Rapid Differential Diagnosis of Graves Disease and Painless Thyroiditis Using Total T3/T4 Ratio, TSH, and Total Alkaline Phosphatase Activity

Transcription

1 Endocrine Journal 2005, 52 (1), Rapid Differential Diagnosis of Graves Disease and Painless Thyroiditis Using Total T3/T4 Ratio, TSH, and Total Alkaline Phosphatase Activity TETSUO YANAGISAWA, KANJI SATO, YOSHIYUKI KATO, SATORU SHIMIZU* AND KAZUE TAKANO Department of Medicine, Institute of Clinical Endocrinology, Tokyo Women s Medical University, Tokyo, Japan *Department of Hygiene and Public Health, Tokyo Women s Medical University, Tokyo, Japan Abstract. When thyrotoxic patients are first seen in an outpatient clinic, it is important to make a differential diagnosis of Graves disease (GD) and painless thyroiditis (PT). Using the three parameters of total T 3 /T 4 ratio, TSH and T-ALP activity, all of which can be obtained within one hour in our hospital, 173 untreated patients with thyrotoxicosis were evaluated for the ratio of each parameter. For GD vs. PT, total T 3 /T 4 (ng/ g) ratio (>20), TSH (<0.005 U/ml) and elevated T-ALP had a likelihood ratio of 2.14, 2.12 and 4.07, respectively. When a patient had all three positive parameters, the likelihood ratio increased to 5.81, which showed effective synergy. These data suggest that, in addition to total T 3 /T 4 ratio and TSH value, T-ALP activity is a useful parameter for the rapid differentiation of GD. The lower T- ALP activity seen in PT is probably due to the fact that patients develop less severe thyrotoxicosis or visit hospital earlier than patients with GD. Therefore, the diagnostic accuracy for Graves disease might be increased by using the three parameters in combination. Key words: Graves disease, Painless thyroiditis, Subacute thyroiditis, Alkaline phosphatase (Endocrine Journal 52: 29 36, 2005) WHEN thyrotoxic patients undergo their first medical examination, it is important to differentiate stimulation-induced thyrotoxicosis (Graves disease) from destruction-induced thyrotoxicosis (painless thyroiditis and subacute thyroiditis), since the latter disease does not require treatment with antithyroid agents [1]. Diagnosis of typical cases of subacute thyroiditis is not so difficult, but it is often problematic to differentiate Graves disease from painless thyroiditis, unless radioactive iodine uptake (RAIU) is determined. The RAIU test is contraindicated, however, when patients are lactating, and it is not practical for all patients with thyrotoxicosis to have an RAIU test to differentiate between the two types. Received: August 23, 2004 Accepted: October 20, 2004 Correspondence to: Kanji SATO, M.D., Ph.D., Institute of Clinical Endocrinology, Tokyo Women s Medical University Kawada-cho 8-1, Shinjuku-ku, Tokyo, Japan In a previous study, Amino et al. [2] showed that a total T 3 /T 4 ratio of <20 ng/dl/ g/dl was indicative of destruction-induced thyrotoxicosis. Recently, these authors also suggested that destruction-induced thyrotoxicosis is indicated when the peripheral eosinophil/ monocyte ratio multiplied by the free T 3 concentration gives a value of <4.5 [3]. Kasagi et al. [4] demonstrated that incomplete suppression of thyroid stimulating hormone (TSH; >0.005~ U/ml) was also indicative of destruction-induced thyrotoxicosis. Furthermore, the zinc concentration in red blood cells has also been reported to be useful for such differentiation [5]. Measurement of TSH receptor antibodies (TSHbinding inhibitory immunoglobulins; TBII) is useful for the diagnosis of Graves disease, but approximately 2 10% of Graves patients are negative for TBII [6, 7], and conversely some patients with painless thyroiditis may be positive [8]. Although thyroid stimulating antibody (TASb) is most pathognomonic for stimulationinduced hyperthyroidism, it is not always positive in

