Clinical Endocrinology (2010) 72, 551557

doi: 10.1111/j.1365-2265.2009.03682.x

ORIGINAL ARTICLE

Severe hyperthyroidism: aetiology, clinical features and treatment outcome

P. Iglesias*,, O. Devora*, J. Garca*, P. Tajada, C. Garca-Arevalo and J. J. Dez *Departments of Endocrinology, Biochemistry, Hospital General, Segovia and Department of Endocrinology, Hospital Ramon y Cajal, Madrid, Spain

Summary
Background Severe hyperthyroidism (SH) is a serious medical disorder that can compromise life. There have not been systematic studies in which SH has been evaluated in detail. Here, our aims were: (1) to analyse both clinical and analytical features and outcome in patients with SH and (2) to compare these data with those found in more usual forms of hyperthyroidism. Patients and methods All patients diagnosed of SH (free thyroxine, FT4 > 100 pmol/l, NR: 1123) seen in our endocrinology clinic in the last 15 years were studied and compared with a sample of patients with mild (mH; FT4, 2350 pmol/l) and moderate (MH; FT4, 51100 pmol/l) hyperthyroidism. Aetiology, clinical analytical and imaging data at diagnosis, therapeutic response and outcome were registered. Results A total of 107 patients with overt hyperthyroidism (81 females, mean age SD 469 161 years) were evaluated. We studied a historic group with SH (n = 21; 14 females, 409 172 years) and, as a comparator group, we analyszed the data of 86 hyperthyroid patients (67 females, 484 1556 years, NS) comparable in age and gender. The comparator group was classied in MH (n = 37, 26 females, 472 166 years) and mH (n = 49, 41 females, 494 148 years). In comparison with mH group, SH patients were signicantly (P < 005) younger and showed a greater proportion of rst episode of thyroid hyperfunction (P < 005). Graves disease was the main aetiology in the three groups, but patients with SH showed the highest titre of TSH-receptor antibodies (TRAb) (P < 0001). Heart rate and size of goitre were higher in SH group than in mH and MH groups (P < 001). Atrial brillation was more frequently reported in SH group than in MH and mH groups (158% vs. 54% and 0%, respectively, P < 005). Results from logistic regression analysis showed that younger age [OR 0958 (95% CI, 09230995), P = 0026], presence of asthenia [OR 435 (1481278), P = 0008] and higher heart rate [OR 103 (101106), P = 0013] were independent clinical variables associated to SH. SH patients showed

similar biochemical parameters in comparison with mH group, except for increased serum aspartate aminotransferase (AST) (P < 001) and calcium (P < 005) levels, and decreased serum cholesterol (P < 005) and albumin (P < 005) concentrations. Logistic regression analysis showed that only AST [OR 107 (102 111), P = 0005] was an independent biochemical variable associated to SH. No differences in the type of therapy, cure rate and time in achieving cure were found in SH subjects in relation to patients with milder forms of hyperthyroidism. FT4 was the only independent predictor of cure [OR 098 (CI 95%, 097099), P < 005]. Conclusions Graves disease is the most common aetiology in patients with SH. This type of hyperthyroidism is usually de novo and is accompanied by more clinical signs, symptoms, and analytical derangements, as well as higher titres of TRAb at diagnosis than milder forms of hyperthyroidism. The present data are not able to show differences in treatment modality, time to achieve cure, and remission rate among patients with mild, moderate and severe hyperthyroidism. (Received 11 February 2009; returned for revision 4 March 2009; nally revised 5 May 2009; accepted 25 July 2009)

Introduction
Thyroid dysfunction affects a signicant portion of the general population. The prevalence of hyperthyroidism ranges between 05% and 30%.13 When subclinical hyperthyroidism is included the prevalence is more than 5%, and higher than this in women with increasing age.3 In contrast, the prevalence of severe hyperthyroidism (SH), especially its most serious form of presentation, i.e. thyrotoxic storm, is very low, about 12% among patients with overt hyperthyroidism.4,5 The three most common causes of thyroid hyperfunction are Graves disease, multinodular toxic goitre, and autonomous hyperfunctioning thyroid nodule.6,7 Clinical presentation of hyperthyroidism is usually mild or moderate. The most severe forms of hyperthyroidism are very rare and, sometimes, accompanied by high morbidity, even mortality in some cases. SH is a serious medical disorder that can compromise patient life, especially in the elderly and in patients with underlying cardiovascular disease.8

