Role of Plasma Exchange in the Thyroid Storm
Abstract
Inadequately treated thyroid storm can lead to death. Therapeutic plasma exchange (TPE) is a suggested treatment when conventional treatments fail, but its indication is not well codified. We report our experience through three explicit cases. Three elderly patients were admitted to our hospital for cardiac or neurologic symptoms due to thyroid storm. After initiation of conventional therapy, TPE was performed with clinical and biological improvement. The speed of symptom resolution varies depending on the severity. This technique must be carried out by experienced medical staff as many complications can occur; nevertheless, in our patients with severe comorbidities, no complications occurred. The action of TPE mainly results from plasma removal of cytokines, putative antibodies, and thyroid hormones and their bound proteins. TPE has a transitory effect and thus should be associated with other thyroid blockers. When there are threatening symptoms, TPE should be done early, without waiting for the efficiency of conventional treatment, since it is the fastest method known for the improvement of the clinical condition. We also suggest starting TPE in case of neurologic symptoms because of very slow and incomplete regression. The Burch and Wartofsky score seems to be a helpful tool in establishing the diagnosis of thyroid storm and for deciding on when to initiate TPE.
Therapeutic plasma exchange (TPE) is a potential therapeutic option for thyroid storm, but its indication is not clearly defined. The literature contains fewer than 100 reported cases and no randomized studies, and clear guidelines about indication criteria, the best time to start and duration of this procedure are lacking.
The thyroid storm is an abnormal state, characterized biologically by extremely elevated serum levels of circulating hormones leading to multisystem involvement with specific appearances in every organ. Its contributing factor is often a physical stress, such as infection or surgery. Even with treatment, thyroid storm-related mortality stays high at approximately 30% (1), mainly due to cardiac arrhythmias, intractable heart failure, and progressive cardiomyopathy. In 1993, Burch and Wartofsky (2) created a scoring system to standardize the diagnosis of thyroid storm using the following parameters: body temperature, central nervous system involvement, gastrointestinal–hepatic dysfunction, heart rate, and the presence or absence of congestive heart failure and/or atrial fibrillation. The severity of each symptom correlates with a number of points, with a maximum score of 140. A score of 25–44 is considered high risk for an imminent storm, and a score >45 is defined as being diagnostic of thyroid storm.
The three most common treatments for hyperthyroidism are antithyroid drugs, radioactive iodine, and thyroidectomy. TPE is an alternative treatment that has been proposed since the 1970s for hyperthyroidism and, more precisely, for the thyroid storm, whatever the etiology (3). TPE has been shown to have a clear benefit in the thyroid storm induced by Graves' disease, the most common cause of hyperthyroidism and found especially in the life-threatening forms. Its efficiency here can be explained by the rapid decline of plasma hormones and antibodies responsible for Graves' disease (4). Its use was also reported in cases of amiodarone-induced hyperthyroidism, inefficiency or toxicity of conventional treatments (such as leukopenia due to propylthiouracil), Hashimoto's encephalopathy (5), hyperthyroidism caused by molar pregnancy (6) and Graves' ophthalmopathy (7–13). Thyroid storm can be also caused by surgery or through the use of radioiodine therapy in unprepared hyperthyroid patients, and TPE is efficient in preparing for emergency surgery in severe cases (14). This technique is not indicated when hyperthyroidism is well-tolerated.
We report herein three patients hospitalized at Colmar Hospital who developed a thyroid storm and were treated with TPE. We used the filtration mode with a dialyzer and a central venovenous access. The replacement solution was mainly albumin. After describing our experience in the resolution of these cases and reviewing the literature (Table 1), we propose some recommendations.
