g/L
RBC	3.82 tera/L	3.80–5.20 tera/L
Hematocrit	30%	41–50%
MCV	78 μm3	82–98 μm3
RDW	13.1%	11.0–15.0%
Platelets	149 giga/L	150–400 giga/L
Cerebral Spinal Fluid
Glucose	2.9 mmol/L	2.4–4.4 mmol/L
Protein	0.23 g/L	0.15–0.45 g/L
WBC	5.0 mega/L	0–5 mega/L
RBC	6.0 mega/L	0–10 mega/L

To estimate fT3, the patient’s serum was diluted with serum from a euthyroid patient at 10× and 20× dilutions, recognizing that the estimate would be crude due to differences in binding globulin concentrations. The approximate fT3 concentration was 330 pmol/L. Tachycardia and agitation were managed with propranolol and midazolam and the patient was transferred ICU. As with previous reports [5], [6], [7] she was treated with plasmapheresis and underwent seven rounds of albumin plasma exchange over a period of eight days in which time the fT3 decreased to 21.7 pmol/L and her mental status improved. She was discharged on day 8 after admission. During a follow-up visit 2 weeks post-discharge the patient had discontinued her thyroid medications completely and her fT3 was in the physiologically normal range (Table 2). In the follow up period after her discharge, anti tissue peroxidase antibodies were found to be normal at 10 IU/mL (N <35).

Table 2.

Thyroid function of the patient prior, during, and after the course of her admission. The reference intervals for fT3, fT4 and TSH are 3.5–6.5 pmol/L, 10–22 pmol/L and 0.27–4.2 mU/L, respectively. D/C = discharge. ^∗Estimated by spiking into normal patient serum.

Timing	fT3 (pmol/L)	fT4 (pmol/L)	TSH (mU/L)
3 months prior	3.8	11	3.00
Day 1	$\simeq {330}^{\ast }$	11	<0.02
Day 2	$\simeq {106}^{\ast }$	9	<0.02
Day 3	$\simeq {43.0}^{\ast }$	–	–
Day 4	29.2	–	–
Day 5	25.5	–	–
Day 6	21.7	–	–
Day 7	21.7	–	–
Two weeks post D/C	3.2	10	2.56

2.1. Mass spectrometric analysis

The compounded liothyronine capsules were obtained from the patient’s family. Conversations with the family indicated that she had manifested no intent to self-harm and a pill count showed that she had not taken more than she had been instructed.

A simple fit-for-purpose T3 assay was developed as follows. A single capsule was emptied into a 20 mL glass scintillation vial. The contents were initially diluted in 3 mL of 1:1 methanol:H₂O and further diluted to a total volume of 15 mL with water. Because the solubility of triiodothyronine (T3) is poor in neutral conditions, an additional 1 mL of NH₄OH (30%) was added for a final total volume of 16 mL. The sample was vortex mixed, sonicated and filtered with glass fiber filter and again diluted 1:10 with methanol.

Both T3 and ${}_{}{}^{13}\text{C}_6^{}$-T3 internal standard (IS) were obtained from Cerilliant (Round Rock, TX) and stock solutions of each were prepared at concentrations of 1 μg/mL, 10 μg/mL and 100 μg/mL by serial dilution in 0.1 M NH₄OH in methanol. Multiple reaction monitoring (MRM) transitions for T3 of $651.8\to 605.8,651.8\to 507.9,651.8\to 478.9$ m/z and corresponding transitions for the IS were selected based on infusion studies on a SCIEX 5500 QTrap triple quadrupole liquid chromatography and tandem mass spectrometry (LC-MS/MS) system (Concord, ON Canada). Chromatography was performed using a simple 3.2 min chromatographic gradient (Fig. 1) of water (solvent A) and methanol (solvent B) both containing 0.1% formic acid and 2 mmol/L ammonium acetate at 0.650 mL/min on a 4 × 3.0 mm C18 (Phenomenex SecurityGuard^TM) guard cartridge. Internal standard was added standards and capsule extracts to a concentration of 1 μg/mL. The LC injection volume was 10 μL.