Correspondence: Dr Pedro Iglesias, Department of Endocrinology, Hospital Ramon y Cajal., Ctra. de Colmenar, Km 9,100, 28034 Madrid, Spain. Tel.: +34 913368343; E-mail: piglo65@gmail.com

2010 Blackwell Publishing Ltd

551

552 P. Iglesias et al. To our knowledge, no systematic and detailed evaluation of clinical features of patients with SH has been reported so far. We performed this cross-sectional study in a cohort of hyperthyroid patients with the aim of assessing both clinical features and treatment outcome in SH patients. We have compared these data with those obtained in patients with moderate and mild forms of hyperthyroidism. Parameters evaluated at diagnosis We reviewed the medical records of our hyperthyroid patients and collected clinical and laboratory data at diagnosis of hyperthyroidism. The presence of goitre was classied according to the WHOs 1960 criteria, i.e, a thyroid gland with palpable lobes larger than terminal phalanges of the thumbs of the person examined. Grade of goitre was classied based on clinical evaluation in grades 1A (palpable lobes larger than terminal phalanges of individuals thumbs), 1B (visible with neck extended), 2 (visible with head in normal position) and 3 (visible at a distance).10 All clinical data related to the presence of previous thyroid pathology and used therapies, personal background, complaints at rst visit, symptoms (dyspnoea, heat intolerance, tremour, sweating, palpitations, nervousness, depression, decreased weight, asthenia, increased appetite and eye changes) and signs (presence or absence of goitre, grade of goitre, intrathoracic goitre, thyroid murmur, distal tremor, tachycardia, sweating, lid lag, exophtalmos and stare) of hyperthyroidism, analytical data (biochemical, hormonal and immunological study) and imaging (thyroid ultrasound and scintigraphy) studies were registered in each patient. The aetiology of hyperthyroidism was determined taking into account the clinical, analytical and morphological data at the time of diagnosis of hyperthyroidism. Type of hyperthyroid episode was classied as de novo or as a consequence of a hyperthyroidism relapse. Parameters evaluated during follow-up In all hyperthyroid patients who were followed up for more than 6 months, the following variables were analysed: therapies used from diagnosis to last visit, time of follow-up, cure rate, time to achieve cure, hormonal data in the last visit, number and type of sequelae of the treatment employed, and thyroid functional status in the last visit. Our criteria of cure for patients with hyperthyroidism were: (1) normal serum TSH, FT4 and FT3 concentrations without therapy for at least 1 year of follow-up, (2) post-surgical hypothyroidism and (3) post-radioiodine hypothyroidism. We took into account the following sequelae of therapy: post-surgical hypothyroidism, post-radioiodine hypothyroidism and post-surgical hypoparathyroidism. Finally, thyroid function status at the last visit was classied as normal thyroid function, subclinical or overt hypothyroidism, and subclinical or overt hyperthyroidism with or without therapy (antithyroid drugs or levothyroxine therapy respectively). Hormone assays Serum TSH, FT4 and FT3 were measured by an electrochemiluminescence immunoassay (ECLIA) by using an E170 analyzer (Roche Diagnostics, Mannheim, Germany). Maximal intra- and interassay coefcients of variation were 135% and 116% for TSH, 204% and 389% for FT4 and 287% and 1017% for FT3. The sensitivity of the TSH, FT4 and FT3 assays was 0005 mU/ml, 03 pmol/l and 04 pmol/l respectively. Values of reference were: TSH 0450 mU/ml; FT4 1123 pmol/l and FT3 3968 pmol/l. Thyroid autoimmune status was studied by the measurement of