| Authors | Number of cases | Median number of TPE/median volume replacement per patient | Indication | Response |
|---|---|---|---|---|
| Ashkar et al. 1970 (3) | 2 | 6/3 L | Cardiac symptoms | Clinical improvement in 48–72 h and diminution of TT3 by 22–80% |
| Herrmann 1973 (15) | 2 | 5/2.65 L | Cardiac and neurological symptoms | Clinical and biological amelioration, patients successfully treated by the last plasma exchange; 40% decrease in TT3 and 56.6% decrease in TT4 |
| Horn 1976 (16) | 1 | 1/5 L | Neurological symptoms | Clinical improvement and biological improvement with significant fall in the free hormones (633 µg FT4 and 13.6 µg T3) |
| Gabry 1981 (17) | 1 | 2/2.5 L | Before surgery | Clinical improvement in 48 h |
| Martin 1981 (18) | 1 | 1/4 L | Cardiac, neurological and ophthalmological symptoms after surgery | Clinical improvement in less than 24 h; 25% decrease in TT4 |
| Ogriseg 1981 (19) | 1 | 3/8.1 L | Cardiac and neurological symptoms | Clinical and biological improvement with a decrease of 18.5 µg and 10 µg of TT3 and TT4, respectively |
| Van de Vyver 1982 (20) | 1 | 3/8.4 L | Cardiac and neurological symptoms | Clinical and biological improvement with the extraction of 530 µg of T4 and 2.8 µg of T3 |
| Shigematsu 1982 (21) | 1 | 1/3.8 L | Resistance to conventional antithyroid drugs, agranulocytosis | Clinical improvement after TPE, but no change observed in TT3 or TT4 |
| May 1983 (22) | 1 | 1/4.8 L | Thyroid hormone intoxication | Clinical improvement immediately after TPE and biological improvement (43% decrease in TT4 and 68% decrease in TT3) |
| Patte et al. 1983 (23) | 1 | 2/3.4 L | Cardiac and neurological symptoms with conventional drug resistance | No amelioration, plasma exchange not well tolerated, with underlying cardiac disease threatening |
| Newcomer 1983 (24) | 1 | 1/4 L | Cardiac and neurological symptoms | Biological improvement (decrease of 64% TT3 and 54% TT4) and clinical improvement in 24 h |
| Tajiri 1984 (25) | 1 | 2/7.2 L | Cardiac and ophthalmological symptoms | Clinical improvement in 48 h, biological improvement (decreases of 8.5% FT4, 27.4% FT3, and 15% TT3, but no decrease in TT4) |
| Sprenger 1985 (26) | 2 | 3.5/2.6–3.8 L per exchange | Thyrotoxicosis | 39% decrease in TT4 in the first case; 53% decrease in TT4 immediately after TPE in the second case, with 80% after a few days |
| Schlienger et al. 1985 (27) | 8 | 2.75/40 mL/kg | Thyrotoxicosis | 7/8 showed clinical improvement, biological improvement with almost no improvement in FT3 or FT4, but amelioration with decreases of 43% of TT3 and 37% of TT4 |
| Braithwaite 1986 (28) | 1 | 3/5.5 L | Thyrotoxicosis after thyroid iodine irradiation and conventional drugs | Clinical improvement with cure in four days and biological improvement (decreases of 76% of T3, 56% of T4, and 39% of FT4) |
| Geissler 1987 (29) | 1 | 2/6 L | Hematological symptoms and before surgery | Clinical and biological amelioration; decreases of 25% of TT3 and 17% of TT4 |
| Binimelis et al. 1987 (30) | 4 | 3.25/3.8–4.5 L | Thyrotoxicosis with neurological and cardiac symptoms | Biological and clinical improvement; thyroid hormone normalization in approximately15 days; >50% decreases in TT3 and TT4 |
| Segers 1988 (31) | 5 | 4.6/2.5 L each time | Thyrotoxicosis | Clinical and biological improvement: approximately 63.5% decrease in FT3 and 57.8% in FT4; one relapse after one week |
| De Rosa et al. 1991 (4) | 1 | 3/6.3 L | Thyrotoxicosis with ophthalmopathy, and before surgery | Biological improvement with decreases of 51% FT3, 47% FT4, 60% TT3, and 53% of TT4 |
| Uzzan 1991 (32) | 1 | 5/22.5 L | Thyrotoxicosis with cardiac and muscular symptoms | Biological improvement: decreases of 75% TT4, 75% FT4, 78% FT3 (nothing for TT3); amiodarone rate decreased too, but then it rose again after TPE stopped |