Fig. 1.

HPLC gradient. Solvent A: water. Solvent B: methanol.

Based on the prescription details, the expected concentration of T3 in the filtered solution was 0.09 μg/mL meaning that the peak-area of an injection of the unknown should have been about one tenth the 1 μg/mL standard. However, initial injections of the unknown showed peaks much higher than this standard, saturating the ion multiplier of the LC-MS/MS instrument (see Fig. 2).

Fig. 2.

Upper pane: Ion chromatograms of initial extraction of compounded capsule showing saturation of the ion multiplier in all three MRM transitions. Green trace: $651.8\to 605.8$ m/z; Red trace: $651.8\to 507.9$ m/z; Blue trace: $651.8\to 478.9$ m/z. Lower pane: Ion chromatograms of ${}_{}{}^{13}\text{C}_6^{}$-T3 internal standard. Green trace: $657.8\to 611.8$ m/z; Red trace: $657.8\to 513.8$ m/z; Blue trace: $657.8\to 474.9$ m/z.

For this reason, the solution obtained from the capsule was diluted a further 100-fold and 1000-fold and compared to a 3-point calibration curve prepared from 100× dilutions of the stock standards (0.01, 0.1 and 1 μg/mL). The estimated T3 concentrations of solutions prepared from the capsule were an average of 1070-fold the expected value. This result was confirmed by duplicate analysis of a second capsule. The set-up of the T3 method and the accompanying analysis took approximately one work day for an experienced mass spectrometrist. The data were reported to the pharmacy that had performed the compounding which voluntarily reported the error to the British Columbia College of Pharmacists.

3. Discussion

3.1. Clinical presentation

The patient’s liothyronine prescription was compounded at approximately 1000-fold the prescribed dose. The first presentation to the ED was 13 days after the prescription fill-date and differed from classic picture of thyroid storm. In particular, she was noted to be afebrile on first visit and there was no mention of tachycardia, hypertension, sweating, palpitations or agitation, the dominant complaint being headache, nausea and vomiting explaining the initial presumption of gastroenteritis.

By the time she developed agitation and confusion returning to the ED for a fourth time, 17 days after the presumed initiation of the prescription, her blood pressure was noted to be normal 121/60 mmHg, she remained afebrile (36.8 °C), was not tachypneic, though was tachycardic at 121 min⁻¹. The clinical features typical of thyroid storm were confusion, anxiety, abdominal pain, nausea and vomiting.

As far as we are aware, this is the fourth report of thyrotoxicosis due to incorrectly compounded thyroid mediations and the third report of liothyronine in particular [5], [6], [8]. This case and previous reports of liothyronine overdose suggest that fever may be unexpectedly absent [7], [8] and that gastrointestinal and neurological symptoms dominate [5], [8].

3.2. Treatment

Successful use of plasmapheresis for the management of thyroid hormone overdose is previously reported with good outcomes. Jha et al. [6] describe a case of thyroid storm after an ingestion of 720 mg of incorrectly compounded T3 and T4 extract presenting with a TSH = 0.07 mU/L, fT3 > 46 pmol/L and fT4 = 78 pmol/L. Plasmapheresis performed on two consecutive days resolved the patient’s symptoms and improved fT3 to 9.5 pmol/L and fT4 to 64 pmol/L, but TSH remained low at 0.02 mU/L.

Shah et al. [5] described thyroid storm due to ingestion of compounded liothyronine presenting with a fT3 > 43.0 pmol/L, fT4 = 5.1 pmol/L and TSH <0.01 mU/L. The patient initially received medical therapy (propranolol, cholestyramine, and glucocorticoids), which proved ineffective. One round of plasmapheresis resulted in a sharp decrease in fT3 to 13.8 pmol/L and the patient’s mental status returned to near-normal five days later.