Patients and methods

Patients A descriptive, observational, cross-sectional and retrospective study in hyperthyroid patients seen in our thyroid clinic from a low-sufcient iodine intake area (centre region of Spain) was carried out. We retrospectively studied all patients diagnosed of SH in the last 15 years. SH group was compared with patients with mild (mH) and moderate (MH) hyperthyroidism (previously diagnosed or de novo), consecutively recruited as they attended our clinic for a period of 6 months from 1st January 2006 to 31st June 2006. Patients with both subclinical hyperthyroidism (normal serum free thyroxine, FT4, and free triiodothyronine, FT3, values with suppressed thyrotropin, TSH, concentrations) and clinical criteria of thyroid storm (score 25, according to diagnostic criteria for thyroid storm)9 were excluded. Criteria for diagnosis and study groups We considered the diagnosis of overt hyperthyroidism when patients showed simultaneously low-serum concentrations of TSH levels (normal range for our laboratory: 045500 mU/ml) and elevated serum FT4 (>23 pmol/l) and/or FT3 (>68 pmol/l) concentrations. Hyperthyroid patients were classied according to their serum FT4 concentrations in mH (FT4, 2450 pmol/l), MH (FT4, 51100 pmol/l) and SH (FT4, >100 pmol/l). The aetiology of hyperthyroidism was determined taking into account the clinical and analytical data at the time of diagnosis. The presence of hyperthyroidism associated to elevated titres (>10 mU/ml) of TSHreceptor antibody (TRAb) and/or diffuse uptake in the thyroid scan with 131I or 99mTc together with at least one of the clinical ndings (signs of thyrotoxicosis, diffuse enlargement of the thyroid gland, and exophthalmos and/or specic ophthalmopathy) were required to diagnose Graves disease. Toxic multinodular goitre was diagnosed in the presence of hyperthyroidism with several nodules on the radioisotope scan or thyroid ultrasound. When the thyroid radioisotope scan showed intense uptake of the isotope at the location of a nodule with a reduction in uptake in the remainder of the thyroid gland, the diagnosis of toxic adenoma or autonomously hyperfunctioning thyroid nodule was made. Iatrogenic thyrotoxicosis was diagnosed in the presence of suppressed TSH as a consequence of exogenously administered levothyroxine for the treatment of nodular thyroid disease. Other causes of hyperthyroidism, such as iodine-induced thyrotoxicosis, subacute thyroiditis, silent thyroiditis and factitious thyrotoxicosis were also considered. Usual clinical, analytical and morphological procedures were employed to diagnose these disorders.

2010 Blackwell Publishing Ltd, Clinical Endocrinology, 72, 551557

Severe hyperthyroidism serum levels of thyroid peroxidase autoantibodies (TPOAb) and anti-thyroglobulin autoantibodies (TGAb) using an enzyme-linked immunosorbent assay (Aeskulisa, Aesku Diagnostics, Germany). TRAb was assessed by a radioisotopic method using a Wallac gamma counter. Positivity for TPOAb, TgAb and TRAb was considered when the titre of these autoantibodies was at least 150 U/ ml, 50 U/ml and 10 mU/ml respectively. Statistical analysis For quantitative variables, results are expressed as mean SD for normally distributed data, and as median (range) for nonparametric data. Conformity to a normal distribution was tested by the KolmogorovSmirnov test. Categorical variables are described as percentages. For comparisons of means between two groups of hyperthyroid patients (SH and non-SH groups), a Students t-test for normally distributed data and MannWhitney U-test for nonnormal data was used. For comparisons of means between more than two groups of hyperthyroid patients (mH, MH and SH groups) one-way anova for independent samples were used for normally distributed data, and the KruskalWallis test was employed for non-normal data. Dunnet post-hoc test and Mann Whitney U-test were used to detect whether mH or MH groups were signicantly different from the single reference group (SH). For comparisons of proportions the chi-square test or Fisher exact test was used. Variables with ordered categories were evaluated by linear by linear association. Association between quantitative variables was examined using Pearsons correlation coefcient for normal variables and Spearmans rank correlation coefcient for non-normal variables. A logistic regression analysis was used to assess: (1) the association of several covariates (demographical, clinical and analytical variables) with SH and (2) the association of several factors related with the prediction of cure in our population of thyrotoxic patients. Differences were considered signicant when P < 005.