2 30 YANAGISAWA et al. thyrotoxic patients with Graves disease [9, 10]. Moreover, several days are required before these results are known. Since thyroid hormone receptors are relatively abundant in osteoblasts [11, 12], thyroid hormones stimulate bone metabolism [13] through increased expression of the receptor activator of NF-kB ligand (RANKL) on osteoblasts [14], which interacts with its receptor (RANK) expressed on osteoclast precursors, leading to enhanced osteoclast formation and bone resorption [15]. From a histomorphometric study of tetracycline double-labeled iliac crest bone biopsies, osteoblastic bone formation begins 3 weeks after the resorption period in the state of thyrotoxicosis [16]. Serum levels of total (T)-ALP or bone-derived alkaline phosphatase (B-ALP) activity in normal subjects have been documented to take at least one month to increase significantly after exogenous triiodothyronine or thyroxine treatment [17, 18]. Indeed, B-ALP is increased in thyrotoxic patients with Graves disease, particularly after treatment with antithyroid agents [19, 20]. Since Graves disease develops insidiously and destruction-induced thyrotoxicosis develops subacutely, there may be significant differences in levels of B-ALP activity at the patient s first visit to hospital. Although B-ALP activity reflects bone formation better than T-ALP activity, it cannot be measured by multi-autoanalyzer and it takes a few days to obtain the results. In the current study, therefore, T-ALP activity was measured instead, giving a result within one hour. The purpose of the study was to determine to what extent the three parameters, total T 3 /T 4 ratio, TSH value, T-ALP activity, were useful for the rapid differential diagnosis of thyrotoxicosis, either when taken alone or in various combinations. All three parameters can be measured within one hour in our hospital. Subjects and Methods Three groups of patients with untreated thyrotoxicosis who visited our outpatient clinic between 2000 and 2003 were enrolled. They included 126 patients with Graves disease (female/male ratio, 96/30; age, 41.2 ± 15.3 years), 20 with painless thyroiditis (female/ male ratio 17/3; age, 36.0 ± 12.9 years), and 27 with subacute thyroiditis (female/male ratio 21/6; age, 48.8 ± 11.4 years). Graves disease was diagnosed on the basis of clinical findings and laboratory tests showing a high value of total T 4 (T 4 ) and total T 3 (T 3 ), a low level of TSH, increased TBII and/or TSAb activity, and/or high thyroidal uptake of 123 I or 99m Tc. Diagnosis of painless thyroiditis was made on the basis of increased T 4 and T 3 levels for less than 3 months, with suppressed TSH, and/or later development of transient hypothyroidism, and/or a low RAIU test value. Diagnosis of subacute thyroiditis was made on the basis of clinical features of fever and thyroid tenderness, and laboratory findings of increased C-reactive protein, T 4 and T 3, and/or presence of hypoechoic areas in the thyroid on ultrasonography. All serum samples were obtained in the thyrotoxic phase before treatment. No patient received medication with a known influence on bone metabolism or liver function, or had a history or present signs or symptoms of bone or liver disease. Since the reference range of T-ALP activity differs according to age and sex, the mean value and standard deviation (SD) of T-ALP activity were determined according to each sex and age using serum obtained from apparently healthy participants. Mean value and SD of T-ALP activity obtained from normal samples (n = 150) was 230 ± 53 IU/l for younger males (<50 years), 236 ± 51 IU/l for older males ( 50 years), 190 ± 56 IU/l for younger females (<50 years), and 249 ± 63 IU/l for older females ( 50 years). The study was approved by the scientific committee of the hospital, and informed consent was obtained from every subject who participated in the study. Assays Serum values of T 4 and T 3 were measured using commercial kits (T 4, Ecuresis T 4 II; T 