The case described by De La Calzada-Jeanlouie et al. [8] was treated medically without the use of plasmapheresis.

The appropriate duration plasmapheresis is unclear but stabilization of mental status near normalization of the free thyroid hormones is advised [6]. Liothyronine overdoses may resolve faster by virtue of the shorter half-life of liothyronine [7], [9].

3.3. Lab considerations

Total T3 testing is unavailable in the province of British Columbia necessitating dilution of the patient specimen into normal serum to estimate the fT3 concentration, which was about 60× the upper limit of normal. Because of the absence of binding proteins, use of onboard commercial instrument diluent is not possible for free hormone analyses. For this reason, dilution studies facilitated the biochemical monitoring of the patient throughout her admission. After the purchase of necessary reagents, we were able to determine the approximate concentration of the compounded T3 using an LC-MS/MS developed in an ad hoc fashion for this particular case.

It is important to emphasize that the LC-MS/MS method used in this case was crude in nature, being unvalidated, but adequate to estimate the capsule T3 content and to establish the cause of overdose. Generally, accurate quantitation of drug concentration in tablets or capsules requires similar methodological rigor to clinical assays [10]. Other similar case reports are silent on the laboratory methodology used to confirm compounding error [3], [4], [8], or state that it was inferred from pharmacy records [2], [6].

3.4. Regulatory issues

Drug compounding is sometimes necessary as it affords dispensing of prescriptions in dosages or preparations not commercially available. Unlike commercial pharmaceutical manufacturers, drug compounders in the United States are not required to register with the US Food and Drug Administration (FDA) meaning that ‘Good Manufacturing Practices’ (GMP)—conditions ensuring sterility, purity and correct potency of drugs—may not be met [1].

In 1992 the FDA clarified the role of compounding pharmacies in Section 460.200 of the Compliance Policy Guides. Subsequently, in 1997 congress enacted Section 503A of the Food Drug and Cosmetic Act (FDCA) of 1938 as part of the Food and Drug Administration Modernization Act banning compounding pharmacies from promoting or advertising their products. This, however, was deemed unconstitutional by compounding pharmacies due to violation of the First Amendment and was overturned in 2002 [11].

In 2012 an outbreak of fungal meningitis in Massachusetts caused by Exserohilum rostratum was traced to contamination of a batch of compounded methylprednisone prepared for epidural use, leaving some 64 dead and 750 infected. The tragedy prompted congress to amend The Drug Quality and Security Act granting the FDA more authority to regulate and monitor compounding pharmacies. Since then, the FDA has conducted more than 425 inspections of compounding pharmacies and recalled over 140 drugs [11].

The Canadian government has raised similar compounding pharmacies making efforts to define and limit their role since 1997. Here again, the federal regulator, Health Canada, makes a distinction between drug manufacturing, which it regulates under the Food and Drugs Act, and drug compounding which it delegates to the respective provincial authorities under the “Policy on Manufacturing and Compounding Drug Products in Canada” (2009) [12].

3.5. Perspective

It should be emphasized that this patient had only weak biochemical evidence of hypothyroidism. Prior to her presentation and after her discharge, we found no TSH above the upper limit of normal from any reporting laboratory in British Columbia. While some may argue for a diagnosis of subclinical hypothyroidism based on debate as to the appropriate upper limit of normal for TSH [13], the case demonstrates that the diagnosis of hypothyroidism, the prescription of thyroid hormone supplementation, particularly in manually compounded form, should not be considered entirely innocuous.

4. Conclusions

We present a case of thyrotoxicosis caused by incorrectly compounded liothyronine, containing 1000-times the stated T3 content as elucidated by LC-MS/MS. Regulation of compounding pharmacies remains an active interest of both the FDA and Health Canada. In the event of a life-threatening overdose, plasmapheresis is an effective therapy.

Conflict of interest statement

The authors declared that there is no conflict of interest.