553

novo hyperthyroidism was signicantly (P < 005) higher in SH group than in mH group. Clinical features of the patients according to severity of hyperthyroidism are summarized in Table 1. Aetiological diagnosis Although Graves disease was the most frequent aetiology of thyrotoxicosis in all studied groups (n = 79, 738%), the proportion of patients with Graves disease was signicantly (P < 005) higher in SH group (n = 18, 857%) than in mH (n = 31, 632%) and in MH (n = 30, 811%) groups. Contrary to that found in mH and MH groups, no patient with nodular thyroid disease (toxic multinodular goitre or toxic thyroid adenoma) was found in SH group. The rest of the aetiologies in each group are summarized in Table 2. Signs and symptoms Severe hyperthyroidism patients showed a higher number of symptoms at diagnosis than mH patients (Table 3). The most common complaint in SH group was asthenia followed by nervousness, dyspnoea and weight loss. SH patients lost more weight than MH and mH groups (71 77 kg vs. 58 53 kg and 39 44 kg, respectively, P < 005). Heart rate and grade of goitre were higher in SH group than in mH and MH groups (P < 001) (Table 3). No signicant differences were observed among groups in relation to the presence of goitre, intrathoracic goitre, thyroid murmur, distal tremor and exophtalmos. However, atrial brillation was more frequently reported in SH group than in mH and MH groups (158% vs. 54% and 0%, respectively, P < 005). Results from logistic regression analysis showed that younger age [OR 0958 (95% CI, 0923 0995), P = 0026], presence of asthenia [OR 435 (1481278), P = 0008] and higher heart rate [OR 103 (101106), P = 0013] were independent clinical variables associated to SH. Analytical data

Results
Studied population A total of 107 patients with overt hyperthyroidism (81 females, mean age SD 469 161 years) were evaluated. We studied a historic group with SH (n = 21; 14 females, 409 172 years) and, as a comparator group, we analysed the data of 86 hyperthyroid patients (67 females, 484 155 years, NS) comparable in age and gender. Comparator group was classied in MH (n = 37, 26 females, 472 166 years) and mH (n = 49, 41 females, 494 148 years). Although the percentage of males was higher in SH group in relation with mH and MH groups, no signicant difference in gender distribution was found. SH patients were signicantly (P < 005) younger than mH patients. We did not nd any statistically signicant difference in family history of thyroid disease and personal background among groups. The presence of both previous thyroid disease and thyroid therapy was more frequently observed in patients with milder form of hyperthyroidism (P < 001). On the other hand, the percentage of patients with de

Severe hyperthyroidism patients showed similar biochemical parameters in comparison with mH group, except for serum aspartate aminotransferase (AST) (P < 001) and calcium (P < 005) levels, that were increased, and serum cholesterol (P < 005) and albumin (P < 005) concentrations that were decreased (Table 3). A simple correlation analysis showed a positive association between FT4 levels and heart rate (r = 0309; P < 005), AST (r = 0353; P < 0001), alanine aminotransferase (ALT) (r = 0275; P < 001), and TRAb (r = 0283; P < 001), and a negative correlation with cholesterol (r = )0313; P < 001), and albumin (r = )381; P < 005) in the whole group of patients. Logistic regression analysis showed that only AST [OR 107 (102111), P = 0005] was an independent biochemical variable associated to SH. Treatment, cure rate and adequacy of therapy Eighty patients [35 (714%) mH, 27 (730%) MH and 18 (947%) SH patients] were evaluated in the follow-up study (>6 months).

2010 Blackwell Publishing Ltd, Clinical Endocrinology, 72, 551557

554 P. Iglesias et al.

Table 1. Clinical features of 107 hyperthyroid patients classied according to the severity of hyperthyroidism Comparator groups Mild hyperthyroidism No of patients (%) Age (year) Gender (male/female) Thyroid disease in relatives (yes/no) Personal history Diabetes mellitus (yes/no) Hypertension (yes/no) Hyperlipidaemia (yes/no) Atrial brillation (yes/no) Smoking (yes/no) Previous thyroid disease (yes/no) Hyperthyroidism/hypothyroidism imple goitre/ nodular goitre/STN Previous thyroid therapy (yes/no) Thyroxine (yes/no) Thionamides (yes/no) Radioiodine (yes/no) Surgery (yes/no) Type of hyperthyroidism episode De novo/relapse 49 (458) 494 148* 8/41 19/30 2/47 12/37 5/44 1/48 7/42 14/34** 12/1 1/3/2 12/37* 1/48 10/39* 1/48 1/48 33/16* Moderate hyperthyroidism 37 (356) 472 166 11/26 12/25 4/33 8/29 5/32 0/37 3/34 2/35 1/1 0/0/0 2/35 1/36 1/36 0/37 0/37 33/4 Study group Severe hyperthyroidism 21 (196) 409 172 7/14 7/14 2/19 3/18 0/21 2/19 5/16 2/19 2/1 0/0/0 1/20 1/20 1/20 0/21 0/21 19/2

Total 86 (804) 484 155 19/67 31/55 6/80 20/66 10/76 1/85 10/76 16/69 13/2 1/3/2 14/72 2/84 11/75 1/85 1/85 66/20

Data indicate mean SD or the number of patients (% of the 107 studied patients) in each studied group. STN, solitary thyroid nodule. *P < 005; **P < 001 vs. severe hyperthyroidism.