3, Ecuresis T 3 ; Roche Diagnostic Co., Ltd., Tokyo, Japan). Normal ranges of T 4 and T 3 were g/dl and ng/dl, respectively. Serum values of TSH were measured with a third-generation ECLIA kit (Ecuresis TSH; Roche Diagnostic Co., Ltd., Tokyo, Japan). The normal range of TSH was U/ml. Serum values of TBII were measured by radioreceptor assay using a commercial kit (Dynotest TRAb human; Yamasa Co., Ltd., Japan). The results were expressed as percentage inhibition of binding of labeled TSH. The normal value was <10%. TSAb activities were measured in terms of the amount of camp produced in cultured porcine thyrocytes [9]. The normal range was less than 180%. T-ALP activity was measured with a commercial kit (Pureauto S ALP; Daiichi Pure Chemi-

3 RAPID DIFFERENTIAL DIAGNOSIS OF THYROTOXICOSIS 31 cals Co., Ltd., Tokyo, Japan), using nitrophenylphosphate as a substrate. The normal range for each parameter is indicated above. Statistical analysis To evaluate statistically to what extent these positive parameters (total T 3 /T 4 ratio>20, TSH value completely suppressed (<0.005 U/ml), and elevated T-ALP activity) were useful for the differential diagnosis of Graves disease, the likelihood ratio of each parameter was calculated [21]. To improve the differential diagnosis, the likelihood ratio using the three parameters in combination was also calculated. Patients for whom all three parameters were determined were assigned a score of 0 3 according to the number of positive parameters, and the likelihood ratio of each score was then calculated. The results of all variables are reported as the mean ± SD. Significant prognostic valuables were identified using analysis of variance models. A p-value of less than 0.05 was regarded as indicating a statistically significant difference. Statistical analyses were performed using ANOVA procedure, SAS system Version 8 (SAS Institute, Cary, NC, USA). Results The total T 3 concentration in patients with Graves disease (467.1 ± ng/dl, n = 94) was significantly (P<0.05) higher than that in patients with subacute thyroiditis (287.8 ± ng/dl, n = 15), but not those with painless thyroiditis (340.2 ± ng/dl, n = 17). As shown in Fig. 1, the total T 3 /T 4 ratio in patients with Graves disease (24.0 ± 5.6 ng/ g, n = 94) was significantly (P<0.05) higher than that in patients with painless thyroiditis (19.3 ± 6.2 ng/ g, n = 17) or subacute thyroiditis (17.5 ± 5.5 ng/ g, n = 15). It should be pointed out, however, that one patient with painless thyroiditis had an extremely high T 3 /T 4 ratio (39.3), with serum levels of T 3 and T 4 of 1060 ng/dl and 27 g/ dl, respectively. To analyze the relationship between T 3 /T 4 ratio and T 4 value, serum levels of total T 3 and T 4 in patients with thyrotoxicosis were plotted (Fig. 2). In Graves disease, 75.5% (71/94) of patients had a T 3 /T 4 ratio of >20, which became more evident as serum T 4 levels exceeded 20 g/dl. In contrast, the T 3 /T 4 ratio was generally <20 in patients with destruction-induced Fig. 1. Total T 3 /T 4 ratio in the three groups of patients with thyrotoxicosis. One patient with painless thyroiditis had an extremely high ratio (38.3). thyrotoxicosis (painless thyroiditis 64.7%; subacute thyroiditis 66.7%), but the ratio increased in a T 4 - dependent manner, and became >20 in four of nine patients with a T 4 level >20 g/dl. As shown in Fig. 3, most of the patients with thyrotoxicosis had a serum TSH level which was less than the normal range (<0.380 U/ml). Consistent with previous reports [4, 22], TSH was suppressed to undetectable levels (<0.005 U/ml) in the majority of patients with Graves disease (80/113, 70.8%), whereas it remained in the detectable range ( U/ml) in painless thyroiditis (12/18, 66.7%) and subacute thyroiditis (14/19, 73.7%). Fig. 4 shows T-ALP activity in patients with Graves disease, painless thyroiditis, and subacute thyroiditis. T-ALP activity in Graves disease (440 ± 207 IU/l, n = 123) was significantly (P<0.05) higher than in patients with painless thyroiditis (232 ± 114 IU/l, n = 19), or those with subacute thyroiditis (332 ± 152 IU/l, n = 24). To evaluate statistically to what extent these positive parameters (total T 3 /T 4 ratio >20, completely suppressed TSH levels, and elevated T-ALP activity) were useful for the differential diagnosis of Graves disease, the likelihood ratio of each parameter was calculated (Table 1). Among the three parameters, elevated T- ALP activity was most closely related with a likelihood