| Puy et al. 1992 (33) | 2 | 3/3.5 L | Severe hyperthyroidism and cardiac failure | Biological amelioration with decreases of 40% TT3, 31% FT3, 27.2% TT4, and 14.4% FT4 |
| Schlienger 1992 (34) | 11 | 2.9/ND | Thyroid storm or persistent thyrotoxicosis | Controls were six euthyroid patients: decreases in TT4 and TT3 blood levels (39% and 47%, respectively), more marked for TT3, related to the removal of TBG (thyroxine-binding globulin); FT3 and FT4 remained unchanged (6% decrease in FT4) |
| Aghini-Lombardi et al. 1993 (35) | 2 | 2/4 L | Thyrotoxicosis | Transient clinical improvement, decreases of approximately 70–75% FT3, and 30% FT4, no range normalization; and decreases of 60% TT3 and 40–45% TT4 |
| Henderson et al. 1994 (36) | 1 | 2/4 L | Thyroxine overdose | No clinical or biological improvement; decreases of 84% FT4 and 58.5% FT3 immediately after TPE, but with subsequent rises afterwards |
| Samaras and Marel 1996 (37) | 1 | 5/ND | Thyrotoxicosis with cardiac symptoms | Failure of treatment due to the death of the patient; at first biological amelioration of TT3 and TT4 was experienced, but afterwards the rate rose again to levels higher than previously |
| Ligtenberg et al. 1999 (38) | 2 | ND/ND | Before surgery and tachycardia | Little effect on plasma level (0% and 33% decreases in FT4, and 7% and 18% decreases in FT3) and no effect on the hyperthyroid state |
| Boers and Colebatch 2001 (5) | 1 | 3/21 L | Resistance to conventional antithyroid drugs with neurological manifestations | Clinical and biological improvement, with a subsequent relapse; no FT3, FT4, TT3, or TT4 samples taken |
| Nieuwenhuis 2004 (39) | 1 | 3/ND | Severe neurological symptoms | Clinical and biological improvement (ND); six months' recuperation; reduction in the level of antibodies to normal |
| Diamond 2004 (40) | 3 | 2.3/7.5 L | Before surgery, TIA | Clinical stabilization before surgery; one patient had no biological amelioration, the other two had average decreases of 40% FT3 and 35% FT4 (TT3 and TT4 ND) |
| Kokuho et al. 2004 (41) | 1 | 1/3.4 L | Resistance to conventional antithyroid drugs | Clinical amelioration in 3 h, with decreases of 25% FT4 and 15% FT3 |
| Ozbey 2004 (42) | 4 | 3/1–1.5 L per exchange | Before surgery | Biological amelioration: mean decreases of 65% TT3 and 75% FT4 |
| Guvenc 2004 (43) | 1 | 3/7.5 L | Resistance to conventional antithyroid drugs and heart failure | Clinical and biological amelioration; decreases of 58% FT3 and 67% FT4 |
| Erbil 2006 (44) | 1 | 3/ND | Preparation for surgery | Biological improvement, with decreases of 43% FT3 and 26% FT4 |
| Azezli 2006 (45) | 1 | 3/6 L | Preparation for surgery | Biological improvement, with decreases of 75% FT3, 64% FT4, 66% TT3, and 51% TT4 |
| Pasimeni 2008 (46) | 1 | 2/6 L | Resistance to conventional antithyroid drugs | Clinical and biological amelioration: decrease of 29% FT4, but no amelioration of FT3 (TT3 and TT4 ND) |
| Adali et al. 2009 (6) | 1 | 2/ND | Resistance to conventional antithyroid drugs, major bleeding | Biological amelioration: decreases of 33% FT3 and 17.5% FT4 |
| Ezer et al. 2009 (14) | 11 | 3.3/ | Before surgery | Improvement in thyroid symptoms; safe preparation of patients before surgery, with mean decreases of 76% FT3 and 44% FT4 (TT3 and TT4 ND) |
| Muller 2011 | 3 | 4.6/2.5 L per exchange | Resistance to conventional drugs, cardiac and neurological symptoms | Biological and clinical amelioration: mean decreases of 55% FT4 and 26% FT3 immediately after TPE sessions |
 |
 |
 |
 |
 |
- TIA, transient ischemic attack; FT3, free triiodothyronine; FT4, free thyroxine; ND, not determined; TPE, therapeutic plasma exchange; TT3, total triiodothyronine; TT4, total thyroxine.