Table 2. Aetiology of thyroid hyperfunction in the three groups of studied patients Mild hyperthyroidism, n = 49 Aetiology Graves disease Toxic multinodular goitre Toxic thyroid adenoma Iodine-induced thyrotoxicosis Subacute thyroiditis Silent thyroiditis Iatrogenic thyrotoxicosis Factitious thyrotoxicosis Unknown Moderate hyperthyroidism, n = 37 Severe hyperthyroidism*, n = 21

Total, n = 107

31 (632) 11 (224) 2 (41) 0 (0) 1 (20) 1 (20) 0 (0) 1 (20) 2 (41)

30 (811) 0 (0) 1 (27) 1 (27) 3 (81) 0 (0) 1 (27) 0 (0) 1 (27)

18 (857) 0 (0) 0 (0) 0 (0) 0 (0) 1 (48) 0 (0) 2 (95) 0 (0)

79 (738) 11 (103) 3 (28) 1 (09) 4 (37) 2 (19) 1 (09) 3 (28) 3 (28)

Data indicate the number of patients and percentage (%) of the total of each group of patients. *P < 005 vs. mild and moderate hyperthyroidism groups

There was no difference in time of follow-up among three studied groups (mH 427 311 months; MH 498 434 months and SH 287 187 months, NS). During the follow-up we did not nd any signicant difference in the type of therapy, in both medical (type, number of cycles and months of therapy with thyroid drugs) and denitive treatments (radioiodine and thyroidectomy) for hyperthyroidism in the three groups of patients. The percentage of patients treated with radioiodine, as well as the number of doses and total dose of radioiodine was similar in all groups of patients. Similarly, the number of patients treated by thyroid surgery and

the type of surgery was also comparable. Lastly, the number and type of sequelae (post-surgical hypothyroidism and post-radioiodine hypothyroidism) of the treatment employed was not different in the three groups of patients. Time of follow-up since rst visit, time to achieve cure and cure criteria were also comparable among three groups. SH patients showed a slightly lower cure rate than that found in the other groups of studied patients (SH, 461%; MH, 666% and mH 686%) (Table 4). In logistic regression analysis, FT4 was the only variable that was an independent predictor of cure in our

2010 Blackwell Publishing Ltd, Clinical Endocrinology, 72, 551557

Severe hyperthyroidism
Table 3. Symptoms, signs and analytical parameters at diagnosis in 107 hyperthyroid patients classied according to the severity of hyperthyroidism Mild hyperthyroidism, n = 49 No. symptoms of hyperthyroidism <4 29 (592) 4 20 (408) Physical examination Weight (kg) 644 109 BMI (kg/m2) 255 33 SBP (mmHg) 1364 212 DBP (mmHg) 780 102 Heart rate (lpm) 842 156 Grade of goitre 1A and 1B 35 (714) 2 and 3 3 (61) Biochemical parameters Glucose (mmol/l) 56 07 Creatinine (mmol/l) 672 124 Cholesterol (mmol/l) 46 12 Triglyceride (mmol/l) 10 05 Albumin (g/l) 430 30 Calcium (mmol/l) 24 01 AST (U/l) 223 84 ALT (U/l) 304 188 GGT (U/l) 289 213 Alkaline phosphatase (U/l) 1375 769 Thyroid function tests TSH (mU/l) 0005 [00050100] FT4 (pmol/l)) 333 119 FT3 (pmol/l) 101 72 Immunological study TPOAb 500 [5002045] TgAb 500 [500650] TRAb 50 [50110] Moderate hyperthyroidism, n = 37

555

Severe hyperthyroidism, n = 21

17 (459) 20 (541) 635 116 251 41 1447 208 743 107 929 225 22 (595) 7 (189) 57 10 672 141 40 15* 09 03 410 30 24 01* 244 92 411 282 355 184 1769 1282 0010 [00050200] 704 200*** 156 126*** 500 [5004595] 500 [5001510] 160 [50320]**