4 32 YANAGISAWA et al. Fig. 2. Relationship between serum T 3 /T 4 ratio and T 4 value in patients with Graves disease (A) and destruction-induced thyrotoxicosis (B). In (B), the clear circles correspond to painless thyroiditis and solid dots correspond to subacute thyroiditis. Upper-left areas above the line represent a T 3 /T 4 ratio of >20. Fig. 3. TSH values in the three groups patients with thyrotoxicosis. Dotted area represents the normal range. Broken line represents the detection limit of the assay. of 4.07, followed by T 3 /T 4 ratio (2.14) and complete TSH suppression (2.12). Since a likelihood ratio based on a single parameter was not significant, the likelihood ratio using all three parameters in combination was then calculated. One hundred and seven of the 173 patients for whom all three parameters were determined were assigned a score of 0 3 according to the number of positive parameters they had. As shown in Table 2, the likelihood ratio for Graves disease vs. painless thyroiditis was Fig. 4. Total alkaline phosphatase activity in the three groups of patients with thyrotoxicosis. Shaded areas represent the normal range for each sex and age group. M, male; F, female for a score of 0, 0.36 for a score of 1 and 2.22 for a score of 2; however, it increased significantly to 5.81

5 RAPID DIFFERENTIAL DIAGNOSIS OF THYROTOXICOSIS 33 Table 1. Frequency and likelihood ratio of each parameter in differentiation of Graves disease vs. painless thyroiditis GD PT ST Sensitivity Sensitivity Sensitivity Likelihood ratio for GD vs. PT T 3 /T 4 20 (ng / g) (71/94) (6/17) (5/15) TSH<0.005 ( U/ml) (80/113) (6/18) (5/19) T-ALP (IU/l) (79/123) (3/19) (7/24) Table 2. Frequency and likelihood ratio of each score in differentiation of Graves disease vs. painless thyroiditis GD PT ST Sensitivity Sensitivity Sensitivity Likelihood ratio for GD vs. PT 0 point (3/82) (2/14) (5/11) point (19/82) (9/14) (3/11) point (26/82) (2/14) (3/11) point (34/82) (1/14) (0/11) when thyrotoxic patients had a score of 3. One patient with severe thyrotoxicosis with a serum T 3 level of 1060 ng/dl and a T 4 level of 27 g/dl showed an extremely high T 3 /T 4 ratio (39.3); however, in this patient the serum level of TSH was not suppressed completely (0.012 U/ml) and T-ALP was also within the normal range. Thyroidal 99m Tc uptake was suppressed to <1% at 20 min, and a diagnosis of painless thyroiditis was made in this patient with only one positive parameter. Another thyrotoxic patient had a high T 3 /T 4 ratio of >20, TSH was suppressed to undetectable levels, and T-ALP was elevated. Although TRAb and TSAb were both negative [10], thyroidal 123 I uptake was increased to 35% at 4 h, and therefore a diagnosis of Graves disease was confirmed in this patient with three positive parameters. Discussion In accordance with previous reports [2], total T 3 /T 4 ratio was usually >20 in Graves patients, and was relatively useful for the differentiation of thyrotoxicosis (Fig. 1). One reason for the increased T 3 /T 4 ratio in Graves patients is that type II deiodinase, which is expressed abundantly in human thyroid tissue [23], is stimulated by TSH receptor-stimulating antibodies [24]. However, one patient with painless thyroiditis had an extremely high T 3 /T 4 ratio of >30 (Fig. 1), with a high serum level of T 3 (1060 ng/dl) and