PATIENTS AND METHODS
Patient 1
A 61-year-old man was hospitalized in Colmar Hospital for thoracic pain and dyspnea on June 2009. He had a past history of ischemic cardiomyopathy and had undergone double by-pass graft surgery and multiple stenting. An episode of recurrent ventricular tachycardia two years before was treated by amiodarone, with a normal TSH level at this time. A few weeks before hospitalization, a flutter appeared while on bisoprolol and amiodarone. Hyperthyroidism was detected and treated with carbimazole 20 mg three times a day and potassium perchlorate 600 mg per day.
On admission, the arterial pressure was 150/78 mm Hg, the heart rate was 87/min with arrhythmia, and he was afebrile. His cardiac symptoms were related to hyperthyroidism. The serum concentrations for free triiodothyronine (FT3) were 10.78 pmol/L (laboratory reference range (RR), 3.1–6.2) and for free thyroxine (FT4) 99.8 pmol/L (RR, 9–21.9) (Immunochemiluminescence; Beckmann, Palo Alto, CA, USA). He had no renal insufficiency or electrolyte abnormalities, and echocardiography showed a ventricular ejection fraction of 66% (using Teichholz's formula). His score according to Burch and Wartofsky (2) was 20 to 140, so he was not considered to be at risk of a thyroid storm.
Since admission, antithyroid drugs were optimized with carbimazole 20 mg six times a day, potassium perchlorate 1000 mg per day, corticosteroid 30 mg per day, and propranolol was given instead of bisoprolol. TPE was started 16 days after admission as the conventional treatment was showing no efficiency in controlling his life-threatening and crippling cardiac symptoms (that is, the severe arrhythmia). The score was 30–140 at this time, as a moderate hepatic dysfunction had appeared. Four successive TPE sessions were performed in eight days (Fig. 1). The FT4 level decreased sharply after each session, with a total reduction of 65% at the end of the treatment, whereas the TSH increased progressively. No complications occurred. Two weeks later, his clinical state was stable without any cardiac arrhythmias, and a thyroidectomy was performed under good conditions and without complications. The entire thyroid was removed and found to be slightly increased in size, hard, and fibrous.
Free triiodothyronine (FT3) and free thyroxine (FT4) serum concentrations during hospitalization in the intensive care unit for each patient. B, before a session of therapeutic plasma exchange (TPE); and A, after TPE. RR: FT3: 3.1–6.2 pmol/L; FT4: 9–21.9 pmol/L.
This patient presented with type 2 amiodarone-induced hyperthyroidism without an admission score considered at risk. Nevertheless, he developed a thyroid storm with cardiac symptoms in the foreground. Because of failure of conventional therapy, TPE was performed and succeeded in significantly reducing the thyroid hormone serum concentrations and in controlling the cardiac symptoms. Thanks to this technique, a stable state was quickly obtained, which enabled the radical surgical treatment to be performed safely.
Patient 2
A 74-year-old man with a past history of hypertension, atrial fibrillation treated with amiodarone, and two ischemic strokes was admitted on February 2001 for a confusional state with anorexia and insomnia. A cerebral CT scan at admission showed only the sequelae of his previous strokes. Several days later he developed fever and loss of consciousness. Despite antibiotic therapy, his temperature increased to 39.5°C and he also developed dyspnea with a respiratory rate of 44 breaths/min, tachyarrhythmia at 150 beats/min, clinical signs of dehydration, and blood pressure of 105/75 mm Hg. The neurological examination showed amyotrophia, absent reflexes, hypotonia, and minimal response to pain stimuli.