7 (333) 14 (666)* 617 99 240 49 1317 171 736 98 1030 237** 11 (524) 6 (286)** 55 05 645 150 39 09* 12 08 400 40* 25 01* 318 157* 367 227 398 340 1654 1018 0005 [00050100]* 1010 000***,b 438 151***,b 890 [5004907] 500 [5001728] 230 [110690] ***,a

Data indicate the number of patients (%), mean SD and/or the median [interquartile range] in each group of study. Grade of goitre was dened according WHOs 1960 criteria (10). Ranges of reference: glucose, 4461 mmol/l; creatinine, 5301061 mmol/l; cholesterol < 52 mmol/l; triglyceride 0<17 mmol/l; albumin, 340540 g/l; calcium, 2126 mmol/l; AST, 1034 U/I; ALT, 3050 U/l; GGT, 245 U/l; Alkaline phosphatase, 44147 U/l. aP<001 bP<0001 vs. moderate hyperthyroidism group. *P < 005; **P < 001; ***P < 0001 vs. mild hyperthyroidism group. BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; AST, aspartate aminotransferase; ALT, alanine aminotransferase; GGT, gamma glutamyltranspeptidasa; TPOAb, thyroid peroxidase antibody; TgAb, thyroglobulin antibody; TSHRAb, TSH-receptor antibody. Table 4. Functional thyroid status and need of therapy in cured and not cured patients at the end of the study classied according to the severity of hyperthyroidism Cure Yes Without sequelae With sequelae No No treatment Antithyroid drugs

Mild hyperthyroidism, n = 35

Moderate hyperthyroidism, n = 27

Severe hyperthyroidism, n = 18

Total, n = 80

24 (686) 13 (542) 11 (458) 11 (314) 10 (909) 1 (91)

18 (666) 7 (389) 11 (611) 9 (333) 5 (555) 4 (445)

8 (461) 4 (500) 4 (500) 10 (555) 2 (20) 8 (80)**

50 (625) 24 (480) 26 (520) 30 (375) 17 (567) 13 (433)

Data indicate the number of patients (%). Cure without sequelae indicates thyroid normofunction without therapy for >1 year; cure with sequelae indicates postradioiodine or postsurgical hypothyroidism and the need of thyroxine therapy. **P < 001 vs. mild and moderate hyperthyroidism groups.

2010 Blackwell Publishing Ltd, Clinical Endocrinology, 72, 551557

556 P. Iglesias et al. population of thyrotoxic patients [OR 0983 (CI 95%, 0967 0998), P < 005]. At the end of follow-up thyroid functional status was similar among three studied groups. At this time, the percentage of patients with normal thyroid function tests (whole group, n = 58, 725%; 38 without and 20 with therapy) was similar in all studied groups: mH group (743%; 19 without and seven with therapy), MH group (741%; 13 without and seven with therapy) and SH group (667%; six without and six with therapy). However, the percentage of these patients treated with antithyroid drugs was signicantly higher in SH group in comparison with patients with milder forms of hyperthyroidism (P < 0001) (Table 4). these effects, as SH patients showed not only a higher degree of goitre, but also a more elevated titre of TSHR-Ab. Lastly, although atrial brillation is uncommon in Graves patients younger than 50 years and usually occurs in up to 20% of older patients,11 our study revealed a high prevalence (16%) of this arrhythmia in a middle age group of patients with SH. Hyperthyroidism can lead to a glucose intolerance or impair glycaemic control in diabetic patients.13 We found that severity of thyroid hyperfunction was not related to baseline serum glucose levels. This could be explained by differences in age and, possibly, in insulin sensitivity of the studied subgroups. In our survey, the presence of a SH was associated with a signicant decreased in cholesterol compared with mH group; however, serum triglyceride levels were not statistically different, probably becuase of the different age, as well as body fat content and distribution of the different studied groups. SH patients showed signicantly decreased serum albumin, thus suggesting a catabolic action of thyroid hormone excess. The association of hypercalcaemia and thyrotoxicosis is well known.14,15 Our study showed that serum calcium levels was an independent biochemical variable associated to SH. Serum calcium rose in line with the severity of hyperthyroidism. In this setting, the percentage of patients with hypercalcaemia (Ca >26 mmol/l) was 143%, 79% and 0% in SH, MH and mH respectively. Therefore, it is important to know this association because of the symptoms of hypercalcaemia (anorexia, nausea, vomiting, polyuria and occasionally impairment of kidney function) may complicate the clinical course of SH. Liver function tests are altered in three-quarters of Graves disease patients.16 Our results conrmed the alterations in liver function tests associated with hyperthyroidism and showed a positive correlation between serum levels of AST and ALT and those of FT4. AST was an independent covariable associated to SH. The selection of the different therapeutic modalities for hyperthyroidism depends on several factors such as patients age, aetiology, size of goitre, associated symptoms, preferences of patient or physician, and availability of nuclear medicine facilities and an experienced endocrine surgeon.6,1719 In our study, the severity of hyperthyroidism did not inuence on the type of treatment used, both medical (antithyroid drugs) and denitive (radioiodine or surgery), or on the number and type of sequelae. No signicant difference in the cure rate was found among the three groups of studied patients. According to these results, we treated hyperthyroid patients regardless of the severity of hyperthyroidism, and they were cured in the same proportion. Although no signicant difference in cure rate was found among the three groups of studied patients, SH patients showed a slightly lower cure rate. This nding, together with the presence of signicantly higher titres of TRAb, the positive association of baseline T4 levels with the probability of cure, and the higher need for long-term treatment with antithyroid drugs to keep thyroid normofunction in SH group, suggests that the use of more aggressive therapeutic measures in these patients might be appropriate. The main limitation of our study is related to the small number of patients with SH in our survey. Obviously, this fact limits the generalisability of the results. The limited number of patients