T 4 (27 g/dl). Since type I deiodinase is stimulated by thyroid hormone in the peripheral organs (liver, kidney) [25], increased type I deiodinase activity was involved, at least in part, in the elevated T 3 /T 4 ratio in this patient. As reported previously [4], serum levels of TSH were completely suppressed (<0.005 U/ml) in most untreated Graves patients, whereas they usually were not completely suppressed in patients with painless thyroiditis or subacute thyroiditis (Fig. 3). Incomplete suppression of serum TSH in destruction-induced thyrotoxicosis is probably due to less intense thyrotoxicosis in these patients. Another likely explanation is that patients with destruction-induced thyrotoxicosis visit hospital earlier after the thyrotoxicosis has developed than Graves patients. Because onset of the former condition is subacute, patients would readily notice the onset of symptoms. It has been reported that it takes approximately 3 4 weeks for TSH to be suppressed completely in thyrotoxic patients after serum T 4 levels increase to thyrotoxic levels [22]. Although serum levels of T-ALP were usually above the upper limit in Graves disease, they were often within the reference range in destruction-induced thyrotoxicosis (Fig. 4). Thyroid hormone receptors are abundant in osteoblast-like cells [11, 12]. Recent studies have revealed that thyroid hormones activate bone remodeling by acting on osteoblasts[13], and stimulate expression of RANKL on their cell surface [14]. RANKL on osteoblasts binds to the RANK receptor expressed on osteoclast progenitor cells and induces their differentiation into multi-nucleated osteoclasts, leading to enhanced bone resorption, which is followed by increased osteoblastic bone formation [15, 26]. Alkaline phosphatase is an ectoenzyme attached to the

6 34 YANAGISAWA et al. cell membrane by a hydrophobic glycosyl-phosphatidylinositol anchor, and is released into the blood stream as a response to several stimuli [27, 28]. A recent in vitro study has shown that triiodothyronine has a stimulatory effect on the release of membrane-bound ALP by osteoblastic cells [29]. Although bone resorption is stimulated in thyrotoxicosis, osteoblastic bone formation begins approximately 3 weeks after the resorption period [16], and serum levels of T-ALP or B-ALP have been reported to increase significantly from 6 9 weeks after exogenous triiodothyronine or thyroxine treatment [17, 18], but not to increase significantly before 3 weeks [30]. Therefore, it will take at least one month for serum T-ALP or B-ALP activity to increase significantly after thyrotoxicosis has developed. Because Graves disease develops insidiously, patients would not readily become aware of subjective symptoms compared with patients with painless thyroiditis, which develops subacutely. Therefore, when a thyrotoxic patient with diffuse and painless goiter accompanied by elevated T-ALP activity visits hospital, it is likely that the thyrotoxicosis will have developed at least one month earlier. Serum T-ALP activity was lower in patients with subacute thyroiditis than in those with Graves disease. According to the hypothesis presented above, T-ALP activities can be expected to be lowest in subacute thyroiditis, because patients with subacute onset of painful thyroiditis would visit hospital the earliest among the three groups. Subacute thyroiditis is supposedly caused by viral infection of undetermined origin [31], and the affected organs include not only the thyroid but also the pancreas and liver [32]; elevated T-ALP activity is thought to be derived from extraskeletal tissue [33]. In general, likelihood ratios of >10 or <0.1 generate large and often conclusive changes from pre- to posttest probability, but ratios of 1 2 and alter probability to a small degree [21]. Although T-ALP activity had the highest likelihood ratio of 4.07 (Table 1), the three parameters were used