Biological investigations revealed the following: TSH < 0.01 mIU/L (RR 0.34–3.80), FT4 > 103 pmol/L, FT3 > 38.5 pmol/L, hypernatremia (Na+ 158 meq/L), hyperkalemia (K+ 5 mmol/L), acute renal insufficiency (serum creatinine 1.6 mg/dL, urea 84.5 mg/dL), cytolysis at twice the normal level, and respiratory acidosis (pH 7.31, pCO2 55 mm Hg). The Burch and Wartofsky score was high at 100–140. Since thyroid antibodies were negative and the ultrasound was normal, type 2 amiodarone-induced hyperthyroidism was suspected.
Mechanical ventilation was required due to his respiratory distress and neurological problems. Carbimazole was introduced after admission (as methimazole is not available in France) and TPE was carried out two days later for a total of six exchanges—every day for three days and then every 48 h. The level of FT4, which was initially very high, decreased under the superior threshold only after six TPE sessions. A rapid FT3 reduction was also seen (Fig. 1). Steroids and potassium perchlorate were introduced, as well as propranolol, after the exchange procedure to control mild tachycardia. The patient left the intensive care unit (ICU) on 5 mg prednisone per day and propranolol; the potassium perchlorate and carbimazole were discontinued.
His clinical picture slowly improved; the patient recovered consciousness within three weeks after TPE and his muscle tone returned one month later. Nevertheless, his respiratory status required a tracheostomy one month after admission. No radical treatment was planned after this event. The normalization of the thyroid hormone serum concentrations was obtained after more than one month and thyroid treatments were stopped. No more complications occurred during this period.
In this second case of amiodarone-induced thyroid storm with a very high risk score, TPE was initiated as first-line therapy in conjunction with conventional treatment. It offered benefits with a rapid resolution of cardiac symptoms and a slow (over one month) but effective recovery of his initially severe neurologic state.
Patient 3
An 85-year-old woman was admitted for hyperthyroidism on September 2002. She had a history of giant cell arteritis, polymyalgia rheumatica, hypertension, and valvular and ischemic cardiomyopathy with conduction and rhythm impairment. A multinodular goiter with hyperthyroidism had been treated with radioactive iodine in 1992 with stable hormonal levels thereafter.
One month before admission, an aortic valve replacement and a mitral valvuloplasty were performed because of recurrent pulmonary edema. In the postoperative days, the patient became asthenic and blood tests revealed a high FT4 at 42.3 pmol/L and a low TSH concentration at 0.005 mU/L. Despite the introduction of carbimazole, she presented with tachycardia (120 beats/min) and severe amyotrophia. Propranolol was introduced; however, her clinical status worsened with the onset of confusion, requiring admission to the ICU nine days later.
In the ICU, she was found to be thin with mild exophthalmia. She presented with atrial fibrillation, tachypnea (50/min), with rales in the right lung. Her blood pressure was 130/80 mm Hg. Neurologic examination found an isolated confusion. Biologically, there was no acidosis or electrolyte imbalance. FT4 was increased to 80 pmol/L and FT3 to 14.3 pmol/L; TSH was undetectable; thyroid antibodies were negative. The Burch and Wartofsky score was high at 75–140. Because of history of multinodular goiter and discrete exophthalmia, a recurrence of the goiter or Graves' disease was suspected.
As her respiratory state worsened, with severe hypoxemia and persistent tachyarrhythmia, daily TPE was started with four exchanges. Ten days after admission, as the levels of thyroid hormone normalized and the clinical outcome was favorable, treatment with carbimazole was switched for propylthiouracil with a tapering dose (Fig. 1). No radical treatment was performed.
In this case, TPE was justified by the high score, the failure of one months' conventional therapy, and the appearance of precarious respiratory and cardiac states. In this elderly patient with severe cardiomyopathy, TPE allowed a favorable outcome without complications.