Discussion
Our data suggest that, in comparison with patients with milder forms of thyrotoxicosis, patients with SH at the time of diagnosis show a lower frequency of previous thyroid disease and thyroid therapy and a higher incidence of de novo hyperthyroidism, have Graves disease as the main aetiology and exhibit higher heart rate, larger goitre, more analytical alterations and higher titres of TRAb. Cure rate and grade of control of thyroid function in SH patients is similar or slightly lower to the milder forms of hyperthyroidism. From this study we can estimate the percentage of patients with SH with respect to our population with overt hyperthyroidism. Because of we attended 86 patients with overt hyperthyroidism in a 6-month period, we would have attended 2580 patients in a 15-year period, assuming similar epidemiological conditions. We found 21 patients with SH, i.e., about 081% of this hypothetical group of hyperthyroid patients. Therefore, we can estimate that the proportion of SH in our population is about 08% of subjects with overt hyperthyroidism (<1%). The aetiology of hyperthyroidism is related to the age, gender and iodine intake. Graves disease has a peak of incidence at 20 40 years, whereas nodular thyroid disease is more frequently reported with advanced age.6,11 Our study, performed in population living in a low-sufcient iodine intake area (central region of Spain, median urinary iodine excretion 110 mg/l),12 showed an inverse relationship between age and the severity of hyperthyroidism. On the other hand, the main aetiology in SH group was Graves disease. Another nding from our study was that most SH patients showed a de novo hyperthyroid episode. This nding suggests that it is truly exceptional, on one hand, the progression towards a SH from milder forms of hyperthyroidism and, on the other hand, the development of SH as a relapse of previous hyperthyroidism. Severe hyperthyroidism was accompanied by a signicantly higher number of symptoms at diagnosis than mH. SH was clinically characterized by the presence of asthenia followed by nervousness, dyspnoea and weight loss. Loss of weight in SH patients, was also higher than in other milder forms of hyperthyroidism. In relation to signs, only heart rate and degree of goitre were signicantly different in SH patients compared with subjects with milder forms of hyperthyroidism. The titre of TRAb was related with goitre size and disease activity.6,7,11 Our results also are in agreement with

2010 Blackwell Publishing Ltd, Clinical Endocrinology, 72, 551557

Severe hyperthyroidism studied could have caused a lack of statistical signicance in the differences found in treatment modalities and outcomes. Further studies with larger samples are warranted, although the low prevalence of SH makes such studies difcult. Another limitation is the time factor in diagnosis in the different study groups. However, many patients in the comparator group were diagnosed with hyperthyroidism many years before the study. This fact might cut down the differences among groups in terms of time. On the other hand, although the use of radioiodine may have increased in recent years, especially in patients with Graves disease, in our clinic it has always been considered as rst-line therapy in patients with hyperthyroidism caused by nodular thyroid disease and second-line therapy in patients with Graves disease presenting with relapse after a 6- to 12-month cycle of antithyroid drug treatment. In summary, Graves disease is the most common aetiology in patients with SH. This type of hyperthyroidism is usually de novo and is accompanied by more clinical signs, symptoms, and analytical derangements, as well as higher titres of TRAb at diagnosis than milder forms of hyperthyroidism. The present data are not able to show differences in treatment modality, time to achieve cure and remission rate among patients with mild, moderate and severe hyperthyroidism.