in combination to improve the pretest probability to differentiate Graves disease from painless thyroiditis. As shown in Table 2, patients who had three positive parameters had a likelihood ratio of 5.81 for Graves disease vs. painless thyroiditis. This means that the diagnostic accuracy for Graves disease was increased by using the parameters in combination; thus, an untreated thyrotoxic patient with three positive parameters would be diagnosed with Graves disease rather than painless thyroiditis. It is considered that these simple parameters may be useful for the differentiation of Graves disease and painless thyroiditis, although a definite diagnosis would be made with the RAIU test or more simply by clinical observation for a few weeks without prescribing antithyroidal agents. In summary, in addition to total T 3 /T 4 ratio and TSH values, T-ALP activity also appears to be useful for the diagnosis of Graves disease. Since all three parameters can be measured within one hour in large hospitals, measurement of total T 3 /T 4 ratio, TSH and T-ALP would be of value for the rapid differential diagnosis of Graves disease and painless thyroiditis, although a definite diagnosis can only be made with the thyroidal radioisotope uptake test. Reference 1. Cooper DS (2001) Treatment of thyrotoxicosis. In: Braverman LR, Utiger RD (eds) Werner & Ingbar s Thyroid. 8th ed, J. B. Lippincott Co., Philadelphia, Amino N, Yabu Y, Miki T, Morimoto S, Kumahara Y, Mori H, Iwatani Y, Nishi K, Nakatani K, Miyai K (1981) Serum ratio of triiodothyronine to thyroxine, and thyroxine-binding globulin and calcitonin concentrations in Graves disease and destruction-induced thyrotoxicosis. J Clin Endocrinol Metab 53: Izumi Y, Hidaka Y, Tada H, Takano T, Kashiwai T, Tatsumi KI, Ichihara K, Amino N (2002) Simple and practical parameters for differentiation between destruction-induced thyrotoxicosis and Graves thyrotoxicosis. Clin Endocrinol 57: Kasagi K, Kousaka T, Misaki T, Iwata M, Alam MS, Konishi J (1999) Comparison of serum thyrotrophin concentrations determined by a third generation assay in patients with various types of overt and subclinical thyrotoxicosis. Clin Endocrinol 50: Sayama N, Yoshida K, Mori K, Fukazawa H, Hori H, Nakazato N, Tani J, Nakagawa Y, Ito S (1998) Measurement of red blood cell zinc concentration with Zntest kit: discrimination between hyperthyroid Graves disease and transient thyrotoxicosis. Endocr J 45:

7 RAPID DIFFERENTIAL DIAGNOSIS OF THYROTOXICOSIS Schott M, Feldkamp J, Bathan C, Fritzen R, Scherbaum WA, Seissler J (2000) Detecting TSHreceptor antibodies with the recombinant TBII assay: technical and clinical evaluation. Horm Metab Res 32: Costagliola S, Morgenthaler NG, Hoermann R, Badenhoop K, Struck J, Freitag D, Poertl S, Weglohner W, Hollidt JM, Quadbeck B, Dumont JE, Schumm- Draeger PM, Bergmann A, Mann K, Vassart G, Usadel KH (1999) Second generation assay for thyrotropin receptor antibodies has superior diagnostic sensitivity for Graves disease. J Clin Endocrinol Metab 84: Morita T, Tamai H, Oshima A, Mukuta T, Fukata S, Kuma K, Kumagai LF, Nagataki S (1990) The occurrence of thyrotropin binding-inhibiting immunoglobulins and thyroid-stimulating antibodies in patients with silent thyroiditis. J Clin Endocrinol Metab 71: Yoshimura-Noh J, Ito K, Uchimura H, Onaya T, Takasu N, Kasagi K, Yamashita K, Ochi Y, Yokoyama N, Nagataki S (1994) Clinical evaluation of thyroid stimulating antibody (TSAb) by Yamasa TSAb kit. Horumon to Rinsho 42: (in Japanese). 