DISCUSSION
The therapeutic benefit of TPE results from the removal of the potential pathological substances from the thyroid storm: hormones, autoantibodies (i.e. Graves' disease), catecholamines (released by the sympathetic system), cytokines (which mediate inflammation), toxins, and so on. The metabolically active part of the hormones is free, but over 99.5% of T4 (Thyroxine) and T3 (Triiodothyronine) are protein bound in the serum. TBG binds approximately 80% of circulating hormones; transthyretin and albumin are minor protein binders (49); albumin binds 10% of T4 and 20% of T3 (33). This technique decreases the effects of thyroid hormones by removing free thyroid hormones and, moreover, by replacing carrier proteins by unsaturated bound proteins of the replacement solution. TPE also removes 5′-monodesiodase which converts T4 to T3 and this removal lessens T3 production (50). Nevertheless TPE has to be associated with an adjunct treatment to stop the synthesis of these components.
These three cases illustrate the efficiency and rapidity of action of plasma exchange in the thyroid storm. TPE has allowed us to observe favorable biological and clinical effects. After analyzing these examples and reviewing the literature (Table 1) we are able to suggest several recommendations.
In the first and third cases, we noticed a dramatic reduction of the free T3 and T4 serum concentrations after the TPE procedures (with a total reduction of 65% of FT4 in the first case, and of 46% for FT4 and 27% for FT3 in the third case). In one study, Puy et al. report a total extraction of FT4 and FT3 from between 20 to 60% and between 20 to 40%, respectively, for TBG after three exchanges (33). Binimelis et al., in a small study on six patients, depict a disappearance rate of total thyroxine (TT4) that is 30-fold greater with TPE than with standard medical treatment. This effect was proportional to the serum level of TT4 (30). In the same way, we saw in the second patient that FT4 could not be quantified, hiding the fact that FT4 was extremely high, thus explaining why reaching normal levels and TPE efficacy need several weeks.
In cases 1 and 2, with amiodarone-induced hyperthyroidism, the efficiency of TPE can also be explained by the fact that it removes amiodarone from the circulation. Amiodarone and its active metabolite are highly bound to plasma proteins and cannot be cleared by hemodialysis, but only by plasma exchange.
In our observations, thyroid hormone levels have a tendency to rise again in the next day after TPE. In some cases the FT3, FT4, and TSH do not decrease during the procedure and may even increase (30,33). This phenomenon is explained by a mobilization effect from the extravascular compartment and the establishment of a new equilibrium between free and intracellular rates (27). Nevertheless, even if the T3 and T4 concentrations still stay high, clinical improvement can be seen (1).
In our first and third cases we simultaneously observed the biological and clinical amelioration of cardiac and neurological symptoms at the end of the TPE procedures. In all three cases reported herein there were cardiac manifestations with problems in reducing tachycardia and managing the arrhythmia. TPE decreased both the thyroid rate and controlled the cardiac symptoms within a few days, when conventional therapy had failed. We observed that cardiac symptoms are more likely to regress faster than the neurological ones.
In our patient who was in a coma (patient 2), we observed slower recuperation requiring three weeks for the coma to resolve. We hypothesize that thyroid hormone removal is greater in the circulation, but that the blood–brain barrier could prevent intracerebral hormone removal. We speculate that in such cases with severe neurological impairment, improvement requires a longer time after a TPE session. The mechanisms underlying reversible neurological symptoms are as yet unknown, and the recuperation seems to be variable between patients. As was the case with our patients, some authors describe very late improvement, up to six months after the procedure (23).
An unexpected action of TPE with clinical improvement a few hours after the first TPE session is frequently reported (27,35). TPE improves the clinical symptoms (especially cardiac ones) of the thyroid storm rather than decreasing the thyroid hormone rates (3). Kokuho et al. reported their experience with a woman with a thyroid storm that was resistant to methimazole, Lugol's solution, propranolol, and digoxin: as she had a body temperature of 39°C and tachycardia, TPE was performed and immediately, after 3 h, the cardiac rhythm slowed and the fever decreased (41).
Some reports have described the failure of TPE to control thyroid storm in different cases such as Graves' disease, amiodarone-induced thyroid storm, or hormone overload (10,23,36–38). TPE appears to be inefficient in hormone overload in Henderson et al.'s study as it had removed only 7% of the ingested dose (36). In Samara et al.'s report, renal and hepatic impairment would have played a role in the failure of TPE (37), while in Kuzuya et al.'s case, TPE was partially efficient, but exophthalmos was certainly too much evolved (10), and in Patte et al.'s study, TPE was stopped because of cardiac instability during the exchange (23). In several cases, TPE has failed because of the great severity of symptoms already present at diagnosis. TPE should be conducted as early as possible in order to be efficient.
The usual association with other drugs, such as glucocorticoids and/or cytotoxic medications, makes it difficult to differentiate the contribution of TPE in these clinical improvements. In the second patient, we could speculate as to the particular extent of each of the separate actions of TPE and carbimazole; however, they were probably both effective and their effects synergistic.
TPE is a recommendation of grade IIc and a category III in the last American Society for Apheresis (ASFA) document (51). TPE should be performed in the following indications: severe symptoms (such as cardiothyrotoxicosis, as in the first case, neurological manifestations, impairment of consciousness, as in the second case, and severe myopathy), rapid clinical degradation, contraindications to other therapies (agranulocytosis, renal insufficiency, asthma, heart failure, as so on), and failure of conventional therapeutics. Before emergency surgery, it can also be used to stabilize the patient.
When treating a thyroid storm, we suggest calculating the Burch and Wartofsky score (2), and if it is higher than 45, to envisage early TPE before the emergence of life-threatening cardiac symptoms or neurological symptoms, especially as regression is slow and sometimes incomplete in the latter. If the score is less than 45, conventional therapies can be tried first. As in the first case, TPE was performed after failure of conventional treatment and as the score worsened. In our cases, TPE was beneficial, even if the patients were elderly (patient 3 was 85 years old). The past history of the patient is not a limiting factor and the technique, when well done, seems to be safe no matter what the patient's age and comorbidities are.
As per the ASFA recommendations (51), we recommend that TPE should be performed daily until clinical improvement is noted and with 40–50 mL/kg of replacement solution. FT3 and FT4 should be sampled before and after each session, without stopping TPE if no decrease occurs, because of the biologico–clinical dissociation.
The overall incidence of adverse effects of this technique, which are mostly reversible, is about 5%. Notable side effects include transfusion reaction, citrate-related nausea and vomiting, vasovagal or hypotensive reactions, respiratory distress, and tetany or seizure. Death is rare and usually due to the underlying disease (52). TPE remains a very harsh technique (mainly infectious complications are reported (30)) that can be implemented in case of thyroid storm, and not in simple hyperthyroidism cases.
CONCLUSIONS
We report herein three cases that emphasize the efficiency and safety of TPE treatment in the case of severe hyperthyroidism symptoms with cardiac or neurological complications, even in an elderly patient. Retrospectively, we calculated the Burch and Wartofsky score, which seems to be a helpful tool in making a diagnosis of thyroid storm and for deciding when to initiate TPE. We suggest also starting TPE if neurological symptoms are present because of their very slow and incomplete regression, and, moreover, if manifestations are severe.
It can, of course, be used as a rescue treatment if traditional medical treatments are contraindicated or ineffective. When there are threatening symptoms, the efficiency of conventional treatments should not be awaited, because TPE is the fastest method to improve the patient's clinical condition. It is also particularly useful in amiodarone-induced pathology. TPE has a transitory effect and thus several sessions should be performed, and it should be associated with other thyroid hormone antagonists. It is a relatively safe therapy, although it has a high associated cost. Monitoring the thyroid hormones is interesting, but clinical improvements are often dissociated from the hormonal levels and must be mainly judged on the patient's condition. A randomized study is needed to explore the benefits of initiating treatment earlier, at diagnosis of a thyroid storm, according to the symptoms and the Burch and Wartofsky score.