557

Competing interests/nancial disclosure

There is no conict of interest.

References
1 Tunbridge, W.M., Evered, D.C., Hall, R. et al. (1977) The spectrum of thyroid disease in a community: the Whickham survey. Clinical Endocrinology (Oxford), 7, 481493. 2 Hollowell, J.G., Staehling, N.W., Flanders, W.D. et al. (2002) Serum TSH, T(4), and thyroid antibodies in the United States population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III). Journal of Clinical Endocrinology and Metabolism, 87, 489499. 3 Aghini-Lombardi, F., Antonangeli, L., Martino, E. et al. (1999) The spectrum of thyroid disorders in an iodine-decient community: the Pescopagano survey. Journal of Clinical Endocrinology and Metabolism, 84, 561566.

4 Karger, S. & Fuhrer, D. (2008) Thyroid storm thyrotoxic crisis: an update. Deutsche medizinische Wochenschrift, 1946, 133. 5 Nayak, B. & Burman, K. (2006) Thyrotoxicosis and thyroid storm. Endocrinology and Metabolism Clinics of North America, 35, 663 686. 6 Cooper, D.S. (2003) Hyperthyroidism. Lancet, 362, 459468. 7 Nayak, B. & Hodak, S.P. (2007) Hyperthyroidism. Endocrinology and Metabolism Clinics of North America, 36, 617656. 8 Mieny, C.J., Franz, R.C. & Venter, I.D. (1982) The management of severe hyperthyroidism. World Journal of Surgery, 6, 689695. 9 Burch, H.B. & Wartofsky, L. (1993) Life-threatening thyrotoxicosis. Thyroid storm. Endocrinology and Metabolism Clinics of North America, 22, 263277. 10 Perez, C., Scrimshaw, S. & Munoz, A. Technique of endemic goitre surveys. (1960) Endemic Goitre. WHO, Geneva, 369383. 11 Weetman, A.P. (2000) Graves disease. New England Journal of Medicine, 343, 12361248. 12 Morreale de Escobar, G. & Escobar del Rey, F. (2003) Consequences of iodine deciency for brain development. In: G. Morreale de Escobar, J.J.M. de Vijler, S. Butz & U. Hostalek eds. Thyroid and Brain. Schattauer Verlag, Stuttgart, 3356. 13 Maxon, H.R., Kreines, K.W., Goldsmith, R.E. et al. (1975) Longterm observations of glucose tolerance in thyrotoxic patients. Archives of Internal Medicine, 135, 14771480. 14 Diamond, T., Vine, J., Smart, R. et al. (1994) Thyrotoxic bone disease in women: a potentially reversible disorder. Annals of Internal Medicine, 120, 811. 15 Iqbal, A.A., Burgess, E.H., Gallina, D.L. et al. (2003) Hypercalcemia in hyperthyroidism: patterns of serum calcium, parathyroid hormone, and 1,25-dihydroxyvitamin D3 levels during management of thyrotoxicosis. Endocrine Practice, 9, 517521. 16 Kubota, S., Amino, N., Matsumoto, Y. et al. (2008) Serial changes in liver function tests in patients with thyrotoxicosis induced by Graves disease and painless thyroiditis. Thyroid, 18, 283287. 17 Cooper, D.S. (2005) Antithyroid drugs. New England Journal of Medicine, 352, 905917. 18 Kobe, C., Weber, I., Eschner, W. et al. (2008) Graves disease and radioiodine therapy. Is success of ablation dependent on the choice of thyreostatic medication?. Nuklearmedizin, 47, 153156. 19 de Ruijter, S.H., van Dalen, T. & Muller, A.F. (2006) The surgical treatment of Graves disease. NederlandsTijdschrift Voor Geneeskd, 150, 13211325.

2010 Blackwell Publishing Ltd, Clinical Endocrinology, 72, 551557

Copyright of Clinical Endocrinology is the property of Blackwell Publishing Limited and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.