10. Sato K, Yamazaki K, Yamada E, Kanaji Y, Miura M, Obara T (1999) Immunoglobulin G of untreated Graves patients with or without TSH receptor antibody (determined by porcine thyrocytes) universally elicit potent thyroid hormone-releasing activity in cultured human thyroid follicles. Thyroid 9: Sato K, Han DC, Fujii Y, Tsushima T, Shizume K (1987) Thyroid hormone stimulates alkaline phosphatase activity in cultured rat osteoblastic cells (ROS 17/2.8) through 3,5,3'-triiodo-L-thyronine nuclear receptors. Endocrinology 120: Abu SO, Bord S, Horner A, Chatterjee VK, Compaton JE (1997) The expression of thyroid hormone receptors in human bone. Bone 21: Britto JM, Fenton AJ, Holloway WR, Nicholson GC (1994) Osteoblasts mediate thyroid hormone stimulation of osteoclastic bone resorption. Endocrinology 134: Miura M, Tanaka K, Komatsu Y, Suda M, Yasoda A, Sakuma Y, Ozaka A, Nakao K (2002) A novel interaction between thyroid hormones and 1,25(OH) 2 D 3 in osteoclast formation. Biochem Biophys Res Commun 291: Suda T, Takahashi N, Udagawa N, Jimi E, Gillespie MT, Martin TJ (1999) Modulation of osteoclast differentiation and function by the new members of the tumor necrosis factor receptor and ligand families. Endocr Rev 20: Eriksen EF, Mosekilde L, Melsen F (1985) Trabecular bone remodeling and bone balance in hyperthyroidism. Bone 6: Hasling C, Eriksen EF, Charles P, Mosekilde L (1987) Exogenous triiodothyronine activates bone remodeling. Bone 8: Smith SR, Lovejoy JC, Rood J, Most M, Wickersham PJ, Volaufova J, Ryan D, Tulley R, Bray GA (2003) The effects of triiodothyronine on bone metabolism in healthy ambulatory men. Thyroid 13: Cooper DS, Kaplan MM, Ridgway EC, Maloof F, Daniels GH (1979) Alkaline phosphatase isoenzyme patterns in hyperthyroidism. Ann Intern Med 90: Siddiqi A, Burrin JM, Noonan K, James I, Wood DF, Price CP, Monson JP (1997) A longitudinal study of markers of bone turnover in Graves disease and their value in predicting bone mineral density. J Clin Endocrinol Metab 82: Jaeschke R, Guyatt G, Lijmer J (2002) Diagnosis tests. In Guyatt G, Rennie D (eds) User s Guides to the Medical Literature, ch.1c2, Chicago: American Medical Association Ito M, Takamatsu J, Yoshida S, Murakami Y, Sakane S, Kuma K, Ohsawa N (1997) Incomplete thyrotroph suppression determined by third generation thyrotropin assay in subacute thyroiditis compared to silent thyroiditis or hyperthyroid Graves disease. J Clin Endocrinol Metab 82: Salvatore D, Tu H, Harney JW, Larsen PR (1996) Type 2 iodothyronine deiodinase is highly expressed in human thyroid. J Clin Invest 15: Murakami M, Araki O, Hosoi Y, Kamiya Y, Morimura T, Ogiwara T, Mizuma H, Mori M (2001) Expression and regulation of type II iodothyronine deiodinase in human thyroid gland. Endocrinology 142: Bianco AC, Salvatore D, Gereben B, Berry MJ, Larsen PR (2002) Biochemistry, cellular and molecular biology, and physiological roles of the iodothyronine selenodeiodinases. Endocr Rev 23: Hofbauer LC, Heufelder AE (2000) The role of receptor activator of nuclear factor-kb ligand and osteoprotegerin in the pathogenesis and treatment of metabolic bone disease. J Clin Endocrinol Metab 85: Mose DW (1997) Physicochemical and pathophysiological factors in the release of membrane-bound alkaline phosphatase from cells. Clin Chim Acta 3: Henthorn PS, Paula SH (1996) Alkaline phosphatase. In Bilezikian JP, Raisz LG, Rodan GA (eds), Principles of Bone Biology, first ed, San Diego, CA, Academic Press, Banovac K, Koren E (2000) Triiodothyronine stimulates the release of membrane-bound alkaline phosphatase in osteoblastic cells. Calcif Tissue Int 67: Frevert EU, Biester A, Muller MJ, Schmidt-Gayk H, von zur Muhlen A, Brabant G (1994) Markers of bone metabolism during short-term administration of thyrox-

8 36 YANAGISAWA et al. ine in healthy volunteers. Eur J Endocrinol 131: Emerson CH, Farwell AP (2001) Sporadic silent thyroiditis, postpartum thyroiditis, and subacute thyroiditis. In: Braverman LE, Utiger RD (eds) Werner & Ingbar s the Thyroid, ed 8, Philadelphia, Lippincott, Kimura M, Amino N, Takeda K, Miyai K (1989) Subacute thyroiditis associated with systemic multi-organ disorders. Endocrinol Jpn 36: Potasman I, Bassan H, Rosenfeld T (1983) Alkaline phosphatase activity in subacute thyroiditis and hyperthyroidism. Isr J Med Sci 19: